Skip to content

Instantly share code, notes, and snippets.

@vanzod
Created April 20, 2022 15:31
Show Gist options
  • Star 0 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save vanzod/4e4d51d76c0acc1081256174a482bd4a to your computer and use it in GitHub Desktop.
Save vanzod/4e4d51d76c0acc1081256174a482bd4a to your computer and use it in GitHub Desktop.
UCX 1.11.2 debug log for failing osu_scatter
This file has been truncated, but you can view the full file.
[1650465527.778184] [ndv4:13452:0] debug.c:1198 UCX DEBUG using signal stack 0x2ae65c744000 size 141824
[1650465527.778267] [ndv4:13452:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465527.778291] [ndv4:13452:0] init.c:113 UCX DEBUG /cvmfs/ndv4.azure/CentOS-HPC/7.9.2021052401/EasyBuild/software/UCX/1.11.2-GCCcore-11.2.0/lib64/libucs.so.0 loaded at 0x2ae65c598000
[1650465527.778320] [ndv4:13452:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650465527.778329] [ndv4:13452:0] module.c:69 UCX DEBUG ucs library path: /cvmfs/ndv4.azure/CentOS-HPC/7.9.2021052401/EasyBuild/software/UCX/1.11.2-GCCcore-11.2.0/lib64/libucs.so.0
[1650465527.778336] [ndv4:13452:0] module.c:253 UCX DEBUG loading modules for ucs
[1650465527.781699] [ndv4:13452:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465527.781723] [ndv4:13452:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650465527.781758] [ndv4:13452:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650465527.781761] [ndv4:13452:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650465527.781768] [ndv4:13452:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650465527.781775] [ndv4:13452:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650465527.781778] [ndv4:13452:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650465527.781782] [ndv4:13452:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650465527.781785] [ndv4:13452:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650465527.781787] [ndv4:13452:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650465527.781790] [ndv4:13452:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650465527.781792] [ndv4:13452:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650465527.781800] [ndv4:13452:0] module.c:253 UCX DEBUG loading modules for uct
[1650465527.783148] [ndv4:13452:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650465527.784670] [ndv4:13452:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650465527.784697] [ndv4:13452:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650465527.784709] [ndv4:13452:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650465527.784721] [ndv4:13452:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650465527.784732] [ndv4:13452:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650465527.784743] [ndv4:13452:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650465527.784755] [ndv4:13452:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650465527.785278] [ndv4:13452:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650465527.790275] [ndv4:15226:0] debug.c:1198 UCX DEBUG using signal stack 0x2b17886a0000 size 141824
[1650465527.790348] [ndv4:15226:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465527.790366] [ndv4:15226:0] init.c:113 UCX DEBUG /cvmfs/ndv4.azure/CentOS-HPC/7.9.2021052401/EasyBuild/software/UCX/1.11.2-GCCcore-11.2.0/lib64/libucs.so.0 loaded at 0x2b17884f4000
[1650465527.790390] [ndv4:15226:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650465527.790399] [ndv4:15226:0] module.c:69 UCX DEBUG ucs library path: /cvmfs/ndv4.azure/CentOS-HPC/7.9.2021052401/EasyBuild/software/UCX/1.11.2-GCCcore-11.2.0/lib64/libucs.so.0
[1650465527.790405] [ndv4:15226:0] module.c:253 UCX DEBUG loading modules for ucs
[1650465527.793470] [ndv4:15226:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465527.793491] [ndv4:15226:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650465527.793524] [ndv4:15226:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650465527.793527] [ndv4:15226:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650465527.793533] [ndv4:15226:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650465527.793540] [ndv4:15226:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650465527.793543] [ndv4:15226:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650465527.793546] [ndv4:15226:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650465527.793549] [ndv4:15226:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650465527.793551] [ndv4:15226:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650465527.793553] [ndv4:15226:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650465527.793556] [ndv4:15226:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650465527.793563] [ndv4:15226:0] module.c:253 UCX DEBUG loading modules for uct
[1650465527.794454] [ndv4:15226:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650465527.795265] [ndv4:15226:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650465527.795282] [ndv4:15226:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650465527.795292] [ndv4:15226:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650465527.795303] [ndv4:15226:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650465527.795313] [ndv4:15226:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650465527.795323] [ndv4:15226:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650465527.795333] [ndv4:15226:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650465527.795875] [ndv4:13452:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650465527.795972] [ndv4:15226:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650465527.796347] [ndv4:13452:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650465527.803344] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x16121e0 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650465527.803464] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650465527.803472] [ndv4:13452:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650465527.804006] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465527.804017] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465527.804041] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465527.804100] [ndv4:13452:0] module.c:253 UCX DEBUG loading modules for ucm
[1650465527.804122] [ndv4:13452:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650465527.805517] [ndv4:13452:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465527.805526] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465527.806360] [ndv4:15226:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650465527.805837] [ndv4:13452:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465527.805895] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465527.805960] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465527.806962] [ndv4:15226:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650465527.822454] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0xf571e0 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650465527.822559] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650465527.822567] [ndv4:15226:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650465527.823160] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465527.823196] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465527.823215] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465527.823276] [ndv4:15226:0] module.c:253 UCX DEBUG loading modules for ucm
[1650465527.823295] [ndv4:15226:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650465527.825920] [ndv4:15226:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465527.825929] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465527.826112] [ndv4:15226:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465527.826137] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465527.826153] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465527.830454] [ndv4:13452:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650465527.830463] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465527.839965] [ndv4:15226:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650465527.839974] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465527.841862] [ndv4:15226:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465527.841868] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465527.843785] [ndv4:13452:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465527.843794] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465527.847356] [ndv4:13452:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465527.847362] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465527.849888] [ndv4:13452:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465527.849894] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465527.853278] [ndv4:15226:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465527.853286] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465527.856061] [ndv4:15226:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465527.856068] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465527.857394] [ndv4:15226:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465527.857400] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465527.857756] [ndv4:15226:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465527.862370] [ndv4:13452:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465527.862377] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465527.862770] [ndv4:13452:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465527.880765] [ndv4:13452:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650465527.881428] [ndv4:13452:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650465527.889613] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x160af50 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650465527.889670] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650465527.889673] [ndv4:13452:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650465527.890179] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465527.890279] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465527.890284] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465527.890327] [ndv4:13452:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650465527.891910] [ndv4:15226:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650465527.892244] [ndv4:13452:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465527.892251] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465527.892670] [ndv4:15226:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650465527.892699] [ndv4:13452:0] ib_md.c:1604 UCX DEBUG mlx5_ib1: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465527.892733] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465527.892739] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465527.894938] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0xf4ff50 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650465527.894975] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650465527.894979] [ndv4:15226:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650465527.895171] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465527.895179] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465527.895185] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465527.895227] [ndv4:15226:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650465527.897045] [ndv4:15226:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465527.897052] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465527.897247] [ndv4:15226:0] ib_md.c:1604 UCX DEBUG mlx5_ib1: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465527.897265] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465527.897270] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465527.899195] [ndv4:15226:0] topo.c:99 UCX DEBUG bus id 0x102000000 doesn't exist. sys_dev = 1
[1650465527.899201] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465527.903846] [ndv4:13452:0] topo.c:99 UCX DEBUG bus id 0x102000000 doesn't exist. sys_dev = 1
[1650465527.903854] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465527.907228] [ndv4:15226:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465527.907235] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465527.915451] [ndv4:13452:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465527.915460] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465527.916069] [ndv4:13452:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465527.916073] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465527.918893] [ndv4:15226:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465527.918902] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465527.924006] [ndv4:13452:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465527.924014] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465527.927121] [ndv4:15226:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465527.927129] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465527.928507] [ndv4:13452:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465527.928514] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465527.928922] [ndv4:13452:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465527.934184] [ndv4:15226:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465527.934191] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465527.934431] [ndv4:15226:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465527.940794] [ndv4:13452:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib2 vendor_id: 0x15b3 device_id: 4124
[1650465527.941792] [ndv4:13452:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib2: disable ODP because it's not supported for DevX QP
[1650465527.942970] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x1612ce0 [id=61 ref 1] uct_ib_async_event_handler() to hash
[1650465527.943008] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 61 events 0x1 mode thread_spinlock
[1650465527.943011] [ndv4:13452:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib2' (InfiniBand channel adapter) with 1 ports
[1650465527.943378] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465527.943454] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465527.943460] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465527.943500] [ndv4:13452:0] ib_md.c:1319 UCX DEBUG mlx5_ib2: using registration cache
[1650465527.947804] [ndv4:13452:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465527.947811] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465527.949381] [ndv4:15226:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib2 vendor_id: 0x15b3 device_id: 4124
[1650465527.948135] [ndv4:13452:0] ib_md.c:1604 UCX DEBUG mlx5_ib2: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465527.948376] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465527.948450] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465527.949948] [ndv4:15226:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib2: disable ODP because it's not supported for DevX QP
[1650465527.954402] [ndv4:13452:0] topo.c:99 UCX DEBUG bus id 0x103000000 doesn't exist. sys_dev = 2
[1650465527.954409] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465527.963970] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0xf57ce0 [id=61 ref 1] uct_ib_async_event_handler() to hash
[1650465527.964008] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 61 events 0x1 mode thread_spinlock
[1650465527.964012] [ndv4:15226:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib2' (InfiniBand channel adapter) with 1 ports
[1650465527.964494] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465527.964655] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465527.964661] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465527.964703] [ndv4:15226:0] ib_md.c:1319 UCX DEBUG mlx5_ib2: using registration cache
[1650465527.966274] [ndv4:15226:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465527.966281] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465527.966721] [ndv4:15226:0] ib_md.c:1604 UCX DEBUG mlx5_ib2: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465527.966821] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465527.966863] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465527.971194] [ndv4:13452:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465527.971202] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465527.976002] [ndv4:15226:0] topo.c:99 UCX DEBUG bus id 0x103000000 doesn't exist. sys_dev = 2
[1650465527.976010] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465527.980193] [ndv4:13452:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465527.980210] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465527.989042] [ndv4:15226:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465527.989049] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465527.991972] [ndv4:13452:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465527.991980] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465527.993753] [ndv4:15226:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465527.993759] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465528.004241] [ndv4:15226:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465528.004249] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465528.006526] [ndv4:13452:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465528.006534] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465528.007634] [ndv4:13452:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465528.012093] [ndv4:15226:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465528.012102] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465528.013110] [ndv4:15226:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465528.024456] [ndv4:13452:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib3 vendor_id: 0x15b3 device_id: 4124
[1650465528.025074] [ndv4:13452:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib3: disable ODP because it's not supported for DevX QP
[1650465528.026862] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x160a2b0 [id=63 ref 1] uct_ib_async_event_handler() to hash
[1650465528.026898] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 63 events 0x1 mode thread_spinlock
[1650465528.026902] [ndv4:13452:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib3' (InfiniBand channel adapter) with 1 ports
[1650465528.027518] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.027571] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.027695] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.027757] [ndv4:13452:0] ib_md.c:1319 UCX DEBUG mlx5_ib3: using registration cache
[1650465528.028916] [ndv4:13452:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.028923] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.029212] [ndv4:13452:0] ib_md.c:1604 UCX DEBUG mlx5_ib3: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.029311] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.029425] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.031079] [ndv4:15226:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib3 vendor_id: 0x15b3 device_id: 4124
[1650465528.031469] [ndv4:15226:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib3: disable ODP because it's not supported for DevX QP
[1650465528.033716] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0xf4f2b0 [id=63 ref 1] uct_ib_async_event_handler() to hash
[1650465528.033749] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 63 events 0x1 mode thread_spinlock
[1650465528.033752] [ndv4:15226:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib3' (InfiniBand channel adapter) with 1 ports
[1650465528.034401] [ndv4:13452:0] topo.c:99 UCX DEBUG bus id 0x104000000 doesn't exist. sys_dev = 3
[1650465528.034408] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465528.034055] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.034250] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.034255] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.034297] [ndv4:15226:0] ib_md.c:1319 UCX DEBUG mlx5_ib3: using registration cache
[1650465528.042554] [ndv4:15226:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.042562] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.042903] [ndv4:15226:0] ib_md.c:1604 UCX DEBUG mlx5_ib3: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.043084] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.043113] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.044906] [ndv4:13452:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465528.044913] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465528.048758] [ndv4:13452:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465528.048765] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465528.050701] [ndv4:15226:0] topo.c:99 UCX DEBUG bus id 0x104000000 doesn't exist. sys_dev = 3
[1650465528.050708] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465528.052220] [ndv4:13452:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465528.052226] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465528.054683] [ndv4:15226:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465528.054689] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465528.068885] [ndv4:13452:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465528.068892] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465528.070232] [ndv4:13452:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465528.077990] [ndv4:15226:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465528.077999] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465528.092454] [ndv4:15226:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465528.092462] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465528.096412] [ndv4:15226:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465528.096417] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465528.097500] [ndv4:15226:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465528.106003] [ndv4:13452:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib4 vendor_id: 0x15b3 device_id: 4124
[1650465528.106428] [ndv4:13452:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib4: disable ODP because it's not supported for DevX QP
[1650465528.117245] [ndv4:15226:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib4 vendor_id: 0x15b3 device_id: 4124
[1650465528.117178] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x1cf3bf0 [id=65 ref 1] uct_ib_async_event_handler() to hash
[1650465528.117202] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 65 events 0x1 mode thread_spinlock
[1650465528.117205] [ndv4:13452:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib4' (InfiniBand channel adapter) with 1 ports
[1650465528.117810] [ndv4:15226:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib4: disable ODP because it's not supported for DevX QP
[1650465528.117769] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465528.117810] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465528.117816] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.117861] [ndv4:13452:0] ib_md.c:1319 UCX DEBUG mlx5_ib4: using registration cache
[1650465528.119305] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x1638bf0 [id=65 ref 1] uct_ib_async_event_handler() to hash
[1650465528.119334] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 65 events 0x1 mode thread_spinlock
[1650465528.119337] [ndv4:15226:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib4' (InfiniBand channel adapter) with 1 ports
[1650465528.119767] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465528.119835] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465528.119841] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.119886] [ndv4:15226:0] ib_md.c:1319 UCX DEBUG mlx5_ib4: using registration cache
[1650465528.121357] [ndv4:15226:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.121364] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.121674] [ndv4:15226:0] ib_md.c:1604 UCX DEBUG mlx5_ib4: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.122061] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465528.122115] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465528.129258] [ndv4:13452:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.129266] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.129433] [ndv4:13452:0] ib_md.c:1604 UCX DEBUG mlx5_ib4: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.129691] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465528.129835] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465528.133507] [ndv4:13452:0] topo.c:99 UCX DEBUG bus id 0x105000000 doesn't exist. sys_dev = 4
[1650465528.133513] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465528.136468] [ndv4:13452:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465528.136474] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465528.140939] [ndv4:15226:0] topo.c:99 UCX DEBUG bus id 0x105000000 doesn't exist. sys_dev = 4
[1650465528.140948] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465528.142095] [ndv4:13452:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465528.142102] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465528.151860] [ndv4:13452:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465528.151868] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465528.152896] [ndv4:15226:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465528.152903] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465528.165486] [ndv4:13452:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465528.165494] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465528.166664] [ndv4:13452:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465528.165619] [ndv4:15226:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465528.165625] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465528.177971] [ndv4:15226:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465528.177978] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465528.180022] [ndv4:13452:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib5 vendor_id: 0x15b3 device_id: 4124
[1650465528.180704] [ndv4:13452:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib5: disable ODP because it's not supported for DevX QP
[1650465528.194999] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x160e710 [id=67 ref 1] uct_ib_async_event_handler() to hash
[1650465528.195026] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 67 events 0x1 mode thread_spinlock
[1650465528.195030] [ndv4:13452:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib5' (InfiniBand channel adapter) with 1 ports
[1650465528.196012] [ndv4:15226:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465528.196020] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465528.195508] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465528.195622] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465528.195627] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.195676] [ndv4:13452:0] ib_md.c:1319 UCX DEBUG mlx5_ib5: using registration cache
[1650465528.197183] [ndv4:13452:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.197190] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.197679] [ndv4:15226:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465528.197436] [ndv4:13452:0] ib_md.c:1604 UCX DEBUG mlx5_ib5: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.197804] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465528.197848] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465528.208895] [ndv4:15226:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib5 vendor_id: 0x15b3 device_id: 4124
[1650465528.209469] [ndv4:15226:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib5: disable ODP because it's not supported for DevX QP
[1650465528.209863] [ndv4:13452:0] topo.c:99 UCX DEBUG bus id 0x106000000 doesn't exist. sys_dev = 5
[1650465528.209871] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465528.213555] [ndv4:13452:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465528.213561] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465528.216860] [ndv4:13452:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465528.216867] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465528.221055] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0xf53710 [id=67 ref 1] uct_ib_async_event_handler() to hash
[1650465528.221087] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 67 events 0x1 mode thread_spinlock
[1650465528.221090] [ndv4:15226:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib5' (InfiniBand channel adapter) with 1 ports
[1650465528.221672] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465528.221813] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465528.221819] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.221869] [ndv4:15226:0] ib_md.c:1319 UCX DEBUG mlx5_ib5: using registration cache
[1650465528.225923] [ndv4:13452:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465528.225929] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465528.228895] [ndv4:15226:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.228903] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.229121] [ndv4:15226:0] ib_md.c:1604 UCX DEBUG mlx5_ib5: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.229259] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465528.229379] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465528.229785] [ndv4:13452:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465528.229791] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465528.231186] [ndv4:13452:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465528.243317] [ndv4:13452:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib6 vendor_id: 0x15b3 device_id: 4124
[1650465528.243782] [ndv4:13452:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib6: disable ODP because it's not supported for DevX QP
[1650465528.250538] [ndv4:15226:0] topo.c:99 UCX DEBUG bus id 0x106000000 doesn't exist. sys_dev = 5
[1650465528.250546] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465528.254330] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x160e560 [id=69 ref 1] uct_ib_async_event_handler() to hash
[1650465528.254360] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 69 events 0x1 mode thread_spinlock
[1650465528.254364] [ndv4:13452:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib6' (InfiniBand channel adapter) with 1 ports
[1650465528.254558] [ndv4:15226:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465528.254564] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465528.254880] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465528.255012] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465528.255017] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.255059] [ndv4:13452:0] ib_md.c:1319 UCX DEBUG mlx5_ib6: using registration cache
[1650465528.259169] [ndv4:15226:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465528.259175] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465528.261384] [ndv4:13452:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.261391] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.261762] [ndv4:13452:0] ib_md.c:1604 UCX DEBUG mlx5_ib6: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.262350] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465528.262423] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465528.274012] [ndv4:15226:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465528.274019] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465528.277020] [ndv4:13452:0] topo.c:99 UCX DEBUG bus id 0x107000000 doesn't exist. sys_dev = 6
[1650465528.277028] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465528.278621] [ndv4:15226:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465528.278628] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465528.279396] [ndv4:15226:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465528.282646] [ndv4:13452:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465528.282652] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465528.293469] [ndv4:15226:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib6 vendor_id: 0x15b3 device_id: 4124
[1650465528.293989] [ndv4:15226:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib6: disable ODP because it's not supported for DevX QP
[1650465528.296875] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0xf53560 [id=69 ref 1] uct_ib_async_event_handler() to hash
[1650465528.296911] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 69 events 0x1 mode thread_spinlock
[1650465528.296915] [ndv4:15226:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib6' (InfiniBand channel adapter) with 1 ports
[1650465528.297369] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465528.297477] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465528.297482] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.297525] [ndv4:15226:0] ib_md.c:1319 UCX DEBUG mlx5_ib6: using registration cache
[1650465528.299094] [ndv4:15226:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.299101] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.299298] [ndv4:15226:0] ib_md.c:1604 UCX DEBUG mlx5_ib6: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.299830] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465528.300061] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465528.302285] [ndv4:13452:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465528.302292] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465528.304780] [ndv4:14205:0] debug.c:1198 UCX DEBUG using signal stack 0x2afda2764000 size 141824
[1650465528.305112] [ndv4:14205:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465528.305135] [ndv4:14205:0] init.c:113 UCX DEBUG /cvmfs/ndv4.azure/CentOS-HPC/7.9.2021052401/EasyBuild/software/UCX/1.11.2-GCCcore-11.2.0/lib64/libucs.so.0 loaded at 0x2afda25d2000
[1650465528.305502] [ndv4:14205:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650465528.305511] [ndv4:14205:0] module.c:69 UCX DEBUG ucs library path: /cvmfs/ndv4.azure/CentOS-HPC/7.9.2021052401/EasyBuild/software/UCX/1.11.2-GCCcore-11.2.0/lib64/libucs.so.0
[1650465528.305518] [ndv4:14205:0] module.c:253 UCX DEBUG loading modules for ucs
[1650465528.312444] [ndv4:15226:0] topo.c:99 UCX DEBUG bus id 0x107000000 doesn't exist. sys_dev = 6
[1650465528.312451] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465528.312220] [ndv4:14205:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465528.312241] [ndv4:14205:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650465528.312276] [ndv4:14205:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650465528.312279] [ndv4:14205:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650465528.312286] [ndv4:14205:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650465528.312293] [ndv4:14205:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650465528.312296] [ndv4:14205:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650465528.312300] [ndv4:14205:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650465528.312302] [ndv4:14205:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650465528.312305] [ndv4:14205:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650465528.312307] [ndv4:14205:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650465528.312309] [ndv4:14205:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650465528.312318] [ndv4:14205:0] module.c:253 UCX DEBUG loading modules for uct
[1650465528.316147] [ndv4:15226:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465528.316152] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465528.318759] [ndv4:14205:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650465528.320312] [ndv4:13452:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465528.320319] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465528.321055] [ndv4:14205:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650465528.321072] [ndv4:14205:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650465528.321084] [ndv4:14205:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650465528.321095] [ndv4:14205:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650465528.321106] [ndv4:14205:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650465528.321117] [ndv4:14205:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650465528.321127] [ndv4:14205:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650465528.323881] [ndv4:13452:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465528.323887] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465528.323841] [ndv4:14205:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650465528.325155] [ndv4:13452:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465528.326841] [ndv4:15226:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465528.326850] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465528.330030] [ndv4:15226:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465528.330037] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465528.341640] [ndv4:13452:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib7 vendor_id: 0x15b3 device_id: 4124
[1650465528.342075] [ndv4:13452:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib7: disable ODP because it's not supported for DevX QP
[1650465528.353919] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x1cf4940 [id=71 ref 1] uct_ib_async_event_handler() to hash
[1650465528.353953] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 71 events 0x1 mode thread_spinlock
[1650465528.353956] [ndv4:13452:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib7' (InfiniBand channel adapter) with 1 ports
[1650465528.354268] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465528.354539] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465528.354545] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.354652] [ndv4:13452:0] ib_md.c:1319 UCX DEBUG mlx5_ib7: using registration cache
[1650465528.356097] [ndv4:13452:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.356104] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.356447] [ndv4:13452:0] ib_md.c:1604 UCX DEBUG mlx5_ib7: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.356770] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465528.357176] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465528.361357] [ndv4:14205:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650465528.362047] [ndv4:14205:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650465528.369768] [ndv4:13452:0] topo.c:99 UCX DEBUG bus id 0x108000000 doesn't exist. sys_dev = 7
[1650465528.369777] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465528.374665] [ndv4:15226:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465528.374674] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465528.375000] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x1990360 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650465528.375099] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650465528.375106] [ndv4:14205:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650465528.376036] [ndv4:15226:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465528.381220] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.381443] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.381467] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.381529] [ndv4:14205:0] module.c:253 UCX DEBUG loading modules for ucm
[1650465528.381552] [ndv4:14205:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650465528.389779] [ndv4:13452:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465528.389787] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465528.393364] [ndv4:14205:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.393379] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.393640] [ndv4:14205:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.393872] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.393977] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.397564] [ndv4:15226:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib7 vendor_id: 0x15b3 device_id: 4124
[1650465528.398504] [ndv4:15226:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib7: disable ODP because it's not supported for DevX QP
[1650465528.398898] [ndv4:13452:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465528.398905] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465528.399995] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x1639940 [id=71 ref 1] uct_ib_async_event_handler() to hash
[1650465528.400030] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 71 events 0x1 mode thread_spinlock
[1650465528.400035] [ndv4:15226:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib7' (InfiniBand channel adapter) with 1 ports
[1650465528.400648] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465528.400727] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465528.400735] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.400808] [ndv4:15226:0] ib_md.c:1319 UCX DEBUG mlx5_ib7: using registration cache
[1650465528.402497] [ndv4:13452:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465528.402503] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465528.404765] [ndv4:13452:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465528.404771] [ndv4:13452:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465528.405044] [ndv4:13452:0] ucp_context.c:1556 UCX DEBUG created ucp context 0x15f7c90 0x15f7c90 [13 mds 47 tls] features 0x1 tl bitmap 0x7fffffffffff 0x0
[1650465528.405640] [ndv4:15226:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.405649] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.405919] [ndv4:15226:0] ib_md.c:1604 UCX DEBUG mlx5_ib7: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.406203] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465528.406331] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465528.414198] [ndv4:14205:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650465528.414207] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465528.426276] [ndv4:14205:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465528.426283] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465528.431665] [ndv4:14205:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465528.431671] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465528.433699] [ndv4:15226:0] topo.c:99 UCX DEBUG bus id 0x108000000 doesn't exist. sys_dev = 7
[1650465528.433707] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465528.435816] [ndv4:14205:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465528.435822] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465528.437721] [ndv4:15226:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465528.437728] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465528.439106] [ndv4:14949:0] debug.c:1198 UCX DEBUG using signal stack 0x2ad4b6430000 size 141824
[1650465528.439496] [ndv4:14949:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465528.439516] [ndv4:14949:0] init.c:113 UCX DEBUG /cvmfs/ndv4.azure/CentOS-HPC/7.9.2021052401/EasyBuild/software/UCX/1.11.2-GCCcore-11.2.0/lib64/libucs.so.0 loaded at 0x2ad4b629e000
[1650465528.439571] [ndv4:14949:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650465528.439678] [ndv4:14949:0] module.c:69 UCX DEBUG ucs library path: /cvmfs/ndv4.azure/CentOS-HPC/7.9.2021052401/EasyBuild/software/UCX/1.11.2-GCCcore-11.2.0/lib64/libucs.so.0
[1650465528.439685] [ndv4:14949:0] module.c:253 UCX DEBUG loading modules for ucs
[1650465528.444391] [ndv4:14205:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465528.444397] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465528.446069] [ndv4:14949:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465528.446092] [ndv4:14949:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650465528.446127] [ndv4:14949:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650465528.446130] [ndv4:14949:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650465528.446136] [ndv4:14949:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650465528.446143] [ndv4:14949:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650465528.446146] [ndv4:14949:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650465528.446150] [ndv4:14949:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650465528.446153] [ndv4:14949:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650465528.446155] [ndv4:14949:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650465528.446157] [ndv4:14949:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650465528.446159] [ndv4:14949:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650465528.446167] [ndv4:14949:0] module.c:253 UCX DEBUG loading modules for uct
[1650465528.446097] [ndv4:14205:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465528.446877] [ndv4:15226:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465528.446885] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465528.452836] [ndv4:14949:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650465528.454979] [ndv4:14949:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650465528.454996] [ndv4:14949:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650465528.455007] [ndv4:14949:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650465528.455018] [ndv4:14949:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650465528.455028] [ndv4:14949:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650465528.455038] [ndv4:14949:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650465528.455048] [ndv4:14949:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650465528.457279] [ndv4:14949:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650465528.459481] [ndv4:15226:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465528.459490] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465528.465834] [ndv4:14205:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650465528.466176] [ndv4:14205:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650465528.467136] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x1986db0 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650465528.467168] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650465528.467172] [ndv4:14205:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650465528.472462] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.472725] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.472732] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.472775] [ndv4:14205:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650465528.474254] [ndv4:14205:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.474262] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.474496] [ndv4:14205:0] ib_md.c:1604 UCX DEBUG mlx5_ib1: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.477047] [ndv4:14949:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650465528.477364] [ndv4:15226:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465528.477372] [ndv4:15226:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465528.477745] [ndv4:15226:0] ucp_context.c:1556 UCX DEBUG created ucp context 0xf3cc90 0xf3cc90 [13 mds 47 tls] features 0x1 tl bitmap 0x7fffffffffff 0x0
[1650465528.477561] [ndv4:14949:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650465528.481298] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.481390] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.484564] [ndv4:13452:0] mm_posix.c:518 UCX DEBUG allocated posix shared memory at 0x2ae65fa2a000 length 12288
[1650465528.484702] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool mm_recv_desc: align 64, maxelems 4294967295, elemsize 8368
[1650465528.483866] [ndv4:14205:0] topo.c:99 UCX DEBUG bus id 0x102000000 doesn't exist. sys_dev = 1
[1650465528.483873] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465528.486287] [ndv4:13443:0] debug.c:1198 UCX DEBUG using signal stack 0x2b750459c000 size 141824
[1650465528.486516] [ndv4:13443:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465528.486539] [ndv4:13443:0] init.c:113 UCX DEBUG /cvmfs/ndv4.azure/CentOS-HPC/7.9.2021052401/EasyBuild/software/UCX/1.11.2-GCCcore-11.2.0/lib64/libucs.so.0 loaded at 0x2b75043f0000
[1650465528.486748] [ndv4:13443:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650465528.486760] [ndv4:13443:0] module.c:69 UCX DEBUG ucs library path: /cvmfs/ndv4.azure/CentOS-HPC/7.9.2021052401/EasyBuild/software/UCX/1.11.2-GCCcore-11.2.0/lib64/libucs.so.0
[1650465528.486766] [ndv4:13443:0] module.c:253 UCX DEBUG loading modules for ucs
[1650465528.486139] [ndv4:13452:0] mm_posix.c:305 UCX DEBUG shared memory mmap(addr=(nil), length=6291456, flags= HUGETLB, fd=76) failed: Invalid argument
[1650465528.486148] [ndv4:13452:0] mm_posix.c:518 UCX DEBUG allocated posix shared memory at 0x2ae65fa2d000 length 4296704
[1650465528.486153] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool mm_recv_desc: allocated chunk 0x2ae65fa2d018 of 4296680 bytes with 512 elements
[1650465528.486415] [ndv4:13452:0] mm_iface.c:600 UCX DEBUG created mm iface 0x1d74d10 FIFO id 0x400000002ed74ed9 va 0x2ae65fa2a000 size 12288 (128 x 64 elems)
[1650465528.486675] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[0]=0x1d74d10 using posix/memory on worker 0x264e8d0
[1650465528.486701] [ndv4:13452:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 8447 bytes with hugetlb
[1650465528.486753] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool mm_recv_desc: align 64, maxelems 4294967295, elemsize 8368
[1650465528.488010] [ndv4:13452:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 4292720 bytes with hugetlb
[1650465528.488025] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool mm_recv_desc: allocated chunk 0x2ae664000018 of 4296680 bytes with 512 elements
[1650465528.488723] [ndv4:13452:0] mm_iface.c:600 UCX DEBUG created mm iface 0x1d752e0 FIFO id 0x64003b va 0x2ae65fe46000 size 12288 (128 x 64 elems)
[1650465528.488734] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[1]=0x1d752e0 using sysv/memory on worker 0x264e8d0
[1650465528.488747] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool self_msg_desc: align 64, maxelems 4294967295, elemsize 8200
[1650465528.488751] [ndv4:13452:0] self.c:220 UCX DEBUG created self iface id 0x8c9f25a5686245c7 send_size 8192
[1650465528.488757] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[2]=0x1d88ec0 using self/memory0 on worker 0x264e8d0
[1650465528.488781] [ndv4:13452:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465528.488786] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465528.488789] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465528.490728] [ndv4:14205:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465528.490735] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465528.491928] [ndv4:13443:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465528.491949] [ndv4:13443:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650465528.491986] [ndv4:13443:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650465528.491989] [ndv4:13443:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650465528.491997] [ndv4:13443:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650465528.492004] [ndv4:13443:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650465528.492007] [ndv4:13443:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650465528.492012] [ndv4:13443:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650465528.492015] [ndv4:13443:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650465528.492017] [ndv4:13443:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650465528.492019] [ndv4:13443:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650465528.492022] [ndv4:13443:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650465528.492030] [ndv4:13443:0] module.c:253 UCX DEBUG loading modules for uct
[1650465528.493151] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x1d8b000 [id=78 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465528.493194] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 78 events 0x5 mode thread_spinlock
[1650465528.493213] [ndv4:13452:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x1d89820: listening for connections (fd=78) on 10.5.0.5:60081
[1650465528.493557] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[3]=0x1d89820 using tcp/eth0 on worker 0x264e8d0
[1650465528.493669] [ndv4:13452:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465528.493674] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465528.493677] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465528.493723] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x1cd75c0 [id=80 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465528.493746] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 80 events 0x5 mode thread_spinlock
[1650465528.493753] [ndv4:13452:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x1d89ea0: listening for connections (fd=80) on 127.0.0.1:60264
[1650465528.493885] [ndv4:13452:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465528.493894] [ndv4:13452:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465528.494152] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[4]=0x1d89ea0 using tcp/lo on worker 0x264e8d0
[1650465528.494170] [ndv4:13452:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465528.494173] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465528.494175] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465528.494469] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x1ce5960 [id=82 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465528.494490] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 82 events 0x5 mode thread_spinlock
[1650465528.494494] [ndv4:13452:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x1d71530: listening for connections (fd=82) on 172.16.1.242:41147
[1650465528.494890] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[5]=0x1d71530 using tcp/ib0 on worker 0x264e8d0
[1650465528.495093] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.495389] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.495771] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.495889] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.496007] [ndv4:14949:0] async.c:228 UCX DEBUG added async handler 0x972270 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650465528.496165] [ndv4:14949:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650465528.496178] [ndv4:14949:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650465528.496664] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.496836] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.496870] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.496951] [ndv4:14949:0] module.c:253 UCX DEBUG loading modules for ucm
[1650465528.496977] [ndv4:14949:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650465528.496710] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465528.498148] [ndv4:14949:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.498159] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.498633] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465528.498672] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465528.498675] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465528.498688] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465528.499206] [ndv4:14205:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465528.499215] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465528.499704] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465528.499713] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465528.501039] [ndv4:13443:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650465528.500640] [ndv4:14949:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.500812] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.500887] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.503089] [ndv4:13443:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650465528.503105] [ndv4:13443:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650465528.503117] [ndv4:13443:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650465528.503127] [ndv4:13443:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650465528.503137] [ndv4:13443:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650465528.503147] [ndv4:13443:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650465528.503158] [ndv4:13443:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650465528.503934] [ndv4:13452:0] ib_iface.c:994 UCX DEBUG iface=0x1d7f670: created RC QP 0xdb0a on mlx5_ib0:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465528.507210] [ndv4:13443:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650465528.509842] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[6]=0x1d7f670 using rc_verbs/mlx5_ib0:1 on worker 0x264e8d0
[1650465528.514474] [ndv4:14205:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465528.514483] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465528.518482] [ndv4:13443:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650465528.517892] [ndv4:14949:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650465528.517902] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465528.518999] [ndv4:13443:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650465528.519299] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.519463] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.522169] [ndv4:14949:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465528.522176] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465528.524976] [ndv4:13443:0] async.c:228 UCX DEBUG added async handler 0x1d3b080 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650465528.525077] [ndv4:13443:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650465528.525085] [ndv4:13443:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650465528.525762] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.525999] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.526023] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.526082] [ndv4:13443:0] module.c:253 UCX DEBUG loading modules for ucm
[1650465528.526104] [ndv4:13443:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650465528.525830] [ndv4:14205:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465528.525841] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465528.526052] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.526276] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.527014] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465528.527420] [ndv4:13452:0] ib_device.c:1394 UCX DEBUG max IB CQE size is 128
[1650465528.527692] [ndv4:14205:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465528.527695] [ndv4:13443:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.527705] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.527957] [ndv4:13443:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.528046] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.528120] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.528639] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465528.528657] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465528.528661] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465528.528718] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465528.529232] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x2718010 of 8176 bytes with 127 elements
[1650465528.529533] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465528.529566] [ndv4:13452:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib0:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465528.529668] [ndv4:13452:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650465528.529674] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465528.533146] [ndv4:13443:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650465528.533154] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465528.537623] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x1ce4c70 [id=85 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465528.537664] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 85 events 0x1 mode thread_spinlock
[1650465528.537690] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[7]=0x1d8d0c0 using rc_mlx5/mlx5_ib0:1 on worker 0x264e8d0
[1650465528.538342] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.538484] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.538725] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.539182] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.543686] [ndv4:14949:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465528.543705] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465528.544099] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465528.544799] [ndv4:14205:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib2 vendor_id: 0x15b3 device_id: 4124
[1650465528.545210] [ndv4:14205:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib2: disable ODP because it's not supported for DevX QP
[1650465528.545956] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x1986e00 [id=61 ref 1] uct_ib_async_event_handler() to hash
[1650465528.545992] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 61 events 0x1 mode thread_spinlock
[1650465528.545995] [ndv4:14205:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib2' (InfiniBand channel adapter) with 1 ports
[1650465528.546404] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.546817] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.546829] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.546914] [ndv4:14205:0] ib_md.c:1319 UCX DEBUG mlx5_ib2: using registration cache
[1650465528.547812] [ndv4:14205:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.547819] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.548071] [ndv4:14205:0] ib_md.c:1604 UCX DEBUG mlx5_ib2: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.548415] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.548799] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.550718] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465528.550729] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465528.550732] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465528.550747] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465528.551230] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465528.551242] [ndv4:13452:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib0:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465528.551273] [ndv4:13452:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650465528.551277] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465528.551289] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x1ce3b40 [id=87 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465528.551313] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 87 events 0x1 mode thread_spinlock
[1650465528.553073] [ndv4:13452:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465528.553481] [ndv4:14949:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465528.553489] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465528.555115] [ndv4:13443:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465528.555123] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465528.556885] [ndv4:14949:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465528.556892] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465528.559726] [ndv4:14949:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465528.563437] [ndv4:13452:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x2982010: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xdbb1
[1650465528.564490] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.564725] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.564751] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2ae65fe4b008 of 151544 bytes with 1052 elements
[1650465528.566397] [ndv4:15226:0] mm_posix.c:518 UCX DEBUG allocated posix shared memory at 0x2b178b99c000 length 12288
[1650465528.566463] [ndv4:14205:0] topo.c:99 UCX DEBUG bus id 0x103000000 doesn't exist. sys_dev = 2
[1650465528.566475] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465528.566800] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool mm_recv_desc: align 64, maxelems 4294967295, elemsize 8368
[1650465528.568358] [ndv4:15226:0] mm_posix.c:305 UCX DEBUG shared memory mmap(addr=(nil), length=6291456, flags= HUGETLB, fd=76) failed: Invalid argument
[1650465528.568376] [ndv4:15226:0] mm_posix.c:518 UCX DEBUG allocated posix shared memory at 0x2b178b99f000 length 4296704
[1650465528.568385] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool mm_recv_desc: allocated chunk 0x2b178b99f018 of 4296680 bytes with 512 elements
[1650465528.568860] [ndv4:13452:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae664600000..0x2ae666c00000 on mlx5_ib0 lkey 0x80400 rkey 0x80400 access 0xf flags 0x3e4
[1650465528.568881] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2ae664600018 of 39845864 bytes with 4752 elements
[1650465528.569015] [ndv4:13452:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x2982010
[1650465528.569051] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[8]=0x2982010 using dc_mlx5/mlx5_ib0:1 on worker 0x264e8d0
[1650465528.569523] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.569669] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.569798] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.569940] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.570427] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465528.571685] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465528.569774] [ndv4:15226:0] mm_iface.c:600 UCX DEBUG created mm iface 0x16b9d10 FIFO id 0x400000000f80c7c5 va 0x2b178b99c000 size 12288 (128 x 64 elems)
[1650465528.569895] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[0]=0x16b9d10 using posix/memory on worker 0x1f938d0
[1650465528.569931] [ndv4:15226:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 8447 bytes with hugetlb
[1650465528.569993] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool mm_recv_desc: align 64, maxelems 4294967295, elemsize 8368
[1650465528.570012] [ndv4:15226:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 4292720 bytes with hugetlb
[1650465528.570023] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool mm_recv_desc: allocated chunk 0x2b1790000018 of 4296680 bytes with 512 elements
[1650465528.570713] [ndv4:15226:0] mm_iface.c:600 UCX DEBUG created mm iface 0x16ba2e0 FIFO id 0x648002 va 0x2b178bdb8000 size 12288 (128 x 64 elems)
[1650465528.570723] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[1]=0x16ba2e0 using sysv/memory on worker 0x1f938d0
[1650465528.570735] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool self_msg_desc: align 64, maxelems 4294967295, elemsize 8200
[1650465528.570738] [ndv4:15226:0] self.c:220 UCX DEBUG created self iface id 0x45a973c9f245d83d send_size 8192
[1650465528.570744] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[2]=0x16cdec0 using self/memory0 on worker 0x1f938d0
[1650465528.570767] [ndv4:15226:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465528.570773] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465528.570775] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465528.572471] [ndv4:13452:0] ib_iface.c:994 UCX DEBUG iface=0x1d95a30: created UD QP 0xdbd1 on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465528.573493] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x16d0000 [id=78 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465528.573522] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 78 events 0x5 mode thread_spinlock
[1650465528.573541] [ndv4:15226:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x16ce820: listening for connections (fd=78) on 10.5.0.5:56827
[1650465528.573882] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465528.574922] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.574993] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.575194] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.575665] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.576188] [ndv4:13452:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae65fe70000..0x2ae65fef5000 on mlx5_ib0 lkey 0x80500 rkey 0x80500 access 0xf flags 0x3e4
[1650465528.576194] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae65fe70018 of 544744 bytes with 128 elements
[1650465528.576199] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465528.573916] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[3]=0x16ce820 using tcp/eth0 on worker 0x1f938d0
[1650465528.573938] [ndv4:15226:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465528.573942] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465528.573944] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465528.574008] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x161c5c0 [id=80 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465528.574030] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 80 events 0x5 mode thread_spinlock
[1650465528.574034] [ndv4:15226:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x16ceea0: listening for connections (fd=80) on 127.0.0.1:36308
[1650465528.574050] [ndv4:15226:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465528.574056] [ndv4:15226:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465528.574206] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[4]=0x16ceea0 using tcp/lo on worker 0x1f938d0
[1650465528.574222] [ndv4:15226:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465528.574225] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465528.574229] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465528.574268] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x162a960 [id=82 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465528.574292] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 82 events 0x5 mode thread_spinlock
[1650465528.574296] [ndv4:15226:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x16b6530: listening for connections (fd=82) on 172.16.1.242:50416
[1650465528.574750] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[5]=0x16b6530 using tcp/ib0 on worker 0x1f938d0
[1650465528.574918] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.574964] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.575195] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.575375] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.575652] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465528.576215] [ndv4:14205:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465528.576225] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465528.577475] [ndv4:13443:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465528.577485] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465528.577446] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465528.577484] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465528.577488] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465528.577500] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465528.577813] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465528.577830] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465528.579366] [ndv4:15226:0] ib_iface.c:994 UCX DEBUG iface=0x16c4670: created RC QP 0xdbe3 on mlx5_ib0:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465528.581440] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x1d95a30: adding gid fe80::15:5dff:fd33:fffb to hash on device mlx5_ib0 port 1 index 0)
[1650465528.582505] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x1d95a30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650465528.583101] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x1d95a30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650465528.587527] [ndv4:14949:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650465528.587837] [ndv4:13443:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465528.587847] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465528.588170] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x1d95a30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650465528.588074] [ndv4:14949:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650465528.589073] [ndv4:14949:0] async.c:228 UCX DEBUG added async handler 0x968cc0 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650465528.589104] [ndv4:14949:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650465528.589108] [ndv4:14949:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650465528.590076] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[6]=0x16c4670 using rc_verbs/mlx5_ib0:1 on worker 0x1f938d0
[1650465528.590241] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.590565] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.589687] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.589774] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.589780] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.589845] [ndv4:14949:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650465528.591035] [ndv4:14949:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.591042] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.591241] [ndv4:14949:0] ib_md.c:1604 UCX DEBUG mlx5_ib1: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.591388] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.591470] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.590823] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.591122] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.591638] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465528.592398] [ndv4:15226:0] ib_device.c:1394 UCX DEBUG max IB CQE size is 128
[1650465528.591962] [ndv4:13552:0] debug.c:1198 UCX DEBUG using signal stack 0x2b264ffd1000 size 141824
[1650465528.592104] [ndv4:13552:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465528.592126] [ndv4:13552:0] init.c:113 UCX DEBUG /cvmfs/ndv4.azure/CentOS-HPC/7.9.2021052401/EasyBuild/software/UCX/1.11.2-GCCcore-11.2.0/lib64/libucs.so.0 loaded at 0x2b26599af000
[1650465528.592291] [ndv4:13552:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650465528.592300] [ndv4:13552:0] module.c:69 UCX DEBUG ucs library path: /cvmfs/ndv4.azure/CentOS-HPC/7.9.2021052401/EasyBuild/software/UCX/1.11.2-GCCcore-11.2.0/lib64/libucs.so.0
[1650465528.592307] [ndv4:13552:0] module.c:253 UCX DEBUG loading modules for ucs
[1650465528.593626] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x1d95a30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650465528.593807] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x1d95a30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650465528.593833] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x1d95a30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650465528.593953] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x1d95a30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650465528.593487] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465528.593496] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465528.593499] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465528.593553] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465528.594106] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x205d010 of 8176 bytes with 127 elements
[1650465528.594357] [ndv4:13452:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465528.594370] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465528.594398] [ndv4:15226:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib0:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465528.594449] [ndv4:15226:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650465528.594455] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465528.597444] [ndv4:13552:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465528.597465] [ndv4:13552:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650465528.597501] [ndv4:13552:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650465528.597504] [ndv4:13552:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650465528.597511] [ndv4:13552:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650465528.597518] [ndv4:13552:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650465528.597522] [ndv4:13552:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650465528.597527] [ndv4:13552:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650465528.597529] [ndv4:13552:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650465528.597532] [ndv4:13552:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650465528.597534] [ndv4:13552:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650465528.597536] [ndv4:13552:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650465528.597544] [ndv4:13552:0] module.c:253 UCX DEBUG loading modules for uct
[1650465528.601759] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x1cd52f0 [id=88 ref 1] uct_ud_iface_async_handler() to hash
[1650465528.601795] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 88 events 0x5 mode thread_spinlock
[1650465528.601825] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[9]=0x1d95a30 using ud_verbs/mlx5_ib0:1 on worker 0x264e8d0
[1650465528.602008] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.602091] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.602256] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.602363] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.603045] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x1629c70 [id=85 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465528.603078] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 85 events 0x1 mode thread_spinlock
[1650465528.603105] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[7]=0x16d20c0 using rc_mlx5/mlx5_ib0:1 on worker 0x1f938d0
[1650465528.603332] [ndv4:14205:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465528.603350] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465528.603080] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465528.603916] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465528.604088] [ndv4:14949:0] topo.c:99 UCX DEBUG bus id 0x102000000 doesn't exist. sys_dev = 1
[1650465528.604096] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465528.605117] [ndv4:13452:0] ib_iface.c:994 UCX DEBUG iface=0x266e750: created UD QP 0xdc27 on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465528.605130] [ndv4:13452:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465528.605487] [ndv4:13443:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465528.605496] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465528.606525] [ndv4:14205:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465528.606532] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465528.606123] [ndv4:13552:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650465528.605895] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465528.606307] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.606659] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.606766] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.606910] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.606671] [ndv4:13443:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465528.607365] [ndv4:13452:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae65fef5000..0x2ae65ff7a000 on mlx5_ib0 lkey 0x80600 rkey 0x80600 access 0xf flags 0x3e4
[1650465528.607371] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae65fef5018 of 544744 bytes with 128 elements
[1650465528.607376] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465528.607794] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x266e750: adding gid fe80::15:5dff:fd33:fffb to hash on device mlx5_ib0 port 1 index 0)
[1650465528.608180] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x266e750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650465528.607931] [ndv4:13552:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650465528.607948] [ndv4:13552:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650465528.607959] [ndv4:13552:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650465528.607970] [ndv4:13552:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650465528.607980] [ndv4:13552:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650465528.607990] [ndv4:13552:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650465528.608001] [ndv4:13552:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650465528.608688] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x266e750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650465528.608920] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x266e750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650465528.609045] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x266e750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650465528.608863] [ndv4:14949:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465528.608870] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465528.610462] [ndv4:13552:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650465528.613721] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.613876] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.614246] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.614528] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.615133] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465528.617061] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x266e750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650465528.617445] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x266e750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650465528.617717] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x266e750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650465528.617725] [ndv4:13452:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib0:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465528.617756] [ndv4:13452:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465528.617761] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x1ce36c0 [id=89 ref 1] uct_ud_iface_async_handler() to hash
