Skip to content

Instantly share code, notes, and snippets.

@vanzod
Created April 20, 2022 15:06
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/3a8d04f14614d8a0914b0bbfb1ecafca to your computer and use it in GitHub Desktop.
Save vanzod/3a8d04f14614d8a0914b0bbfb1ecafca to your computer and use it in GitHub Desktop.
UCX 1.11.2 debug log for successful osu_scatter
This file has been truncated, but you can view the full file.
[1650465628.842779] [ndv4:54965:0] debug.c:1198 UCX DEBUG using signal stack 0x2b4b1d33f000 size 141824
[1650465628.842855] [ndv4:54965:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465628.842874] [ndv4:54965:0] 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 0x2b4b1d1ad000
[1650465628.842890] [ndv4:54965:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650465628.842899] [ndv4:54965:0] 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
[1650465628.842905] [ndv4:54965:0] module.c:253 UCX DEBUG loading modules for ucs
[1650465628.844926] [ndv4:54965:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465628.844945] [ndv4:54965:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650465628.844977] [ndv4:54965:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650465628.844980] [ndv4:54965:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650465628.844987] [ndv4:54965:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650465628.844994] [ndv4:54965:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650465628.844996] [ndv4:54965:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650465628.845001] [ndv4:54965:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650465628.845003] [ndv4:54965:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650465628.845006] [ndv4:54965:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650465628.845008] [ndv4:54965:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650465628.845010] [ndv4:54965:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650465628.845018] [ndv4:54965:0] module.c:253 UCX DEBUG loading modules for uct
[1650465628.845871] [ndv4:54965:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650465628.846460] [ndv4:54965:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650465628.846473] [ndv4:54965:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650465628.846485] [ndv4:54965:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650465628.846495] [ndv4:54965:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650465628.846506] [ndv4:54965:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650465628.846516] [ndv4:54965:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650465628.846526] [ndv4:54965:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650465628.846950] [ndv4:54965:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650465628.871052] [ndv4:54965:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650465628.871410] [ndv4:54965:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650465628.876869] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x13a1770 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650465628.876968] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650465628.876976] [ndv4:54965:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650465628.877132] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465628.877144] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465628.877170] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465628.877226] [ndv4:54965:0] module.c:253 UCX DEBUG loading modules for ucm
[1650465628.877248] [ndv4:54965:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650465628.877792] [ndv4:54965:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465628.877798] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465628.878043] [ndv4:54965:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465628.886764] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465628.886776] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465628.906117] [ndv4:54965:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650465628.906125] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465628.915312] [ndv4:54965:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465628.915319] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465628.916888] [ndv4:54965:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465628.916895] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465628.917160] [ndv4:54965:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465628.917164] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465628.926198] [ndv4:54965:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465628.926205] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465628.926420] [ndv4:54965:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465628.933862] [ndv4:54965:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650465628.934181] [ndv4:54965:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650465628.934712] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x1a84230 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650465628.934736] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650465628.934740] [ndv4:54965:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650465628.935038] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465628.935046] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465628.935051] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465628.935092] [ndv4:54965:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650465628.945862] [ndv4:54965:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465628.945870] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465628.946088] [ndv4:54965:0] ib_md.c:1604 UCX DEBUG mlx5_ib1: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465628.946180] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465628.946186] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465628.946684] [ndv4:54965:0] topo.c:99 UCX DEBUG bus id 0x102000000 doesn't exist. sys_dev = 1
[1650465628.946689] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465628.947385] [ndv4:54965:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465628.947389] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465628.947864] [ndv4:54965:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465628.947870] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465628.957359] [ndv4:54965:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465628.957366] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465628.978279] [ndv4:54965:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465628.978287] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465628.978495] [ndv4:54965:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465628.984708] [ndv4:54965:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib2 vendor_id: 0x15b3 device_id: 4124
[1650465628.985074] [ndv4:54965:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib2: disable ODP because it's not supported for DevX QP
[1650465628.986231] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x1a841c0 [id=61 ref 1] uct_ib_async_event_handler() to hash
[1650465628.986256] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 61 events 0x1 mode thread_spinlock
[1650465628.986259] [ndv4:54965:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib2' (InfiniBand channel adapter) with 1 ports
[1650465628.986504] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465628.986513] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465628.986518] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465628.986562] [ndv4:54965:0] ib_md.c:1319 UCX DEBUG mlx5_ib2: using registration cache
[1650465628.987418] [ndv4:54965:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465628.987423] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465628.987660] [ndv4:54965:0] ib_md.c:1604 UCX DEBUG mlx5_ib2: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465628.987747] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465628.987754] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.000031] [ndv4:54965:0] topo.c:99 UCX DEBUG bus id 0x103000000 doesn't exist. sys_dev = 2
[1650465629.000039] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465629.000114] [ndv4:54965:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465629.000117] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465629.001343] [ndv4:54965:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465629.001347] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465629.003045] [ndv4:54965:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465629.003050] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465629.017740] [ndv4:54965:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465629.017749] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465629.017960] [ndv4:54965:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.034393] [ndv4:54965:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib3 vendor_id: 0x15b3 device_id: 4124
[1650465629.034778] [ndv4:54965:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib3: disable ODP because it's not supported for DevX QP
[1650465629.035460] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x139aeb0 [id=63 ref 1] uct_ib_async_event_handler() to hash
[1650465629.035484] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 63 events 0x1 mode thread_spinlock
[1650465629.035487] [ndv4:54965:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib3' (InfiniBand channel adapter) with 1 ports
[1650465629.035678] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465629.035686] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465629.035691] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.035731] [ndv4:54965:0] ib_md.c:1319 UCX DEBUG mlx5_ib3: using registration cache
[1650465629.037434] [ndv4:54965:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.037439] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.037676] [ndv4:54965:0] ib_md.c:1604 UCX DEBUG mlx5_ib3: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.038147] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465629.038153] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465629.050057] [ndv4:54965:0] topo.c:99 UCX DEBUG bus id 0x104000000 doesn't exist. sys_dev = 3
[1650465629.050072] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465629.050784] [ndv4:54965:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465629.050789] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465629.061545] [ndv4:54756:0] debug.c:1198 UCX DEBUG using signal stack 0x2b3b50a32000 size 141824
[1650465629.061711] [ndv4:54756:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465629.061736] [ndv4:54756:0] 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 0x2b3b50886000
[1650465629.061754] [ndv4:54756:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650465629.061764] [ndv4:54756:0] 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
[1650465629.061770] [ndv4:54756:0] module.c:253 UCX DEBUG loading modules for ucs
[1650465629.061746] [ndv4:54965:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465629.061754] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465629.065140] [ndv4:54756:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465629.065160] [ndv4:54756:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650465629.065199] [ndv4:54756:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650465629.065202] [ndv4:54756:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650465629.065209] [ndv4:54756:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650465629.065217] [ndv4:54756:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650465629.065220] [ndv4:54756:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650465629.065225] [ndv4:54756:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650465629.065228] [ndv4:54756:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650465629.065230] [ndv4:54756:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650465629.065232] [ndv4:54756:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650465629.065234] [ndv4:54756:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650465629.065243] [ndv4:54756:0] module.c:253 UCX DEBUG loading modules for uct
[1650465629.066028] [ndv4:54756:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650465629.068450] [ndv4:55152:0] debug.c:1198 UCX DEBUG using signal stack 0x2b3b93d85000 size 141824
[1650465629.068526] [ndv4:55152:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465629.068545] [ndv4:55152:0] 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 0x2b3b93bf3000
[1650465629.068563] [ndv4:55152:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650465629.068683] [ndv4:55152:0] 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
[1650465629.068693] [ndv4:55152:0] module.c:253 UCX DEBUG loading modules for ucs
[1650465629.071186] [ndv4:55152:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465629.071204] [ndv4:55152:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650465629.071236] [ndv4:55152:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650465629.071240] [ndv4:55152:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650465629.071247] [ndv4:55152:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650465629.071253] [ndv4:55152:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650465629.071256] [ndv4:55152:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650465629.071261] [ndv4:55152:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650465629.071264] [ndv4:55152:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650465629.071266] [ndv4:55152:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650465629.071268] [ndv4:55152:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650465629.071270] [ndv4:55152:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650465629.071278] [ndv4:55152:0] module.c:253 UCX DEBUG loading modules for uct
[1650465629.071966] [ndv4:55152:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650465629.073920] [ndv4:54965:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465629.073928] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465629.075239] [ndv4:55152:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650465629.075255] [ndv4:55152:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650465629.075266] [ndv4:55152:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650465629.075277] [ndv4:55152:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650465629.075286] [ndv4:55152:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650465629.075297] [ndv4:55152:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650465629.075307] [ndv4:55152:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650465629.075753] [ndv4:55152:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650465629.074913] [ndv4:54756:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650465629.074929] [ndv4:54756:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650465629.074940] [ndv4:54756:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650465629.074950] [ndv4:54756:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650465629.074960] [ndv4:54756:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650465629.074970] [ndv4:54756:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650465629.074981] [ndv4:54756:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650465629.075645] [ndv4:54756:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650465629.077694] [ndv4:54965:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465629.077699] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465629.077984] [ndv4:54965:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.083669] [ndv4:54756:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650465629.083993] [ndv4:54756:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650465629.084750] [ndv4:54965:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib4 vendor_id: 0x15b3 device_id: 4124
[1650465629.085175] [ndv4:54965:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib4: disable ODP because it's not supported for DevX QP
[1650465629.090063] [ndv4:54756:0] async.c:228 UCX DEBUG added async handler 0x1e5f7f0 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650465629.090170] [ndv4:54756:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650465629.090178] [ndv4:54756:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650465629.090747] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.090758] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.090784] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.090844] [ndv4:54756:0] module.c:253 UCX DEBUG loading modules for ucm
[1650465629.090867] [ndv4:54756:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650465629.093041] [ndv4:54756:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.093049] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.093203] [ndv4:54756:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.093258] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.093265] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.094532] [ndv4:55152:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650465629.094917] [ndv4:55152:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650465629.099563] [ndv4:55152:0] async.c:228 UCX DEBUG added async handler 0x25102a0 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650465629.099777] [ndv4:55152:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650465629.099786] [ndv4:55152:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650465629.099964] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.099978] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.100004] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.100063] [ndv4:55152:0] module.c:253 UCX DEBUG loading modules for ucm
[1650465629.100085] [ndv4:55152:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650465629.113420] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x13a2090 [id=65 ref 1] uct_ib_async_event_handler() to hash
[1650465629.113451] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 65 events 0x1 mode thread_spinlock
[1650465629.113455] [ndv4:54965:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib4' (InfiniBand channel adapter) with 1 ports
[1650465629.112466] [ndv4:54756:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650465629.112475] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.113360] [ndv4:54756:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465629.113364] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.113752] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465629.113766] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465629.113775] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.113834] [ndv4:54965:0] ib_md.c:1319 UCX DEBUG mlx5_ib4: using registration cache
[1650465629.117339] [ndv4:54965:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.117347] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.117590] [ndv4:54965:0] ib_md.c:1604 UCX DEBUG mlx5_ib4: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.117686] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465629.117693] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465629.118790] [ndv4:54965:0] topo.c:99 UCX DEBUG bus id 0x105000000 doesn't exist. sys_dev = 4
[1650465629.118794] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465629.127865] [ndv4:55152:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.127876] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.128053] [ndv4:55152:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.128369] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.128376] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.131875] [ndv4:54756:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465629.131883] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.143769] [ndv4:54965:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465629.143776] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465629.144283] [ndv4:54965:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465629.144286] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465629.145399] [ndv4:55152:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650465629.145407] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.147500] [ndv4:54756:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465629.147508] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.152107] [ndv4:54756:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465629.152114] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.152379] [ndv4:54756:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.153670] [ndv4:54965:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465629.153677] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465629.154301] [ndv4:55152:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465629.154309] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.164640] [ndv4:55152:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465629.164647] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.168360] [ndv4:54756:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650465629.168696] [ndv4:54756:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650465629.169737] [ndv4:54756:0] async.c:228 UCX DEBUG added async handler 0x2540b40 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650465629.169764] [ndv4:54756:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650465629.169768] [ndv4:54756:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650465629.170094] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.170104] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.170114] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.170175] [ndv4:54756:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650465629.171549] [ndv4:54965:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465629.171557] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465629.171797] [ndv4:54965:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.172645] [ndv4:54756:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.172652] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.172790] [ndv4:54756:0] ib_md.c:1604 UCX DEBUG mlx5_ib1: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.172944] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.172950] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.174590] [ndv4:55152:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465629.174597] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.175991] [ndv4:55152:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465629.175997] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.176222] [ndv4:55152:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.184025] [ndv4:55152:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650465629.184404] [ndv4:55152:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650465629.183925] [ndv4:54756:0] topo.c:99 UCX DEBUG bus id 0x102000000 doesn't exist. sys_dev = 1
[1650465629.183934] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.186506] [ndv4:55152:0] async.c:228 UCX DEBUG added async handler 0x2536130 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650465629.186531] [ndv4:55152:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650465629.186534] [ndv4:55152:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650465629.186862] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.186871] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.186876] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.186917] [ndv4:55152:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650465629.187284] [ndv4:54965:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib5 vendor_id: 0x15b3 device_id: 4124
[1650465629.187733] [ndv4:54965:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib5: disable ODP because it's not supported for DevX QP
[1650465629.189194] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x139ab00 [id=67 ref 1] uct_ib_async_event_handler() to hash
[1650465629.189219] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 67 events 0x1 mode thread_spinlock
[1650465629.189223] [ndv4:54965:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib5' (InfiniBand channel adapter) with 1 ports
[1650465629.189492] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465629.189507] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465629.189518] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.189654] [ndv4:54965:0] ib_md.c:1319 UCX DEBUG mlx5_ib5: using registration cache
[1650465629.189514] [ndv4:55152:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.189522] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.189676] [ndv4:55152:0] ib_md.c:1604 UCX DEBUG mlx5_ib1: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.189981] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.189989] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.191762] [ndv4:54965:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.191770] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.192260] [ndv4:54965:0] ib_md.c:1604 UCX DEBUG mlx5_ib5: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.192318] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465629.192326] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465629.197904] [ndv4:54756:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465629.197911] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.202291] [ndv4:55152:0] topo.c:99 UCX DEBUG bus id 0x102000000 doesn't exist. sys_dev = 1
[1650465629.202300] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.206769] [ndv4:54756:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465629.206777] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.213673] [ndv4:55152:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465629.213680] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.214761] [ndv4:55152:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465629.214766] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.217321] [ndv4:55152:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465629.217327] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.216182] [ndv4:54756:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465629.216189] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.218392] [ndv4:54756:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465629.218398] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.218716] [ndv4:54756:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.219680] [ndv4:54965:0] topo.c:99 UCX DEBUG bus id 0x106000000 doesn't exist. sys_dev = 5
[1650465629.219690] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465629.224211] [ndv4:54756:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib2 vendor_id: 0x15b3 device_id: 4124
[1650465629.224568] [ndv4:54756:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib2: disable ODP because it's not supported for DevX QP
[1650465629.225194] [ndv4:54756:0] async.c:228 UCX DEBUG added async handler 0x1e5e970 [id=61 ref 1] uct_ib_async_event_handler() to hash
[1650465629.225220] [ndv4:54756:0] async.c:506 UCX DEBUG listening to async event fd 61 events 0x1 mode thread_spinlock
[1650465629.225224] [ndv4:54756:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib2' (InfiniBand channel adapter) with 1 ports
[1650465629.225503] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.225517] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.225527] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.225608] [ndv4:54756:0] ib_md.c:1319 UCX DEBUG mlx5_ib2: using registration cache
[1650465629.230023] [ndv4:55152:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465629.230031] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.230243] [ndv4:55152:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.235484] [ndv4:55152:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib2 vendor_id: 0x15b3 device_id: 4124
[1650465629.235791] [ndv4:55152:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib2: disable ODP because it's not supported for DevX QP
[1650465629.236088] [ndv4:55152:0] async.c:228 UCX DEBUG added async handler 0x2536c60 [id=61 ref 1] uct_ib_async_event_handler() to hash
[1650465629.236112] [ndv4:55152:0] async.c:506 UCX DEBUG listening to async event fd 61 events 0x1 mode thread_spinlock
[1650465629.236116] [ndv4:55152:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib2' (InfiniBand channel adapter) with 1 ports
[1650465629.236267] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.236276] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.236281] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.236328] [ndv4:55152:0] ib_md.c:1319 UCX DEBUG mlx5_ib2: using registration cache
[1650465629.238518] [ndv4:55152:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.238533] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.238715] [ndv4:55152:0] ib_md.c:1604 UCX DEBUG mlx5_ib2: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.238878] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.238884] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.242099] [ndv4:54756:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.242107] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.242245] [ndv4:54756:0] ib_md.c:1604 UCX DEBUG mlx5_ib2: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.242414] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.242422] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.243888] [ndv4:54756:0] topo.c:99 UCX DEBUG bus id 0x103000000 doesn't exist. sys_dev = 2
[1650465629.243895] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465629.247740] [ndv4:54965:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465629.247749] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465629.249244] [ndv4:54965:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465629.249250] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465629.254333] [ndv4:54756:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465629.254342] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465629.255218] [ndv4:54756:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465629.255223] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465629.255684] [ndv4:54756:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465629.255690] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465629.255988] [ndv4:54756:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465629.255992] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465629.256272] [ndv4:54756:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.259163] [ndv4:55152:0] topo.c:99 UCX DEBUG bus id 0x103000000 doesn't exist. sys_dev = 2
[1650465629.259171] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465629.260460] [ndv4:55152:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465629.260465] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465629.263204] [ndv4:54965:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465629.263212] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465629.273165] [ndv4:54756:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib3 vendor_id: 0x15b3 device_id: 4124
[1650465629.273450] [ndv4:54756:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib3: disable ODP because it's not supported for DevX QP
[1650465629.274790] [ndv4:54756:0] async.c:228 UCX DEBUG added async handler 0x1e5f2f0 [id=63 ref 1] uct_ib_async_event_handler() to hash
[1650465629.274814] [ndv4:54756:0] async.c:506 UCX DEBUG listening to async event fd 63 events 0x1 mode thread_spinlock
[1650465629.274817] [ndv4:54756:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib3' (InfiniBand channel adapter) with 1 ports
[1650465629.275089] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465629.275099] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465629.275104] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.275147] [ndv4:54756:0] ib_md.c:1319 UCX DEBUG mlx5_ib3: using registration cache
[1650465629.280135] [ndv4:55152:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465629.280143] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465629.283382] [ndv4:55152:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465629.283389] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465629.283031] [ndv4:54965:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465629.283039] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465629.283356] [ndv4:54965:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.285804] [ndv4:54756:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.285826] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.287425] [ndv4:54756:0] ib_md.c:1604 UCX DEBUG mlx5_ib3: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.287536] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465629.287548] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465629.289755] [ndv4:54965:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib6 vendor_id: 0x15b3 device_id: 4124
[1650465629.290097] [ndv4:54965:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib6: disable ODP because it's not supported for DevX QP
[1650465629.298882] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x139ce10 [id=69 ref 1] uct_ib_async_event_handler() to hash
[1650465629.298907] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 69 events 0x1 mode thread_spinlock
[1650465629.298910] [ndv4:54965:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib6' (InfiniBand channel adapter) with 1 ports
[1650465629.299154] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465629.299165] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465629.299170] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.299215] [ndv4:54965:0] ib_md.c:1319 UCX DEBUG mlx5_ib6: using registration cache
[1650465629.300013] [ndv4:54965:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.300018] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.300643] [ndv4:54965:0] ib_md.c:1604 UCX DEBUG mlx5_ib6: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.300759] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465629.300766] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465629.302412] [ndv4:55152:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465629.302419] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465629.302653] [ndv4:55152:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.309811] [ndv4:55516:0] debug.c:1198 UCX DEBUG using signal stack 0x2ae248e6a000 size 141824
[1650465629.309894] [ndv4:55516:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465629.309917] [ndv4:55516:0] 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 0x2ae248cbe000
[1650465629.309939] [ndv4:55516:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650465629.309948] [ndv4:55516:0] 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
[1650465629.309954] [ndv4:55516:0] module.c:253 UCX DEBUG loading modules for ucs
[1650465629.312273] [ndv4:55512:0] debug.c:1198 UCX DEBUG using signal stack 0x2aea6cc40000 size 141824
[1650465629.312349] [ndv4:55512:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465629.312371] [ndv4:55512:0] 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 0x2aea6ca94000
[1650465629.312389] [ndv4:55512:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650465629.312398] [ndv4:55512:0] 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
[1650465629.312404] [ndv4:55512:0] module.c:253 UCX DEBUG loading modules for ucs
[1650465629.312644] [ndv4:55516:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465629.312664] [ndv4:55516:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650465629.312704] [ndv4:55516:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650465629.312708] [ndv4:55516:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650465629.312716] [ndv4:55516:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650465629.312722] [ndv4:55516:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650465629.312725] [ndv4:55516:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650465629.312730] [ndv4:55516:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650465629.312732] [ndv4:55516:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650465629.312734] [ndv4:55516:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650465629.312737] [ndv4:55516:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650465629.312739] [ndv4:55516:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650465629.312747] [ndv4:55516:0] module.c:253 UCX DEBUG loading modules for uct
[1650465629.310936] [ndv4:54965:0] topo.c:99 UCX DEBUG bus id 0x107000000 doesn't exist. sys_dev = 6
[1650465629.310944] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465629.311368] [ndv4:54965:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465629.311372] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465629.311718] [ndv4:54965:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465629.311722] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465629.312855] [ndv4:54965:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465629.312860] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465629.313123] [ndv4:54965:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465629.313126] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465629.313340] [ndv4:54965:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.314760] [ndv4:55512:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465629.314778] [ndv4:55512:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650465629.314811] [ndv4:55512:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650465629.314815] [ndv4:55512:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650465629.314823] [ndv4:55512:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650465629.314830] [ndv4:55512:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650465629.314833] [ndv4:55512:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650465629.314839] [ndv4:55512:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650465629.314842] [ndv4:55512:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650465629.314844] [ndv4:55512:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650465629.314847] [ndv4:55512:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650465629.314849] [ndv4:55512:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650465629.314857] [ndv4:55512:0] module.c:253 UCX DEBUG loading modules for uct
[1650465629.315530] [ndv4:55512:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650465629.313572] [ndv4:55516:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650465629.314161] [ndv4:55516:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650465629.314175] [ndv4:55516:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650465629.314187] [ndv4:55516:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650465629.314198] [ndv4:55516:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650465629.314208] [ndv4:55516:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650465629.314218] [ndv4:55516:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650465629.314228] [ndv4:55516:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650465629.314698] [ndv4:55516:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650465629.314383] [ndv4:54756:0] topo.c:99 UCX DEBUG bus id 0x104000000 doesn't exist. sys_dev = 3
[1650465629.314393] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465629.314701] [ndv4:54756:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465629.314706] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465629.315527] [ndv4:54756:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465629.315531] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465629.317874] [ndv4:54756:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465629.317880] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465629.318854] [ndv4:55152:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib3 vendor_id: 0x15b3 device_id: 4124
[1650465629.319168] [ndv4:55152:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib3: disable ODP because it's not supported for DevX QP
[1650465629.320782] [ndv4:55152:0] async.c:228 UCX DEBUG added async handler 0x2536360 [id=63 ref 1] uct_ib_async_event_handler() to hash
[1650465629.320805] [ndv4:55152:0] async.c:506 UCX DEBUG listening to async event fd 63 events 0x1 mode thread_spinlock
[1650465629.320808] [ndv4:55152:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib3' (InfiniBand channel adapter) with 1 ports
[1650465629.321279] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465629.321288] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465629.321293] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.321336] [ndv4:55152:0] ib_md.c:1319 UCX DEBUG mlx5_ib3: using registration cache
[1650465629.323124] [ndv4:55516:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650465629.323422] [ndv4:55516:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650465629.323528] [ndv4:55512:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650465629.323546] [ndv4:55512:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650465629.323557] [ndv4:55512:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650465629.323567] [ndv4:55512:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650465629.323617] [ndv4:55512:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650465629.323628] [ndv4:55512:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650465629.323639] [ndv4:55512:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650465629.324007] [ndv4:55512:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650465629.324434] [ndv4:54965:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib7 vendor_id: 0x15b3 device_id: 4124
[1650465629.324796] [ndv4:54965:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib7: disable ODP because it's not supported for DevX QP
[1650465629.326239] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x139af10 [id=71 ref 1] uct_ib_async_event_handler() to hash
[1650465629.326262] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 71 events 0x1 mode thread_spinlock
[1650465629.326265] [ndv4:54965:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib7' (InfiniBand channel adapter) with 1 ports
[1650465629.326556] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465629.326566] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465629.326571] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.326642] [ndv4:54965:0] ib_md.c:1319 UCX DEBUG mlx5_ib7: using registration cache
[1650465629.327840] [ndv4:54756:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465629.327850] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465629.328215] [ndv4:54756:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.328712] [ndv4:54965:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.328720] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.328985] [ndv4:54965:0] ib_md.c:1604 UCX DEBUG mlx5_ib7: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.329023] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465629.329028] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465629.329405] [ndv4:54965:0] topo.c:99 UCX DEBUG bus id 0x108000000 doesn't exist. sys_dev = 7
[1650465629.329409] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465629.329555] [ndv4:54965:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465629.329559] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465629.329687] [ndv4:55512:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650465629.329993] [ndv4:55512:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650465629.330214] [ndv4:55516:0] async.c:228 UCX DEBUG added async handler 0x12771f0 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650465629.330311] [ndv4:55516:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650465629.330319] [ndv4:55516:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650465629.330496] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.330506] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.330532] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.330615] [ndv4:55516:0] module.c:253 UCX DEBUG loading modules for ucm
[1650465629.330638] [ndv4:55516:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650465629.331815] [ndv4:55516:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.331823] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.331975] [ndv4:55516:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.332035] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.332041] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.332789] [ndv4:55516:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650465629.332794] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.333475] [ndv4:55516:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465629.333480] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.333258] [ndv4:55152:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.333265] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.333417] [ndv4:55152:0] ib_md.c:1604 UCX DEBUG mlx5_ib3: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.333473] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465629.333480] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465629.333973] [ndv4:55152:0] topo.c:99 UCX DEBUG bus id 0x104000000 doesn't exist. sys_dev = 3
[1650465629.333978] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465629.334285] [ndv4:55152:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465629.334289] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465629.334436] [ndv4:55152:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465629.334440] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465629.334554] [ndv4:55152:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465629.334557] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465629.335022] [ndv4:55152:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465629.335027] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465629.335237] [ndv4:55152:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.334364] [ndv4:55512:0] async.c:228 UCX DEBUG added async handler 0x128d1f0 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650465629.334463] [ndv4:55512:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650465629.334470] [ndv4:55512:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650465629.334620] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.334628] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.334651] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.334707] [ndv4:55512:0] module.c:253 UCX DEBUG loading modules for ucm
[1650465629.334729] [ndv4:55512:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650465629.335310] [ndv4:55512:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.335317] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.335477] [ndv4:55512:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.335501] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.335508] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.338843] [ndv4:54756:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib4 vendor_id: 0x15b3 device_id: 4124
[1650465629.339146] [ndv4:54756:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib4: disable ODP because it's not supported for DevX QP
[1650465629.341725] [ndv4:55152:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib4 vendor_id: 0x15b3 device_id: 4124
[1650465629.342047] [ndv4:55152:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib4: disable ODP because it's not supported for DevX QP
[1650465629.342976] [ndv4:54965:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465629.342983] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465629.342972] [ndv4:55152:0] async.c:228 UCX DEBUG added async handler 0x253af90 [id=65 ref 1] uct_ib_async_event_handler() to hash
[1650465629.342994] [ndv4:55152:0] async.c:506 UCX DEBUG listening to async event fd 65 events 0x1 mode thread_spinlock
[1650465629.342997] [ndv4:55152:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib4' (InfiniBand channel adapter) with 1 ports
[1650465629.343339] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465629.343347] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465629.343353] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.343392] [ndv4:55152:0] ib_md.c:1319 UCX DEBUG mlx5_ib4: using registration cache
[1650465629.344771] [ndv4:55152:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.344777] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.344936] [ndv4:55152:0] ib_md.c:1604 UCX DEBUG mlx5_ib4: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.345201] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465629.345207] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465629.344729] [ndv4:55516:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465629.344737] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.347206] [ndv4:55512:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650465629.347215] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.348357] [ndv4:54756:0] async.c:228 UCX DEBUG added async handler 0x1e58d20 [id=65 ref 1] uct_ib_async_event_handler() to hash
[1650465629.348381] [ndv4:54756:0] async.c:506 UCX DEBUG listening to async event fd 65 events 0x1 mode thread_spinlock
[1650465629.348385] [ndv4:54756:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib4' (InfiniBand channel adapter) with 1 ports
[1650465629.348767] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465629.348777] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465629.348783] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.348829] [ndv4:54756:0] ib_md.c:1319 UCX DEBUG mlx5_ib4: using registration cache
[1650465629.350324] [ndv4:54756:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.350329] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.350462] [ndv4:54756:0] ib_md.c:1604 UCX DEBUG mlx5_ib4: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.350682] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465629.350688] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465629.350547] [ndv4:55512:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465629.350558] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.354037] [ndv4:54965:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465629.354044] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465629.356157] [ndv4:55152:0] topo.c:99 UCX DEBUG bus id 0x105000000 doesn't exist. sys_dev = 4
[1650465629.356165] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465629.355662] [ndv4:55516:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465629.355670] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.356282] [ndv4:55516:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465629.356286] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.356505] [ndv4:55516:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.356025] [ndv4:54965:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465629.356032] [ndv4:54965:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465629.356126] [ndv4:54965:0] ucp_context.c:1556 UCX DEBUG created ucp context 0x1385b10 0x1385b10 [13 mds 47 tls] features 0x1 tl bitmap 0x7fffffffffff 0x0
[1650465629.360922] [ndv4:55512:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465629.360931] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.360742] [ndv4:54756:0] topo.c:99 UCX DEBUG bus id 0x105000000 doesn't exist. sys_dev = 4