[1650465528.617780] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 89 events 0x5 mode thread_spinlock
[1650465528.617797] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[10]=0x266e750 using ud_mlx5/mlx5_ib0:1 on worker 0x264e8d0
[1650465528.618145] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.618406] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.618431] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465528.618440] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465528.618444] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465528.618459] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465528.618881] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.618923] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.619086] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465528.619094] [ndv4:15226:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib0:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465528.619127] [ndv4:15226:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650465528.619131] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465528.619142] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x1628b40 [id=87 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465528.619168] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 87 events 0x1 mode thread_spinlock
[1650465528.619468] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465528.619725] [ndv4:15226:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465528.620830] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465528.620840] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465528.620843] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465528.620862] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465528.621899] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465528.621907] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465528.622491] [ndv4:13452:0] ib_iface.c:994 UCX DEBUG iface=0x2f1f050: created RC QP 0xbfb5 on mlx5_ib1:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465528.621774] [ndv4:14205:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465528.621782] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465528.624099] [ndv4:14205:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465528.623762] [ndv4:13552:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650465528.624216] [ndv4:13552:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650465528.624190] [ndv4:14949:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465528.624198] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465528.628177] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[11]=0x2f1f050 using rc_verbs/mlx5_ib1:1 on worker 0x264e8d0
[1650465528.628248] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.628368] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.628819] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.628832] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.629049] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465528.628491] [ndv4:15226:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x22c7010: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xdc69
[1650465528.629737] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.629864] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.629887] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b178bdbd008 of 151544 bytes with 1052 elements
[1650465528.630309] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465528.630320] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465528.630323] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465528.630340] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465528.630921] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x2f32010 of 8176 bytes with 127 elements
[1650465528.631156] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465528.631165] [ndv4:13452:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib1:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465528.631200] [ndv4:13452:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650465528.631205] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465528.631217] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x1d76f70 [id=92 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465528.631246] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 92 events 0x1 mode thread_spinlock
[1650465528.631257] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[12]=0x2da0010 using rc_mlx5/mlx5_ib1:1 on worker 0x264e8d0
[1650465528.630163] [ndv4:13552:0] async.c:228 UCX DEBUG added async handler 0x13d41e0 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650465528.630272] [ndv4:13552:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650465528.630280] [ndv4:13552:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650465528.630678] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.631087] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.631111] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.631166] [ndv4:13552:0] module.c:253 UCX DEBUG loading modules for ucm
[1650465528.631188] [ndv4:13552:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650465528.632604] [ndv4:13443:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650465528.633052] [ndv4:13552:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.633062] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.633271] [ndv4:13552:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.633903] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.634019] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.634382] [ndv4:14949:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465528.634391] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465528.634193] [ndv4:15226:0] ib_md.c:812 UCX DEBUG registered memory 0x2b1790600000..0x2b1792c00000 on mlx5_ib0 lkey 0x80800 rkey 0x80800 access 0xf flags 0x3e4
[1650465528.634211] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b1790600018 of 39845864 bytes with 4752 elements
[1650465528.634356] [ndv4:15226:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x22c7010
[1650465528.634395] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[8]=0x22c7010 using dc_mlx5/mlx5_ib0:1 on worker 0x1f938d0
[1650465528.634809] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.635036] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.635370] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.635396] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.634418] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.634641] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.634944] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.635034] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.635768] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465528.633876] [ndv4:13443:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650465528.636391] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465528.637511] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465528.637571] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465528.637631] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465528.637635] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465528.637705] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465528.638111] [ndv4:14205:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib3 vendor_id: 0x15b3 device_id: 4124
[1650465528.638109] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465528.638118] [ndv4:13452:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib1:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465528.638145] [ndv4:13452:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650465528.638148] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465528.638157] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x1d88dd0 [id=94 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465528.638180] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 94 events 0x1 mode thread_spinlock
[1650465528.638688] [ndv4:13443:0] async.c:228 UCX DEBUG added async handler 0x1d3bd70 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650465528.638724] [ndv4:13443:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650465528.638730] [ndv4:13443:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650465528.638826] [ndv4:13452:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465528.639085] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.639329] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.639339] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.639409] [ndv4:13443:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650465528.639773] [ndv4:14205:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib3: disable ODP because it's not supported for DevX QP
[1650465528.639719] [ndv4:15226:0] ib_iface.c:994 UCX DEBUG iface=0x16daa30: created UD QP 0xdc87 on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465528.641074] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x20705a0 [id=63 ref 1] uct_ib_async_event_handler() to hash
[1650465528.641095] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 63 events 0x1 mode thread_spinlock
[1650465528.641098] [ndv4:14205:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib3' (InfiniBand channel adapter) with 1 ports
[1650465528.641342] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.641536] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.641544] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.641681] [ndv4:14205:0] ib_md.c:1319 UCX DEBUG mlx5_ib3: using registration cache
[1650465528.640409] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465528.641232] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.641326] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.641566] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.641786] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.642230] [ndv4:15226:0] ib_md.c:812 UCX DEBUG registered memory 0x2b178bde2000..0x2b178be67000 on mlx5_ib0 lkey 0x80a00 rkey 0x80a00 access 0xf flags 0x3e4
[1650465528.642235] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b178bde2018 of 544744 bytes with 128 elements
[1650465528.642239] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465528.641519] [ndv4:13443:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.641527] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.641788] [ndv4:13443:0] ib_md.c:1604 UCX DEBUG mlx5_ib1: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.641935] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.642419] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.643101] [ndv4:14205:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.643110] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.643371] [ndv4:14205:0] ib_md.c:1604 UCX DEBUG mlx5_ib3: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.643744] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.644243] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.647207] [ndv4:13452:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x30ca010: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xbfe5
[1650465528.646782] [ndv4:13443:0] topo.c:99 UCX DEBUG bus id 0x102000000 doesn't exist. sys_dev = 1
[1650465528.646789] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465528.648174] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.648392] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.648402] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2ae65ff7c008 of 151544 bytes with 1052 elements
[1650465528.647865] [ndv4:14205:0] topo.c:99 UCX DEBUG bus id 0x104000000 doesn't exist. sys_dev = 3
[1650465528.647873] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465528.652171] [ndv4:13452:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae666e00000..0x2ae669400000 on mlx5_ib1 lkey 0x80800 rkey 0x80800 access 0xf flags 0x3e4
[1650465528.652196] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2ae666e00018 of 39845864 bytes with 4752 elements
[1650465528.652334] [ndv4:13452:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x30ca010
[1650465528.652370] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[13]=0x30ca010 using dc_mlx5/mlx5_ib1:1 on worker 0x264e8d0
[1650465528.652733] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.652800] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.653245] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.653558] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.652657] [ndv4:14205:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465528.652664] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465528.653946] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465528.654810] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465528.655163] [ndv4:13452:0] ib_iface.c:994 UCX DEBUG iface=0x2fc93d0: created UD QP 0xbfe9 on mlx5_ib1:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465528.655102] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x16daa30: adding gid fe80::15:5dff:fd33:fffb to hash on device mlx5_ib0 port 1 index 0)
[1650465528.655486] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x16daa30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650465528.655711] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x16daa30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650465528.655739] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x16daa30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650465528.655886] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x16daa30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650465528.655847] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465528.656099] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.656388] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.656899] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.656960] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.657455] [ndv4:13452:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae66941b000..0x2ae6694a0000 on mlx5_ib1 lkey 0x80900 rkey 0x80900 access 0xf flags 0x3e4
[1650465528.657461] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae66941b018 of 544744 bytes with 128 elements
[1650465528.657466] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465528.657643] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x2fc93d0: adding gid fe80::15:5dff:fd33:fffc to hash on device mlx5_ib1 port 1 index 0)
[1650465528.657684] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x2fc93d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 1)
[1650465528.657773] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x2fc93d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 2)
[1650465528.657913] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x2fc93d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 3)
[1650465528.657973] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x2fc93d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 4)
[1650465528.658234] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x2fc93d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 5)
[1650465528.658564] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x2fc93d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 6)
[1650465528.658837] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x2fc93d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 7)
[1650465528.659182] [ndv4:13452:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465528.659191] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x1d7d2b0 [id=95 ref 1] uct_ud_iface_async_handler() to hash
[1650465528.659227] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 95 events 0x5 mode thread_spinlock
[1650465528.659238] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[14]=0x2fc93d0 using ud_verbs/mlx5_ib1:1 on worker 0x264e8d0
[1650465528.661181] [ndv4:13443:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465528.661190] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465528.661709] [ndv4:12741:0] debug.c:1198 UCX DEBUG using signal stack 0x2b7350987000 size 141824
[1650465528.662254] [ndv4:12741:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465528.662276] [ndv4:12741:0] init.c:113 UCX DEBUG /cvmfs/ndv4.azure/CentOS-HPC/7.9.2021052401/EasyBuild/software/UCX/1.11.2-GCCcore-11.2.0/lib64/libucs.so.0 loaded at 0x2b73507db000
[1650465528.662304] [ndv4:12741:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650465528.662313] [ndv4:12741:0] module.c:69 UCX DEBUG ucs library path: /cvmfs/ndv4.azure/CentOS-HPC/7.9.2021052401/EasyBuild/software/UCX/1.11.2-GCCcore-11.2.0/lib64/libucs.so.0
[1650465528.662319] [ndv4:12741:0] module.c:253 UCX DEBUG loading modules for ucs
[1650465528.664733] [ndv4:13552:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650465528.664742] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465528.665519] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x16daa30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650465528.665869] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x16daa30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650465528.666144] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x16daa30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650465528.666712] [ndv4:15226:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465528.664732] [ndv4:14205:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465528.664738] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465528.667392] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.667673] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.668073] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.668154] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.667455] [ndv4:12741:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465528.667479] [ndv4:12741:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650465528.667512] [ndv4:12741:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650465528.667515] [ndv4:12741:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650465528.667521] [ndv4:12741:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650465528.667528] [ndv4:12741:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650465528.667531] [ndv4:12741:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650465528.667535] [ndv4:12741:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650465528.667537] [ndv4:12741:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650465528.667540] [ndv4:12741:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650465528.667542] [ndv4:12741:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650465528.667544] [ndv4:12741:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650465528.667552] [ndv4:12741:0] module.c:253 UCX DEBUG loading modules for uct
[1650465528.668520] [ndv4:14421:0] debug.c:1198 UCX DEBUG using signal stack 0x2ab786fdd000 size 141824
[1650465528.668843] [ndv4:14421:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465528.668864] [ndv4:14421:0] init.c:113 UCX DEBUG /cvmfs/ndv4.azure/CentOS-HPC/7.9.2021052401/EasyBuild/software/UCX/1.11.2-GCCcore-11.2.0/lib64/libucs.so.0 loaded at 0x2ab786e31000
[1650465528.668920] [ndv4:14421:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650465528.668929] [ndv4:14421:0] module.c:69 UCX DEBUG ucs library path: /cvmfs/ndv4.azure/CentOS-HPC/7.9.2021052401/EasyBuild/software/UCX/1.11.2-GCCcore-11.2.0/lib64/libucs.so.0
[1650465528.668936] [ndv4:14421:0] module.c:253 UCX DEBUG loading modules for ucs
[1650465528.668801] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465528.669714] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465528.670173] [ndv4:13452:0] ib_iface.c:994 UCX DEBUG iface=0x2e90280: created UD QP 0xc000 on mlx5_ib1:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465528.670185] [ndv4:13452:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465528.670812] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465528.671182] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.671407] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.671863] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.671959] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.672145] [ndv4:14949:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465528.672155] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465528.672366] [ndv4:13452:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae6694a0000..0x2ae669525000 on mlx5_ib1 lkey 0x80a00 rkey 0x80a00 access 0xf flags 0x3e4
[1650465528.672373] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae6694a0018 of 544744 bytes with 128 elements
[1650465528.672377] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465528.672683] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x2e90280: adding gid fe80::15:5dff:fd33:fffc to hash on device mlx5_ib1 port 1 index 0)
[1650465528.672926] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x2e90280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 1)
[1650465528.673190] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x2e90280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 2)
[1650465528.673949] [ndv4:14949:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465528.674051] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x161a2f0 [id=88 ref 1] uct_ud_iface_async_handler() to hash
[1650465528.674090] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 88 events 0x5 mode thread_spinlock
[1650465528.674126] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[9]=0x16daa30 using ud_verbs/mlx5_ib0:1 on worker 0x1f938d0
[1650465528.674670] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.674999] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.675366] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.675412] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.674155] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x2e90280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 3)
[1650465528.675382] [ndv4:14421:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465528.675405] [ndv4:14421:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650465528.675435] [ndv4:14421:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650465528.675438] [ndv4:14421:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650465528.675446] [ndv4:14421:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650465528.675453] [ndv4:14421:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650465528.675456] [ndv4:14421:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650465528.675460] [ndv4:14421:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650465528.675462] [ndv4:14421:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650465528.675465] [ndv4:14421:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650465528.675467] [ndv4:14421:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650465528.675469] [ndv4:14421:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650465528.675476] [ndv4:14421:0] module.c:253 UCX DEBUG loading modules for uct
[1650465528.676733] [ndv4:12741:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650465528.676095] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465528.677496] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465528.679159] [ndv4:12741:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650465528.679176] [ndv4:12741:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650465528.679370] [ndv4:12741:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650465528.679405] [ndv4:12741:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650465528.679569] [ndv4:12741:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650465528.679669] [ndv4:12741:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650465528.679680] [ndv4:12741:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650465528.679848] [ndv4:15226:0] ib_iface.c:994 UCX DEBUG iface=0x1fb3750: created UD QP 0xdcee on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465528.679863] [ndv4:15226:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465528.678708] [ndv4:13443:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465528.678716] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465528.680122] [ndv4:14205:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465528.680130] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465528.679782] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x2e90280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 4)
[1650465528.680202] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x2e90280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 5)
[1650465528.680922] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465528.681247] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.681730] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.681939] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.682072] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.682425] [ndv4:12741:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650465528.682533] [ndv4:15226:0] ib_md.c:812 UCX DEBUG registered memory 0x2b178be67000..0x2b178beec000 on mlx5_ib0 lkey 0x80b00 rkey 0x80b00 access 0xf flags 0x3e4
[1650465528.682540] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b178be67018 of 544744 bytes with 128 elements
[1650465528.682545] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465528.682825] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x1fb3750: adding gid fe80::15:5dff:fd33:fffb to hash on device mlx5_ib0 port 1 index 0)
[1650465528.682970] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x1fb3750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650465528.683143] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x1fb3750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650465528.682987] [ndv4:13443:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465528.682994] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465528.683637] [ndv4:14421:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650465528.687231] [ndv4:13443:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465528.687237] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465528.689528] [ndv4:13443:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465528.692066] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x2e90280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 6)
[1650465528.692681] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x2e90280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 7)
[1650465528.692690] [ndv4:13452:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib1:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465528.692722] [ndv4:13452:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465528.692728] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x2da8c30 [id=96 ref 1] uct_ud_iface_async_handler() to hash
[1650465528.692755] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 96 events 0x5 mode thread_spinlock
[1650465528.692771] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[15]=0x2e90280 using ud_mlx5/mlx5_ib1:1 on worker 0x264e8d0
[1650465528.692984] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.693102] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.693431] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.693467] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.694564] [ndv4:13552:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465528.694661] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465528.695669] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465528.695802] [ndv4:14205:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465528.695810] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465528.696993] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465528.697012] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465528.697016] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465528.697067] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465528.697100] [ndv4:13552:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465528.697106] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465528.696956] [ndv4:14205:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465528.698189] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465528.698196] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465528.699617] [ndv4:13552:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465528.699624] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465528.698713] [ndv4:13452:0] ib_iface.c:994 UCX DEBUG iface=0x34b90a0: created RC QP 0xbff3 on mlx5_ib2:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465528.700293] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[16]=0x34b90a0 using rc_verbs/mlx5_ib2:1 on worker 0x264e8d0
[1650465528.701158] [ndv4:14421:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650465528.701175] [ndv4:14421:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650465528.701187] [ndv4:14421:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650465528.701199] [ndv4:14421:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650465528.701209] [ndv4:14421:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650465528.701219] [ndv4:14421:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650465528.701230] [ndv4:14421:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650465528.701659] [ndv4:13443:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib2 vendor_id: 0x15b3 device_id: 4124
[1650465528.700822] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x1fb3750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650465528.700894] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x1fb3750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650465528.701106] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x1fb3750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650465528.701836] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x1fb3750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650465528.702185] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x1fb3750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650465528.702192] [ndv4:15226:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib0:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465528.702223] [ndv4:15226:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465528.702228] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x16286c0 [id=89 ref 1] uct_ud_iface_async_handler() to hash
[1650465528.702259] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 89 events 0x5 mode thread_spinlock
[1650465528.702273] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[10]=0x1fb3750 using ud_mlx5/mlx5_ib0:1 on worker 0x1f938d0
[1650465528.702410] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.702074] [ndv4:13443:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib2: disable ODP because it's not supported for DevX QP
[1650465528.703315] [ndv4:13443:0] async.c:228 UCX DEBUG added async handler 0x1d43c00 [id=61 ref 1] uct_ib_async_event_handler() to hash
[1650465528.703349] [ndv4:13443:0] async.c:506 UCX DEBUG listening to async event fd 61 events 0x1 mode thread_spinlock
[1650465528.703352] [ndv4:13443:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib2' (InfiniBand channel adapter) with 1 ports
[1650465528.702201] [ndv4:14949:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib2 vendor_id: 0x15b3 device_id: 4124
[1650465528.702722] [ndv4:14949:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib2: disable ODP because it's not supported for DevX QP
[1650465528.702748] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.703269] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.703404] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.703832] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.704204] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.704210] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.704252] [ndv4:13443:0] ib_md.c:1319 UCX DEBUG mlx5_ib2: using registration cache
[1650465528.704276] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465528.704354] [ndv4:14421:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650465528.703879] [ndv4:14949:0] async.c:228 UCX DEBUG added async handler 0x970ad0 [id=61 ref 1] uct_ib_async_event_handler() to hash
[1650465528.703906] [ndv4:14949:0] async.c:506 UCX DEBUG listening to async event fd 61 events 0x1 mode thread_spinlock
[1650465528.703911] [ndv4:14949:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib2' (InfiniBand channel adapter) with 1 ports
[1650465528.704164] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.704284] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.704293] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.704370] [ndv4:14949:0] ib_md.c:1319 UCX DEBUG mlx5_ib2: using registration cache
[1650465528.705119] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465528.705131] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465528.705133] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465528.705148] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465528.705400] [ndv4:14949:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.705408] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.705787] [ndv4:12741:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650465528.706035] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465528.706043] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465528.706256] [ndv4:12741:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650465528.706548] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.706808] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.707094] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.707663] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.706764] [ndv4:14949:0] ib_md.c:1604 UCX DEBUG mlx5_ib2: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.706930] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.707168] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.706704] [ndv4:15226:0] ib_iface.c:994 UCX DEBUG iface=0x2864050: created RC QP 0xc041 on mlx5_ib1:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465528.708279] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465528.709932] [ndv4:14205:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib4 vendor_id: 0x15b3 device_id: 4124
[1650465528.710375] [ndv4:14205:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib4: disable ODP because it's not supported for DevX QP
[1650465528.709660] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[11]=0x2864050 using rc_verbs/mlx5_ib1:1 on worker 0x1f938d0
[1650465528.710745] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.711076] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.710813] [ndv4:13552:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465528.710825] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465528.711889] [ndv4:13552:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465528.711139] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x206fd30 [id=65 ref 1] uct_ib_async_event_handler() to hash
[1650465528.711163] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 65 events 0x1 mode thread_spinlock
[1650465528.711167] [ndv4:14205:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib4' (InfiniBand channel adapter) with 1 ports
[1650465528.711673] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465528.711826] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465528.711832] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.711873] [ndv4:14205:0] ib_md.c:1319 UCX DEBUG mlx5_ib4: using registration cache
[1650465528.711074] [ndv4:13443:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.711086] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.711298] [ndv4:13443:0] ib_md.c:1604 UCX DEBUG mlx5_ib2: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.711748] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.711829] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.711766] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.711825] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.712274] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465528.712630] [ndv4:12741:0] async.c:228 UCX DEBUG added async handler 0x26f61e0 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650465528.712747] [ndv4:12741:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650465528.712754] [ndv4:12741:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650465528.713131] [ndv4:14205:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.713139] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.713334] [ndv4:14205:0] ib_md.c:1604 UCX DEBUG mlx5_ib4: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.713625] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465528.713663] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465528.713019] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.713082] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.713104] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.713165] [ndv4:12741:0] module.c:253 UCX DEBUG loading modules for ucm
[1650465528.713187] [ndv4:12741:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650465528.713490] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465528.713499] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465528.713502] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465528.713521] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465528.714023] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x2877010 of 8176 bytes with 127 elements
[1650465528.714325] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465528.714331] [ndv4:15226:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib1:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465528.714367] [ndv4:15226:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650465528.714371] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465528.714381] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x16bbf70 [id=92 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465528.714421] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 92 events 0x1 mode thread_spinlock
[1650465528.714430] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[12]=0x26e5010 using rc_mlx5/mlx5_ib1:1 on worker 0x1f938d0
[1650465528.714750] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.714794] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.715521] [ndv4:12741:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.715531] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.716047] [ndv4:12741:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.716103] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.716146] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.715947] [ndv4:14949:0] topo.c:99 UCX DEBUG bus id 0x103000000 doesn't exist. sys_dev = 2
[1650465528.715957] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465528.717102] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465528.717120] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465528.717124] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465528.717175] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465528.717719] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x37da010 of 8176 bytes with 127 elements
[1650465528.718505] [ndv4:14421:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650465528.717966] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465528.717974] [ndv4:13452:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib2:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465528.718010] [ndv4:13452:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib2 length=2048) failed: Invalid argument
[1650465528.718014] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465528.718024] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x1d975f0 [id=99 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465528.718048] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 99 events 0x1 mode thread_spinlock
[1650465528.718073] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[17]=0x3606030 using rc_mlx5/mlx5_ib2:1 on worker 0x264e8d0
[1650465528.718190] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.718393] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.718688] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.718847] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.718936] [ndv4:14421:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650465528.719308] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465528.720970] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.721028] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.721102] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465528.721120] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465528.721123] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465528.721177] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465528.721421] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465528.722839] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465528.722854] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465528.722858] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465528.722912] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465528.722062] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465528.722070] [ndv4:13452:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib2:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465528.722102] [ndv4:13452:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib2 length=2048) failed: Invalid argument
[1650465528.722105] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465528.722113] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x1d774c0 [id=101 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465528.722142] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 101 events 0x1 mode thread_spinlock
[1650465528.722787] [ndv4:13452:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465528.721958] [ndv4:13443:0] topo.c:99 UCX DEBUG bus id 0x103000000 doesn't exist. sys_dev = 2
[1650465528.721967] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465528.723368] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465528.723375] [ndv4:15226:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib1:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465528.723409] [ndv4:15226:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650465528.723412] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465528.723419] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x16cddd0 [id=94 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465528.723444] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 94 events 0x1 mode thread_spinlock
[1650465528.723781] [ndv4:14205:0] topo.c:99 UCX DEBUG bus id 0x105000000 doesn't exist. sys_dev = 4
[1650465528.723789] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465528.723957] [ndv4:15226:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465528.726757] [ndv4:12741:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650465528.726765] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465528.729154] [ndv4:14949:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465528.729166] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465528.730973] [ndv4:14421:0] async.c:228 UCX DEBUG added async handler 0xf35280 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650465528.731086] [ndv4:14421:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650465528.731094] [ndv4:14421:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650465528.731554] [ndv4:12741:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465528.731560] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465528.731731] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.731766] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.731788] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.731851] [ndv4:14421:0] module.c:253 UCX DEBUG loading modules for ucm
[1650465528.731873] [ndv4:14421:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650465528.731805] [ndv4:13552:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650465528.732315] [ndv4:13552:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650465528.733361] [ndv4:14421:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.733369] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.733621] [ndv4:14421:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.734103] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465528.734500] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465528.732367] [ndv4:13452:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x37dc050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xc018
[1650465528.732888] [ndv4:14949:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465528.732895] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465528.734644] [ndv4:13443:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465528.734672] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465528.734717] [ndv4:14205:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465528.734724] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465528.732840] [ndv4:15226:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x2a0f010: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xc07a
[1650465528.734112] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.734172] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.734182] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b178beee008 of 151544 bytes with 1052 elements
[1650465528.735754] [ndv4:12741:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465528.735761] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465528.737717] [ndv4:14949:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465528.737725] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465528.738137] [ndv4:15226:0] ib_md.c:812 UCX DEBUG registered memory 0x2b1792e00000..0x2b1795400000 on mlx5_ib1 lkey 0x80b00 rkey 0x80b00 access 0xf flags 0x3e4
[1650465528.738159] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b1792e00018 of 39845864 bytes with 4752 elements
[1650465528.738297] [ndv4:15226:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x2a0f010
[1650465528.738334] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[13]=0x2a0f010 using dc_mlx5/mlx5_ib1:1 on worker 0x1f938d0
[1650465528.738692] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.738775] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.739068] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.739101] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.738791] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.739140] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.739152] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2ae65ffa3008 of 151544 bytes with 1052 elements
[1650465528.739955] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465528.740085] [ndv4:14421:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650465528.740092] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465528.740774] [ndv4:14949:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465528.740780] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465528.741265] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465528.742518] [ndv4:14949:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465528.742977] [ndv4:15226:0] ib_iface.c:994 UCX DEBUG iface=0x290e3d0: created UD QP 0xc08a on mlx5_ib1:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465528.743545] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465528.743911] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.744047] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.743161] [ndv4:14205:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465528.743170] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465528.743265] [ndv4:13452:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae669600000..0x2ae66bc00000 on mlx5_ib2 lkey 0x80800 rkey 0x80800 access 0xf flags 0x3e4
[1650465528.743294] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2ae669600018 of 39845864 bytes with 4752 elements
[1650465528.743439] [ndv4:13452:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x37dc050
[1650465528.743473] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[18]=0x37dc050 using dc_mlx5/mlx5_ib2:1 on worker 0x264e8d0
[1650465528.744191] [ndv4:12741:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465528.744200] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465528.744754] [ndv4:13443:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465528.744763] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465528.744698] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.744709] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.745123] [ndv4:15226:0] ib_md.c:812 UCX DEBUG registered memory 0x2b178bf13000..0x2b178bf98000 on mlx5_ib1 lkey 0x80c00 rkey 0x80c00 access 0xf flags 0x3e4
[1650465528.745129] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b178bf13018 of 544744 bytes with 128 elements
[1650465528.745134] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465528.745472] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x290e3d0: adding gid fe80::15:5dff:fd33:fffc to hash on device mlx5_ib1 port 1 index 0)
[1650465528.745566] [ndv4:13552:0] async.c:228 UCX DEBUG added async handler 0x13ccf50 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650465528.745662] [ndv4:13552:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650465528.745666] [ndv4:13552:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650465528.745894] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x290e3d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 1)
[1650465528.745891] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.746143] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.746150] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.746197] [ndv4:13552:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650465528.747017] [ndv4:13552:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.747025] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.747660] [ndv4:13552:0] ib_md.c:1604 UCX DEBUG mlx5_ib1: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.748019] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.748335] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.748223] [ndv4:12741:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465528.748230] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465528.750171] [ndv4:12741:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465528.751796] [ndv4:13552:0] topo.c:99 UCX DEBUG bus id 0x102000000 doesn't exist. sys_dev = 1
[1650465528.751804] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465528.754860] [ndv4:13552:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465528.754866] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465528.758293] [ndv4:13552:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465528.758299] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465528.759734] [ndv4:14421:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465528.759741] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465528.759684] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.759932] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.760256] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.760481] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.760885] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465528.760729] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x290e3d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 2)
[1650465528.761849] [ndv4:14205:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465528.761857] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465528.761959] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465528.762287] [ndv4:13452:0] ib_iface.c:994 UCX DEBUG iface=0x39b8060: created UD QP 0xc028 on mlx5_ib2:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465528.763066] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465528.763398] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.763921] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.764214] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.764569] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.762876] [ndv4:13443:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465528.762884] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465528.765170] [ndv4:13452:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae66bd26000..0x2ae66bdab000 on mlx5_ib2 lkey 0x80900 rkey 0x80900 access 0xf flags 0x3e4
[1650465528.765177] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae66bd26018 of 544744 bytes with 128 elements
[1650465528.765181] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465528.765243] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x39b8060: adding gid fe80::15:5dff:fd33:fffd to hash on device mlx5_ib2 port 1 index 0)
[1650465528.765466] [ndv4:14205:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465528.765476] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465528.767647] [ndv4:14205:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465528.771885] [ndv4:14949:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib3 vendor_id: 0x15b3 device_id: 4124
[1650465528.772352] [ndv4:14949:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib3: disable ODP because it's not supported for DevX QP
[1650465528.772730] [ndv4:14949:0] async.c:228 UCX DEBUG added async handler 0x9709c0 [id=63 ref 1] uct_ib_async_event_handler() to hash
[1650465528.772767] [ndv4:14949:0] async.c:506 UCX DEBUG listening to async event fd 63 events 0x1 mode thread_spinlock
[1650465528.772771] [ndv4:14949:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib3' (InfiniBand channel adapter) with 1 ports
[1650465528.774257] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x290e3d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 3)
[1650465528.774389] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x290e3d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 4)
[1650465528.774625] [ndv4:13552:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465528.774632] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465528.774562] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.774795] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.774804] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.774858] [ndv4:14949:0] ib_md.c:1319 UCX DEBUG mlx5_ib3: using registration cache
[1650465528.776189] [ndv4:14421:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465528.776197] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465528.777074] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x39b8060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 1)
[1650465528.777297] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x39b8060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 2)
[1650465528.777535] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x39b8060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 3)
[1650465528.777551] [ndv4:13552:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465528.777558] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465528.778414] [ndv4:13443:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465528.778422] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465528.778140] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x39b8060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 4)
[1650465528.778442] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x39b8060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 5)
[1650465528.778573] [ndv4:14949:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.778635] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.778791] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x39b8060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 6)
[1650465528.778903] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x39b8060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 7)
[1650465528.779210] [ndv4:13452:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465528.779219] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x2f35e70 [id=102 ref 1] uct_ud_iface_async_handler() to hash
[1650465528.779240] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 102 events 0x5 mode thread_spinlock
[1650465528.779255] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[19]=0x39b8060 using ud_verbs/mlx5_ib2:1 on worker 0x264e8d0
[1650465528.779174] [ndv4:13552:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465528.779724] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.779809] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.780466] [ndv4:13443:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465528.780649] [ndv4:14949:0] ib_md.c:1604 UCX DEBUG mlx5_ib3: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.780771] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.780820] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.782078] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x290e3d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 5)
[1650465528.782638] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x290e3d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 6)
[1650465528.782979] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x290e3d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 7)
[1650465528.783320] [ndv4:12741:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650465528.783929] [ndv4:12741:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650465528.784368] [ndv4:12741:0] async.c:228 UCX DEBUG added async handler 0x26eef50 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650465528.784393] [ndv4:12741:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650465528.784397] [ndv4:12741:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650465528.784716] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.784930] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.784937] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.784979] [ndv4:12741:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650465528.783325] [ndv4:15226:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465528.783338] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x16c22b0 [id=95 ref 1] uct_ud_iface_async_handler() to hash
[1650465528.783372] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 95 events 0x5 mode thread_spinlock
[1650465528.783385] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[14]=0x290e3d0 using ud_verbs/mlx5_ib1:1 on worker 0x1f938d0
[1650465528.783802] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.783909] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.784167] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.784276] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.785280] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465528.784026] [ndv4:14949:0] topo.c:99 UCX DEBUG bus id 0x104000000 doesn't exist. sys_dev = 3
[1650465528.784034] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465528.786167] [ndv4:12741:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.786174] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.786404] [ndv4:12741:0] ib_md.c:1604 UCX DEBUG mlx5_ib1: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.786687] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.787128] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.788006] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.788136] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.788557] [ndv4:14205:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib5 vendor_id: 0x15b3 device_id: 4124
[1650465528.788762] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465528.789105] [ndv4:14205:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib5: disable ODP because it's not supported for DevX QP
[1650465528.789796] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465528.789972] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x198af90 [id=67 ref 1] uct_ib_async_event_handler() to hash
[1650465528.790002] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 67 events 0x1 mode thread_spinlock
[1650465528.790005] [ndv4:14205:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib5' (InfiniBand channel adapter) with 1 ports
[1650465528.790107] [ndv4:13452:0] ib_iface.c:994 UCX DEBUG iface=0x3ad6050: created UD QP 0xc03c on mlx5_ib2:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465528.790115] [ndv4:13452:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465528.790851] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465528.790446] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465528.790972] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465528.790980] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.791028] [ndv4:14205:0] ib_md.c:1319 UCX DEBUG mlx5_ib5: using registration cache
[1650465528.791124] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.791500] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.792850] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.793011] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.794524] [ndv4:14421:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465528.794537] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465528.793524] [ndv4:13452:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae66bdab000..0x2ae66be30000 on mlx5_ib2 lkey 0x80a00 rkey 0x80a00 access 0xf flags 0x3e4
[1650465528.793530] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae66bdab018 of 544744 bytes with 128 elements
[1650465528.793535] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465528.794035] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x3ad6050: adding gid fe80::15:5dff:fd33:fffd to hash on device mlx5_ib2 port 1 index 0)
[1650465528.794313] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x3ad6050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 1)
[1650465528.794523] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x3ad6050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 2)
[1650465528.794779] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x3ad6050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 3)
[1650465528.795133] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x3ad6050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 4)
[1650465528.795068] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465528.795947] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x3ad6050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 5)
[1650465528.796162] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x3ad6050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 6)
[1650465528.796543] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x3ad6050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 7)
[1650465528.796551] [ndv4:13452:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib2:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465528.796668] [ndv4:13452:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465528.796674] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x34c1c60 [id=103 ref 1] uct_ud_iface_async_handler() to hash
[1650465528.796697] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 103 events 0x5 mode thread_spinlock
[1650465528.796707] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[20]=0x3ad6050 using ud_mlx5/mlx5_ib2:1 on worker 0x264e8d0
[1650465528.796957] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.797226] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.797890] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.797941] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.798658] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465528.797070] [ndv4:13443:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib3 vendor_id: 0x15b3 device_id: 4124
[1650465528.797494] [ndv4:13443:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib3: disable ODP because it's not supported for DevX QP
[1650465528.798465] [ndv4:13443:0] async.c:228 UCX DEBUG added async handler 0x1d3b3a0 [id=63 ref 1] uct_ib_async_event_handler() to hash
[1650465528.798494] [ndv4:13443:0] async.c:506 UCX DEBUG listening to async event fd 63 events 0x1 mode thread_spinlock
[1650465528.798498] [ndv4:13443:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib3' (InfiniBand channel adapter) with 1 ports
[1650465528.798994] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.799164] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.799170] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.799223] [ndv4:13443:0] ib_md.c:1319 UCX DEBUG mlx5_ib3: using registration cache
[1650465528.796796] [ndv4:15226:0] ib_iface.c:994 UCX DEBUG iface=0x27d5280: created UD QP 0xc101 on mlx5_ib1:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465528.796805] [ndv4:15226:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465528.797676] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465528.798420] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.798757] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.798806] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.798828] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.799545] [ndv4:15226:0] ib_md.c:812 UCX DEBUG registered memory 0x2b179541b000..0x2b17954a0000 on mlx5_ib1 lkey 0x80d00 rkey 0x80d00 access 0xf flags 0x3e4
[1650465528.799552] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b179541b018 of 544744 bytes with 128 elements
[1650465528.799557] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465528.799970] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x27d5280: adding gid fe80::15:5dff:fd33:fffc to hash on device mlx5_ib1 port 1 index 0)
[1650465528.798489] [ndv4:12741:0] topo.c:99 UCX DEBUG bus id 0x102000000 doesn't exist. sys_dev = 1
[1650465528.798496] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465528.801621] [ndv4:12741:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465528.801627] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465528.799949] [ndv4:14205:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.799957] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.800177] [ndv4:14205:0] ib_md.c:1604 UCX DEBUG mlx5_ib5: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.800677] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465528.800836] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465528.800161] [ndv4:13443:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.800168] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.800321] [ndv4:13443:0] ib_md.c:1604 UCX DEBUG mlx5_ib3: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.800692] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.800879] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.800119] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465528.800135] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465528.800138] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465528.800188] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465528.801240] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465528.801246] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465528.801744] [ndv4:13452:0] ib_iface.c:994 UCX DEBUG iface=0x3bd60a0: created RC QP 0xbf5b on mlx5_ib3:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465528.797775] [ndv4:13552:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib2 vendor_id: 0x15b3 device_id: 4124
[1650465528.798193] [ndv4:13552:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib2: disable ODP because it's not supported for DevX QP
[1650465528.798892] [ndv4:13552:0] async.c:228 UCX DEBUG added async handler 0x13d4ce0 [id=61 ref 1] uct_ib_async_event_handler() to hash
[1650465528.798915] [ndv4:13552:0] async.c:506 UCX DEBUG listening to async event fd 61 events 0x1 mode thread_spinlock
[1650465528.798919] [ndv4:13552:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib2' (InfiniBand channel adapter) with 1 ports
[1650465528.799316] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.799381] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.799387] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.799434] [ndv4:13552:0] ib_md.c:1319 UCX DEBUG mlx5_ib2: using registration cache
[1650465528.804068] [ndv4:14205:0] topo.c:99 UCX DEBUG bus id 0x106000000 doesn't exist. sys_dev = 5
[1650465528.804076] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465528.804014] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[21]=0x3bd60a0 using rc_verbs/mlx5_ib3:1 on worker 0x264e8d0
[1650465528.804103] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.804144] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.804507] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.805606] [ndv4:14949:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465528.805614] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465528.805102] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.806120] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465528.806483] [ndv4:13443:0] topo.c:99 UCX DEBUG bus id 0x104000000 doesn't exist. sys_dev = 3
[1650465528.806490] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465528.810473] [ndv4:13552:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.810482] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.809724] [ndv4:14949:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465528.809731] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465528.809842] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x27d5280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 1)
[1650465528.810304] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x27d5280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 2)
[1650465528.810777] [ndv4:13552:0] ib_md.c:1604 UCX DEBUG mlx5_ib2: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.810969] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.811248] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.810857] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x27d5280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 3)
[1650465528.811627] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x27d5280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 4)
[1650465528.811729] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x27d5280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 5)
[1650465528.812090] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x27d5280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 6)
[1650465528.812165] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x27d5280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 7)
[1650465528.812175] [ndv4:15226:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib1:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465528.812279] [ndv4:15226:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465528.812286] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x26edc30 [id=96 ref 1] uct_ud_iface_async_handler() to hash
[1650465528.812301] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 96 events 0x5 mode thread_spinlock
[1650465528.812323] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[15]=0x27d5280 using ud_mlx5/mlx5_ib1:1 on worker 0x1f938d0
[1650465528.812711] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.813530] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.814199] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.814322] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.814658] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465528.815242] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465528.815261] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465528.815265] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465528.815320] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465528.815725] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465528.815742] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465528.815745] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465528.815798] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465528.815932] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x3d24010 of 8176 bytes with 127 elements
[1650465528.815531] [ndv4:14205:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465528.815540] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465528.816269] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465528.816276] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465528.816909] [ndv4:15226:0] ib_iface.c:994 UCX DEBUG iface=0x2dfe0a0: created RC QP 0xc052 on mlx5_ib2:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465528.816529] [ndv4:14421:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465528.816543] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465528.816169] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465528.816178] [ndv4:13452:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib3:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465528.816214] [ndv4:13452:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib3 length=2048) failed: Invalid argument
[1650465528.816218] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465528.816232] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x2b463b0 [id=106 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465528.816251] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 106 events 0x1 mode thread_spinlock
[1650465528.816269] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[22]=0x3d2b100 using rc_mlx5/mlx5_ib3:1 on worker 0x264e8d0
[1650465528.816334] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.816970] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.817300] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.817461] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.819144] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[16]=0x2dfe0a0 using rc_verbs/mlx5_ib2:1 on worker 0x1f938d0
[1650465528.819285] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.819364] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.820066] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.820226] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.820318] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465528.818826] [ndv4:14421:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465528.819245] [ndv4:14949:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465528.819252] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465528.821231] [ndv4:13443:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465528.821238] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465528.820788] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465528.822484] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465528.822498] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465528.822501] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465528.822556] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465528.823056] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x311f010 of 8176 bytes with 127 elements
[1650465528.822948] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465528.822967] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465528.822971] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465528.823028] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465528.824787] [ndv4:13552:0] topo.c:99 UCX DEBUG bus id 0x103000000 doesn't exist. sys_dev = 2
[1650465528.824796] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465528.823549] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465528.823557] [ndv4:13452:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib3:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465528.823631] [ndv4:13452:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib3 length=2048) failed: Invalid argument
[1650465528.823635] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465528.823645] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x1d72f00 [id=108 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465528.823676] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 108 events 0x1 mode thread_spinlock
[1650465528.824340] [ndv4:13452:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465528.823364] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465528.823370] [ndv4:15226:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib2:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465528.823406] [ndv4:15226:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib2 length=2048) failed: Invalid argument
[1650465528.823410] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465528.823422] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x16dc5f0 [id=99 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465528.823450] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 99 events 0x1 mode thread_spinlock
[1650465528.823477] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[17]=0x2f4b030 using rc_mlx5/mlx5_ib2:1 on worker 0x1f938d0
[1650465528.823750] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.824390] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.824826] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.824863] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.826017] [ndv4:12741:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465528.826026] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465528.827743] [ndv4:13443:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465528.827751] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465528.831526] [ndv4:14205:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465528.831535] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465528.831136] [ndv4:12741:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465528.831143] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465528.832341] [ndv4:14421:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650465528.832895] [ndv4:14421:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650465528.833553] [ndv4:12741:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465528.833559] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465528.834138] [ndv4:14421:0] async.c:228 UCX DEBUG added async handler 0xf2bcf0 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650465528.834164] [ndv4:14421:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650465528.834168] [ndv4:14421:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650465528.834852] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.835777] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.835786] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.835846] [ndv4:14421:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650465528.835281] [ndv4:12741:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465528.834632] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465528.836035] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465528.836050] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465528.836054] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465528.836112] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465528.836777] [ndv4:13552:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465528.836785] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465528.836660] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465528.836666] [ndv4:15226:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib2:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465528.836700] [ndv4:15226:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib2 length=2048) failed: Invalid argument
[1650465528.836704] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465528.836712] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x16bc4c0 [id=101 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465528.836732] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 101 events 0x1 mode thread_spinlock
[1650465528.837249] [ndv4:15226:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465528.836761] [ndv4:13452:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x3efe010: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xbf8c
[1650465528.837314] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.837488] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.837501] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2ae65ffca008 of 151544 bytes with 1052 elements
[1650465528.837885] [ndv4:14205:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465528.837891] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465528.839694] [ndv4:13443:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465528.839705] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465528.842096] [ndv4:13452:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae66c000000..0x2ae66e600000 on mlx5_ib3 lkey 0x80400 rkey 0x80400 access 0xf flags 0x3e4
[1650465528.842117] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2ae66c000018 of 39845864 bytes with 4752 elements
[1650465528.842261] [ndv4:13452:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x3efe010
[1650465528.842293] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[23]=0x3efe010 using dc_mlx5/mlx5_ib3:1 on worker 0x264e8d0
[1650465528.842509] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.842753] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.843548] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.843712] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.842337] [ndv4:14421:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.842346] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.842702] [ndv4:14421:0] ib_md.c:1604 UCX DEBUG mlx5_ib1: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.843556] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465528.843739] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465528.841748] [ndv4:14205:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465528.841754] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465528.843751] [ndv4:14205:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465528.844080] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465528.845212] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465528.845633] [ndv4:13452:0] ib_iface.c:994 UCX DEBUG iface=0x40da060: created UD QP 0xbf90 on mlx5_ib3:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465528.846130] [ndv4:14949:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465528.846144] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465528.846376] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465528.846798] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.846825] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.846914] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.847098] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.846486] [ndv4:12741:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib2 vendor_id: 0x15b3 device_id: 4124
[1650465528.847021] [ndv4:12741:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib2: disable ODP because it's not supported for DevX QP
[1650465528.847557] [ndv4:13452:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae66e631000..0x2ae66e6b6000 on mlx5_ib3 lkey 0x80500 rkey 0x80500 access 0xf flags 0x3e4
[1650465528.847563] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae66e631018 of 544744 bytes with 128 elements
[1650465528.847567] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465528.847924] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x40da060: adding gid fe80::15:5dff:fd33:fffe to hash on device mlx5_ib3 port 1 index 0)
[1650465528.848083] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x40da060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 1)
[1650465528.848405] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x40da060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 2)
[1650465528.847396] [ndv4:14949:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465528.848636] [ndv4:13552:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465528.848643] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465528.847938] [ndv4:12741:0] async.c:228 UCX DEBUG added async handler 0x26f6ce0 [id=61 ref 1] uct_ib_async_event_handler() to hash
[1650465528.847965] [ndv4:12741:0] async.c:506 UCX DEBUG listening to async event fd 61 events 0x1 mode thread_spinlock
[1650465528.847969] [ndv4:12741:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib2' (InfiniBand channel adapter) with 1 ports
[1650465528.848217] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.848255] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.848261] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.848302] [ndv4:12741:0] ib_md.c:1319 UCX DEBUG mlx5_ib2: using registration cache
[1650465528.849631] [ndv4:12741:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.849637] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.848024] [ndv4:15226:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x3121050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xc07b
[1650465528.848737] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.849032] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.849044] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b178bf9a008 of 151544 bytes with 1052 elements
[1650465528.849803] [ndv4:12741:0] ib_md.c:1604 UCX DEBUG mlx5_ib2: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.850025] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.850345] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.854433] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x40da060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 3)
[1650465528.855118] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x40da060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 4)
[1650465528.855089] [ndv4:14421:0] topo.c:99 UCX DEBUG bus id 0x102000000 doesn't exist. sys_dev = 1
[1650465528.855097] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465528.854480] [ndv4:12741:0] topo.c:99 UCX DEBUG bus id 0x103000000 doesn't exist. sys_dev = 2
[1650465528.854488] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465528.855888] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x40da060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 5)
[1650465528.856352] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x40da060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 6)
[1650465528.856010] [ndv4:15226:0] ib_md.c:812 UCX DEBUG registered memory 0x2b1795600000..0x2b1797c00000 on mlx5_ib2 lkey 0x80b00 rkey 0x80b00 access 0xf flags 0x3e4
[1650465528.856025] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b1795600018 of 39845864 bytes with 4752 elements
[1650465528.856194] [ndv4:15226:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x3121050
[1650465528.856219] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[18]=0x3121050 using dc_mlx5/mlx5_ib2:1 on worker 0x1f938d0
[1650465528.856792] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.856905] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.857009] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.857251] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.858098] [ndv4:13552:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465528.858107] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465528.857855] [ndv4:13443:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465528.857865] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465528.859119] [ndv4:13443:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465528.862676] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465528.864125] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465528.864400] [ndv4:15226:0] ib_iface.c:994 UCX DEBUG iface=0x32fd060: created UD QP 0xc08c on mlx5_ib2:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465528.863306] [ndv4:14421:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465528.863314] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465528.863619] [ndv4:12741:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465528.863626] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465528.862377] [ndv4:13552:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465528.862384] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465528.864171] [ndv4:13552:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465528.861183] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x40da060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 7)
[1650465528.861539] [ndv4:13452:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465528.861551] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x40dae60 [id=109 ref 1] uct_ud_iface_async_handler() to hash
[1650465528.861656] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 109 events 0x5 mode thread_spinlock
[1650465528.861685] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[24]=0x40da060 using ud_verbs/mlx5_ib3:1 on worker 0x264e8d0
[1650465528.862379] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.862825] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.863207] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.863407] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.863982] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465528.863711] [ndv4:14205:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib6 vendor_id: 0x15b3 device_id: 4124
[1650465528.864143] [ndv4:14205:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib6: disable ODP because it's not supported for DevX QP
[1650465528.865095] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465528.865026] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x198a720 [id=69 ref 1] uct_ib_async_event_handler() to hash
[1650465528.865058] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 69 events 0x1 mode thread_spinlock
[1650465528.865063] [ndv4:14205:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib6' (InfiniBand channel adapter) with 1 ports
[1650465528.865778] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.865982] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.866182] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.866217] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.866676] [ndv4:15226:0] ib_md.c:812 UCX DEBUG registered memory 0x2b1797ca1000..0x2b1797d26000 on mlx5_ib2 lkey 0x80c00 rkey 0x80c00 access 0xf flags 0x3e4
[1650465528.866682] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b1797ca1018 of 544744 bytes with 128 elements
[1650465528.866686] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465528.865790] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465528.865982] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465528.865987] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.866034] [ndv4:14205:0] ib_md.c:1319 UCX DEBUG mlx5_ib6: using registration cache
[1650465528.867641] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465528.867343] [ndv4:14421:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465528.867349] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465528.867809] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x32fd060: adding gid fe80::15:5dff:fd33:fffd to hash on device mlx5_ib2 port 1 index 0)
[1650465528.868447] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x32fd060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 1)
[1650465528.867888] [ndv4:14205:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.867895] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.868114] [ndv4:14205:0] ib_md.c:1604 UCX DEBUG mlx5_ib6: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.868453] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465528.868715] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465528.869254] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x32fd060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 2)
[1650465528.869024] [ndv4:13452:0] ib_iface.c:994 UCX DEBUG iface=0x33e6440: created UD QP 0xbf9e on mlx5_ib3:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465528.869033] [ndv4:13452:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465528.870007] [ndv4:14949:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib4 vendor_id: 0x15b3 device_id: 4124
[1650465528.870141] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465528.870560] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.870445] [ndv4:14949:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib4: disable ODP because it's not supported for DevX QP
[1650465528.870820] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.870978] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.871116] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.871516] [ndv4:14949:0] async.c:228 UCX DEBUG added async handler 0x970150 [id=65 ref 1] uct_ib_async_event_handler() to hash
[1650465528.871552] [ndv4:14949:0] async.c:506 UCX DEBUG listening to async event fd 65 events 0x1 mode thread_spinlock
[1650465528.871555] [ndv4:14949:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib4' (InfiniBand channel adapter) with 1 ports
[1650465528.871698] [ndv4:13452:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae66e6b6000..0x2ae66e73b000 on mlx5_ib3 lkey 0x80600 rkey 0x80600 access 0xf flags 0x3e4
[1650465528.871706] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae66e6b6018 of 544744 bytes with 128 elements
[1650465528.871711] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465528.871877] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x33e6440: adding gid fe80::15:5dff:fd33:fffe to hash on device mlx5_ib3 port 1 index 0)
[1650465528.871964] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x33e6440: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 1)
[1650465528.872396] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x33e6440: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 2)
[1650465528.872534] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x33e6440: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 3)
[1650465528.872680] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x33e6440: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 4)
[1650465528.872892] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x33e6440: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 5)
[1650465528.871803] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465528.871942] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465528.871948] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.871999] [ndv4:14949:0] ib_md.c:1319 UCX DEBUG mlx5_ib4: using registration cache
[1650465528.873091] [ndv4:14949:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.873099] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.871844] [ndv4:12741:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465528.871852] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465528.874219] [ndv4:14949:0] ib_md.c:1604 UCX DEBUG mlx5_ib4: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.873799] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x33e6440: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 6)
[1650465528.874413] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x33e6440: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 7)
[1650465528.874420] [ndv4:13452:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib3:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465528.874450] [ndv4:13452:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465528.874455] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x2b461e0 [id=110 ref 1] uct_ud_iface_async_handler() to hash
[1650465528.874481] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 110 events 0x5 mode thread_spinlock
[1650465528.874495] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[25]=0x33e6440 using ud_mlx5/mlx5_ib3:1 on worker 0x264e8d0
[1650465528.875547] [ndv4:12741:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465528.875554] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465528.875039] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465528.875110] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465528.874808] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465528.875112] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465528.875502] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465528.875689] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465528.876321] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465528.878659] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465528.878679] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465528.878683] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465528.878736] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465528.877449] [ndv4:13552:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib3 vendor_id: 0x15b3 device_id: 4124
[1650465528.877864] [ndv4:13552:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib3: disable ODP because it's not supported for DevX QP
[1650465528.877477] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x32fd060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 3)
[1650465528.877866] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x32fd060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 4)
[1650465528.878029] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x32fd060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 5)
[1650465528.878530] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x32fd060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 6)
[1650465528.879875] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465528.879881] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465528.880689] [ndv4:14205:0] topo.c:99 UCX DEBUG bus id 0x107000000 doesn't exist. sys_dev = 6
[1650465528.880696] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465528.881921] [ndv4:13452:0] ib_iface.c:994 UCX DEBUG iface=0x42f3020: created RC QP 0xbf0b on mlx5_ib4:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465528.882002] [ndv4:14421:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465528.882009] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465528.883768] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[26]=0x42f3020 using rc_verbs/mlx5_ib4:1 on worker 0x264e8d0
[1650465528.883867] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465528.883988] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465528.884660] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465528.884745] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465528.885230] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465528.885967] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x32fd060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 7)
[1650465528.886170] [ndv4:13552:0] async.c:228 UCX DEBUG added async handler 0x13cc2b0 [id=63 ref 1] uct_ib_async_event_handler() to hash
[1650465528.886202] [ndv4:13552:0] async.c:506 UCX DEBUG listening to async event fd 63 events 0x1 mode thread_spinlock
[1650465528.886206] [ndv4:13552:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib3' (InfiniBand channel adapter) with 1 ports
[1650465528.886401] [ndv4:15226:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465528.886409] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x287ae70 [id=102 ref 1] uct_ud_iface_async_handler() to hash
[1650465528.886436] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 102 events 0x5 mode thread_spinlock
[1650465528.886449] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[19]=0x32fd060 using ud_verbs/mlx5_ib2:1 on worker 0x1f938d0
[1650465528.886782] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.886876] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.887069] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.887472] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.886919] [ndv4:12741:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465528.886927] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465528.886663] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.886780] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.886786] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.886830] [ndv4:13552:0] ib_md.c:1319 UCX DEBUG mlx5_ib3: using registration cache
[1650465528.887908] [ndv4:13552:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.887915] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.887745] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465528.887774] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465528.887778] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465528.887834] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465528.888128] [ndv4:12741:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465528.888175] [ndv4:13552:0] ib_md.c:1604 UCX DEBUG mlx5_ib3: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.888285] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.888362] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.888803] [ndv4:13443:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib4 vendor_id: 0x15b3 device_id: 4124
[1650465528.889381] [ndv4:13443:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib4: disable ODP because it's not supported for DevX QP
[1650465528.890650] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x4614010 of 8176 bytes with 127 elements
[1650465528.890964] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465528.890972] [ndv4:13452:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib4:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465528.891009] [ndv4:13452:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib4 length=2048) failed: Invalid argument
[1650465528.891013] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465528.891025] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x3b03690 [id=113 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465528.891053] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 113 events 0x1 mode thread_spinlock
[1650465528.891064] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[27]=0x4440030 using rc_mlx5/mlx5_ib4:1 on worker 0x264e8d0
[1650465528.891504] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465528.891741] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465528.892074] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465528.892354] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465528.890921] [ndv4:14205:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465528.890929] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465528.890898] [ndv4:13443:0] async.c:228 UCX DEBUG added async handler 0x1d43280 [id=65 ref 1] uct_ib_async_event_handler() to hash
[1650465528.890937] [ndv4:13443:0] async.c:506 UCX DEBUG listening to async event fd 65 events 0x1 mode thread_spinlock
[1650465528.890942] [ndv4:13443:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib4' (InfiniBand channel adapter) with 1 ports
[1650465528.892125] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465528.892355] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465528.892366] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.892435] [ndv4:13443:0] ib_md.c:1319 UCX DEBUG mlx5_ib4: using registration cache
[1650465528.892553] [ndv4:14949:0] topo.c:99 UCX DEBUG bus id 0x105000000 doesn't exist. sys_dev = 4
[1650465528.892563] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465528.892556] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465528.893870] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465528.894165] [ndv4:15226:0] ib_iface.c:994 UCX DEBUG iface=0x341b050: created UD QP 0xc0ad on mlx5_ib2:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465528.894173] [ndv4:15226:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465528.894416] [ndv4:13443:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.894425] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.894684] [ndv4:13443:0] ib_md.c:1604 UCX DEBUG mlx5_ib4: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.894825] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465528.894880] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465528.895614] [ndv4:14421:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465528.895622] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465528.895807] [ndv4:14205:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465528.895813] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465528.894858] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465528.895455] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.895668] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.895826] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.896445] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.896884] [ndv4:15226:0] ib_md.c:812 UCX DEBUG registered memory 0x2b1797d26000..0x2b1797dab000 on mlx5_ib2 lkey 0x80d00 rkey 0x80d00 access 0xf flags 0x3e4
[1650465528.896890] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b1797d26018 of 544744 bytes with 128 elements
[1650465528.896895] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465528.897232] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x341b050: adding gid fe80::15:5dff:fd33:fffd to hash on device mlx5_ib2 port 1 index 0)
[1650465528.897536] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x341b050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 1)
[1650465528.897679] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x341b050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 2)
[1650465528.897746] [ndv4:14421:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465528.898117] [ndv4:14949:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465528.898124] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465528.899919] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465528.899914] [ndv4:13443:0] topo.c:99 UCX DEBUG bus id 0x105000000 doesn't exist. sys_dev = 4
[1650465528.899921] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465528.905294] [ndv4:13552:0] topo.c:99 UCX DEBUG bus id 0x104000000 doesn't exist. sys_dev = 3
[1650465528.905305] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465528.904477] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x341b050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 3)
[1650465528.904983] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x341b050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 4)
[1650465528.905262] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x341b050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 5)
[1650465528.905404] [ndv4:14205:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465528.905411] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465528.908706] [ndv4:13443:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465528.908715] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465528.909434] [ndv4:13552:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465528.909441] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465528.910452] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x341b050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 6)
[1650465528.910513] [ndv4:14205:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465528.910520] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465528.912112] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465528.912131] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465528.912135] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465528.912192] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465528.912316] [ndv4:12741:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib3 vendor_id: 0x15b3 device_id: 4124
[1650465528.912468] [ndv4:14205:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465528.912643] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465528.912650] [ndv4:13452:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib4:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465528.912684] [ndv4:13452:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib4 length=2048) failed: Invalid argument
[1650465528.912688] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465528.912697] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x4448f60 [id=115 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465528.912720] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 115 events 0x1 mode thread_spinlock
[1650465528.912936] [ndv4:12741:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib3: disable ODP because it's not supported for DevX QP
[1650465528.913365] [ndv4:13452:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465528.913456] [ndv4:13552:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465528.913462] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465528.913617] [ndv4:12741:0] async.c:228 UCX DEBUG added async handler 0x26ee2b0 [id=63 ref 1] uct_ib_async_event_handler() to hash
[1650465528.913655] [ndv4:12741:0] async.c:506 UCX DEBUG listening to async event fd 63 events 0x1 mode thread_spinlock
[1650465528.913658] [ndv4:12741:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib3' (InfiniBand channel adapter) with 1 ports
[1650465528.913859] [ndv4:13443:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465528.913865] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465528.913933] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.914541] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.914546] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.914642] [ndv4:12741:0] ib_md.c:1319 UCX DEBUG mlx5_ib3: using registration cache
[1650465528.914793] [ndv4:14421:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib2 vendor_id: 0x15b3 device_id: 4124
[1650465528.915253] [ndv4:14421:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib2: disable ODP because it's not supported for DevX QP
[1650465528.915812] [ndv4:14421:0] async.c:228 UCX DEBUG added async handler 0xf33cb0 [id=61 ref 1] uct_ib_async_event_handler() to hash
[1650465528.915836] [ndv4:14421:0] async.c:506 UCX DEBUG listening to async event fd 61 events 0x1 mode thread_spinlock
[1650465528.915839] [ndv4:14421:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib2' (InfiniBand channel adapter) with 1 ports
[1650465528.916420] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.916555] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.916560] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.916675] [ndv4:14421:0] ib_md.c:1319 UCX DEBUG mlx5_ib2: using registration cache
[1650465528.917864] [ndv4:14421:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.917871] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.918110] [ndv4:14421:0] ib_md.c:1604 UCX DEBUG mlx5_ib2: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.918224] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465528.918296] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465528.918892] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x341b050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 7)
[1650465528.918902] [ndv4:15226:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib2:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465528.918936] [ndv4:15226:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465528.918941] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x2e06c60 [id=103 ref 1] uct_ud_iface_async_handler() to hash
[1650465528.918965] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 103 events 0x5 mode thread_spinlock
[1650465528.918976] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[20]=0x341b050 using ud_mlx5/mlx5_ib2:1 on worker 0x1f938d0
[1650465528.919195] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.919383] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.919808] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.919888] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.921262] [ndv4:12741:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.921270] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.921473] [ndv4:12741:0] ib_md.c:1604 UCX DEBUG mlx5_ib3: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.920495] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465528.922127] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465528.922143] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465528.922146] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465528.922198] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465528.922672] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465528.922678] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465528.923414] [ndv4:14949:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465528.923422] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465528.923504] [ndv4:13552:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465528.923511] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465528.923175] [ndv4:15226:0] ib_iface.c:994 UCX DEBUG iface=0x351b0a0: created RC QP 0xbfc6 on mlx5_ib3:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465528.924279] [ndv4:13452:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x4616050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xbf64
[1650465528.925148] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465528.925220] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465528.925229] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2ae670f3c008 of 151544 bytes with 1052 elements
[1650465528.924438] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[21]=0x351b0a0 using rc_verbs/mlx5_ib3:1 on worker 0x1f938d0
[1650465528.924787] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.924952] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.925182] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.925205] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.925512] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465528.924815] [ndv4:14205:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib7 vendor_id: 0x15b3 device_id: 4124
[1650465528.925190] [ndv4:14205:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib7: disable ODP because it's not supported for DevX QP
[1650465528.926318] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x2070cd0 [id=71 ref 1] uct_ib_async_event_handler() to hash
[1650465528.926349] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 71 events 0x1 mode thread_spinlock
[1650465528.926352] [ndv4:14205:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib7' (InfiniBand channel adapter) with 1 ports
[1650465528.924315] [ndv4:13443:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465528.924326] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465528.926238] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.926314] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.927151] [ndv4:13552:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465528.927158] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465528.928876] [ndv4:13552:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465528.928774] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465528.928784] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465528.928787] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465528.928841] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465528.929206] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x3669010 of 8176 bytes with 127 elements
[1650465528.929049] [ndv4:13452:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae66e800000..0x2ae670e00000 on mlx5_ib4 lkey 0x80800 rkey 0x80800 access 0xf flags 0x3e4
[1650465528.929070] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2ae66e800018 of 39845864 bytes with 4752 elements
[1650465528.929209] [ndv4:13452:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x4616050
[1650465528.929246] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[28]=0x4616050 using dc_mlx5/mlx5_ib4:1 on worker 0x264e8d0
[1650465528.929947] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465528.930074] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465528.930240] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465528.930435] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465528.929421] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465528.929427] [ndv4:15226:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib3:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465528.929460] [ndv4:15226:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib3 length=2048) failed: Invalid argument
[1650465528.929464] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465528.929473] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x248b3b0 [id=106 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465528.929499] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 106 events 0x1 mode thread_spinlock
[1650465528.929509] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[22]=0x3670100 using rc_mlx5/mlx5_ib3:1 on worker 0x1f938d0
[1650465528.929679] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.929757] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.929934] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.930061] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.930979] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465528.932045] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465528.932108] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465528.932114] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.932161] [ndv4:14205:0] ib_md.c:1319 UCX DEBUG mlx5_ib7: using registration cache
[1650465528.932177] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465528.932538] [ndv4:13452:0] ib_iface.c:994 UCX DEBUG iface=0x47f21f0: created UD QP 0xbf91 on mlx5_ib4:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465528.933240] [ndv4:14205:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.933247] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.933509] [ndv4:14205:0] ib_md.c:1604 UCX DEBUG mlx5_ib7: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.933379] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465528.933030] [ndv4:14421:0] topo.c:99 UCX DEBUG bus id 0x103000000 doesn't exist. sys_dev = 2
[1650465528.933040] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465528.934680] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465528.934736] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465528.933865] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465528.934017] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465528.934132] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465528.934716] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465528.935324] [ndv4:13452:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae670f61000..0x2ae670fe6000 on mlx5_ib4 lkey 0x80900 rkey 0x80900 access 0xf flags 0x3e4
[1650465528.935329] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae670f61018 of 544744 bytes with 128 elements
[1650465528.935334] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465528.935704] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x47f21f0: adding gid fe80::15:5dff:fd33:ffff to hash on device mlx5_ib4 port 1 index 0)
[1650465528.935697] [ndv4:12741:0] topo.c:99 UCX DEBUG bus id 0x104000000 doesn't exist. sys_dev = 3
[1650465528.935705] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465528.934871] [ndv4:14949:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465528.934879] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465528.936674] [ndv4:14421:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465528.936680] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465528.940563] [ndv4:13443:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465528.940572] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465528.939486] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x47f21f0: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 1)
[1650465528.938789] [ndv4:14205:0] topo.c:99 UCX DEBUG bus id 0x108000000 doesn't exist. sys_dev = 7
[1650465528.938795] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465528.940858] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465528.941985] [ndv4:13443:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465528.942183] [ndv4:12741:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465528.942191] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465528.942775] [ndv4:14421:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465528.942782] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465528.943124] [ndv4:13552:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib4 vendor_id: 0x15b3 device_id: 4124
[1650465528.943664] [ndv4:13552:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib4: disable ODP because it's not supported for DevX QP
[1650465528.944322] [ndv4:13552:0] async.c:228 UCX DEBUG added async handler 0x1ab5bf0 [id=65 ref 1] uct_ib_async_event_handler() to hash
[1650465528.944349] [ndv4:13552:0] async.c:506 UCX DEBUG listening to async event fd 65 events 0x1 mode thread_spinlock
[1650465528.944352] [ndv4:13552:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib4' (InfiniBand channel adapter) with 1 ports
[1650465528.944151] [ndv4:14949:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465528.944159] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465528.944781] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465528.944853] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465528.944859] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.944904] [ndv4:13552:0] ib_md.c:1319 UCX DEBUG mlx5_ib4: using registration cache
[1650465528.945669] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x47f21f0: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 2)
[1650465528.946062] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x47f21f0: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 3)
[1650465528.946171] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x47f21f0: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 4)
[1650465528.946305] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x47f21f0: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 5)
[1650465528.946480] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x47f21f0: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 6)
[1650465528.946704] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x47f21f0: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 7)
[1650465528.945866] [ndv4:14949:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465528.947028] [ndv4:13452:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465528.947036] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x42cea70 [id=116 ref 1] uct_ud_iface_async_handler() to hash
[1650465528.947070] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 116 events 0x5 mode thread_spinlock
[1650465528.947083] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[29]=0x47f21f0 using ud_verbs/mlx5_ib4:1 on worker 0x264e8d0
[1650465528.947326] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465528.947412] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465528.947770] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465528.948052] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465528.948297] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465528.949287] [ndv4:14205:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465528.949296] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465528.949642] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465528.951100] [ndv4:13452:0] ib_iface.c:994 UCX DEBUG iface=0x4910050: created UD QP 0xbfba on mlx5_ib4:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465528.951108] [ndv4:13452:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465528.951880] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465528.951899] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465528.951904] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465528.951960] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465528.952170] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465528.953278] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465528.953479] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465528.953777] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465528.953875] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465528.954242] [ndv4:13452:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae670fe6000..0x2ae67106b000 on mlx5_ib4 lkey 0x80a00 rkey 0x80a00 access 0xf flags 0x3e4
[1650465528.954247] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae670fe6018 of 544744 bytes with 128 elements
[1650465528.954252] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465528.955145] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x4910050: adding gid fe80::15:5dff:fd33:ffff to hash on device mlx5_ib4 port 1 index 0)
[1650465528.956383] [ndv4:12741:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465528.956393] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465528.955795] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x4910050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 1)
[1650465528.954754] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465528.954764] [ndv4:15226:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib3:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465528.954793] [ndv4:15226:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib3 length=2048) failed: Invalid argument
[1650465528.954797] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465528.954806] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x16b7f00 [id=108 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465528.954834] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 108 events 0x1 mode thread_spinlock
[1650465528.955626] [ndv4:15226:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465528.955491] [ndv4:13552:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.955501] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.955733] [ndv4:13552:0] ib_md.c:1604 UCX DEBUG mlx5_ib4: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.956092] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465528.956656] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465528.960928] [ndv4:14949:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib5 vendor_id: 0x15b3 device_id: 4124
[1650465528.961318] [ndv4:14949:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib5: disable ODP because it's not supported for DevX QP
[1650465528.963983] [ndv4:14205:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465528.963992] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465528.964416] [ndv4:14421:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465528.964425] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465528.964170] [ndv4:15226:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x3843010: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xbff4
[1650465528.964691] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x4910050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 2)
[1650465528.968876] [ndv4:13443:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib5 vendor_id: 0x15b3 device_id: 4124
[1650465528.969405] [ndv4:13443:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib5: disable ODP because it's not supported for DevX QP
[1650465528.969880] [ndv4:13552:0] topo.c:99 UCX DEBUG bus id 0x105000000 doesn't exist. sys_dev = 4
[1650465528.969888] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465528.970416] [ndv4:12741:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465528.970424] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465528.970363] [ndv4:13443:0] async.c:228 UCX DEBUG added async handler 0x1d41f90 [id=67 ref 1] uct_ib_async_event_handler() to hash
[1650465528.970395] [ndv4:13443:0] async.c:506 UCX DEBUG listening to async event fd 67 events 0x1 mode thread_spinlock
[1650465528.970399] [ndv4:13443:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib5' (InfiniBand channel adapter) with 1 ports
[1650465528.971446] [ndv4:14949:0] async.c:228 UCX DEBUG added async handler 0x970710 [id=67 ref 1] uct_ib_async_event_handler() to hash
[1650465528.971480] [ndv4:14949:0] async.c:506 UCX DEBUG listening to async event fd 67 events 0x1 mode thread_spinlock
[1650465528.971484] [ndv4:14949:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib5' (InfiniBand channel adapter) with 1 ports
[1650465528.971523] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465528.971769] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465528.971779] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.971845] [ndv4:13443:0] ib_md.c:1319 UCX DEBUG mlx5_ib5: using registration cache
[1650465528.972072] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465528.972265] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465528.972271] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465528.972336] [ndv4:14949:0] ib_md.c:1319 UCX DEBUG mlx5_ib5: using registration cache
[1650465528.972012] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.972555] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.972566] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b178bfc1008 of 151544 bytes with 1052 elements
[1650465528.972330] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x4910050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 3)
[1650465528.972741] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x4910050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 4)
[1650465528.972911] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x4910050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 5)
[1650465528.973199] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x4910050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 6)
[1650465528.973448] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x4910050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 7)
[1650465528.973455] [ndv4:13452:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib4:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465528.973488] [ndv4:13452:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465528.973493] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x42ce410 [id=117 ref 1] uct_ud_iface_async_handler() to hash
[1650465528.973516] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 117 events 0x5 mode thread_spinlock
[1650465528.973528] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[30]=0x4910050 using ud_mlx5/mlx5_ib4:1 on worker 0x264e8d0
[1650465528.973681] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465528.973754] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465528.974223] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465528.974251] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465528.973984] [ndv4:12741:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465528.973991] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465528.975040] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib5:1
[1650465528.974016] [ndv4:14949:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.974023] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.974271] [ndv4:14949:0] ib_md.c:1604 UCX DEBUG mlx5_ib5: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.974783] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465528.974999] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465528.975740] [ndv4:12741:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465528.976208] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465528.976224] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465528.976227] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465528.976287] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465528.976547] [ndv4:14421:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465528.976557] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465528.977452] [ndv4:14421:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465528.976567] [ndv4:15226:0] ib_md.c:812 UCX DEBUG registered memory 0x2b1797e00000..0x2b179a400000 on mlx5_ib3 lkey 0x80700 rkey 0x80700 access 0xf flags 0x3e4
[1650465528.976633] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b1797e00018 of 39845864 bytes with 4752 elements
[1650465528.976775] [ndv4:15226:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x3843010
[1650465528.976812] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[23]=0x3843010 using dc_mlx5/mlx5_ib3:1 on worker 0x1f938d0
[1650465528.977257] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.977392] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.977437] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465528.977443] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465528.978187] [ndv4:13452:0] ib_iface.c:994 UCX DEBUG iface=0x4a100a0: created RC QP 0xbde6 on mlx5_ib5:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465528.980935] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[31]=0x4a100a0 using rc_verbs/mlx5_ib5:1 on worker 0x264e8d0
[1650465528.981267] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465528.981442] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465528.981886] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465528.982186] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465528.984641] [ndv4:14949:0] topo.c:99 UCX DEBUG bus id 0x106000000 doesn't exist. sys_dev = 5
[1650465528.984650] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465528.985548] [ndv4:13443:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465528.985560] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465528.984344] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.984377] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.984837] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465528.985750] [ndv4:13552:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465528.985759] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465528.986256] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465528.986572] [ndv4:15226:0] ib_iface.c:994 UCX DEBUG iface=0x3a1f060: created UD QP 0xc009 on mlx5_ib3:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465528.987444] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465528.988027] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.988434] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.988696] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465528.988826] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465528.986667] [ndv4:13443:0] ib_md.c:1604 UCX DEBUG mlx5_ib5: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465528.986778] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465528.987032] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465528.988744] [ndv4:14949:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465528.988750] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465528.989426] [ndv4:14205:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465528.989438] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465528.990758] [ndv4:15226:0] ib_md.c:812 UCX DEBUG registered memory 0x2b179a5ac000..0x2b179a631000 on mlx5_ib3 lkey 0x80800 rkey 0x80800 access 0xf flags 0x3e4
[1650465528.990772] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b179a5ac018 of 544744 bytes with 128 elements
[1650465528.990781] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465528.991005] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x3a1f060: adding gid fe80::15:5dff:fd33:fffe to hash on device mlx5_ib3 port 1 index 0)
[1650465528.991443] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x3a1f060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 1)
[1650465528.990902] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib5:1
[1650465528.991377] [ndv4:13443:0] topo.c:99 UCX DEBUG bus id 0x106000000 doesn't exist. sys_dev = 5
[1650465528.991384] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465528.997102] [ndv4:13552:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465528.997111] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465528.997982] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x3a1f060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 2)
[1650465528.998236] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x3a1f060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 3)
[1650465528.998285] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465528.998303] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465528.998307] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465528.998357] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465528.998914] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x3a1f060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 4)
[1650465528.999291] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x3a1f060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 5)
[1650465528.998864] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x4d31010 of 8176 bytes with 127 elements
[1650465528.999104] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465528.999112] [ndv4:13452:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib5:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465528.999152] [ndv4:13452:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib5 length=2048) failed: Invalid argument
[1650465528.999156] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465528.999169] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x3d23e10 [id=120 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465528.999201] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 120 events 0x1 mode thread_spinlock
[1650465528.999213] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[32]=0x4b5d030 using rc_mlx5/mlx5_ib5:1 on worker 0x264e8d0
[1650465528.999289] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465528.999545] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465528.999286] [ndv4:14949:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465528.999294] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465529.000090] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x3a1f060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 6)
[1650465528.999852] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.000264] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.000524] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib5:1
[1650465529.002038] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.002052] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.002055] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.002110] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.002555] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465529.002561] [ndv4:13452:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib5:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.002641] [ndv4:13452:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib5 length=2048) failed: Invalid argument
[1650465529.002645] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.002653] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x3d23f80 [id=122 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.002681] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 122 events 0x1 mode thread_spinlock
[1650465529.003307] [ndv4:13452:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465529.003641] [ndv4:14205:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465529.003648] [ndv4:14205:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465529.004181] [ndv4:14205:0] ucp_context.c:1556 UCX DEBUG created ucp context 0x1973dc0 0x1973dc0 [13 mds 47 tls] features 0x1 tl bitmap 0x7fffffffffff 0x0
[1650465529.007651] [ndv4:12741:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib4 vendor_id: 0x15b3 device_id: 4124
[1650465529.007109] [ndv4:13443:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465529.007118] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465529.008236] [ndv4:12741:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib4: disable ODP because it's not supported for DevX QP
[1650465529.008953] [ndv4:14421:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib3 vendor_id: 0x15b3 device_id: 4124
[1650465529.009644] [ndv4:12741:0] async.c:228 UCX DEBUG added async handler 0x2dd7bf0 [id=65 ref 1] uct_ib_async_event_handler() to hash
[1650465529.009674] [ndv4:12741:0] async.c:506 UCX DEBUG listening to async event fd 65 events 0x1 mode thread_spinlock
[1650465529.009678] [ndv4:12741:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib4' (InfiniBand channel adapter) with 1 ports
[1650465529.009658] [ndv4:14421:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib3: disable ODP because it's not supported for DevX QP
[1650465529.010645] [ndv4:14421:0] async.c:228 UCX DEBUG added async handler 0xf2b280 [id=63 ref 1] uct_ib_async_event_handler() to hash
[1650465529.010679] [ndv4:14421:0] async.c:506 UCX DEBUG listening to async event fd 63 events 0x1 mode thread_spinlock
[1650465529.010683] [ndv4:14421:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib3' (InfiniBand channel adapter) with 1 ports
[1650465529.011724] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.011797] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.011806] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465529.011873] [ndv4:14421:0] ib_md.c:1319 UCX DEBUG mlx5_ib3: using registration cache
[1650465529.011889] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x3a1f060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 7)
[1650465529.012304] [ndv4:15226:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.012316] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x3a1fe60 [id=109 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.012355] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 109 events 0x5 mode thread_spinlock
[1650465529.012378] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[24]=0x3a1f060 using ud_verbs/mlx5_ib3:1 on worker 0x1f938d0
[1650465529.012742] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.012978] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.012238] [ndv4:13552:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465529.012250] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465529.013690] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.014409] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.013843] [ndv4:14949:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465529.013851] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465529.014020] [ndv4:14421:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465529.014028] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465529.014246] [ndv4:14421:0] ib_md.c:1604 UCX DEBUG mlx5_ib3: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465529.013801] [ndv4:13452:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x4d33050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xbe32
[1650465529.014871] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465529.014878] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.015108] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.015121] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2ae67386c008 of 151544 bytes with 1052 elements
[1650465529.015910] [ndv4:13552:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465529.015917] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465529.017244] [ndv4:13552:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465529.018138] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.018546] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.018552] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465529.018679] [ndv4:12741:0] ib_md.c:1319 UCX DEBUG mlx5_ib4: using registration cache
[1650465529.018858] [ndv4:13452:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae671200000..0x2ae673800000 on mlx5_ib5 lkey 0x80600 rkey 0x80600 access 0xf flags 0x3e4
[1650465529.018882] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2ae671200018 of 39845864 bytes with 4752 elements
[1650465529.019027] [ndv4:13452:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x4d33050
[1650465529.019061] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[33]=0x4d33050 using dc_mlx5/mlx5_ib5:1 on worker 0x264e8d0
[1650465529.019767] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.019822] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.020568] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.020703] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.021184] [ndv4:12741:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465529.021192] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465529.022116] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.022355] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.022034] [ndv4:13443:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465529.022043] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465529.021927] [ndv4:12741:0] ib_md.c:1604 UCX DEBUG mlx5_ib4: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465529.022264] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.022360] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.022899] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib5:1
[1650465529.025079] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.025766] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.025887] [ndv4:13452:0] ib_iface.c:994 UCX DEBUG iface=0x4f0f060: created UD QP 0xbe5c on mlx5_ib5:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.026163] [ndv4:15226:0] ib_iface.c:994 UCX DEBUG iface=0x2d2b440: created UD QP 0xc032 on mlx5_ib3:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.026170] [ndv4:15226:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465529.026690] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.027211] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.028107] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.028489] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.027121] [ndv4:13443:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465529.027128] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465529.028930] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.029312] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.029633] [ndv4:12741:0] topo.c:99 UCX DEBUG bus id 0x105000000 doesn't exist. sys_dev = 4
[1650465529.029639] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465529.029793] [ndv4:15226:0] ib_md.c:812 UCX DEBUG registered memory 0x2b179a631000..0x2b179a6b6000 on mlx5_ib3 lkey 0x80900 rkey 0x80900 access 0xf flags 0x3e4
[1650465529.029800] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b179a631018 of 544744 bytes with 128 elements
[1650465529.029804] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.030261] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x2d2b440: adding gid fe80::15:5dff:fd33:fffe to hash on device mlx5_ib3 port 1 index 0)
[1650465529.034684] [ndv4:13443:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465529.034691] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465529.035959] [ndv4:13552:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib5 vendor_id: 0x15b3 device_id: 4124
[1650465529.036443] [ndv4:13552:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib5: disable ODP because it's not supported for DevX QP
[1650465529.036707] [ndv4:13443:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465529.035816] [ndv4:14949:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465529.035827] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465529.036895] [ndv4:14949:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465529.037371] [ndv4:13552:0] async.c:228 UCX DEBUG added async handler 0x13d0710 [id=67 ref 1] uct_ib_async_event_handler() to hash
[1650465529.037402] [ndv4:13552:0] async.c:506 UCX DEBUG listening to async event fd 67 events 0x1 mode thread_spinlock
[1650465529.037405] [ndv4:13552:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib5' (InfiniBand channel adapter) with 1 ports
[1650465529.037942] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.037982] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.037987] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465529.038043] [ndv4:13552:0] ib_md.c:1319 UCX DEBUG mlx5_ib5: using registration cache
[1650465529.038958] [ndv4:13552:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465529.038966] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465529.039193] [ndv4:13552:0] ib_md.c:1604 UCX DEBUG mlx5_ib5: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465529.039682] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.039802] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.041259] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.041400] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.041489] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.041880] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.042371] [ndv4:13452:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae673891000..0x2ae673916000 on mlx5_ib5 lkey 0x80700 rkey 0x80700 access 0xf flags 0x3e4
[1650465529.042377] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae673891018 of 544744 bytes with 128 elements
[1650465529.042381] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.043030] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x4f0f060: adding gid fe80::15:5dff:fd34:0 to hash on device mlx5_ib5 port 1 index 0)
[1650465529.043508] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x4f0f060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 1)
[1650465529.043110] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x2d2b440: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 1)
[1650465529.043335] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x2d2b440: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 2)
[1650465529.044207] [ndv4:13552:0] topo.c:99 UCX DEBUG bus id 0x106000000 doesn't exist. sys_dev = 5
[1650465529.044214] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465529.048756] [ndv4:14421:0] topo.c:99 UCX DEBUG bus id 0x104000000 doesn't exist. sys_dev = 3
[1650465529.048765] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465529.051177] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x4f0f060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 2)
[1650465529.051420] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x4f0f060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 3)
[1650465529.051163] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x2d2b440: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 3)
[1650465529.051754] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x2d2b440: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 4)
[1650465529.052067] [ndv4:12741:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465529.052074] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465529.051863] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x4f0f060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 4)
[1650465529.052987] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x2d2b440: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 5)
[1650465529.053251] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x2d2b440: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 6)
[1650465529.052915] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x4f0f060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 5)
[1650465529.053248] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x4f0f060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 6)
[1650465529.053562] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x4f0f060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 7)
[1650465529.053146] [ndv4:13443:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib6 vendor_id: 0x15b3 device_id: 4124
[1650465529.053546] [ndv4:13443:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib6: disable ODP because it's not supported for DevX QP
[1650465529.053979] [ndv4:13452:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.053989] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x49eb970 [id=123 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.054017] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 123 events 0x5 mode thread_spinlock
[1650465529.054030] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[34]=0x4f0f060 using ud_verbs/mlx5_ib5:1 on worker 0x264e8d0
[1650465529.054297] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.054818] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.054992] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.055550] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.055005] [ndv4:13443:0] async.c:228 UCX DEBUG added async handler 0x1d41f40 [id=69 ref 1] uct_ib_async_event_handler() to hash
[1650465529.055033] [ndv4:13443:0] async.c:506 UCX DEBUG listening to async event fd 69 events 0x1 mode thread_spinlock
[1650465529.055037] [ndv4:13443:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib6' (InfiniBand channel adapter) with 1 ports
[1650465529.055643] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.056038] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.056044] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465529.056092] [ndv4:13443:0] ib_md.c:1319 UCX DEBUG mlx5_ib6: using registration cache
[1650465529.055831] [ndv4:13552:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465529.055841] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465529.056298] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib5:1
[1650465529.057167] [ndv4:13443:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465529.057175] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465529.057435] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.057420] [ndv4:13443:0] ib_md.c:1604 UCX DEBUG mlx5_ib6: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465529.057622] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.058069] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.060293] [ndv4:13452:0] ib_iface.c:994 UCX DEBUG iface=0x502d460: created UD QP 0xbe9f on mlx5_ib5:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.060301] [ndv4:13452:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465529.061436] [ndv4:14421:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465529.061445] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465529.061070] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x2d2b440: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 7)
[1650465529.061082] [ndv4:15226:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib3:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.061123] [ndv4:15226:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.061128] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x248b1e0 [id=110 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.061153] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 110 events 0x5 mode thread_spinlock
[1650465529.061167] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[25]=0x2d2b440 using ud_mlx5/mlx5_ib3:1 on worker 0x1f938d0
[1650465529.061373] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.061461] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.061907] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.062700] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.062787] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.062836] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.062869] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.063316] [ndv4:13452:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae673916000..0x2ae67399b000 on mlx5_ib5 lkey 0x80800 rkey 0x80800 access 0xf flags 0x3e4
[1650465529.063322] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae673916018 of 544744 bytes with 128 elements
[1650465529.063326] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.062026] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.062511] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.062938] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465529.065001] [ndv4:13443:0] topo.c:99 UCX DEBUG bus id 0x107000000 doesn't exist. sys_dev = 6
[1650465529.065008] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465529.065491] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465529.065506] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.065510] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.065560] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.064749] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x502d460: adding gid fe80::15:5dff:fd34:0 to hash on device mlx5_ib5 port 1 index 0)
[1650465529.065291] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x502d460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 1)
[1650465529.065493] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x502d460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 2)
[1650465529.065813] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x502d460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 3)
[1650465529.065955] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x502d460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 4)
[1650465529.066531] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x502d460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 5)
[1650465529.066635] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.066642] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465529.066927] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x502d460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 6)
[1650465529.067169] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x502d460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 7)
[1650465529.067177] [ndv4:13452:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib5:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.067210] [ndv4:13452:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.067215] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x502df60 [id=124 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.067248] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 124 events 0x5 mode thread_spinlock
[1650465529.067259] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[35]=0x502d460 using ud_mlx5/mlx5_ib5:1 on worker 0x264e8d0
[1650465529.067818] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.067912] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.067214] [ndv4:15226:0] ib_iface.c:994 UCX DEBUG iface=0x3c38020: created RC QP 0xc025 on mlx5_ib4:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465529.068007] [ndv4:14421:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465529.068014] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465529.068682] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[26]=0x3c38020 using rc_verbs/mlx5_ib4:1 on worker 0x1f938d0
[1650465529.068876] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.069361] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.070181] [ndv4:13443:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465529.070188] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465529.074960] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.075055] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.075867] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib6:1
[1650465529.076549] [ndv4:12741:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465529.076556] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465529.076391] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.076421] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.077097] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465529.077651] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465529.077669] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.077672] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.077727] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.077284] [ndv4:13552:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465529.077292] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465529.078624] [ndv4:14421:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465529.078631] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465529.079083] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.079097] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.079101] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.079155] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.079569] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x3f59010 of 8176 bytes with 127 elements
[1650465529.078797] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.078803] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465529.079313] [ndv4:13452:0] ib_iface.c:994 UCX DEBUG iface=0x512d0a0: created RC QP 0xbdaa on mlx5_ib6:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465529.080171] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[36]=0x512d0a0 using rc_verbs/mlx5_ib6:1 on worker 0x264e8d0
[1650465529.080319] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.080636] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.081106] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.082342] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.079949] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.079956] [ndv4:15226:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib4:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.079990] [ndv4:15226:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib4 length=2048) failed: Invalid argument
[1650465529.079994] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.080005] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x3448690 [id=113 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.080039] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 113 events 0x1 mode thread_spinlock
[1650465529.080049] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[27]=0x3d85030 using rc_mlx5/mlx5_ib4:1 on worker 0x1f938d0
[1650465529.080272] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.080491] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.080908] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.081327] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.082560] [ndv4:13443:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465529.082630] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465529.082327] [ndv4:14949:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib6 vendor_id: 0x15b3 device_id: 4124
[1650465529.083802] [ndv4:14949:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib6: disable ODP because it's not supported for DevX QP
[1650465529.083595] [ndv4:13552:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465529.083601] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465529.085073] [ndv4:14421:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465529.085080] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465529.087689] [ndv4:14421:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465529.087398] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465529.088413] [ndv4:13552:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465529.088420] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465529.088335] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib6:1