[1650465629.360751] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465629.363512] [ndv4:54756:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465629.363518] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465629.362643] [ndv4:55512:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465629.362649] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.364730] [ndv4:55516:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650465629.365054] [ndv4:55516:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650465629.364878] [ndv4:55512:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465629.364884] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.365115] [ndv4:55512:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.366114] [ndv4:55516:0] async.c:228 UCX DEBUG added async handler 0x126ff60 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650465629.366137] [ndv4:55516:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650465629.366140] [ndv4:55516:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650465629.366390] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.366400] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.366405] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.366450] [ndv4:55516:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650465629.367662] [ndv4:55152:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465629.367670] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465629.368660] [ndv4:55516:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.368668] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.368801] [ndv4:55516:0] ib_md.c:1604 UCX DEBUG mlx5_ib1: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.368954] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.368960] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.371045] [ndv4:55516:0] topo.c:99 UCX DEBUG bus id 0x102000000 doesn't exist. sys_dev = 1
[1650465629.371053] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.372795] [ndv4:55512:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650465629.372914] [ndv4:55516:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465629.372921] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.374278] [ndv4:55516:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465629.374283] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.373106] [ndv4:55512:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650465629.373966] [ndv4:55512:0] async.c:228 UCX DEBUG added async handler 0x1285f60 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650465629.373987] [ndv4:55512:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650465629.373991] [ndv4:55512:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650465629.374179] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.374188] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.374193] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.374233] [ndv4:55512:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650465629.375792] [ndv4:55512:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.375798] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.375955] [ndv4:55512:0] ib_md.c:1604 UCX DEBUG mlx5_ib1: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.376142] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.376149] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.379836] [ndv4:55152:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465629.379844] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465629.381304] [ndv4:55152:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465629.381310] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465629.384243] [ndv4:54756:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465629.384251] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465629.385599] [ndv4:55516:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465629.385607] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.392869] [ndv4:55152:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465629.392877] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465629.393091] [ndv4:55152:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.399992] [ndv4:55516:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465629.400001] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.400224] [ndv4:55516:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.398544] [ndv4:54756:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465629.398551] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465629.400524] [ndv4:54756:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465629.400532] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465629.400854] [ndv4:54756:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.400147] [ndv4:55512:0] topo.c:99 UCX DEBUG bus id 0x102000000 doesn't exist. sys_dev = 1
[1650465629.400155] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.400958] [ndv4:55512:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465629.400963] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.401163] [ndv4:55512:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465629.401166] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.406677] [ndv4:54756:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib5 vendor_id: 0x15b3 device_id: 4124
[1650465629.406972] [ndv4:54756:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib5: disable ODP because it's not supported for DevX QP
[1650465629.407443] [ndv4:54756:0] async.c:228 UCX DEBUG added async handler 0x1e5adf0 [id=67 ref 1] uct_ib_async_event_handler() to hash
[1650465629.407466] [ndv4:54756:0] async.c:506 UCX DEBUG listening to async event fd 67 events 0x1 mode thread_spinlock
[1650465629.407469] [ndv4:54756:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib5' (InfiniBand channel adapter) with 1 ports
[1650465629.412262] [ndv4:55512:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465629.412268] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.415353] [ndv4:55512:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465629.415359] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.415568] [ndv4:55512:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.416911] [ndv4:55516:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib2 vendor_id: 0x15b3 device_id: 4124
[1650465629.417198] [ndv4:55516:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib2: disable ODP because it's not supported for DevX QP
[1650465629.417454] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465629.417466] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465629.417471] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.417519] [ndv4:54756:0] ib_md.c:1319 UCX DEBUG mlx5_ib5: using registration cache
[1650465629.418651] [ndv4:54756:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.418657] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.418835] [ndv4:54756:0] ib_md.c:1604 UCX DEBUG mlx5_ib5: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.419020] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465629.419027] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465629.417880] [ndv4:55516:0] async.c:228 UCX DEBUG added async handler 0x1277cf0 [id=61 ref 1] uct_ib_async_event_handler() to hash
[1650465629.417928] [ndv4:55516:0] async.c:506 UCX DEBUG listening to async event fd 61 events 0x1 mode thread_spinlock
[1650465629.417932] [ndv4:55516:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib2' (InfiniBand channel adapter) with 1 ports
[1650465629.418068] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.418076] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.418081] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.418152] [ndv4:55516:0] ib_md.c:1319 UCX DEBUG mlx5_ib2: using registration cache
[1650465629.419677] [ndv4:55516:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.419683] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.419816] [ndv4:55516:0] ib_md.c:1604 UCX DEBUG mlx5_ib2: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.420235] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.420242] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.419998] [ndv4:55152:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib5 vendor_id: 0x15b3 device_id: 4124
[1650465629.420335] [ndv4:55152:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib5: disable ODP because it's not supported for DevX QP
[1650465629.421262] [ndv4:54756:0] topo.c:99 UCX DEBUG bus id 0x106000000 doesn't exist. sys_dev = 5
[1650465629.421268] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465629.422223] [ndv4:55152:0] async.c:228 UCX DEBUG added async handler 0x253cf40 [id=67 ref 1] uct_ib_async_event_handler() to hash
[1650465629.422247] [ndv4:55152:0] async.c:506 UCX DEBUG listening to async event fd 67 events 0x1 mode thread_spinlock
[1650465629.422250] [ndv4:55152:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib5' (InfiniBand channel adapter) with 1 ports
[1650465629.422569] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465629.422614] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465629.422619] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.422663] [ndv4:55152:0] ib_md.c:1319 UCX DEBUG mlx5_ib5: using registration cache
[1650465629.423827] [ndv4:55152:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.423834] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.423986] [ndv4:55152:0] ib_md.c:1604 UCX DEBUG mlx5_ib5: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.424098] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465629.424104] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465629.426091] [ndv4:54965:0] mm_posix.c:518 UCX DEBUG allocated posix shared memory at 0x2b4b1ff61000 length 12288
[1650465629.426745] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool mm_recv_desc: align 64, maxelems 4294967295, elemsize 8368
[1650465629.427806] [ndv4:54965:0] mm_posix.c:305 UCX DEBUG shared memory mmap(addr=(nil), length=6291456, flags= HUGETLB, fd=76) failed: Invalid argument
[1650465629.427815] [ndv4:54965:0] mm_posix.c:518 UCX DEBUG allocated posix shared memory at 0x2b4b246c5000 length 4296704
[1650465629.427819] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool mm_recv_desc: allocated chunk 0x2b4b246c5018 of 4296680 bytes with 512 elements
[1650465629.428085] [ndv4:54965:0] mm_iface.c:600 UCX DEBUG created mm iface 0x1b02d20 FIFO id 0x4000000036dc426d va 0x2b4b1ff61000 size 12288 (128 x 64 elems)
[1650465629.428132] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[0]=0x1b02d20 using posix/memory on worker 0x23dc840
[1650465629.428157] [ndv4:54965:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 8447 bytes with hugetlb
[1650465629.428193] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool mm_recv_desc: align 64, maxelems 4294967295, elemsize 8368
[1650465629.429135] [ndv4:54965:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 4292720 bytes with hugetlb
[1650465629.429148] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool mm_recv_desc: allocated chunk 0x2b4b24ade018 of 4296680 bytes with 512 elements
[1650465629.429798] [ndv4:54965:0] mm_iface.c:600 UCX DEBUG created mm iface 0x1b032f0 FIFO id 0x718004 va 0x2b4b1ff64000 size 12288 (128 x 64 elems)
[1650465629.429806] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[1]=0x1b032f0 using sysv/memory on worker 0x23dc840
[1650465629.429818] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool self_msg_desc: align 64, maxelems 4294967295, elemsize 8200
[1650465629.429821] [ndv4:54965:0] self.c:220 UCX DEBUG created self iface id 0xc4ee86fb0bd28fe8 send_size 8192
[1650465629.429827] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[2]=0x1b16ed0 using self/memory0 on worker 0x23dc840
[1650465629.429851] [ndv4:54965:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465629.429856] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465629.429858] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465629.432614] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x1b24ef0 [id=78 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465629.432643] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 78 events 0x5 mode thread_spinlock
[1650465629.432662] [ndv4:54965:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x1b17830: listening for connections (fd=78) on 10.5.0.5:52109
[1650465629.432785] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[3]=0x1b17830 using tcp/eth0 on worker 0x23dc840
[1650465629.432805] [ndv4:54965:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465629.432808] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465629.432811] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465629.432847] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x1a732b0 [id=80 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465629.432867] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 80 events 0x5 mode thread_spinlock
[1650465629.432870] [ndv4:54965:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x1b17eb0: listening for connections (fd=80) on 127.0.0.1:46545
[1650465629.432887] [ndv4:54965:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465629.432893] [ndv4:54965:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465629.432950] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[4]=0x1b17eb0 using tcp/lo on worker 0x23dc840
[1650465629.432966] [ndv4:54965:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465629.432968] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465629.432971] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465629.433010] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x1a72fe0 [id=82 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465629.433029] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 82 events 0x5 mode thread_spinlock
[1650465629.433034] [ndv4:54965:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x1aff540: listening for connections (fd=82) on 172.16.1.242:34126
[1650465629.430967] [ndv4:55516:0] topo.c:99 UCX DEBUG bus id 0x103000000 doesn't exist. sys_dev = 2
[1650465629.430976] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465629.433239] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[5]=0x1aff540 using tcp/ib0 on worker 0x23dc840
[1650465629.433389] [ndv4:55512:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib2 vendor_id: 0x15b3 device_id: 4124
[1650465629.433742] [ndv4:55512:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib2: disable ODP because it's not supported for DevX QP
[1650465629.434285] [ndv4:55512:0] async.c:228 UCX DEBUG added async handler 0x128dcf0 [id=61 ref 1] uct_ib_async_event_handler() to hash
[1650465629.434308] [ndv4:55512:0] async.c:506 UCX DEBUG listening to async event fd 61 events 0x1 mode thread_spinlock
[1650465629.434311] [ndv4:55512:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib2' (InfiniBand channel adapter) with 1 ports
[1650465629.434447] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.434456] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.434461] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.434501] [ndv4:55512:0] ib_md.c:1319 UCX DEBUG mlx5_ib2: using registration cache
[1650465629.435063] [ndv4:55512:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.435068] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.435217] [ndv4:55512:0] ib_md.c:1604 UCX DEBUG mlx5_ib2: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.435244] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.435249] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.435980] [ndv4:55152:0] topo.c:99 UCX DEBUG bus id 0x106000000 doesn't exist. sys_dev = 5
[1650465629.435988] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465629.440058] [ndv4:54756:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465629.440066] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465629.440965] [ndv4:54756:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465629.440970] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465629.440501] [ndv4:55516:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465629.440510] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465629.441310] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.441321] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.444628] [ndv4:55152:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465629.444635] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465629.443854] [ndv4:55512:0] topo.c:99 UCX DEBUG bus id 0x103000000 doesn't exist. sys_dev = 2
[1650465629.443861] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465629.450697] [ndv4:55516:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465629.450705] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465629.458273] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.458285] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.459304] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465629.459931] [ndv4:55516:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465629.459942] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465629.459128] [ndv4:54756:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465629.459137] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465629.460569] [ndv4:55512:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465629.460612] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465629.461183] [ndv4:55152:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465629.461190] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465629.462152] [ndv4:55516:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465629.462160] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465629.462387] [ndv4:55516:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.467675] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465629.467714] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.467718] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.467731] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.468673] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465629.468682] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465629.469217] [ndv4:54965:0] ib_iface.c:994 UCX DEBUG iface=0x1b0d680: created RC QP 0xf475 on mlx5_ib0:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465629.470360] [ndv4:54756:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465629.470369] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465629.470659] [ndv4:54756:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.471295] [ndv4:55152:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465629.471302] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465629.472384] [ndv4:55152:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465629.472388] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465629.472672] [ndv4:55152:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.478271] [ndv4:55152:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib6 vendor_id: 0x15b3 device_id: 4124
[1650465629.478630] [ndv4:55152:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib6: disable ODP because it's not supported for DevX QP
[1650465629.479073] [ndv4:55152:0] async.c:228 UCX DEBUG added async handler 0x2536cb0 [id=69 ref 1] uct_ib_async_event_handler() to hash
[1650465629.479095] [ndv4:55152:0] async.c:506 UCX DEBUG listening to async event fd 69 events 0x1 mode thread_spinlock
[1650465629.479098] [ndv4:55152:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib6' (InfiniBand channel adapter) with 1 ports
[1650465629.479252] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465629.479262] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465629.479267] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.479307] [ndv4:55152:0] ib_md.c:1319 UCX DEBUG mlx5_ib6: using registration cache
[1650465629.482039] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[6]=0x1b0d680 using rc_verbs/mlx5_ib0:1 on worker 0x23dc840
[1650465629.482132] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.482138] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.482308] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.482313] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.482503] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465629.483013] [ndv4:54965:0] ib_device.c:1394 UCX DEBUG max IB CQE size is 128
[1650465629.481708] [ndv4:55512:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465629.481716] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465629.483275] [ndv4:55512:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465629.483279] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465629.484034] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465629.484043] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.484046] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.484098] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.484554] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x24a6010 of 8176 bytes with 127 elements
[1650465629.484834] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465629.484858] [ndv4:54965:0] 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> }
[1650465629.484909] [ndv4:54965:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650465629.484914] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465629.489527] [ndv4:55516:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib3 vendor_id: 0x15b3 device_id: 4124
[1650465629.489826] [ndv4:55516:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib3: disable ODP because it's not supported for DevX QP
[1650465629.490225] [ndv4:55152:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.490233] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.490398] [ndv4:55152:0] ib_md.c:1604 UCX DEBUG mlx5_ib6: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.490828] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465629.490835] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465629.490705] [ndv4:55516:0] async.c:228 UCX DEBUG added async handler 0x126f2c0 [id=63 ref 1] uct_ib_async_event_handler() to hash
[1650465629.490728] [ndv4:55516:0] async.c:506 UCX DEBUG listening to async event fd 63 events 0x1 mode thread_spinlock
[1650465629.490731] [ndv4:55516:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib3' (InfiniBand channel adapter) with 1 ports
[1650465629.491006] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465629.491015] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465629.491020] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.491060] [ndv4:55516:0] ib_md.c:1319 UCX DEBUG mlx5_ib3: using registration cache
[1650465629.491923] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x1a64060 [id=85 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465629.491950] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 85 events 0x1 mode thread_spinlock
[1650465629.491968] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[7]=0x1b1b0d0 using rc_mlx5/mlx5_ib0:1 on worker 0x23dc840
[1650465629.492371] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.492379] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.492645] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.492651] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.493440] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465629.495318] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465629.495327] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.495330] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.495344] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.495661] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465629.495669] [ndv4:54965:0] 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> }
[1650465629.495700] [ndv4:54965:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650465629.495704] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465629.495712] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x1395710 [id=87 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465629.495730] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 87 events 0x1 mode thread_spinlock
[1650465629.496339] [ndv4:54965:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465629.496728] [ndv4:54756:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib6 vendor_id: 0x15b3 device_id: 4124
[1650465629.497079] [ndv4:54756:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib6: disable ODP because it's not supported for DevX QP
[1650465629.497652] [ndv4:54756:0] async.c:228 UCX DEBUG added async handler 0x1e5e9c0 [id=69 ref 1] uct_ib_async_event_handler() to hash
[1650465629.497686] [ndv4:54756:0] async.c:506 UCX DEBUG listening to async event fd 69 events 0x1 mode thread_spinlock
[1650465629.497690] [ndv4:54756:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib6' (InfiniBand channel adapter) with 1 ports
[1650465629.497824] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465629.497837] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465629.497845] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.497902] [ndv4:54756:0] ib_md.c:1319 UCX DEBUG mlx5_ib6: using registration cache
[1650465629.498924] [ndv4:54756:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.498930] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.499072] [ndv4:54756:0] ib_md.c:1604 UCX DEBUG mlx5_ib6: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.499109] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465629.499117] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465629.501569] [ndv4:55516:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.501645] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.501794] [ndv4:55516:0] ib_md.c:1604 UCX DEBUG mlx5_ib3: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.502084] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465629.502091] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465629.502917] [ndv4:54965:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x2710010: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xf498
[1650465629.503438] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.503445] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.503465] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b4b1ff69008 of 151544 bytes with 1052 elements
[1650465629.502643] [ndv4:55512:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465629.502651] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465629.502862] [ndv4:55512:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.505819] [ndv4:55516:0] topo.c:99 UCX DEBUG bus id 0x104000000 doesn't exist. sys_dev = 3
[1650465629.505826] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465629.507161] [ndv4:54965:0] ib_md.c:812 UCX DEBUG registered memory 0x2b4b25000000..0x2b4b27600000 on mlx5_ib0 lkey 0x80600 rkey 0x80600 access 0xf flags 0x3e4
[1650465629.507180] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b4b25000018 of 39845864 bytes with 4752 elements
[1650465629.507318] [ndv4:54965:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x2710010
[1650465629.507352] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[8]=0x2710010 using dc_mlx5/mlx5_ib0:1 on worker 0x23dc840
[1650465629.507412] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.507422] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.507490] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.507496] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.507822] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465629.508592] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465629.508939] [ndv4:54965:0] ib_iface.c:994 UCX DEBUG iface=0x23fc6c0: created UD QP 0xf47e on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465629.509449] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465629.509763] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.509771] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.510049] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.510054] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.509654] [ndv4:55512:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib3 vendor_id: 0x15b3 device_id: 4124
[1650465629.509958] [ndv4:55512:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib3: disable ODP because it's not supported for DevX QP
[1650465629.510195] [ndv4:55512:0] async.c:228 UCX DEBUG added async handler 0x12852c0 [id=63 ref 1] uct_ib_async_event_handler() to hash
[1650465629.510219] [ndv4:55512:0] async.c:506 UCX DEBUG listening to async event fd 63 events 0x1 mode thread_spinlock
[1650465629.510222] [ndv4:55512:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib3' (InfiniBand channel adapter) with 1 ports
[1650465629.510392] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465629.510401] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465629.510406] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.510445] [ndv4:55512:0] ib_md.c:1319 UCX DEBUG mlx5_ib3: using registration cache
[1650465629.511786] [ndv4:55512:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.511793] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.510481] [ndv4:54965:0] ib_md.c:812 UCX DEBUG registered memory 0x2b4b276f8000..0x2b4b2777d000 on mlx5_ib0 lkey 0x80700 rkey 0x80700 access 0xf flags 0x3e4
[1650465629.510488] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b4b276f8018 of 544744 bytes with 128 elements
[1650465629.510493] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465629.511193] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x23fc6c0: adding gid fe80::15:5dff:fd33:fffb to hash on device mlx5_ib0 port 1 index 0)
[1650465629.511223] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x23fc6c0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650465629.511566] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x23fc6c0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650465629.511653] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x23fc6c0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650465629.511672] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x23fc6c0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650465629.511692] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x23fc6c0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650465629.511737] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x23fc6c0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650465629.511963] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x23fc6c0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650465629.512241] [ndv4:54965:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465629.511940] [ndv4:55512:0] ib_md.c:1604 UCX DEBUG mlx5_ib3: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.512021] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465629.512026] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465629.512329] [ndv4:55512:0] topo.c:99 UCX DEBUG bus id 0x104000000 doesn't exist. sys_dev = 3
[1650465629.512333] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465629.512870] [ndv4:55512:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465629.512874] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465629.513205] [ndv4:55512:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465629.513208] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465629.511719] [ndv4:55152:0] topo.c:99 UCX DEBUG bus id 0x107000000 doesn't exist. sys_dev = 6
[1650465629.511728] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465629.511828] [ndv4:55152:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465629.511832] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465629.513972] [ndv4:55512:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465629.513979] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465629.514838] [ndv4:55512:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465629.514842] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465629.515080] [ndv4:55512:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.517860] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x1a73a90 [id=88 ref 1] uct_ud_iface_async_handler() to hash
[1650465629.517894] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 88 events 0x5 mode thread_spinlock
[1650465629.517915] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[9]=0x23fc6c0 using ud_verbs/mlx5_ib0:1 on worker 0x23dc840
[1650465629.518048] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.518056] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.518153] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.518157] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.518971] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465629.519655] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465629.519911] [ndv4:54965:0] ib_iface.c:994 UCX DEBUG iface=0x1b252d0: created UD QP 0xf47f on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465629.519922] [ndv4:54965:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465629.519152] [ndv4:55516:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465629.519162] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465629.520412] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465629.523329] [ndv4:55152:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465629.523339] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465629.529562] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.529570] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.529735] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.529740] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.530328] [ndv4:55516:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465629.530337] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465629.530086] [ndv4:54965:0] ib_md.c:812 UCX DEBUG registered memory 0x2b4b2777d000..0x2b4b27802000 on mlx5_ib0 lkey 0x80800 rkey 0x80800 access 0xf flags 0x3e4
[1650465629.530092] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b4b2777d018 of 544744 bytes with 128 elements
[1650465629.530097] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465629.530811] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x1b252d0: adding gid fe80::15:5dff:fd33:fffb to hash on device mlx5_ib0 port 1 index 0)
[1650465629.531262] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x1b252d0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650465629.531240] [ndv4:55512:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib4 vendor_id: 0x15b3 device_id: 4124
[1650465629.531560] [ndv4:55512:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib4: disable ODP because it's not supported for DevX QP
[1650465629.533205] [ndv4:55512:0] async.c:228 UCX DEBUG added async handler 0x12877e0 [id=65 ref 1] uct_ib_async_event_handler() to hash
[1650465629.533230] [ndv4:55512:0] async.c:506 UCX DEBUG listening to async event fd 65 events 0x1 mode thread_spinlock
[1650465629.533233] [ndv4:55512:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib4' (InfiniBand channel adapter) with 1 ports
[1650465629.532066] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x1b252d0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650465629.533167] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x1b252d0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650465629.533742] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465629.533752] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465629.533757] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.533794] [ndv4:55512:0] ib_md.c:1319 UCX DEBUG mlx5_ib4: using registration cache
[1650465629.536093] [ndv4:55512:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.536101] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.536252] [ndv4:55512:0] ib_md.c:1604 UCX DEBUG mlx5_ib4: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.536447] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465629.536453] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465629.535708] [ndv4:55152:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465629.535717] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465629.539078] [ndv4:54756:0] topo.c:99 UCX DEBUG bus id 0x107000000 doesn't exist. sys_dev = 6
[1650465629.539088] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465629.541511] [ndv4:55516:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465629.541520] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465629.541640] [ndv4:55516:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465629.541643] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465629.541854] [ndv4:55516:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.543483] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x1b252d0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650465629.547005] [ndv4:55512:0] topo.c:99 UCX DEBUG bus id 0x105000000 doesn't exist. sys_dev = 4
[1650465629.547012] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465629.548716] [ndv4:55512:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465629.548723] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465629.548628] [ndv4:54756:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465629.548638] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465629.551958] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x1b252d0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650465629.554962] [ndv4:55152:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465629.554971] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465629.555184] [ndv4:55152:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.557985] [ndv4:55516:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib4 vendor_id: 0x15b3 device_id: 4124
[1650465629.558266] [ndv4:55516:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib4: disable ODP because it's not supported for DevX QP
[1650465629.559386] [ndv4:55512:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465629.559393] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465629.559547] [ndv4:55512:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465629.559550] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465629.559319] [ndv4:55516:0] async.c:228 UCX DEBUG added async handler 0x12717e0 [id=65 ref 1] uct_ib_async_event_handler() to hash
[1650465629.559344] [ndv4:55516:0] async.c:506 UCX DEBUG listening to async event fd 65 events 0x1 mode thread_spinlock
[1650465629.559347] [ndv4:55516:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib4' (InfiniBand channel adapter) with 1 ports
[1650465629.559457] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465629.559464] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465629.559469] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.559510] [ndv4:55516:0] ib_md.c:1319 UCX DEBUG mlx5_ib4: using registration cache
[1650465629.560990] [ndv4:54756:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465629.560998] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465629.560241] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x1b252d0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650465629.561009] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x1b252d0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650465629.561015] [ndv4:54965:0] 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> }
[1650465629.561046] [ndv4:54965:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465629.561050] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x1b02c00 [id=89 ref 1] uct_ud_iface_async_handler() to hash
[1650465629.561070] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 89 events 0x5 mode thread_spinlock
[1650465629.561083] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[10]=0x1b252d0 using ud_mlx5/mlx5_ib0:1 on worker 0x23dc840
[1650465629.561233] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.561239] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.569474] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.569491] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.570422] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465629.570418] [ndv4:55516:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.570426] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.570558] [ndv4:55516:0] ib_md.c:1604 UCX DEBUG mlx5_ib4: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.570838] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465629.570845] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465629.571860] [ndv4:55152:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib7 vendor_id: 0x15b3 device_id: 4124
[1650465629.572207] [ndv4:55152:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib7: disable ODP because it's not supported for DevX QP
[1650465629.571854] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465629.571863] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.571865] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.571879] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.571777] [ndv4:54756:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465629.571785] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465629.572938] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465629.572944] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465629.573342] [ndv4:54965:0] ib_iface.c:994 UCX DEBUG iface=0x2cad050: created RC QP 0xda68 on mlx5_ib1:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465629.573673] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[11]=0x2cad050 using rc_verbs/mlx5_ib1:1 on worker 0x23dc840
[1650465629.579071] [ndv4:55512:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465629.579079] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465629.579291] [ndv4:55512:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.580977] [ndv4:55152:0] async.c:228 UCX DEBUG added async handler 0x25387e0 [id=71 ref 1] uct_ib_async_event_handler() to hash
[1650465629.581001] [ndv4:55152:0] async.c:506 UCX DEBUG listening to async event fd 71 events 0x1 mode thread_spinlock
[1650465629.581005] [ndv4:55152:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib7' (InfiniBand channel adapter) with 1 ports
[1650465629.581260] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465629.581270] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465629.581275] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.581317] [ndv4:55152:0] ib_md.c:1319 UCX DEBUG mlx5_ib7: using registration cache
[1650465629.582090] [ndv4:54756:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465629.582097] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465629.582382] [ndv4:54756:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.583677] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.583688] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.583827] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.583832] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.588997] [ndv4:55512:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib5 vendor_id: 0x15b3 device_id: 4124
[1650465629.589322] [ndv4:55512:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib5: disable ODP because it's not supported for DevX QP
[1650465629.590196] [ndv4:54756:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib7 vendor_id: 0x15b3 device_id: 4124
[1650465629.590509] [ndv4:54756:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib7: disable ODP because it's not supported for DevX QP
[1650465629.589850] [ndv4:55516:0] topo.c:99 UCX DEBUG bus id 0x105000000 doesn't exist. sys_dev = 4
[1650465629.589859] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465629.592048] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465629.592123] [ndv4:55512:0] async.c:228 UCX DEBUG added async handler 0x196ebf0 [id=67 ref 1] uct_ib_async_event_handler() to hash
[1650465629.592154] [ndv4:55512:0] async.c:506 UCX DEBUG listening to async event fd 67 events 0x1 mode thread_spinlock
[1650465629.592158] [ndv4:55512:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib5' (InfiniBand channel adapter) with 1 ports
[1650465629.592658] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465629.592666] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465629.592671] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.592714] [ndv4:55512:0] ib_md.c:1319 UCX DEBUG mlx5_ib5: using registration cache
[1650465629.591428] [ndv4:55516:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465629.591435] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465629.592071] [ndv4:54756:0] async.c:228 UCX DEBUG added async handler 0x1e5ad90 [id=71 ref 1] uct_ib_async_event_handler() to hash
[1650465629.592095] [ndv4:54756:0] async.c:506 UCX DEBUG listening to async event fd 71 events 0x1 mode thread_spinlock
[1650465629.592098] [ndv4:54756:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib7' (InfiniBand channel adapter) with 1 ports
[1650465629.592401] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465629.592410] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465629.592418] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.592469] [ndv4:54756:0] ib_md.c:1319 UCX DEBUG mlx5_ib7: using registration cache
[1650465629.593826] [ndv4:54756:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.593834] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.593849] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465629.593858] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.593861] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.593879] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.594001] [ndv4:54756:0] ib_md.c:1604 UCX DEBUG mlx5_ib7: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.594038] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465629.594046] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465629.594324] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x2cc0010 of 8176 bytes with 127 elements
[1650465629.594551] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465629.594563] [ndv4:54965:0] 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> }
[1650465629.594618] [ndv4:54965:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650465629.594623] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465629.594635] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x1b0a9a0 [id=92 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465629.594656] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 92 events 0x1 mode thread_spinlock
[1650465629.594669] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[12]=0x2b2e010 using rc_mlx5/mlx5_ib1:1 on worker 0x23dc840
[1650465629.594732] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.594739] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.594879] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.594884] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.595308] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465629.596673] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465629.596696] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.596700] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.596757] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.597196] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465629.597203] [ndv4:54965:0] 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> }
[1650465629.597238] [ndv4:54965:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650465629.597242] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465629.597249] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x1b16de0 [id=94 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465629.597271] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 94 events 0x1 mode thread_spinlock
[1650465629.597864] [ndv4:54965:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465629.599779] [ndv4:55152:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.599793] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.599955] [ndv4:55152:0] ib_md.c:1604 UCX DEBUG mlx5_ib7: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.600083] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465629.600090] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465629.601665] [ndv4:55152:0] topo.c:99 UCX DEBUG bus id 0x108000000 doesn't exist. sys_dev = 7
[1650465629.601673] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465629.602046] [ndv4:55152:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465629.602050] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465629.602477] [ndv4:55512:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.602486] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.602675] [ndv4:55512:0] ib_md.c:1604 UCX DEBUG mlx5_ib5: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.602692] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465629.602698] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465629.603230] [ndv4:55512:0] topo.c:99 UCX DEBUG bus id 0x106000000 doesn't exist. sys_dev = 5
[1650465629.603235] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465629.603898] [ndv4:55512:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465629.603902] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465629.605126] [ndv4:55432:0] debug.c:1198 UCX DEBUG using signal stack 0x2aee39a40000 size 141824
[1650465629.605202] [ndv4:55432:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465629.605224] [ndv4:55432:0] 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 0x2aee398ae000
[1650465629.605241] [ndv4:55432:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650465629.605250] [ndv4:55432:0] 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
[1650465629.605256] [ndv4:55432:0] module.c:253 UCX DEBUG loading modules for ucs
[1650465629.605219] [ndv4:54965:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x2e58010: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xda8b
[1650465629.605300] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.605306] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.605314] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b4b1ff90008 of 151544 bytes with 1052 elements
[1650465629.609236] [ndv4:54965:0] ib_md.c:812 UCX DEBUG registered memory 0x2b4b27a00000..0x2b4b2a000000 on mlx5_ib1 lkey 0x80400 rkey 0x80400 access 0xf flags 0x3e4
[1650465629.609258] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b4b27a00018 of 39845864 bytes with 4752 elements
[1650465629.609398] [ndv4:54965:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x2e58010
[1650465629.609432] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[13]=0x2e58010 using dc_mlx5/mlx5_ib1:1 on worker 0x23dc840
[1650465629.609533] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.609545] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.609667] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.609672] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.607726] [ndv4:55432:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465629.607745] [ndv4:55432:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650465629.607777] [ndv4:55432:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650465629.607780] [ndv4:55432:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650465629.607787] [ndv4:55432:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650465629.607794] [ndv4:55432:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650465629.607796] [ndv4:55432:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650465629.607802] [ndv4:55432:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650465629.607804] [ndv4:55432:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650465629.607806] [ndv4:55432:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650465629.607809] [ndv4:55432:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650465629.607811] [ndv4:55432:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650465629.607819] [ndv4:55432:0] module.c:253 UCX DEBUG loading modules for uct