[1650465529.089833] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.089850] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.089853] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.089926] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.090029] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.090058] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.090063] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.090122] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.090482] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465529.090490] [ndv4:15226:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib4:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.090524] [ndv4:15226:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib4 length=2048) failed: Invalid argument
[1650465529.090527] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.090536] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x3d8df60 [id=115 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.090566] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 115 events 0x1 mode thread_spinlock
[1650465529.090871] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x544e010 of 8176 bytes with 127 elements
[1650465529.090839] [ndv4:13552:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465529.091738] [ndv4:14949:0] async.c:228 UCX DEBUG added async handler 0x10532d0 [id=69 ref 1] uct_ib_async_event_handler() to hash
[1650465529.091774] [ndv4:14949:0] async.c:506 UCX DEBUG listening to async event fd 69 events 0x1 mode thread_spinlock
[1650465529.091778] [ndv4:14949:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib6' (InfiniBand channel adapter) with 1 ports
[1650465529.092256] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.092410] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.092416] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465529.092469] [ndv4:14949:0] ib_md.c:1319 UCX DEBUG mlx5_ib6: using registration cache
[1650465529.091182] [ndv4:15226:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465529.091198] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.091209] [ndv4:13452:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib6:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.091251] [ndv4:13452:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib6 length=2048) failed: Invalid argument
[1650465529.091255] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.091270] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x5282c00 [id=127 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.091300] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 127 events 0x1 mode thread_spinlock
[1650465529.091318] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[37]=0x527a030 using rc_mlx5/mlx5_ib6:1 on worker 0x264e8d0
[1650465529.091652] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.091808] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.092274] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.092775] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.093834] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib6:1
[1650465529.094181] [ndv4:14205:0] mm_posix.c:518 UCX DEBUG allocated posix shared memory at 0x2afda3df3000 length 12288
[1650465529.094273] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool mm_recv_desc: align 64, maxelems 4294967295, elemsize 8368
[1650465529.095016] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.095033] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.095036] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.095093] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.095528] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465529.095535] [ndv4:13452:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib6:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.095571] [ndv4:13452:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib6 length=2048) failed: Invalid argument
[1650465529.095625] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.095633] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x49eb330 [id=129 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.095660] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 129 events 0x1 mode thread_spinlock
[1650465529.096292] [ndv4:13452:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465529.096113] [ndv4:14205:0] mm_posix.c:305 UCX DEBUG shared memory mmap(addr=(nil), length=6291456, flags= HUGETLB, fd=76) failed: Invalid argument
[1650465529.096122] [ndv4:14205:0] mm_posix.c:518 UCX DEBUG allocated posix shared memory at 0x2afda9c64000 length 4296704
[1650465529.096128] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool mm_recv_desc: allocated chunk 0x2afda9c64018 of 4296680 bytes with 512 elements
[1650465529.096382] [ndv4:14205:0] mm_iface.c:600 UCX DEBUG created mm iface 0x20f3b20 FIFO id 0x4000000046425bb9 va 0x2afda3df3000 size 12288 (128 x 64 elems)
[1650465529.096457] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[0]=0x20f3b20 using posix/memory on worker 0x29ca770
[1650465529.096482] [ndv4:14205:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 8447 bytes with hugetlb
[1650465529.096525] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool mm_recv_desc: align 64, maxelems 4294967295, elemsize 8368
[1650465529.096542] [ndv4:14205:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 4292720 bytes with hugetlb
[1650465529.096551] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool mm_recv_desc: allocated chunk 0x2afdaa07d018 of 4296680 bytes with 512 elements
[1650465529.097204] [ndv4:14205:0] mm_iface.c:600 UCX DEBUG created mm iface 0x2105870 FIFO id 0x64802a va 0x2afda3df6000 size 12288 (128 x 64 elems)
[1650465529.097215] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[1]=0x2105870 using sysv/memory on worker 0x29ca770
[1650465529.097227] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool self_msg_desc: align 64, maxelems 4294967295, elemsize 8200
[1650465529.097230] [ndv4:14205:0] self.c:220 UCX DEBUG created self iface id 0xb55d03b1f3198649 send_size 8192
[1650465529.097236] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[2]=0x2104f10 using self/memory0 on worker 0x29ca770
[1650465529.097259] [ndv4:14205:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465529.097264] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465529.097267] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465529.100141] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x2112bf0 [id=78 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465529.100184] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 78 events 0x5 mode thread_spinlock
[1650465529.100203] [ndv4:14205:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x2105e40: listening for connections (fd=78) on 10.5.0.5:35403
[1650465529.100509] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[3]=0x2105e40 using tcp/eth0 on worker 0x29ca770
[1650465529.100529] [ndv4:14205:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465529.100532] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465529.100535] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465529.100664] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x2061700 [id=80 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465529.100687] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 80 events 0x5 mode thread_spinlock
[1650465529.100691] [ndv4:14205:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x20ed5c0: listening for connections (fd=80) on 127.0.0.1:52878
[1650465529.100709] [ndv4:14205:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465529.100715] [ndv4:14205:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465529.100823] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[4]=0x20ed5c0 using tcp/lo on worker 0x29ca770
[1650465529.100842] [ndv4:14205:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465529.100845] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465529.100848] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465529.100886] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x2060fe0 [id=82 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465529.100921] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 82 events 0x5 mode thread_spinlock
[1650465529.100925] [ndv4:14205:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x20edc40: listening for connections (fd=82) on 172.16.1.242:55307
[1650465529.101334] [ndv4:14949:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465529.101343] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465529.101546] [ndv4:14949:0] ib_md.c:1604 UCX DEBUG mlx5_ib6: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465529.101783] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.102051] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.101288] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[5]=0x20edc40 using tcp/ib0 on worker 0x29ca770
[1650465529.101435] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.101520] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.101778] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.101873] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.101967] [ndv4:15226:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x3f5b050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xc061
[1650465529.102810] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.102831] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.102842] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b179ceb7008 of 151544 bytes with 1052 elements
[1650465529.102815] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465529.102809] [ndv4:13443:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465529.102819] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465529.104132] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465529.104169] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.104172] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.104185] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.104843] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.104852] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465529.106058] [ndv4:13452:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x5450050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xbdd7
[1650465529.106571] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.106882] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.106896] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2ae67619c008 of 151544 bytes with 1052 elements
[1650465529.106892] [ndv4:14949:0] topo.c:99 UCX DEBUG bus id 0x107000000 doesn't exist. sys_dev = 6
[1650465529.106899] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465529.106070] [ndv4:14205:0] ib_iface.c:994 UCX DEBUG iface=0x20fb6c0: created RC QP 0xde53 on mlx5_ib0:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465529.106751] [ndv4:15226:0] ib_md.c:812 UCX DEBUG registered memory 0x2b179a800000..0x2b179ce00000 on mlx5_ib4 lkey 0x80b00 rkey 0x80b00 access 0xf flags 0x3e4
[1650465529.106765] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b179a800018 of 39845864 bytes with 4752 elements
[1650465529.106902] [ndv4:15226:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x3f5b050
[1650465529.106937] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[28]=0x3f5b050 using dc_mlx5/mlx5_ib4:1 on worker 0x1f938d0
[1650465529.107298] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.107393] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.107660] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.107790] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.108801] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465529.110738] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.111111] [ndv4:13452:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae673a00000..0x2ae676000000 on mlx5_ib6 lkey 0x80400 rkey 0x80400 access 0xf flags 0x3e4
[1650465529.111127] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2ae673a00018 of 39845864 bytes with 4752 elements
[1650465529.111272] [ndv4:13452:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x5450050
[1650465529.111305] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[38]=0x5450050 using dc_mlx5/mlx5_ib6:1 on worker 0x264e8d0
[1650465529.112161] [ndv4:15226:0] ib_iface.c:994 UCX DEBUG iface=0x41371f0: created UD QP 0xc07d on mlx5_ib4:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.112170] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.112283] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.112283] [ndv4:13552:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib6 vendor_id: 0x15b3 device_id: 4124
[1650465529.112922] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.112969] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.112980] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.113069] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.113078] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.112894] [ndv4:13552:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib6: disable ODP because it's not supported for DevX QP
[1650465529.113756] [ndv4:15226:0] ib_md.c:812 UCX DEBUG registered memory 0x2b179cedc000..0x2b179cf61000 on mlx5_ib4 lkey 0x80c00 rkey 0x80c00 access 0xf flags 0x3e4
[1650465529.113763] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b179cedc018 of 544744 bytes with 128 elements
[1650465529.113767] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.114220] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x41371f0: adding gid fe80::15:5dff:fd33:ffff to hash on device mlx5_ib4 port 1 index 0)
[1650465529.114528] [ndv4:13443:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465529.114536] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465529.114748] [ndv4:13552:0] async.c:228 UCX DEBUG added async handler 0x13d0560 [id=69 ref 1] uct_ib_async_event_handler() to hash
[1650465529.114782] [ndv4:13552:0] async.c:506 UCX DEBUG listening to async event fd 69 events 0x1 mode thread_spinlock
[1650465529.114786] [ndv4:13552:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib6' (InfiniBand channel adapter) with 1 ports
[1650465529.115653] [ndv4:12741:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465529.115667] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465529.115802] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.116041] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.116047] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465529.116105] [ndv4:13552:0] ib_md.c:1319 UCX DEBUG mlx5_ib6: using registration cache
[1650465529.114804] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x41371f0: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 1)
[1650465529.117038] [ndv4:13443:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465529.116990] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[6]=0x20fb6c0 using rc_verbs/mlx5_ib0:1 on worker 0x29ca770
[1650465529.117127] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.117684] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.117916] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.118346] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.118671] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465529.121109] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.121216] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.121735] [ndv4:14421:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib4 vendor_id: 0x15b3 device_id: 4124
[1650465529.122301] [ndv4:14421:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib4: disable ODP because it's not supported for DevX QP
[1650465529.121811] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib6:1
[1650465529.122967] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x41371f0: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 2)
[1650465529.123307] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x41371f0: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 3)
[1650465529.122983] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.123354] [ndv4:13452:0] ib_iface.c:994 UCX DEBUG iface=0x562c2e0: created UD QP 0xbde1 on mlx5_ib6:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.123310] [ndv4:14421:0] async.c:228 UCX DEBUG added async handler 0x1617280 [id=65 ref 1] uct_ib_async_event_handler() to hash
[1650465529.123337] [ndv4:14421:0] async.c:506 UCX DEBUG listening to async event fd 65 events 0x1 mode thread_spinlock
[1650465529.123341] [ndv4:14421:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib4' (InfiniBand channel adapter) with 1 ports
[1650465529.124051] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.124489] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.124770] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.125036] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.125181] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.124411] [ndv4:13552:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465529.124420] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465529.124695] [ndv4:13552:0] ib_md.c:1604 UCX DEBUG mlx5_ib6: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465529.125016] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.125180] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.123890] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x41371f0: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 4)
[1650465529.124827] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x41371f0: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 5)
[1650465529.125150] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x41371f0: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 6)
[1650465529.123835] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.124410] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.124415] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465529.124461] [ndv4:14421:0] ib_md.c:1319 UCX DEBUG mlx5_ib4: using registration cache
[1650465529.125777] [ndv4:13452:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae6761c1000..0x2ae676246000 on mlx5_ib6 lkey 0x80500 rkey 0x80500 access 0xf flags 0x3e4
[1650465529.125782] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae6761c1018 of 544744 bytes with 128 elements
[1650465529.125787] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.125847] [ndv4:14421:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465529.125854] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465529.126131] [ndv4:14421:0] ib_md.c:1604 UCX DEBUG mlx5_ib4: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465529.126382] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.126791] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.127901] [ndv4:14205:0] ib_device.c:1394 UCX DEBUG max IB CQE size is 128
[1650465529.129054] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.129063] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.129066] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.129119] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.129482] [ndv4:12741:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465529.129490] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465529.130173] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x2a94010 of 8176 bytes with 127 elements
[1650465529.130801] [ndv4:14421:0] topo.c:99 UCX DEBUG bus id 0x105000000 doesn't exist. sys_dev = 4
[1650465529.130807] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465529.130526] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.130553] [ndv4:14205:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib0:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.130658] [ndv4:14205:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650465529.130665] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.131557] [ndv4:13443:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib7 vendor_id: 0x15b3 device_id: 4124
[1650465529.131922] [ndv4:13443:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib7: disable ODP because it's not supported for DevX QP
[1650465529.132520] [ndv4:12741:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465529.132540] [ndv4:13443:0] async.c:228 UCX DEBUG added async handler 0x1d43820 [id=71 ref 1] uct_ib_async_event_handler() to hash
[1650465529.132671] [ndv4:13443:0] async.c:506 UCX DEBUG listening to async event fd 71 events 0x1 mode thread_spinlock
[1650465529.132677] [ndv4:13443:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib7' (InfiniBand channel adapter) with 1 ports
[1650465529.132988] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.133131] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.133137] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465529.133197] [ndv4:13443:0] ib_md.c:1319 UCX DEBUG mlx5_ib7: using registration cache
[1650465529.138435] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x205f380 [id=85 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.138468] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 85 events 0x1 mode thread_spinlock
[1650465529.138488] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[7]=0x2109110 using rc_mlx5/mlx5_ib0:1 on worker 0x29ca770
[1650465529.139346] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.139528] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.139834] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.139977] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.140360] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465529.139973] [ndv4:14949:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465529.139980] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465529.141684] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.141693] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.141696] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.141709] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.142138] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465529.142145] [ndv4:14205:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib0:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.142177] [ndv4:14205:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650465529.142181] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.142190] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x205fe00 [id=87 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.142217] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 87 events 0x1 mode thread_spinlock
[1650465529.142673] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x41371f0: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 7)
[1650465529.143651] [ndv4:15226:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.143664] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x3c13a70 [id=116 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.143695] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 116 events 0x5 mode thread_spinlock
[1650465529.143714] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[29]=0x41371f0 using ud_verbs/mlx5_ib4:1 on worker 0x1f938d0
[1650465529.143912] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.144111] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.144446] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.144538] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.142747] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x562c2e0: adding gid fe80::15:5dff:fd34:1 to hash on device mlx5_ib6 port 1 index 0)
[1650465529.143124] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x562c2e0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 1)
[1650465529.143376] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x562c2e0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 2)
[1650465529.143520] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x562c2e0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 3)
[1650465529.143785] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x562c2e0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 4)
[1650465529.143918] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x562c2e0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 5)
[1650465529.144124] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x562c2e0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 6)
[1650465529.144453] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x562c2e0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 7)
[1650465529.142987] [ndv4:14205:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465529.144789] [ndv4:13452:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.144797] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x51088d0 [id=130 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.144824] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 130 events 0x5 mode thread_spinlock
[1650465529.144837] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[39]=0x562c2e0 using ud_verbs/mlx5_ib6:1 on worker 0x264e8d0
[1650465529.145065] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.145075] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.145203] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.145661] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.145444] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465529.146516] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib6:1
[1650465529.146556] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.146955] [ndv4:15226:0] ib_iface.c:994 UCX DEBUG iface=0x4255050: created UD QP 0xc0c7 on mlx5_ib4:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.146963] [ndv4:15226:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465529.148656] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.149041] [ndv4:13452:0] ib_iface.c:994 UCX DEBUG iface=0x574a460: created UD QP 0xbe1d on mlx5_ib6:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.149048] [ndv4:13452:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465529.149953] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.147840] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.148688] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.149079] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.149523] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.150246] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.148943] [ndv4:13443:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465529.148953] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465529.149206] [ndv4:13443:0] ib_md.c:1604 UCX DEBUG mlx5_ib7: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465529.149456] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.150247] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.150937] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.151379] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.151453] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.151670] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.150730] [ndv4:15226:0] ib_md.c:812 UCX DEBUG registered memory 0x2b179cf61000..0x2b179cfe6000 on mlx5_ib4 lkey 0x80d00 rkey 0x80d00 access 0xf flags 0x3e4
[1650465529.150736] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b179cf61018 of 544744 bytes with 128 elements
[1650465529.150740] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.152271] [ndv4:13452:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae676246000..0x2ae6762cb000 on mlx5_ib6 lkey 0x80600 rkey 0x80600 access 0xf flags 0x3e4
[1650465529.152279] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae676246018 of 544744 bytes with 128 elements
[1650465529.152285] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.152999] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x574a460: adding gid fe80::15:5dff:fd34:1 to hash on device mlx5_ib6 port 1 index 0)
[1650465529.153035] [ndv4:14949:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465529.153042] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465529.152235] [ndv4:14205:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x2cfe010: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xde76
[1650465529.153941] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.153999] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.154023] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2afda3dfb008 of 151544 bytes with 1052 elements
[1650465529.155021] [ndv4:12741:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib5 vendor_id: 0x15b3 device_id: 4124
[1650465529.155674] [ndv4:12741:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib5: disable ODP because it's not supported for DevX QP
[1650465529.157067] [ndv4:12741:0] async.c:228 UCX DEBUG added async handler 0x26f2710 [id=67 ref 1] uct_ib_async_event_handler() to hash
[1650465529.157099] [ndv4:12741:0] async.c:506 UCX DEBUG listening to async event fd 67 events 0x1 mode thread_spinlock
[1650465529.157103] [ndv4:12741:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib5' (InfiniBand channel adapter) with 1 ports
[1650465529.157404] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.157970] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.157976] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465529.158047] [ndv4:12741:0] ib_md.c:1319 UCX DEBUG mlx5_ib5: using registration cache
[1650465529.158054] [ndv4:14205:0] ib_md.c:812 UCX DEBUG registered memory 0x2afdaa600000..0x2afdacc00000 on mlx5_ib0 lkey 0x80c00 rkey 0x80c00 access 0xf flags 0x3e4
[1650465529.158074] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2afdaa600018 of 39845864 bytes with 4752 elements
[1650465529.158213] [ndv4:14205:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x2cfe010
[1650465529.158249] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[8]=0x2cfe010 using dc_mlx5/mlx5_ib0:1 on worker 0x29ca770
[1650465529.158381] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.158716] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.158957] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.159051] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.159723] [ndv4:12741:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465529.159729] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465529.158652] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x574a460: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 1)
[1650465529.159253] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x574a460: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 2)
[1650465529.158221] [ndv4:14949:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465529.158228] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465529.160012] [ndv4:12741:0] ib_md.c:1604 UCX DEBUG mlx5_ib5: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465529.160273] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.161021] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.162960] [ndv4:13443:0] topo.c:99 UCX DEBUG bus id 0x108000000 doesn't exist. sys_dev = 7
[1650465529.162968] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465529.164570] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x574a460: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 3)
[1650465529.165327] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x574a460: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 4)
[1650465529.164885] [ndv4:12741:0] topo.c:99 UCX DEBUG bus id 0x106000000 doesn't exist. sys_dev = 5
[1650465529.164892] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465529.163273] [ndv4:14421:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465529.163283] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465529.164751] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465529.166302] [ndv4:14949:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465529.166309] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465529.168556] [ndv4:14949:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465529.168887] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x4255050: adding gid fe80::15:5dff:fd33:ffff to hash on device mlx5_ib4 port 1 index 0)
[1650465529.169421] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x4255050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 1)
[1650465529.170363] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x4255050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 2)
[1650465529.171160] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x4255050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 3)
[1650465529.171283] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x4255050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 4)
[1650465529.172172] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.172397] [ndv4:13552:0] topo.c:99 UCX DEBUG bus id 0x107000000 doesn't exist. sys_dev = 6
[1650465529.172407] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465529.172704] [ndv4:14205:0] ib_iface.c:994 UCX DEBUG iface=0x29ea5f0: created UD QP 0xde5c on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.172468] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x4255050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 5)
[1650465529.173399] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.173923] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.174567] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.173303] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x4255050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 6)
[1650465529.175106] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.175190] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.176096] [ndv4:13443:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465529.176106] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465529.176458] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x574a460: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 5)
[1650465529.176737] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x574a460: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 6)
[1650465529.176320] [ndv4:14421:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465529.176329] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465529.175755] [ndv4:14205:0] ib_md.c:812 UCX DEBUG registered memory 0x2afda3e20000..0x2afda3ea5000 on mlx5_ib0 lkey 0x80d00 rkey 0x80d00 access 0xf flags 0x3e4
[1650465529.175763] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2afda3e20018 of 544744 bytes with 128 elements
[1650465529.175767] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.176207] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x29ea5f0: adding gid fe80::15:5dff:fd33:fffb to hash on device mlx5_ib0 port 1 index 0)
[1650465529.176726] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x29ea5f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650465529.177175] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x29ea5f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650465529.177501] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x29ea5f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650465529.178089] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x29ea5f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650465529.178190] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x29ea5f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650465529.178460] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x29ea5f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650465529.179347] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x4255050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 7)
[1650465529.179359] [ndv4:15226:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib4:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.179407] [ndv4:15226:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.179412] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x3c13410 [id=117 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.179438] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 117 events 0x5 mode thread_spinlock
[1650465529.179453] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[30]=0x4255050 using ud_mlx5/mlx5_ib4:1 on worker 0x1f938d0
[1650465529.179748] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.180141] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.180773] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.180838] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.181200] [ndv4:14421:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465529.181207] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465529.181293] [ndv4:12741:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465529.181303] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465529.181366] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib5:1
[1650465529.182060] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x574a460: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 7)
[1650465529.182070] [ndv4:13452:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib6:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.182166] [ndv4:13452:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.182172] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x5108a40 [id=131 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.182205] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 131 events 0x5 mode thread_spinlock
[1650465529.182218] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[40]=0x574a460 using ud_mlx5/mlx5_ib6:1 on worker 0x264e8d0
[1650465529.182518] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465529.182536] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.182540] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.182638] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.183031] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.183364] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.183790] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.183868] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.184301] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib7:1
[1650465529.183194] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.183203] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465529.183715] [ndv4:15226:0] ib_iface.c:994 UCX DEBUG iface=0x43550a0: created RC QP 0xbfb1 on mlx5_ib5:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465529.185497] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465529.185512] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.185515] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.185566] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.185792] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[31]=0x43550a0 using rc_verbs/mlx5_ib5:1 on worker 0x1f938d0
[1650465529.185984] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.186216] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.185096] [ndv4:14421:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465529.185161] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465529.187171] [ndv4:14421:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465529.187837] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x29ea5f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650465529.188442] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.188451] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465529.188255] [ndv4:14205:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.189046] [ndv4:13452:0] ib_iface.c:994 UCX DEBUG iface=0x584a0a0: created RC QP 0xbd08 on mlx5_ib7:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465529.190127] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[41]=0x584a0a0 using rc_verbs/mlx5_ib7:1 on worker 0x264e8d0
[1650465529.190415] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.190663] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.190904] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.191249] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.194301] [ndv4:13552:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465529.194311] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465529.194408] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x28fc370 [id=88 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.194447] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 88 events 0x5 mode thread_spinlock
[1650465529.194472] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[9]=0x29ea5f0 using ud_verbs/mlx5_ib0:1 on worker 0x29ca770
[1650465529.194823] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.195159] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.195543] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.195648] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.195250] [ndv4:13443:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465529.195261] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465529.195699] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.196816] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.197709] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465529.199390] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib7:1
[1650465529.199963] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.200527] [ndv4:14205:0] ib_iface.c:994 UCX DEBUG iface=0x2111a80: created UD QP 0xde5d on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.200542] [ndv4:14205:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465529.203292] [ndv4:14421:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib5 vendor_id: 0x15b3 device_id: 4124
[1650465529.203187] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.203485] [ndv4:13443:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465529.203493] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465529.203865] [ndv4:14421:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib5: disable ODP because it's not supported for DevX QP
[1650465529.204261] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.204293] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.204296] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.204314] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.204318] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.204372] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.204738] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.204967] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.204987] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x5b6b010 of 8176 bytes with 127 elements
[1650465529.205203] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib5:1
[1650465529.205004] [ndv4:14421:0] async.c:228 UCX DEBUG added async handler 0xf2f710 [id=67 ref 1] uct_ib_async_event_handler() to hash
[1650465529.205035] [ndv4:14421:0] async.c:506 UCX DEBUG listening to async event fd 67 events 0x1 mode thread_spinlock
[1650465529.205039] [ndv4:14421:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib5' (InfiniBand channel adapter) with 1 ports
[1650465529.205696] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.205882] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.205887] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465529.205936] [ndv4:14421:0] ib_md.c:1319 UCX DEBUG mlx5_ib5: using registration cache
[1650465529.207055] [ndv4:14421:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465529.207062] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465529.207318] [ndv4:14421:0] ib_md.c:1604 UCX DEBUG mlx5_ib5: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465529.207638] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.208263] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.205648] [ndv4:14205:0] ib_md.c:812 UCX DEBUG registered memory 0x2afda3ea5000..0x2afda3f2a000 on mlx5_ib0 lkey 0x80e00 rkey 0x80e00 access 0xf flags 0x3e4
[1650465529.205656] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2afda3ea5018 of 544744 bytes with 128 elements
[1650465529.205660] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.206636] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.206646] [ndv4:13452:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib7:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.206687] [ndv4:13452:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib7 length=2048) failed: Invalid argument
[1650465529.206692] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.206706] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x57778f0 [id=134 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.206738] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 134 events 0x1 mode thread_spinlock
[1650465529.206753] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[42]=0x5997030 using rc_mlx5/mlx5_ib7:1 on worker 0x264e8d0
[1650465529.207042] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.207468] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.207964] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.208286] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.206146] [ndv4:14949:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib7 vendor_id: 0x15b3 device_id: 4124
[1650465529.206552] [ndv4:14949:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib7: disable ODP because it's not supported for DevX QP
[1650465529.207815] [ndv4:14949:0] async.c:228 UCX DEBUG added async handler 0x95cfa0 [id=71 ref 1] uct_ib_async_event_handler() to hash
[1650465529.207840] [ndv4:14949:0] async.c:506 UCX DEBUG listening to async event fd 71 events 0x1 mode thread_spinlock
[1650465529.207843] [ndv4:14949:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib7' (InfiniBand channel adapter) with 1 ports
[1650465529.208355] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.208390] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.208395] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465529.208442] [ndv4:14949:0] ib_md.c:1319 UCX DEBUG mlx5_ib7: using registration cache
[1650465529.205297] [ndv4:12741:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465529.205307] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465529.208999] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib7:1
[1650465529.209388] [ndv4:14949:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465529.209395] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465529.209609] [ndv4:14949:0] ib_md.c:1604 UCX DEBUG mlx5_ib7: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465529.210274] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.210291] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.210295] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.210351] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.210832] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465529.210841] [ndv4:13452:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib7:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.210877] [ndv4:13452:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib7 length=2048) failed: Invalid argument
[1650465529.210880] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.210889] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x1d97df0 [id=136 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.210916] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 136 events 0x1 mode thread_spinlock
[1650465529.211549] [ndv4:13452:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465529.210721] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.210907] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.213560] [ndv4:14421:0] topo.c:99 UCX DEBUG bus id 0x106000000 doesn't exist. sys_dev = 5
[1650465529.213567] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465529.215739] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x2111a80: adding gid fe80::15:5dff:fd33:fffb to hash on device mlx5_ib0 port 1 index 0)
[1650465529.216172] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x2111a80: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650465529.216723] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.216744] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.216748] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.216801] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.217252] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x4676010 of 8176 bytes with 127 elements
[1650465529.218170] [ndv4:13552:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465529.218236] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465529.217513] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.217521] [ndv4:15226:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib5:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.217563] [ndv4:15226:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib5 length=2048) failed: Invalid argument
[1650465529.217567] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.217632] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x3668e10 [id=120 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.217644] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 120 events 0x1 mode thread_spinlock
[1650465529.217661] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[32]=0x44a2030 using rc_mlx5/mlx5_ib5:1 on worker 0x1f938d0
[1650465529.217926] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.218149] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.218934] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.219253] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.220698] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib5:1
[1650465529.219607] [ndv4:12741:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465529.219618] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465529.218822] [ndv4:13443:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465529.218829] [ndv4:13443:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465529.219346] [ndv4:13443:0] ucp_context.c:1556 UCX DEBUG created ucp context 0x1d28d80 0x1d28d80 [13 mds 47 tls] features 0x1 tl bitmap 0x7fffffffffff 0x0
[1650465529.221076] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x2111a80: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650465529.222102] [ndv4:14949:0] topo.c:99 UCX DEBUG bus id 0x108000000 doesn't exist. sys_dev = 7
[1650465529.222111] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465529.224006] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.224023] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.224026] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.224082] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.224228] [ndv4:13452:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x5b6d050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xbd3d
[1650465529.225021] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465529.225028] [ndv4:15226:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib5:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.225063] [ndv4:15226:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib5 length=2048) failed: Invalid argument
[1650465529.225066] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.225074] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x3668f80 [id=122 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.225103] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 122 events 0x1 mode thread_spinlock
[1650465529.225696] [ndv4:15226:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465529.226032] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.226264] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.226284] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2ae678acc008 of 151544 bytes with 1052 elements
[1650465529.228164] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x2111a80: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650465529.228709] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x2111a80: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650465529.229418] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x2111a80: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650465529.229174] [ndv4:13552:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465529.229184] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465529.229241] [ndv4:12741:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465529.229249] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465529.230207] [ndv4:13452:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae676400000..0x2ae678a00000 on mlx5_ib7 lkey 0x80a00 rkey 0x80a00 access 0xf flags 0x3e4
[1650465529.230231] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2ae676400018 of 39845864 bytes with 4752 elements
[1650465529.230371] [ndv4:13452:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x5b6d050
[1650465529.230407] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[43]=0x5b6d050 using dc_mlx5/mlx5_ib7:1 on worker 0x264e8d0
[1650465529.230467] [ndv4:14949:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465529.230478] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465529.230954] [ndv4:12741:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465529.230862] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.230932] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.231515] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.231893] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.232498] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib7:1
[1650465529.231994] [ndv4:14421:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465529.232003] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465529.233874] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.234203] [ndv4:13452:0] ib_iface.c:994 UCX DEBUG iface=0x5d49060: created UD QP 0xbd48 on mlx5_ib7:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.234897] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.235432] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.235507] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.236079] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.235272] [ndv4:15226:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x4678050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xc01e
[1650465529.235472] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.235779] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.235924] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b179f7e7008 of 151544 bytes with 1052 elements
[1650465529.237135] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x2111a80: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650465529.237048] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.237550] [ndv4:13452:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae678af1000..0x2ae678b76000 on mlx5_ib7 lkey 0x80b00 rkey 0x80b00 access 0xf flags 0x3e4
[1650465529.237557] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae678af1018 of 544744 bytes with 128 elements
[1650465529.237561] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.238044] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x5d49060: adding gid fe80::15:5dff:fd34:2 to hash on device mlx5_ib7 port 1 index 0)
[1650465529.238075] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x5d49060: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 1)
[1650465529.238241] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x5d49060: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 2)
[1650465529.237890] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x2111a80: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650465529.237899] [ndv4:14205:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib0:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.238015] [ndv4:14205:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.238021] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x2051df0 [id=89 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.238048] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 89 events 0x5 mode thread_spinlock
[1650465529.238072] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[10]=0x2111a80 using ud_mlx5/mlx5_ib0:1 on worker 0x29ca770
[1650465529.238477] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.239289] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.239719] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.239825] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.237526] [ndv4:14421:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465529.237532] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465529.238662] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x5d49060: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 3)
[1650465529.239259] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x5d49060: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 4)
[1650465529.240093] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x5d49060: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 5)
[1650465529.240568] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x5d49060: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 6)
[1650465529.240351] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465529.241817] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465529.241830] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.241833] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.241856] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.242664] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.242671] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465529.239893] [ndv4:15226:0] ib_md.c:812 UCX DEBUG registered memory 0x2b179d000000..0x2b179f600000 on mlx5_ib5 lkey 0x80900 rkey 0x80900 access 0xf flags 0x3e4
[1650465529.239915] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b179d000018 of 39845864 bytes with 4752 elements
[1650465529.240055] [ndv4:15226:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x4678050
[1650465529.240089] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[33]=0x4678050 using dc_mlx5/mlx5_ib5:1 on worker 0x1f938d0
[1650465529.240413] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.240753] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.241147] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.241175] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.242136] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib5:1
[1650465529.241124] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x5d49060: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 7)
[1650465529.241491] [ndv4:13452:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.241498] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x5d49f40 [id=137 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.241534] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 137 events 0x5 mode thread_spinlock
[1650465529.241546] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[44]=0x5d49060 using ud_verbs/mlx5_ib7:1 on worker 0x264e8d0
[1650465529.241874] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.242645] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.242881] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.243009] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.243307] [ndv4:13452:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib7:1
[1650465529.244714] [ndv4:13452:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.245031] [ndv4:13452:0] ib_iface.c:994 UCX DEBUG iface=0x33b97b0: created UD QP 0xbd58 on mlx5_ib7:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.245038] [ndv4:13452:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465529.243738] [ndv4:14205:0] ib_iface.c:994 UCX DEBUG iface=0x329b050: created RC QP 0xc450 on mlx5_ib1:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465529.246223] [ndv4:14949:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465529.246234] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465529.245700] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.246401] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.246474] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.245964] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[11]=0x329b050 using rc_verbs/mlx5_ib1:1 on worker 0x29ca770
[1650465529.246389] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.246737] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.247027] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.247268] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.246823] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.246908] [ndv4:13452:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.247316] [ndv4:13452:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae678b76000..0x2ae678bfb000 on mlx5_ib7 lkey 0x80c00 rkey 0x80c00 access 0xf flags 0x3e4
[1650465529.247321] [ndv4:13452:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae678b76018 of 544744 bytes with 128 elements
[1650465529.247326] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.247767] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x33b97b0: adding gid fe80::15:5dff:fd34:2 to hash on device mlx5_ib7 port 1 index 0)
[1650465529.248155] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x33b97b0: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 1)
[1650465529.248343] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x33b97b0: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 2)
[1650465529.248822] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x33b97b0: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 3)
[1650465529.248986] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x33b97b0: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 4)
[1650465529.249538] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x33b97b0: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 5)
[1650465529.250027] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x33b97b0: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 6)
[1650465529.251564] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.251982] [ndv4:15226:0] ib_iface.c:994 UCX DEBUG iface=0x4854060: created UD QP 0xc055 on mlx5_ib5:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.252680] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.253029] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.253260] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.253711] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.254217] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.255163] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465529.254779] [ndv4:15226:0] ib_md.c:812 UCX DEBUG registered memory 0x2b179f80c000..0x2b179f891000 on mlx5_ib5 lkey 0x80a00 rkey 0x80a00 access 0xf flags 0x3e4
[1650465529.254785] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b179f80c018 of 544744 bytes with 128 elements
[1650465529.254790] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.255426] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x4854060: adding gid fe80::15:5dff:fd34:0 to hash on device mlx5_ib5 port 1 index 0)
[1650465529.254698] [ndv4:14421:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465529.254705] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465529.255198] [ndv4:13452:0] ud_iface.c:393 UCX DEBUG iface 0x33b97b0: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 7)
[1650465529.255205] [ndv4:13452:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib7:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.255234] [ndv4:13452:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.255238] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x1d759b0 [id=138 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.255269] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 138 events 0x5 mode thread_spinlock
[1650465529.255281] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[45]=0x33b97b0 using ud_mlx5/mlx5_ib7:1 on worker 0x264e8d0
[1650465529.255346] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool uct_scopy_iface_tx_mp: align 64, maxelems 4294967295, elemsize 736
[1650465529.256567] [ndv4:13452:0] ucp_worker.c:1159 UCX DEBUG created interface[46]=0x5e67880 using cma/memory on worker 0x264e8d0
[1650465529.256715] [ndv4:13452:0] ucp_worker.c:982 UCX DEBUG selected scalable tl bitmap: 0x7fffffffffff 0x0 (47 tls)
[1650465529.256685] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.256696] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.256699] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.256722] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.257312] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x32ae010 of 8176 bytes with 127 elements
[1650465529.257679] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.257686] [ndv4:14205:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib1:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.257720] [ndv4:14205:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650465529.257724] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.257733] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x20f8460 [id=92 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.257765] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 92 events 0x1 mode thread_spinlock
[1650465529.257775] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[12]=0x311c010 using rc_mlx5/mlx5_ib1:1 on worker 0x29ca770
[1650465529.258078] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.258215] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.260354] [ndv4:13552:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465529.260367] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465529.259953] [ndv4:14421:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465529.259960] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465529.261642] [ndv4:14421:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465529.262510] [ndv4:13552:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465529.264323] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.264452] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.266402] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x58257a0 [id=75 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.266434] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 75 events 0x0 mode thread_spinlock
[1650465529.266486] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x58255d0 [id=76 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.266532] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 76 events 0x0 mode thread_spinlock
[1650465529.266551] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x5825610 [id=77 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.266558] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 77 events 0x0 mode thread_spinlock
[1650465529.265835] [ndv4:14949:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465529.265845] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465529.265473] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465529.267259] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.267277] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.267281] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.267344] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.267850] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465529.267858] [ndv4:14205:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib1:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.267890] [ndv4:14205:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650465529.267895] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.267903] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x20f7fa0 [id=94 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.267928] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 94 events 0x1 mode thread_spinlock
[1650465529.268667] [ndv4:14205:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465529.268650] [ndv4:12741:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib6 vendor_id: 0x15b3 device_id: 4124
[1650465529.269213] [ndv4:12741:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib6: disable ODP because it's not supported for DevX QP
[1650465529.270265] [ndv4:12741:0] async.c:228 UCX DEBUG added async handler 0x26f2560 [id=69 ref 1] uct_ib_async_event_handler() to hash
[1650465529.270303] [ndv4:12741:0] async.c:506 UCX DEBUG listening to async event fd 69 events 0x1 mode thread_spinlock
[1650465529.270308] [ndv4:12741:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib6' (InfiniBand channel adapter) with 1 ports
[1650465529.269751] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x4854060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 1)
[1650465529.270759] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x4854060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 2)
[1650465529.271064] [ndv4:14949:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465529.271070] [ndv4:14949:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465529.271335] [ndv4:14949:0] ucp_context.c:1556 UCX DEBUG created ucp context 0x955cd0 0x955cd0 [13 mds 47 tls] features 0x1 tl bitmap 0x7fffffffffff 0x0
[1650465529.271700] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x33b9600 [id=79 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.271739] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 79 events 0x0 mode thread_spinlock
[1650465529.271920] [ndv4:13452:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465529.271927] [ndv4:13452:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465529.272372] [ndv4:13452:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465529.272381] [ndv4:13452:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465529.272788] [ndv4:13452:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465529.272795] [ndv4:13452:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465529.271889] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x4854060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 3)
[1650465529.272816] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x4854060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 4)
[1650465529.273458] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x4854060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 5)
[1650465529.270873] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.271011] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.271017] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465529.271090] [ndv4:12741:0] ib_md.c:1319 UCX DEBUG mlx5_ib6: using registration cache
[1650465529.272244] [ndv4:12741:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465529.272252] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465529.272511] [ndv4:12741:0] ib_md.c:1604 UCX DEBUG mlx5_ib6: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465529.272895] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.273535] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.273749] [ndv4:13452:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465529.273756] [ndv4:13452:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465529.274202] [ndv4:13452:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465529.274207] [ndv4:13452:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465529.274816] [ndv4:13452:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465529.274824] [ndv4:13452:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465529.275140] [ndv4:13452:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465529.275145] [ndv4:13452:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465529.275310] [ndv4:13452:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465529.275315] [ndv4:13452:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465529.275386] [ndv4:13452:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465529.275391] [ndv4:13452:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465529.275833] [ndv4:13452:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465529.275840] [ndv4:13452:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465529.275950] [ndv4:13452:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465529.275955] [ndv4:13452:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465529.276233] [ndv4:13452:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465529.276240] [ndv4:13452:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465529.276925] [ndv4:13452:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465529.276933] [ndv4:13452:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465529.277207] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x33b9640 [id=81 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.277242] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 81 events 0x0 mode thread_spinlock
[1650465529.280262] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x4854060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 6)
[1650465529.280567] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x4854060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 7)
[1650465529.281021] [ndv4:15226:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.281030] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x4330970 [id=123 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.281068] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 123 events 0x5 mode thread_spinlock
[1650465529.281079] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[34]=0x4854060 using ud_verbs/mlx5_ib5:1 on worker 0x1f938d0
[1650465529.280827] [ndv4:14205:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x3446010: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xc473
[1650465529.281804] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.282303] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.282401] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.282436] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.283246] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib5:1
[1650465529.282741] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.283256] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.283268] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2afda3f2c008 of 151544 bytes with 1052 elements
[1650465529.284376] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.284713] [ndv4:15226:0] ib_iface.c:994 UCX DEBUG iface=0x4972460: created UD QP 0xc084 on mlx5_ib5:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.284721] [ndv4:15226:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465529.285064] [ndv4:12741:0] topo.c:99 UCX DEBUG bus id 0x107000000 doesn't exist. sys_dev = 6
[1650465529.285073] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465529.285367] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.285974] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.286243] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.286540] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.286736] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.288516] [ndv4:13552:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib7 vendor_id: 0x15b3 device_id: 4124
[1650465529.289161] [ndv4:13552:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib7: disable ODP because it's not supported for DevX QP
[1650465529.287858] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x33b9680 [id=83 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.287895] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 83 events 0x0 mode thread_spinlock
[1650465529.287982] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x33b96c0 [id=84 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.288004] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 84 events 0x0 mode thread_spinlock
[1650465529.288024] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x33b9730 [id=86 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.288045] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 86 events 0x0 mode thread_spinlock
[1650465529.288094] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x33b9770 [id=90 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.288113] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 90 events 0x0 mode thread_spinlock
[1650465529.288129] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x40daec0 [id=91 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.288177] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 91 events 0x0 mode thread_spinlock
[1650465529.288192] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x40daf00 [id=93 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.288238] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 93 events 0x0 mode thread_spinlock
[1650465529.288276] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x40daf40 [id=97 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.288296] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 97 events 0x0 mode thread_spinlock
[1650465529.288312] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x40daf80 [id=98 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.288352] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 98 events 0x0 mode thread_spinlock
[1650465529.288369] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x3494b00 [id=100 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.288413] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 100 events 0x0 mode thread_spinlock
[1650465529.288453] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x3494b40 [id=104 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.288479] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 104 events 0x0 mode thread_spinlock
[1650465529.288495] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x3494b80 [id=105 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.288515] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 105 events 0x0 mode thread_spinlock
[1650465529.288530] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x3494bc0 [id=107 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.288552] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 107 events 0x0 mode thread_spinlock
[1650465529.288653] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x3494c00 [id=111 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.288682] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 111 events 0x0 mode thread_spinlock
[1650465529.288702] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x3494c40 [id=112 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.288721] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 112 events 0x0 mode thread_spinlock
[1650465529.288737] [ndv4:13452:0] async.c:228 UCX DE[1650465529.287165] [ndv4:15226:0] ib_md.c:812 UCX DEBUG registered memory 0x2b179f891000..0x2b179f916000 on mlx5_ib5 lkey 0x80b00 rkey 0x80b00 access 0xf flags 0x3e4
[1650465529.287172] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b179f891018 of 544744 bytes with 128 elements
[1650465529.287177] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.287436] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x4972460: adding gid fe80::15:5dff:fd34:0 to hash on device mlx5_ib5 port 1 index 0)
[1650465529.288144] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x4972460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 1)
[1650465529.289168] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x4972460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 2)
[1650465529.289392] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x4972460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 3)
[1650465529.289738] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x4972460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 4)
[1650465529.290121] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x4972460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 5)
[1650465529.287368] [ndv4:14205:0] ib_md.c:812 UCX DEBUG registered memory 0x2afdace00000..0x2afdaf400000 on mlx5_ib1 lkey 0x80e00 rkey 0x80e00 access 0xf flags 0x3e4
[1650465529.287392] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2afdace00018 of 39845864 bytes with 4752 elements
[1650465529.287536] [ndv4:14205:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x3446010
[1650465529.287570] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[13]=0x3446010 using dc_mlx5/mlx5_ib1:1 on worker 0x29ca770
[1650465529.287907] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.288307] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.289148] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.289687] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.290323] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
BUG added async handler 0x3494c80 [id=114 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.288756] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 114 events 0x0 mode thread_spinlock
[1650465529.288796] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x3494cc0 [id=118 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.288803] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 118 events 0x0 mode thread_spinlock
[1650465529.288819] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x5e67dd0 [id=119 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.288823] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 119 events 0x0 mode thread_spinlock
[1650465529.288838] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x5e67e10 [id=121 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.288841] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 121 events 0x0 mode thread_spinlock
[1650465529.288879] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x5e67e50 [id=125 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.288904] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 125 events 0x0 mode thread_spinlock
[1650465529.288920] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x5e67e90 [id=126 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.288938] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 126 events 0x0 mode thread_spinlock
[1650465529.288952] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x5e67ed0 [id=128 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.288970] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 128 events 0x0 mode thread_spinlock
[1650465529.289006] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x5e67f10 [id=132 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.289030] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 132 events 0x0 mode thread_spinlock
[1650465529.289045] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x5e67f50 [id=133 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.289060] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 133 events 0x0 mode thread_spinlock
[1650465529.289074] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x5e67f90 [id=135 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.289094] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 135 events 0x0 mode thread_spinlock
[1650465529.290636] [ndv4:12741:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465529.290643] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465529.291650] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.292041] [ndv4:14205:0] ib_iface.c:994 UCX DEBUG iface=0x3345450: created UD QP 0xc459 on mlx5_ib1:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.292679] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.291983] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x4972460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 6)
[1650465529.292370] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x4972460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 7)
[1650465529.292377] [ndv4:15226:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib5:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.292416] [ndv4:15226:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.292421] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x4972f60 [id=124 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.292430] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 124 events 0x5 mode thread_spinlock
[1650465529.292443] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[35]=0x4972460 using ud_mlx5/mlx5_ib5:1 on worker 0x1f938d0
[1650465529.292794] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.293169] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.293303] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.294129] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.293792] [ndv4:13452:0] async.c:228 UCX DEBUG added async handler 0x5f42af0 [id=139 ref 1] uct_rdmacm_cm_event_handler() to hash
[1650465529.293807] [ndv4:13452:0] async.c:506 UCX DEBUG listening to async event fd 139 events 0x1 mode thread_spinlock
[1650465529.293820] [ndv4:13452:0] rdmacm_cm.c:922 UCX DEBUG created rdmacm_cm 0x5f424f0 with event_channel 0x5e67fd0 (fd=139)
[1650465529.293850] [ndv4:13452:0] tcp_sockcm.c:186 UCX DEBUG created tcp_sockcm 0x3edd7e0
[1650465529.293862] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool ucp_requests: align 64, maxelems 4294967295, elemsize 440
[1650465529.293865] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool ucp_rkeys: align 64, maxelems 4294967295, elemsize 168
[1650465529.293870] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool ucp_am_bufs: align 64, maxelems 4294967295, elemsize 65624
[1650465529.293873] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool ucp_reg_bufs: align 64, maxelems 4294967295, elemsize 8208
[1650465529.293876] [ndv4:13452:0] mpool.c:88 UCX DEBUG mpool ucp_rndv_frags: align 512, maxelems 4294967295, elemsize 524304
[1650465529.293959] [ndv4:13452:0] parser.c:1893 UCX INFO UCX_* env variables: UCX_POSIX_USE_PROC_LINK=n UCX_LOG_LEVEL=debug
[1650465529.295205] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib6:1
[1650465529.294542] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.294893] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.295319] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.295872] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.296293] [ndv4:14205:0] ib_md.c:812 UCX DEBUG registered memory 0x2afda3f51000..0x2afda3fd6000 on mlx5_ib1 lkey 0x80f00 rkey 0x80f00 access 0xf flags 0x3e4
[1650465529.296300] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2afda3f51018 of 544744 bytes with 128 elements
[1650465529.296305] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.296719] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x3345450: adding gid fe80::15:5dff:fd33:fffc to hash on device mlx5_ib1 port 1 index 0)
[1650465529.296861] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x3345450: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 1)
[1650465529.297007] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x3345450: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 2)
[1650465529.296107] [ndv4:12741:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465529.296114] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465529.297869] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x3345450: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 3)
[1650465529.298112] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x3345450: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 4)
[1650465529.298880] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x3345450: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 5)
[1650465529.299286] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x3345450: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 6)
[1650465529.299825] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x3345450: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 7)
[1650465529.300120] [ndv4:14205:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.300127] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x2113600 [id=95 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.300143] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 95 events 0x5 mode thread_spinlock
[1650465529.300153] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[14]=0x3345450 using ud_verbs/mlx5_ib1:1 on worker 0x29ca770
[1650465529.300681] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.300968] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.301119] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.301236] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.299720] [ndv4:12741:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465529.299727] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465529.301484] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465529.304874] [ndv4:12741:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465529.304880] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465529.308037] [ndv4:12741:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465529.310632] [ndv4:14421:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib6 vendor_id: 0x15b3 device_id: 4124
[1650465529.311246] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465529.311264] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.311268] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.311321] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.311718] [ndv4:14421:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib6: disable ODP because it's not supported for DevX QP
[1650465529.311956] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.311964] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465529.312903] [ndv4:15226:0] ib_iface.c:994 UCX DEBUG iface=0x4a720a0: created RC QP 0xbfc1 on mlx5_ib6:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465529.312757] [ndv4:14421:0] async.c:228 UCX DEBUG added async handler 0x1619890 [id=69 ref 1] uct_ib_async_event_handler() to hash
[1650465529.312795] [ndv4:14421:0] async.c:506 UCX DEBUG listening to async event fd 69 events 0x1 mode thread_spinlock
[1650465529.312799] [ndv4:14421:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib6' (InfiniBand channel adapter) with 1 ports
[1650465529.313542] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.313707] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.313713] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465529.313769] [ndv4:14421:0] ib_md.c:1319 UCX DEBUG mlx5_ib6: using registration cache
[1650465529.314173] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.314570] [ndv4:14205:0] ib_iface.c:994 UCX DEBUG iface=0x320c280: created UD QP 0xc45a on mlx5_ib1:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.314623] [ndv4:14205:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465529.315269] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.316240] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.316548] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.316875] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.317059] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.316192] [ndv4:13552:0] async.c:228 UCX DEBUG added async handler 0x1ab6940 [id=71 ref 1] uct_ib_async_event_handler() to hash
[1650465529.316225] [ndv4:13552:0] async.c:506 UCX DEBUG listening to async event fd 71 events 0x1 mode thread_spinlock
[1650465529.316229] [ndv4:13552:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib7' (InfiniBand channel adapter) with 1 ports
[1650465529.316999] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.317061] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.317066] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465529.317117] [ndv4:13552:0] ib_md.c:1319 UCX DEBUG mlx5_ib7: using registration cache
[1650465529.316470] [ndv4:13443:0] mm_posix.c:518 UCX DEBUG allocated posix shared memory at 0x2b7507896000 length 12288
[1650465529.316557] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool mm_recv_desc: align 64, maxelems 4294967295, elemsize 8368
[1650465529.315799] [ndv4:14421:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465529.315806] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465529.316012] [ndv4:14421:0] ib_md.c:1604 UCX DEBUG mlx5_ib6: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465529.314180] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[36]=0x4a720a0 using rc_verbs/mlx5_ib6:1 on worker 0x1f938d0
[1650465529.314395] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.314438] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.314522] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.314529] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.316476] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib6:1
[1650465529.318617] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.318631] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.318634] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.318688] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.317462] [ndv4:14205:0] ib_md.c:812 UCX DEBUG registered memory 0x2afdaf498000..0x2afdaf51d000 on mlx5_ib1 lkey 0x81000 rkey 0x81000 access 0xf flags 0x3e4
[1650465529.317468] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2afdaf498018 of 544744 bytes with 128 elements
[1650465529.317473] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.318185] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x320c280: adding gid fe80::15:5dff:fd33:fffc to hash on device mlx5_ib1 port 1 index 0)
[1650465529.318403] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x320c280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 1)
[1650465529.318605] [ndv4:13552:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465529.318613] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465529.318885] [ndv4:13552:0] ib_md.c:1604 UCX DEBUG mlx5_ib7: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465529.319246] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.319665] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.317920] [ndv4:13443:0] mm_posix.c:305 UCX DEBUG shared memory mmap(addr=(nil), length=6291456, flags= HUGETLB, fd=76) failed: Invalid argument
[1650465529.317930] [ndv4:13443:0] mm_posix.c:518 UCX DEBUG allocated posix shared memory at 0x2b7507899000 length 4296704
[1650465529.317935] [ndv4:13443:0] mpool.c:205 UCX DEBUG mpool mm_recv_desc: allocated chunk 0x2b7507899018 of 4296680 bytes with 512 elements
[1650465529.318199] [ndv4:13443:0] mm_iface.c:600 UCX DEBUG created mm iface 0x24a5d50 FIFO id 0x400000005e8b39bc va 0x2b7507896000 size 12288 (128 x 64 elems)
[1650465529.318548] [ndv4:13443:0] ucp_worker.c:1159 UCX DEBUG created interface[0]=0x24a5d50 using posix/memory on worker 0x2d7f8d0
[1650465529.318572] [ndv4:13443:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 8447 bytes with hugetlb
[1650465529.318701] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool mm_recv_desc: align 64, maxelems 4294967295, elemsize 8368
[1650465529.318717] [ndv4:13443:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 4292720 bytes with hugetlb
[1650465529.318727] [ndv4:13443:0] mpool.c:205 UCX DEBUG mpool mm_recv_desc: allocated chunk 0x2b750c000018 of 4296680 bytes with 512 elements
[1650465529.319309] [ndv4:13443:0] mm_iface.c:600 UCX DEBUG created mm iface 0x24a6320 FIFO id 0x648035 va 0x2b7507cb2000 size 12288 (128 x 64 elems)
[1650465529.319316] [ndv4:13443:0] ucp_worker.c:1159 UCX DEBUG created interface[1]=0x24a6320 using sysv/memory on worker 0x2d7f8d0
[1650465529.319328] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool self_msg_desc: align 64, maxelems 4294967295, elemsize 8200
[1650465529.319331] [ndv4:13443:0] self.c:220 UCX DEBUG created self iface id 0x5dd80def44468893 send_size 8192
[1650465529.319338] [ndv4:13443:0] ucp_worker.c:1159 UCX DEBUG created interface[2]=0x24b9f00 using self/memory0 on worker 0x2d7f8d0
[1650465529.319361] [ndv4:13443:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465529.319366] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465529.319369] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465529.319130] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x4d93010 of 8176 bytes with 127 elements
[1650465529.319402] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.319408] [ndv4:15226:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib6:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.319445] [ndv4:15226:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib6 length=2048) failed: Invalid argument
[1650465529.319450] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.319460] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x4bc7c00 [id=127 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.319471] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 127 events 0x1 mode thread_spinlock
[1650465529.319479] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[37]=0x4bbf030 using rc_mlx5/mlx5_ib6:1 on worker 0x1f938d0
[1650465529.319356] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x320c280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 2)
[1650465529.319995] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x320c280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 3)
[1650465529.320313] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x320c280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 4)
[1650465529.320627] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x320c280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 5)
[1650465529.321292] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x320c280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 6)
[1650465529.322147] [ndv4:12741:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib7 vendor_id: 0x15b3 device_id: 4124
[1650465529.322690] [ndv4:12741:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib7: disable ODP because it's not supported for DevX QP
[1650465529.322612] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.323030] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.324294] [ndv4:13552:0] topo.c:99 UCX DEBUG bus id 0x108000000 doesn't exist. sys_dev = 7
[1650465529.324301] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465529.323926] [ndv4:13443:0] async.c:228 UCX DEBUG added async handler 0x24bc040 [id=78 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465529.323966] [ndv4:13443:0] async.c:506 UCX DEBUG listening to async event fd 78 events 0x5 mode thread_spinlock
[1650465529.323985] [ndv4:13443:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x24ba860: listening for connections (fd=78) on 10.5.0.5:35601
[1650465529.324737] [ndv4:13443:0] ucp_worker.c:1159 UCX DEBUG created interface[3]=0x24ba860 using tcp/eth0 on worker 0x2d7f8d0
[1650465529.324759] [ndv4:13443:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465529.324763] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465529.324766] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465529.324925] [ndv4:13443:0] async.c:228 UCX DEBUG added async handler 0x1d1f7b0 [id=80 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465529.324949] [ndv4:13443:0] async.c:506 UCX DEBUG listening to async event fd 80 events 0x5 mode thread_spinlock
[1650465529.324953] [ndv4:13443:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x24baee0: listening for connections (fd=80) on 127.0.0.1:60761
[1650465529.325015] [ndv4:13443:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465529.325021] [ndv4:13443:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465529.325471] [ndv4:13443:0] ucp_worker.c:1159 UCX DEBUG created interface[4]=0x24baee0 using tcp/lo on worker 0x2d7f8d0
[1650465529.325487] [ndv4:13443:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465529.325490] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465529.325493] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465529.325523] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.325756] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.326240] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.326497] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.326410] [ndv4:13443:0] async.c:228 UCX DEBUG added async handler 0x24177d0 [id=82 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465529.326431] [ndv4:13443:0] async.c:506 UCX DEBUG listening to async event fd 82 events 0x5 mode thread_spinlock
[1650465529.326435] [ndv4:13443:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x24a2570: listening for connections (fd=82) on 172.16.1.242:42120
[1650465529.326979] [ndv4:13443:0] ucp_worker.c:1159 UCX DEBUG created interface[5]=0x24a2570 using tcp/ib0 on worker 0x2d7f8d0
[1650465529.327523] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.328082] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.328520] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.328642] [ndv4:12741:0] async.c:228 UCX DEBUG added async handler 0x2dd8940 [id=71 ref 1] uct_ib_async_event_handler() to hash
[1650465529.328682] [ndv4:12741:0] async.c:506 UCX DEBUG listening to async event fd 71 events 0x1 mode thread_spinlock
[1650465529.328686] [ndv4:12741:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib7' (InfiniBand channel adapter) with 1 ports
[1650465529.329338] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.329551] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.329557] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465529.329695] [ndv4:12741:0] ib_md.c:1319 UCX DEBUG mlx5_ib7: using registration cache
[1650465529.328997] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.330255] [ndv4:13443:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465529.330541] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib6:1
[1650465529.331712] [ndv4:13443:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465529.331753] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.331757] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.331769] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.331830] [ndv4:12741:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465529.331837] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465529.332053] [ndv4:12741:0] ib_md.c:1604 UCX DEBUG mlx5_ib7: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465529.332201] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.332210] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465529.333050] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x320c280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 7)
[1650465529.333061] [ndv4:14205:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib1:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.333095] [ndv4:14205:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.333099] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x33452c0 [id=96 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.333134] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 96 events 0x5 mode thread_spinlock
[1650465529.333147] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[15]=0x320c280 using ud_mlx5/mlx5_ib1:1 on worker 0x29ca770
[1650465529.333802] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.334164] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.333020] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.333039] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.333042] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.333104] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.334224] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465529.334231] [ndv4:15226:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib6:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.334266] [ndv4:15226:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib6 length=2048) failed: Invalid argument
[1650465529.334269] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.334277] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x4330330 [id=129 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.334306] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 129 events 0x1 mode thread_spinlock
[1650465529.333212] [ndv4:13443:0] ib_iface.c:994 UCX DEBUG iface=0x24b06b0: created RC QP 0xde5e on mlx5_ib0:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465529.332830] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.333522] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.334958] [ndv4:15226:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465529.334953] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.335113] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.335956] [ndv4:13552:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465529.335964] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465529.341779] [ndv4:12741:0] topo.c:99 UCX DEBUG bus id 0x108000000 doesn't exist. sys_dev = 7
[1650465529.341786] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465529.342860] [ndv4:14421:0] topo.c:99 UCX DEBUG bus id 0x107000000 doesn't exist. sys_dev = 6
[1650465529.342870] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465529.344257] [ndv4:15226:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x4d95050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xc02b
[1650465529.345156] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.345198] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.345208] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b17a2117008 of 151544 bytes with 1052 elements
[1650465529.349245] [ndv4:15226:0] ib_md.c:812 UCX DEBUG registered memory 0x2b179fa00000..0x2b17a2000000 on mlx5_ib6 lkey 0x80700 rkey 0x80700 access 0xf flags 0x3e4
[1650465529.349268] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b179fa00018 of 39845864 bytes with 4752 elements
[1650465529.349410] [ndv4:15226:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x4d95050
[1650465529.349449] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[38]=0x4d95050 using dc_mlx5/mlx5_ib6:1 on worker 0x1f938d0
[1650465529.349815] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.349989] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.350360] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.350407] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.349122] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465529.351425] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib6:1
[1650465529.352777] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.353995] [ndv4:15226:0] ib_iface.c:994 UCX DEBUG iface=0x4f712e0: created UD QP 0xc04b on mlx5_ib6:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.354789] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.356247] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.356328] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.356488] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.357299] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.357783] [ndv4:15226:0] ib_md.c:812 UCX DEBUG registered memory 0x2b17a213c000..0x2b17a21c1000 on mlx5_ib6 lkey 0x80800 rkey 0x80800 access 0xf flags 0x3e4
[1650465529.357790] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b17a213c018 of 544744 bytes with 128 elements
[1650465529.357794] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.358397] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x4f712e0: adding gid fe80::15:5dff:fd34:1 to hash on device mlx5_ib6 port 1 index 0)
[1650465529.358690] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465529.358711] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.358715] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.358776] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.359541] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.359550] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465529.359456] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x4f712e0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 1)
[1650465529.360088] [ndv4:13443:0] ucp_worker.c:1159 UCX DEBUG created interface[6]=0x24b06b0 using rc_verbs/mlx5_ib0:1 on worker 0x2d7f8d0
[1650465529.360297] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.361076] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.361491] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.361922] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.362951] [ndv4:13443:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465529.360516] [ndv4:14205:0] ib_iface.c:994 UCX DEBUG iface=0x38350a0: created RC QP 0xc4aa on mlx5_ib2:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465529.359764] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x4f712e0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 2)
[1650465529.360219] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x4f712e0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 3)
[1650465529.360937] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x4f712e0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 4)
[1650465529.362067] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x4f712e0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 5)
[1650465529.362371] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x4f712e0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 6)
[1650465529.362424] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x4f712e0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 7)
[1650465529.362710] [ndv4:15226:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.362719] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x4a4d8d0 [id=130 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.362755] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 130 events 0x5 mode thread_spinlock
[1650465529.362771] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[39]=0x4f712e0 using ud_verbs/mlx5_ib6:1 on worker 0x1f938d0
[1650465529.362999] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.363552] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.361096] [ndv4:12741:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465529.361106] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465529.363954] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.364182] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.363923] [ndv4:13443:0] ib_device.c:1394 UCX DEBUG max IB CQE size is 128
[1650465529.365136] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib6:1
[1650465529.365151] [ndv4:13443:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.365163] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.365167] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.365222] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.366566] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.365871] [ndv4:13443:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x2e49010 of 8176 bytes with 127 elements
[1650465529.366167] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.366190] [ndv4:13443:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib0:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.366239] [ndv4:13443:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650465529.366244] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.365554] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[16]=0x38350a0 using rc_verbs/mlx5_ib2:1 on worker 0x29ca770
[1650465529.366792] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.367687] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.367992] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.367519] [ndv4:15226:0] ib_iface.c:994 UCX DEBUG iface=0x508f460: created UD QP 0xc067 on mlx5_ib6:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.367527] [ndv4:15226:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465529.368692] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.369326] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465529.369889] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.370735] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.370800] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.370915] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.371166] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.371062] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.371077] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.371080] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.371131] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.371685] [ndv4:15226:0] ib_md.c:812 UCX DEBUG registered memory 0x2b17a21c1000..0x2b17a2246000 on mlx5_ib6 lkey 0x80900 rkey 0x80900 access 0xf flags 0x3e4
[1650465529.371691] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b17a21c1018 of 544744 bytes with 128 elements
[1650465529.371696] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.371928] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x508f460: adding gid fe80::15:5dff:fd34:1 to hash on device mlx5_ib6 port 1 index 0)
[1650465529.372335] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x508f460: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 1)
[1650465529.372401] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x508f460: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 2)
[1650465529.372453] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x508f460: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 3)
[1650465529.371681] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x3b56010 of 8176 bytes with 127 elements
[1650465529.371937] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.371944] [ndv4:14205:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib2:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.371979] [ndv4:14205:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib2 length=2048) failed: Invalid argument
[1650465529.371982] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.371995] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x32a3bb0 [id=99 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.372004] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 99 events 0x1 mode thread_spinlock
[1650465529.372027] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[17]=0x3982030 using rc_mlx5/mlx5_ib2:1 on worker 0x29ca770
[1650465529.372254] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.372665] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.372881] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.373196] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.374057] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465529.373696] [ndv4:13443:0] async.c:228 UCX DEBUG added async handler 0x2416090 [id=85 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.373728] [ndv4:13443:0] async.c:506 UCX DEBUG listening to async event fd 85 events 0x1 mode thread_spinlock
[1650465529.373748] [ndv4:13443:0] ucp_worker.c:1159 UCX DEBUG created interface[7]=0x24be100 using rc_mlx5/mlx5_ib0:1 on worker 0x2d7f8d0
[1650465529.374175] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.374854] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.375130] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.375698] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.375714] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.375717] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.375773] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.375934] [ndv4:14949:0] mm_posix.c:518 UCX DEBUG allocated posix shared memory at 0x2ad4b7eeb000 length 12288
[1650465529.376013] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool mm_recv_desc: align 64, maxelems 4294967295, elemsize 8368
[1650465529.376172] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465529.376181] [ndv4:14205:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib2:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.376215] [ndv4:14205:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib2 length=2048) failed: Invalid argument
[1650465529.376218] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.376226] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x20f33e0 [id=101 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.376246] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 101 events 0x1 mode thread_spinlock
[1650465529.375883] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.376519] [ndv4:13443:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465529.376925] [ndv4:14205:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465529.377210] [ndv4:14949:0] mm_posix.c:305 UCX DEBUG shared memory mmap(addr=(nil), length=6291456, flags= HUGETLB, fd=76) failed: Invalid argument
[1650465529.377219] [ndv4:14949:0] mm_posix.c:518 UCX DEBUG allocated posix shared memory at 0x2ad4bd841000 length 4296704
[1650465529.377224] [ndv4:14949:0] mpool.c:205 UCX DEBUG mpool mm_recv_desc: allocated chunk 0x2ad4bd841018 of 4296680 bytes with 512 elements
[1650465529.377480] [ndv4:14949:0] mm_iface.c:600 UCX DEBUG created mm iface 0x10d5a10 FIFO id 0x400000000c4c9420 va 0x2ad4b7eeb000 size 12288 (128 x 64 elems)
[1650465529.377673] [ndv4:13443:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.377683] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.377686] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.377699] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.378113] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465529.378120] [ndv4:13443:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib0:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.378153] [ndv4:13443:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650465529.378156] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.378165] [ndv4:13443:0] async.c:228 UCX DEBUG added async handler 0x2417480 [id=87 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.378187] [ndv4:13443:0] async.c:506 UCX DEBUG listening to async event fd 87 events 0x1 mode thread_spinlock
[1650465529.377823] [ndv4:14949:0] ucp_worker.c:1159 UCX DEBUG created interface[0]=0x10d5a10 using posix/memory on worker 0x19ac660
[1650465529.377852] [ndv4:14949:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 8447 bytes with hugetlb
[1650465529.377895] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool mm_recv_desc: align 64, maxelems 4294967295, elemsize 8368
[1650465529.377914] [ndv4:14949:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 4292720 bytes with hugetlb
[1650465529.377924] [ndv4:14949:0] mpool.c:205 UCX DEBUG mpool mm_recv_desc: allocated chunk 0x2ad4bdc5a018 of 4296680 bytes with 512 elements
[1650465529.378500] [ndv4:14949:0] mm_iface.c:600 UCX DEBUG created mm iface 0x10e7760 FIFO id 0x648038 va 0x2ad4b7eee000 size 12288 (128 x 64 elems)
[1650465529.378509] [ndv4:14949:0] ucp_worker.c:1159 UCX DEBUG created interface[1]=0x10e7760 using sysv/memory on worker 0x19ac660
[1650465529.378521] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool self_msg_desc: align 64, maxelems 4294967295, elemsize 8200
[1650465529.378524] [ndv4:14949:0] self.c:220 UCX DEBUG created self iface id 0xa7861ef0e5659aa9 send_size 8192
[1650465529.378530] [ndv4:14949:0] ucp_worker.c:1159 UCX DEBUG created interface[2]=0x10e6e00 using self/memory0 on worker 0x19ac660
[1650465529.378553] [ndv4:14949:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465529.378558] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465529.378561] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465529.378876] [ndv4:13443:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465529.380233] [ndv4:13552:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465529.380241] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465529.379806] [ndv4:14421:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465529.379817] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465529.382017] [ndv4:14949:0] async.c:228 UCX DEBUG added async handler 0x10f4ae0 [id=78 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465529.382061] [ndv4:14949:0] async.c:506 UCX DEBUG listening to async event fd 78 events 0x5 mode thread_spinlock
[1650465529.382081] [ndv4:14949:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x10e7d30: listening for connections (fd=78) on 10.5.0.5:34331
[1650465529.382717] [ndv4:14949:0] ucp_worker.c:1159 UCX DEBUG created interface[3]=0x10e7d30 using tcp/eth0 on worker 0x19ac660
[1650465529.382740] [ndv4:14949:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465529.382744] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465529.382747] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465529.382872] [ndv4:14949:0] async.c:228 UCX DEBUG added async handler 0x10435f0 [id=80 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465529.382904] [ndv4:14949:0] async.c:506 UCX DEBUG listening to async event fd 80 events 0x5 mode thread_spinlock
[1650465529.382908] [ndv4:14949:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x10cf4b0: listening for connections (fd=80) on 127.0.0.1:37455
[1650465529.382939] [ndv4:14949:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465529.382945] [ndv4:14949:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465529.383144] [ndv4:14949:0] ucp_worker.c:1159 UCX DEBUG created interface[4]=0x10cf4b0 using tcp/lo on worker 0x19ac660
[1650465529.383160] [ndv4:14949:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465529.383163] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465529.383165] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465529.383442] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x508f460: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 4)
[1650465529.384013] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x508f460: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 5)
[1650465529.384072] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x508f460: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 6)
[1650465529.384286] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x508f460: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 7)
[1650465529.384294] [ndv4:15226:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib6:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.384398] [ndv4:15226:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.384404] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x4a4da40 [id=131 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.384432] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 131 events 0x5 mode thread_spinlock
[1650465529.384451] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[40]=0x508f460 using ud_mlx5/mlx5_ib6:1 on worker 0x1f938d0
[1650465529.384712] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.385490] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.383374] [ndv4:14949:0] async.c:228 UCX DEBUG added async handler 0x1042ed0 [id=82 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465529.383403] [ndv4:14949:0] async.c:506 UCX DEBUG listening to async event fd 82 events 0x5 mode thread_spinlock
[1650465529.383407] [ndv4:14949:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x10cfb30: listening for connections (fd=82) on 172.16.1.242:45429
[1650465529.384048] [ndv4:14949:0] ucp_worker.c:1159 UCX DEBUG created interface[5]=0x10cfb30 using tcp/ib0 on worker 0x19ac660
[1650465529.384265] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.384927] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.385043] [ndv4:12741:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465529.385052] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465529.385638] [ndv4:14205:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x3b58050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xc4bd
[1650465529.386063] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.386508] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.386516] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2afda3fd8008 of 151544 bytes with 1052 elements
[1650465529.385859] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.386194] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.386546] [ndv4:14949:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465529.385760] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.386376] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.388063] [ndv4:13443:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x30b3010: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xde84
[1650465529.389327] [ndv4:14949:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465529.389395] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.389399] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.389431] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.389768] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.389778] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465529.389346] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.389800] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.389823] [ndv4:13443:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b7507cb7008 of 151544 bytes with 1052 elements
[1650465529.390411] [ndv4:14205:0] ib_md.c:812 UCX DEBUG registered memory 0x2afdaf600000..0x2afdb1c00000 on mlx5_ib2 lkey 0x81000 rkey 0x81000 access 0xf flags 0x3e4
[1650465529.390430] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2afdaf600018 of 39845864 bytes with 4752 elements
[1650465529.390566] [ndv4:14205:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x3b58050
[1650465529.390658] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[18]=0x3b58050 using dc_mlx5/mlx5_ib2:1 on worker 0x29ca770
[1650465529.391375] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.390340] [ndv4:14949:0] ib_iface.c:994 UCX DEBUG iface=0x10dd5b0: created RC QP 0xde67 on mlx5_ib0:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465529.392170] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.392680] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.392931] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.393380] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465529.393800] [ndv4:13443:0] ib_md.c:812 UCX DEBUG registered memory 0x2b750c600000..0x2b750ec00000 on mlx5_ib0 lkey 0x80f00 rkey 0x80f00 access 0xf flags 0x3e4
[1650465529.393822] [ndv4:13443:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b750c600018 of 39845864 bytes with 4752 elements
[1650465529.393959] [ndv4:13443:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x30b3010
[1650465529.393996] [ndv4:13443:0] ucp_worker.c:1159 UCX DEBUG created interface[8]=0x30b3010 using dc_mlx5/mlx5_ib0:1 on worker 0x2d7f8d0
[1650465529.394430] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.394469] [ndv4:14949:0] ucp_worker.c:1159 UCX DEBUG created interface[6]=0x10dd5b0 using rc_verbs/mlx5_ib0:1 on worker 0x19ac660
[1650465529.395035] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.395385] [ndv4:14205:0] ib_iface.c:994 UCX DEBUG iface=0x3d34060: created UD QP 0xc4b3 on mlx5_ib2:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.394841] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.395383] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.395765] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.394875] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.395105] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.395568] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.395806] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.395992] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.396656] [ndv4:13443:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465529.396697] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.397146] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.397298] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.397314] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.397967] [ndv4:13443:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.398357] [ndv4:13443:0] ib_iface.c:994 UCX DEBUG iface=0x24c6a70: created UD QP 0xde68 on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.397739] [ndv4:14205:0] ib_md.c:812 UCX DEBUG registered memory 0x2afdb1d1e000..0x2afdb1da3000 on mlx5_ib2 lkey 0x81100 rkey 0x81100 access 0xf flags 0x3e4
[1650465529.397746] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2afdb1d1e018 of 544744 bytes with 128 elements
[1650465529.397752] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.398373] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x3d34060: adding gid fe80::15:5dff:fd33:fffd to hash on device mlx5_ib2 port 1 index 0)
[1650465529.399089] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.399718] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.400033] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.400427] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.400915] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.401422] [ndv4:13443:0] ib_md.c:812 UCX DEBUG registered memory 0x2b7507cdc000..0x2b7507d61000 on mlx5_ib0 lkey 0x81000 rkey 0x81000 access 0xf flags 0x3e4
[1650465529.401436] [ndv4:13443:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b7507cdc018 of 544744 bytes with 128 elements
[1650465529.401442] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.402289] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x24c6a70: adding gid fe80::15:5dff:fd33:fffb to hash on device mlx5_ib0 port 1 index 0)
[1650465529.403165] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x24c6a70: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650465529.404506] [ndv4:13552:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465529.404514] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465529.408911] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib7:1
[1650465529.408879] [ndv4:14949:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465529.409969] [ndv4:14949:0] ib_device.c:1394 UCX DEBUG max IB CQE size is 128
[1650465529.409654] [ndv4:14421:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465529.409663] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465529.410174] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x3d34060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 1)
[1650465529.410357] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x3d34060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 2)
[1650465529.410342] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465529.410360] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.410364] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.410418] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.410936] [ndv4:12741:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465529.410945] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465529.411110] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.411117] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465529.411076] [ndv4:14949:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.411090] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.411094] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.411153] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.411676] [ndv4:14949:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x1a76010 of 8176 bytes with 127 elements
[1650465529.411952] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.411984] [ndv4:14949:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib0:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.412044] [ndv4:14949:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650465529.412050] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.412372] [ndv4:15226:0] ib_iface.c:994 UCX DEBUG iface=0x518f0a0: created RC QP 0xbe92 on mlx5_ib7:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465529.413822] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[41]=0x518f0a0 using rc_verbs/mlx5_ib7:1 on worker 0x1f938d0
[1650465529.414299] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.414308] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.414521] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.414658] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.414855] [ndv4:14421:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465529.414863] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465529.414706] [ndv4:12741:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465529.414712] [ndv4:12741:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465529.414881] [ndv4:12741:0] ucp_context.c:1556 UCX DEBUG created ucp context 0x26dbc90 0x26dbc90 [13 mds 47 tls] features 0x1 tl bitmap 0x7fffffffffff 0x0
[1650465529.415086] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib7:1
[1650465529.415844] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x24c6a70: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650465529.416455] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x24c6a70: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650465529.417000] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x24c6a70: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650465529.417375] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.417392] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.417395] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.417445] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.417904] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x24c6a70: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650465529.419014] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x24c6a70: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650465529.419174] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x24c6a70: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650465529.419499] [ndv4:13443:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.418029] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x54b0010 of 8176 bytes with 127 elements
[1650465529.418506] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.418512] [ndv4:15226:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib7:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.418554] [ndv4:15226:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib7 length=2048) failed: Invalid argument
[1650465529.418558] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.418569] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x50bc8f0 [id=134 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.418683] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 134 events 0x1 mode thread_spinlock
[1650465529.418696] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[42]=0x52dc030 using rc_mlx5/mlx5_ib7:1 on worker 0x1f938d0
[1650465529.418929] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.419047] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.419382] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.419451] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.419562] [ndv4:14949:0] async.c:228 UCX DEBUG added async handler 0x1035490 [id=85 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.419675] [ndv4:14949:0] async.c:506 UCX DEBUG listening to async event fd 85 events 0x1 mode thread_spinlock
[1650465529.419790] [ndv4:14949:0] ucp_worker.c:1159 UCX DEBUG created interface[7]=0x10eb000 using rc_mlx5/mlx5_ib0:1 on worker 0x19ac660
[1650465529.419939] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.420012] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.420216] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.420224] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.419784] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x3d34060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 3)
[1650465529.419960] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x3d34060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 4)
[1650465529.420794] [ndv4:13552:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465529.420802] [ndv4:13552:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465529.421086] [ndv4:13552:0] ucp_context.c:1556 UCX DEBUG created ucp context 0x13b9c90 0x13b9c90 [13 mds 47 tls] features 0x1 tl bitmap 0x7fffffffffff 0x0
[1650465529.420915] [ndv4:14949:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465529.422066] [ndv4:14949:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.422077] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.422080] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.422094] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.422531] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465529.422539] [ndv4:14949:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib0:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.422619] [ndv4:14949:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650465529.422623] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.422632] [ndv4:14949:0] async.c:228 UCX DEBUG added async handler 0x10447d0 [id=87 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.422660] [ndv4:14949:0] async.c:506 UCX DEBUG listening to async event fd 87 events 0x1 mode thread_spinlock
[1650465529.423275] [ndv4:14949:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465529.425807] [ndv4:13443:0] async.c:228 UCX DEBUG added async handler 0x24ae4a0 [id=88 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.425846] [ndv4:13443:0] async.c:506 UCX DEBUG listening to async event fd 88 events 0x5 mode thread_spinlock
[1650465529.425888] [ndv4:13443:0] ucp_worker.c:1159 UCX DEBUG created interface[9]=0x24c6a70 using ud_verbs/mlx5_ib0:1 on worker 0x2d7f8d0
[1650465529.426135] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.426393] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.428194] [ndv4:14421:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465529.428203] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465529.427023] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.427236] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.427827] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib7:1
[1650465529.427778] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x3d34060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 5)
[1650465529.428154] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x3d34060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 6)
[1650465529.428902] [ndv4:13443:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465529.429654] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.429673] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.429677] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.429734] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.429949] [ndv4:13443:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.430258] [ndv4:13443:0] ib_iface.c:994 UCX DEBUG iface=0x2d9f750: created UD QP 0xde6d on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.430271] [ndv4:13443:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465529.430356] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465529.430364] [ndv4:15226:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib7:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.430398] [ndv4:15226:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib7 length=2048) failed: Invalid argument
[1650465529.430402] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.430410] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x16dcdf0 [id=136 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.430437] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 136 events 0x1 mode thread_spinlock
[1650465529.430394] [ndv4:14421:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465529.430960] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.431137] [ndv4:15226:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465529.432149] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.432373] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.432734] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.432908] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.433621] [ndv4:14949:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x1ce0010: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xde91
[1650465529.434712] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.435085] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.435111] [ndv4:14949:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2ad4b7ef3008 of 151544 bytes with 1052 elements
[1650465529.433395] [ndv4:13443:0] ib_md.c:812 UCX DEBUG registered memory 0x2b7507d61000..0x2b7507de6000 on mlx5_ib0 lkey 0x81100 rkey 0x81100 access 0xf flags 0x3e4
[1650465529.433402] [ndv4:13443:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b7507d61018 of 544744 bytes with 128 elements
[1650465529.433406] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.437928] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x3d34060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 7)
[1650465529.438922] [ndv4:14949:0] ib_md.c:812 UCX DEBUG registered memory 0x2ad4be200000..0x2ad4c0800000 on mlx5_ib0 lkey 0x81200 rkey 0x81200 access 0xf flags 0x3e4
[1650465529.438947] [ndv4:14949:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2ad4be200018 of 39845864 bytes with 4752 elements
[1650465529.439086] [ndv4:14949:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x1ce0010
[1650465529.439124] [ndv4:14949:0] ucp_worker.c:1159 UCX DEBUG created interface[8]=0x1ce0010 using dc_mlx5/mlx5_ib0:1 on worker 0x19ac660
[1650465529.438290] [ndv4:14205:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.438301] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x2113f20 [id=102 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.438332] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 102 events 0x5 mode thread_spinlock
[1650465529.438352] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[19]=0x3d34060 using ud_verbs/mlx5_ib2:1 on worker 0x29ca770
[1650465529.438671] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.439281] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.439532] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.440288] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.440439] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.440435] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.441375] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.441742] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.442281] [ndv4:14949:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465529.442513] [ndv4:15226:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x54b2050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xbef3
[1650465529.442940] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.443190] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.443201] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b17a4a47008 of 151544 bytes with 1052 elements
[1650465529.445084] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x2d9f750: adding gid fe80::15:5dff:fd33:fffb to hash on device mlx5_ib0 port 1 index 0)
[1650465529.446634] [ndv4:14949:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.446143] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x2d9f750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650465529.446909] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x2d9f750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650465529.447511] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x2d9f750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650465529.447388] [ndv4:14949:0] ib_iface.c:994 UCX DEBUG iface=0x19cc4e0: created UD QP 0xde72 on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.447073] [ndv4:15226:0] ib_md.c:812 UCX DEBUG registered memory 0x2b17a2400000..0x2b17a4a00000 on mlx5_ib7 lkey 0x80f00 rkey 0x80f00 access 0xf flags 0x3e4
[1650465529.447096] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b17a2400018 of 39845864 bytes with 4752 elements
[1650465529.447235] [ndv4:15226:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x54b2050
[1650465529.447273] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[43]=0x54b2050 using dc_mlx5/mlx5_ib7:1 on worker 0x1f938d0
[1650465529.447879] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.447981] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.448458] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.448651] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.448101] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x2d9f750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650465529.449070] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x2d9f750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650465529.448021] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.448848] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.448878] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.449262] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.449396] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.449891] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465529.450069] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib7:1
[1650465529.450199] [ndv4:14949:0] ib_md.c:812 UCX DEBUG registered memory 0x2ad4b7f18000..0x2ad4b7f9d000 on mlx5_ib0 lkey 0x81300 rkey 0x81300 access 0xf flags 0x3e4
[1650465529.450207] [ndv4:14949:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ad4b7f18018 of 544744 bytes with 128 elements
[1650465529.450212] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.451066] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x19cc4e0: adding gid fe80::15:5dff:fd33:fffb to hash on device mlx5_ib0 port 1 index 0)
[1650465529.451394] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x19cc4e0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650465529.450105] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x2d9f750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650465529.451755] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.452420] [ndv4:14205:0] ib_iface.c:994 UCX DEBUG iface=0x2071930: created UD QP 0xc4b4 on mlx5_ib2:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.452429] [ndv4:14205:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465529.451871] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x19cc4e0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650465529.451964] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x19cc4e0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650465529.452395] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x19cc4e0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650465529.452278] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.452717] [ndv4:15226:0] ib_iface.c:994 UCX DEBUG iface=0x568e060: created UD QP 0xbf3c on mlx5_ib7:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.452886] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x19cc4e0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650465529.453232] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x19cc4e0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650465529.454199] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x19cc4e0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650465529.454475] [ndv4:14949:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.453424] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.454963] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.455137] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.453344] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.455982] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.456205] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.456645] [ndv4:15226:0] ib_md.c:812 UCX DEBUG registered memory 0x2b17a4a6c000..0x2b17a4af1000 on mlx5_ib7 lkey 0x81000 rkey 0x81000 access 0xf flags 0x3e4
[1650465529.456652] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b17a4a6c018 of 544744 bytes with 128 elements
[1650465529.456656] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.457668] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x568e060: adding gid fe80::15:5dff:fd34:2 to hash on device mlx5_ib7 port 1 index 0)
[1650465529.458067] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x568e060: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 1)
[1650465529.457951] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x2d9f750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650465529.457961] [ndv4:13443:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib0:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.458022] [ndv4:13443:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.458027] [ndv4:13443:0] async.c:228 UCX DEBUG added async handler 0x2408220 [id=89 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.458056] [ndv4:13443:0] async.c:506 UCX DEBUG listening to async event fd 89 events 0x5 mode thread_spinlock
[1650465529.458071] [ndv4:13443:0] ucp_worker.c:1159 UCX DEBUG created interface[10]=0x2d9f750 using ud_mlx5/mlx5_ib0:1 on worker 0x2d7f8d0
[1650465529.459263] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x568e060: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 2)
[1650465529.459662] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x568e060: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 3)
[1650465529.460665] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x568e060: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 4)
[1650465529.459020] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.459307] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.460280] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.461141] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.460980] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.461127] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.461569] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.460849] [ndv4:14949:0] async.c:228 UCX DEBUG added async handler 0x18e3970 [id=88 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.460893] [ndv4:14949:0] async.c:506 UCX DEBUG listening to async event fd 88 events 0x5 mode thread_spinlock
[1650465529.460916] [ndv4:14949:0] ucp_worker.c:1159 UCX DEBUG created interface[9]=0x19cc4e0 using ud_verbs/mlx5_ib0:1 on worker 0x19ac660
[1650465529.461337] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.461725] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.462271] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.462457] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.462171] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.462772] [ndv4:14205:0] ib_md.c:812 UCX DEBUG registered memory 0x2afdb1da3000..0x2afdb1e28000 on mlx5_ib2 lkey 0x81200 rkey 0x81200 access 0xf flags 0x3e4
[1650465529.462778] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2afdb1da3018 of 544744 bytes with 128 elements
[1650465529.462784] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.463508] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x2071930: adding gid fe80::15:5dff:fd33:fffd to hash on device mlx5_ib2 port 1 index 0)
[1650465529.463828] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x2071930: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 1)
[1650465529.463958] [ndv4:14949:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465529.464994] [ndv4:14949:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.465284] [ndv4:14421:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib7 vendor_id: 0x15b3 device_id: 4124
[1650465529.465519] [ndv4:14949:0] ib_iface.c:994 UCX DEBUG iface=0x10f3970: created UD QP 0xde74 on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.465530] [ndv4:14949:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465529.466631] [ndv4:14421:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib7: disable ODP because it's not supported for DevX QP
[1650465529.466142] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.466976] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.467094] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.468381] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x568e060: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 5)
[1650465529.468885] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x568e060: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 6)
[1650465529.468966] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x568e060: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 7)
[1650465529.469411] [ndv4:15226:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.469419] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x568ef40 [id=137 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.469451] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 137 events 0x5 mode thread_spinlock
[1650465529.469463] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[44]=0x568e060 using ud_verbs/mlx5_ib7:1 on worker 0x1f938d0
[1650465529.469880] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.470458] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.469008] [ndv4:13443:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465529.470747] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.470760] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.471365] [ndv4:15226:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib7:1
[1650465529.472316] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x2071930: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 2)
[1650465529.473269] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x2071930: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 3)
[1650465529.473020] [ndv4:15226:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.473333] [ndv4:15226:0] ib_iface.c:994 UCX DEBUG iface=0x2cfe7b0: created UD QP 0xbf7c on mlx5_ib7:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.473340] [ndv4:15226:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465529.474122] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.474383] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.474948] [ndv4:14949:0] ib_md.c:812 UCX DEBUG registered memory 0x2ad4c0874000..0x2ad4c08f9000 on mlx5_ib0 lkey 0x81400 rkey 0x81400 access 0xf flags 0x3e4
[1650465529.474957] [ndv4:14949:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ad4c0874018 of 544744 bytes with 128 elements
[1650465529.474963] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.475833] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x10f3970: adding gid fe80::15:5dff:fd33:fffb to hash on device mlx5_ib0 port 1 index 0)
[1650465529.476349] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x10f3970: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650465529.474130] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.475022] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.476069] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.474060] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x2071930: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 4)
[1650465529.474590] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x2071930: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 5)
[1650465529.475213] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x2071930: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 6)
[1650465529.475654] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x2071930: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 7)
[1650465529.475663] [ndv4:14205:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib2:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.475706] [ndv4:14205:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.475711] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x3d34ef0 [id=103 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.475740] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 103 events 0x5 mode thread_spinlock
[1650465529.475758] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[20]=0x2071930 using ud_mlx5/mlx5_ib2:1 on worker 0x29ca770
[1650465529.476709] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x10f3970: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650465529.477711] [ndv4:14421:0] async.c:228 UCX DEBUG added async handler 0xf330a0 [id=71 ref 1] uct_ib_async_event_handler() to hash
[1650465529.477729] [ndv4:14421:0] async.c:506 UCX DEBUG listening to async event fd 71 events 0x1 mode thread_spinlock
[1650465529.477733] [ndv4:14421:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib7' (InfiniBand channel adapter) with 1 ports
[1650465529.478644] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.478967] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.478974] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465529.479027] [ndv4:14421:0] ib_md.c:1319 UCX DEBUG mlx5_ib7: using registration cache
[1650465529.478949] [ndv4:13443:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465529.478960] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.478964] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.478981] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.480129] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.480135] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465529.480892] [ndv4:14421:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465529.480899] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465529.481256] [ndv4:13443:0] ib_iface.c:994 UCX DEBUG iface=0x3650050: created RC QP 0xc45b on mlx5_ib1:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465529.481206] [ndv4:14421:0] ib_md.c:1604 UCX DEBUG mlx5_ib7: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465529.481878] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.481906] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.486418] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.486620] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.486988] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.487303] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.485556] [ndv4:13443:0] ucp_worker.c:1159 UCX DEBUG created interface[11]=0x3650050 using rc_verbs/mlx5_ib1:1 on worker 0x2d7f8d0
[1650465529.485896] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.486367] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.487323] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.486847] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x10f3970: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650465529.486460] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.486621] [ndv4:15226:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.487390] [ndv4:15226:0] ib_md.c:812 UCX DEBUG registered memory 0x2b17a4af1000..0x2b17a4b76000 on mlx5_ib7 lkey 0x81100 rkey 0x81100 access 0xf flags 0x3e4
[1650465529.487396] [ndv4:15226:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b17a4af1018 of 544744 bytes with 128 elements
[1650465529.487401] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.488229] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x2cfe7b0: adding gid fe80::15:5dff:fd34:2 to hash on device mlx5_ib7 port 1 index 0)
[1650465529.487710] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.488777] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x2cfe7b0: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 1)
[1650465529.488437] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465529.489680] [ndv4:13443:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465529.489948] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465529.489966] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.489969] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.490020] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.491113] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.491119] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465529.491687] [ndv4:13443:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.491705] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.491709] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.491734] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.492091] [ndv4:14205:0] ib_iface.c:994 UCX DEBUG iface=0x3f520a0: created RC QP 0xc43b on mlx5_ib3:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465529.492247] [ndv4:13443:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x3663010 of 8176 bytes with 127 elements
[1650465529.492679] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.492687] [ndv4:13443:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib1:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.492724] [ndv4:13443:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650465529.492729] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.492742] [ndv4:13443:0] async.c:228 UCX DEBUG added async handler 0x24ade90 [id=92 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.492772] [ndv4:13443:0] async.c:506 UCX DEBUG listening to async event fd 92 events 0x1 mode thread_spinlock
[1650465529.492797] [ndv4:13443:0] ucp_worker.c:1159 UCX DEBUG created interface[12]=0x34d1010 using rc_mlx5/mlx5_ib1:1 on worker 0x2d7f8d0
[1650465529.493333] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.493679] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.494682] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.495042] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.495400] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[21]=0x3f520a0 using rc_verbs/mlx5_ib3:1 on worker 0x29ca770
[1650465529.495741] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.495834] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.496465] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.496714] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.496342] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x10f3970: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650465529.497140] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x10f3970: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650465529.497427] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x10f3970: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650465529.496309] [ndv4:14421:0] topo.c:99 UCX DEBUG bus id 0x108000000 doesn't exist. sys_dev = 7
[1650465529.496318] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465529.496282] [ndv4:13443:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465529.497006] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x2cfe7b0: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 2)
[1650465529.497569] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x2cfe7b0: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 3)
[1650465529.498136] [ndv4:13443:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.498154] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.498157] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.498215] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.498387] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x10f3970: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650465529.498396] [ndv4:14949:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib0:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.498430] [ndv4:14949:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.498435] [ndv4:14949:0] async.c:228 UCX DEBUG added async handler 0x1041670 [id=89 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.498466] [ndv4:14949:0] async.c:506 UCX DEBUG listening to async event fd 89 events 0x5 mode thread_spinlock
[1650465529.498481] [ndv4:14949:0] ucp_worker.c:1159 UCX DEBUG created interface[10]=0x10f3970 using ud_mlx5/mlx5_ib0:1 on worker 0x19ac660
[1650465529.498675] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.499419] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.498683] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465529.498690] [ndv4:13443:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib1:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.498722] [ndv4:13443:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650465529.498725] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.498733] [ndv4:13443:0] async.c:228 UCX DEBUG added async handler 0x24a9170 [id=94 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.498760] [ndv4:13443:0] async.c:506 UCX DEBUG listening to async event fd 94 events 0x1 mode thread_spinlock