[1650465629.608495] [ndv4:55432:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650465629.609063] [ndv4:55432:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650465629.609076] [ndv4:55432:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650465629.609088] [ndv4:55432:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650465629.609098] [ndv4:55432:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650465629.609109] [ndv4:55432:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650465629.609119] [ndv4:55432:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650465629.609129] [ndv4:55432:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650465629.609487] [ndv4:55432:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650465629.610070] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465629.610937] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465629.611250] [ndv4:54965:0] ib_iface.c:994 UCX DEBUG iface=0x2d57460: created UD QP 0xda71 on mlx5_ib1:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465629.611788] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465629.611868] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.611874] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.611919] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.611924] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.613363] [ndv4:55516:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465629.613371] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465629.612365] [ndv4:54965:0] ib_md.c:812 UCX DEBUG registered memory 0x2b4b2a003000..0x2b4b2a088000 on mlx5_ib1 lkey 0x80500 rkey 0x80500 access 0xf flags 0x3e4
[1650465629.612371] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b4b2a003018 of 544744 bytes with 128 elements
[1650465629.612375] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465629.612758] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x2d57460: adding gid fe80::15:5dff:fd33:fffc to hash on device mlx5_ib1 port 1 index 0)
[1650465629.613073] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x2d57460: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 1)
[1650465629.613339] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x2d57460: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 2)
[1650465629.615129] [ndv4:55512:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465629.615137] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465629.615425] [ndv4:54756:0] topo.c:99 UCX DEBUG bus id 0x108000000 doesn't exist. sys_dev = 7
[1650465629.615436] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465629.619285] [ndv4:55432:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650465629.619663] [ndv4:55432:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650465629.623945] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x2d57460: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 3)
[1650465629.624195] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x2d57460: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 4)
[1650465629.624358] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x2d57460: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 5)
[1650465629.624675] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x2d57460: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 6)
[1650465629.626343] [ndv4:54756:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465629.626351] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465629.625454] [ndv4:55516:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465629.625462] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465629.626448] [ndv4:55516:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465629.626452] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465629.626754] [ndv4:55516:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.630032] [ndv4:54756:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465629.630039] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465629.633515] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x2d57460: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 7)
[1650465629.633872] [ndv4:54965:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465629.633882] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x1b09820 [id=95 ref 1] uct_ud_iface_async_handler() to hash
[1650465629.633909] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 95 events 0x5 mode thread_spinlock
[1650465629.633921] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[14]=0x2d57460 using ud_verbs/mlx5_ib1:1 on worker 0x23dc840
[1650465629.634016] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.634022] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.634170] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.634174] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.635378] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465629.634295] [ndv4:55152:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465629.634304] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465629.635444] [ndv4:55432:0] async.c:228 UCX DEBUG added async handler 0x210e320 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650465629.635540] [ndv4:55432:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650465629.635548] [ndv4:55432:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650465629.634390] [ndv4:55512:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465629.634398] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465629.635993] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.636003] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.636027] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.636085] [ndv4:55432:0] module.c:253 UCX DEBUG loading modules for ucm
[1650465629.636107] [ndv4:55432:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650465629.637468] [ndv4:55432:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.637475] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.635502] [ndv4:55516:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib5 vendor_id: 0x15b3 device_id: 4124
[1650465629.635893] [ndv4:55516:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib5: disable ODP because it's not supported for DevX QP
[1650465629.637206] [ndv4:55516:0] async.c:228 UCX DEBUG added async handler 0x1958bf0 [id=67 ref 1] uct_ib_async_event_handler() to hash
[1650465629.637241] [ndv4:55516:0] async.c:506 UCX DEBUG listening to async event fd 67 events 0x1 mode thread_spinlock
[1650465629.637245] [ndv4:55516:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib5' (InfiniBand channel adapter) with 1 ports
[1650465629.637482] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465629.637494] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465629.637505] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.637609] [ndv4:55516:0] ib_md.c:1319 UCX DEBUG mlx5_ib5: using registration cache
[1650465629.637736] [ndv4:55432:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.637893] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.637899] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.636829] [ndv4:55152:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465629.636835] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465629.639314] [ndv4:55152:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465629.639321] [ndv4:55152:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465629.639407] [ndv4:55152:0] ucp_context.c:1556 UCX DEBUG created ucp context 0x2523d70 0x2523d70 [13 mds 47 tls] features 0x1 tl bitmap 0x7fffffffffff 0x0
[1650465629.639961] [ndv4:55516:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.639971] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.640125] [ndv4:55516:0] ib_md.c:1604 UCX DEBUG mlx5_ib5: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.640159] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465629.640167] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465629.641513] [ndv4:55516:0] topo.c:99 UCX DEBUG bus id 0x106000000 doesn't exist. sys_dev = 5
[1650465629.641518] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465629.640696] [ndv4:54756:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465629.640729] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465629.644787] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465629.644477] [ndv4:55512:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465629.644484] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465629.644758] [ndv4:55512:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.645244] [ndv4:54965:0] ib_iface.c:994 UCX DEBUG iface=0x2c1e280: created UD QP 0xda72 on mlx5_ib1:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465629.645251] [ndv4:54965:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465629.645932] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465629.646214] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.646222] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.646358] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.646363] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.647179] [ndv4:55432:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650465629.647187] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.646753] [ndv4:54965:0] ib_md.c:812 UCX DEBUG registered memory 0x2b4b2a088000..0x2b4b2a10d000 on mlx5_ib1 lkey 0x80600 rkey 0x80600 access 0xf flags 0x3e4
[1650465629.646759] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b4b2a088018 of 544744 bytes with 128 elements
[1650465629.646764] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465629.647113] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x2c1e280: adding gid fe80::15:5dff:fd33:fffc to hash on device mlx5_ib1 port 1 index 0)
[1650465629.647366] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x2c1e280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 1)
[1650465629.649230] [ndv4:55432:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465629.649237] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.649659] [ndv4:55691:0] debug.c:1198 UCX DEBUG using signal stack 0x2b871c5e3000 size 141824
[1650465629.649731] [ndv4:55691:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465629.649750] [ndv4:55691:0] 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 0x2b871c437000
[1650465629.649771] [ndv4:55691:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650465629.649779] [ndv4:55691:0] 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
[1650465629.649785] [ndv4:55691:0] module.c:253 UCX DEBUG loading modules for ucs
[1650465629.650472] [ndv4:54756:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465629.650483] [ndv4:54756:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465629.650795] [ndv4:54756:0] ucp_context.c:1556 UCX DEBUG created ucp context 0x1e43b50 0x1e43b50 [13 mds 47 tls] features 0x1 tl bitmap 0x7fffffffffff 0x0
[1650465629.652420] [ndv4:55691:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465629.652436] [ndv4:55691:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650465629.652472] [ndv4:55691:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650465629.652475] [ndv4:55691:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650465629.652481] [ndv4:55691:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650465629.652488] [ndv4:55691:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650465629.652491] [ndv4:55691:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650465629.652495] [ndv4:55691:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650465629.652498] [ndv4:55691:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650465629.652501] [ndv4:55691:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650465629.652504] [ndv4:55691:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650465629.652506] [ndv4:55691:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650465629.652514] [ndv4:55691:0] module.c:253 UCX DEBUG loading modules for uct
[1650465629.653260] [ndv4:55691:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650465629.654170] [ndv4:55691:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650465629.654188] [ndv4:55691:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650465629.654200] [ndv4:55691:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650465629.654212] [ndv4:55691:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650465629.654357] [ndv4:55691:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650465629.654511] [ndv4:55691:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650465629.654528] [ndv4:55691:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650465629.654940] [ndv4:55691:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650465629.657259] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x2c1e280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 2)
[1650465629.657690] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x2c1e280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 3)
[1650465629.658091] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x2c1e280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 4)
[1650465629.658369] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x2c1e280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 5)
[1650465629.661807] [ndv4:55516:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465629.661817] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465629.661786] [ndv4:55512:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib6 vendor_id: 0x15b3 device_id: 4124
[1650465629.662107] [ndv4:55512:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib6: disable ODP because it's not supported for DevX QP
[1650465629.663182] [ndv4:55512:0] async.c:228 UCX DEBUG added async handler 0x128da10 [id=69 ref 1] uct_ib_async_event_handler() to hash
[1650465629.663204] [ndv4:55512:0] async.c:506 UCX DEBUG listening to async event fd 69 events 0x1 mode thread_spinlock
[1650465629.663207] [ndv4:55512:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib6' (InfiniBand channel adapter) with 1 ports
[1650465629.664428] [ndv4:55691:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650465629.663491] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465629.663500] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465629.663505] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.663545] [ndv4:55512:0] ib_md.c:1319 UCX DEBUG mlx5_ib6: using registration cache
[1650465629.664728] [ndv4:55512:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.664734] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.664790] [ndv4:55691:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650465629.664899] [ndv4:55512:0] ib_md.c:1604 UCX DEBUG mlx5_ib6: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.665070] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465629.665077] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465629.666607] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x2c1e280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 6)
[1650465629.666889] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x2c1e280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 7)
[1650465629.666896] [ndv4:54965:0] 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> }
[1650465629.666928] [ndv4:54965:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465629.666932] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x1b18b40 [id=96 ref 1] uct_ud_iface_async_handler() to hash
[1650465629.666953] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 96 events 0x5 mode thread_spinlock
[1650465629.666965] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[15]=0x2c1e280 using ud_mlx5/mlx5_ib1:1 on worker 0x23dc840
[1650465629.667120] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.667126] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.667260] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.667265] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.667710] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465629.669124] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465629.669143] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.669146] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.669206] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.669194] [ndv4:55432:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465629.669205] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.670232] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465629.670238] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465629.670752] [ndv4:54965:0] ib_iface.c:994 UCX DEBUG iface=0x32470a0: created RC QP 0xdacc on mlx5_ib2:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465629.671077] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[16]=0x32470a0 using rc_verbs/mlx5_ib2:1 on worker 0x23dc840
[1650465629.671265] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.671271] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.671989] [ndv4:55691:0] async.c:228 UCX DEBUG added async handler 0x23c5280 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650465629.672086] [ndv4:55691:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650465629.672093] [ndv4:55691:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650465629.672451] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.672462] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.672486] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.672571] [ndv4:55691:0] module.c:253 UCX DEBUG loading modules for ucm
[1650465629.672628] [ndv4:55691:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650465629.671666] [ndv4:55516:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465629.671674] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465629.676694] [ndv4:55512:0] topo.c:99 UCX DEBUG bus id 0x107000000 doesn't exist. sys_dev = 6
[1650465629.676706] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465629.680029] [ndv4:55152:0] mm_posix.c:518 UCX DEBUG allocated posix shared memory at 0x2b3b9706c000 length 12288
[1650465629.680097] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool mm_recv_desc: align 64, maxelems 4294967295, elemsize 8368
[1650465629.681185] [ndv4:55152:0] mm_posix.c:305 UCX DEBUG shared memory mmap(addr=(nil), length=6291456, flags= HUGETLB, fd=76) failed: Invalid argument
[1650465629.681195] [ndv4:55152:0] mm_posix.c:518 UCX DEBUG allocated posix shared memory at 0x2b3b9706f000 length 4296704
[1650465629.681200] [ndv4:55152:0] mpool.c:205 UCX DEBUG mpool mm_recv_desc: allocated chunk 0x2b3b9706f018 of 4296680 bytes with 512 elements
[1650465629.681456] [ndv4:55152:0] mm_iface.c:600 UCX DEBUG created mm iface 0x2ca0d60 FIFO id 0x400000002cee43e8 va 0x2b3b9706c000 size 12288 (128 x 64 elems)
[1650465629.681507] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[0]=0x2ca0d60 using posix/memory on worker 0x357a8c0
[1650465629.681533] [ndv4:55152:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 8447 bytes with hugetlb
[1650465629.681571] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool mm_recv_desc: align 64, maxelems 4294967295, elemsize 8368
[1650465629.681606] [ndv4:55152:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 4292720 bytes with hugetlb
[1650465629.681616] [ndv4:55152:0] mpool.c:205 UCX DEBUG mpool mm_recv_desc: allocated chunk 0x2b3b9748b018 of 4296680 bytes with 512 elements
[1650465629.682172] [ndv4:55152:0] mm_iface.c:600 UCX DEBUG created mm iface 0x2ca1330 FIFO id 0x718018 va 0x2b3b97488000 size 12288 (128 x 64 elems)
[1650465629.682181] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[1]=0x2ca1330 using sysv/memory on worker 0x357a8c0
[1650465629.682193] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool self_msg_desc: align 64, maxelems 4294967295, elemsize 8200
[1650465629.682196] [ndv4:55152:0] self.c:220 UCX DEBUG created self iface id 0xcf9893745fe6be74 send_size 8192
[1650465629.682202] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[2]=0x2cb4f10 using self/memory0 on worker 0x357a8c0
[1650465629.682224] [ndv4:55152:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465629.682230] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465629.682233] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465629.679925] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.679935] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.680376] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465629.681723] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465629.681739] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.681743] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.681794] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.682267] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x3568010 of 8176 bytes with 127 elements
[1650465629.682484] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465629.682495] [ndv4:54965:0] 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> }
[1650465629.682531] [ndv4:54965:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib2 length=2048) failed: Invalid argument
[1650465629.682537] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465629.682550] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x1b00f60 [id=99 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465629.682570] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 99 events 0x1 mode thread_spinlock
[1650465629.682616] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[17]=0x3394030 using rc_mlx5/mlx5_ib2:1 on worker 0x23dc840
[1650465629.682726] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.682733] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.682811] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.682816] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.680993] [ndv4:55516:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465629.681001] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465629.682909] [ndv4:55516:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465629.682916] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465629.683216] [ndv4:55516:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.682942] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465629.684068] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465629.684084] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.684087] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.684142] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.684507] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465629.684514] [ndv4:54965:0] 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> }
[1650465629.684546] [ndv4:54965:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib2 length=2048) failed: Invalid argument
[1650465629.684550] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465629.684557] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x28e9ef0 [id=101 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465629.684608] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 101 events 0x1 mode thread_spinlock
[1650465629.685205] [ndv4:54965:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465629.685452] [ndv4:55152:0] async.c:228 UCX DEBUG added async handler 0x2cb7050 [id=78 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465629.685487] [ndv4:55152:0] async.c:506 UCX DEBUG listening to async event fd 78 events 0x5 mode thread_spinlock
[1650465629.685505] [ndv4:55152:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x2cb5870: listening for connections (fd=78) on 10.5.0.5:46789
[1650465629.685721] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[3]=0x2cb5870 using tcp/eth0 on worker 0x357a8c0
[1650465629.685743] [ndv4:55152:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465629.685747] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465629.685750] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465629.685790] [ndv4:55152:0] async.c:228 UCX DEBUG added async handler 0x2c119a0 [id=80 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465629.685810] [ndv4:55152:0] async.c:506 UCX DEBUG listening to async event fd 80 events 0x5 mode thread_spinlock
[1650465629.685814] [ndv4:55152:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x2cb5ef0: listening for connections (fd=80) on 127.0.0.1:59444
[1650465629.685829] [ndv4:55152:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465629.685834] [ndv4:55152:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465629.685891] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[4]=0x2cb5ef0 using tcp/lo on worker 0x357a8c0
[1650465629.685908] [ndv4:55152:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465629.685911] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465629.685914] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465629.685950] [ndv4:55152:0] async.c:228 UCX DEBUG added async handler 0x2c0fb10 [id=82 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465629.685974] [ndv4:55152:0] async.c:506 UCX DEBUG listening to async event fd 82 events 0x5 mode thread_spinlock
[1650465629.685978] [ndv4:55152:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x2c9d580: listening for connections (fd=82) on 172.16.1.242:39739
[1650465629.686197] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[5]=0x2c9d580 using tcp/ib0 on worker 0x357a8c0
[1650465629.686387] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.686395] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.686560] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.686565] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.687243] [ndv4:55152:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465629.688034] [ndv4:55152:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465629.688072] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.688076] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.688088] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.688479] [ndv4:55432:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465629.688489] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.688526] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465629.688534] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465629.689697] [ndv4:55152:0] ib_iface.c:994 UCX DEBUG iface=0x2cab6c0: created RC QP 0xf480 on mlx5_ib0:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465629.691626] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[6]=0x2cab6c0 using rc_verbs/mlx5_ib0:1 on worker 0x357a8c0
[1650465629.691668] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.691674] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.691772] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.691776] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.690406] [ndv4:55512:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465629.690420] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465629.692232] [ndv4:54965:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x356a050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xdadf
[1650465629.692290] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.692296] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.692305] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b4b1ffb7008 of 151544 bytes with 1052 elements
[1650465629.692641] [ndv4:55152:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465629.692870] [ndv4:55152:0] ib_device.c:1394 UCX DEBUG max IB CQE size is 128
[1650465629.694006] [ndv4:55152:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465629.694016] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.694019] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.694072] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.694498] [ndv4:55152:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x3644010 of 8176 bytes with 127 elements
[1650465629.695644] [ndv4:55691:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.695659] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.695817] [ndv4:55691:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.695974] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.695981] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.694774] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465629.694797] [ndv4:55152:0] 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> }
[1650465629.694848] [ndv4:55152:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650465629.694854] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465629.695972] [ndv4:54965:0] ib_md.c:812 UCX DEBUG registered memory 0x2b4b2a200000..0x2b4b2c800000 on mlx5_ib2 lkey 0x80600 rkey 0x80600 access 0xf flags 0x3e4
[1650465629.695990] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b4b2a200018 of 39845864 bytes with 4752 elements
[1650465629.696129] [ndv4:54965:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x356a050
[1650465629.696160] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[18]=0x356a050 using dc_mlx5/mlx5_ib2:1 on worker 0x23dc840
[1650465629.696407] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.696418] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.696691] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.696696] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.697604] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465629.698909] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465629.699211] [ndv4:54965:0] ib_iface.c:994 UCX DEBUG iface=0x3746060: created UD QP 0xdad5 on mlx5_ib2:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465629.699781] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465629.699879] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.699886] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.699921] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.699926] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.699778] [ndv4:55432:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465629.699786] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.700007] [ndv4:55432:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.700379] [ndv4:54965:0] ib_md.c:812 UCX DEBUG registered memory 0x2b4b2c90e000..0x2b4b2c993000 on mlx5_ib2 lkey 0x80700 rkey 0x80700 access 0xf flags 0x3e4
[1650465629.700385] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b4b2c90e018 of 544744 bytes with 128 elements
[1650465629.700389] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465629.700434] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x3746060: adding gid fe80::15:5dff:fd33:fffd to hash on device mlx5_ib2 port 1 index 0)
[1650465629.700460] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x3746060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 1)
[1650465629.700476] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x3746060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 2)
[1650465629.700497] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x3746060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 3)
[1650465629.700525] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x3746060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 4)
[1650465629.700553] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x3746060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 5)
[1650465629.700643] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x3746060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 6)
[1650465629.700672] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x3746060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 7)
[1650465629.700943] [ndv4:54965:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465629.700952] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x2cc39c0 [id=102 ref 1] uct_ud_iface_async_handler() to hash
[1650465629.700981] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 102 events 0x5 mode thread_spinlock
[1650465629.700992] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[19]=0x3746060 using ud_verbs/mlx5_ib2:1 on worker 0x23dc840
[1650465629.701003] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.701008] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.701050] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.701054] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.701197] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465629.701899] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465629.702210] [ndv4:54965:0] ib_iface.c:994 UCX DEBUG iface=0x139f1a0: created UD QP 0xdad6 on mlx5_ib2:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465629.702217] [ndv4:54965:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465629.701931] [ndv4:55152:0] async.c:228 UCX DEBUG added async handler 0x2c01350 [id=85 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465629.701961] [ndv4:55152:0] async.c:506 UCX DEBUG listening to async event fd 85 events 0x1 mode thread_spinlock
[1650465629.701980] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[7]=0x2cb9110 using rc_mlx5/mlx5_ib0:1 on worker 0x357a8c0
[1650465629.702003] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.702010] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.702087] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.702093] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.702278] [ndv4:55152:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465629.702508] [ndv4:55516:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib6 vendor_id: 0x15b3 device_id: 4124
[1650465629.702874] [ndv4:55516:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib6: disable ODP because it's not supported for DevX QP
[1650465629.702774] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465629.702848] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.702854] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.702865] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.702869] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.702995] [ndv4:55152:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465629.703006] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.703010] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.703024] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.703198] [ndv4:54965:0] ib_md.c:812 UCX DEBUG registered memory 0x2b4b2c993000..0x2b4b2ca18000 on mlx5_ib2 lkey 0x80800 rkey 0x80800 access 0xf flags 0x3e4
[1650465629.703205] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b4b2c993018 of 544744 bytes with 128 elements
[1650465629.703210] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465629.703391] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465629.703399] [ndv4:55152:0] 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> }
[1650465629.703431] [ndv4:55152:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650465629.703435] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465629.703444] [ndv4:55152:0] async.c:228 UCX DEBUG added async handler 0x2c0fb80 [id=87 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465629.703464] [ndv4:55152:0] async.c:506 UCX DEBUG listening to async event fd 87 events 0x1 mode thread_spinlock
[1650465629.704067] [ndv4:55152:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465629.704179] [ndv4:55512:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465629.704190] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465629.708180] [ndv4:55432:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650465629.708537] [ndv4:55432:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650465629.709052] [ndv4:55432:0] async.c:228 UCX DEBUG added async handler 0x2134fc0 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650465629.709077] [ndv4:55432:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650465629.709081] [ndv4:55432:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650465629.709269] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.709278] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.709283] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.709325] [ndv4:55432:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650465629.709938] [ndv4:55432:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.709943] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.711535] [ndv4:55152:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x38ae010: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xf4a6
[1650465629.711884] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.711891] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.711911] [ndv4:55152:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b3b978a6008 of 151544 bytes with 1052 elements
[1650465629.712100] [ndv4:55691:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650465629.712109] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.712502] [ndv4:55691:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465629.712506] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.710162] [ndv4:55432:0] ib_md.c:1604 UCX DEBUG mlx5_ib1: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.710244] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.710250] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.711188] [ndv4:55432:0] topo.c:99 UCX DEBUG bus id 0x102000000 doesn't exist. sys_dev = 1
[1650465629.711192] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.714875] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x139f1a0: adding gid fe80::15:5dff:fd33:fffd to hash on device mlx5_ib2 port 1 index 0)
[1650465629.715002] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x139f1a0: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 1)
[1650465629.715092] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x139f1a0: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 2)
[1650465629.715183] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x139f1a0: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 3)
[1650465629.715250] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x139f1a0: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 4)
[1650465629.715338] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x139f1a0: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 5)
[1650465629.715427] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x139f1a0: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 6)
[1650465629.715515] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x139f1a0: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 7)
[1650465629.715521] [ndv4:54965:0] 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> }
[1650465629.715552] [ndv4:54965:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465629.715556] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x3746fc0 [id=103 ref 1] uct_ud_iface_async_handler() to hash
[1650465629.715594] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 103 events 0x5 mode thread_spinlock
[1650465629.715607] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[20]=0x139f1a0 using ud_mlx5/mlx5_ib2:1 on worker 0x23dc840
[1650465629.715623] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465629.715629] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465629.715712] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465629.715718] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465629.715864] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465629.715533] [ndv4:55516:0] async.c:228 UCX DEBUG added async handler 0x1277a10 [id=69 ref 1] uct_ib_async_event_handler() to hash
[1650465629.715587] [ndv4:55516:0] async.c:506 UCX DEBUG listening to async event fd 69 events 0x1 mode thread_spinlock
[1650465629.715593] [ndv4:55516:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib6' (InfiniBand channel adapter) with 1 ports
[1650465629.715828] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465629.715838] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465629.715844] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.715909] [ndv4:55516:0] ib_md.c:1319 UCX DEBUG mlx5_ib6: using registration cache
[1650465629.715594] [ndv4:54756:0] mm_posix.c:518 UCX DEBUG allocated posix shared memory at 0x2b3b53d21000 length 12288
[1650465629.715665] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool mm_recv_desc: align 64, maxelems 4294967295, elemsize 8368
[1650465629.716781] [ndv4:54756:0] mm_posix.c:305 UCX DEBUG shared memory mmap(addr=(nil), length=6291456, flags= HUGETLB, fd=76) failed: Invalid argument
[1650465629.716790] [ndv4:54756:0] mm_posix.c:518 UCX DEBUG allocated posix shared memory at 0x2b3b58000000 length 4296704
[1650465629.716796] [ndv4:54756:0] mpool.c:205 UCX DEBUG mpool mm_recv_desc: allocated chunk 0x2b3b58000018 of 4296680 bytes with 512 elements
[1650465629.717053] [ndv4:54756:0] mm_iface.c:600 UCX DEBUG created mm iface 0x25c0d70 FIFO id 0x400000004ce22b49 va 0x2b3b53d21000 size 12288 (128 x 64 elems)
[1650465629.717103] [ndv4:54756:0] ucp_worker.c:1159 UCX DEBUG created interface[0]=0x25c0d70 using posix/memory on worker 0x2e9a8d0
[1650465629.717130] [ndv4:54756:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 8447 bytes with hugetlb
[1650465629.717162] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool mm_recv_desc: align 64, maxelems 4294967295, elemsize 8368
[1650465629.717178] [ndv4:54756:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 4292720 bytes with hugetlb
[1650465629.717186] [ndv4:54756:0] mpool.c:205 UCX DEBUG mpool mm_recv_desc: allocated chunk 0x2b3b58419018 of 4296680 bytes with 512 elements
[1650465629.717847] [ndv4:54756:0] mm_iface.c:600 UCX DEBUG created mm iface 0x25c1340 FIFO id 0x71801c va 0x2b3b53d24000 size 12288 (128 x 64 elems)
[1650465629.717856] [ndv4:54756:0] ucp_worker.c:1159 UCX DEBUG created interface[1]=0x25c1340 using sysv/memory on worker 0x2e9a8d0
[1650465629.717868] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool self_msg_desc: align 64, maxelems 4294967295, elemsize 8200
[1650465629.717871] [ndv4:54756:0] self.c:220 UCX DEBUG created self iface id 0x98bcd82d5f8a642c send_size 8192
[1650465629.717877] [ndv4:54756:0] ucp_worker.c:1159 UCX DEBUG created interface[2]=0x25d4f20 using self/memory0 on worker 0x2e9a8d0
[1650465629.717901] [ndv4:54756:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465629.717906] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465629.717909] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465629.716986] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465629.717001] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.717004] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.717053] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.718161] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465629.718166] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465629.718669] [ndv4:54965:0] ib_iface.c:994 UCX DEBUG iface=0x39640a0: created RC QP 0xda53 on mlx5_ib3:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465629.719012] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[21]=0x39640a0 using rc_verbs/mlx5_ib3:1 on worker 0x23dc840
[1650465629.719135] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465629.719141] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465629.719261] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465629.719266] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465629.719649] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465629.717809] [ndv4:55516:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.717817] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.717986] [ndv4:55516:0] ib_md.c:1604 UCX DEBUG mlx5_ib6: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.718171] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465629.718178] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465629.719792] [ndv4:55516:0] topo.c:99 UCX DEBUG bus id 0x107000000 doesn't exist. sys_dev = 6
[1650465629.719798] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465629.717435] [ndv4:55152:0] ib_md.c:812 UCX DEBUG registered memory 0x2b3b9c200000..0x2b3b9e800000 on mlx5_ib0 lkey 0x80900 rkey 0x80900 access 0xf flags 0x3e4
[1650465629.717456] [ndv4:55152:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b3b9c200018 of 39845864 bytes with 4752 elements
[1650465629.717674] [ndv4:55152:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x38ae010
[1650465629.717706] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[8]=0x38ae010 using dc_mlx5/mlx5_ib0:1 on worker 0x357a8c0
[1650465629.717912] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.717924] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.718167] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.718173] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.719007] [ndv4:55152:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465629.720005] [ndv4:55152:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465629.720350] [ndv4:55152:0] ib_iface.c:994 UCX DEBUG iface=0x2cc1a80: created UD QP 0xf489 on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465629.716732] [ndv4:55512:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465629.716751] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465629.720887] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465629.720900] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.720903] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.720954] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.721275] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x3c85010 of 8176 bytes with 127 elements
[1650465629.720922] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465629.721155] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.721161] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.721225] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.721230] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.720439] [ndv4:54756:0] async.c:228 UCX DEBUG added async handler 0x25e2f40 [id=78 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465629.720469] [ndv4:54756:0] async.c:506 UCX DEBUG listening to async event fd 78 events 0x5 mode thread_spinlock
[1650465629.720488] [ndv4:54756:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x25d5880: listening for connections (fd=78) on 10.5.0.5:49260
[1650465629.720635] [ndv4:54756:0] ucp_worker.c:1159 UCX DEBUG created interface[3]=0x25d5880 using tcp/eth0 on worker 0x2e9a8d0
[1650465629.720655] [ndv4:54756:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465629.720659] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465629.720662] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465629.720725] [ndv4:54756:0] async.c:228 UCX DEBUG added async handler 0x25319c0 [id=80 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465629.720745] [ndv4:54756:0] async.c:506 UCX DEBUG listening to async event fd 80 events 0x5 mode thread_spinlock
[1650465629.720749] [ndv4:54756:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x25d5f00: listening for connections (fd=80) on 127.0.0.1:33573
[1650465629.720767] [ndv4:54756:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465629.720772] [ndv4:54756:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465629.720837] [ndv4:54756:0] ucp_worker.c:1159 UCX DEBUG created interface[4]=0x25d5f00 using tcp/lo on worker 0x2e9a8d0
[1650465629.720853] [ndv4:54756:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465629.720855] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465629.720858] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465629.720895] [ndv4:54756:0] async.c:228 UCX DEBUG added async handler 0x252fb30 [id=82 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465629.720917] [ndv4:54756:0] async.c:506 UCX DEBUG listening to async event fd 82 events 0x5 mode thread_spinlock
[1650465629.720920] [ndv4:54756:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x25bd590: listening for connections (fd=82) on 172.16.1.242:52612
[1650465629.721282] [ndv4:54756:0] ucp_worker.c:1159 UCX DEBUG created interface[5]=0x25bd590 using tcp/ib0 on worker 0x2e9a8d0
[1650465629.721410] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.721418] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.721594] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.721601] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.721773] [ndv4:54756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465629.721665] [ndv4:55152:0] ib_md.c:812 UCX DEBUG registered memory 0x2b3b978cb000..0x2b3b97950000 on mlx5_ib0 lkey 0x80a00 rkey 0x80a00 access 0xf flags 0x3e4
[1650465629.721672] [ndv4:55152:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b3b978cb018 of 544744 bytes with 128 elements
[1650465629.721676] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465629.721484] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465629.721489] [ndv4:54965:0] 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> }
[1650465629.721526] [ndv4:54965:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib3 length=2048) failed: Invalid argument
[1650465629.721530] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465629.721540] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x3864ae0 [id=106 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465629.721565] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 106 events 0x1 mode thread_spinlock
[1650465629.721573] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[22]=0x3ab1030 using rc_mlx5/mlx5_ib3:1 on worker 0x23dc840
[1650465629.722241] [ndv4:55432:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465629.722249] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.722674] [ndv4:54756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465629.722722] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.722727] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.722766] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.723070] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465629.723079] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465629.723623] [ndv4:54756:0] ib_iface.c:994 UCX DEBUG iface=0x25cb6d0: created RC QP 0xf48a on mlx5_ib0:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465629.727771] [ndv4:55512:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465629.727777] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465629.728073] [ndv4:55512:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.729286] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465629.729295] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465629.729518] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465629.729523] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465629.730286] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465629.729451] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x2cc1a80: adding gid fe80::15:5dff:fd33:fffb to hash on device mlx5_ib0 port 1 index 0)
[1650465629.729814] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x2cc1a80: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650465629.730558] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x2cc1a80: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650465629.731122] [ndv4:55432:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465629.731129] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.731859] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465629.731877] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.731881] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.731935] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.732386] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465629.732393] [ndv4:54965:0] 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> }
[1650465629.732425] [ndv4:54965:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib3 length=2048) failed: Invalid argument