[1650465529.499418] [ndv4:13443:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465529.498368] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x2cfe7b0: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 4)
[1650465529.498569] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x2cfe7b0: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 5)
[1650465529.498941] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x2cfe7b0: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 6)
[1650465529.499787] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.499943] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.500483] [ndv4:14949:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465529.501699] [ndv4:14949:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465529.501709] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.501712] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.501726] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.501983] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.501988] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465529.502406] [ndv4:14949:0] ib_iface.c:994 UCX DEBUG iface=0x227d050: created RC QP 0xc45e on mlx5_ib1:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465529.503491] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465529.504497] [ndv4:15226:0] ud_iface.c:393 UCX DEBUG iface 0x2cfe7b0: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 7)
[1650465529.504518] [ndv4:15226:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib7:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.503154] [ndv4:14949:0] ucp_worker.c:1159 UCX DEBUG created interface[11]=0x227d050 using rc_verbs/mlx5_ib1:1 on worker 0x19ac660
[1650465529.503371] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.503633] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.504106] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.504207] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.504696] [ndv4:14949:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465529.504714] [ndv4:15226:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.504724] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x16ba9b0 [id=138 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.504763] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 138 events 0x5 mode thread_spinlock
[1650465529.504796] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[45]=0x2cfe7b0 using ud_mlx5/mlx5_ib7:1 on worker 0x1f938d0
[1650465529.504868] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool uct_scopy_iface_tx_mp: align 64, maxelems 4294967295, elemsize 736
[1650465529.505167] [ndv4:15226:0] ucp_worker.c:1159 UCX DEBUG created interface[46]=0x57ac880 using cma/memory on worker 0x1f938d0
[1650465529.505175] [ndv4:15226:0] ucp_worker.c:982 UCX DEBUG selected scalable tl bitmap: 0x7fffffffffff 0x0 (47 tls)
[1650465529.505664] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.505685] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.505689] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.505741] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.506027] [ndv4:14949:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.506037] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.506040] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.506055] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.506351] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x4273010 of 8176 bytes with 127 elements
[1650465529.506399] [ndv4:14949:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x2290010 of 8176 bytes with 127 elements
[1650465529.506686] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.506694] [ndv4:14949:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib1:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.506732] [ndv4:14949:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650465529.506736] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.506747] [ndv4:14949:0] async.c:228 UCX DEBUG added async handler 0x2285fc0 [id=92 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.506775] [ndv4:14949:0] async.c:506 UCX DEBUG listening to async event fd 92 events 0x1 mode thread_spinlock
[1650465529.506784] [ndv4:14949:0] ucp_worker.c:1159 UCX DEBUG created interface[12]=0x20fe010 using rc_mlx5/mlx5_ib1:1 on worker 0x19ac660
[1650465529.507121] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.507144] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.507394] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.507418] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.506698] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.506709] [ndv4:14205:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib3:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.506773] [ndv4:14205:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib3 length=2048) failed: Invalid argument
[1650465529.506777] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.506793] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x3e52ae0 [id=106 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.506822] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 106 events 0x1 mode thread_spinlock
[1650465529.506844] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[22]=0x409f030 using rc_mlx5/mlx5_ib3:1 on worker 0x29ca770
[1650465529.507222] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.507235] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.507389] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.507418] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.508271] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465529.508440] [ndv4:12741:0] mm_posix.c:518 UCX DEBUG allocated posix shared memory at 0x2b7353c83000 length 12288
[1650465529.508507] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool mm_recv_desc: align 64, maxelems 4294967295, elemsize 8368
[1650465529.508962] [ndv4:13443:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x37fb010: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xc481
[1650465529.510127] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.510266] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.510279] [ndv4:13443:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b7507de8008 of 151544 bytes with 1052 elements
[1650465529.509900] [ndv4:12741:0] mm_posix.c:305 UCX DEBUG shared memory mmap(addr=(nil), length=6291456, flags= HUGETLB, fd=76) failed: Invalid argument
[1650465529.509909] [ndv4:12741:0] mm_posix.c:518 UCX DEBUG allocated posix shared memory at 0x2b7358000000 length 4296704
[1650465529.509915] [ndv4:12741:0] mpool.c:205 UCX DEBUG mpool mm_recv_desc: allocated chunk 0x2b7358000018 of 4296680 bytes with 512 elements
[1650465529.510170] [ndv4:12741:0] mm_iface.c:600 UCX DEBUG created mm iface 0x2e58d10 FIFO id 0x400000006f2aa999 va 0x2b7353c83000 size 12288 (128 x 64 elems)
[1650465529.510322] [ndv4:12741:0] ucp_worker.c:1159 UCX DEBUG created interface[0]=0x2e58d10 using posix/memory on worker 0x37328d0
[1650465529.510345] [ndv4:12741:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 8447 bytes with hugetlb
[1650465529.510384] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool mm_recv_desc: align 64, maxelems 4294967295, elemsize 8368
[1650465529.510402] [ndv4:12741:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 4292720 bytes with hugetlb
[1650465529.510411] [ndv4:12741:0] mpool.c:205 UCX DEBUG mpool mm_recv_desc: allocated chunk 0x2b7358419018 of 4296680 bytes with 512 elements
[1650465529.511050] [ndv4:12741:0] mm_iface.c:600 UCX DEBUG created mm iface 0x2e592e0 FIFO id 0x64803d va 0x2b7353c86000 size 12288 (128 x 64 elems)
[1650465529.511060] [ndv4:12741:0] ucp_worker.c:1159 UCX DEBUG created interface[1]=0x2e592e0 using sysv/memory on worker 0x37328d0
[1650465529.511072] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool self_msg_desc: align 64, maxelems 4294967295, elemsize 8200
[1650465529.511075] [ndv4:12741:0] self.c:220 UCX DEBUG created self iface id 0x19986a38e0b4fb7d send_size 8192
[1650465529.511081] [ndv4:12741:0] ucp_worker.c:1159 UCX DEBUG created interface[2]=0x2e6cec0 using self/memory0 on worker 0x37328d0
[1650465529.511103] [ndv4:12741:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465529.511108] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465529.511111] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465529.515348] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x516a7a0 [id=75 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.515389] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 75 events 0x0 mode thread_spinlock
[1650465529.515437] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x516a5d0 [id=76 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.515443] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 76 events 0x0 mode thread_spinlock
[1650465529.515458] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x516a610 [id=77 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.515465] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 77 events 0x0 mode thread_spinlock
[1650465529.514360] [ndv4:12741:0] async.c:228 UCX DEBUG added async handler 0x2e6f000 [id=78 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465529.514398] [ndv4:12741:0] async.c:506 UCX DEBUG listening to async event fd 78 events 0x5 mode thread_spinlock
[1650465529.514415] [ndv4:12741:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x2e6d820: listening for connections (fd=78) on 10.5.0.5:39592
[1650465529.514886] [ndv4:12741:0] ucp_worker.c:1159 UCX DEBUG created interface[3]=0x2e6d820 using tcp/eth0 on worker 0x37328d0
[1650465529.514906] [ndv4:12741:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465529.514910] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465529.514913] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465529.514952] [ndv4:12741:0] async.c:228 UCX DEBUG added async handler 0x2dbb5c0 [id=80 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465529.514979] [ndv4:12741:0] async.c:506 UCX DEBUG listening to async event fd 80 events 0x5 mode thread_spinlock
[1650465529.514983] [ndv4:12741:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x2e6dea0: listening for connections (fd=80) on 127.0.0.1:39725
[1650465529.514999] [ndv4:12741:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465529.515005] [ndv4:12741:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465529.515669] [ndv4:12741:0] ucp_worker.c:1159 UCX DEBUG created interface[4]=0x2e6dea0 using tcp/lo on worker 0x37328d0
[1650465529.515687] [ndv4:12741:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465529.515690] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465529.515692] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465529.515733] [ndv4:12741:0] async.c:228 UCX DEBUG added async handler 0x2dc9960 [id=82 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465529.515766] [ndv4:12741:0] async.c:506 UCX DEBUG listening to async event fd 82 events 0x5 mode thread_spinlock
[1650465529.515770] [ndv4:12741:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x2e55530: listening for connections (fd=82) on 172.16.1.242:41065
[1650465529.514229] [ndv4:13443:0] ib_md.c:812 UCX DEBUG registered memory 0x2b750ee00000..0x2b7511400000 on mlx5_ib1 lkey 0x81100 rkey 0x81100 access 0xf flags 0x3e4
[1650465529.514248] [ndv4:13443:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b750ee00018 of 39845864 bytes with 4752 elements
[1650465529.514386] [ndv4:13443:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x37fb010
[1650465529.514424] [ndv4:13443:0] ucp_worker.c:1159 UCX DEBUG created interface[13]=0x37fb010 using dc_mlx5/mlx5_ib1:1 on worker 0x2d7f8d0
[1650465529.514805] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.515525] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.515893] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.516325] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.516361] [ndv4:12741:0] ucp_worker.c:1159 UCX DEBUG created interface[5]=0x2e55530 using tcp/ib0 on worker 0x37328d0
[1650465529.517559] [ndv4:14949:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465529.519112] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x2cfe600 [id=79 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.519151] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 79 events 0x0 mode thread_spinlock
[1650465529.519169] [ndv4:15226:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465529.519175] [ndv4:15226:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465529.519274] [ndv4:15226:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465529.519279] [ndv4:15226:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465529.519346] [ndv4:15226:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465529.519352] [ndv4:15226:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465529.519715] [ndv4:15226:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465529.519722] [ndv4:15226:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465529.520196] [ndv4:15226:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465529.520202] [ndv4:15226:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465529.519059] [ndv4:14949:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.519075] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.519078] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.519137] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.519661] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465529.519668] [ndv4:14949:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib1:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.519701] [ndv4:14949:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650465529.519706] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.519713] [ndv4:14949:0] async.c:228 UCX DEBUG added async handler 0x10d6590 [id=94 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.519746] [ndv4:14949:0] async.c:506 UCX DEBUG listening to async event fd 94 events 0x1 mode thread_spinlock
[1650465529.520165] [ndv4:14949:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465529.519534] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.519556] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.519559] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.519670] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.520151] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465529.520159] [ndv4:14205:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib3:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.520200] [ndv4:14205:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib3 length=2048) failed: Invalid argument
[1650465529.520203] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.520216] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x20f1c90 [id=108 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.520248] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 108 events 0x1 mode thread_spinlock
[1650465529.520834] [ndv4:15226:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465529.520841] [ndv4:15226:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465529.520926] [ndv4:15226:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465529.520931] [ndv4:15226:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465529.521161] [ndv4:15226:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465529.521166] [ndv4:15226:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465529.521002] [ndv4:14205:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465529.521807] [ndv4:15226:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465529.521814] [ndv4:15226:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465529.521883] [ndv4:15226:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465529.521887] [ndv4:15226:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465529.522279] [ndv4:15226:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465529.522284] [ndv4:15226:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465529.522723] [ndv4:15226:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465529.522730] [ndv4:15226:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465529.523001] [ndv4:15226:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465529.523005] [ndv4:15226:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465529.523341] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x2cfe640 [id=81 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.523381] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 81 events 0x0 mode thread_spinlock
[1650465529.525389] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.525761] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.526339] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.526402] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.527225] [ndv4:12741:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465529.528967] [ndv4:12741:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465529.529005] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.529009] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.529027] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.528351] [ndv4:14949:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x2428010: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xc48d
[1650465529.529442] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.529451] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465529.530078] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x2cfe680 [id=83 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.530107] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 83 events 0x0 mode thread_spinlock
[1650465529.530189] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x2cfe6c0 [id=84 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.530213] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 84 events 0x0 mode thread_spinlock
[1650465529.530233] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x2cfe730 [id=86 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.530257] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 86 events 0x0 mode thread_spinlock
[1650465529.530306] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x2cfe770 [id=90 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.530327] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 90 events 0x0 mode thread_spinlock
[1650465529.530343] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x3a1fec0 [id=91 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.530365] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 91 events 0x0 mode thread_spinlock
[1650465529.530381] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x3a1ff00 [id=93 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.530407] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 93 events 0x0 mode thread_spinlock
[1650465529.530444] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x3a1ff40 [id=97 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.530470] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 97 events 0x0 mode thread_spinlock
[1650465529.530486] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x3a1ff80 [id=98 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.530508] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 98 events 0x0 mode thread_spinlock
[1650465529.530523] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x2dd9b00 [id=100 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.530543] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 100 events 0x0 mode thread_spinlock
[1650465529.530672] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x2dd9b40 [id=104 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.530707] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 104 events 0x0 mode thread_spinlock
[1650465529.530724] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x2dd9b80 [id=105 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.530751] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 105 events 0x0 mode thread_spinlock
[1650465529.530767] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x2dd9bc0 [id=107 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.530792] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 107 events 0x0 mode thread_spinlock
[1650465529.530829] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x2dd9c00 [id=111 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.530850] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 111 events 0x0 mode thread_spinlock
[1650465529.530867] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x2dd9c40 [id=112 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.530891] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 112 events 0x0 mode thread_spinlock
[1650465529.530905] [ndv4:15226:0] async.c:228 UCX DE[1650465529.530963] [ndv4:14205:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x4275050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xc44e
[1650465529.530331] [ndv4:12741:0] ib_iface.c:994 UCX DEBUG iface=0x2e63670: created RC QP 0xde78 on mlx5_ib0:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
BUG added async handler 0x2dd9c80 [id=114 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.530939] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 114 events 0x0 mode thread_spinlock
[1650465529.530975] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x2dd9cc0 [id=118 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.530999] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 118 events 0x0 mode thread_spinlock
[1650465529.531015] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x57acdd0 [id=119 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.531034] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 119 events 0x0 mode thread_spinlock
[1650465529.531048] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x57ace10 [id=121 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.531074] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 121 events 0x0 mode thread_spinlock
[1650465529.531110] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x57ace50 [id=125 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.531133] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 125 events 0x0 mode thread_spinlock
[1650465529.531149] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x57ace90 [id=126 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.531171] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 126 events 0x0 mode thread_spinlock
[1650465529.531187] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x57aced0 [id=128 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.531211] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 128 events 0x0 mode thread_spinlock
[1650465529.531248] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x57acf10 [id=132 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.531266] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 132 events 0x0 mode thread_spinlock
[1650465529.531281] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x57acf50 [id=133 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.531299] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 133 events 0x0 mode thread_spinlock
[1650465529.531314] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x57acf90 [id=135 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465529.531332] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 135 events 0x0 mode thread_spinlock
[1650465529.531750] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.531844] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.531856] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2afdb4629008 of 151544 bytes with 1052 elements
[1650465529.532635] [ndv4:14421:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465529.532645] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465529.533180] [ndv4:13443:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465529.534719] [ndv4:13443:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.535101] [ndv4:13443:0] ib_iface.c:994 UCX DEBUG iface=0x36fa420: created UD QP 0xc46d on mlx5_ib1:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.534869] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.535121] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.535152] [ndv4:14949:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2ad4b7f9f008 of 151544 bytes with 1052 elements
[1650465529.535705] [ndv4:15226:0] async.c:228 UCX DEBUG added async handler 0x5887af0 [id=139 ref 1] uct_rdmacm_cm_event_handler() to hash
[1650465529.535769] [ndv4:15226:0] async.c:506 UCX DEBUG listening to async event fd 139 events 0x1 mode thread_spinlock
[1650465529.535797] [ndv4:15226:0] rdmacm_cm.c:922 UCX DEBUG created rdmacm_cm 0x58874f0 with event_channel 0x57acfd0 (fd=139)
[1650465529.535861] [ndv4:15226:0] tcp_sockcm.c:186 UCX DEBUG created tcp_sockcm 0x38227e0
[1650465529.535882] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool ucp_requests: align 64, maxelems 4294967295, elemsize 440
[1650465529.535886] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool ucp_rkeys: align 64, maxelems 4294967295, elemsize 168
[1650465529.535891] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool ucp_am_bufs: align 64, maxelems 4294967295, elemsize 65624
[1650465529.535894] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool ucp_reg_bufs: align 64, maxelems 4294967295, elemsize 8208
[1650465529.535896] [ndv4:15226:0] mpool.c:88 UCX DEBUG mpool ucp_rndv_frags: align 512, maxelems 4294967295, elemsize 524304
[1650465529.535989] [ndv4:15226:0] parser.c:1893 UCX INFO UCX_* env variables: UCX_POSIX_USE_PROC_LINK=n UCX_LOG_LEVEL=debug
[1650465529.535733] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.536134] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.536627] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.537167] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.537683] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.535890] [ndv4:14205:0] ib_md.c:812 UCX DEBUG registered memory 0x2afdb2000000..0x2afdb4600000 on mlx5_ib3 lkey 0x80a00 rkey 0x80a00 access 0xf flags 0x3e4
[1650465529.535905] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2afdb2000018 of 39845864 bytes with 4752 elements
[1650465529.536049] [ndv4:14205:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x4275050
[1650465529.536081] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[23]=0x4275050 using dc_mlx5/mlx5_ib3:1 on worker 0x29ca770
[1650465529.536534] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.536728] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.537153] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.537372] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.538434] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465529.538210] [ndv4:13552:0] mm_posix.c:518 UCX DEBUG allocated posix shared memory at 0x2b265bf47000 length 12288
[1650465529.538297] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool mm_recv_desc: align 64, maxelems 4294967295, elemsize 8368
[1650465529.539998] [ndv4:13552:0] mm_posix.c:305 UCX DEBUG shared memory mmap(addr=(nil), length=6291456, flags= HUGETLB, fd=76) failed: Invalid argument
[1650465529.540011] [ndv4:13552:0] mm_posix.c:518 UCX DEBUG allocated posix shared memory at 0x2b2660ec7000 length 4296704
[1650465529.540018] [ndv4:13552:0] mpool.c:205 UCX DEBUG mpool mm_recv_desc: allocated chunk 0x2b2660ec7018 of 4296680 bytes with 512 elements
[1650465529.540305] [ndv4:13552:0] mm_iface.c:600 UCX DEBUG created mm iface 0x1b36d10 FIFO id 0x400000002adac201 va 0x2b265bf47000 size 12288 (128 x 64 elems)
[1650465529.538161] [ndv4:13443:0] ib_md.c:812 UCX DEBUG registered memory 0x2b7507e0d000..0x2b7507e92000 on mlx5_ib1 lkey 0x81200 rkey 0x81200 access 0xf flags 0x3e4
[1650465529.538168] [ndv4:13443:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b7507e0d018 of 544744 bytes with 128 elements
[1650465529.538173] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.538455] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x36fa420: adding gid fe80::15:5dff:fd33:fffc to hash on device mlx5_ib1 port 1 index 0)
[1650465529.539302] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x36fa420: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 1)
[1650465529.539735] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x36fa420: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 2)
[1650465529.540288] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x36fa420: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 3)
[1650465529.539134] [ndv4:14949:0] ib_md.c:812 UCX DEBUG registered memory 0x2ad4c0a00000..0x2ad4c3000000 on mlx5_ib1 lkey 0x81300 rkey 0x81300 access 0xf flags 0x3e4
[1650465529.539148] [ndv4:14949:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2ad4c0a00018 of 39845864 bytes with 4752 elements
[1650465529.539300] [ndv4:14949:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x2428010
[1650465529.539337] [ndv4:14949:0] ucp_worker.c:1159 UCX DEBUG created interface[13]=0x2428010 using dc_mlx5/mlx5_ib1:1 on worker 0x19ac660
[1650465529.539702] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.539914] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.540084] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.540118] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.541109] [ndv4:14949:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465529.540737] [ndv4:13552:0] ucp_worker.c:1159 UCX DEBUG created interface[0]=0x1b36d10 using posix/memory on worker 0x24108d0
[1650465529.540767] [ndv4:13552:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 8447 bytes with hugetlb
[1650465529.540823] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool mm_recv_desc: align 64, maxelems 4294967295, elemsize 8368
[1650465529.540845] [ndv4:13552:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 4292720 bytes with hugetlb
[1650465529.540855] [ndv4:13552:0] mpool.c:205 UCX DEBUG mpool mm_recv_desc: allocated chunk 0x2b26612e0018 of 4296680 bytes with 512 elements
[1650465529.541452] [ndv4:13552:0] mm_iface.c:600 UCX DEBUG created mm iface 0x1b372e0 FIFO id 0x64803f va 0x2b265bf4a000 size 12288 (128 x 64 elems)
[1650465529.541461] [ndv4:13552:0] ucp_worker.c:1159 UCX DEBUG created interface[1]=0x1b372e0 using sysv/memory on worker 0x24108d0
[1650465529.541475] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool self_msg_desc: align 64, maxelems 4294967295, elemsize 8200
[1650465529.541478] [ndv4:13552:0] self.c:220 UCX DEBUG created self iface id 0xdd7e8e124eead7c send_size 8192
[1650465529.541486] [ndv4:13552:0] ucp_worker.c:1159 UCX DEBUG created interface[2]=0x1b4aec0 using self/memory0 on worker 0x24108d0
[1650465529.541513] [ndv4:13552:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465529.541519] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465529.541522] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465529.540844] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x36fa420: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 4)
[1650465529.541408] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x36fa420: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 5)
[1650465529.542212] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x36fa420: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 6)
[1650465529.542346] [ndv4:14949:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.542796] [ndv4:14949:0] ib_iface.c:994 UCX DEBUG iface=0x2327300: created UD QP 0xc46e on mlx5_ib1:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.543226] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x36fa420: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 7)
[1650465529.544301] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.544817] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.545178] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.545911] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.546180] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.544969] [ndv4:13552:0] async.c:228 UCX DEBUG added async handler 0x1b4d000 [id=78 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465529.545015] [ndv4:13552:0] async.c:506 UCX DEBUG listening to async event fd 78 events 0x5 mode thread_spinlock
[1650465529.545038] [ndv4:13552:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x1b4b820: listening for connections (fd=78) on 10.5.0.5:45569
[1650465529.545957] [ndv4:13552:0] ucp_worker.c:1159 UCX DEBUG created interface[3]=0x1b4b820 using tcp/eth0 on worker 0x24108d0
[1650465529.545981] [ndv4:13552:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465529.545984] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465529.545987] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465529.546274] [ndv4:13552:0] async.c:228 UCX DEBUG added async handler 0x1a995c0 [id=80 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465529.546303] [ndv4:13552:0] async.c:506 UCX DEBUG listening to async event fd 80 events 0x5 mode thread_spinlock
[1650465529.546307] [ndv4:13552:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x1b4bea0: listening for connections (fd=80) on 127.0.0.1:53720
[1650465529.546323] [ndv4:13552:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465529.546329] [ndv4:13552:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465529.546661] [ndv4:13552:0] ucp_worker.c:1159 UCX DEBUG created interface[4]=0x1b4bea0 using tcp/lo on worker 0x24108d0
[1650465529.546679] [ndv4:13552:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465529.546683] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465529.546685] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465529.546727] [ndv4:13552:0] async.c:228 UCX DEBUG added async handler 0x1aa7960 [id=82 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465529.546755] [ndv4:13552:0] async.c:506 UCX DEBUG listening to async event fd 82 events 0x5 mode thread_spinlock
[1650465529.546759] [ndv4:13552:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x1b33530: listening for connections (fd=82) on 172.16.1.242:37867
[1650465529.544808] [ndv4:12741:0] ucp_worker.c:1159 UCX DEBUG created interface[6]=0x2e63670 using rc_verbs/mlx5_ib0:1 on worker 0x37328d0
[1650465529.545164] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.545565] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.546149] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.546364] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.543493] [ndv4:13443:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.543502] [ndv4:13443:0] async.c:228 UCX DEBUG added async handler 0x24a8500 [id=95 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.543533] [ndv4:13443:0] async.c:506 UCX DEBUG listening to async event fd 95 events 0x5 mode thread_spinlock
[1650465529.543556] [ndv4:13443:0] ucp_worker.c:1159 UCX DEBUG created interface[14]=0x36fa420 using ud_verbs/mlx5_ib1:1 on worker 0x2d7f8d0
[1650465529.543643] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.543886] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.544712] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.545089] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.545662] [ndv4:13443:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465529.546671] [ndv4:14949:0] ib_md.c:812 UCX DEBUG registered memory 0x2ad4c30fa000..0x2ad4c317f000 on mlx5_ib1 lkey 0x80600 rkey 0x80600 access 0xf flags 0x3e4
[1650465529.546678] [ndv4:14949:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ad4c30fa018 of 544744 bytes with 128 elements
[1650465529.546682] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.546995] [ndv4:12741:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465529.547259] [ndv4:12741:0] ib_device.c:1394 UCX DEBUG max IB CQE size is 128
[1650465529.548703] [ndv4:12741:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.548716] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.548720] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.548777] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.547065] [ndv4:13443:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.547410] [ndv4:13443:0] ib_iface.c:994 UCX DEBUG iface=0x35c1280: created UD QP 0xc46f on mlx5_ib1:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.547417] [ndv4:13443:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465529.548042] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.547202] [ndv4:13552:0] ucp_worker.c:1159 UCX DEBUG created interface[5]=0x1b33530 using tcp/ib0 on worker 0x24108d0
[1650465529.547700] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.548016] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.548475] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.548863] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.549086] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.549143] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.549315] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.549537] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.550729] [ndv4:13443:0] ib_md.c:812 UCX DEBUG registered memory 0x2b7507e92000..0x2b7507f17000 on mlx5_ib1 lkey 0x80700 rkey 0x80700 access 0xf flags 0x3e4
[1650465529.550741] [ndv4:13443:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b7507e92018 of 544744 bytes with 128 elements
[1650465529.550747] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.550856] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x35c1280: adding gid fe80::15:5dff:fd33:fffc to hash on device mlx5_ib1 port 1 index 0)
[1650465529.551164] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x35c1280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 1)
[1650465529.549862] [ndv4:13552:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465529.550637] [ndv4:13552:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465529.550678] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.550682] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.550701] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.550966] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.550974] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465529.551637] [ndv4:13552:0] ib_iface.c:994 UCX DEBUG iface=0x1b41670: created RC QP 0xde79 on mlx5_ib0:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465529.549257] [ndv4:12741:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x37fc010 of 8176 bytes with 127 elements
[1650465529.549533] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.549559] [ndv4:12741:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib0:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.549668] [ndv4:12741:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650465529.549675] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.552026] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x35c1280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 2)
[1650465529.549991] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.550466] [ndv4:14205:0] ib_iface.c:994 UCX DEBUG iface=0x4451060: created UD QP 0xc444 on mlx5_ib3:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.551146] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.552209] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.552692] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.552860] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.552957] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.553389] [ndv4:14205:0] ib_md.c:812 UCX DEBUG registered memory 0x2afdb464e000..0x2afdb46d3000 on mlx5_ib3 lkey 0x80b00 rkey 0x80b00 access 0xf flags 0x3e4
[1650465529.553396] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2afdb464e018 of 544744 bytes with 128 elements
[1650465529.553401] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.557175] [ndv4:12741:0] async.c:228 UCX DEBUG added async handler 0x2dc8c70 [id=85 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.557212] [ndv4:12741:0] async.c:506 UCX DEBUG listening to async event fd 85 events 0x1 mode thread_spinlock
[1650465529.557236] [ndv4:12741:0] ucp_worker.c:1159 UCX DEBUG created interface[7]=0x2e710c0 using rc_mlx5/mlx5_ib0:1 on worker 0x37328d0
[1650465529.557885] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.558081] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.558390] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.558782] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.558721] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x2327300: adding gid fe80::15:5dff:fd33:fffc to hash on device mlx5_ib1 port 1 index 0)
[1650465529.558852] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x2327300: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 1)
[1650465529.559084] [ndv4:12741:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465529.560087] [ndv4:12741:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.560096] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.560099] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.560113] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.560568] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465529.560627] [ndv4:12741:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib0:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.560659] [ndv4:12741:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650465529.560663] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.560673] [ndv4:12741:0] async.c:228 UCX DEBUG added async handler 0x2dc7b40 [id=87 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.560699] [ndv4:12741:0] async.c:506 UCX DEBUG listening to async event fd 87 events 0x1 mode thread_spinlock
[1650465529.561234] [ndv4:12741:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465529.567210] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x35c1280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 3)
[1650465529.568080] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x35c1280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 4)
[1650465529.568668] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x35c1280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 5)
[1650465529.568827] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x4451060: adding gid fe80::15:5dff:fd33:fffe to hash on device mlx5_ib3 port 1 index 0)
[1650465529.569541] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x4451060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 1)
[1650465529.571134] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x2327300: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 2)
[1650465529.570189] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x4451060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 2)
[1650465529.570888] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x4451060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 3)
[1650465529.571118] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x4451060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 4)
[1650465529.571418] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x4451060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 5)
[1650465529.570310] [ndv4:12741:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x3a66010: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xdea3
[1650465529.571049] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.571286] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.571308] [ndv4:12741:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b7353c8b008 of 151544 bytes with 1052 elements
[1650465529.572697] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x4451060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 6)
[1650465529.572722] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x4451060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 7)
[1650465529.573094] [ndv4:14205:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.573112] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x3810fc0 [id=109 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.573149] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 109 events 0x5 mode thread_spinlock
[1650465529.573180] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[24]=0x4451060 using ud_verbs/mlx5_ib3:1 on worker 0x29ca770
[1650465529.573400] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.573479] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.573985] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.574007] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.574622] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465529.575319] [ndv4:12741:0] ib_md.c:812 UCX DEBUG registered memory 0x2b7358a00000..0x2b735b000000 on mlx5_ib0 lkey 0x81900 rkey 0x81900 access 0xf flags 0x3e4
[1650465529.575332] [ndv4:12741:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b7358a00018 of 39845864 bytes with 4752 elements
[1650465529.575476] [ndv4:12741:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x3a66010
[1650465529.575505] [ndv4:12741:0] ucp_worker.c:1159 UCX DEBUG created interface[8]=0x3a66010 using dc_mlx5/mlx5_ib0:1 on worker 0x37328d0
[1650465529.575745] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.575811] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.576138] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.576332] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.577065] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.576542] [ndv4:14421:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465529.576551] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465529.577642] [ndv4:14205:0] ib_iface.c:994 UCX DEBUG iface=0x456f460: created UD QP 0xc445 on mlx5_ib3:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.577691] [ndv4:14205:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465529.578071] [ndv4:12741:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465529.578359] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.578940] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x2327300: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 3)
[1650465529.579078] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x2327300: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 4)
[1650465529.579558] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x2327300: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 5)
[1650465529.578897] [ndv4:13552:0] ucp_worker.c:1159 UCX DEBUG created interface[6]=0x1b41670 using rc_verbs/mlx5_ib0:1 on worker 0x24108d0
[1650465529.579033] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.579086] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.579257] [ndv4:12741:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.579649] [ndv4:12741:0] ib_iface.c:994 UCX DEBUG iface=0x2e79a30: created UD QP 0xde86 on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.581016] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x2327300: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 6)
[1650465529.580314] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.581290] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.581691] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.582274] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.582351] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.581764] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x2327300: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 7)
[1650465529.582054] [ndv4:14949:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.582060] [ndv4:14949:0] async.c:228 UCX DEBUG added async handler 0x23270c0 [id=95 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.582092] [ndv4:14949:0] async.c:506 UCX DEBUG listening to async event fd 95 events 0x5 mode thread_spinlock
[1650465529.582106] [ndv4:14949:0] ucp_worker.c:1159 UCX DEBUG created interface[14]=0x2327300 using ud_verbs/mlx5_ib1:1 on worker 0x19ac660
[1650465529.582774] [ndv4:12741:0] ib_md.c:812 UCX DEBUG registered memory 0x2b7353cb0000..0x2b7353d35000 on mlx5_ib0 lkey 0x81a00 rkey 0x81a00 access 0xf flags 0x3e4
[1650465529.582782] [ndv4:12741:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b7353cb0018 of 544744 bytes with 128 elements
[1650465529.582787] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.584405] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x2e79a30: adding gid fe80::15:5dff:fd33:fffb to hash on device mlx5_ib0 port 1 index 0)
[1650465529.585211] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x2e79a30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650465529.585405] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x2e79a30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650465529.585558] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x2e79a30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650465529.589830] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x35c1280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 6)
[1650465529.590457] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x35c1280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 7)
[1650465529.590468] [ndv4:13443:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib1:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.590673] [ndv4:13443:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.590683] [ndv4:13443:0] async.c:228 UCX DEBUG added async handler 0x34d9be0 [id=96 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.590716] [ndv4:13443:0] async.c:506 UCX DEBUG listening to async event fd 96 events 0x5 mode thread_spinlock
[1650465529.590734] [ndv4:13443:0] ucp_worker.c:1159 UCX DEBUG created interface[15]=0x35c1280 using ud_mlx5/mlx5_ib1:1 on worker 0x2d7f8d0
[1650465529.591169] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.591236] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.591440] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.591446] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.590814] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.590968] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.591018] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.591108] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.591556] [ndv4:14205:0] ib_md.c:812 UCX DEBUG registered memory 0x2afdb46d3000..0x2afdb4758000 on mlx5_ib3 lkey 0x80c00 rkey 0x80c00 access 0xf flags 0x3e4
[1650465529.591562] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2afdb46d3018 of 544744 bytes with 128 elements
[1650465529.591566] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.591177] [ndv4:14419:0] debug.c:1198 UCX DEBUG using signal stack 0x2b79b7fdb000 size 141824
[1650465529.591264] [ndv4:14419:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465529.591284] [ndv4:14419:0] init.c:113 UCX DEBUG /cvmfs/ndv4.azure/CentOS-HPC/7.9.2021052401/EasyBuild/software/UCX/1.11.2-GCCcore-11.2.0/lib64/libucs.so.0 loaded at 0x2b79c23f9000
[1650465529.591314] [ndv4:14419:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650465529.591323] [ndv4:14419:0] module.c:69 UCX DEBUG ucs library path: /cvmfs/ndv4.azure/CentOS-HPC/7.9.2021052401/EasyBuild/software/UCX/1.11.2-GCCcore-11.2.0/lib64/libucs.so.0
[1650465529.591330] [ndv4:14419:0] module.c:253 UCX DEBUG loading modules for ucs
[1650465529.591896] [ndv4:13443:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465529.592305] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.592329] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.592785] [ndv4:13552:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465529.593060] [ndv4:13552:0] ib_device.c:1394 UCX DEBUG max IB CQE size is 128
[1650465529.594465] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.594477] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.594555] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.594565] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.594206] [ndv4:13552:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.594216] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.594220] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.594278] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.594943] [ndv4:13552:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x24da010 of 8176 bytes with 127 elements
[1650465529.595303] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.595328] [ndv4:13552:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib0:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.595387] [ndv4:13552:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650465529.595393] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.594811] [ndv4:14419:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465529.594834] [ndv4:14419:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650465529.594869] [ndv4:14419:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650465529.594872] [ndv4:14419:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650465529.594879] [ndv4:14419:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650465529.594887] [ndv4:14419:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650465529.594889] [ndv4:14419:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650465529.594894] [ndv4:14419:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650465529.594896] [ndv4:14419:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650465529.594899] [ndv4:14419:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650465529.594902] [ndv4:14419:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650465529.594904] [ndv4:14419:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650465529.594912] [ndv4:14419:0] module.c:253 UCX DEBUG loading modules for uct
[1650465529.594822] [ndv4:14949:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465529.595864] [ndv4:14419:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650465529.596861] [ndv4:14419:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650465529.597044] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x2e79a30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650465529.597071] [ndv4:14419:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650465529.597720] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x2e79a30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650465529.598726] [ndv4:14419:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650465529.598748] [ndv4:14419:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650465529.598972] [ndv4:14419:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650465529.599228] [ndv4:14419:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650465529.599243] [ndv4:14419:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650465529.599891] [ndv4:14419:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650465529.604958] [ndv4:14421:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465529.605021] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465529.604163] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x456f460: adding gid fe80::15:5dff:fd33:fffe to hash on device mlx5_ib3 port 1 index 0)
[1650465529.604326] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x456f460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 1)
[1650465529.604367] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x456f460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 2)
[1650465529.604383] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x456f460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 3)
[1650465529.604410] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x456f460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 4)
[1650465529.604429] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x456f460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 5)
[1650465529.604468] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x456f460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 6)
[1650465529.604900] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x456f460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 7)
[1650465529.604907] [ndv4:14205:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib3:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.604951] [ndv4:14205:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.604957] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x456ff60 [id=110 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.604993] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 110 events 0x5 mode thread_spinlock
[1650465529.605009] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[25]=0x456f460 using ud_mlx5/mlx5_ib3:1 on worker 0x29ca770
[1650465529.605039] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.605046] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.605181] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.605196] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.605257] [ndv4:13552:0] async.c:228 UCX DEBUG added async handler 0x1aa6c70 [id=85 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.605299] [ndv4:13552:0] async.c:506 UCX DEBUG listening to async event fd 85 events 0x1 mode thread_spinlock
[1650465529.605324] [ndv4:13552:0] ucp_worker.c:1159 UCX DEBUG created interface[7]=0x1b4f0c0 using rc_mlx5/mlx5_ib0:1 on worker 0x24108d0
[1650465529.605390] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.605404] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.605496] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.605501] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.605607] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465529.605812] [ndv4:13552:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465529.606848] [ndv4:13552:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.606859] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.606862] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.606880] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.607350] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465529.607358] [ndv4:13552:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib0:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.607399] [ndv4:13552:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650465529.607403] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.607413] [ndv4:13552:0] async.c:228 UCX DEBUG added async handler 0x1aa5b40 [id=87 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.607439] [ndv4:13552:0] async.c:506 UCX DEBUG listening to async event fd 87 events 0x1 mode thread_spinlock
[1650465529.607810] [ndv4:14419:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650465529.608037] [ndv4:13552:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465529.608270] [ndv4:14419:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650465529.608921] [ndv4:14949:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.609419] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x2e79a30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650465529.609902] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x2e79a30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650465529.609410] [ndv4:14949:0] ib_iface.c:994 UCX DEBUG iface=0x2293070: created UD QP 0xc471 on mlx5_ib1:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.609418] [ndv4:14949:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465529.610217] [ndv4:12741:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.610198] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.610669] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.610676] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.610693] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.610707] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.611248] [ndv4:14949:0] ib_md.c:812 UCX DEBUG registered memory 0x2ad4c317f000..0x2ad4c3204000 on mlx5_ib1 lkey 0x81400 rkey 0x81400 access 0xf flags 0x3e4
[1650465529.611255] [ndv4:14949:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ad4c317f018 of 544744 bytes with 128 elements
[1650465529.611260] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.612442] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x2293070: adding gid fe80::15:5dff:fd33:fffc to hash on device mlx5_ib1 port 1 index 0)
[1650465529.612887] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x2293070: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 1)
[1650465529.616305] [ndv4:12741:0] async.c:228 UCX DEBUG added async handler 0x2db92f0 [id=88 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.616343] [ndv4:12741:0] async.c:506 UCX DEBUG listening to async event fd 88 events 0x5 mode thread_spinlock
[1650465529.616365] [ndv4:12741:0] ucp_worker.c:1159 UCX DEBUG created interface[9]=0x2e79a30 using ud_verbs/mlx5_ib0:1 on worker 0x37328d0
[1650465529.616555] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.616567] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.616728] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.616735] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.617019] [ndv4:12741:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465529.618013] [ndv4:12741:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.618377] [ndv4:12741:0] ib_iface.c:994 UCX DEBUG iface=0x3752750: created UD QP 0xde8f on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.618390] [ndv4:12741:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465529.619162] [ndv4:13443:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465529.619181] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.619185] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.619240] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.619144] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.619175] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.619181] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.619215] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.619220] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.619693] [ndv4:12741:0] ib_md.c:812 UCX DEBUG registered memory 0x2b7353d35000..0x2b7353dba000 on mlx5_ib0 lkey 0x81b00 rkey 0x81b00 access 0xf flags 0x3e4
[1650465529.619699] [ndv4:12741:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b7353d35018 of 544744 bytes with 128 elements
[1650465529.619703] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.620050] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x3752750: adding gid fe80::15:5dff:fd33:fffb to hash on device mlx5_ib0 port 1 index 0)
[1650465529.620082] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x3752750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650465529.620545] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x3752750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650465529.618804] [ndv4:13552:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x2744010: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xdeaf
[1650465529.619008] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.619023] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.619046] [ndv4:13552:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b265bf4f008 of 151544 bytes with 1052 elements
[1650465529.619872] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.619878] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465529.619943] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465529.619965] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.619969] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.620020] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.620727] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.620732] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465529.621033] [ndv4:13443:0] ib_iface.c:994 UCX DEBUG iface=0x3bea0a0: created RC QP 0xc4b5 on mlx5_ib2:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465529.620846] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x3752750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650465529.621670] [ndv4:14205:0] ib_iface.c:994 UCX DEBUG iface=0x466f0a0: created RC QP 0xc4a6 on mlx5_ib4:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465529.623079] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x2293070: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 2)
[1650465529.622399] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[26]=0x466f0a0 using rc_verbs/mlx5_ib4:1 on worker 0x29ca770
[1650465529.622553] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.622625] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.622026] [ndv4:13443:0] ucp_worker.c:1159 UCX DEBUG created interface[16]=0x3bea0a0 using rc_verbs/mlx5_ib2:1 on worker 0x2d7f8d0
[1650465529.622134] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.622144] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.622518] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.622549] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.623306] [ndv4:13443:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465529.624034] [ndv4:13552:0] ib_md.c:812 UCX DEBUG registered memory 0x2b2661800000..0x2b2663e00000 on mlx5_ib0 lkey 0x81c00 rkey 0x81c00 access 0xf flags 0x3e4
[1650465529.624062] [ndv4:13552:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b2661800018 of 39845864 bytes with 4752 elements
[1650465529.624204] [ndv4:13552:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x2744010
[1650465529.624241] [ndv4:13552:0] ucp_worker.c:1159 UCX DEBUG created interface[8]=0x2744010 using dc_mlx5/mlx5_ib0:1 on worker 0x24108d0
[1650465529.624268] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.624278] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.624535] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.624540] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.624714] [ndv4:13443:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.624731] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.624735] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.624785] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.625224] [ndv4:13443:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x3f0b010 of 8176 bytes with 127 elements
[1650465529.625646] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.625652] [ndv4:13443:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib2:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.625687] [ndv4:13443:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib2 length=2048) failed: Invalid argument
[1650465529.625692] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.625702] [ndv4:13443:0] async.c:228 UCX DEBUG added async handler 0x2e53440 [id=99 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.625730] [ndv4:13443:0] async.c:506 UCX DEBUG listening to async event fd 99 events 0x1 mode thread_spinlock
[1650465529.625752] [ndv4:13443:0] ucp_worker.c:1159 UCX DEBUG created interface[17]=0x3d37030 using rc_mlx5/mlx5_ib2:1 on worker 0x2d7f8d0
[1650465529.624836] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x2293070: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 3)
[1650465529.624861] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x2293070: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 4)
[1650465529.625243] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x2293070: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 5)
[1650465529.625457] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x2293070: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 6)
[1650465529.625486] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x2293070: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 7)
[1650465529.625492] [ndv4:14949:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib1:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.625535] [ndv4:14949:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.625541] [ndv4:14949:0] async.c:228 UCX DEBUG added async handler 0x2106c50 [id=96 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.625571] [ndv4:14949:0] async.c:506 UCX DEBUG listening to async event fd 96 events 0x5 mode thread_spinlock
[1650465529.625665] [ndv4:14949:0] ucp_worker.c:1159 UCX DEBUG created interface[15]=0x2293070 using ud_mlx5/mlx5_ib1:1 on worker 0x19ac660
[1650465529.625680] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.625703] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.626346] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.626351] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.626662] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.626670] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.626936] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.626971] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.627969] [ndv4:13443:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465529.628651] [ndv4:14421:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465529.628660] [ndv4:14421:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465529.628766] [ndv4:14421:0] ucp_context.c:1556 UCX DEBUG created ucp context 0xf18cd0 0xf18cd0 [13 mds 47 tls] features 0x1 tl bitmap 0x7fffffffffff 0x0
[1650465529.629849] [ndv4:13443:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.629867] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.629870] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.629938] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.630386] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x3752750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650465529.631506] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x3752750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650465529.631535] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x3752750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650465529.630345] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465529.630352] [ndv4:13443:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib2:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.630387] [ndv4:13443:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib2 length=2048) failed: Invalid argument
[1650465529.630391] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.630398] [ndv4:13443:0] async.c:228 UCX DEBUG added async handler 0x3d3ff60 [id=101 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.630427] [ndv4:13443:0] async.c:506 UCX DEBUG listening to async event fd 101 events 0x1 mode thread_spinlock
[1650465529.631061] [ndv4:13443:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465529.633960] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.633984] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.634416] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465529.635688] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.635704] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.635707] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.635761] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.635698] [ndv4:13552:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465529.636362] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x4990010 of 8176 bytes with 127 elements
[1650465529.636274] [ndv4:14419:0] async.c:228 UCX DEBUG added async handler 0x8d61e0 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650465529.636390] [ndv4:14419:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650465529.636399] [ndv4:14419:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650465529.636469] [ndv4:14949:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465529.636610] [ndv4:14419:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.636620] [ndv4:14419:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.636642] [ndv4:14419:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465529.636702] [ndv4:14419:0] module.c:253 UCX DEBUG loading modules for ucm
[1650465529.636723] [ndv4:14419:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650465529.636523] [ndv4:13552:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.637009] [ndv4:13552:0] ib_iface.c:994 UCX DEBUG iface=0x1b57a30: created UD QP 0xde93 on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.636678] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.636685] [ndv4:14205:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib4:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.636724] [ndv4:14205:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib4 length=2048) failed: Invalid argument
[1650465529.636728] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.636740] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x4677f60 [id=113 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.636769] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 113 events 0x1 mode thread_spinlock
[1650465529.636781] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[27]=0x47bc030 using rc_mlx5/mlx5_ib4:1 on worker 0x29ca770
[1650465529.637035] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.637053] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.637221] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.637232] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.637744] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.637796] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.637803] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.638355] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.638393] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.638274] [ndv4:14949:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465529.638292] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.638309] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.638361] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.637879] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465529.638710] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.638717] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465529.639343] [ndv4:14949:0] ib_iface.c:994 UCX DEBUG iface=0x28170a0: created RC QP 0xc4c1 on mlx5_ib2:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465529.638785] [ndv4:13552:0] ib_md.c:812 UCX DEBUG registered memory 0x2b265bf74000..0x2b265bff9000 on mlx5_ib0 lkey 0x81d00 rkey 0x81d00 access 0xf flags 0x3e4
[1650465529.638792] [ndv4:13552:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b265bf74018 of 544744 bytes with 128 elements
[1650465529.638797] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.639410] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x1b57a30: adding gid fe80::15:5dff:fd33:fffb to hash on device mlx5_ib0 port 1 index 0)
[1650465529.639478] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.639495] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.639498] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.639556] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.639880] [ndv4:14949:0] ucp_worker.c:1159 UCX DEBUG created interface[16]=0x28170a0 using rc_verbs/mlx5_ib2:1 on worker 0x19ac660
[1650465529.639927] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.639963] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.640064] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.640077] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.640029] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465529.640036] [ndv4:14205:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib4:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.640069] [ndv4:14205:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib4 length=2048) failed: Invalid argument
[1650465529.640072] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.640080] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x3762580 [id=115 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.640107] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 115 events 0x1 mode thread_spinlock
[1650465529.639570] [ndv4:13443:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x3f0d050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xc4cb
[1650465529.639884] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.639955] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.639965] [ndv4:13443:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b7507f19008 of 151544 bytes with 1052 elements
[1650465529.640888] [ndv4:14949:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465529.640828] [ndv4:14205:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465529.641936] [ndv4:14949:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.641950] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.641953] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.642005] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.642387] [ndv4:14949:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x2b38010 of 8176 bytes with 127 elements
[1650465529.642690] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.642697] [ndv4:14949:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib2:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.642736] [ndv4:14949:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib2 length=2048) failed: Invalid argument
[1650465529.642739] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.642749] [ndv4:14949:0] async.c:228 UCX DEBUG added async handler 0x10d4e50 [id=99 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.642786] [ndv4:14949:0] async.c:506 UCX DEBUG listening to async event fd 99 events 0x1 mode thread_spinlock
[1650465529.642812] [ndv4:14949:0] ucp_worker.c:1159 UCX DEBUG created interface[17]=0x2964030 using rc_mlx5/mlx5_ib2:1 on worker 0x19ac660
[1650465529.642938] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.642945] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.643112] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.643129] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.642501] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x3752750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650465529.642512] [ndv4:12741:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib0:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.642621] [ndv4:12741:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.642626] [ndv4:12741:0] async.c:228 UCX DEBUG added async handler 0x2dc76c0 [id=89 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.642659] [ndv4:12741:0] async.c:506 UCX DEBUG listening to async event fd 89 events 0x5 mode thread_spinlock
[1650465529.642676] [ndv4:12741:0] ucp_worker.c:1159 UCX DEBUG created interface[10]=0x3752750 using ud_mlx5/mlx5_ib0:1 on worker 0x37328d0
[1650465529.642905] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.642926] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.643103] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.643129] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.643483] [ndv4:12741:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465529.643897] [ndv4:14949:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465529.643886] [ndv4:13443:0] ib_md.c:812 UCX DEBUG registered memory 0x2b7511600000..0x2b7513c00000 on mlx5_ib2 lkey 0x81300 rkey 0x81300 access 0xf flags 0x3e4
[1650465529.643911] [ndv4:13443:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b7511600018 of 39845864 bytes with 4752 elements
[1650465529.644051] [ndv4:13443:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x3f0d050
[1650465529.644087] [ndv4:13443:0] ucp_worker.c:1159 UCX DEBUG created interface[18]=0x3f0d050 using dc_mlx5/mlx5_ib2:1 on worker 0x2d7f8d0
[1650465529.644269] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.644292] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.644523] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.644552] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.644989] [ndv4:12741:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465529.644998] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.645001] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.645019] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.645483] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.645523] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465529.646339] [ndv4:12741:0] ib_iface.c:994 UCX DEBUG iface=0x4003050: created RC QP 0xc474 on mlx5_ib1:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465529.647845] [ndv4:14419:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465529.647857] [ndv4:14419:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465529.647116] [ndv4:12741:0] ucp_worker.c:1159 UCX DEBUG created interface[11]=0x4003050 using rc_verbs/mlx5_ib1:1 on worker 0x37328d0
[1650465529.647536] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.647570] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.647865] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.647871] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.648063] [ndv4:12741:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465529.649392] [ndv4:12741:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.649401] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.649404] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.649422] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.649863] [ndv4:12741:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x4016010 of 8176 bytes with 127 elements
[1650465529.650089] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.650095] [ndv4:12741:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib1:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.650133] [ndv4:12741:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650465529.650136] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.650145] [ndv4:12741:0] async.c:228 UCX DEBUG added async handler 0x2e5af70 [id=92 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.650167] [ndv4:12741:0] async.c:506 UCX DEBUG listening to async event fd 92 events 0x1 mode thread_spinlock
[1650465529.650176] [ndv4:12741:0] ucp_worker.c:1159 UCX DEBUG created interface[12]=0x3e84010 using rc_mlx5/mlx5_ib1:1 on worker 0x37328d0
[1650465529.650297] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.650311] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.650443] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.650509] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.649357] [ndv4:14419:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465529.649428] [ndv4:14419:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.649445] [ndv4:14419:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.649322] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x1b57a30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650465529.649895] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x1b57a30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650465529.650229] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x1b57a30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650465529.650489] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x1b57a30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650465529.651075] [ndv4:12741:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465529.651249] [ndv4:14419:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650465529.651257] [ndv4:14419:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465529.650501] [ndv4:14205:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x4992050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xc4b9
[1650465529.650745] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.650833] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.650842] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2afdb6f5a008 of 151544 bytes with 1052 elements
[1650465529.652681] [ndv4:12741:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.652713] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.652717] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.652778] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.653215] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465529.653230] [ndv4:12741:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib1:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.653305] [ndv4:12741:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650465529.653310] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.653329] [ndv4:12741:0] async.c:228 UCX DEBUG added async handler 0x2e6cdd0 [id=94 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.653369] [ndv4:12741:0] async.c:506 UCX DEBUG listening to async event fd 94 events 0x1 mode thread_spinlock
[1650465529.653935] [ndv4:12741:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465529.655249] [ndv4:13443:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465529.654912] [ndv4:14205:0] ib_md.c:812 UCX DEBUG registered memory 0x2afdb4800000..0x2afdb6e00000 on mlx5_ib4 lkey 0x80600 rkey 0x80600 access 0xf flags 0x3e4
[1650465529.654925] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2afdb4800018 of 39845864 bytes with 4752 elements
[1650465529.655064] [ndv4:14205:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x4992050
[1650465529.655097] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[28]=0x4992050 using dc_mlx5/mlx5_ib4:1 on worker 0x29ca770
[1650465529.655263] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.655273] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.655698] [ndv4:14949:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.655718] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.655721] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.655778] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.656186] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465529.656193] [ndv4:14949:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib2:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.656231] [ndv4:14949:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib2 length=2048) failed: Invalid argument
[1650465529.656236] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.656245] [ndv4:14949:0] async.c:228 UCX DEBUG added async handler 0x296ce80 [id=101 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.656272] [ndv4:14949:0] async.c:506 UCX DEBUG listening to async event fd 101 events 0x1 mode thread_spinlock
[1650465529.656378] [ndv4:13443:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.656728] [ndv4:14949:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465529.656826] [ndv4:13443:0] ib_iface.c:994 UCX DEBUG iface=0x40e9060: created UD QP 0xc4c2 on mlx5_ib2:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.657486] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.657593] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.657600] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.657638] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.657643] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.658054] [ndv4:13443:0] ib_md.c:812 UCX DEBUG registered memory 0x2b7507f3e000..0x2b7507fc3000 on mlx5_ib2 lkey 0x81400 rkey 0x81400 access 0xf flags 0x3e4
[1650465529.658060] [ndv4:13443:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b7507f3e018 of 544744 bytes with 128 elements
[1650465529.658065] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.658307] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x40e9060: adding gid fe80::15:5dff:fd33:fffd to hash on device mlx5_ib2 port 1 index 0)
[1650465529.658715] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x40e9060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 1)
[1650465529.658907] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x40e9060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 2)
[1650465529.658989] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x40e9060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 3)
[1650465529.659120] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x40e9060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 4)
[1650465529.659227] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x40e9060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 5)
[1650465529.659252] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x40e9060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 6)
[1650465529.659276] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x40e9060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 7)
[1650465529.659658] [ndv4:13443:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.659669] [ndv4:13443:0] async.c:228 UCX DEBUG added async handler 0x24b9630 [id=102 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.659706] [ndv4:13443:0] async.c:506 UCX DEBUG listening to async event fd 102 events 0x5 mode thread_spinlock
[1650465529.659718] [ndv4:13443:0] ucp_worker.c:1159 UCX DEBUG created interface[19]=0x40e9060 using ud_verbs/mlx5_ib2:1 on worker 0x2d7f8d0
[1650465529.659729] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.659734] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.659832] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.659837] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.660686] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x1b57a30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650465529.660962] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x1b57a30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650465529.661278] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x1b57a30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650465529.661536] [ndv4:13552:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.661629] [ndv4:14419:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465529.661638] [ndv4:14419:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465529.664303] [ndv4:12741:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x41ae010: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xc49d
[1650465529.664743] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.664751] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.664762] [ndv4:12741:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b7353dbc008 of 151544 bytes with 1052 elements
[1650465529.664191] [ndv4:14949:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x2b3a050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xc4d7
[1650465529.664647] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.664656] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.664667] [ndv4:14949:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2ad4b7fc6008 of 151544 bytes with 1052 elements
[1650465529.664891] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.664903] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.665412] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465529.666503] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.666935] [ndv4:14205:0] ib_iface.c:994 UCX DEBUG iface=0x4b6e250: created UD QP 0xc4af on mlx5_ib4:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.667541] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.667859] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.667875] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.667904] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.667920] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.667467] [ndv4:13552:0] async.c:228 UCX DEBUG added async handler 0x1a972f0 [id=88 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.667506] [ndv4:13552:0] async.c:506 UCX DEBUG listening to async event fd 88 events 0x5 mode thread_spinlock
[1650465529.667532] [ndv4:13552:0] ucp_worker.c:1159 UCX DEBUG created interface[9]=0x1b57a30 using ud_verbs/mlx5_ib0:1 on worker 0x24108d0
[1650465529.668358] [ndv4:14949:0] ib_md.c:812 UCX DEBUG registered memory 0x2ad4c3400000..0x2ad4c5a00000 on mlx5_ib2 lkey 0x81500 rkey 0x81500 access 0xf flags 0x3e4
[1650465529.668378] [ndv4:14949:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2ad4c3400018 of 39845864 bytes with 4752 elements
[1650465529.668521] [ndv4:14949:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x2b3a050
[1650465529.668558] [ndv4:14949:0] ucp_worker.c:1159 UCX DEBUG created interface[18]=0x2b3a050 using dc_mlx5/mlx5_ib2:1 on worker 0x19ac660
[1650465529.668766] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.668777] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.668907] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.668914] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.669493] [ndv4:14949:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465529.668389] [ndv4:14205:0] ib_md.c:812 UCX DEBUG registered memory 0x2afdb6f7f000..0x2afdb7004000 on mlx5_ib4 lkey 0x80700 rkey 0x80700 access 0xf flags 0x3e4
[1650465529.668397] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2afdb6f7f018 of 544744 bytes with 128 elements
[1650465529.668402] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.668944] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x4b6e250: adding gid fe80::15:5dff:fd33:ffff to hash on device mlx5_ib4 port 1 index 0)
[1650465529.668971] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x4b6e250: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 1)
[1650465529.669191] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x4b6e250: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 2)
[1650465529.669246] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x4b6e250: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 3)
[1650465529.669319] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x4b6e250: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 4)
[1650465529.669731] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x4b6e250: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 5)
[1650465529.670266] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x4b6e250: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 6)
[1650465529.669328] [ndv4:12741:0] ib_md.c:812 UCX DEBUG registered memory 0x2b735b200000..0x2b735d800000 on mlx5_ib1 lkey 0x81500 rkey 0x81500 access 0xf flags 0x3e4
[1650465529.669343] [ndv4:12741:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b735b200018 of 39845864 bytes with 4752 elements
[1650465529.669483] [ndv4:12741:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x41ae010
[1650465529.669514] [ndv4:12741:0] ucp_worker.c:1159 UCX DEBUG created interface[13]=0x41ae010 using dc_mlx5/mlx5_ib1:1 on worker 0x37328d0
[1650465529.669540] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.669547] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.670343] [ndv4:13443:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465529.670870] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x4b6e250: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 7)
[1650465529.671502] [ndv4:13443:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.671832] [ndv4:13443:0] ib_iface.c:994 UCX DEBUG iface=0x24269c0: created UD QP 0xc4ce on mlx5_ib2:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.671839] [ndv4:13443:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465529.671200] [ndv4:14205:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.671208] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x3e7f3f0 [id=116 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.671236] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 116 events 0x5 mode thread_spinlock
[1650465529.671256] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[29]=0x4b6e250 using ud_verbs/mlx5_ib4:1 on worker 0x29ca770
[1650465529.671340] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.671347] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.671429] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.671434] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.671966] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465529.672548] [ndv4:14419:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465529.672666] [ndv4:14419:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465529.672366] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.672622] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.672629] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.672754] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.672814] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.673180] [ndv4:13443:0] ib_md.c:812 UCX DEBUG registered memory 0x2b7513c1c000..0x2b7513ca1000 on mlx5_ib2 lkey 0x81600 rkey 0x81600 access 0xf flags 0x3e4
[1650465529.673192] [ndv4:13443:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b7513c1c018 of 544744 bytes with 128 elements
[1650465529.673198] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.673434] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x24269c0: adding gid fe80::15:5dff:fd33:fffd to hash on device mlx5_ib2 port 1 index 0)
[1650465529.673616] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x24269c0: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 1)
[1650465529.673761] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x24269c0: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 2)
[1650465529.677598] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.677621] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.677683] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.677694] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.679026] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.679036] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.679988] [ndv4:12741:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465529.679895] [ndv4:14949:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.680236] [ndv4:14949:0] ib_iface.c:994 UCX DEBUG iface=0x2d16060: created UD QP 0xc4cf on mlx5_ib2:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.680817] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.681442] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.681450] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.681471] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.681475] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.681678] [ndv4:12741:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.682500] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.682997] [ndv4:14205:0] ib_iface.c:994 UCX DEBUG iface=0x4c8c050: created UD QP 0xc4b0 on mlx5_ib4:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.683004] [ndv4:14205:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465529.682020] [ndv4:12741:0] ib_iface.c:994 UCX DEBUG iface=0x40ad3d0: created UD QP 0xc47d on mlx5_ib1:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.681931] [ndv4:14949:0] ib_md.c:812 UCX DEBUG registered memory 0x2ad4c5a05000..0x2ad4c5a8a000 on mlx5_ib2 lkey 0x81700 rkey 0x81700 access 0xf flags 0x3e4
[1650465529.681938] [ndv4:14949:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ad4c5a05018 of 544744 bytes with 128 elements
[1650465529.681942] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.682782] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x2d16060: adding gid fe80::15:5dff:fd33:fffd to hash on device mlx5_ib2 port 1 index 0)
[1650465529.683820] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.683956] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x24269c0: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 3)
[1650465529.684737] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x24269c0: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 4)
[1650465529.683534] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.683808] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.683827] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.684088] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.684093] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.684510] [ndv4:12741:0] ib_md.c:812 UCX DEBUG registered memory 0x2b7353de1000..0x2b7353e66000 on mlx5_ib1 lkey 0x81600 rkey 0x81600 access 0xf flags 0x3e4
[1650465529.684516] [ndv4:12741:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b7353de1018 of 544744 bytes with 128 elements
[1650465529.684520] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.684834] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x40ad3d0: adding gid fe80::15:5dff:fd33:fffc to hash on device mlx5_ib1 port 1 index 0)
[1650465529.685888] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x40ad3d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 1)
[1650465529.685745] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x24269c0: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 5)
[1650465529.685776] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x24269c0: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 6)
[1650465529.687226] [ndv4:13552:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465529.688613] [ndv4:13552:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.689076] [ndv4:13552:0] ib_iface.c:994 UCX DEBUG iface=0x2430750: created UD QP 0xde94 on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.689091] [ndv4:13552:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465529.689826] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.689965] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.689973] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.690059] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.690068] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.690550] [ndv4:13552:0] ib_md.c:812 UCX DEBUG registered memory 0x2b2663efa000..0x2b2663f7f000 on mlx5_ib0 lkey 0x81e00 rkey 0x81e00 access 0xf flags 0x3e4
[1650465529.690557] [ndv4:13552:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b2663efa018 of 544744 bytes with 128 elements
[1650465529.690561] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.690828] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x2430750: adding gid fe80::15:5dff:fd33:fffb to hash on device mlx5_ib0 port 1 index 0)
[1650465529.691012] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x2430750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650465529.691131] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x2430750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650465529.691417] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x2430750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650465529.692231] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.692242] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.692344] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.692350] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.692111] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x2d16060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 1)
[1650465529.692743] [ndv4:14205:0] ib_md.c:812 UCX DEBUG registered memory 0x2afdb7004000..0x2afdb7089000 on mlx5_ib4 lkey 0x80e00 rkey 0x80e00 access 0xf flags 0x3e4
[1650465529.692752] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2afdb7004018 of 544744 bytes with 128 elements
[1650465529.692757] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.693168] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x4c8c050: adding gid fe80::15:5dff:fd33:ffff to hash on device mlx5_ib4 port 1 index 0)
[1650465529.693376] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x4c8c050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 1)
[1650465529.693433] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x4c8c050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 2)
[1650465529.693518] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x4c8c050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 3)
[1650465529.693757] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x4c8c050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 4)
[1650465529.693825] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x4c8c050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 5)
[1650465529.693984] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x4c8c050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 6)
[1650465529.694164] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x4c8c050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 7)
[1650465529.694171] [ndv4:14205:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib4:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.694205] [ndv4:14205:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.694210] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x464aa10 [id=117 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.694235] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 117 events 0x5 mode thread_spinlock
[1650465529.694248] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[30]=0x4c8c050 using ud_mlx5/mlx5_ib4:1 on worker 0x29ca770
[1650465529.694312] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.694318] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.694396] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.694415] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.694807] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x24269c0: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 7)
[1650465529.694819] [ndv4:13443:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib2:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.694852] [ndv4:13443:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.694858] [ndv4:13443:0] async.c:228 UCX DEBUG added async handler 0x36666b0 [id=103 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.694879] [ndv4:13443:0] async.c:506 UCX DEBUG listening to async event fd 103 events 0x5 mode thread_spinlock
[1650465529.694890] [ndv4:13443:0] ucp_worker.c:1159 UCX DEBUG created interface[20]=0x24269c0 using ud_mlx5/mlx5_ib2:1 on worker 0x2d7f8d0
[1650465529.694907] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.694914] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.695058] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.695068] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.694864] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib5:1
[1650465529.694737] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x40ad3d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 2)
[1650465529.695412] [ndv4:13443:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465529.696087] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465529.696104] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.696108] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.696159] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.696551] [ndv4:13443:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465529.696568] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.696571] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.696643] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.696864] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.696870] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465529.697175] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.697183] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465529.697794] [ndv4:14205:0] ib_iface.c:994 UCX DEBUG iface=0x4d8c0a0: created RC QP 0xc443 on mlx5_ib5:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465529.698112] [ndv4:13443:0] ib_iface.c:994 UCX DEBUG iface=0x43070a0: created RC QP 0xc446 on mlx5_ib3:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465529.699010] [ndv4:13443:0] ucp_worker.c:1159 UCX DEBUG created interface[21]=0x43070a0 using rc_verbs/mlx5_ib3:1 on worker 0x2d7f8d0
[1650465529.699207] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.699215] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.699284] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.699288] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.698610] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[31]=0x4d8c0a0 using rc_verbs/mlx5_ib5:1 on worker 0x29ca770
[1650465529.698858] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.698866] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.699042] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.699047] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.700142] [ndv4:13443:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465529.700188] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib5:1
[1650465529.700792] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x2430750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650465529.701175] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x2430750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650465529.702115] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x2d16060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 2)
[1650465529.702071] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.702085] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.702088] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.702137] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.702697] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x50ad010 of 8176 bytes with 127 elements
[1650465529.701906] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x2430750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650465529.702356] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x2430750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650465529.702364] [ndv4:13552:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib0:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.702399] [ndv4:13552:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.702404] [ndv4:13552:0] async.c:228 UCX DEBUG added async handler 0x1aa56c0 [id=89 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.702430] [ndv4:13552:0] async.c:506 UCX DEBUG listening to async event fd 89 events 0x5 mode thread_spinlock
[1650465529.702445] [ndv4:13552:0] ucp_worker.c:1159 UCX DEBUG created interface[10]=0x2430750 using ud_mlx5/mlx5_ib0:1 on worker 0x24108d0
[1650465529.702656] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.702663] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.702981] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.702988] [ndv4:14205:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib5:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.703023] [ndv4:14205:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib5 length=2048) failed: Invalid argument
[1650465529.703026] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.703039] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x3e7f740 [id=120 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.703061] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 120 events 0x1 mode thread_spinlock
[1650465529.703072] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[32]=0x4ed9030 using rc_mlx5/mlx5_ib5:1 on worker 0x29ca770
[1650465529.703192] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.703199] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.703363] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.703368] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.704194] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x40ad3d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 3)
[1650465529.704496] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x40ad3d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 4)
[1650465529.704997] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x40ad3d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 5)
[1650465529.705356] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x40ad3d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 6)
[1650465529.705393] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x40ad3d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 7)
[1650465529.705710] [ndv4:12741:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.705718] [ndv4:12741:0] async.c:228 UCX DEBUG added async handler 0x2e612b0 [id=95 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.705748] [ndv4:12741:0] async.c:506 UCX DEBUG listening to async event fd 95 events 0x5 mode thread_spinlock
[1650465529.705760] [ndv4:12741:0] ucp_worker.c:1159 UCX DEBUG created interface[14]=0x40ad3d0 using ud_verbs/mlx5_ib1:1 on worker 0x37328d0
[1650465529.705876] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.705883] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.706018] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.706025] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.708504] [ndv4:13443:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.708519] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.708522] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.708629] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.709075] [ndv4:13443:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x4628010 of 8176 bytes with 127 elements
[1650465529.710302] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.710315] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.711642] [ndv4:13552:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465529.709988] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x2d16060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 3)
[1650465529.710604] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x2d16060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 4)
[1650465529.711313] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x2d16060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 5)
[1650465529.712063] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x2d16060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 6)
[1650465529.712729] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x2d16060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 7)
[1650465529.709360] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.709367] [ndv4:13443:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib3:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.709403] [ndv4:13443:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib3 length=2048) failed: Invalid argument
[1650465529.709407] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.709418] [ndv4:13443:0] async.c:228 UCX DEBUG added async handler 0x4207ae0 [id=106 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.709446] [ndv4:13443:0] async.c:506 UCX DEBUG listening to async event fd 106 events 0x1 mode thread_spinlock
[1650465529.709456] [ndv4:13443:0] ucp_worker.c:1159 UCX DEBUG created interface[22]=0x4454030 using rc_mlx5/mlx5_ib3:1 on worker 0x2d7f8d0
[1650465529.709530] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.709537] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.709696] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.709702] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.710496] [ndv4:13443:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465529.711468] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib5:1
[1650465529.712938] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.712954] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.712957] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.713015] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.712838] [ndv4:13552:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465529.712848] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.712851] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.712868] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.712988] [ndv4:14949:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.712997] [ndv4:14949:0] async.c:228 UCX DEBUG added async handler 0x1ea4f00 [id=102 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.713032] [ndv4:14949:0] async.c:506 UCX DEBUG listening to async event fd 102 events 0x5 mode thread_spinlock
[1650465529.713043] [ndv4:14949:0] ucp_worker.c:1159 UCX DEBUG created interface[19]=0x2d16060 using ud_verbs/mlx5_ib2:1 on worker 0x19ac660
[1650465529.713143] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.713149] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.713254] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.713259] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.713209] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.713215] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465529.713458] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465529.713467] [ndv4:14205:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib5:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.713501] [ndv4:14205:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib5 length=2048) failed: Invalid argument
[1650465529.713505] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.713514] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x4ee1fc0 [id=122 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.713542] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 122 events 0x1 mode thread_spinlock
[1650465529.713863] [ndv4:13552:0] ib_iface.c:994 UCX DEBUG iface=0x2ce1050: created RC QP 0xc47f on mlx5_ib1:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465529.714298] [ndv4:14205:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465529.714653] [ndv4:13552:0] ucp_worker.c:1159 UCX DEBUG created interface[11]=0x2ce1050 using rc_verbs/mlx5_ib1:1 on worker 0x24108d0
[1650465529.714696] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.714702] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.714839] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.714844] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.715153] [ndv4:13552:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465529.716539] [ndv4:13552:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.716549] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.716552] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.716570] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.717064] [ndv4:13552:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x2cf4010 of 8176 bytes with 127 elements
[1650465529.717277] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.717282] [ndv4:13552:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib1:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.717321] [ndv4:13552:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650465529.717325] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.717336] [ndv4:13552:0] async.c:228 UCX DEBUG added async handler 0x1b38f70 [id=92 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.717364] [ndv4:13552:0] async.c:506 UCX DEBUG listening to async event fd 92 events 0x1 mode thread_spinlock
[1650465529.717374] [ndv4:13552:0] ucp_worker.c:1159 UCX DEBUG created interface[12]=0x2b62010 using rc_mlx5/mlx5_ib1:1 on worker 0x24108d0
[1650465529.717415] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.717422] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.717527] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.717533] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.719325] [ndv4:14419:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465529.719337] [ndv4:14419:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465529.720835] [ndv4:14419:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465529.720842] [ndv4:14419:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465529.721162] [ndv4:14419:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465529.722361] [ndv4:13443:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.722381] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.722384] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.722440] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.722760] [ndv4:14205:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x50af050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xc456
[1650465529.723010] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.723017] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.723026] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2afdb988c008 of 151544 bytes with 1052 elements
[1650465529.722900] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465529.722908] [ndv4:13443:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib3:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.722940] [ndv4:13443:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib3 length=2048) failed: Invalid argument
[1650465529.722944] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.722951] [ndv4:13443:0] async.c:228 UCX DEBUG added async handler 0x24a3f40 [id=108 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.722974] [ndv4:13443:0] async.c:506 UCX DEBUG listening to async event fd 108 events 0x1 mode thread_spinlock
[1650465529.723586] [ndv4:13443:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465529.724718] [ndv4:14949:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465529.725714] [ndv4:14949:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.726005] [ndv4:14949:0] ib_iface.c:994 UCX DEBUG iface=0x2e34050: created UD QP 0xc4d2 on mlx5_ib2:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.726013] [ndv4:14949:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465529.726322] [ndv4:12741:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465529.726918] [ndv4:14205:0] ib_md.c:812 UCX DEBUG registered memory 0x2afdb7200000..0x2afdb9800000 on mlx5_ib5 lkey 0x80e00 rkey 0x80e00 access 0xf flags 0x3e4
[1650465529.726935] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2afdb7200018 of 39845864 bytes with 4752 elements
[1650465529.727121] [ndv4:14205:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x50af050
[1650465529.727155] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[33]=0x50af050 using dc_mlx5/mlx5_ib5:1 on worker 0x29ca770
[1650465529.726691] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.726782] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.726788] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.726815] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.726820] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.727269] [ndv4:14949:0] ib_md.c:812 UCX DEBUG registered memory 0x2ad4c5a8a000..0x2ad4c5b0f000 on mlx5_ib2 lkey 0x81800 rkey 0x81800 access 0xf flags 0x3e4
[1650465529.727276] [ndv4:14949:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ad4c5a8a018 of 544744 bytes with 128 elements
[1650465529.727280] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.727552] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x2e34050: adding gid fe80::15:5dff:fd33:fffd to hash on device mlx5_ib2 port 1 index 0)
[1650465529.727304] [ndv4:12741:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.727807] [ndv4:12741:0] ib_iface.c:994 UCX DEBUG iface=0x3f74280: created UD QP 0xc483 on mlx5_ib1:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.727816] [ndv4:12741:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465529.728684] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.728834] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.728844] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.728932] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.728937] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.729436] [ndv4:12741:0] ib_md.c:812 UCX DEBUG registered memory 0x2b7353e66000..0x2b7353eeb000 on mlx5_ib1 lkey 0x81700 rkey 0x81700 access 0xf flags 0x3e4
[1650465529.729443] [ndv4:12741:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b7353e66018 of 544744 bytes with 128 elements
[1650465529.729447] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.729803] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x3f74280: adding gid fe80::15:5dff:fd33:fffc to hash on device mlx5_ib1 port 1 index 0)
[1650465529.729892] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x3f74280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 1)
[1650465529.730106] [ndv4:13552:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465529.731425] [ndv4:13443:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x462a050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xc45c
[1650465529.731490] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.731497] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.731506] [ndv4:13443:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b7507fc5008 of 151544 bytes with 1052 elements
[1650465529.731392] [ndv4:13552:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.731412] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.731415] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.731472] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.731998] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465529.732005] [ndv4:13552:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib1:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.732038] [ndv4:13552:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650465529.732042] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.732050] [ndv4:13552:0] async.c:228 UCX DEBUG added async handler 0x1b4add0 [id=94 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.732074] [ndv4:13552:0] async.c:506 UCX DEBUG listening to async event fd 94 events 0x1 mode thread_spinlock
[1650465529.732518] [ndv4:13552:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465529.735331] [ndv4:13443:0] ib_md.c:812 UCX DEBUG registered memory 0x2b7513e00000..0x2b7516400000 on mlx5_ib3 lkey 0x80d00 rkey 0x80d00 access 0xf flags 0x3e4
[1650465529.735347] [ndv4:13443:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b7513e00018 of 39845864 bytes with 4752 elements
[1650465529.735487] [ndv4:13443:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x462a050
[1650465529.735522] [ndv4:13443:0] ucp_worker.c:1159 UCX DEBUG created interface[23]=0x462a050 using dc_mlx5/mlx5_ib3:1 on worker 0x2d7f8d0
[1650465529.735817] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.735828] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.735895] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.735900] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.736289] [ndv4:13443:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465529.737734] [ndv4:13443:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.738173] [ndv4:13443:0] ib_iface.c:994 UCX DEBUG iface=0x4806060: created UD QP 0xc452 on mlx5_ib3:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.738815] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.738958] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.738966] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.738998] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.739002] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.739934] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.739966] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.740076] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.740081] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.740487] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib5:1
[1650465529.739494] [ndv4:13552:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x2e8c010: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xc4ab
[1650465529.739612] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.739620] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.739629] [ndv4:13552:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b2666780008 of 151544 bytes with 1052 elements
[1650465529.739377] [ndv4:13443:0] ib_md.c:812 UCX DEBUG registered memory 0x2b75164a2000..0x2b7516527000 on mlx5_ib3 lkey 0x80e00 rkey 0x80e00 access 0xf flags 0x3e4
[1650465529.739384] [ndv4:13443:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b75164a2018 of 544744 bytes with 128 elements
[1650465529.739388] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.739548] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x4806060: adding gid fe80::15:5dff:fd33:fffe to hash on device mlx5_ib3 port 1 index 0)
[1650465529.739927] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x4806060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 1)
[1650465529.740146] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x4806060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 2)
[1650465529.740324] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x4806060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 3)
[1650465529.740470] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x4806060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 4)
[1650465529.740650] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x4806060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 5)
[1650465529.740710] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x4806060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 6)
[1650465529.740747] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x4806060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 7)
[1650465529.740958] [ndv4:13443:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.740968] [ndv4:13443:0] async.c:228 UCX DEBUG added async handler 0x3bc5e10 [id=109 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.740993] [ndv4:13443:0] async.c:506 UCX DEBUG listening to async event fd 109 events 0x5 mode thread_spinlock
[1650465529.741005] [ndv4:13443:0] ucp_worker.c:1159 UCX DEBUG created interface[24]=0x4806060 using ud_verbs/mlx5_ib3:1 on worker 0x2d7f8d0
[1650465529.741024] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.741031] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.741081] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.741085] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.741243] [ndv4:13443:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465529.742316] [ndv4:13443:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.742711] [ndv4:13443:0] ib_iface.c:994 UCX DEBUG iface=0x4924460: created UD QP 0xc453 on mlx5_ib3:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.742719] [ndv4:13443:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465529.742814] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x3f74280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 2)
[1650465529.743060] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x3f74280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 3)
[1650465529.743251] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x3f74280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 4)
[1650465529.743487] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x3f74280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 5)
[1650465529.743548] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x3f74280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 6)
[1650465529.743684] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x3f74280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 7)
[1650465529.743693] [ndv4:12741:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib1:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.743734] [ndv4:12741:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.743742] [ndv4:12741:0] async.c:228 UCX DEBUG added async handler 0x3e8cc30 [id=96 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.743772] [ndv4:12741:0] async.c:506 UCX DEBUG listening to async event fd 96 events 0x5 mode thread_spinlock
[1650465529.743793] [ndv4:12741:0] ucp_worker.c:1159 UCX DEBUG created interface[15]=0x3f74280 using ud_mlx5/mlx5_ib1:1 on worker 0x37328d0
[1650465529.743873] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.743880] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.743940] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.743944] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.744302] [ndv4:12741:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465529.745319] [ndv4:12741:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465529.745335] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.745338] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.745390] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.743195] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.743325] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.743331] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.743342] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.743346] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.743709] [ndv4:13443:0] ib_md.c:812 UCX DEBUG registered memory 0x2b7516527000..0x2b75165ac000 on mlx5_ib3 lkey 0x80f00 rkey 0x80f00 access 0xf flags 0x3e4
[1650465529.743714] [ndv4:13443:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b7516527018 of 544744 bytes with 128 elements
[1650465529.743718] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.744200] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x4924460: adding gid fe80::15:5dff:fd33:fffe to hash on device mlx5_ib3 port 1 index 0)
[1650465529.744557] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x4924460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 1)
[1650465529.744597] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x4924460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 2)
[1650465529.744611] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x4924460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 3)
[1650465529.744624] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x4924460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 4)
[1650465529.744637] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x4924460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 5)
[1650465529.744651] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x4924460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 6)
[1650465529.744868] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x4924460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 7)
[1650465529.744874] [ndv4:13443:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib3:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.744904] [ndv4:13443:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.744909] [ndv4:13443:0] async.c:228 UCX DEBUG added async handler 0x24b8f70 [id=110 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.744936] [ndv4:13443:0] async.c:506 UCX DEBUG listening to async event fd 110 events 0x5 mode thread_spinlock
[1650465529.744946] [ndv4:13443:0] ucp_worker.c:1159 UCX DEBUG created interface[25]=0x4924460 using ud_mlx5/mlx5_ib3:1 on worker 0x2d7f8d0
[1650465529.745003] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.745010] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.745085] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.745091] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.745794] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.745801] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465529.743283] [ndv4:13552:0] ib_md.c:812 UCX DEBUG registered memory 0x2b2664000000..0x2b2666600000 on mlx5_ib1 lkey 0x81800 rkey 0x81800 access 0xf flags 0x3e4
[1650465529.743301] [ndv4:13552:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b2664000018 of 39845864 bytes with 4752 elements
[1650465529.743443] [ndv4:13552:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x2e8c010
[1650465529.743478] [ndv4:13552:0] ucp_worker.c:1159 UCX DEBUG created interface[13]=0x2e8c010 using dc_mlx5/mlx5_ib1:1 on worker 0x24108d0
[1650465529.743551] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.743562] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.743658] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.743663] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.743963] [ndv4:13552:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465529.745180] [ndv4:13552:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.745504] [ndv4:13552:0] ib_iface.c:994 UCX DEBUG iface=0x2d8b3d0: created UD QP 0xc48e on mlx5_ib1:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.745962] [ndv4:13443:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465529.745997] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.746475] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.746483] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.746520] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.746528] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.746478] [ndv4:12741:0] ib_iface.c:994 UCX DEBUG iface=0x459d0a0: created RC QP 0xc4d3 on mlx5_ib2:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465529.746926] [ndv4:13552:0] ib_md.c:812 UCX DEBUG registered memory 0x2b26667a5000..0x2b266682a000 on mlx5_ib1 lkey 0x81900 rkey 0x81900 access 0xf flags 0x3e4
[1650465529.746932] [ndv4:13552:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b26667a5018 of 544744 bytes with 128 elements
[1650465529.746936] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.747409] [ndv4:13443:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465529.747427] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.747430] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.747482] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.747302] [ndv4:12741:0] ucp_worker.c:1159 UCX DEBUG created interface[16]=0x459d0a0 using rc_verbs/mlx5_ib2:1 on worker 0x37328d0
[1650465529.747483] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.747490] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.747959] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.747966] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465529.748819] [ndv4:14421:0] mm_posix.c:518 UCX DEBUG allocated posix shared memory at 0x2ab787e7e000 length 12288
[1650465529.748907] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool mm_recv_desc: align 64, maxelems 4294967295, elemsize 8368
[1650465529.748761] [ndv4:13443:0] ib_iface.c:994 UCX DEBUG iface=0x4a240a0: created RC QP 0xc4b1 on mlx5_ib4:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465529.749689] [ndv4:13443:0] ucp_worker.c:1159 UCX DEBUG created interface[26]=0x4a240a0 using rc_verbs/mlx5_ib4:1 on worker 0x2d7f8d0
[1650465529.749795] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.749802] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.749961] [ndv4:14421:0] mm_posix.c:305 UCX DEBUG shared memory mmap(addr=(nil), length=6291456, flags= HUGETLB, fd=76) failed: Invalid argument
[1650465529.749971] [ndv4:14421:0] mm_posix.c:518 UCX DEBUG allocated posix shared memory at 0x2ab78e454000 length 4296704
[1650465529.749976] [ndv4:14421:0] mpool.c:205 UCX DEBUG mpool mm_recv_desc: allocated chunk 0x2ab78e454018 of 4296680 bytes with 512 elements
[1650465529.750229] [ndv4:14421:0] mm_iface.c:600 UCX DEBUG created mm iface 0x16ac390 FIFO id 0x400000002dc1aa42 va 0x2ab787e7e000 size 12288 (128 x 64 elems)
[1650465529.750282] [ndv4:14421:0] ucp_worker.c:1159 UCX DEBUG created interface[0]=0x16ac390 using posix/memory on worker 0x1f71f60
[1650465529.750308] [ndv4:14421:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 8447 bytes with hugetlb
[1650465529.750347] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool mm_recv_desc: align 64, maxelems 4294967295, elemsize 8368
[1650465529.750365] [ndv4:14421:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 4292720 bytes with hugetlb
[1650465529.750374] [ndv4:14421:0] mpool.c:205 UCX DEBUG mpool mm_recv_desc: allocated chunk 0x2ab78e86d018 of 4296680 bytes with 512 elements
[1650465529.750974] [ndv4:14421:0] mm_iface.c:600 UCX DEBUG created mm iface 0x1698370 FIFO id 0x650006 va 0x2ab787e81000 size 12288 (128 x 64 elems)
[1650465529.750982] [ndv4:14421:0] ucp_worker.c:1159 UCX DEBUG created interface[1]=0x1698370 using sysv/memory on worker 0x1f71f60
[1650465529.750994] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool self_msg_desc: align 64, maxelems 4294967295, elemsize 8200
[1650465529.750998] [ndv4:14421:0] self.c:220 UCX DEBUG created self iface id 0x28eaabbf9b96ab84 send_size 8192
[1650465529.751004] [ndv4:14421:0] ucp_worker.c:1159 UCX DEBUG created interface[2]=0x1698940 using self/memory0 on worker 0x1f71f60
[1650465529.751027] [ndv4:14421:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465529.751032] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465529.751035] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465529.751974] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.752415] [ndv4:14205:0] ib_iface.c:994 UCX DEBUG iface=0x528b060: created UD QP 0xc44c on mlx5_ib5:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.753037] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.753160] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.753167] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.753224] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.753229] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.753593] [ndv4:14419:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650465529.753507] [ndv4:14421:0] async.c:228 UCX DEBUG added async handler 0x169d3a0 [id=78 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465529.753543] [ndv4:14421:0] async.c:506 UCX DEBUG listening to async event fd 78 events 0x5 mode thread_spinlock
[1650465529.753562] [ndv4:14421:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x16ace80: listening for connections (fd=78) on 10.5.0.5:52989
[1650465529.753711] [ndv4:14205:0] ib_md.c:812 UCX DEBUG registered memory 0x2afdb98b1000..0x2afdb9936000 on mlx5_ib5 lkey 0x80f00 rkey 0x80f00 access 0xf flags 0x3e4
[1650465529.753719] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2afdb98b1018 of 544744 bytes with 128 elements
[1650465529.753725] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.754323] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x528b060: adding gid fe80::15:5dff:fd34:0 to hash on device mlx5_ib5 port 1 index 0)
[1650465529.754418] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x528b060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 1)
[1650465529.754439] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x528b060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 2)
[1650465529.754602] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x528b060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 3)
[1650465529.754695] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x528b060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 4)
[1650465529.754925] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x528b060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 5)
[1650465529.755039] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x528b060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 6)
[1650465529.755209] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x528b060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 7)
[1650465529.755564] [ndv4:14205:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.755571] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x4d679b0 [id=123 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.755647] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 123 events 0x5 mode thread_spinlock
[1650465529.755660] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[34]=0x528b060 using ud_verbs/mlx5_ib5:1 on worker 0x29ca770
[1650465529.755860] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.755866] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.756027] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.756032] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.754225] [ndv4:14419:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650465529.755208] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x2d8b3d0: adding gid fe80::15:5dff:fd33:fffc to hash on device mlx5_ib1 port 1 index 0)
[1650465529.755236] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x2d8b3d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 1)
[1650465529.755252] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x2d8b3d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 2)
[1650465529.755267] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x2d8b3d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 3)
[1650465529.755511] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x2d8b3d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 4)
[1650465529.755761] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x2d8b3d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 5)
[1650465529.755783] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x2d8b3d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 6)
[1650465529.755799] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x2d8b3d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 7)
[1650465529.756060] [ndv4:13552:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.756069] [ndv4:13552:0] async.c:228 UCX DEBUG added async handler 0x1b3f2b0 [id=95 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.756098] [ndv4:13552:0] async.c:506 UCX DEBUG listening to async event fd 95 events 0x5 mode thread_spinlock
[1650465529.756109] [ndv4:13552:0] ucp_worker.c:1159 UCX DEBUG created interface[14]=0x2d8b3d0 using ud_verbs/mlx5_ib1:1 on worker 0x24108d0
[1650465529.756242] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.756248] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.756405] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.756410] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.753758] [ndv4:14421:0] ucp_worker.c:1159 UCX DEBUG created interface[3]=0x16ace80 using tcp/eth0 on worker 0x1f71f60
[1650465529.753780] [ndv4:14421:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465529.753784] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465529.753786] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465529.753826] [ndv4:14421:0] async.c:228 UCX DEBUG added async handler 0x1608900 [id=80 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465529.753847] [ndv4:14421:0] async.c:506 UCX DEBUG listening to async event fd 80 events 0x5 mode thread_spinlock
[1650465529.753852] [ndv4:14421:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x16ad500: listening for connections (fd=80) on 127.0.0.1:60246
[1650465529.753868] [ndv4:14421:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465529.753873] [ndv4:14421:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465529.753930] [ndv4:14421:0] ucp_worker.c:1159 UCX DEBUG created interface[4]=0x16ad500 using tcp/lo on worker 0x1f71f60
[1650465529.753947] [ndv4:14421:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465529.753950] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465529.753953] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465529.753991] [ndv4:14421:0] async.c:228 UCX DEBUG added async handler 0x1608630 [id=82 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465529.754009] [ndv4:14421:0] async.c:506 UCX DEBUG listening to async event fd 82 events 0x5 mode thread_spinlock
[1650465529.754013] [ndv4:14421:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x1694b90: listening for connections (fd=82) on 172.16.1.242:56708
[1650465529.754283] [ndv4:14421:0] ucp_worker.c:1159 UCX DEBUG created interface[5]=0x1694b90 using tcp/ib0 on worker 0x1f71f60
[1650465529.754312] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.754320] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.754466] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.754471] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.756170] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib5:1
[1650465529.757161] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.755856] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.755866] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.756462] [ndv4:12741:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465529.757471] [ndv4:14205:0] ib_iface.c:994 UCX DEBUG iface=0x53a9460: created UD QP 0xc44d on mlx5_ib5:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.757477] [ndv4:14205:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465529.758052] [ndv4:12741:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.758068] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.758071] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.758122] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.758073] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.758237] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.758244] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.758379] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.758384] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.758647] [ndv4:12741:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x48be010 of 8176 bytes with 127 elements
[1650465529.760023] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.760033] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.760501] [ndv4:13443:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465529.758910] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.758916] [ndv4:12741:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib2:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.758952] [ndv4:12741:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib2 length=2048) failed: Invalid argument
[1650465529.758956] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.758968] [ndv4:12741:0] async.c:228 UCX DEBUG added async handler 0x2e7b5f0 [id=99 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.759002] [ndv4:12741:0] async.c:506 UCX DEBUG listening to async event fd 99 events 0x1 mode thread_spinlock
[1650465529.759029] [ndv4:12741:0] ucp_worker.c:1159 UCX DEBUG created interface[17]=0x46ea030 using rc_mlx5/mlx5_ib2:1 on worker 0x37328d0
[1650465529.759216] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.759222] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.759398] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.759403] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.759734] [ndv4:12741:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465529.758750] [ndv4:14205:0] ib_md.c:812 UCX DEBUG registered memory 0x2afdb9936000..0x2afdb99bb000 on mlx5_ib5 lkey 0x81000 rkey 0x81000 access 0xf flags 0x3e4
[1650465529.758755] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2afdb9936018 of 544744 bytes with 128 elements
[1650465529.758760] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.759331] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x53a9460: adding gid fe80::15:5dff:fd34:0 to hash on device mlx5_ib5 port 1 index 0)
[1650465529.759496] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x53a9460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 1)
[1650465529.759512] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x53a9460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 2)
[1650465529.759944] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x53a9460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 3)
[1650465529.759970] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x53a9460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 4)
[1650465529.759984] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x53a9460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 5)
[1650465529.760543] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x53a9460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 6)
[1650465529.760689] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x53a9460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 7)
[1650465529.760695] [ndv4:14205:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib5:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.760734] [ndv4:14205:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.760738] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x528bf20 [id=124 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.760761] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 124 events 0x5 mode thread_spinlock
[1650465529.760771] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[35]=0x53a9460 using ud_mlx5/mlx5_ib5:1 on worker 0x29ca770
[1650465529.760915] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.760921] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.761104] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.761109] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.761114] [ndv4:12741:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.761132] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.761135] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.761191] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.761540] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465529.761546] [ndv4:12741:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib2:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.761607] [ndv4:12741:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib2 length=2048) failed: Invalid argument
[1650465529.761611] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.761618] [ndv4:12741:0] async.c:228 UCX DEBUG added async handler 0x2e5b4c0 [id=101 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.761649] [ndv4:12741:0] async.c:506 UCX DEBUG listening to async event fd 101 events 0x1 mode thread_spinlock
[1650465529.761667] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib6:1
[1650465529.761634] [ndv4:13443:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.761649] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.761653] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.761704] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.762145] [ndv4:13443:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x4d45010 of 8176 bytes with 127 elements
[1650465529.762211] [ndv4:12741:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465529.762856] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465529.762875] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.762878] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.762927] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.762397] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x2e34050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 1)
[1650465529.762402] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.762408] [ndv4:13443:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib4:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.762443] [ndv4:13443:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib4 length=2048) failed: Invalid argument
[1650465529.762447] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.762458] [ndv4:13443:0] async.c:228 UCX DEBUG added async handler 0x4b79da0 [id=113 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.762478] [ndv4:13443:0] async.c:506 UCX DEBUG listening to async event fd 113 events 0x1 mode thread_spinlock
[1650465529.762489] [ndv4:13443:0] ucp_worker.c:1159 UCX DEBUG created interface[27]=0x4b71030 using rc_mlx5/mlx5_ib4:1 on worker 0x2d7f8d0
[1650465529.762654] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.762660] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.762756] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.762761] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.763148] [ndv4:13443:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465529.763644] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.763650] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465529.764505] [ndv4:13443:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.764523] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.764527] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.764635] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.764506] [ndv4:14205:0] ib_iface.c:994 UCX DEBUG iface=0x54a90a0: created RC QP 0xc494 on mlx5_ib6:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465529.764989] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465529.764996] [ndv4:13443:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib4:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.765030] [ndv4:13443:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib4 length=2048) failed: Invalid argument
[1650465529.765034] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.765041] [ndv4:13443:0] async.c:228 UCX DEBUG added async handler 0x4b79fc0 [id=115 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.765068] [ndv4:13443:0] async.c:506 UCX DEBUG listening to async event fd 115 events 0x1 mode thread_spinlock
[1650465529.765137] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[36]=0x54a90a0 using rc_verbs/mlx5_ib6:1 on worker 0x29ca770
[1650465529.765273] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.765280] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.765440] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.765445] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.765643] [ndv4:13443:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465529.766340] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib6:1
[1650465529.767700] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.767716] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.767719] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.767771] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.768376] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x57ca010 of 8176 bytes with 127 elements
[1650465529.768662] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.768668] [ndv4:14205:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib6:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.768706] [ndv4:14205:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib6 length=2048) failed: Invalid argument
[1650465529.768710] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.768720] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x55fec40 [id=127 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.768744] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 127 events 0x1 mode thread_spinlock
[1650465529.768754] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[37]=0x55f6030 using rc_mlx5/mlx5_ib6:1 on worker 0x29ca770
[1650465529.768769] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.768774] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.768891] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.768897] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.768863] [ndv4:13552:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465529.769163] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib6:1
[1650465529.770432] [ndv4:12741:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x48c0050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xc4e8
[1650465529.770727] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.770735] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.770744] [ndv4:12741:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b7353eed008 of 151544 bytes with 1052 elements
[1650465529.773084] [ndv4:13443:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x4d47050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xc4c7
[1650465529.773183] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.773190] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.773199] [ndv4:13443:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b7518dad008 of 151544 bytes with 1052 elements
[1650465529.774468] [ndv4:12741:0] ib_md.c:812 UCX DEBUG registered memory 0x2b735da00000..0x2b7360000000 on mlx5_ib2 lkey 0x81b00 rkey 0x81b00 access 0xf flags 0x3e4
[1650465529.774495] [ndv4:12741:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b735da00018 of 39845864 bytes with 4752 elements
[1650465529.774663] [ndv4:12741:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x48c0050
[1650465529.774697] [ndv4:12741:0] ucp_worker.c:1159 UCX DEBUG created interface[18]=0x48c0050 using dc_mlx5/mlx5_ib2:1 on worker 0x37328d0
[1650465529.774827] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.774840] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.774932] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.774937] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.775320] [ndv4:12741:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465529.776352] [ndv4:12741:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.776898] [ndv4:13443:0] ib_md.c:812 UCX DEBUG registered memory 0x2b7516600000..0x2b7518c00000 on mlx5_ib4 lkey 0x80f00 rkey 0x80f00 access 0xf flags 0x3e4
[1650465529.776915] [ndv4:13443:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b7516600018 of 39845864 bytes with 4752 elements
[1650465529.777051] [ndv4:13443:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x4d47050
[1650465529.777087] [ndv4:13443:0] ucp_worker.c:1159 UCX DEBUG created interface[28]=0x4d47050 using dc_mlx5/mlx5_ib4:1 on worker 0x2d7f8d0
[1650465529.777228] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.777239] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.777327] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.777332] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.776764] [ndv4:12741:0] ib_iface.c:994 UCX DEBUG iface=0x4a9c060: created UD QP 0xc4de on mlx5_ib2:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.777331] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.777502] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.777507] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.777640] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.777645] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.776676] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x2e34050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 2)
[1650465529.778001] [ndv4:12741:0] ib_md.c:812 UCX DEBUG registered memory 0x2b7353f12000..0x2b7353f97000 on mlx5_ib2 lkey 0x81c00 rkey 0x81c00 access 0xf flags 0x3e4
[1650465529.778007] [ndv4:12741:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b7353f12018 of 544744 bytes with 128 elements
[1650465529.778012] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.779744] [ndv4:14421:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465529.780384] [ndv4:14419:0] async.c:228 UCX DEBUG added async handler 0x8cef50 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650465529.780414] [ndv4:14419:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650465529.780420] [ndv4:14419:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650465529.782937] [ndv4:13552:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.783306] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.783333] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.783337] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.783395] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.783400] [ndv4:13552:0] ib_iface.c:994 UCX DEBUG iface=0x2c52280: created UD QP 0xc48f on mlx5_ib1:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.783408] [ndv4:13552:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465529.783867] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465529.783879] [ndv4:14205:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib6:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.783921] [ndv4:14205:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib6 length=2048) failed: Invalid argument
[1650465529.783924] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.783937] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x5484360 [id=129 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.783964] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 129 events 0x1 mode thread_spinlock
[1650465529.784046] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.784640] [ndv4:14205:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465529.786763] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x4a9c060: adding gid fe80::15:5dff:fd33:fffd to hash on device mlx5_ib2 port 1 index 0)
[1650465529.786796] [ndv4:13443:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465529.786340] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x2e34050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 3)
[1650465529.786912] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x2e34050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 4)
[1650465529.787129] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x2e34050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 5)
[1650465529.787457] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x2e34050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 6)