[1650465629.732428] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465629.732435] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x1b05480 [id=108 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465629.732454] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 108 events 0x1 mode thread_spinlock
[1650465629.732808] [ndv4:55691:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465629.732817] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.733025] [ndv4:54965:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465629.734355] [ndv4:55691:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465629.734361] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.737356] [ndv4:55691:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465629.737362] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.737679] [ndv4:55691:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.740192] [ndv4:54965:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x3c87050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xda66
[1650465629.740561] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465629.740569] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465629.740633] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b4b2f219008 of 151544 bytes with 1052 elements
[1650465629.739538] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x2cc1a80: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650465629.740143] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x2cc1a80: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650465629.741014] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x2cc1a80: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650465629.741525] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x2cc1a80: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650465629.742380] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x2cc1a80: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650465629.742533] [ndv4:55512:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib7 vendor_id: 0x15b3 device_id: 4124
[1650465629.741706] [ndv4:54756:0] ucp_worker.c:1159 UCX DEBUG created interface[6]=0x25cb6d0 using rc_verbs/mlx5_ib0:1 on worker 0x2e9a8d0
[1650465629.741882] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.741889] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.741961] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.741965] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.742744] [ndv4:54756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465629.742915] [ndv4:55512:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib7: disable ODP because it's not supported for DevX QP
[1650465629.742722] [ndv4:55152:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465629.744295] [ndv4:54965:0] ib_md.c:812 UCX DEBUG registered memory 0x2b4b2cc00000..0x2b4b2f200000 on mlx5_ib3 lkey 0x80800 rkey 0x80800 access 0xf flags 0x3e4
[1650465629.744314] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b4b2cc00018 of 39845864 bytes with 4752 elements
[1650465629.744452] [ndv4:54965:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x3c87050
[1650465629.744486] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[23]=0x3c87050 using dc_mlx5/mlx5_ib3:1 on worker 0x23dc840
[1650465629.744787] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465629.744798] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465629.744997] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465629.745002] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465629.745738] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465629.749423] [ndv4:55432:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465629.749431] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.748526] [ndv4:55152:0] async.c:228 UCX DEBUG added async handler 0x2ca0c40 [id=88 ref 1] uct_ud_iface_async_handler() to hash
[1650465629.748558] [ndv4:55152:0] async.c:506 UCX DEBUG listening to async event fd 88 events 0x5 mode thread_spinlock
[1650465629.748657] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[9]=0x2cc1a80 using ud_verbs/mlx5_ib0:1 on worker 0x357a8c0
[1650465629.748817] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.748824] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.749078] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.749083] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.749949] [ndv4:55152:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465629.750518] [ndv4:54756:0] ib_device.c:1394 UCX DEBUG max IB CQE size is 128
[1650465629.750967] [ndv4:55152:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465629.751381] [ndv4:55152:0] ib_iface.c:994 UCX DEBUG iface=0x359a740: created UD QP 0xf48b on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465629.751397] [ndv4:55152:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465629.752041] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465629.751510] [ndv4:54756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465629.751524] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.751527] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.751621] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.752036] [ndv4:54756:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x2f64010 of 8176 bytes with 127 elements
[1650465629.752249] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465629.752272] [ndv4:54756:0] 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> }
[1650465629.752323] [ndv4:54756:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650465629.752329] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465629.752346] [ndv4:55512:0] async.c:228 UCX DEBUG added async handler 0x196f4b0 [id=71 ref 1] uct_ib_async_event_handler() to hash
[1650465629.752372] [ndv4:55512:0] async.c:506 UCX DEBUG listening to async event fd 71 events 0x1 mode thread_spinlock
[1650465629.752375] [ndv4:55512:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib7' (InfiniBand channel adapter) with 1 ports
[1650465629.752633] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465629.752643] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465629.752649] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.752704] [ndv4:55512:0] ib_md.c:1319 UCX DEBUG mlx5_ib7: using registration cache
[1650465629.754838] [ndv4:55512:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.754845] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.755007] [ndv4:55512:0] ib_md.c:1604 UCX DEBUG mlx5_ib7: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.755213] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465629.755220] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465629.752929] [ndv4:55516:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465629.752940] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465629.754944] [ndv4:55516:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465629.754950] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465629.753717] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465629.754029] [ndv4:54965:0] ib_iface.c:994 UCX DEBUG iface=0x3e63060: created UD QP 0xda5c on mlx5_ib3:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465629.754516] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465629.754728] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465629.754736] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465629.754892] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465629.754896] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465629.755340] [ndv4:54965:0] ib_md.c:812 UCX DEBUG registered memory 0x2b4b2f23e000..0x2b4b2f2c3000 on mlx5_ib3 lkey 0x80900 rkey 0x80900 access 0xf flags 0x3e4
[1650465629.755350] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b4b2f23e018 of 544744 bytes with 128 elements
[1650465629.755355] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465629.755497] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x3e63060: adding gid fe80::15:5dff:fd33:fffe to hash on device mlx5_ib3 port 1 index 0)
[1650465629.756485] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x3e63060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 1)
[1650465629.757709] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x3e63060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 2)
[1650465629.758191] [ndv4:55512:0] topo.c:99 UCX DEBUG bus id 0x108000000 doesn't exist. sys_dev = 7
[1650465629.758198] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465629.758527] [ndv4:55516:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465629.758534] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465629.758947] [ndv4:55432:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465629.758955] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.759173] [ndv4:55432:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.758979] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x3e63060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 3)
[1650465629.759839] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x3e63060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 4)
[1650465629.759302] [ndv4:54756:0] async.c:228 UCX DEBUG added async handler 0x25213a0 [id=85 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465629.759337] [ndv4:54756:0] async.c:506 UCX DEBUG listening to async event fd 85 events 0x1 mode thread_spinlock
[1650465629.759360] [ndv4:54756:0] ucp_worker.c:1159 UCX DEBUG created interface[7]=0x25d9120 using rc_mlx5/mlx5_ib0:1 on worker 0x2e9a8d0
[1650465629.759672] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.759680] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.759770] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.759775] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.759994] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.760004] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.760115] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.760121] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.760535] [ndv4:54756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465629.760903] [ndv4:55512:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465629.760909] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465629.760637] [ndv4:55152:0] ib_md.c:812 UCX DEBUG registered memory 0x2b3b97950000..0x2b3b979d5000 on mlx5_ib0 lkey 0x80b00 rkey 0x80b00 access 0xf flags 0x3e4
[1650465629.760644] [ndv4:55152:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b3b97950018 of 544744 bytes with 128 elements
[1650465629.760648] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465629.761540] [ndv4:54756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465629.761552] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.761555] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.761571] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.761794] [ndv4:55691:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650465629.761988] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465629.761995] [ndv4:54756:0] 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> }
[1650465629.762041] [ndv4:54756:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650465629.762045] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465629.762056] [ndv4:54756:0] async.c:228 UCX DEBUG added async handler 0x1e53fb0 [id=87 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465629.762076] [ndv4:54756:0] async.c:506 UCX DEBUG listening to async event fd 87 events 0x1 mode thread_spinlock
[1650465629.762171] [ndv4:55691:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650465629.762570] [ndv4:54756:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465629.763154] [ndv4:55691:0] async.c:228 UCX DEBUG added async handler 0x23bbc50 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650465629.763183] [ndv4:55691:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650465629.763187] [ndv4:55691:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650465629.763434] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.763444] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.763450] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.763496] [ndv4:55691:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650465629.765852] [ndv4:55691:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.765860] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.766047] [ndv4:55691:0] ib_md.c:1604 UCX DEBUG mlx5_ib1: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.766137] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.766143] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.767435] [ndv4:55691:0] topo.c:99 UCX DEBUG bus id 0x102000000 doesn't exist. sys_dev = 1
[1650465629.767439] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.768569] [ndv4:54756:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x31ce010: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xf4b4
[1650465629.768831] [ndv4:55691:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465629.768837] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.769455] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x3e63060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 5)
[1650465629.770082] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x3e63060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 6)
[1650465629.770533] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x3e63060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 7)
[1650465629.769838] [ndv4:55516:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465629.769846] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465629.770130] [ndv4:55516:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.770779] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x359a740: adding gid fe80::15:5dff:fd33:fffb to hash on device mlx5_ib0 port 1 index 0)
[1650465629.770866] [ndv4:54965:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465629.770879] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x3222fc0 [id=109 ref 1] uct_ud_iface_async_handler() to hash
[1650465629.770907] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 109 events 0x5 mode thread_spinlock
[1650465629.770921] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[24]=0x3e63060 using ud_verbs/mlx5_ib3:1 on worker 0x23dc840
[1650465629.770940] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465629.770946] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465629.771059] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465629.771064] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465629.771457] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465629.776945] [ndv4:55516:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib7 vendor_id: 0x15b3 device_id: 4124
[1650465629.777233] [ndv4:55516:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib7: disable ODP because it's not supported for DevX QP
[1650465629.777699] [ndv4:55432:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib2 vendor_id: 0x15b3 device_id: 4124
[1650465629.777967] [ndv4:55516:0] async.c:228 UCX DEBUG added async handler 0x19594b0 [id=71 ref 1] uct_ib_async_event_handler() to hash
[1650465629.777993] [ndv4:55516:0] async.c:506 UCX DEBUG listening to async event fd 71 events 0x1 mode thread_spinlock
[1650465629.777998] [ndv4:55516:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib7' (InfiniBand channel adapter) with 1 ports
[1650465629.778075] [ndv4:55432:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib2: disable ODP because it's not supported for DevX QP
[1650465629.778177] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465629.778187] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465629.778193] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.778245] [ndv4:55516:0] ib_md.c:1319 UCX DEBUG mlx5_ib7: using registration cache
[1650465629.780125] [ndv4:55691:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465629.780132] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.780432] [ndv4:55691:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465629.780436] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.780694] [ndv4:55691:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465629.780702] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.780929] [ndv4:55691:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.783957] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x359a740: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650465629.784076] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x359a740: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650465629.784130] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x359a740: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650465629.784161] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x359a740: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650465629.784204] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x359a740: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650465629.784363] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x359a740: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650465629.784552] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x359a740: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650465629.784558] [ndv4:55152:0] 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> }
[1650465629.784604] [ndv4:55152:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465629.784609] [ndv4:55152:0] async.c:228 UCX DEBUG added async handler 0x2533420 [id=89 ref 1] uct_ud_iface_async_handler() to hash
[1650465629.784633] [ndv4:55152:0] async.c:506 UCX DEBUG listening to async event fd 89 events 0x5 mode thread_spinlock
[1650465629.784647] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[10]=0x359a740 using ud_mlx5/mlx5_ib0:1 on worker 0x357a8c0
[1650465629.784683] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.784689] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.784807] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.784811] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.785215] [ndv4:55152:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465629.785984] [ndv4:55152:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465629.785993] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.785996] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.786009] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.784821] [ndv4:55512:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465629.784828] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465629.785469] [ndv4:55512:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465629.785474] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465629.785894] [ndv4:55512:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465629.785898] [ndv4:55512:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465629.785996] [ndv4:55512:0] ucp_context.c:1556 UCX DEBUG created ucp context 0x1272ca0 0x1272ca0 [13 mds 47 tls] features 0x1 tl bitmap 0x7fffffffffff 0x0
[1650465629.785516] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465629.785900] [ndv4:54965:0] ib_iface.c:994 UCX DEBUG iface=0x3f81460: created UD QP 0xda5d on mlx5_ib3:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465629.785910] [ndv4:54965:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465629.786622] [ndv4:55691:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib2 vendor_id: 0x15b3 device_id: 4124
[1650465629.786658] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465629.786665] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465629.786475] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465629.786609] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465629.786616] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465629.786679] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465629.786684] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465629.786996] [ndv4:55691:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib2: disable ODP because it's not supported for DevX QP
[1650465629.787758] [ndv4:55152:0] ib_iface.c:994 UCX DEBUG iface=0x3e4b050: created RC QP 0xda73 on mlx5_ib1:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465629.787684] [ndv4:54965:0] ib_md.c:812 UCX DEBUG registered memory 0x2b4b2f2c3000..0x2b4b2f348000 on mlx5_ib3 lkey 0x80a00 rkey 0x80a00 access 0xf flags 0x3e4
[1650465629.787693] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b4b2f2c3018 of 544744 bytes with 128 elements
[1650465629.787698] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465629.788170] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x3f81460: adding gid fe80::15:5dff:fd33:fffe to hash on device mlx5_ib3 port 1 index 0)
[1650465629.788603] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[11]=0x3e4b050 using rc_verbs/mlx5_ib1:1 on worker 0x357a8c0
[1650465629.788663] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.788669] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.788733] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.788737] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.789276] [ndv4:55152:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465629.789982] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.789992] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.790021] [ndv4:54756:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b3b53d29008 of 151544 bytes with 1052 elements
[1650465629.791829] [ndv4:55516:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.791853] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.792039] [ndv4:55516:0] ib_md.c:1604 UCX DEBUG mlx5_ib7: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.793740] [ndv4:54756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b3b58a00000..0x2b3b5b000000 on mlx5_ib0 lkey 0x80c00 rkey 0x80c00 access 0xf flags 0x3e4
[1650465629.793769] [ndv4:54756:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b3b58a00018 of 39845864 bytes with 4752 elements
[1650465629.794186] [ndv4:54756:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x31ce010
[1650465629.794265] [ndv4:54756:0] ucp_worker.c:1159 UCX DEBUG created interface[8]=0x31ce010 using dc_mlx5/mlx5_ib0:1 on worker 0x2e9a8d0
[1650465629.794348] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.794357] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.794638] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.794644] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.795669] [ndv4:54756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465629.796393] [ndv4:54756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465629.796876] [ndv4:54756:0] ib_iface.c:994 UCX DEBUG iface=0x25e1a90: created UD QP 0xf494 on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465629.797337] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465629.797882] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.797895] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.798025] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.798031] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.796540] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x3f81460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 1)
[1650465629.796684] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x3f81460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 2)
[1650465629.796860] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x3f81460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 3)
[1650465629.797187] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x3f81460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 4)
[1650465629.797466] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x3f81460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 5)
[1650465629.797531] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x3f81460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 6)
[1650465629.797563] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x3f81460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 7)
[1650465629.797569] [ndv4:54965:0] 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> }
[1650465629.797672] [ndv4:54965:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465629.797679] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x3f81f60 [id=110 ref 1] uct_ud_iface_async_handler() to hash
[1650465629.797697] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 110 events 0x5 mode thread_spinlock
[1650465629.797708] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[25]=0x3f81460 using ud_mlx5/mlx5_ib3:1 on worker 0x23dc840
[1650465629.797897] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465629.797904] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465629.798070] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465629.798075] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465629.799248] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465629.798398] [ndv4:54756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b3b53d4e000..0x2b3b53dd3000 on mlx5_ib0 lkey 0x80d00 rkey 0x80d00 access 0xf flags 0x3e4
[1650465629.798412] [ndv4:54756:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b3b53d4e018 of 544744 bytes with 128 elements
[1650465629.798429] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465629.799392] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x25e1a90: adding gid fe80::15:5dff:fd33:fffb to hash on device mlx5_ib0 port 1 index 0)
[1650465629.799936] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x25e1a90: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650465629.800273] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x25e1a90: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650465629.800538] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x25e1a90: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650465629.800810] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x25e1a90: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650465629.801055] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x25e1a90: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650465629.801217] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x25e1a90: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650465629.801304] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x25e1a90: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650465629.800401] [ndv4:55152:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465629.800412] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.800415] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.800431] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.800811] [ndv4:55152:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x3e5e010 of 8176 bytes with 127 elements
[1650465629.801046] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465629.801052] [ndv4:55152:0] 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> }
[1650465629.801086] [ndv4:55152:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650465629.801090] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465629.801100] [ndv4:55152:0] async.c:228 UCX DEBUG added async handler 0x3e53fb0 [id=92 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465629.801123] [ndv4:55152:0] async.c:506 UCX DEBUG listening to async event fd 92 events 0x1 mode thread_spinlock
[1650465629.801133] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[12]=0x3ccc010 using rc_mlx5/mlx5_ib1:1 on worker 0x357a8c0
[1650465629.801207] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.801213] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.801305] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.801310] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.801590] [ndv4:55152:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465629.801556] [ndv4:54756:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465629.803724] [ndv4:55432:0] async.c:228 UCX DEBUG added async handler 0x2134ce0 [id=61 ref 1] uct_ib_async_event_handler() to hash
[1650465629.803753] [ndv4:55432:0] async.c:506 UCX DEBUG listening to async event fd 61 events 0x1 mode thread_spinlock
[1650465629.803758] [ndv4:55432:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib2' (InfiniBand channel adapter) with 1 ports
[1650465629.803965] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.803977] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.803982] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.804025] [ndv4:55432:0] ib_md.c:1319 UCX DEBUG mlx5_ib2: using registration cache
[1650465629.804647] [ndv4:55432:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.804652] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.804866] [ndv4:55432:0] ib_md.c:1604 UCX DEBUG mlx5_ib2: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.804910] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.804916] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.805499] [ndv4:55432:0] topo.c:99 UCX DEBUG bus id 0x103000000 doesn't exist. sys_dev = 2
[1650465629.805504] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465629.806165] [ndv4:55432:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465629.806169] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465629.803837] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465629.803856] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465629.804262] [ndv4:55516:0] topo.c:99 UCX DEBUG bus id 0x108000000 doesn't exist. sys_dev = 7
[1650465629.804267] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465629.805062] [ndv4:55516:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465629.805067] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465629.805787] [ndv4:55516:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465629.805793] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465629.806933] [ndv4:55516:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465629.806938] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465629.807288] [ndv4:54756:0] async.c:228 UCX DEBUG added async handler 0x25c94c0 [id=88 ref 1] uct_ud_iface_async_handler() to hash
[1650465629.807321] [ndv4:54756:0] async.c:506 UCX DEBUG listening to async event fd 88 events 0x5 mode thread_spinlock
[1650465629.807354] [ndv4:54756:0] ucp_worker.c:1159 UCX DEBUG created interface[9]=0x25e1a90 using ud_verbs/mlx5_ib0:1 on worker 0x2e9a8d0
[1650465629.807800] [ndv4:55691:0] async.c:228 UCX DEBUG added async handler 0x23c3cb0 [id=61 ref 1] uct_ib_async_event_handler() to hash
[1650465629.807829] [ndv4:55691:0] async.c:506 UCX DEBUG listening to async event fd 61 events 0x1 mode thread_spinlock
[1650465629.807834] [ndv4:55691:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib2' (InfiniBand channel adapter) with 1 ports
[1650465629.808019] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.808034] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.808044] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.808100] [ndv4:55691:0] ib_md.c:1319 UCX DEBUG mlx5_ib2: using registration cache
[1650465629.809019] [ndv4:55691:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.809025] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.809170] [ndv4:55691:0] ib_md.c:1604 UCX DEBUG mlx5_ib2: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.809301] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.809308] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.809972] [ndv4:55691:0] topo.c:99 UCX DEBUG bus id 0x103000000 doesn't exist. sys_dev = 2
[1650465629.809977] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465629.810368] [ndv4:55691:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465629.810372] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465629.810487] [ndv4:55691:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465629.810490] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465629.810697] [ndv4:55691:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465629.810704] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465629.810821] [ndv4:55691:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465629.810825] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465629.811049] [ndv4:55691:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.813645] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465629.813667] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.813670] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.813726] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.814765] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465629.814771] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465629.815317] [ndv4:54965:0] ib_iface.c:994 UCX DEBUG iface=0x40810a0: created RC QP 0xdac5 on mlx5_ib4:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465629.815668] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[26]=0x40810a0 using rc_verbs/mlx5_ib4:1 on worker 0x23dc840
[1650465629.815690] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465629.815695] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465629.815554] [ndv4:55152:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465629.815594] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.815598] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.815653] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.816073] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465629.816083] [ndv4:55152:0] 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> }
[1650465629.816116] [ndv4:55152:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650465629.816121] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465629.816130] [ndv4:55152:0] async.c:228 UCX DEBUG added async handler 0x2cb4e80 [id=94 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465629.816152] [ndv4:55152:0] async.c:506 UCX DEBUG listening to async event fd 94 events 0x1 mode thread_spinlock
[1650465629.816790] [ndv4:55152:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465629.818126] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.818136] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.818296] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.818302] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.819015] [ndv4:54756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465629.818852] [ndv4:55516:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465629.818860] [ndv4:55516:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465629.818978] [ndv4:55516:0] ucp_context.c:1556 UCX DEBUG created ucp context 0x125cca0 0x125cca0 [13 mds 47 tls] features 0x1 tl bitmap 0x7fffffffffff 0x0
[1650465629.820672] [ndv4:54756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465629.821138] [ndv4:54756:0] ib_iface.c:994 UCX DEBUG iface=0x25e3320: created UD QP 0xf496 on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465629.821156] [ndv4:54756:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465629.821666] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465629.821971] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.821978] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.822088] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.822094] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.822470] [ndv4:54756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b3b53dd3000..0x2b3b53e58000 on mlx5_ib0 lkey 0x80e00 rkey 0x80e00 access 0xf flags 0x3e4
[1650465629.822484] [ndv4:54756:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b3b53dd3018 of 544744 bytes with 128 elements
[1650465629.822500] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465629.823084] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x25e3320: adding gid fe80::15:5dff:fd33:fffb to hash on device mlx5_ib0 port 1 index 0)
[1650465629.823447] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x25e3320: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650465629.823791] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x25e3320: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650465629.824006] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x25e3320: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650465629.824199] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x25e3320: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650465629.824372] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x25e3320: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650465629.824544] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x25e3320: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650465629.824806] [ndv4:55152:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x3ff6010: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xda99
[1650465629.825072] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.825079] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.825088] [ndv4:55152:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b3b979d7008 of 151544 bytes with 1052 elements
[1650465629.826658] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465629.826670] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465629.826857] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465629.828238] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465629.828252] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.828255] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.828308] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.828896] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x43a2010 of 8176 bytes with 127 elements
[1650465629.829166] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465629.829176] [ndv4:54965:0] 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> }
[1650465629.829210] [ndv4:54965:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib4 length=2048) failed: Invalid argument
[1650465629.829214] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465629.829227] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x4089f60 [id=113 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465629.829248] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 113 events 0x1 mode thread_spinlock
[1650465629.829258] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[27]=0x41ce030 using rc_mlx5/mlx5_ib4:1 on worker 0x23dc840
[1650465629.829274] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465629.829280] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465629.829381] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465629.829386] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465629.828988] [ndv4:55152:0] ib_md.c:812 UCX DEBUG registered memory 0x2b3b9ea00000..0x2b3ba1000000 on mlx5_ib1 lkey 0x80700 rkey 0x80700 access 0xf flags 0x3e4
[1650465629.829004] [ndv4:55152:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b3b9ea00018 of 39845864 bytes with 4752 elements
[1650465629.829143] [ndv4:55152:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x3ff6010
[1650465629.829175] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[13]=0x3ff6010 using dc_mlx5/mlx5_ib1:1 on worker 0x357a8c0
[1650465629.829339] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.829349] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.829526] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.829531] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.830772] [ndv4:55152:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465629.830400] [ndv4:55432:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465629.830408] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465629.832092] [ndv4:55152:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465629.832538] [ndv4:55152:0] ib_iface.c:994 UCX DEBUG iface=0x3ef5500: created UD QP 0xda7c on mlx5_ib1:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465629.833229] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465629.833642] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.833648] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.833878] [ndv4:55432:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465629.833884] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465629.835161] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x25e3320: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650465629.835180] [ndv4:54756:0] 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> }
[1650465629.835749] [ndv4:54756:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465629.835767] [ndv4:54756:0] async.c:228 UCX DEBUG added async handler 0x25d4300 [id=89 ref 1] uct_ud_iface_async_handler() to hash
[1650465629.835798] [ndv4:54756:0] async.c:506 UCX DEBUG listening to async event fd 89 events 0x5 mode thread_spinlock
[1650465629.835829] [ndv4:54756:0] ucp_worker.c:1159 UCX DEBUG created interface[10]=0x25e3320 using ud_mlx5/mlx5_ib0:1 on worker 0x2e9a8d0
[1650465629.836009] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.836017] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.836291] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.836297] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.836972] [ndv4:55432:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465629.836978] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465629.837192] [ndv4:55432:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.838031] [ndv4:55691:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib3 vendor_id: 0x15b3 device_id: 4124
[1650465629.838254] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465629.838447] [ndv4:55691:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib3: disable ODP because it's not supported for DevX QP
[1650465629.839268] [ndv4:55691:0] async.c:228 UCX DEBUG added async handler 0x23bb280 [id=63 ref 1] uct_ib_async_event_handler() to hash
[1650465629.839300] [ndv4:55691:0] async.c:506 UCX DEBUG listening to async event fd 63 events 0x1 mode thread_spinlock
[1650465629.839303] [ndv4:55691:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib3' (InfiniBand channel adapter) with 1 ports
[1650465629.839396] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465629.839413] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.839416] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.839476] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.839879] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465629.839886] [ndv4:54965:0] 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> }
[1650465629.839918] [ndv4:54965:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib4 length=2048) failed: Invalid argument
[1650465629.839922] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465629.839929] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x3174580 [id=115 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465629.839952] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 115 events 0x1 mode thread_spinlock
[1650465629.840672] [ndv4:54965:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465629.839482] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465629.839492] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465629.839497] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.839539] [ndv4:55691:0] ib_md.c:1319 UCX DEBUG mlx5_ib3: using registration cache
[1650465629.841974] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.841993] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.842477] [ndv4:55691:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.842485] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.842747] [ndv4:55691:0] ib_md.c:1604 UCX DEBUG mlx5_ib3: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.843069] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465629.843076] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465629.842410] [ndv4:55152:0] ib_md.c:812 UCX DEBUG registered memory 0x2b3b979fc000..0x2b3b97a81000 on mlx5_ib1 lkey 0x80a00 rkey 0x80a00 access 0xf flags 0x3e4
[1650465629.842418] [ndv4:55152:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b3b979fc018 of 544744 bytes with 128 elements
[1650465629.842425] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465629.843259] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x3ef5500: adding gid fe80::15:5dff:fd33:fffc to hash on device mlx5_ib1 port 1 index 0)
[1650465629.843868] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x3ef5500: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 1)
[1650465629.844307] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x3ef5500: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 2)
[1650465629.844849] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x3ef5500: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 3)
[1650465629.845316] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x3ef5500: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 4)
[1650465629.845687] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x3ef5500: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 5)
[1650465629.845863] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x3ef5500: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 6)
[1650465629.845974] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x3ef5500: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 7)
[1650465629.845239] [ndv4:55691:0] topo.c:99 UCX DEBUG bus id 0x104000000 doesn't exist. sys_dev = 3
[1650465629.845245] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465629.846271] [ndv4:55152:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465629.846282] [ndv4:55152:0] async.c:228 UCX DEBUG added async handler 0x2c9f930 [id=95 ref 1] uct_ud_iface_async_handler() to hash
[1650465629.846317] [ndv4:55152:0] async.c:506 UCX DEBUG listening to async event fd 95 events 0x5 mode thread_spinlock
[1650465629.846346] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[14]=0x3ef5500 using ud_verbs/mlx5_ib1:1 on worker 0x357a8c0
[1650465629.846371] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.846381] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.846444] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.846449] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.846678] [ndv4:55152:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465629.847499] [ndv4:55152:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465629.847922] [ndv4:55152:0] ib_iface.c:994 UCX DEBUG iface=0x3dbc280: created UD QP 0xda7e on mlx5_ib1:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465629.847930] [ndv4:55152:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465629.848287] [ndv4:54756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465629.848464] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465629.848917] [ndv4:54965:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x43a4050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xdad8
[1650465629.849155] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465629.849168] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465629.849181] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b4b31b49008 of 151544 bytes with 1052 elements
[1650465629.849867] [ndv4:55512:0] mm_posix.c:518 UCX DEBUG allocated posix shared memory at 0x2aea6ff3a000 length 12288
[1650465629.849930] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool mm_recv_desc: align 64, maxelems 4294967295, elemsize 8368
[1650465629.851153] [ndv4:55512:0] mm_posix.c:305 UCX DEBUG shared memory mmap(addr=(nil), length=6291456, flags= HUGETLB, fd=76) failed: Invalid argument
[1650465629.851162] [ndv4:55512:0] mm_posix.c:518 UCX DEBUG allocated posix shared memory at 0x2aea74000000 length 4296704
[1650465629.851168] [ndv4:55512:0] mpool.c:205 UCX DEBUG mpool mm_recv_desc: allocated chunk 0x2aea74000018 of 4296680 bytes with 512 elements
[1650465629.851427] [ndv4:55512:0] mm_iface.c:600 UCX DEBUG created mm iface 0x19efd10 FIFO id 0x4000000050eb928a va 0x2aea6ff3a000 size 12288 (128 x 64 elems)
[1650465629.851474] [ndv4:55512:0] ucp_worker.c:1159 UCX DEBUG created interface[0]=0x19efd10 using posix/memory on worker 0x22c98d0
[1650465629.851497] [ndv4:55512:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 8447 bytes with hugetlb
[1650465629.851533] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool mm_recv_desc: align 64, maxelems 4294967295, elemsize 8368
[1650465629.851547] [ndv4:55512:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 4292720 bytes with hugetlb
[1650465629.851556] [ndv4:55512:0] mpool.c:205 UCX DEBUG mpool mm_recv_desc: allocated chunk 0x2aea74419018 of 4296680 bytes with 512 elements
[1650465629.849263] [ndv4:54756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465629.849284] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.849288] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.849312] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.849662] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465629.849673] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465629.850102] [ndv4:54756:0] ib_iface.c:994 UCX DEBUG iface=0x376b050: created RC QP 0xda7f on mlx5_ib1:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465629.850747] [ndv4:54756:0] ucp_worker.c:1159 UCX DEBUG created interface[11]=0x376b050 using rc_verbs/mlx5_ib1:1 on worker 0x2e9a8d0
[1650465629.852235] [ndv4:55512:0] mm_iface.c:600 UCX DEBUG created mm iface 0x19f02e0 FIFO id 0x718024 va 0x2aea6ff3d000 size 12288 (128 x 64 elems)
[1650465629.852246] [ndv4:55512:0] ucp_worker.c:1159 UCX DEBUG created interface[1]=0x19f02e0 using sysv/memory on worker 0x22c98d0
[1650465629.852258] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool self_msg_desc: align 64, maxelems 4294967295, elemsize 8200
[1650465629.852261] [ndv4:55512:0] self.c:220 UCX DEBUG created self iface id 0x74b2d79bf49b778b send_size 8192
[1650465629.852267] [ndv4:55512:0] ucp_worker.c:1159 UCX DEBUG created interface[2]=0x1a03ec0 using self/memory0 on worker 0x22c98d0
[1650465629.852290] [ndv4:55512:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465629.852296] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465629.852298] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465629.853967] [ndv4:54965:0] ib_md.c:812 UCX DEBUG registered memory 0x2b4b2f400000..0x2b4b31a00000 on mlx5_ib4 lkey 0x80600 rkey 0x80600 access 0xf flags 0x3e4
[1650465629.853994] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b4b2f400018 of 39845864 bytes with 4752 elements
[1650465629.854175] [ndv4:54965:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x43a4050
[1650465629.854211] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[28]=0x43a4050 using dc_mlx5/mlx5_ib4:1 on worker 0x23dc840
[1650465629.854335] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465629.854348] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465629.854413] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465629.854417] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465629.854007] [ndv4:54861:0] debug.c:1198 UCX DEBUG using signal stack 0x2ae884b7c000 size 141824
[1650465629.854086] [ndv4:54861:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465629.854109] [ndv4:54861:0] 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 0x2ae8849d0000
[1650465629.854126] [ndv4:54861:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650465629.854135] [ndv4:54861:0] 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
[1650465629.854141] [ndv4:54861:0] module.c:253 UCX DEBUG loading modules for ucs
[1650465629.855426] [ndv4:55512:0] async.c:228 UCX DEBUG added async handler 0x1a06000 [id=78 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465629.855461] [ndv4:55512:0] async.c:506 UCX DEBUG listening to async event fd 78 events 0x5 mode thread_spinlock
[1650465629.855479] [ndv4:55512:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x1a04820: listening for connections (fd=78) on 10.5.0.5:52013
[1650465629.855706] [ndv4:55512:0] ucp_worker.c:1159 UCX DEBUG created interface[3]=0x1a04820 using tcp/eth0 on worker 0x22c98d0
[1650465629.855728] [ndv4:55512:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465629.855732] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465629.855735] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465629.855776] [ndv4:55512:0] async.c:228 UCX DEBUG added async handler 0x1952250 [id=80 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465629.855801] [ndv4:55512:0] async.c:506 UCX DEBUG listening to async event fd 80 events 0x5 mode thread_spinlock
[1650465629.855805] [ndv4:55512:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x1a04ea0: listening for connections (fd=80) on 127.0.0.1:41660
[1650465629.855821] [ndv4:55512:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465629.855826] [ndv4:55512:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465629.855885] [ndv4:55512:0] ucp_worker.c:1159 UCX DEBUG created interface[4]=0x1a04ea0 using tcp/lo on worker 0x22c98d0
[1650465629.855901] [ndv4:55512:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465629.855904] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465629.855907] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465629.855946] [ndv4:55512:0] async.c:228 UCX DEBUG added async handler 0x19605b0 [id=82 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465629.855969] [ndv4:55512:0] async.c:506 UCX DEBUG listening to async event fd 82 events 0x5 mode thread_spinlock
[1650465629.855973] [ndv4:55512:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x19ec530: listening for connections (fd=82) on 172.16.1.242:43172
[1650465629.856203] [ndv4:55512:0] ucp_worker.c:1159 UCX DEBUG created interface[5]=0x19ec530 using tcp/ib0 on worker 0x22c98d0
[1650465629.856413] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.856422] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.856650] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.856655] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.857649] [ndv4:55512:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465629.856838] [ndv4:54861:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465629.856861] [ndv4:54861:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650465629.856899] [ndv4:54861:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650465629.856903] [ndv4:54861:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650465629.856910] [ndv4:54861:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650465629.856917] [ndv4:54861:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650465629.856919] [ndv4:54861:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650465629.856925] [ndv4:54861:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650465629.856927] [ndv4:54861:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650465629.856930] [ndv4:54861:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650465629.856933] [ndv4:54861:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650465629.856935] [ndv4:54861:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650465629.856944] [ndv4:54861:0] module.c:253 UCX DEBUG loading modules for uct
[1650465629.857706] [ndv4:54861:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650465629.858940] [ndv4:54861:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650465629.858958] [ndv4:54861:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650465629.858969] [ndv4:54861:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650465629.858979] [ndv4:54861:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650465629.858990] [ndv4:54861:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650465629.859001] [ndv4:54861:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650465629.859013] [ndv4:54861:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650465629.858944] [ndv4:55512:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465629.858984] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.858988] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.859000] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.859354] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465629.859365] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465629.859487] [ndv4:55691:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465629.859496] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465629.859508] [ndv4:54861:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650465629.860017] [ndv4:55512:0] ib_iface.c:994 UCX DEBUG iface=0x19fa670: created RC QP 0xf49a on mlx5_ib0:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465629.859859] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.859872] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.860055] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.860071] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.860488] [ndv4:55152:0] ib_md.c:812 UCX DEBUG registered memory 0x2b3b97a81000..0x2b3b97b06000 on mlx5_ib1 lkey 0x80b00 rkey 0x80b00 access 0xf flags 0x3e4
[1650465629.860494] [ndv4:55152:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b3b97a81018 of 544744 bytes with 128 elements
[1650465629.860499] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465629.860892] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x3dbc280: adding gid fe80::15:5dff:fd33:fffc to hash on device mlx5_ib1 port 1 index 0)
[1650465629.861193] [ndv4:55691:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465629.861199] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465629.861636] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.861649] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.861781] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.861787] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.862012] [ndv4:54756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465629.863050] [ndv4:54756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465629.863061] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.863064] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.863083] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.863436] [ndv4:54756:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x377e010 of 8176 bytes with 127 elements
[1650465629.863671] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465629.863680] [ndv4:54756:0] 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> }
[1650465629.863715] [ndv4:54756:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650465629.863719] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465629.863731] [ndv4:54756:0] async.c:228 UCX DEBUG added async handler 0x25c4190 [id=92 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465629.863755] [ndv4:54756:0] async.c:506 UCX DEBUG listening to async event fd 92 events 0x1 mode thread_spinlock
[1650465629.863765] [ndv4:54756:0] ucp_worker.c:1159 UCX DEBUG created interface[12]=0x35ec010 using rc_mlx5/mlx5_ib1:1 on worker 0x2e9a8d0
[1650465629.863815] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.863821] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.863895] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.863901] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.864465] [ndv4:55324:0] debug.c:1198 UCX DEBUG using signal stack 0x2b6f65e62000 size 141824
[1650465629.864539] [ndv4:55324:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465629.864562] [ndv4:55324:0] 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 0x2b6f65cd0000
[1650465629.864655] [ndv4:55324:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650465629.864666] [ndv4:55324:0] 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
[1650465629.864672] [ndv4:55324:0] module.c:253 UCX DEBUG loading modules for ucs
[1650465629.864443] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465629.865707] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465629.866023] [ndv4:54965:0] ib_iface.c:994 UCX DEBUG iface=0x4580250: created UD QP 0xdace on mlx5_ib4:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465629.866616] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465629.866856] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465629.866864] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465629.866981] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465629.866986] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465629.867565] [ndv4:54965:0] ib_md.c:812 UCX DEBUG registered memory 0x2b4b31b6e000..0x2b4b31bf3000 on mlx5_ib4 lkey 0x80700 rkey 0x80700 access 0xf flags 0x3e4
[1650465629.867572] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b4b31b6e018 of 544744 bytes with 128 elements
[1650465629.867612] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465629.867933] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x4580250: adding gid fe80::15:5dff:fd33:ffff to hash on device mlx5_ib4 port 1 index 0)
[1650465629.868116] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x4580250: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 1)
[1650465629.868337] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x4580250: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 2)
[1650465629.868937] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x4580250: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 3)
[1650465629.869275] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x4580250: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 4)
[1650465629.867167] [ndv4:55324:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465629.867186] [ndv4:55324:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650465629.867223] [ndv4:55324:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650465629.867227] [ndv4:55324:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650465629.867234] [ndv4:55324:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650465629.867240] [ndv4:55324:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650465629.867243] [ndv4:55324:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650465629.867248] [ndv4:55324:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650465629.867251] [ndv4:55324:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650465629.867254] [ndv4:55324:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650465629.867256] [ndv4:55324:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650465629.867258] [ndv4:55324:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650465629.867266] [ndv4:55324:0] module.c:253 UCX DEBUG loading modules for uct
[1650465629.868014] [ndv4:55324:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650465629.868742] [ndv4:55324:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650465629.868758] [ndv4:55324:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650465629.868769] [ndv4:55324:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650465629.868780] [ndv4:55324:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650465629.868790] [ndv4:55324:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650465629.868799] [ndv4:55324:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650465629.868810] [ndv4:55324:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650465629.869183] [ndv4:55324:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650465629.869387] [ndv4:55432:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib3 vendor_id: 0x15b3 device_id: 4124
[1650465629.869092] [ndv4:55516:0] mm_posix.c:518 UCX DEBUG allocated posix shared memory at 0x2ae24bd37000 length 12288
[1650465629.869153] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool mm_recv_desc: align 64, maxelems 4294967295, elemsize 8368
[1650465629.869858] [ndv4:55432:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib3: disable ODP because it's not supported for DevX QP
[1650465629.870979] [ndv4:55432:0] async.c:228 UCX DEBUG added async handler 0x2134ec0 [id=63 ref 1] uct_ib_async_event_handler() to hash
[1650465629.871001] [ndv4:55432:0] async.c:506 UCX DEBUG listening to async event fd 63 events 0x1 mode thread_spinlock
[1650465629.871004] [ndv4:55432:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib3' (InfiniBand channel adapter) with 1 ports
[1650465629.871308] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465629.871317] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465629.871322] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.871364] [ndv4:55432:0] ib_md.c:1319 UCX DEBUG mlx5_ib3: using registration cache
[1650465629.870346] [ndv4:55516:0] mm_posix.c:305 UCX DEBUG shared memory mmap(addr=(nil), length=6291456, flags= HUGETLB, fd=76) failed: Invalid argument
[1650465629.870356] [ndv4:55516:0] mm_posix.c:518 UCX DEBUG allocated posix shared memory at 0x2ae250423000 length 4296704
[1650465629.870361] [ndv4:55516:0] mpool.c:205 UCX DEBUG mpool mm_recv_desc: allocated chunk 0x2ae250423018 of 4296680 bytes with 512 elements
[1650465629.870685] [ndv4:55516:0] mm_iface.c:600 UCX DEBUG created mm iface 0x19d9d10 FIFO id 0x400000007110a1e4 va 0x2ae24bd37000 size 12288 (128 x 64 elems)
[1650465629.870741] [ndv4:55516:0] ucp_worker.c:1159 UCX DEBUG created interface[0]=0x19d9d10 using posix/memory on worker 0x22b38d0
[1650465629.870766] [ndv4:55516:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 8447 bytes with hugetlb
[1650465629.870801] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool mm_recv_desc: align 64, maxelems 4294967295, elemsize 8368
[1650465629.870815] [ndv4:55516:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 4292720 bytes with hugetlb
[1650465629.870825] [ndv4:55516:0] mpool.c:205 UCX DEBUG mpool mm_recv_desc: allocated chunk 0x2ae25083c018 of 4296680 bytes with 512 elements
[1650465629.871445] [ndv4:55516:0] mm_iface.c:600 UCX DEBUG created mm iface 0x19da2e0 FIFO id 0x718028 va 0x2ae24bd3a000 size 12288 (128 x 64 elems)
[1650465629.871454] [ndv4:55516:0] ucp_worker.c:1159 UCX DEBUG created interface[1]=0x19da2e0 using sysv/memory on worker 0x22b38d0
[1650465629.871467] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool self_msg_desc: align 64, maxelems 4294967295, elemsize 8200
[1650465629.871470] [ndv4:55516:0] self.c:220 UCX DEBUG created self iface id 0xafb26d37f61159f8 send_size 8192
[1650465629.871476] [ndv4:55516:0] ucp_worker.c:1159 UCX DEBUG created interface[2]=0x19edec0 using self/memory0 on worker 0x22b38d0
[1650465629.871499] [ndv4:55516:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465629.871537] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465629.871540] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465629.871896] [ndv4:55512:0] ucp_worker.c:1159 UCX DEBUG created interface[6]=0x19fa670 using rc_verbs/mlx5_ib0:1 on worker 0x22c98d0
[1650465629.871954] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.871962] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.872115] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.872120] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.872766] [ndv4:55512:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465629.873075] [ndv4:55512:0] ib_device.c:1394 UCX DEBUG max IB CQE size is 128
[1650465629.872851] [ndv4:55691:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465629.872859] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465629.873459] [ndv4:55691:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465629.873463] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465629.874102] [ndv4:55512:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465629.874111] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.874115] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.874171] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.874669] [ndv4:55512:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x2393010 of 8176 bytes with 127 elements
[1650465629.875002] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465629.875025] [ndv4:55512:0] 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> }
[1650465629.875069] [ndv4:55512:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650465629.875074] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465629.874311] [ndv4:54756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465629.875671] [ndv4:54756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465629.875690] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.875693] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.875751] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.876096] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465629.876103] [ndv4:54756:0] 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> }
[1650465629.876134] [ndv4:54756:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650465629.876137] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465629.876145] [ndv4:54756:0] async.c:228 UCX DEBUG added async handler 0x25c4fb0 [id=94 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465629.876164] [ndv4:54756:0] async.c:506 UCX DEBUG listening to async event fd 94 events 0x1 mode thread_spinlock
[1650465629.876648] [ndv4:54756:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465629.873750] [ndv4:55691:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.875172] [ndv4:55516:0] async.c:228 UCX DEBUG added async handler 0x19f0000 [id=78 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465629.875203] [ndv4:55516:0] async.c:506 UCX DEBUG listening to async event fd 78 events 0x5 mode thread_spinlock
[1650465629.875225] [ndv4:55516:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x19ee820: listening for connections (fd=78) on 10.5.0.5:47929
[1650465629.875368] [ndv4:55516:0] ucp_worker.c:1159 UCX DEBUG created interface[3]=0x19ee820 using tcp/eth0 on worker 0x22b38d0
[1650465629.875388] [ndv4:55516:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465629.875392] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465629.875395] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465629.875434] [ndv4:55516:0] async.c:228 UCX DEBUG added async handler 0x193c250 [id=80 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465629.875458] [ndv4:55516:0] async.c:506 UCX DEBUG listening to async event fd 80 events 0x5 mode thread_spinlock
[1650465629.875487] [ndv4:55516:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x19eeea0: listening for connections (fd=80) on 127.0.0.1:33777
[1650465629.875504] [ndv4:55516:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465629.875509] [ndv4:55516:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465629.875608] [ndv4:55516:0] ucp_worker.c:1159 UCX DEBUG created interface[4]=0x19eeea0 using tcp/lo on worker 0x22b38d0
[1650465629.875628] [ndv4:55516:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465629.875631] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465629.875634] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465629.875675] [ndv4:55516:0] async.c:228 UCX DEBUG added async handler 0x194a5b0 [id=82 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465629.875699] [ndv4:55516:0] async.c:506 UCX DEBUG listening to async event fd 82 events 0x5 mode thread_spinlock
[1650465629.875703] [ndv4:55516:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x19d6530: listening for connections (fd=82) on 172.16.1.242:44411
[1650465629.875897] [ndv4:55516:0] ucp_worker.c:1159 UCX DEBUG created interface[5]=0x19d6530 using tcp/ib0 on worker 0x22b38d0
[1650465629.876738] [ndv4:55324:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650465629.877130] [ndv4:55324:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650465629.878071] [ndv4:54861:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650465629.878452] [ndv4:54861:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650465629.881837] [ndv4:55324:0] async.c:228 UCX DEBUG added async handler 0x16372a0 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650465629.881938] [ndv4:55324:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650465629.881944] [ndv4:55324:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650465629.882089] [ndv4:55512:0] async.c:228 UCX DEBUG added async handler 0x195e710 [id=85 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465629.882118] [ndv4:55512:0] async.c:506 UCX DEBUG listening to async event fd 85 events 0x1 mode thread_spinlock
[1650465629.882141] [ndv4:55512:0] ucp_worker.c:1159 UCX DEBUG created interface[7]=0x1a080c0 using rc_mlx5/mlx5_ib0:1 on worker 0x22c98d0
[1650465629.882205] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.882211] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.882285] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.882289] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.882777] [ndv4:55512:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465629.882182] [ndv4:55324:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.882193] [ndv4:55324:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.882213] [ndv4:55324:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.882273] [ndv4:55324:0] module.c:253 UCX DEBUG loading modules for ucm
[1650465629.882295] [ndv4:55324:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650465629.883420] [ndv4:55324:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.883434] [ndv4:55324:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.883938] [ndv4:55512:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465629.883948] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.883952] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.883965] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.884322] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465629.884329] [ndv4:55512:0] 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> }
[1650465629.884360] [ndv4:55512:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650465629.884364] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465629.884374] [ndv4:55512:0] async.c:228 UCX DEBUG added async handler 0x1282ab0 [id=87 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465629.884399] [ndv4:55512:0] async.c:506 UCX DEBUG listening to async event fd 87 events 0x1 mode thread_spinlock
[1650465629.885146] [ndv4:55512:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465629.883316] [ndv4:54861:0] async.c:228 UCX DEBUG added async handler 0x18c5360 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650465629.883413] [ndv4:54861:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650465629.883421] [ndv4:54861:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650465629.883819] [ndv4:54861:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.883829] [ndv4:54861:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.883852] [ndv4:54861:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.883912] [ndv4:54861:0] module.c:253 UCX DEBUG loading modules for ucm
[1650465629.883934] [ndv4:54861:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650465629.884328] [ndv4:55432:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.884338] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.884555] [ndv4:55432:0] ib_md.c:1604 UCX DEBUG mlx5_ib3: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.884582] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465629.884589] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465629.885107] [ndv4:55432:0] topo.c:99 UCX DEBUG bus id 0x104000000 doesn't exist. sys_dev = 3
[1650465629.885112] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465629.885519] [ndv4:55432:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465629.885523] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465629.883022] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x3dbc280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 1)
[1650465629.883376] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x3dbc280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 2)
[1650465629.883444] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x3dbc280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 3)
[1650465629.883461] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x3dbc280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 4)
[1650465629.883475] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x3dbc280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 5)
[1650465629.883705] [ndv4:55324:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.883830] [ndv4:55324:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.883836] [ndv4:55324:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.882981] [ndv4:54756:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x3916010: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xdaa8
[1650465629.883362] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.883369] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.883379] [ndv4:54756:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b3b53e5a008 of 151544 bytes with 1052 elements
[1650465629.886418] [ndv4:55432:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465629.886424] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465629.886907] [ndv4:55432:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465629.886912] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465629.887071] [ndv4:54756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b3b5b200000..0x2b3b5d800000 on mlx5_ib1 lkey 0x80c00 rkey 0x80c00 access 0xf flags 0x3e4
[1650465629.887086] [ndv4:54756:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b3b5b200018 of 39845864 bytes with 4752 elements
[1650465629.887250] [ndv4:54756:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x3916010
[1650465629.887270] [ndv4:54756:0] ucp_worker.c:1159 UCX DEBUG created interface[13]=0x3916010 using dc_mlx5/mlx5_ib1:1 on worker 0x2e9a8d0
[1650465629.887344] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.887351] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.887413] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.887419] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.886984] [ndv4:54715:0] debug.c:1198 UCX DEBUG using signal stack 0x2ac8ab71e000 size 141824
[1650465629.887058] [ndv4:54715:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465629.887079] [ndv4:54715:0] 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 0x2ac8ab58c000
[1650465629.887101] [ndv4:54715:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650465629.887110] [ndv4:54715:0] 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
[1650465629.887116] [ndv4:54715:0] module.c:253 UCX DEBUG loading modules for ucs
[1650465629.887956] [ndv4:54756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465629.888875] [ndv4:54756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465629.889219] [ndv4:54756:0] ib_iface.c:994 UCX DEBUG iface=0x3815510: created UD QP 0xda89 on mlx5_ib1:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465629.889669] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465629.889729] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.889739] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.889761] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.889765] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.890078] [ndv4:54756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b3b53e7f000..0x2b3b53f04000 on mlx5_ib1 lkey 0x80d00 rkey 0x80d00 access 0xf flags 0x3e4
[1650465629.890086] [ndv4:54756:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b3b53e7f018 of 544744 bytes with 128 elements
[1650465629.890093] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465629.890134] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x3815510: adding gid fe80::15:5dff:fd33:fffc to hash on device mlx5_ib1 port 1 index 0)
[1650465629.890182] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x3815510: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 1)
[1650465629.890232] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x3815510: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 2)
[1650465629.890285] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x3815510: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 3)
[1650465629.890316] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x3815510: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 4)
[1650465629.890340] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x3815510: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 5)
[1650465629.890354] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x3815510: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 6)
[1650465629.890369] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x3815510: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 7)
[1650465629.890538] [ndv4:54756:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465629.890548] [ndv4:54756:0] async.c:228 UCX DEBUG added async handler 0x25bf940 [id=95 ref 1] uct_ud_iface_async_handler() to hash
[1650465629.890590] [ndv4:54756:0] async.c:506 UCX DEBUG listening to async event fd 95 events 0x5 mode thread_spinlock
[1650465629.890615] [ndv4:54756:0] ucp_worker.c:1159 UCX DEBUG created interface[14]=0x3815510 using ud_verbs/mlx5_ib1:1 on worker 0x2e9a8d0
[1650465629.890627] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.890631] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.890686] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.890691] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.889592] [ndv4:54715:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465629.889610] [ndv4:54715:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650465629.889642] [ndv4:54715:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650465629.889645] [ndv4:54715:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650465629.889653] [ndv4:54715:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650465629.889659] [ndv4:54715:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650465629.889662] [ndv4:54715:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650465629.889666] [ndv4:54715:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650465629.889669] [ndv4:54715:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650465629.889671] [ndv4:54715:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650465629.889673] [ndv4:54715:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650465629.889675] [ndv4:54715:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650465629.889685] [ndv4:54715:0] module.c:253 UCX DEBUG loading modules for uct
[1650465629.890386] [ndv4:54715:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650465629.890961] [ndv4:54715:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650465629.890978] [ndv4:54715:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650465629.890989] [ndv4:54715:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650465629.891000] [ndv4:54715:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650465629.891010] [ndv4:54715:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650465629.891020] [ndv4:54715:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650465629.891031] [ndv4:54715:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650465629.891402] [ndv4:54715:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650465629.890919] [ndv4:54756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465629.891833] [ndv4:55512:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x25fd010: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xf4c4
[1650465629.892030] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.892038] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.892056] [ndv4:55512:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2aea6ff42008 of 151544 bytes with 1052 elements
[1650465629.891680] [ndv4:54756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465629.891965] [ndv4:54756:0] ib_iface.c:994 UCX DEBUG iface=0x3392010: created UD QP 0xda8c on mlx5_ib1:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465629.891971] [ndv4:54756:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465629.892353] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465629.892785] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.892792] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.892893] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.892897] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.893289] [ndv4:54756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b3b53f04000..0x2b3b53f89000 on mlx5_ib1 lkey 0x80e00 rkey 0x80e00 access 0xf flags 0x3e4
[1650465629.893297] [ndv4:54756:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b3b53f04018 of 544744 bytes with 128 elements
[1650465629.893303] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465629.893906] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x3392010: adding gid fe80::15:5dff:fd33:fffc to hash on device mlx5_ib1 port 1 index 0)
[1650465629.894390] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x3392010: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 1)
[1650465629.891523] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.891538] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.891976] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.891985] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.892387] [ndv4:55516:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465629.893407] [ndv4:55516:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465629.893445] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.893449] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.893462] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.893717] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465629.893726] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465629.894201] [ndv4:55516:0] ib_iface.c:994 UCX DEBUG iface=0x19e4670: created RC QP 0xf4a7 on mlx5_ib0:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465629.895761] [ndv4:55512:0] ib_md.c:812 UCX DEBUG registered memory 0x2aea74a00000..0x2aea77000000 on mlx5_ib0 lkey 0x81300 rkey 0x81300 access 0xf flags 0x3e4
[1650465629.895778] [ndv4:55512:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2aea74a00018 of 39845864 bytes with 4752 elements
[1650465629.895913] [ndv4:55512:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x25fd010
[1650465629.895946] [ndv4:55512:0] ucp_worker.c:1159 UCX DEBUG created interface[8]=0x25fd010 using dc_mlx5/mlx5_ib0:1 on worker 0x22c98d0
[1650465629.896007] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.896017] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.896087] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.896092] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.896314] [ndv4:55512:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465629.896978] [ndv4:55512:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465629.896924] [ndv4:55691:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib4 vendor_id: 0x15b3 device_id: 4124
[1650465629.897060] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x4580250: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 5)
[1650465629.897173] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x4580250: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 6)
[1650465629.897306] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x4580250: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 7)
[1650465629.897345] [ndv4:55512:0] ib_iface.c:994 UCX DEBUG iface=0x1a10a30: created UD QP 0xf4a8 on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465629.897706] [ndv4:54965:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465629.897713] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x38913f0 [id=116 ref 1] uct_ud_iface_async_handler() to hash
[1650465629.897737] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 116 events 0x5 mode thread_spinlock
[1650465629.897755] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[29]=0x4580250 using ud_verbs/mlx5_ib4:1 on worker 0x23dc840
[1650465629.897847] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465629.897854] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465629.897916] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465629.897921] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465629.897969] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465629.898187] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.898195] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.898241] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.898245] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.898642] [ndv4:55512:0] ib_md.c:812 UCX DEBUG registered memory 0x2aea6ff67000..0x2aea6ffec000 on mlx5_ib0 lkey 0x81400 rkey 0x81400 access 0xf flags 0x3e4
[1650465629.898649] [ndv4:55512:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2aea6ff67018 of 544744 bytes with 128 elements
[1650465629.898653] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465629.898858] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465629.898835] [ndv4:55691:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib4: disable ODP because it's not supported for DevX QP
[1650465629.899604] [ndv4:55691:0] async.c:228 UCX DEBUG added async handler 0x23c30b0 [id=65 ref 1] uct_ib_async_event_handler() to hash
[1650465629.899634] [ndv4:55691:0] async.c:506 UCX DEBUG listening to async event fd 65 events 0x1 mode thread_spinlock
[1650465629.899638] [ndv4:55691:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib4' (InfiniBand channel adapter) with 1 ports
[1650465629.899886] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465629.899895] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465629.899900] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.899947] [ndv4:55691:0] ib_md.c:1319 UCX DEBUG mlx5_ib4: using registration cache
[1650465629.901212] [ndv4:55691:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.901218] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.900004] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465629.900506] [ndv4:54965:0] ib_iface.c:994 UCX DEBUG iface=0x469e050: created UD QP 0xdacf on mlx5_ib4:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465629.900514] [ndv4:54965:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465629.900644] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x3dbc280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 6)
[1650465629.901158] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x3dbc280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 7)
[1650465629.901168] [ndv4:55152:0] 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> }
[1650465629.901254] [ndv4:55152:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465629.901263] [ndv4:55152:0] async.c:228 UCX DEBUG added async handler 0x2cb6b10 [id=96 ref 1] uct_ud_iface_async_handler() to hash
[1650465629.901295] [ndv4:55152:0] async.c:506 UCX DEBUG listening to async event fd 96 events 0x5 mode thread_spinlock
[1650465629.901317] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[15]=0x3dbc280 using ud_mlx5/mlx5_ib1:1 on worker 0x357a8c0
[1650465629.901336] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.901342] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.901404] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.901409] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.901594] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465629.901626] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465629.901631] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465629.901644] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465629.901649] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465629.902219] [ndv4:54965:0] ib_md.c:812 UCX DEBUG registered memory 0x2b4b31bf3000..0x2b4b31c78000 on mlx5_ib4 lkey 0x80800 rkey 0x80800 access 0xf flags 0x3e4
[1650465629.902225] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b4b31bf3018 of 544744 bytes with 128 elements
[1650465629.902230] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465629.902607] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x469e050: adding gid fe80::15:5dff:fd33:ffff to hash on device mlx5_ib4 port 1 index 0)
[1650465629.901508] [ndv4:55324:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650465629.901516] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.901793] [ndv4:55324:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465629.901798] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.901568] [ndv4:55691:0] ib_md.c:1604 UCX DEBUG mlx5_ib4: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.901602] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465629.901608] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465629.902085] [ndv4:55691:0] topo.c:99 UCX DEBUG bus id 0x105000000 doesn't exist. sys_dev = 4
[1650465629.902090] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465629.901873] [ndv4:55152:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465629.902055] [ndv4:54861:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.902068] [ndv4:54861:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.902261] [ndv4:54861:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.902443] [ndv4:54861:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.902451] [ndv4:54861:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.903038] [ndv4:55152:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465629.903054] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.903057] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.903107] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.902791] [ndv4:55432:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465629.902798] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465629.903016] [ndv4:55432:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.902731] [ndv4:55324:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465629.902737] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.903155] [ndv4:55324:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465629.903159] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.903703] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465629.903709] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465629.904463] [ndv4:55152:0] ib_iface.c:994 UCX DEBUG iface=0x43e50a0: created RC QP 0xdad9 on mlx5_ib2:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465629.905475] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[16]=0x43e50a0 using rc_verbs/mlx5_ib2:1 on worker 0x357a8c0
[1650465629.905524] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.905529] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.905611] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.905615] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.905812] [ndv4:55152:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465629.906420] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x3392010: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 2)
[1650465629.907012] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x3392010: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 3)
[1650465629.907391] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x3392010: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 4)
[1650465629.907075] [ndv4:55152:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465629.907089] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.907093] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.907143] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.907629] [ndv4:55152:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x4706010 of 8176 bytes with 127 elements
[1650465629.908945] [ndv4:55516:0] ucp_worker.c:1159 UCX DEBUG created interface[6]=0x19e4670 using rc_verbs/mlx5_ib0:1 on worker 0x22b38d0
[1650465629.909131] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.909139] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.909320] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.909326] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.907945] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465629.907952] [ndv4:55152:0] 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> }
[1650465629.907986] [ndv4:55152:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib2 length=2048) failed: Invalid argument
[1650465629.907991] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465629.908001] [ndv4:55152:0] async.c:228 UCX DEBUG added async handler 0x2ca4950 [id=99 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465629.908022] [ndv4:55152:0] async.c:506 UCX DEBUG listening to async event fd 99 events 0x1 mode thread_spinlock
[1650465629.908045] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[17]=0x4532030 using rc_mlx5/mlx5_ib2:1 on worker 0x357a8c0
[1650465629.908127] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.908133] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.908265] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.908271] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.909264] [ndv4:55152:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465629.909903] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x1a10a30: adding gid fe80::15:5dff:fd33:fffb to hash on device mlx5_ib0 port 1 index 0)
[1650465629.910559] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x1a10a30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650465629.910860] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x1a10a30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650465629.910875] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x1a10a30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650465629.910890] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x1a10a30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650465629.910459] [ndv4:55516:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465629.911119] [ndv4:55516:0] ib_device.c:1394 UCX DEBUG max IB CQE size is 128
[1650465629.909200] [ndv4:54715:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650465629.909692] [ndv4:54715:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650465629.912191] [ndv4:55516:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465629.912209] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.912213] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.912267] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.912560] [ndv4:55691:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465629.912661] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465629.912637] [ndv4:55516:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x237d010 of 8176 bytes with 127 elements
[1650465629.912868] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465629.912901] [ndv4:55516:0] 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> }
[1650465629.913070] [ndv4:55516:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650465629.913077] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465629.914113] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x469e050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 1)
[1650465629.914975] [ndv4:55324:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465629.914983] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.915199] [ndv4:55324:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.917050] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x3392010: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 5)
[1650465629.917404] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x3392010: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 6)
[1650465629.917677] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x3392010: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 7)
[1650465629.917684] [ndv4:54756:0] 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> }
[1650465629.917786] [ndv4:54756:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465629.917792] [ndv4:54756:0] async.c:228 UCX DEBUG added async handler 0x3392e80 [id=96 ref 1] uct_ud_iface_async_handler() to hash
[1650465629.917815] [ndv4:54756:0] async.c:506 UCX DEBUG listening to async event fd 96 events 0x5 mode thread_spinlock
[1650465629.917830] [ndv4:54756:0] ucp_worker.c:1159 UCX DEBUG created interface[15]=0x3392010 using ud_mlx5/mlx5_ib1:1 on worker 0x2e9a8d0
[1650465629.917902] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.917908] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.918034] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.918039] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.918244] [ndv4:54717:0] debug.c:1198 UCX DEBUG using signal stack 0x2b7cc1036000 size 141824
[1650465629.918318] [ndv4:54717:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465629.918338] [ndv4:54717:0] 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 0x2b7cc0e8a000
[1650465629.918357] [ndv4:54717:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650465629.918367] [ndv4:54717:0] 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
[1650465629.918373] [ndv4:54717:0] module.c:253 UCX DEBUG loading modules for ucs
[1650465629.918611] [ndv4:55432:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib4 vendor_id: 0x15b3 device_id: 4124
[1650465629.919001] [ndv4:55432:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib4: disable ODP because it's not supported for DevX QP
[1650465629.919524] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x1a10a30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650465629.919787] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x1a10a30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650465629.919804] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x1a10a30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650465629.920095] [ndv4:55152:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465629.920112] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.920115] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.920175] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.920644] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465629.920653] [ndv4:55152:0] 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> }
[1650465629.920692] [ndv4:55152:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib2 length=2048) failed: Invalid argument
[1650465629.920696] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465629.920703] [ndv4:55152:0] async.c:228 UCX DEBUG added async handler 0x3a87ef0 [id=101 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465629.920726] [ndv4:55152:0] async.c:506 UCX DEBUG listening to async event fd 101 events 0x1 mode thread_spinlock
[1650465629.920123] [ndv4:55512:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465629.920610] [ndv4:55432:0] async.c:228 UCX DEBUG added async handler 0x281dd70 [id=65 ref 1] uct_ib_async_event_handler() to hash
[1650465629.920643] [ndv4:55432:0] async.c:506 UCX DEBUG listening to async event fd 65 events 0x1 mode thread_spinlock
[1650465629.920647] [ndv4:55432:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib4' (InfiniBand channel adapter) with 1 ports
[1650465629.921358] [ndv4:55152:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465629.922288] [ndv4:54717:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465629.922313] [ndv4:54717:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650465629.922352] [ndv4:54717:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650465629.922356] [ndv4:54717:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650465629.922364] [ndv4:54717:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650465629.922373] [ndv4:54717:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650465629.922376] [ndv4:54717:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650465629.922382] [ndv4:54717:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650465629.922385] [ndv4:54717:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650465629.922387] [ndv4:54717:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650465629.922390] [ndv4:54717:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650465629.922392] [ndv4:54717:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650465629.922405] [ndv4:54717:0] module.c:253 UCX DEBUG loading modules for uct
[1650465629.923267] [ndv4:54717:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650465629.921803] [ndv4:54861:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650465629.921814] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.923832] [ndv4:54853:0] debug.c:1198 UCX DEBUG using signal stack 0x2b6769a40000 size 141824
[1650465629.923906] [ndv4:54853:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465629.923927] [ndv4:54853:0] 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 0x2b67698ae000
[1650465629.923949] [ndv4:54853:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650465629.923959] [ndv4:54853:0] 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
[1650465629.923965] [ndv4:54853:0] module.c:253 UCX DEBUG loading modules for ucs
[1650465629.923864] [ndv4:54717:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650465629.923880] [ndv4:54717:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650465629.923891] [ndv4:54717:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650465629.923901] [ndv4:54717:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650465629.923911] [ndv4:54717:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650465629.923921] [ndv4:54717:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650465629.923931] [ndv4:54717:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650465629.924311] [ndv4:54717:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650465629.924080] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x469e050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 2)
[1650465629.922774] [ndv4:55516:0] async.c:228 UCX DEBUG added async handler 0x1948710 [id=85 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465629.922806] [ndv4:55516:0] async.c:506 UCX DEBUG listening to async event fd 85 events 0x1 mode thread_spinlock
[1650465629.922845] [ndv4:55516:0] ucp_worker.c:1159 UCX DEBUG created interface[7]=0x19f20c0 using rc_mlx5/mlx5_ib0:1 on worker 0x22b38d0
[1650465629.923181] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.923188] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.923377] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.923383] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.924866] [ndv4:54715:0] async.c:228 UCX DEBUG added async handler 0x2006080 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650465629.924962] [ndv4:54715:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650465629.924970] [ndv4:54715:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650465629.925358] [ndv4:55324:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650465629.925520] [ndv4:54715:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.925532] [ndv4:54715:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.925555] [ndv4:54715:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.925656] [ndv4:54715:0] module.c:253 UCX DEBUG loading modules for ucm
[1650465629.925679] [ndv4:54715:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650465629.925773] [ndv4:55324:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650465629.926950] [ndv4:55512:0] async.c:228 UCX DEBUG added async handler 0x195fc50 [id=88 ref 1] uct_ud_iface_async_handler() to hash
[1650465629.926979] [ndv4:55512:0] async.c:506 UCX DEBUG listening to async event fd 88 events 0x5 mode thread_spinlock
[1650465629.927012] [ndv4:55512:0] ucp_worker.c:1159 UCX DEBUG created interface[9]=0x1a10a30 using ud_verbs/mlx5_ib0:1 on worker 0x22c98d0
[1650465629.927399] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.927407] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.927527] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.927531] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.926657] [ndv4:54853:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465629.926676] [ndv4:54853:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650465629.926711] [ndv4:54853:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650465629.926714] [ndv4:54853:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650465629.926721] [ndv4:54853:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650465629.926728] [ndv4:54853:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650465629.926733] [ndv4:54853:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650465629.926738] [ndv4:54853:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650465629.926740] [ndv4:54853:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650465629.926742] [ndv4:54853:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650465629.926745] [ndv4:54853:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650465629.926747] [ndv4:54853:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650465629.926756] [ndv4:54853:0] module.c:253 UCX DEBUG loading modules for uct
[1650465629.927465] [ndv4:54853:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650465629.927567] [ndv4:55324:0] async.c:228 UCX DEBUG added async handler 0x165d130 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650465629.927662] [ndv4:55324:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650465629.927666] [ndv4:55324:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650465629.928108] [ndv4:55324:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.928116] [ndv4:55324:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.928121] [ndv4:55324:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.928162] [ndv4:55324:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650465629.928324] [ndv4:54853:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650465629.928340] [ndv4:54853:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650465629.928351] [ndv4:54853:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650465629.928362] [ndv4:54853:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650465629.928372] [ndv4:54853:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650465629.928383] [ndv4:54853:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650465629.928395] [ndv4:54853:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650465629.928841] [ndv4:54853:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650465629.927794] [ndv4:55512:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465629.928918] [ndv4:55512:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465629.929354] [ndv4:55512:0] ib_iface.c:994 UCX DEBUG iface=0x22e9750: created UD QP 0xf4aa on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465629.929364] [ndv4:55512:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465629.929935] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465629.930445] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.930452] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.930535] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.930540] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.931004] [ndv4:55512:0] ib_md.c:812 UCX DEBUG registered memory 0x2aea77033000..0x2aea770b8000 on mlx5_ib0 lkey 0x81700 rkey 0x81700 access 0xf flags 0x3e4
[1650465629.931010] [ndv4:55512:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2aea77033018 of 544744 bytes with 128 elements
[1650465629.931015] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465629.931461] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x22e9750: adding gid fe80::15:5dff:fd33:fffb to hash on device mlx5_ib0 port 1 index 0)
[1650465629.929324] [ndv4:54756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465629.931359] [ndv4:54756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465629.931391] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.931395] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.931448] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.929675] [ndv4:55152:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x4708050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xdaef
[1650465629.930570] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.930629] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.930639] [ndv4:55152:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b3b97b08008 of 151544 bytes with 1052 elements
[1650465629.931816] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465629.931823] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465629.932194] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465629.932207] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465629.932213] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.932254] [ndv4:55432:0] ib_md.c:1319 UCX DEBUG mlx5_ib4: using registration cache
[1650465629.932267] [ndv4:54756:0] ib_iface.c:994 UCX DEBUG iface=0x3d060a0: created RC QP 0xdae5 on mlx5_ib2:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465629.932405] [ndv4:54717:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650465629.932701] [ndv4:54756:0] ucp_worker.c:1159 UCX DEBUG created interface[16]=0x3d060a0 using rc_verbs/mlx5_ib2:1 on worker 0x2e9a8d0
[1650465629.932798] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.932806] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.932877] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.932881] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.932753] [ndv4:54717:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650465629.933532] [ndv4:54861:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465629.933540] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.934788] [ndv4:54861:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465629.934793] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.934356] [ndv4:55152:0] ib_md.c:812 UCX DEBUG registered memory 0x2b3ba1200000..0x2b3ba3800000 on mlx5_ib2 lkey 0x80d00 rkey 0x80d00 access 0xf flags 0x3e4
[1650465629.934374] [ndv4:55152:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b3ba1200018 of 39845864 bytes with 4752 elements
[1650465629.934514] [ndv4:55152:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x4708050
[1650465629.934548] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[18]=0x4708050 using dc_mlx5/mlx5_ib2:1 on worker 0x357a8c0
[1650465629.934650] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.934661] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.934943] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.934949] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.933607] [ndv4:55691:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465629.933620] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465629.933155] [ndv4:54756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465629.933951] [ndv4:55516:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465629.935487] [ndv4:55516:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465629.935498] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.935501] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.935517] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.935891] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465629.935898] [ndv4:55516:0] 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> }
[1650465629.935931] [ndv4:55516:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650465629.935935] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465629.935945] [ndv4:55516:0] async.c:228 UCX DEBUG added async handler 0x126cab0 [id=87 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465629.935967] [ndv4:55516:0] async.c:506 UCX DEBUG listening to async event fd 87 events 0x1 mode thread_spinlock
[1650465629.934323] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x469e050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 3)
[1650465629.935021] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x469e050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 4)
[1650465629.935453] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x469e050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 5)
[1650465629.935359] [ndv4:55152:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465629.936370] [ndv4:55516:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465629.936463] [ndv4:55152:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465629.936783] [ndv4:55152:0] ib_iface.c:994 UCX DEBUG iface=0x48e4060: created UD QP 0xdae6 on mlx5_ib2:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465629.937258] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465629.937601] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.937607] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.937793] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.937797] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.937343] [ndv4:54715:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.937354] [ndv4:54715:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.937534] [ndv4:54715:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.937688] [ndv4:54715:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.937694] [ndv4:54715:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.938175] [ndv4:55152:0] ib_md.c:812 UCX DEBUG registered memory 0x2b3b97b2d000..0x2b3b97bb2000 on mlx5_ib2 lkey 0x80e00 rkey 0x80e00 access 0xf flags 0x3e4
[1650465629.938181] [ndv4:55152:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b3b97b2d018 of 544744 bytes with 128 elements
[1650465629.938186] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465629.939183] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x48e4060: adding gid fe80::15:5dff:fd33:fffd to hash on device mlx5_ib2 port 1 index 0)
[1650465629.937864] [ndv4:55324:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.937872] [ndv4:55324:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.938124] [ndv4:55324:0] ib_md.c:1604 UCX DEBUG mlx5_ib1: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.938352] [ndv4:55324:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.938359] [ndv4:55324:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.940044] [ndv4:55324:0] topo.c:99 UCX DEBUG bus id 0x102000000 doesn't exist. sys_dev = 1
[1650465629.940051] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.939904] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x48e4060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 1)
[1650465629.940537] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x22e9750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650465629.940997] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x22e9750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650465629.941544] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x22e9750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650465629.941565] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x22e9750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650465629.941885] [ndv4:55432:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.941894] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.942086] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x22e9750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650465629.942395] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x22e9750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650465629.942140] [ndv4:55432:0] ib_md.c:1604 UCX DEBUG mlx5_ib4: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.942249] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465629.942256] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465629.942680] [ndv4:55432:0] topo.c:99 UCX DEBUG bus id 0x105000000 doesn't exist. sys_dev = 4
[1650465629.942684] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465629.942804] [ndv4:55432:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465629.942807] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465629.942905] [ndv4:55432:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465629.942908] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465629.943014] [ndv4:55432:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465629.943017] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465629.943120] [ndv4:55432:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465629.943123] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465629.943333] [ndv4:55432:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.942076] [ndv4:55516:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x25e7010: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xf4d3
[1650465629.942336] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.942344] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.942360] [ndv4:55516:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2ae24bd3f008 of 151544 bytes with 1052 elements
[1650465629.945948] [ndv4:55516:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae250e00000..0x2ae253400000 on mlx5_ib0 lkey 0x81800 rkey 0x81800 access 0xf flags 0x3e4
[1650465629.945974] [ndv4:55516:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2ae250e00018 of 39845864 bytes with 4752 elements
[1650465629.946145] [ndv4:55516:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x25e7010
[1650465629.946183] [ndv4:55516:0] ucp_worker.c:1159 UCX DEBUG created interface[8]=0x25e7010 using dc_mlx5/mlx5_ib0:1 on worker 0x22b38d0
[1650465629.946241] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.946252] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.946332] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.946337] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.946803] [ndv4:55516:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465629.947002] [ndv4:54756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465629.947027] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.947031] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.947083] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.947441] [ndv4:54861:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465629.947447] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.948171] [ndv4:54861:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465629.948176] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.948396] [ndv4:54861:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.948287] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x469e050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 6)
[1650465629.948991] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x469e050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 7)
[1650465629.948999] [ndv4:54965:0] 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> }
[1650465629.949033] [ndv4:54965:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465629.949038] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x405ca10 [id=117 ref 1] uct_ud_iface_async_handler() to hash
[1650465629.949058] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 117 events 0x5 mode thread_spinlock
[1650465629.949070] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[30]=0x469e050 using ud_mlx5/mlx5_ib4:1 on worker 0x23dc840
[1650465629.949213] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465629.949219] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465629.949357] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465629.949362] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465629.949716] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib5:1
[1650465629.949525] [ndv4:54853:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650465629.949945] [ndv4:54853:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650465629.948515] [ndv4:55691:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465629.948524] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465629.948332] [ndv4:54756:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x4027010 of 8176 bytes with 127 elements
[1650465629.948665] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465629.948675] [ndv4:54756:0] 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> }
[1650465629.948705] [ndv4:54756:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib2 length=2048) failed: Invalid argument
[1650465629.948709] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465629.948723] [ndv4:54756:0] async.c:228 UCX DEBUG added async handler 0x25c4900 [id=99 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465629.948748] [ndv4:54756:0] async.c:506 UCX DEBUG listening to async event fd 99 events 0x1 mode thread_spinlock
[1650465629.948783] [ndv4:54756:0] ucp_worker.c:1159 UCX DEBUG created interface[17]=0x3e53030 using rc_mlx5/mlx5_ib2:1 on worker 0x2e9a8d0
[1650465629.948971] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.948977] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.949232] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.949238] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.949909] [ndv4:54756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465629.947648] [ndv4:55516:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465629.947991] [ndv4:55516:0] ib_iface.c:994 UCX DEBUG iface=0x19faa30: created UD QP 0xf4b6 on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465629.948489] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465629.950627] [ndv4:54715:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650465629.950635] [ndv4:54715:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.951394] [ndv4:54756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465629.951410] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.951413] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.951472] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.951223] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465629.951241] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.951244] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.951296] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.952327] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465629.952334] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465629.951827] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x48e4060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 2)
[1650465629.951812] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465629.951820] [ndv4:54756:0] 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> }
[1650465629.951855] [ndv4:54756:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib2 length=2048) failed: Invalid argument
[1650465629.951859] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465629.951870] [ndv4:54756:0] async.c:228 UCX DEBUG added async handler 0x33a7ef0 [id=101 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465629.951897] [ndv4:54756:0] async.c:506 UCX DEBUG listening to async event fd 101 events 0x1 mode thread_spinlock
[1650465629.952292] [ndv4:54756:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465629.952774] [ndv4:54965:0] ib_iface.c:994 UCX DEBUG iface=0x479e0a0: created RC QP 0xda55 on mlx5_ib5:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465629.953135] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[31]=0x479e0a0 using rc_verbs/mlx5_ib5:1 on worker 0x23dc840
[1650465629.953340] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465629.953346] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465629.953454] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465629.953459] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465629.953799] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib5:1
[1650465629.953633] [ndv4:55324:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465629.953640] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.954899] [ndv4:55324:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465629.954905] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.956796] [ndv4:55324:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465629.956801] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.955028] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x22e9750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650465629.955041] [ndv4:55512:0] 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> }
[1650465629.955137] [ndv4:55512:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465629.955145] [ndv4:55512:0] async.c:228 UCX DEBUG added async handler 0x1951060 [id=89 ref 1] uct_ud_iface_async_handler() to hash
[1650465629.955174] [ndv4:55512:0] async.c:506 UCX DEBUG listening to async event fd 89 events 0x5 mode thread_spinlock
[1650465629.955197] [ndv4:55512:0] ucp_worker.c:1159 UCX DEBUG created interface[10]=0x22e9750 using ud_mlx5/mlx5_ib0:1 on worker 0x22c98d0
[1650465629.955217] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.955224] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.955414] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.955419] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.956532] [ndv4:55512:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465629.957889] [ndv4:55512:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465629.957898] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.957901] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.957921] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.958284] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465629.958290] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465629.955142] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465629.955158] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.955161] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.955212] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.955661] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x4abf010 of 8176 bytes with 127 elements
[1650465629.955909] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465629.955915] [ndv4:54965:0] 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> }
[1650465629.955947] [ndv4:54965:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib5 length=2048) failed: Invalid argument
[1650465629.955951] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465629.955961] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x3891740 [id=120 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465629.955981] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 120 events 0x1 mode thread_spinlock
[1650465629.955991] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[32]=0x48eb030 using rc_mlx5/mlx5_ib5:1 on worker 0x23dc840
[1650465629.956007] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465629.956013] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465629.956206] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465629.956210] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465629.956418] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib5:1
[1650465629.958507] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465629.958522] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.958525] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.958635] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.956590] [ndv4:54861:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650465629.956966] [ndv4:54861:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650465629.957904] [ndv4:54861:0] async.c:228 UCX DEBUG added async handler 0x18bbdb0 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650465629.957930] [ndv4:54861:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650465629.957933] [ndv4:54861:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650465629.958166] [ndv4:54861:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.958173] [ndv4:54861:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.958178] [ndv4:54861:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.958219] [ndv4:54861:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650465629.959567] [ndv4:54861:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.959605] [ndv4:54861:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.958957] [ndv4:55324:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465629.958962] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.959228] [ndv4:55324:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.959132] [ndv4:55512:0] ib_iface.c:994 UCX DEBUG iface=0x2b9a050: created RC QP 0xda8d on mlx5_ib1:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465629.959081] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465629.959089] [ndv4:54965:0] 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> }
[1650465629.959120] [ndv4:54965:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib5 length=2048) failed: Invalid argument
[1650465629.959123] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465629.959130] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x48f3fc0 [id=122 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465629.959149] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 122 events 0x1 mode thread_spinlock
[1650465629.958465] [ndv4:54756:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x4029050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xdafc
[1650465629.959270] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.959279] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.959291] [ndv4:54756:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b3b53f8b008 of 151544 bytes with 1052 elements
[1650465629.957507] [ndv4:55691:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465629.957516] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465629.957849] [ndv4:55691:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.956429] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.956446] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.956803] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.956809] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.957172] [ndv4:55516:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae24bd64000..0x2ae24bde9000 on mlx5_ib0 lkey 0x81900 rkey 0x81900 access 0xf flags 0x3e4
[1650465629.957182] [ndv4:55516:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae24bd64018 of 544744 bytes with 128 elements
[1650465629.957192] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465629.957812] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x19faa30: adding gid fe80::15:5dff:fd33:fffb to hash on device mlx5_ib0 port 1 index 0)
[1650465629.958148] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x19faa30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650465629.958491] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x19faa30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650465629.959752] [ndv4:54965:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465629.959924] [ndv4:55512:0] ucp_worker.c:1159 UCX DEBUG created interface[11]=0x2b9a050 using rc_verbs/mlx5_ib1:1 on worker 0x22c98d0
[1650465629.960110] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.960116] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.959837] [ndv4:54861:0] ib_md.c:1604 UCX DEBUG mlx5_ib1: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.960122] [ndv4:54861:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.960128] [ndv4:54861:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.960604] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x48e4060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 3)
[1650465629.961158] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x48e4060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 4)
[1650465629.961415] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x48e4060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 5)
[1650465629.961430] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x48e4060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 6)
[1650465629.961539] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x48e4060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 7)
[1650465629.962775] [ndv4:54861:0] topo.c:99 UCX DEBUG bus id 0x102000000 doesn't exist. sys_dev = 1
[1650465629.962782] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.961991] [ndv4:55152:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465629.961998] [ndv4:55152:0] async.c:228 UCX DEBUG added async handler 0x2ca9390 [id=102 ref 1] uct_ud_iface_async_handler() to hash
[1650465629.962020] [ndv4:55152:0] async.c:506 UCX DEBUG listening to async event fd 102 events 0x5 mode thread_spinlock
[1650465629.962030] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[19]=0x48e4060 using ud_verbs/mlx5_ib2:1 on worker 0x357a8c0
[1650465629.962342] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.962347] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.962552] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.962558] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.963189] [ndv4:54756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b3b5da00000..0x2b3b60000000 on mlx5_ib2 lkey 0x80f00 rkey 0x80f00 access 0xf flags 0x3e4
[1650465629.963213] [ndv4:54756:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b3b5da00018 of 39845864 bytes with 4752 elements
[1650465629.963387] [ndv4:54756:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x4029050
[1650465629.963441] [ndv4:54756:0] ucp_worker.c:1159 UCX DEBUG created interface[18]=0x4029050 using dc_mlx5/mlx5_ib2:1 on worker 0x2e9a8d0
[1650465629.963793] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.963805] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.964022] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.964028] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.964840] [ndv4:54853:0] async.c:228 UCX DEBUG added async handler 0x138c080 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650465629.964959] [ndv4:54853:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650465629.964967] [ndv4:54853:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650465629.964756] [ndv4:54756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465629.965181] [ndv4:54853:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.965194] [ndv4:54853:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.965229] [ndv4:54853:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.965302] [ndv4:54853:0] module.c:253 UCX DEBUG loading modules for ucm
[1650465629.965329] [ndv4:54853:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650465629.966213] [ndv4:54853:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.966222] [ndv4:54853:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.965825] [ndv4:54756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465629.966220] [ndv4:54756:0] ib_iface.c:994 UCX DEBUG iface=0x4205060: created UD QP 0xdaf2 on mlx5_ib2:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465629.966907] [ndv4:54717:0] async.c:228 UCX DEBUG added async handler 0x23e91e0 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650465629.966996] [ndv4:54717:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650465629.967006] [ndv4:54717:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650465629.967376] [ndv4:54717:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.967386] [ndv4:54717:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.967405] [ndv4:54717:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.967464] [ndv4:54717:0] module.c:253 UCX DEBUG loading modules for ucm
[1650465629.967483] [ndv4:54717:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650465629.966389] [ndv4:54853:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.966719] [ndv4:54853:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.966725] [ndv4:54853:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.967375] [ndv4:54853:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650465629.967380] [ndv4:54853:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.966749] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465629.967155] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.967162] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.967241] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.967246] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.967684] [ndv4:54756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b3b60035000..0x2b3b600ba000 on mlx5_ib2 lkey 0x81000 rkey 0x81000 access 0xf flags 0x3e4
[1650465629.967691] [ndv4:54756:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b3b60035018 of 544744 bytes with 128 elements
[1650465629.967696] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465629.967197] [ndv4:54965:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x4ac1050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xda68
[1650465629.967478] [ndv4:55432:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib5 vendor_id: 0x15b3 device_id: 4124
[1650465629.967905] [ndv4:55432:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib5: disable ODP because it's not supported for DevX QP
[1650465629.969355] [ndv4:55324:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib2 vendor_id: 0x15b3 device_id: 4124
[1650465629.969757] [ndv4:55324:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib2: disable ODP because it's not supported for DevX QP
[1650465629.969021] [ndv4:55432:0] async.c:228 UCX DEBUG added async handler 0x21367b0 [id=67 ref 1] uct_ib_async_event_handler() to hash
[1650465629.969045] [ndv4:55432:0] async.c:506 UCX DEBUG listening to async event fd 67 events 0x1 mode thread_spinlock
[1650465629.969048] [ndv4:55432:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib5' (InfiniBand channel adapter) with 1 ports
[1650465629.969252] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465629.969260] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465629.969265] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.969308] [ndv4:55432:0] ib_md.c:1319 UCX DEBUG mlx5_ib5: using registration cache
[1650465629.970785] [ndv4:55432:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.970791] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.968432] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x19faa30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650465629.968932] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x19faa30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650465629.969255] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x19faa30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650465629.969509] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x19faa30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650465629.970084] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.970093] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.970764] [ndv4:55512:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465629.970964] [ndv4:55432:0] ib_md.c:1604 UCX DEBUG mlx5_ib5: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.971938] [ndv4:55512:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465629.971951] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.971954] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.971972] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.972300] [ndv4:55512:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x2bad010 of 8176 bytes with 127 elements
[1650465629.972529] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465629.972536] [ndv4:55512:0] 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> }
[1650465629.972570] [ndv4:55512:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650465629.972601] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465629.972611] [ndv4:55512:0] async.c:228 UCX DEBUG added async handler 0x23a0fc0 [id=92 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465629.972636] [ndv4:55512:0] async.c:506 UCX DEBUG listening to async event fd 92 events 0x1 mode thread_spinlock
[1650465629.972646] [ndv4:55512:0] ucp_worker.c:1159 UCX DEBUG created interface[12]=0x2a1b010 using rc_mlx5/mlx5_ib1:1 on worker 0x22c98d0
[1650465629.972776] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.972781] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.972967] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.972972] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.973376] [ndv4:55512:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465629.973219] [ndv4:55152:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465629.974078] [ndv4:54861:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465629.974088] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.975249] [ndv4:54861:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465629.975253] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.974470] [ndv4:55152:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465629.974746] [ndv4:55152:0] ib_iface.c:994 UCX DEBUG iface=0x2c21900: created UD QP 0xdaf3 on mlx5_ib2:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465629.974752] [ndv4:55152:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465629.975390] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465629.975526] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.975532] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.975704] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.975709] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.976152] [ndv4:55691:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib5 vendor_id: 0x15b3 device_id: 4124
[1650465629.976460] [ndv4:55691:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib5: disable ODP because it's not supported for DevX QP
[1650465629.976147] [ndv4:55152:0] ib_md.c:812 UCX DEBUG registered memory 0x2b3b97bb2000..0x2b3b97c37000 on mlx5_ib2 lkey 0x81100 rkey 0x81100 access 0xf flags 0x3e4
[1650465629.976153] [ndv4:55152:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b3b97bb2018 of 544744 bytes with 128 elements
[1650465629.976157] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465629.976191] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x2c21900: adding gid fe80::15:5dff:fd33:fffd to hash on device mlx5_ib2 port 1 index 0)
[1650465629.976206] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x2c21900: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 1)
[1650465629.976221] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x2c21900: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 2)
[1650465629.976293] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x2c21900: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 3)
[1650465629.976381] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x2c21900: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 4)
[1650465629.977225] [ndv4:55691:0] async.c:228 UCX DEBUG added async handler 0x2aa4c40 [id=67 ref 1] uct_ib_async_event_handler() to hash
[1650465629.977257] [ndv4:55691:0] async.c:506 UCX DEBUG listening to async event fd 67 events 0x1 mode thread_spinlock
[1650465629.977261] [ndv4:55691:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib5' (InfiniBand channel adapter) with 1 ports
[1650465629.979359] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x19faa30: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650465629.977958] [ndv4:54717:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.977970] [ndv4:54717:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.978174] [ndv4:54717:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.978406] [ndv4:54717:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.978413] [ndv4:54717:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.977361] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465629.977373] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465629.977385] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b4b34479008 of 151544 bytes with 1052 elements
[1650465629.977908] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x4205060: adding gid fe80::15:5dff:fd33:fffd to hash on device mlx5_ib2 port 1 index 0)
[1650465629.978352] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x4205060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 1)
[1650465629.978372] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x4205060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 2)
[1650465629.978938] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x4205060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 3)
[1650465629.979226] [ndv4:54853:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465629.979234] [ndv4:54853:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.980210] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465629.980220] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465629.982388] [ndv4:55432:0] topo.c:99 UCX DEBUG bus id 0x106000000 doesn't exist. sys_dev = 5
[1650465629.982394] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465629.981324] [ndv4:54965:0] ib_md.c:812 UCX DEBUG registered memory 0x2b4b31e00000..0x2b4b34400000 on mlx5_ib5 lkey 0x80400 rkey 0x80400 access 0xf flags 0x3e4
[1650465629.981337] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b4b31e00018 of 39845864 bytes with 4752 elements
[1650465629.981479] [ndv4:54965:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x4ac1050
[1650465629.981509] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[33]=0x4ac1050 using dc_mlx5/mlx5_ib5:1 on worker 0x23dc840
[1650465629.981827] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465629.981834] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465629.981912] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465629.981916] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465629.980361] [ndv4:55324:0] async.c:228 UCX DEBUG added async handler 0x165dc60 [id=61 ref 1] uct_ib_async_event_handler() to hash
[1650465629.980401] [ndv4:55324:0] async.c:506 UCX DEBUG listening to async event fd 61 events 0x1 mode thread_spinlock
[1650465629.980405] [ndv4:55324:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib2' (InfiniBand channel adapter) with 1 ports
[1650465629.980728] [ndv4:55324:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.980742] [ndv4:55324:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.980752] [ndv4:55324:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.980838] [ndv4:55324:0] ib_md.c:1319 UCX DEBUG mlx5_ib2: using registration cache
[1650465629.982243] [ndv4:55324:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.982254] [ndv4:55324:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.982462] [ndv4:55324:0] ib_md.c:1604 UCX DEBUG mlx5_ib2: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.982713] [ndv4:55324:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465629.982721] [ndv4:55324:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465629.979726] [ndv4:55516:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465629.981969] [ndv4:54717:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650465629.981977] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.984395] [ndv4:55512:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465629.984412] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465629.984416] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465629.984470] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465629.984857] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465629.984865] [ndv4:55512:0] 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> }
[1650465629.984898] [ndv4:55512:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650465629.984902] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465629.984909] [ndv4:55512:0] async.c:228 UCX DEBUG added async handler 0x1a03db0 [id=94 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465629.984930] [ndv4:55512:0] async.c:506 UCX DEBUG listening to async event fd 94 events 0x1 mode thread_spinlock
[1650465629.984461] [ndv4:55432:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465629.984468] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465629.984538] [ndv4:55432:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465629.984541] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465629.984645] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib5:1
[1650465629.985648] [ndv4:55516:0] async.c:228 UCX DEBUG added async handler 0x1949c50 [id=88 ref 1] uct_ud_iface_async_handler() to hash
[1650465629.985677] [ndv4:55516:0] async.c:506 UCX DEBUG listening to async event fd 88 events 0x5 mode thread_spinlock
[1650465629.985698] [ndv4:55516:0] ucp_worker.c:1159 UCX DEBUG created interface[9]=0x19faa30 using ud_verbs/mlx5_ib0:1 on worker 0x22b38d0
[1650465629.985726] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.985733] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.985780] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.985785] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.985945] [ndv4:55516:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465629.985664] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x2c21900: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 5)
[1650465629.986298] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x2c21900: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 6)
[1650465629.986322] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x2c21900: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 7)
[1650465629.986327] [ndv4:55152:0] 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> }
[1650465629.986362] [ndv4:55152:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465629.986367] [ndv4:55152:0] async.c:228 UCX DEBUG added async handler 0x48e4ef0 [id=103 ref 1] uct_ud_iface_async_handler() to hash
[1650465629.986387] [ndv4:55152:0] async.c:506 UCX DEBUG listening to async event fd 103 events 0x5 mode thread_spinlock
[1650465629.986402] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[20]=0x2c21900 using ud_mlx5/mlx5_ib2:1 on worker 0x357a8c0
[1650465629.986892] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465629.986899] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465629.987571] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465629.987626] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465629.985663] [ndv4:55512:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465629.985771] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465629.985783] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465629.985790] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465629.985870] [ndv4:55691:0] ib_md.c:1319 UCX DEBUG mlx5_ib5: using registration cache
[1650465629.987527] [ndv4:55691:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465629.987535] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465629.987743] [ndv4:55691:0] ib_md.c:1604 UCX DEBUG mlx5_ib5: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465629.987923] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465629.987930] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465629.986012] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465629.986484] [ndv4:54965:0] ib_iface.c:994 UCX DEBUG iface=0x4c9d060: created UD QP 0xda5e on mlx5_ib5:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465629.987162] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465629.987974] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465629.987981] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465629.988246] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465629.988251] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465629.988706] [ndv4:54965:0] ib_md.c:812 UCX DEBUG registered memory 0x2b4b3449e000..0x2b4b34523000 on mlx5_ib5 lkey 0x80500 rkey 0x80500 access 0xf flags 0x3e4
[1650465629.988712] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b4b3449e018 of 544744 bytes with 128 elements
[1650465629.988716] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465629.989368] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x4c9d060: adding gid fe80::15:5dff:fd34:0 to hash on device mlx5_ib5 port 1 index 0)
[1650465629.989717] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x4c9d060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 1)
[1650465629.989960] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x4c9d060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 2)
[1650465629.990024] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x4c9d060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 3)
[1650465629.990041] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x4c9d060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 4)
[1650465629.990645] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x4c9d060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 5)
[1650465629.990944] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x4c9d060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 6)
[1650465629.991104] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x4c9d060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 7)
[1650465629.988864] [ndv4:54715:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465629.988873] [ndv4:54715:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.989951] [ndv4:54715:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465629.989956] [ndv4:54715:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.990941] [ndv4:54715:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465629.990947] [ndv4:54715:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.990474] [ndv4:55691:0] topo.c:99 UCX DEBUG bus id 0x106000000 doesn't exist. sys_dev = 5
[1650465629.990481] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465629.988001] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x4205060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 4)
[1650465629.989218] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x4205060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 5)
[1650465629.988092] [ndv4:55516:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465629.988552] [ndv4:55516:0] ib_iface.c:994 UCX DEBUG iface=0x22d3750: created UD QP 0xf4b7 on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465629.988563] [ndv4:55516:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465629.989189] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465629.989455] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.989462] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.989740] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465629.989745] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465629.990077] [ndv4:55516:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae24bde9000..0x2ae24be6e000 on mlx5_ib0 lkey 0x81a00 rkey 0x81a00 access 0xf flags 0x3e4
[1650465629.990086] [ndv4:55516:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae24bde9018 of 544744 bytes with 128 elements
[1650465629.990094] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465629.987328] [ndv4:55432:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465629.987334] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465629.986867] [ndv4:54861:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465629.986873] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465629.991381] [ndv4:54965:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465629.991388] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x47799b0 [id=123 ref 1] uct_ud_iface_async_handler() to hash
[1650465629.991410] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 123 events 0x5 mode thread_spinlock
[1650465629.991421] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[34]=0x4c9d060 using ud_verbs/mlx5_ib5:1 on worker 0x23dc840
[1650465629.991560] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465629.991566] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465629.991624] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465629.991628] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465629.991784] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib5:1
[1650465629.991982] [ndv4:54715:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465629.991988] [ndv4:54715:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.992246] [ndv4:54715:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465629.992774] [ndv4:55512:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x2d45010: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xdab6
[1650465629.992932] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.992938] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.992947] [ndv4:55512:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2aea798b9008 of 151544 bytes with 1052 elements
[1650465629.992664] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465629.992944] [ndv4:54965:0] ib_iface.c:994 UCX DEBUG iface=0x4dbb460: created UD QP 0xda5f on mlx5_ib5:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465629.992952] [ndv4:54965:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465629.993392] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465629.993608] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465629.993615] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465629.993646] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465629.993652] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465629.993971] [ndv4:54965:0] ib_md.c:812 UCX DEBUG registered memory 0x2b4b34523000..0x2b4b345a8000 on mlx5_ib5 lkey 0x80600 rkey 0x80600 access 0xf flags 0x3e4
[1650465629.993977] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b4b34523018 of 544744 bytes with 128 elements
[1650465629.993982] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465629.994013] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x4dbb460: adding gid fe80::15:5dff:fd34:0 to hash on device mlx5_ib5 port 1 index 0)
[1650465629.994044] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x4dbb460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 1)
[1650465629.994269] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x4dbb460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 2)
[1650465629.992711] [ndv4:54717:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465629.992719] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465629.995395] [ndv4:55324:0] topo.c:99 UCX DEBUG bus id 0x103000000 doesn't exist. sys_dev = 2
[1650465629.995404] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465629.997443] [ndv4:55512:0] ib_md.c:812 UCX DEBUG registered memory 0x2aea77200000..0x2aea79800000 on mlx5_ib1 lkey 0x80f00 rkey 0x80f00 access 0xf flags 0x3e4
[1650465629.997472] [ndv4:55512:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2aea77200018 of 39845864 bytes with 4752 elements
[1650465629.998165] [ndv4:55152:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465629.997767] [ndv4:55512:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x2d45010
[1650465629.997826] [ndv4:55512:0] ucp_worker.c:1159 UCX DEBUG created interface[13]=0x2d45010 using dc_mlx5/mlx5_ib1:1 on worker 0x22c98d0
[1650465629.998159] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.998171] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.998241] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465629.998246] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465629.999488] [ndv4:55512:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465629.999680] [ndv4:55432:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465629.999688] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465629.999901] [ndv4:55432:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.000020] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x4205060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 6)
[1650465630.000701] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x4205060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 7)
[1650465630.000835] [ndv4:55512:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.001253] [ndv4:55512:0] ib_iface.c:994 UCX DEBUG iface=0x2c443b0: created UD QP 0xda97 on mlx5_ib1:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465630.001893] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465630.001976] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.001983] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.002098] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.002104] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.000947] [ndv4:54756:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.000957] [ndv4:54756:0] async.c:228 UCX DEBUG added async handler 0x3782400 [id=102 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.000981] [ndv4:54756:0] async.c:506 UCX DEBUG listening to async event fd 102 events 0x5 mode thread_spinlock
[1650465630.000995] [ndv4:54756:0] ucp_worker.c:1159 UCX DEBUG created interface[19]=0x4205060 using ud_verbs/mlx5_ib2:1 on worker 0x2e9a8d0
[1650465630.001151] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.001158] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.001466] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.001471] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.001317] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x22d3750: adding gid fe80::15:5dff:fd33:fffb to hash on device mlx5_ib0 port 1 index 0)
[1650465630.002302] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x22d3750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650465630.002487] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x22d3750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650465630.003115] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x22d3750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650465630.003844] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x22d3750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650465630.002435] [ndv4:55512:0] ib_md.c:812 UCX DEBUG registered memory 0x2aea798de000..0x2aea79963000 on mlx5_ib1 lkey 0x81000 rkey 0x81000 access 0xf flags 0x3e4
[1650465630.002442] [ndv4:55512:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2aea798de018 of 544744 bytes with 128 elements
[1650465630.002446] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465630.002479] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x2c443b0: adding gid fe80::15:5dff:fd33:fffc to hash on device mlx5_ib1 port 1 index 0)
[1650465630.002906] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x2c443b0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 1)
[1650465630.003192] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x2c443b0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 2)
[1650465630.003756] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x2c443b0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 3)
[1650465630.003776] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x2c443b0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 4)
[1650465630.003790] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x2c443b0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 5)
[1650465630.003809] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x2c443b0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 6)
[1650465630.002425] [ndv4:54756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465630.001153] [ndv4:54853:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465630.001162] [ndv4:54853:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465630.002523] [ndv4:54853:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465630.002529] [ndv4:54853:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465630.002792] [ndv4:55691:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465630.002800] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465630.003988] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x4dbb460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 3)
[1650465630.003978] [ndv4:54717:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465630.003987] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465630.005792] [ndv4:54853:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465630.005798] [ndv4:54853:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465630.006092] [ndv4:54853:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.011605] [ndv4:54853:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650465630.011928] [ndv4:54853:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650465630.011998] [ndv4:55691:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465630.012006] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465630.012164] [ndv4:55691:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465630.012167] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465630.012359] [ndv4:55691:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465630.012363] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465630.012682] [ndv4:55691:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.007837] [ndv4:55152:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465630.007859] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.007864] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.007915] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.008444] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.008456] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465630.009694] [ndv4:55152:0] ib_iface.c:994 UCX DEBUG iface=0x4b020a0: created RC QP 0xda5e on mlx5_ib3:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465630.011049] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[21]=0x4b020a0 using rc_verbs/mlx5_ib3:1 on worker 0x357a8c0
[1650465630.011097] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.011104] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.011185] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.011190] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.011400] [ndv4:55152:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465630.012444] [ndv4:55152:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.012457] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.012460] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.012508] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.008473] [ndv4:54715:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650465630.008841] [ndv4:54715:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650465630.012713] [ndv4:54756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.012971] [ndv4:55152:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x4e23010 of 8176 bytes with 127 elements
[1650465630.013188] [ndv4:54756:0] ib_iface.c:994 UCX DEBUG iface=0x4323050: created UD QP 0xdaf4 on mlx5_ib2:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465630.013200] [ndv4:54756:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465630.013209] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x2c443b0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 7)
[1650465630.013238] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.013246] [ndv4:55152:0] 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> }
[1650465630.013284] [ndv4:55152:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib3 length=2048) failed: Invalid argument
[1650465630.013289] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.013303] [ndv4:55152:0] async.c:228 UCX DEBUG added async handler 0x4a02ae0 [id=106 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.013328] [ndv4:55152:0] async.c:506 UCX DEBUG listening to async event fd 106 events 0x1 mode thread_spinlock
[1650465630.013339] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[22]=0x4c4f030 using rc_mlx5/mlx5_ib3:1 on worker 0x357a8c0
[1650465630.013370] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.013376] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.013491] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.013497] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.013666] [ndv4:55512:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.013678] [ndv4:55512:0] async.c:228 UCX DEBUG added async handler 0x19f8340 [id=95 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.013700] [ndv4:55512:0] async.c:506 UCX DEBUG listening to async event fd 95 events 0x5 mode thread_spinlock
[1650465630.013712] [ndv4:55512:0] ucp_worker.c:1159 UCX DEBUG created interface[14]=0x2c443b0 using ud_verbs/mlx5_ib1:1 on worker 0x22c98d0
[1650465630.013760] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.013766] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.013811] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.013815] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.014024] [ndv4:55512:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465630.013900] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465630.014173] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.014181] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.014195] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.014198] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.014668] [ndv4:54756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b3b600ba000..0x2b3b6013f000 on mlx5_ib2 lkey 0x81200 rkey 0x81200 access 0xf flags 0x3e4
[1650465630.014675] [ndv4:54756:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b3b600ba018 of 544744 bytes with 128 elements
[1650465630.014679] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465630.014715] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x4323050: adding gid fe80::15:5dff:fd33:fffd to hash on device mlx5_ib2 port 1 index 0)
[1650465630.014732] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x4323050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 1)
[1650465630.014746] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x4323050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 2)
[1650465630.014759] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x4323050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 3)
[1650465630.014948] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x4323050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 4)
[1650465630.015206] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x4323050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 5)
[1650465630.015315] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x4323050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 6)
[1650465630.015571] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x4323050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 7)
[1650465630.015597] [ndv4:54756:0] 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> }
[1650465630.015633] [ndv4:54756:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.015638] [ndv4:54756:0] async.c:228 UCX DEBUG added async handler 0x4323f30 [id=103 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.015662] [ndv4:54756:0] async.c:506 UCX DEBUG listening to async event fd 103 events 0x5 mode thread_spinlock
[1650465630.015673] [ndv4:54756:0] ucp_worker.c:1159 UCX DEBUG created interface[20]=0x4323050 using ud_mlx5/mlx5_ib2:1 on worker 0x2e9a8d0
[1650465630.015688] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.015694] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.015799] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.015804] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.013707] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x22d3750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650465630.014153] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x22d3750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650465630.014193] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x22d3750: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650465630.014201] [ndv4:55516:0] 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> }
[1650465630.014305] [ndv4:55516:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.014310] [ndv4:55516:0] async.c:228 UCX DEBUG added async handler 0x193b060 [id=89 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.014334] [ndv4:55516:0] async.c:506 UCX DEBUG listening to async event fd 89 events 0x5 mode thread_spinlock
[1650465630.014354] [ndv4:55516:0] ucp_worker.c:1159 UCX DEBUG created interface[10]=0x22d3750 using ud_mlx5/mlx5_ib0:1 on worker 0x22b38d0
[1650465630.014371] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.014377] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.014468] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.014472] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.014687] [ndv4:55516:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465630.015498] [ndv4:55516:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465630.015512] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.015516] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.015543] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.015927] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.015932] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465630.013884] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x4dbb460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 4)
[1650465630.013921] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x4dbb460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 5)
[1650465630.013936] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x4dbb460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 6)
[1650465630.013949] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x4dbb460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 7)
[1650465630.013956] [ndv4:54965:0] 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> }
[1650465630.013987] [ndv4:54965:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.013992] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x4c9df20 [id=124 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.014016] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 124 events 0x5 mode thread_spinlock
[1650465630.014028] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[35]=0x4dbb460 using ud_mlx5/mlx5_ib5:1 on worker 0x23dc840
[1650465630.014153] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.014159] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.014233] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.014238] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.014432] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib6:1
[1650465630.015648] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465630.015666] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.015669] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.015718] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.013696] [ndv4:55152:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465630.014674] [ndv4:55152:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.014688] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.014691] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.014747] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.015142] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465630.015149] [ndv4:55152:0] 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> }
[1650465630.015182] [ndv4:55152:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib3 length=2048) failed: Invalid argument
[1650465630.015186] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.015192] [ndv4:55152:0] async.c:228 UCX DEBUG added async handler 0x2c9ef50 [id=108 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.015212] [ndv4:55152:0] async.c:506 UCX DEBUG listening to async event fd 108 events 0x1 mode thread_spinlock
[1650465630.015855] [ndv4:55152:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465630.014795] [ndv4:55324:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465630.014803] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465630.014875] [ndv4:55324:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465630.014879] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465630.015180] [ndv4:55324:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465630.015184] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465630.027544] [ndv4:55324:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465630.027550] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465630.027857] [ndv4:55324:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.035491] [ndv4:55324:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib3 vendor_id: 0x15b3 device_id: 4124
[1650465630.035867] [ndv4:55324:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib3: disable ODP because it's not supported for DevX QP
[1650465630.036784] [ndv4:55324:0] async.c:228 UCX DEBUG added async handler 0x165d360 [id=63 ref 1] uct_ib_async_event_handler() to hash
[1650465630.036809] [ndv4:55324:0] async.c:506 UCX DEBUG listening to async event fd 63 events 0x1 mode thread_spinlock
[1650465630.036813] [ndv4:55324:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib3' (InfiniBand channel adapter) with 1 ports
[1650465630.036973] [ndv4:55324:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.036981] [ndv4:55324:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.036986] [ndv4:55324:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.037034] [ndv4:55324:0] ib_md.c:1319 UCX DEBUG mlx5_ib3: using registration cache
[1650465630.059798] [ndv4:55324:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.059811] [ndv4:55324:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.060076] [ndv4:55324:0] ib_md.c:1604 UCX DEBUG mlx5_ib3: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.060361] [ndv4:55324:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.060368] [ndv4:55324:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.083124] [ndv4:55324:0] topo.c:99 UCX DEBUG bus id 0x104000000 doesn't exist. sys_dev = 3
[1650465630.083134] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465630.094948] [ndv4:55324:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465630.094956] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465630.095375] [ndv4:55324:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465630.095379] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465630.013807] [ndv4:54717:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465630.013819] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465630.014323] [ndv4:54717:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465630.014328] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465630.014695] [ndv4:54717:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.022673] [ndv4:54717:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650465630.023043] [ndv4:54717:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650465630.023512] [ndv4:54717:0] async.c:228 UCX DEBUG added async handler 0x23e1f50 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650465630.023543] [ndv4:54717:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650465630.023547] [ndv4:54717:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650465630.023721] [ndv4:54717:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.023731] [ndv4:54717:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.023738] [ndv4:54717:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.023789] [ndv4:54717:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650465630.024852] [ndv4:54717:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.024857] [ndv4:54717:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.024995] [ndv4:54717:0] ib_md.c:1604 UCX DEBUG mlx5_ib1: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.025016] [ndv4:54717:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.025022] [ndv4:54717:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.036542] [ndv4:54717:0] topo.c:99 UCX DEBUG bus id 0x102000000 doesn't exist. sys_dev = 1
[1650465630.036562] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465630.058841] [ndv4:54717:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465630.058849] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465630.070081] [ndv4:54717:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465630.070089] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465630.071923] [ndv4:54717:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465630.071929] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465630.084768] [ndv4:54717:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465630.084776] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465630.085020] [ndv4:54717:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.093030] [ndv4:54717:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib2 vendor_id: 0x15b3 device_id: 4124
[1650465630.093337] [ndv4:54717:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib2: disable ODP because it's not supported for DevX QP
[1650465630.093775] [ndv4:54717:0] async.c:228 UCX DEBUG added async handler 0x23e9ce0 [id=61 ref 1] uct_ib_async_event_handler() to hash
[1650465630.093805] [ndv4:54717:0] async.c:506 UCX DEBUG listening to async event fd 61 events 0x1 mode thread_spinlock
[1650465630.093808] [ndv4:54717:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib2' (InfiniBand channel adapter) with 1 ports
[1650465630.093965] [ndv4:54717:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.093973] [ndv4:54717:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.093979] [ndv4:54717:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.094052] [ndv4:54717:0] ib_md.c:1319 UCX DEBUG mlx5_ib2: using registration cache
[1650465630.021239] [ndv4:54715:0] async.c:228 UCX DEBUG added async handler 0x2006d70 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650465630.021267] [ndv4:54715:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650465630.021271] [ndv4:54715:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650465630.021418] [ndv4:54715:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.021426] [ndv4:54715:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.021431] [ndv4:54715:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.021477] [ndv4:54715:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650465630.022671] [ndv4:54715:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.022677] [ndv4:54715:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.022869] [ndv4:54715:0] ib_md.c:1604 UCX DEBUG mlx5_ib1: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.022946] [ndv4:54715:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.022952] [ndv4:54715:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.023540] [ndv4:54715:0] topo.c:99 UCX DEBUG bus id 0x102000000 doesn't exist. sys_dev = 1
[1650465630.023545] [ndv4:54715:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465630.036001] [ndv4:54715:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465630.036009] [ndv4:54715:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465630.059476] [ndv4:54715:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465630.059484] [ndv4:54715:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465630.070631] [ndv4:54715:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465630.070638] [ndv4:54715:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465630.083192] [ndv4:54715:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465630.083200] [ndv4:54715:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465630.083426] [ndv4:54715:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.090595] [ndv4:54715:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib2 vendor_id: 0x15b3 device_id: 4124
[1650465630.090946] [ndv4:54715:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib2: disable ODP because it's not supported for DevX QP
[1650465630.091790] [ndv4:54715:0] async.c:228 UCX DEBUG added async handler 0x200ec00 [id=61 ref 1] uct_ib_async_event_handler() to hash
[1650465630.091811] [ndv4:54715:0] async.c:506 UCX DEBUG listening to async event fd 61 events 0x1 mode thread_spinlock
[1650465630.091814] [ndv4:54715:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib2' (InfiniBand channel adapter) with 1 ports
[1650465630.091981] [ndv4:54715:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.091989] [ndv4:54715:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.091994] [ndv4:54715:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.092035] [ndv4:54715:0] ib_md.c:1319 UCX DEBUG mlx5_ib2: using registration cache
[1650465630.094019] [ndv4:54715:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.094026] [ndv4:54715:0] mpool.c:88 UCX D[1650465630.016506] [ndv4:54861:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465630.016515] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465630.016747] [ndv4:54861:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.023498] [ndv4:54861:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib2 vendor_id: 0x15b3 device_id: 4124
[1650465630.023816] [ndv4:54861:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib2: disable ODP because it's not supported for DevX QP
[1650465630.024121] [ndv4:54861:0] async.c:228 UCX DEBUG added async handler 0x18bbe00 [id=61 ref 1] uct_ib_async_event_handler() to hash
[1650465630.024145] [ndv4:54861:0] async.c:506 UCX DEBUG listening to async event fd 61 events 0x1 mode thread_spinlock
[1650465630.024148] [ndv4:54861:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib2' (InfiniBand channel adapter) with 1 ports
[1650465630.024424] [ndv4:54861:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.024433] [ndv4:54861:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.024437] [ndv4:54861:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.024477] [ndv4:54861:0] ib_md.c:1319 UCX DEBUG mlx5_ib2: using registration cache
[1650465630.025023] [ndv4:54861:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.025028] [ndv4:54861:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.025226] [ndv4:54861:0] ib_md.c:1604 UCX DEBUG mlx5_ib2: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.025357] [ndv4:54861:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.025363] [ndv4:54861:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.026110] [ndv4:54861:0] topo.c:99 UCX DEBUG bus id 0x103000000 doesn't exist. sys_dev = 2
[1650465630.026115] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465630.026675] [ndv4:54861:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465630.026680] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465630.050561] [ndv4:54861:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465630.050568] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465630.062301] [ndv4:54861:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465630.062309] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465630.073604] [ndv4:54861:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465630.073610] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465630.073825] [ndv4:54861:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.081010] [ndv4:54861:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib3 vendor_id: 0x15b3 device_id: 4124
[1650465630.081364] [ndv4:54861:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib3: disable ODP because it's not supported for DevX QP
[1650465630.082102] [ndv4:54861:0] async.c:228 UCX DEBUG added async handler 0x1fa55a0 [id=63 ref 1] uct_ib_async_event_handler() to hash
[1650465630.082125] [ndv4:54861:0] async.c:506 UCX DEBUG listening to async event fd 63 events 0x1 mode thread_spinlock
[1650465630.082128] [ndv4:54861:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib3' (InfiniBand channel adapter) with 1 ports
[1650465630.082535] [ndv4:54861:0] ib_md.c:296 UCX D[1650465630.023694] [ndv4:55152:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x4e25050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xda74
[1650465630.023947] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.023954] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.023963] [ndv4:55152:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b3b97c39008 of 151544 bytes with 1052 elements
[1650465630.027773] [ndv4:55152:0] ib_md.c:812 UCX DEBUG registered memory 0x2b3ba3a00000..0x2b3ba6000000 on mlx5_ib3 lkey 0x80b00 rkey 0x80b00 access 0xf flags 0x3e4
[1650465630.027791] [ndv4:55152:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b3ba3a00018 of 39845864 bytes with 4752 elements
[1650465630.027938] [ndv4:55152:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x4e25050
[1650465630.027971] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[23]=0x4e25050 using dc_mlx5/mlx5_ib3:1 on worker 0x357a8c0
[1650465630.028173] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.028183] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.028292] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.028297] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.029110] [ndv4:55152:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465630.040320] [ndv4:55152:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.040602] [ndv4:55152:0] ib_iface.c:994 UCX DEBUG iface=0x5001060: created UD QP 0xda78 on mlx5_ib3:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465630.041069] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465630.041448] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.041455] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.041526] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.041531] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.041894] [ndv4:55152:0] ib_md.c:812 UCX DEBUG registered memory 0x2b3b97c5e000..0x2b3b97ce3000 on mlx5_ib3 lkey 0x80e00 rkey 0x80e00 access 0xf flags 0x3e4
[1650465630.041900] [ndv4:55152:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b3b97c5e018 of 544744 bytes with 128 elements
[1650465630.041904] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465630.051711] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x5001060: adding gid fe80::15:5dff:fd33:fffe to hash on device mlx5_ib3 port 1 index 0)
[1650465630.052188] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x5001060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 1)
[1650465630.053143] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x5001060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 2)
[1650465630.053770] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x5001060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 3)
[1650465630.054253] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x5001060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 4)
[1650465630.061938] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x5001060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 5)
[1650465630.062490] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x5001060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 6)
[1650465630.063043] [ndv4:55152:0] [1650465630.027037] [ndv4:54853:0] async.c:228 UCX DEBUG added async handler 0x138cd70 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650465630.027067] [ndv4:54853:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650465630.027071] [ndv4:54853:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650465630.027393] [ndv4:54853:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.027406] [ndv4:54853:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.027412] [ndv4:54853:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.027460] [ndv4:54853:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650465630.029757] [ndv4:54853:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.029764] [ndv4:54853:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.029921] [ndv4:54853:0] ib_md.c:1604 UCX DEBUG mlx5_ib1: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.030121] [ndv4:54853:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.030128] [ndv4:54853:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.044339] [ndv4:54853:0] topo.c:99 UCX DEBUG bus id 0x102000000 doesn't exist. sys_dev = 1
[1650465630.044351] [ndv4:54853:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465630.044872] [ndv4:54853:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465630.044877] [ndv4:54853:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465630.045054] [ndv4:54853:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465630.045057] [ndv4:54853:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465630.045196] [ndv4:54853:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465630.045199] [ndv4:54853:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465630.045295] [ndv4:54853:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465630.045298] [ndv4:54853:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465630.045708] [ndv4:54853:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.055200] [ndv4:54853:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib2 vendor_id: 0x15b3 device_id: 4124
[1650465630.055511] [ndv4:54853:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib2: disable ODP because it's not supported for DevX QP
[1650465630.064874] [ndv4:54853:0] async.c:228 UCX DEBUG added async handler 0x1394c00 [id=61 ref 1] uct_ib_async_event_handler() to hash
[1650465630.064910] [ndv4:54853:0] async.c:506 UCX DEBUG listening to async event fd 61 events 0x1 mode thread_spinlock
[1650465630.064914] [ndv4:54853:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib2' (InfiniBand channel adapter) with 1 ports
[1650465630.065122] [ndv4:54853:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.065133] [ndv4:54853:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.065139] [ndv4:54853:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.065200] [ndv4:54853:0] ib_md.c:1319 UCX DEBUG mlx5_ib2: using registration cache
[1650465630.066856] [ndv4:54853:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.066863] [ndv4:54853:0] mpool.c:88 UCX D[1650465630.056845] [ndv4:55039:0] debug.c:1198 UCX DEBUG using signal stack 0x2b0b39c2e000 size 141824
[1650465630.056922] [ndv4:55039:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465630.056942] [ndv4:55039:0] 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 0x2b0b39a9c000
[1650465630.056961] [ndv4:55039:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650465630.056969] [ndv4:55039:0] 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
[1650465630.056975] [ndv4:55039:0] module.c:253 UCX DEBUG loading modules for ucs
[1650465630.059510] [ndv4:55039:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465630.059527] [ndv4:55039:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650465630.059565] [ndv4:55039:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650465630.059568] [ndv4:55039:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650465630.059593] [ndv4:55039:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650465630.059601] [ndv4:55039:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650465630.059604] [ndv4:55039:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650465630.059608] [ndv4:55039:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650465630.059611] [ndv4:55039:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650465630.059613] [ndv4:55039:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650465630.059615] [ndv4:55039:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650465630.059618] [ndv4:55039:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650465630.059626] [ndv4:55039:0] module.c:253 UCX DEBUG loading modules for uct
[1650465630.060357] [ndv4:55039:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650465630.062039] [ndv4:55039:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650465630.062055] [ndv4:55039:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650465630.062068] [ndv4:55039:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650465630.062079] [ndv4:55039:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650465630.062089] [ndv4:55039:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650465630.062100] [ndv4:55039:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650465630.062110] [ndv4:55039:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650465630.062481] [ndv4:55039:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650465630.068823] [ndv4:55039:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650465630.069140] [ndv4:55039:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650465630.074085] [ndv4:55039:0] async.c:228 UCX DEBUG added async handler 0x19dd1e0 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650465630.074175] [ndv4:55039:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650465630.074182] [ndv4:55039:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650465630.074417] [[1650465630.019115] [ndv4:55432:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib6 vendor_id: 0x15b3 device_id: 4124
[1650465630.019487] [ndv4:55432:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib6: disable ODP because it's not supported for DevX QP
[1650465630.020966] [ndv4:55432:0] async.c:228 UCX DEBUG added async handler 0x281e640 [id=69 ref 1] uct_ib_async_event_handler() to hash
[1650465630.020991] [ndv4:55432:0] async.c:506 UCX DEBUG listening to async event fd 69 events 0x1 mode thread_spinlock
[1650465630.020995] [ndv4:55432:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib6' (InfiniBand channel adapter) with 1 ports
[1650465630.021326] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.021335] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.021340] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.021381] [ndv4:55432:0] ib_md.c:1319 UCX DEBUG mlx5_ib6: using registration cache
[1650465630.022804] [ndv4:55432:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.022810] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.022974] [ndv4:55432:0] ib_md.c:1604 UCX DEBUG mlx5_ib6: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.022990] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.022996] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.023670] [ndv4:55432:0] topo.c:99 UCX DEBUG bus id 0x107000000 doesn't exist. sys_dev = 6
[1650465630.023675] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465630.024164] [ndv4:55432:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465630.024167] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465630.024478] [ndv4:55432:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465630.024481] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465630.036204] [ndv4:55432:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465630.036211] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465630.058795] [ndv4:55432:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465630.058802] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465630.059020] [ndv4:55432:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.064844] [ndv4:55432:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib7 vendor_id: 0x15b3 device_id: 4124
[1650465630.065168] [ndv4:55432:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib7: disable ODP because it's not supported for DevX QP
[1650465630.066033] [ndv4:55432:0] async.c:228 UCX DEBUG added async handler 0x2136760 [id=71 ref 1] uct_ib_async_event_handler() to hash
[1650465630.066056] [ndv4:55432:0] async.c:506 UCX DEBUG listening to async event fd 71 events 0x1 mode thread_spinlock
[1650465630.066059] [ndv4:55432:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib7' (InfiniBand channel adapter) with 1 ports
[1650465630.066182] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465630.066194] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465630.066200] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.066244] [ndv4:55432:0] ib_md.c:1319 UCX DEBUG mlx5_ib7: usin[1650465630.020999] [ndv4:55691:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib6 vendor_id: 0x15b3 device_id: 4124
[1650465630.021350] [ndv4:55691:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib6: disable ODP because it's not supported for DevX QP
[1650465630.043708] [ndv4:55691:0] async.c:228 UCX DEBUG added async handler 0x2aa6050 [id=69 ref 1] uct_ib_async_event_handler() to hash
[1650465630.043739] [ndv4:55691:0] async.c:506 UCX DEBUG listening to async event fd 69 events 0x1 mode thread_spinlock
[1650465630.043744] [ndv4:55691:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib6' (InfiniBand channel adapter) with 1 ports
[1650465630.043917] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.043927] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.043934] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.044005] [ndv4:55691:0] ib_md.c:1319 UCX DEBUG mlx5_ib6: using registration cache
[1650465630.046476] [ndv4:55691:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.046482] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.047671] [ndv4:55691:0] ib_md.c:1604 UCX DEBUG mlx5_ib6: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.047790] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.047797] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.058394] [ndv4:55691:0] topo.c:99 UCX DEBUG bus id 0x107000000 doesn't exist. sys_dev = 6
[1650465630.058401] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465630.059313] [ndv4:55691:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465630.059318] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465630.072100] [ndv4:55691:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465630.072109] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465630.083381] [ndv4:55691:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465630.083388] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465630.097835] [ndv4:55691:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465630.097844] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465630.098119] [ndv4:55691:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.016755] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.016761] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465630.017287] [ndv4:54965:0] ib_iface.c:994 UCX DEBUG iface=0x4ebb0a0: created RC QP 0xdaba on mlx5_ib6:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465630.017681] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[36]=0x4ebb0a0 using rc_verbs/mlx5_ib6:1 on worker 0x23dc840
[1650465630.017812] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.017819] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.017927] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.017933] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.018564] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib6:1
[1650465630.030946] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.030960] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.030963] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.031013] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.031460] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x51dc010 of 8176 bytes with 127 elements
[1650465630.031676] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.031682] [ndv4:54965:0] 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> }
[1650465630.031717] [ndv4:54965:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib6 length=2048) failed: Invalid argument
[1650465630.031721] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.031732] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x5010c40 [id=127 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.031751] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 127 events 0x1 mode thread_spinlock
[1650465630.031762] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[37]=0x5008030 using rc_mlx5/mlx5_ib6:1 on worker 0x23dc840
[1650465630.031833] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.031839] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.032112] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.032116] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.033100] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib6:1
[1650465630.034522] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.034538] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.034541] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.034639] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.035006] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465630.028017] [ndv4:55512:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.028320] [ndv4:55512:0] ib_iface.c:994 UCX DEBUG iface=0x2b0b280: created UD QP 0xdaa1 on mlx5_ib1:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465630.028327] [ndv4:55512:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465630.028847] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465630.029387] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.029394] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.029684] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.029694] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.030080] [ndv4:55512:0] ib_md.c:812 UCX DEBUG registered memory 0x2aea79963000..0x2aea799e8000 on mlx5_ib1 lkey 0x81100 rkey 0x81100 access 0xf flags 0x3e4
[1650465630.030085] [ndv4:55512:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2aea79963018 of 544744 bytes with 128 elements
[1650465630.030091] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465630.030614] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x2b0b280: adding gid fe80::15:5dff:fd33:fffc to hash on device mlx5_ib1 port 1 index 0)
[1650465630.031138] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x2b0b280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 1)
[1650465630.031300] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x2b0b280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 2)
[1650465630.052211] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x2b0b280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 3)
[1650465630.053036] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x2b0b280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 4)
[1650465630.053778] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x2b0b280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 5)
[1650465630.054323] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x2b0b280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 6)
[1650465630.072772] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x2b0b280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 7)
[1650465630.072783] [ndv4:55512:0] 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> }
[1650465630.072834] [ndv4:55512:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.072840] [ndv4:55512:0] async.c:228 UCX DEBUG added async handler 0x19f74c0 [id=96 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.072864] [ndv4:55512:0] async.c:506 UCX DEBUG listening to async event fd 96 events 0x5 mode thread_spinlock
[1650465630.072885] [ndv4:55512:0] ucp_worker.c:1159 UCX DEBUG created interface[15]=0x2b0b280 using ud_mlx5/mlx5_ib1:1 on worker 0x22c98d0
[1650465630.072974] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.072980] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.073121] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.073126] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.094058] [ndv4:55512:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465630.095473] [ndv4:55512:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t [1650465630.016046] [ndv4:54756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465630.017075] [ndv4:54756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465630.017093] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.017096] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.017152] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.017479] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.017484] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465630.018157] [ndv4:54756:0] ib_iface.c:994 UCX DEBUG iface=0x44230a0: created RC QP 0xda60 on mlx5_ib3:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465630.018518] [ndv4:54756:0] ucp_worker.c:1159 UCX DEBUG created interface[21]=0x44230a0 using rc_verbs/mlx5_ib3:1 on worker 0x2e9a8d0
[1650465630.018885] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.018891] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.019102] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.019107] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.019717] [ndv4:54756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465630.021497] [ndv4:54756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.021535] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.021539] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.021620] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.021967] [ndv4:54756:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x4744010 of 8176 bytes with 127 elements
[1650465630.022145] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.022151] [ndv4:54756:0] 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> }
[1650465630.022186] [ndv4:54756:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib3 length=2048) failed: Invalid argument
[1650465630.022190] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.022200] [ndv4:54756:0] async.c:228 UCX DEBUG added async handler 0x3815f80 [id=106 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.022226] [ndv4:54756:0] async.c:506 UCX DEBUG listening to async event fd 106 events 0x1 mode thread_spinlock
[1650465630.022236] [ndv4:54756:0] ucp_worker.c:1159 UCX DEBUG created interface[22]=0x4570030 using rc_mlx5/mlx5_ib3:1 on worker 0x2e9a8d0
[1650465630.022352] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.022359] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.022522] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.022527] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.022939] [ndv4:54756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465630.024035] [ndv4:5[1650465630.016433] [ndv4:55516:0] ib_iface.c:994 UCX DEBUG iface=0x2b84050: created RC QP 0xda9a on mlx5_ib1:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465630.017064] [ndv4:55516:0] ucp_worker.c:1159 UCX DEBUG created interface[11]=0x2b84050 using rc_verbs/mlx5_ib1:1 on worker 0x22b38d0
[1650465630.017217] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.017225] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.017291] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.017296] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.017784] [ndv4:55516:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465630.019268] [ndv4:55516:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.019278] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.019281] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.019304] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.019643] [ndv4:55516:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x2b97010 of 8176 bytes with 127 elements
[1650465630.019876] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.019882] [ndv4:55516:0] 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> }
[1650465630.019940] [ndv4:55516:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650465630.019944] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.019954] [ndv4:55516:0] async.c:228 UCX DEBUG added async handler 0x238afc0 [id=92 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.019972] [ndv4:55516:0] async.c:506 UCX DEBUG listening to async event fd 92 events 0x1 mode thread_spinlock
[1650465630.019982] [ndv4:55516:0] ucp_worker.c:1159 UCX DEBUG created interface[12]=0x2a05010 using rc_mlx5/mlx5_ib1:1 on worker 0x22b38d0
[1650465630.020223] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.020230] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.020410] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.020416] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.020895] [ndv4:55516:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465630.022244] [ndv4:55516:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.022261] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.022265] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.022318] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.022636] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465630.022642] [ndv4:55516:0] 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> }
[1650465630.022675] [ndv4:55516:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650465630.022678] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.022685] [ndv4:55516:0] async.c:228 UCX DEBUG added async handler 0x19eddb0 [id=94 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.022705] [ndv4:55516:0] async.c:506 UCX DEBUG listening to async event fd 94 events 0x1 mode thread_spinlock
[1650465630.023125] [ndv4:55516:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465630.029890] [ndv4:55516:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x2d2f010: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xdac3
[1650465630.030379] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.030387] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.030396] [ndv4:55516:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2ae24be70008 of 151544 bytes with 1052 elements
[1650465630.034249] [ndv4:55516:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae253600000..0x2ae255c00000 on mlx5_ib1 lkey 0x81200 rkey 0x81200 access 0xf flags 0x3e4
[1650465630.034262] [ndv4:55516:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2ae253600018 of 39845864 bytes with 4752 elements
[1650465630.034430] [ndv4:55516:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x2d2f010
[1650465630.034454] [ndv4:55516:0] ucp_worker.c:1159 UCX DEBUG created interface[13]=0x2d2f010 using dc_mlx5/mlx5_ib1:1 on worker 0x22b38d0
[1650465630.034519] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.034527] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.034677] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.034683] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.048011] [ndv4:55516:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465630.049044] [ndv4:55516:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.049457] [ndv4:55516:0] ib_iface.c:994 UCX DEBUG iface=0x2c2e3b0: created UD QP 0xdaa5 on mlx5_ib1:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465630.050050] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465630.050513] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.050521] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.050701] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.050707] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.051087] [ndv4:55516:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae24be95000..0x2ae24bf1a000 on mlx5_ib1 lkey 0x81300 rkey 0x81300 access 0xf flags 0x3e4
[1650465630.051103] [ndv4:55516:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae24be95018 of 544744 bytes with 128 elements
[1650465630.051114] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465630.052056] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x2c2e3b0: adding gid fe80::15:5dff:fd33:fffc to hash on device mlx5_ib1 port 1 index 0)
[1650465630.052821] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x2c2e3b0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 1)
[1650465630.053658] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x2c2e3b0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 2)
[1650465630.054210] [ndv4:55516:0] EBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.094686] [ndv4:54715:0] ib_md.c:1604 UCX DEBUG mlx5_ib2: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.094827] [ndv4:54715:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.094833] [ndv4:54715:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.096569] [ndv4:54715:0] topo.c:99 UCX DEBUG bus id 0x103000000 doesn't exist. sys_dev = 2
[1650465630.096636] [ndv4:54715:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465630.098641] [ndv4:54715:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465630.098646] [ndv4:54715:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
EBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.082543] [ndv4:54861:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.082548] [ndv4:54861:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.082613] [ndv4:54861:0] ib_md.c:1319 UCX DEBUG mlx5_ib3: using registration cache
[1650465630.084781] [ndv4:54861:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.084787] [ndv4:54861:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.084978] [ndv4:54861:0] ib_md.c:1604 UCX DEBUG mlx5_ib3: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.085388] [ndv4:54861:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.085394] [ndv4:54861:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
ud_iface.c:393 UCX DEBUG iface 0x5001060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 7)
[1650465630.063343] [ndv4:55152:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.063353] [ndv4:55152:0] async.c:228 UCX DEBUG added async handler 0x43c0e10 [id=109 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.063378] [ndv4:55152:0] async.c:506 UCX DEBUG listening to async event fd 109 events 0x5 mode thread_spinlock
[1650465630.063392] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[24]=0x5001060 using ud_verbs/mlx5_ib3:1 on worker 0x357a8c0
[1650465630.063407] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.063413] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.063501] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.063506] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.063879] [ndv4:55152:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465630.064794] [ndv4:55152:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.065144] [ndv4:55152:0] ib_iface.c:994 UCX DEBUG iface=0x511f460: created UD QP 0xda79 on mlx5_ib3:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465630.065150] [ndv4:55152:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465630.065792] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465630.065997] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.066004] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.066128] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.066134] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.066873] [ndv4:55152:0] ib_md.c:812 UCX DEBUG registered memory 0x2b3b97ce3000..0x2b3b97d68000 on mlx5_ib3 lkey 0x80f00 rkey 0x80f00 access 0xf flags 0x3e4
[1650465630.066879] [ndv4:55152:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b3b97ce3018 of 544744 bytes with 128 elements
[1650465630.066884] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465630.079289] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x511f460: adding gid fe80::15:5dff:fd33:fffe to hash on device mlx5_ib3 port 1 index 0)
[1650465630.080004] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x511f460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 1)
[1650465630.080023] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x511f460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 2)
[1650465630.080414] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x511f460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 3)
[1650465630.090716] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x511f460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 4)
[1650465630.090966] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x511f460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 5)
[1650465630.091319] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x511f460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 6)
[1650465630.091846] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x511f460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 7)
[1650465630.091853] [ndv4:55152:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib3:1, SLs with AR support = { <none> }, SLs without AEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.067097] [ndv4:54853:0] ib_md.c:1604 UCX DEBUG mlx5_ib2: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.067178] [ndv4:54853:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.067184] [ndv4:54853:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.067852] [ndv4:54853:0] topo.c:99 UCX DEBUG bus id 0x103000000 doesn't exist. sys_dev = 2
[1650465630.067857] [ndv4:54853:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465630.080443] [ndv4:54853:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465630.080450] [ndv4:54853:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465630.081842] [ndv4:54853:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465630.081849] [ndv4:54853:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465630.095452] [ndv4:54853:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465630.095460] [ndv4:54853:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
ndv4:55039:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465630.074426] [ndv4:55039:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465630.074446] [ndv4:55039:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.074504] [ndv4:55039:0] module.c:253 UCX DEBUG loading modules for ucm
[1650465630.074525] [ndv4:55039:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650465630.075659] [ndv4:55039:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.075667] [ndv4:55039:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.075870] [ndv4:55039:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.076023] [ndv4:55039:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465630.076029] [ndv4:55039:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465630.077436] [ndv4:55039:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650465630.077441] [ndv4:55039:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465630.087953] [ndv4:55039:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465630.087960] [ndv4:55039:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465630.088404] [ndv4:55039:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465630.088408] [ndv4:55039:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
g registration cache
[1650465630.067772] [ndv4:55432:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.067778] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.067928] [ndv4:55432:0] ib_md.c:1604 UCX DEBUG mlx5_ib7: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.067962] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465630.067968] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465630.068164] [ndv4:55432:0] topo.c:99 UCX DEBUG bus id 0x108000000 doesn't exist. sys_dev = 7
[1650465630.068168] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465630.068299] [ndv4:55432:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465630.068303] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465630.068431] [ndv4:55432:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465630.068434] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465630.102067] [ndv4:55432:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465630.102076] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465630.104016] [ndv4:55432:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465630.104022] [ndv4:55432:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465630.104089] [ndv4:55432:0] ucp_context.c:1556 UCX DEBUG created ucp context 0x2121df0 0x2121df0 [13 mds 47 tls] features 0x1 tl bitmap 0x7fffffffffff 0x0
[1650465630.035012] [ndv4:54965:0] 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> }
[1650465630.035046] [ndv4:54965:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib6 length=2048) failed: Invalid argument
[1650465630.035050] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.035057] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x4e96360 [id=129 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.035076] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 129 events 0x1 mode thread_spinlock
[1650465630.035698] [ndv4:54965:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465630.042707] [ndv4:54965:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x51de050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xdadd
[1650465630.042810] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.042817] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.042826] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b4b36da9008 of 151544 bytes with 1052 elements
[1650465630.046424] [ndv4:54965:0] ib_md.c:812 UCX DEBUG registered memory 0x2b4b34600000..0x2b4b36c00000 on mlx5_ib6 lkey 0x80e00 rkey 0x80e00 access 0xf flags 0x3e4
[1650465630.046441] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b4b34600018 of 39845864 bytes with 4752 elements
[1650465630.046633] [ndv4:54965:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x51de050
[1650465630.046665] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[38]=0x51de050 using dc_mlx5/mlx5_ib6:1 on worker 0x23dc840
[1650465630.046786] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.046796] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.046923] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.046929] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.047236] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib6:1
[1650465630.048108] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.048442] [ndv4:54965:0] ib_iface.c:994 UCX DEBUG iface=0x53ba340: created UD QP 0xdac4 on mlx5_ib6:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465630.048959] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465630.049261] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.049268] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.049381] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.049385] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.049838] [ndv4:54965:0] ib_md.c:812 UCX DEBUG registered memory 0x2b4b36dce000..0x2b4b36e53000 on mlx5_ib6 lkey 0x80f00 rkey 0x80f00 access 0xf flags 0x3e4
[1650465630.049844] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b4b36dce018 of 544744 bytes with 128 elements
[1650465630.049849] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465630.050292] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x53ba340: adding gid fe80::15:5dff:fd34:1 to hash on device mlx5_ib6 port 1 index 0)
[1650465630.050634] [ndv4:54965:0] ud_iface.c:393 Uheadroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465630.095489] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.095493] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.095543] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.096132] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.096138] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465630.097051] [ndv4:55512:0] ib_iface.c:994 UCX DEBUG iface=0x31340a0: created RC QP 0xdaf5 on mlx5_ib2:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465630.097907] [ndv4:55512:0] ucp_worker.c:1159 UCX DEBUG created interface[16]=0x31340a0 using rc_verbs/mlx5_ib2:1 on worker 0x22c98d0
[1650465630.098041] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.098048] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.098388] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.098393] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.098853] [ndv4:55512:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465630.099890] [ndv4:55512:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.099904] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.099907] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.099958] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.100318] [ndv4:55512:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x3455010 of 8176 bytes with 127 elements
[1650465630.100558] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.100564] [ndv4:55512:0] 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> }
[1650465630.100670] [ndv4:55512:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib2 length=2048) failed: Invalid argument
[1650465630.100675] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.100686] [ndv4:55512:0] async.c:228 UCX DEBUG added async handler 0x1a12610 [id=99 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.100710] [ndv4:55512:0] async.c:506 UCX DEBUG listening to async event fd 99 events 0x1 mode thread_spinlock
[1650465630.100737] [ndv4:55512:0] ucp_worker.c:1159 UCX DEBUG created interface[17]=0x3281030 using rc_mlx5/mlx5_ib2:1 on worker 0x22c98d0
[1650465630.100930] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.100936] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.101138] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.101144] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.101957] [ndv4:55512:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465630.103548] [ndv4:55512:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.103564] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_4756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.024053] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.024056] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.024114] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.024446] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465630.024452] [ndv4:54756:0] 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> }
[1650465630.024484] [ndv4:54756:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib3 length=2048) failed: Invalid argument
[1650465630.024488] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.024495] [ndv4:54756:0] async.c:228 UCX DEBUG added async handler 0x43fe9d0 [id=108 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.024515] [ndv4:54756:0] async.c:506 UCX DEBUG listening to async event fd 108 events 0x1 mode thread_spinlock
[1650465630.024824] [ndv4:54756:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465630.030990] [ndv4:54756:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x4746050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xda7d
[1650465630.031269] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.031277] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.031289] [ndv4:54756:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b3b53fb2008 of 151544 bytes with 1052 elements
[1650465630.036239] [ndv4:54756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b3b60200000..0x2b3b62800000 on mlx5_ib3 lkey 0x80c00 rkey 0x80c00 access 0xf flags 0x3e4
[1650465630.036262] [ndv4:54756:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b3b60200018 of 39845864 bytes with 4752 elements
[1650465630.036430] [ndv4:54756:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x4746050
[1650465630.036466] [ndv4:54756:0] ucp_worker.c:1159 UCX DEBUG created interface[23]=0x4746050 using dc_mlx5/mlx5_ib3:1 on worker 0x2e9a8d0
[1650465630.036644] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.036656] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.036816] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.036822] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.037918] [ndv4:54756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465630.039001] [ndv4:54756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.039448] [ndv4:54756:0] ib_iface.c:994 UCX DEBUG iface=0x4922060: created UD QP 0xda77 on mlx5_ib3:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465630.040034] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465630.040185] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.040191] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.040327] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.040335] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.040702] [ndv4:5475R support = { <none> }
[1650465630.091887] [ndv4:55152:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.091892] [ndv4:55152:0] async.c:228 UCX DEBUG added async handler 0x2cb3f80 [id=110 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.091914] [ndv4:55152:0] async.c:506 UCX DEBUG listening to async event fd 110 events 0x5 mode thread_spinlock
[1650465630.091925] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[25]=0x511f460 using ud_mlx5/mlx5_ib3:1 on worker 0x357a8c0
[1650465630.092014] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.092020] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.092158] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.092163] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.092779] [ndv4:55152:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465630.094036] [ndv4:55152:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465630.094051] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.101935] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.102001] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.102810] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.102817] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465630.104065] [ndv4:55152:0] ib_iface.c:994 UCX DEBUG iface=0x521f0a0: created RC QP 0xdad0 on mlx5_ib4:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465630.105061] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[26]=0x521f0a0 using rc_verbs/mlx5_ib4:1 on worker 0x357a8c0
[1650465630.105335] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.105342] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.105553] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.105558] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.106897] [ndv4:55152:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
CX DEBUG iface 0x53ba340: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 1)
[1650465630.060071] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x53ba340: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 2)
[1650465630.060532] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x53ba340: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 3)
[1650465630.070978] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x53ba340: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 4)
[1650465630.071279] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x53ba340: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 5)
[1650465630.071298] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x53ba340: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 6)
[1650465630.071313] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x53ba340: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 7)
[1650465630.071633] [ndv4:54965:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.071642] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x53bae90 [id=130 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.071670] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 130 events 0x5 mode thread_spinlock
[1650465630.071684] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[39]=0x53ba340 using ud_verbs/mlx5_ib6:1 on worker 0x23dc840
[1650465630.104841] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.104852] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.104896] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.104900] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.103567] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.103681] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.104083] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465630.104090] [ndv4:55512:0] 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> }
[1650465630.104125] [ndv4:55512:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib2 length=2048) failed: Invalid argument
[1650465630.104130] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.104139] [ndv4:55512:0] async.c:228 UCX DEBUG added async handler 0x27c1ef0 [id=101 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.104162] [ndv4:55512:0] async.c:506 UCX DEBUG listening to async event fd 101 events 0x1 mode thread_spinlock
[1650465630.104734] [ndv4:55512:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
6:0] ib_md.c:812 UCX DEBUG registered memory 0x2b3b62940000..0x2b3b629c5000 on mlx5_ib3 lkey 0x80d00 rkey 0x80d00 access 0xf flags 0x3e4
[1650465630.040715] [ndv4:54756:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b3b62940018 of 544744 bytes with 128 elements
[1650465630.040725] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465630.041393] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x4922060: adding gid fe80::15:5dff:fd33:fffe to hash on device mlx5_ib3 port 1 index 0)
[1650465630.041791] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x4922060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 1)
[1650465630.042100] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x4922060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 2)
[1650465630.042288] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x4922060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 3)
[1650465630.042459] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x4922060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 4)
[1650465630.042607] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x4922060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 5)
[1650465630.042751] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x4922060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 6)
[1650465630.105477] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x4922060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 7)
[1650465630.106287] [ndv4:54756:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.106297] [ndv4:54756:0] async.c:228 UCX DEBUG added async handler 0x3d0ee40 [id=109 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.106327] [ndv4:54756:0] async.c:506 UCX DEBUG listening to async event fd 109 events 0x5 mode thread_spinlock
[1650465630.106340] [ndv4:54756:0] ucp_worker.c:1159 UCX DEBUG created interface[24]=0x4922060 using ud_verbs/mlx5_ib3:1 on worker 0x2e9a8d0
[1650465630.106434] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.106441] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.106818] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.106825] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.107308] [ndv4:54756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
ud_iface.c:393 UCX DEBUG iface 0x2c2e3b0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 3)
[1650465630.054650] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x2c2e3b0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 4)
[1650465630.054941] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x2c2e3b0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 5)
[1650465630.055210] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x2c2e3b0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 6)
[1650465630.055354] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x2c2e3b0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 7)
[1650465630.055566] [ndv4:55516:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.055595] [ndv4:55516:0] async.c:228 UCX DEBUG added async handler 0x19e2340 [id=95 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.055640] [ndv4:55516:0] async.c:506 UCX DEBUG listening to async event fd 95 events 0x5 mode thread_spinlock
[1650465630.055665] [ndv4:55516:0] ucp_worker.c:1159 UCX DEBUG created interface[14]=0x2c2e3b0 using ud_verbs/mlx5_ib1:1 on worker 0x22b38d0
[1650465630.055746] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.104128] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.105054] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.105064] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.106091] [ndv4:55516:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465630.107930] [ndv4:55516:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.108441] [ndv4:55516:0] ib_iface.c:994 UCX DEBUG iface=0x2af5280: created UD QP 0xdaa6 on mlx5_ib1:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465630.108449] [ndv4:55516:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465630.106888] [ndv4:54717:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.106899] [ndv4:54717:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.107066] [ndv4:54717:0] ib_md.c:1604 UCX DEBUG mlx5_ib2: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.107202] [ndv4:54717:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.107209] [ndv4:54717:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.108413] [ndv4:55152:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.108428] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.108432] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.108482] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.108969] [ndv4:55152:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x5540010 of 8176 bytes with 127 elements
[1650465630.109194] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.109201] [ndv4:55152:0] 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> }
[1650465630.109238] [ndv4:55152:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib4 length=2048) failed: Invalid argument
[1650465630.109242] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.109252] [ndv4:55152:0] async.c:228 UCX DEBUG added async handler 0x5374da0 [id=113 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.109280] [ndv4:55152:0] async.c:506 UCX DEBUG listening to async event fd 113 events 0x1 mode thread_spinlock
[1650465630.109292] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[27]=0x536c030 using rc_mlx5/mlx5_ib4:1 on worker 0x357a8c0
[1650465630.109423] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.109432] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.109812] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.109818] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.109158] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465630.109334] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.109341] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.109432] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.109441] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.109920] [ndv4:55516:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae24bf1a000..0x2ae24bf9f000 on mlx5_ib1 lkey 0x81400 rkey 0x81400 access 0xf flags 0x3e4
[1650465630.109930] [ndv4:55516:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae24bf1a018 of 544744 bytes with 128 elements
[1650465630.109939] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465630.110268] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x2af5280: adding gid fe80::15:5dff:fd33:fffc to hash on device mlx5_ib1 port 1 index 0)
[1650465630.110294] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x2af5280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 1)
[1650465630.110528] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x2af5280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 2)
[1650465630.110789] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x2af5280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 3)
[1650465630.111041] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x2af5280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 4)
[1650465630.111369] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x2af5280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 5)
[1650465630.111496] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x2af5280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 6)
[1650465630.109488] [ndv4:54756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.110134] [ndv4:54756:0] ib_iface.c:994 UCX DEBUG iface=0x4a40460: created UD QP 0xda7a on mlx5_ib3:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465630.110143] [ndv4:54756:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465630.110873] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465630.111095] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.111101] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.111174] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.111179] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.111585] [ndv4:54756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b3b629c5000..0x2b3b62a4a000 on mlx5_ib3 lkey 0x81000 rkey 0x81000 access 0xf flags 0x3e4
[1650465630.111596] [ndv4:54756:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b3b629c5018 of 544744 bytes with 128 elements
[1650465630.111604] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465630.111644] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x4a40460: adding gid fe80::15:5dff:fd33:fffe to hash on device mlx5_ib3 port 1 index 0)
[1650465630.111985] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x4a40460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 1)
[1650465630.111487] [ndv4:55039:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465630.111496] [ndv4:55039:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465630.111826] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x2af5280: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 7)
[1650465630.111840] [ndv4:55516:0] 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> }
[1650465630.111934] [ndv4:55516:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.111942] [ndv4:55516:0] async.c:228 UCX DEBUG added async handler 0x19e14c0 [id=96 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.111970] [ndv4:55516:0] async.c:506 UCX DEBUG listening to async event fd 96 events 0x5 mode thread_spinlock
[1650465630.111996] [ndv4:55516:0] ucp_worker.c:1159 UCX DEBUG created interface[15]=0x2af5280 using ud_mlx5/mlx5_ib1:1 on worker 0x22b38d0
[1650465630.112153] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.112160] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.112517] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.112524] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.122248] [ndv4:55516:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465630.123393] [ndv4:55516:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465630.123409] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.123413] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.123466] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.123865] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.123881] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465630.124385] [ndv4:55516:0] ib_iface.c:994 UCX DEBUG iface=0x311e0a0: created RC QP 0xdb01 on mlx5_ib2:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465630.124832] [ndv4:55516:0] ucp_worker.c:1159 UCX DEBUG created interface[16]=0x311e0a0 using rc_verbs/mlx5_ib2:1 on worker 0x22b38d0
[1650465630.124957] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.124964] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.125108] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.125114] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.140302] [ndv4:55516:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465630.150111] [ndv4:55516:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.150131] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.150135] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.150186] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.150539] [ndv4:55516:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x343f010 of 8176 bytes with 127 elements
[1650465630.150773] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.150781] [ndv4:55516:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx[1650465630.122694] [ndv4:55691:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib7 vendor_id: 0x15b3 device_id: 4124
[1650465630.123032] [ndv4:55691:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib7: disable ODP because it's not supported for DevX QP
[1650465630.137606] [ndv4:55691:0] async.c:228 UCX DEBUG added async handler 0x2aa7210 [id=71 ref 1] uct_ib_async_event_handler() to hash
[1650465630.137635] [ndv4:55691:0] async.c:506 UCX DEBUG listening to async event fd 71 events 0x1 mode thread_spinlock
[1650465630.137639] [ndv4:55691:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib7' (InfiniBand channel adapter) with 1 ports
[1650465630.137902] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465630.137910] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465630.137916] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.137964] [ndv4:55691:0] ib_md.c:1319 UCX DEBUG mlx5_ib7: using registration cache
[1650465630.139657] [ndv4:55691:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.139664] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.139797] [ndv4:55691:0] ib_md.c:1604 UCX DEBUG mlx5_ib7: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.140029] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465630.140035] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465630.161613] [ndv4:55691:0] topo.c:99 UCX DEBUG bus id 0x108000000 doesn't exist. sys_dev = 7
[1650465630.161620] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465630.118158] [ndv4:54717:0] topo.c:99 UCX DEBUG bus id 0x103000000 doesn't exist. sys_dev = 2
[1650465630.118166] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465630.118657] [ndv4:54717:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465630.118663] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465630.119030] [ndv4:54717:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465630.119034] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465630.119245] [ndv4:54717:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465630.119248] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465630.119473] [ndv4:54717:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465630.119476] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465630.119765] [ndv4:54717:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.125934] [ndv4:54717:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib3 vendor_id: 0x15b3 device_id: 4124
[1650465630.126231] [ndv4:54717:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib3: disable ODP because it's not supported for DevX QP
[1650465630.126733] [ndv4:54717:0] async.c:228 UCX DEBUG added async handler 0x23e12b0 [id=63 ref 1] uct_ib_async_event_handler() to hash
[1650465630.126774] [ndv4:54717:0] async.c:506 UCX DEBUG listening to async event fd 63 events 0x1 mode thread_spinlock
[1650465630.126778] [ndv4:54717:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib3' (InfiniBand channel adapter) with 1 ports
[1650465630.127004] [ndv4:54717:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.127012] [ndv4:54717:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.127018] [ndv4:54717:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.127062] [ndv4:54717:0] ib_md.c:1319 UCX DEBUG mlx5_ib3: using registration cache
[1650465630.127961] [ndv4:54717:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.127967] [ndv4:54717:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.128129] [ndv4:54717:0] ib_md.c:1604 UCX DEBUG mlx5_ib3: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.128203] [ndv4:54717:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.128209] [ndv4:54717:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.141367] [ndv4:54717:0] topo.c:99 UCX DEBUG bus id 0x104000000 doesn't exist. sys_dev = 3
[1650465630.141380] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465630.142364] [ndv4:54717:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465630.142369] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465630.164446] [ndv4:54717:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465630.164466] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465630.168240] [ndv4:54717:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465630.168247] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465630.171163] [ndv4:54717:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465630.171169] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465630.171507] [ndv4:5471[1650465630.122254] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib6:1
[1650465630.123610] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.124154] [ndv4:54965:0] ib_iface.c:994 UCX DEBUG iface=0x54d8460: created UD QP 0xdac5 on mlx5_ib6:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465630.124161] [ndv4:54965:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465630.124885] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465630.124914] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.124920] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.124932] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.124935] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.125385] [ndv4:54965:0] ib_md.c:812 UCX DEBUG registered memory 0x2b4b36e53000..0x2b4b36ed8000 on mlx5_ib6 lkey 0x81000 rkey 0x81000 access 0xf flags 0x3e4
[1650465630.125390] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b4b36e53018 of 544744 bytes with 128 elements
[1650465630.125395] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465630.125503] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x54d8460: adding gid fe80::15:5dff:fd34:1 to hash on device mlx5_ib6 port 1 index 0)
[1650465630.125522] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x54d8460: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 1)
[1650465630.125651] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x54d8460: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 2)
[1650465630.125711] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x54d8460: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 3)
[1650465630.125749] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x54d8460: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 4)
[1650465630.125771] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x54d8460: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 5)
[1650465630.125804] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x54d8460: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 6)
[1650465630.139002] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x54d8460: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 7)
[1650465630.139013] [ndv4:54965:0] 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> }
[1650465630.139043] [ndv4:54965:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.139048] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x4e96970 [id=131 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.139070] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 131 events 0x5 mode thread_spinlock
[1650465630.139082] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[40]=0x54d8460 using ud_mlx5/mlx5_ib6:1 on worker 0x23dc840
[1650465630.139098] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465630.139103] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465630.139277] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465630.139282] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465630.148176] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on 5_ib2:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465630.150817] [ndv4:55516:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib2 length=2048) failed: Invalid argument
[1650465630.150821] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.150838] [ndv4:55516:0] async.c:228 UCX DEBUG added async handler 0x19fc610 [id=99 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.150868] [ndv4:55516:0] async