[1650465529.787823] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x2e34050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 7)
[1650465529.787833] [ndv4:14949:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib2:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.787867] [ndv4:14949:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.787872] [ndv4:14949:0] async.c:228 UCX DEBUG added async handler 0x2d16e10 [id=103 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.787899] [ndv4:14949:0] async.c:506 UCX DEBUG listening to async event fd 103 events 0x5 mode thread_spinlock
[1650465529.787911] [ndv4:14949:0] ucp_worker.c:1159 UCX DEBUG created interface[20]=0x2e34050 using ud_mlx5/mlx5_ib2:1 on worker 0x19ac660
[1650465529.788245] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.788251] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.788344] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.788348] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.788335] [ndv4:13443:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.788780] [ndv4:13443:0] ib_iface.c:994 UCX DEBUG iface=0x4f23220: created UD QP 0xc4bd on mlx5_ib4:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.789152] [ndv4:14419:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.789169] [ndv4:14419:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.789177] [ndv4:14419:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465529.789256] [ndv4:14419:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650465529.790680] [ndv4:14419:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465529.790686] [ndv4:14419:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465529.790921] [ndv4:14419:0] ib_md.c:1604 UCX DEBUG mlx5_ib1: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465529.791038] [ndv4:14419:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.791044] [ndv4:14419:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.789495] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.789542] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.789549] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.789631] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.789637] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.790035] [ndv4:13443:0] ib_md.c:812 UCX DEBUG registered memory 0x2b7518dd2000..0x2b7518e57000 on mlx5_ib4 lkey 0x81000 rkey 0x81000 access 0xf flags 0x3e4
[1650465529.790041] [ndv4:13443:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b7518dd2018 of 544744 bytes with 128 elements
[1650465529.790046] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.790847] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x4f23220: adding gid fe80::15:5dff:fd33:ffff to hash on device mlx5_ib4 port 1 index 0)
[1650465529.791212] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x4f23220: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 1)
[1650465529.791523] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x4f23220: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 2)
[1650465529.789069] [ndv4:14949:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465529.792653] [ndv4:14205:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x57cc050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xc4b7
[1650465529.792829] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.792836] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.792844] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2afdbc1be008 of 151544 bytes with 1052 elements
[1650465529.793748] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.793761] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.793921] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465529.793926] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465529.794370] [ndv4:13552:0] ib_md.c:812 UCX DEBUG registered memory 0x2b266682a000..0x2b26668af000 on mlx5_ib1 lkey 0x81a00 rkey 0x81a00 access 0xf flags 0x3e4
[1650465529.794376] [ndv4:13552:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b266682a018 of 544744 bytes with 128 elements
[1650465529.794380] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.794917] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x2c52280: adding gid fe80::15:5dff:fd33:fffc to hash on device mlx5_ib1 port 1 index 0)
[1650465529.795408] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x2c52280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 1)
[1650465529.795708] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x2c52280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 2)
[1650465529.796746] [ndv4:14205:0] ib_md.c:812 UCX DEBUG registered memory 0x2afdb9a00000..0x2afdbc000000 on mlx5_ib6 lkey 0x80a00 rkey 0x80a00 access 0xf flags 0x3e4
[1650465529.796767] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2afdb9a00018 of 39845864 bytes with 4752 elements
[1650465529.796905] [ndv4:14205:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x57cc050
[1650465529.796943] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[38]=0x57cc050 using dc_mlx5/mlx5_ib6:1 on worker 0x29ca770
[1650465529.797246] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.797256] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.797341] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.797346] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.797903] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib6:1
[1650465529.798217] [ndv4:14421:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465529.798260] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.798263] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.798278] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.798723] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.798733] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465529.799698] [ndv4:14421:0] ib_iface.c:994 UCX DEBUG iface=0x16a2cd0: created RC QP 0xde95 on mlx5_ib0:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465529.800702] [ndv4:14949:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465529.800721] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.800724] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.800776] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.801240] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.801247] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465529.802007] [ndv4:14949:0] ib_iface.c:994 UCX DEBUG iface=0x2f340a0: created RC QP 0xc454 on mlx5_ib3:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465529.802816] [ndv4:14949:0] ucp_worker.c:1159 UCX DEBUG created interface[21]=0x2f340a0 using rc_verbs/mlx5_ib3:1 on worker 0x19ac660
[1650465529.802894] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.802900] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.802964] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.802969] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.801628] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x4f23220: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 3)
[1650465529.802320] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x4f23220: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 4)
[1650465529.802778] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x4f23220: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 5)
[1650465529.803450] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x4f23220: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 6)
[1650465529.803471] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x4f23220: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 7)
[1650465529.801686] [ndv4:14421:0] ucp_worker.c:1159 UCX DEBUG created interface[6]=0x16a2cd0 using rc_verbs/mlx5_ib0:1 on worker 0x1f71f60
[1650465529.801744] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.801751] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.801913] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.801919] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.802329] [ndv4:14421:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465529.803760] [ndv4:13443:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.803769] [ndv4:13443:0] async.c:228 UCX DEBUG added async handler 0x42343a0 [id=116 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.803799] [ndv4:13443:0] async.c:506 UCX DEBUG listening to async event fd 116 events 0x5 mode thread_spinlock
[1650465529.803811] [ndv4:13443:0] ucp_worker.c:1159 UCX DEBUG created interface[29]=0x4f23220 using ud_verbs/mlx5_ib4:1 on worker 0x2d7f8d0
[1650465529.803910] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.803916] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.804010] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.804015] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.803666] [ndv4:14949:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465529.804862] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x2c52280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 3)
[1650465529.805281] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x2c52280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 4)
[1650465529.804323] [ndv4:13443:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465529.805913] [ndv4:13443:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.806277] [ndv4:13443:0] ib_iface.c:994 UCX DEBUG iface=0x5041050: created UD QP 0xc4be on mlx5_ib4:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.806285] [ndv4:13443:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465529.806785] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.806902] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.806908] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.806922] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.806925] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.807213] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.807655] [ndv4:14205:0] ib_iface.c:994 UCX DEBUG iface=0x59a8340: created UD QP 0xc49d on mlx5_ib6:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.808275] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.808779] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.808788] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.808983] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.808988] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.809417] [ndv4:14205:0] ib_md.c:812 UCX DEBUG registered memory 0x2afdbc1e3000..0x2afdbc268000 on mlx5_ib6 lkey 0x80b00 rkey 0x80b00 access 0xf flags 0x3e4
[1650465529.809423] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2afdbc1e3018 of 544744 bytes with 128 elements
[1650465529.809427] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.810173] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x59a8340: adding gid fe80::15:5dff:fd34:1 to hash on device mlx5_ib6 port 1 index 0)
[1650465529.807290] [ndv4:13443:0] ib_md.c:812 UCX DEBUG registered memory 0x2b7518e57000..0x2b7518edc000 on mlx5_ib4 lkey 0x81100 rkey 0x81100 access 0xf flags 0x3e4
[1650465529.807297] [ndv4:13443:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b7518e57018 of 544744 bytes with 128 elements
[1650465529.807302] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.807835] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x5041050: adding gid fe80::15:5dff:fd33:ffff to hash on device mlx5_ib4 port 1 index 0)
[1650465529.808384] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x5041050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 1)
[1650465529.809889] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x5041050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 2)
[1650465529.810328] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x5041050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 3)
[1650465529.811041] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x5041050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 4)
[1650465529.810566] [ndv4:14419:0] topo.c:99 UCX DEBUG bus id 0x102000000 doesn't exist. sys_dev = 1
[1650465529.810591] [ndv4:14419:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465529.810985] [ndv4:14421:0] ib_device.c:1394 UCX DEBUG max IB CQE size is 128
[1650465529.811858] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x5041050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 5)
[1650465529.812081] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x5041050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 6)
[1650465529.812102] [ndv4:14421:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.812111] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.812114] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.812162] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.812663] [ndv4:14421:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x203b010 of 8176 bytes with 127 elements
[1650465529.814261] [ndv4:14949:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.814275] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.814278] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.814331] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.814793] [ndv4:14949:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x3255010 of 8176 bytes with 127 elements
[1650465529.815014] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.815019] [ndv4:14949:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib3:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.815057] [ndv4:14949:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib3 length=2048) failed: Invalid argument
[1650465529.815061] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.815071] [ndv4:14949:0] async.c:228 UCX DEBUG added async handler 0x10d1d40 [id=106 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.815091] [ndv4:14949:0] async.c:506 UCX DEBUG listening to async event fd 106 events 0x1 mode thread_spinlock
[1650465529.815101] [ndv4:14949:0] ucp_worker.c:1159 UCX DEBUG created interface[22]=0x3081030 using rc_mlx5/mlx5_ib3:1 on worker 0x19ac660
[1650465529.814922] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x4a9c060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 1)
[1650465529.812925] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.812950] [ndv4:14421:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib0:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.813003] [ndv4:14421:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650465529.813009] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.813967] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x2c52280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 5)
[1650465529.814250] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x2c52280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 6)
[1650465529.814516] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x2c52280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 7)
[1650465529.814523] [ndv4:13552:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib1:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.814559] [ndv4:13552:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.814564] [ndv4:13552:0] async.c:228 UCX DEBUG added async handler 0x2b6ac30 [id=96 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.814645] [ndv4:13552:0] async.c:506 UCX DEBUG listening to async event fd 96 events 0x5 mode thread_spinlock
[1650465529.814659] [ndv4:13552:0] ucp_worker.c:1159 UCX DEBUG created interface[15]=0x2c52280 using ud_mlx5/mlx5_ib1:1 on worker 0x24108d0
[1650465529.815001] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.815008] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.815280] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.815286] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.818805] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x59a8340: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 1)
[1650465529.819106] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x59a8340: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 2)
[1650465529.819549] [ndv4:14421:0] async.c:228 UCX DEBUG added async handler 0x15f8390 [id=85 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.819660] [ndv4:14421:0] async.c:506 UCX DEBUG listening to async event fd 85 events 0x1 mode thread_spinlock
[1650465529.819679] [ndv4:14421:0] ucp_worker.c:1159 UCX DEBUG created interface[7]=0x16ae710 using rc_mlx5/mlx5_ib0:1 on worker 0x1f71f60
[1650465529.819860] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.819867] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.820060] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.820065] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.821828] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x5041050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 7)
[1650465529.821837] [ndv4:13443:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib4:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.821867] [ndv4:13443:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.821872] [ndv4:13443:0] async.c:228 UCX DEBUG added async handler 0x4a2cc00 [id=117 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.821894] [ndv4:13443:0] async.c:506 UCX DEBUG listening to async event fd 117 events 0x5 mode thread_spinlock
[1650465529.821904] [ndv4:13443:0] ucp_worker.c:1159 UCX DEBUG created interface[30]=0x5041050 using ud_mlx5/mlx5_ib4:1 on worker 0x2d7f8d0
[1650465529.821972] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.821978] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.822035] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.822040] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.822403] [ndv4:13443:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib5:1
[1650465529.823971] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.823981] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.824052] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.824056] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.823661] [ndv4:13443:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465529.823678] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.823681] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.823732] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.824171] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.824177] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465529.824912] [ndv4:13443:0] ib_iface.c:994 UCX DEBUG iface=0x51410a0: created RC QP 0xc44e on mlx5_ib5:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465529.823616] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x4a9c060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 2)
[1650465529.824072] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x4a9c060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 3)
[1650465529.824459] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x4a9c060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 4)
[1650465529.824922] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x4a9c060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 5)
[1650465529.825352] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x4a9c060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 6)
[1650465529.825709] [ndv4:13443:0] ucp_worker.c:1159 UCX DEBUG created interface[31]=0x51410a0 using rc_verbs/mlx5_ib5:1 on worker 0x2d7f8d0
[1650465529.825984] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.825992] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.826124] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.826129] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.826736] [ndv4:13443:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib5:1
[1650465529.830763] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x59a8340: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 3)
[1650465529.830807] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x59a8340: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 4)
[1650465529.830823] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x59a8340: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 5)
[1650465529.831331] [ndv4:14419:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465529.831339] [ndv4:14419:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465529.834391] [ndv4:13552:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465529.834704] [ndv4:14949:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465529.835366] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x4a9c060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 7)
[1650465529.835669] [ndv4:13552:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465529.835687] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.835690] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.835745] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.836655] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.836663] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465529.835721] [ndv4:12741:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.835732] [ndv4:12741:0] async.c:228 UCX DEBUG added async handler 0x4019e70 [id=102 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.835756] [ndv4:12741:0] async.c:506 UCX DEBUG listening to async event fd 102 events 0x5 mode thread_spinlock
[1650465529.835777] [ndv4:12741:0] ucp_worker.c:1159 UCX DEBUG created interface[19]=0x4a9c060 using ud_verbs/mlx5_ib2:1 on worker 0x37328d0
[1650465529.835843] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.835849] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.835929] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.835935] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.836592] [ndv4:12741:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465529.836145] [ndv4:14949:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.836160] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.836163] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.836219] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.836659] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465529.836666] [ndv4:14949:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib3:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.836699] [ndv4:14949:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib3 length=2048) failed: Invalid argument
[1650465529.836703] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.836709] [ndv4:14949:0] async.c:228 UCX DEBUG added async handler 0x1053e70 [id=108 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.836729] [ndv4:14949:0] async.c:506 UCX DEBUG listening to async event fd 108 events 0x1 mode thread_spinlock
[1650465529.837311] [ndv4:12741:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.837365] [ndv4:13552:0] ib_iface.c:994 UCX DEBUG iface=0x327b0a0: created RC QP 0xc4df on mlx5_ib2:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465529.837267] [ndv4:14949:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465529.837678] [ndv4:12741:0] ib_iface.c:994 UCX DEBUG iface=0x4bba050: created UD QP 0xc4e0 on mlx5_ib2:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.837687] [ndv4:12741:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465529.838109] [ndv4:13552:0] ucp_worker.c:1159 UCX DEBUG created interface[16]=0x327b0a0 using rc_verbs/mlx5_ib2:1 on worker 0x24108d0
[1650465529.838279] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.838286] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.838423] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.838429] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.838396] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.838574] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.838622] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.838641] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.838646] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.838701] [ndv4:13443:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.838716] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.838719] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.838773] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.839256] [ndv4:13443:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x5462010 of 8176 bytes with 127 elements
[1650465529.839524] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.839532] [ndv4:13443:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib5:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.839569] [ndv4:13443:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib5 length=2048) failed: Invalid argument
[1650465529.839635] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.839647] [ndv4:13443:0] async.c:228 UCX DEBUG added async handler 0x4234e50 [id=120 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.839667] [ndv4:13443:0] async.c:506 UCX DEBUG listening to async event fd 120 events 0x1 mode thread_spinlock
[1650465529.839678] [ndv4:13443:0] ucp_worker.c:1159 UCX DEBUG created interface[32]=0x528e030 using rc_mlx5/mlx5_ib5:1 on worker 0x2d7f8d0
[1650465529.839755] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.839762] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.839843] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.839848] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.839058] [ndv4:13552:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465529.839084] [ndv4:12741:0] ib_md.c:812 UCX DEBUG registered memory 0x2b7360035000..0x2b73600ba000 on mlx5_ib2 lkey 0x81d00 rkey 0x81d00 access 0xf flags 0x3e4
[1650465529.839091] [ndv4:12741:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b7360035018 of 544744 bytes with 128 elements
[1650465529.839095] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.839130] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x4bba050: adding gid fe80::15:5dff:fd33:fffd to hash on device mlx5_ib2 port 1 index 0)
[1650465529.839146] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x4bba050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 1)
[1650465529.839222] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x4bba050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 2)
[1650465529.839633] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x4bba050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 3)
[1650465529.840504] [ndv4:13443:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib5:1
[1650465529.840661] [ndv4:13552:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.840679] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.840682] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.840733] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.841206] [ndv4:13552:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x359c010 of 8176 bytes with 127 elements
[1650465529.840747] [ndv4:14421:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465529.842075] [ndv4:14421:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.842086] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.842089] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.842105] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.841797] [ndv4:13443:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.841814] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.841818] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.841876] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.842247] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465529.842254] [ndv4:13443:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib5:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.842288] [ndv4:13443:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib5 length=2048) failed: Invalid argument
[1650465529.842292] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.842300] [ndv4:13443:0] async.c:228 UCX DEBUG added async handler 0x5296f80 [id=122 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.842325] [ndv4:13443:0] async.c:506 UCX DEBUG listening to async event fd 122 events 0x1 mode thread_spinlock
[1650465529.841687] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.841696] [ndv4:13552:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib2:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.841731] [ndv4:13552:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib2 length=2048) failed: Invalid argument
[1650465529.841735] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.841748] [ndv4:13552:0] async.c:228 UCX DEBUG added async handler 0x1b595f0 [id=99 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.841772] [ndv4:13552:0] async.c:506 UCX DEBUG listening to async event fd 99 events 0x1 mode thread_spinlock
[1650465529.841795] [ndv4:13552:0] ucp_worker.c:1159 UCX DEBUG created interface[17]=0x33c8030 using rc_mlx5/mlx5_ib2:1 on worker 0x24108d0
[1650465529.841815] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.841822] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.841978] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.841983] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.842347] [ndv4:13552:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465529.842494] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465529.842502] [ndv4:14421:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib0:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.842534] [ndv4:14421:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650465529.842538] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.842548] [ndv4:14421:0] async.c:228 UCX DEBUG added async handler 0x1606a40 [id=87 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.842569] [ndv4:14421:0] async.c:506 UCX DEBUG listening to async event fd 87 events 0x1 mode thread_spinlock
[1650465529.842934] [ndv4:13443:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465529.843307] [ndv4:14421:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465529.844650] [ndv4:14949:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x3257050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xc46a
[1650465529.844736] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.844743] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.844751] [ndv4:14949:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2ad4c8310008 of 151544 bytes with 1052 elements
[1650465529.848421] [ndv4:14949:0] ib_md.c:812 UCX DEBUG registered memory 0x2ad4c5c00000..0x2ad4c8200000 on mlx5_ib3 lkey 0x81000 rkey 0x81000 access 0xf flags 0x3e4
[1650465529.848440] [ndv4:14949:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2ad4c5c00018 of 39845864 bytes with 4752 elements
[1650465529.848602] [ndv4:14949:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x3257050
[1650465529.848636] [ndv4:14949:0] ucp_worker.c:1159 UCX DEBUG created interface[23]=0x3257050 using dc_mlx5/mlx5_ib3:1 on worker 0x19ac660
[1650465529.848756] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.848768] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.848882] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.848886] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.849484] [ndv4:14949:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465529.850215] [ndv4:13443:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x5464050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xc464
[1650465529.850336] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.850343] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.850351] [ndv4:13443:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b751b6dd008 of 151544 bytes with 1052 elements
[1650465529.852527] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x4bba050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 4)
[1650465529.852557] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x4bba050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 5)
[1650465529.852631] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x4bba050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 6)
[1650465529.852646] [ndv4:12741:0] ud_iface.c:393 UCX DEBUG iface 0x4bba050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 7)
[1650465529.852653] [ndv4:12741:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib2:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.852683] [ndv4:12741:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.852687] [ndv4:12741:0] async.c:228 UCX DEBUG added async handler 0x45a5c60 [id=103 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.852707] [ndv4:12741:0] async.c:506 UCX DEBUG listening to async event fd 103 events 0x5 mode thread_spinlock
[1650465529.852718] [ndv4:12741:0] ucp_worker.c:1159 UCX DEBUG created interface[20]=0x4bba050 using ud_mlx5/mlx5_ib2:1 on worker 0x37328d0
[1650465529.852775] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.852781] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.852922] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.852926] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.853372] [ndv4:12741:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465529.850941] [ndv4:14421:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x22a6010: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xdebe
[1650465529.851064] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.851070] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.851091] [ndv4:14421:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2ab787e86008 of 151544 bytes with 1052 elements
[1650465529.850537] [ndv4:14949:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.850907] [ndv4:14949:0] ib_iface.c:994 UCX DEBUG iface=0x27f2700: created UD QP 0xc460 on mlx5_ib3:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.851379] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.851731] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.851738] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.851894] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.851900] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.852264] [ndv4:14949:0] ib_md.c:812 UCX DEBUG registered memory 0x2ad4c8335000..0x2ad4c83ba000 on mlx5_ib3 lkey 0x81100 rkey 0x81100 access 0xf flags 0x3e4
[1650465529.852270] [ndv4:14949:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ad4c8335018 of 544744 bytes with 128 elements
[1650465529.852275] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.852409] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x27f2700: adding gid fe80::15:5dff:fd33:fffe to hash on device mlx5_ib3 port 1 index 0)
[1650465529.852660] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x27f2700: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 1)
[1650465529.852777] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x27f2700: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 2)
[1650465529.853070] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x27f2700: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 3)
[1650465529.853409] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x27f2700: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 4)
[1650465529.853426] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x27f2700: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 5)
[1650465529.853517] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x27f2700: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 6)
[1650465529.853894] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x27f2700: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 7)
[1650465529.853135] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x59a8340: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 6)
[1650465529.853342] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x59a8340: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 7)
[1650465529.853674] [ndv4:14205:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.853681] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x59a8e90 [id=130 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.853703] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 130 events 0x5 mode thread_spinlock
[1650465529.853717] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[39]=0x59a8340 using ud_verbs/mlx5_ib6:1 on worker 0x29ca770
[1650465529.853877] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.853883] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.854052] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.854058] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.854861] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib6:1
[1650465529.854525] [ndv4:14419:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465529.854533] [ndv4:14419:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465529.854880] [ndv4:14419:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465529.854884] [ndv4:14419:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465529.854131] [ndv4:13443:0] ib_md.c:812 UCX DEBUG registered memory 0x2b7519000000..0x2b751b600000 on mlx5_ib5 lkey 0x81100 rkey 0x81100 access 0xf flags 0x3e4
[1650465529.854146] [ndv4:13443:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b7519000018 of 39845864 bytes with 4752 elements
[1650465529.854282] [ndv4:13443:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x5464050
[1650465529.854317] [ndv4:13443:0] ucp_worker.c:1159 UCX DEBUG created interface[33]=0x5464050 using dc_mlx5/mlx5_ib5:1 on worker 0x2d7f8d0
[1650465529.854766] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.854777] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.854639] [ndv4:12741:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465529.854658] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.854662] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.854714] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.855052] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.855060] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465529.855522] [ndv4:12741:0] ib_iface.c:994 UCX DEBUG iface=0x4cba0a0: created RC QP 0xc461 on mlx5_ib3:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465529.854117] [ndv4:14949:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.854128] [ndv4:14949:0] async.c:228 UCX DEBUG added async handler 0x34338e0 [id=109 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.854151] [ndv4:14949:0] async.c:506 UCX DEBUG listening to async event fd 109 events 0x5 mode thread_spinlock
[1650465529.854161] [ndv4:14949:0] ucp_worker.c:1159 UCX DEBUG created interface[24]=0x27f2700 using ud_verbs/mlx5_ib3:1 on worker 0x19ac660
[1650465529.854465] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.854472] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.854789] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.854796] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.855376] [ndv4:14949:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465529.854881] [ndv4:14421:0] ib_md.c:812 UCX DEBUG registered memory 0x2ab78ee00000..0x2ab791400000 on mlx5_ib0 lkey 0x81f00 rkey 0x81f00 access 0xf flags 0x3e4
[1650465529.854904] [ndv4:14421:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2ab78ee00018 of 39845864 bytes with 4752 elements
[1650465529.855037] [ndv4:14421:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x22a6010
[1650465529.855073] [ndv4:14421:0] ucp_worker.c:1159 UCX DEBUG created interface[8]=0x22a6010 using dc_mlx5/mlx5_ib0:1 on worker 0x1f71f60
[1650465529.855240] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.855251] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.855411] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.855417] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.856054] [ndv4:12741:0] ucp_worker.c:1159 UCX DEBUG created interface[21]=0x4cba0a0 using rc_verbs/mlx5_ib3:1 on worker 0x37328d0
[1650465529.856117] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.856125] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.856128] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.856412] [ndv4:14949:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.856542] [ndv4:14205:0] ib_iface.c:994 UCX DEBUG iface=0x5ac6460: created UD QP 0xc49e on mlx5_ib6:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.856548] [ndv4:14205:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465529.856697] [ndv4:14949:0] ib_iface.c:994 UCX DEBUG iface=0x3551460: created UD QP 0xc462 on mlx5_ib3:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.856705] [ndv4:14949:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465529.857152] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.857341] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.857348] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.857431] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.857435] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.857156] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.857342] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.857348] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.857376] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.857383] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.857820] [ndv4:14949:0] ib_md.c:812 UCX DEBUG registered memory 0x2ad4c83ba000..0x2ad4c843f000 on mlx5_ib3 lkey 0x81200 rkey 0x81200 access 0xf flags 0x3e4
[1650465529.857826] [ndv4:14949:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ad4c83ba018 of 544744 bytes with 128 elements
[1650465529.857831] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.857983] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x3551460: adding gid fe80::15:5dff:fd33:fffe to hash on device mlx5_ib3 port 1 index 0)
[1650465529.858383] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x3551460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 1)
[1650465529.858744] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x3551460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 2)
[1650465529.859237] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x3551460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 3)
[1650465529.857761] [ndv4:14205:0] ib_md.c:812 UCX DEBUG registered memory 0x2afdbc268000..0x2afdbc2ed000 on mlx5_ib6 lkey 0x80c00 rkey 0x80c00 access 0xf flags 0x3e4
[1650465529.857768] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2afdbc268018 of 544744 bytes with 128 elements
[1650465529.857772] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.858201] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x5ac6460: adding gid fe80::15:5dff:fd34:1 to hash on device mlx5_ib6 port 1 index 0)
[1650465529.858340] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x5ac6460: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 1)
[1650465529.858958] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x5ac6460: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 2)
[1650465529.859226] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x5ac6460: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 3)
[1650465529.859386] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x5ac6460: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 4)
[1650465529.857726] [ndv4:13552:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.857745] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.857748] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.857803] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.858233] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465529.858241] [ndv4:13552:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib2:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.858274] [ndv4:13552:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib2 length=2048) failed: Invalid argument
[1650465529.858277] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.858284] [ndv4:13552:0] async.c:228 UCX DEBUG added async handler 0x1b394c0 [id=101 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.858311] [ndv4:13552:0] async.c:506 UCX DEBUG listening to async event fd 101 events 0x1 mode thread_spinlock
[1650465529.858834] [ndv4:13552:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465529.861622] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.861631] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.862507] [ndv4:14421:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465529.862821] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.862833] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.863551] [ndv4:12741:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465529.862616] [ndv4:14419:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465529.862624] [ndv4:14419:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465529.862897] [ndv4:14419:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465529.862561] [ndv4:13443:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib5:1
[1650465529.863965] [ndv4:14421:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.864283] [ndv4:13443:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.864332] [ndv4:14421:0] ib_iface.c:994 UCX DEBUG iface=0x2049050: created UD QP 0xde9f on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.864691] [ndv4:13443:0] ib_iface.c:994 UCX DEBUG iface=0x5640060: created UD QP 0xc45a on mlx5_ib5:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.864602] [ndv4:12741:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.864612] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.864615] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.864665] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.864923] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.865144] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.865151] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.865176] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.865180] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.865538] [ndv4:14421:0] ib_md.c:812 UCX DEBUG registered memory 0x2ab787eab000..0x2ab787f30000 on mlx5_ib0 lkey 0x82000 rkey 0x82000 access 0xf flags 0x3e4
[1650465529.865544] [ndv4:14421:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ab787eab018 of 544744 bytes with 128 elements
[1650465529.865548] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.865627] [ndv4:14421:0] ud_iface.c:393 UCX DEBUG iface 0x2049050: adding gid fe80::15:5dff:fd33:fffb to hash on device mlx5_ib0 port 1 index 0)
[1650465529.865877] [ndv4:14421:0] ud_iface.c:393 UCX DEBUG iface 0x2049050: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650465529.865195] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.865378] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.865386] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.865521] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.865526] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.865076] [ndv4:12741:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x4e08010 of 8176 bytes with 127 elements
[1650465529.865282] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.865288] [ndv4:12741:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib3:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.865324] [ndv4:12741:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib3 length=2048) failed: Invalid argument
[1650465529.865328] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.865338] [ndv4:12741:0] async.c:228 UCX DEBUG added async handler 0x3c2a3b0 [id=106 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.865357] [ndv4:12741:0] async.c:506 UCX DEBUG listening to async event fd 106 events 0x1 mode thread_spinlock
[1650465529.865367] [ndv4:12741:0] ucp_worker.c:1159 UCX DEBUG created interface[22]=0x4e0f100 using rc_mlx5/mlx5_ib3:1 on worker 0x37328d0
[1650465529.865477] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.865483] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.865555] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.865561] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.866027] [ndv4:13443:0] ib_md.c:812 UCX DEBUG registered memory 0x2b751b702000..0x2b751b787000 on mlx5_ib5 lkey 0x81200 rkey 0x81200 access 0xf flags 0x3e4
[1650465529.866034] [ndv4:13443:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b751b702018 of 544744 bytes with 128 elements
[1650465529.866038] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.866233] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x5640060: adding gid fe80::15:5dff:fd34:0 to hash on device mlx5_ib5 port 1 index 0)
[1650465529.866255] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x5640060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 1)
[1650465529.867050] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x5640060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 2)
[1650465529.867500] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x5640060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 3)
[1650465529.868031] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x5640060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 4)
[1650465529.868354] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x5640060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 5)
[1650465529.868860] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x5640060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 6)
[1650465529.869571] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x5640060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 7)
[1650465529.866473] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x5ac6460: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 5)
[1650465529.866973] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x5ac6460: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 6)
[1650465529.867480] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x5ac6460: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 7)
[1650465529.867485] [ndv4:14205:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib6:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.867518] [ndv4:14205:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.867522] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x5484970 [id=131 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.867542] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 131 events 0x5 mode thread_spinlock
[1650465529.867553] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[40]=0x5ac6460 using ud_mlx5/mlx5_ib6:1 on worker 0x29ca770
[1650465529.867717] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.867724] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.867869] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.867874] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.868325] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib7:1
[1650465529.866557] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x3551460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 4)
[1650465529.866969] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x3551460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 5)
[1650465529.867467] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x3551460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 6)
[1650465529.868351] [ndv4:14949:0] ud_iface.c:393 UCX DEBUG iface 0x3551460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 7)
[1650465529.868361] [ndv4:14949:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib3:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.868391] [ndv4:14949:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.868396] [ndv4:14949:0] async.c:228 UCX DEBUG added async handler 0x3551f60 [id=110 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.868418] [ndv4:14949:0] async.c:506 UCX DEBUG listening to async event fd 110 events 0x5 mode thread_spinlock
[1650465529.868430] [ndv4:14949:0] ucp_worker.c:1159 UCX DEBUG created interface[25]=0x3551460 using ud_mlx5/mlx5_ib3:1 on worker 0x19ac660
[1650465529.868812] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.868819] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.869010] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.869015] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.869506] [ndv4:14949:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465529.868567] [ndv4:13552:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x359e050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xc4f6
[1650465529.868933] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.868940] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.868949] [ndv4:13552:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b26690b0008 of 151544 bytes with 1052 elements
[1650465529.870636] [ndv4:14949:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465529.870652] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.870655] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.870709] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.869896] [ndv4:13443:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.869907] [ndv4:13443:0] async.c:228 UCX DEBUG added async handler 0x511c9b0 [id=123 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.869937] [ndv4:13443:0] async.c:506 UCX DEBUG listening to async event fd 123 events 0x5 mode thread_spinlock
[1650465529.869947] [ndv4:13443:0] ucp_worker.c:1159 UCX DEBUG created interface[34]=0x5640060 using ud_verbs/mlx5_ib5:1 on worker 0x2d7f8d0
[1650465529.870108] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.870115] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.870336] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.870342] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.870025] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465529.870045] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.870048] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.870097] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.871195] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.871201] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465529.871032] [ndv4:13443:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib5:1
[1650465529.871173] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.871180] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465529.871926] [ndv4:14949:0] ib_iface.c:994 UCX DEBUG iface=0x36510a0: created RC QP 0xc4bf on mlx5_ib4:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465529.872101] [ndv4:14205:0] ib_iface.c:994 UCX DEBUG iface=0x5bc60a0: created RC QP 0xc43a on mlx5_ib7:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465529.872366] [ndv4:13443:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.872819] [ndv4:14419:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib2 vendor_id: 0x15b3 device_id: 4124
[1650465529.873199] [ndv4:14419:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib2: disable ODP because it's not supported for DevX QP
[1650465529.872752] [ndv4:13552:0] ib_md.c:812 UCX DEBUG registered memory 0x2b2666a00000..0x2b2669000000 on mlx5_ib2 lkey 0x81e00 rkey 0x81e00 access 0xf flags 0x3e4
[1650465529.872774] [ndv4:13552:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b2666a00018 of 39845864 bytes with 4752 elements
[1650465529.872917] [ndv4:13552:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x359e050
[1650465529.872951] [ndv4:13552:0] ucp_worker.c:1159 UCX DEBUG created interface[18]=0x359e050 using dc_mlx5/mlx5_ib2:1 on worker 0x24108d0
[1650465529.872972] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.872981] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.873086] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.873091] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.873472] [ndv4:13552:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465529.872675] [ndv4:13443:0] ib_iface.c:994 UCX DEBUG iface=0x575e460: created UD QP 0xc45b on mlx5_ib5:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.872682] [ndv4:13443:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465529.873227] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.874068] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.874075] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.874090] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465529.874093] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465529.872672] [ndv4:14949:0] ucp_worker.c:1159 UCX DEBUG created interface[26]=0x36510a0 using rc_verbs/mlx5_ib4:1 on worker 0x19ac660
[1650465529.872842] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.872849] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.872920] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.872925] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.873088] [ndv4:14949:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465529.874451] [ndv4:13443:0] ib_md.c:812 UCX DEBUG registered memory 0x2b751b787000..0x2b751b80c000 on mlx5_ib5 lkey 0x81300 rkey 0x81300 access 0xf flags 0x3e4
[1650465529.874457] [ndv4:13443:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b751b787018 of 544744 bytes with 128 elements
[1650465529.874462] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.874761] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x575e460: adding gid fe80::15:5dff:fd34:0 to hash on device mlx5_ib5 port 1 index 0)
[1650465529.873770] [ndv4:12741:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465529.874363] [ndv4:14421:0] ud_iface.c:393 UCX DEBUG iface 0x2049050: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650465529.874740] [ndv4:14421:0] ud_iface.c:393 UCX DEBUG iface 0x2049050: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650465529.875101] [ndv4:14421:0] ud_iface.c:393 UCX DEBUG iface 0x2049050: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650465529.875232] [ndv4:14421:0] ud_iface.c:393 UCX DEBUG iface 0x2049050: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650465529.875021] [ndv4:14949:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.875034] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.875037] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.875090] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.875553] [ndv4:14949:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x3972010 of 8176 bytes with 127 elements
[1650465529.872741] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[41]=0x5bc60a0 using rc_verbs/mlx5_ib7:1 on worker 0x29ca770
[1650465529.872926] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.872933] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.873049] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.873054] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.873679] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib7:1
[1650465529.875896] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.875911] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.875915] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.875967] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.876443] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x5ee7010 of 8176 bytes with 127 elements
[1650465529.875815] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.875823] [ndv4:14949:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib4:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.875859] [ndv4:14949:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib4 length=2048) failed: Invalid argument
[1650465529.875863] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.875875] [ndv4:14949:0] async.c:228 UCX DEBUG added async handler 0x2744690 [id=113 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.875895] [ndv4:14949:0] async.c:506 UCX DEBUG listening to async event fd 113 events 0x1 mode thread_spinlock
[1650465529.875904] [ndv4:14949:0] ucp_worker.c:1159 UCX DEBUG created interface[27]=0x379e030 using rc_mlx5/mlx5_ib4:1 on worker 0x19ac660
[1650465529.876142] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.876149] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.876363] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.876367] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.875769] [ndv4:14421:0] ud_iface.c:393 UCX DEBUG iface 0x2049050: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650465529.876536] [ndv4:14421:0] ud_iface.c:393 UCX DEBUG iface 0x2049050: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650465529.876696] [ndv4:14949:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465529.876938] [ndv4:14421:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.876689] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.876696] [ndv4:14205:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib7:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.876732] [ndv4:14205:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib7 length=2048) failed: Invalid argument
[1650465529.876736] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.876746] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x5af3960 [id=134 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.876766] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 134 events 0x1 mode thread_spinlock
[1650465529.876775] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[42]=0x5d13030 using rc_mlx5/mlx5_ib7:1 on worker 0x29ca770
[1650465529.876857] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.876864] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.877072] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.877078] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.878095] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib7:1
[1650465529.879642] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.879660] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.879663] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.879719] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.880093] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465529.880102] [ndv4:14205:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib7:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.880136] [ndv4:14205:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib7 length=2048) failed: Invalid argument
[1650465529.880139] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.880146] [ndv4:14205:0] async.c:228 UCX DEBUG added async handler 0x20ef2f0 [id=136 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.880164] [ndv4:14205:0] async.c:506 UCX DEBUG listening to async event fd 136 events 0x1 mode thread_spinlock
[1650465529.880834] [ndv4:14205:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465529.881476] [ndv4:13552:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.881814] [ndv4:14419:0] async.c:228 UCX DEBUG added async handler 0x8d6ce0 [id=61 ref 1] uct_ib_async_event_handler() to hash
[1650465529.881840] [ndv4:14419:0] async.c:506 UCX DEBUG listening to async event fd 61 events 0x1 mode thread_spinlock
[1650465529.881844] [ndv4:14419:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib2' (InfiniBand channel adapter) with 1 ports
[1650465529.881980] [ndv4:13552:0] ib_iface.c:994 UCX DEBUG iface=0x377a060: created UD QP 0xc4ec on mlx5_ib2:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.882079] [ndv4:14419:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.882090] [ndv4:14419:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.882096] [ndv4:14419:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465529.882139] [ndv4:14419:0] ib_md.c:1319 UCX DEBUG mlx5_ib2: using registration cache
[1650465529.883475] [ndv4:14419:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465529.883480] [ndv4:14419:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465529.883703] [ndv4:14419:0] ib_md.c:1604 UCX DEBUG mlx5_ib2: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465529.883977] [ndv4:14419:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.883983] [ndv4:14419:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.882569] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.882872] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.882879] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.883005] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.883010] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.883389] [ndv4:13552:0] ib_md.c:812 UCX DEBUG registered memory 0x2b26690d5000..0x2b266915a000 on mlx5_ib2 lkey 0x81f00 rkey 0x81f00 access 0xf flags 0x3e4
[1650465529.883395] [ndv4:13552:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b26690d5018 of 544744 bytes with 128 elements
[1650465529.883399] [ndv4:13552:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.883982] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x377a060: adding gid fe80::15:5dff:fd33:fffd to hash on device mlx5_ib2 port 1 index 0)
[1650465529.884569] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x377a060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 1)
[1650465529.884753] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x377a060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 2)
[1650465529.883223] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x575e460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 1)
[1650465529.883622] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x575e460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 2)
[1650465529.884134] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x575e460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 3)
[1650465529.884974] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x575e460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 4)
[1650465529.885484] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x575e460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 5)
[1650465529.885890] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x575e460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 6)
[1650465529.886229] [ndv4:13443:0] ud_iface.c:393 UCX DEBUG iface 0x575e460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 7)
[1650465529.886236] [ndv4:13443:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib5:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.886272] [ndv4:13443:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.886278] [ndv4:13443:0] async.c:228 UCX DEBUG added async handler 0x5640f20 [id=124 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.886308] [ndv4:13443:0] async.c:506 UCX DEBUG listening to async event fd 124 events 0x5 mode thread_spinlock
[1650465529.886319] [ndv4:13443:0] ucp_worker.c:1159 UCX DEBUG created interface[35]=0x575e460 using ud_mlx5/mlx5_ib5:1 on worker 0x2d7f8d0
[1650465529.886384] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.886390] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.886509] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.886514] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.882192] [ndv4:12741:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.882211] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.882214] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.882272] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.882670] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465529.882677] [ndv4:12741:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib3:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.882711] [ndv4:12741:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib3 length=2048) failed: Invalid argument
[1650465529.882715] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.882723] [ndv4:12741:0] async.c:228 UCX DEBUG added async handler 0x2e56f00 [id=108 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.882742] [ndv4:12741:0] async.c:506 UCX DEBUG listening to async event fd 108 events 0x1 mode thread_spinlock
[1650465529.883193] [ndv4:12741:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465529.882730] [ndv4:14421:0] async.c:228 UCX DEBUG added async handler 0x160a520 [id=88 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.882767] [ndv4:14421:0] async.c:506 UCX DEBUG listening to async event fd 88 events 0x5 mode thread_spinlock
[1650465529.882791] [ndv4:14421:0] ucp_worker.c:1159 UCX DEBUG created interface[9]=0x2049050 using ud_verbs/mlx5_ib0:1 on worker 0x1f71f60
[1650465529.882896] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.882902] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.883065] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.883070] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.883711] [ndv4:14421:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465529.885002] [ndv4:14421:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.885397] [ndv4:14421:0] ib_iface.c:994 UCX DEBUG iface=0x16b9090: created UD QP 0xdea1 on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.885409] [ndv4:14421:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465529.886028] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.886202] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.886209] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.886235] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465529.886240] [ndv4:14421:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465529.886697] [ndv4:14421:0] ib_md.c:812 UCX DEBUG registered memory 0x2ab787f30000..0x2ab787fb5000 on mlx5_ib0 lkey 0x82100 rkey 0x82100 access 0xf flags 0x3e4
[1650465529.886703] [ndv4:14421:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ab787f30018 of 544744 bytes with 128 elements
[1650465529.886707] [ndv4:14421:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.886944] [ndv4:14421:0] ud_iface.c:393 UCX DEBUG iface 0x16b9090: adding gid fe80::15:5dff:fd33:fffb to hash on device mlx5_ib0 port 1 index 0)
[1650465529.885944] [ndv4:14949:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.885963] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.885966] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.886022] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.886449] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465529.886455] [ndv4:14949:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib4:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.886488] [ndv4:14949:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib4 length=2048) failed: Invalid argument
[1650465529.886492] [ndv4:14949:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.886500] [ndv4:14949:0] async.c:228 UCX DEBUG added async handler 0x37a6f10 [id=115 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.886519] [ndv4:14949:0] async.c:506 UCX DEBUG listening to async event fd 115 events 0x1 mode thread_spinlock
[1650465529.887117] [ndv4:14949:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465529.890809] [ndv4:12741:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x4fe2010: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xc478
[1650465529.890885] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.890892] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.890902] [ndv4:12741:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b7353f99008 of 151544 bytes with 1052 elements
[1650465529.889806] [ndv4:14205:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x5ee9050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xc45d
[1650465529.889939] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.889946] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.889955] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2afdbeaf0008 of 151544 bytes with 1052 elements
[1650465529.895331] [ndv4:12741:0] ib_md.c:812 UCX DEBUG registered memory 0x2b7360200000..0x2b7362800000 on mlx5_ib3 lkey 0x81300 rkey 0x81300 access 0xf flags 0x3e4
[1650465529.895344] [ndv4:12741:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b7360200018 of 39845864 bytes with 4752 elements
[1650465529.895485] [ndv4:12741:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x4fe2010
[1650465529.895506] [ndv4:12741:0] ucp_worker.c:1159 UCX DEBUG created interface[23]=0x4fe2010 using dc_mlx5/mlx5_ib3:1 on worker 0x37328d0
[1650465529.895535] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.895545] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.895660] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.895666] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.896054] [ndv4:12741:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465529.896075] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x377a060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 3)
[1650465529.896406] [ndv4:14419:0] topo.c:99 UCX DEBUG bus id 0x103000000 doesn't exist. sys_dev = 2
[1650465529.896414] [ndv4:14419:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465529.894569] [ndv4:14205:0] ib_md.c:812 UCX DEBUG registered memory 0x2afdbc400000..0x2afdbea00000 on mlx5_ib7 lkey 0x81200 rkey 0x81200 access 0xf flags 0x3e4
[1650465529.894618] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2afdbc400018 of 39845864 bytes with 4752 elements
[1650465529.894809] [ndv4:14205:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x5ee9050
[1650465529.894850] [ndv4:14205:0] ucp_worker.c:1159 UCX DEBUG created interface[43]=0x5ee9050 using dc_mlx5/mlx5_ib7:1 on worker 0x29ca770
[1650465529.894527] [ndv4:14949:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x3974050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xc4d5
[1650465529.894720] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.894728] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.894737] [ndv4:14949:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2ad4cac40008 of 151544 bytes with 1052 elements
[1650465529.897997] [ndv4:13443:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib6:1
[1650465529.899180] [ndv4:13443:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465529.899198] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.899201] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.899256] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.898433] [ndv4:14949:0] ib_md.c:812 UCX DEBUG registered memory 0x2ad4c8600000..0x2ad4cac00000 on mlx5_ib4 lkey 0x81200 rkey 0x81200 access 0xf flags 0x3e4
[1650465529.898455] [ndv4:14949:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2ad4c8600018 of 39845864 bytes with 4752 elements
[1650465529.898645] [ndv4:14949:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x3974050
[1650465529.898676] [ndv4:14949:0] ucp_worker.c:1159 UCX DEBUG created interface[28]=0x3974050 using dc_mlx5/mlx5_ib4:1 on worker 0x19ac660
[1650465529.898819] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.898831] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.898970] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465529.898975] [ndv4:14949:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465529.899697] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.899703] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465529.900412] [ndv4:13443:0] ib_iface.c:994 UCX DEBUG iface=0x585e0a0: created RC QP 0xc49f on mlx5_ib6:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465529.901183] [ndv4:13443:0] ucp_worker.c:1159 UCX DEBUG created interface[36]=0x585e0a0 using rc_verbs/mlx5_ib6:1 on worker 0x2d7f8d0
[1650465529.901271] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.901277] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.901441] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.901446] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.901775] [ndv4:13443:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib6:1
[1650465529.903093] [ndv4:13443:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.903111] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.903115] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.903165] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.903592] [ndv4:13443:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x5b7f010 of 8176 bytes with 127 elements
[1650465529.903841] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465529.903847] [ndv4:13443:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib6:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.903889] [ndv4:13443:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib6 length=2048) failed: Invalid argument
[1650465529.903894] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.903906] [ndv4:13443:0] async.c:228 UCX DEBUG added async handler 0x59b3c40 [id=127 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.903926] [ndv4:13443:0] async.c:506 UCX DEBUG listening to async event fd 127 events 0x1 mode thread_spinlock
[1650465529.903937] [ndv4:13443:0] ucp_worker.c:1159 UCX DEBUG created interface[37]=0x59ab030 using rc_mlx5/mlx5_ib6:1 on worker 0x2d7f8d0
[1650465529.904005] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.904011] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.904162] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465529.904167] [ndv4:13443:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465529.904483] [ndv4:13443:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib6:1
[1650465529.905762] [ndv4:13443:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465529.905779] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465529.905782] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465529.905844] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465529.905341] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.905352] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.905433] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.905438] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.905868] [ndv4:14205:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib7:1
[1650465529.906202] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465529.906209] [ndv4:13443:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib6:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465529.906247] [ndv4:13443:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib6 length=2048) failed: Invalid argument
[1650465529.906252] [ndv4:13443:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465529.906260] [ndv4:13443:0] async.c:228 UCX DEBUG added async handler 0x5839360 [id=129 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465529.906284] [ndv4:13443:0] async.c:506 UCX DEBUG listening to async event fd 129 events 0x1 mode thread_spinlock
[1650465529.906373] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x377a060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 4)
[1650465529.906520] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x377a060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 5)
[1650465529.906567] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x377a060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 6)
[1650465529.906726] [ndv4:13552:0] ud_iface.c:393 UCX DEBUG iface 0x377a060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 7)
[1650465529.906838] [ndv4:13443:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465529.906994] [ndv4:13552:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465529.907004] [ndv4:13552:0] async.c:228 UCX DEBUG added async handler 0x2cf7e70 [id=102 ref 1] uct_ud_iface_async_handler() to hash
[1650465529.907033] [ndv4:13552:0] async.c:506 UCX DEBUG listening to async event fd 102 events 0x5 mode thread_spinlock
[1650465529.907044] [ndv4:13552:0] ucp_worker.c:1159 UCX DEBUG created interface[19]=0x377a060 using ud_verbs/mlx5_ib2:1 on worker 0x24108d0
[1650465529.907060] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.907067] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.907153] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465529.907157] [ndv4:13552:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465529.907013] [ndv4:14205:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.907411] [ndv4:14205:0] ib_iface.c:994 UCX DEBUG iface=0x60c5060: created UD QP 0xc443 on mlx5_ib7:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.907148] [ndv4:12741:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.907531] [ndv4:12741:0] ib_iface.c:994 UCX DEBUG iface=0x51be060: created UD QP 0xc46e on mlx5_ib3:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.908007] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.908051] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.908058] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.908104] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465529.908109] [ndv4:14205:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465529.908477] [ndv4:14205:0] ib_md.c:812 UCX DEBUG registered memory 0x2afdbeb15000..0x2afdbeb9a000 on mlx5_ib7 lkey 0x81300 rkey 0x81300 access 0xf flags 0x3e4
[1650465529.908483] [ndv4:14205:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2afdbeb15018 of 544744 bytes with 128 elements
[1650465529.908487] [ndv4:14205:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465529.908792] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x60c5060: adding gid fe80::15:5dff:fd34:2 to hash on device mlx5_ib7 port 1 index 0)
[1650465529.909092] [ndv4:14205:0] ud_iface.c:393 UCX DEBUG iface 0x60c5060: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 1)
[1650465529.908970] [ndv4:14949:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465529.908123] [ndv4:12741:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465529.908227] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465529.908233] [ndv4:12741:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465529.907549] [ndv4:13552:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465529.908500] [ndv4:13552:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465529.908905] [ndv4:13552:0] ib_iface.c:994 UCX DEBUG iface=0x3898050: created UD QP 0xc4ed on mlx5_ib2:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465529.908912] [ndv4:13552:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465529.907788] [ndv4:14421:0] ud_iface.c:393 UCX DEBUG iface 0x16b9090: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650465529.907944] [ndv4:14421:0] ud_iface.c:393 UCX DEBUG iface 0x16b9090: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment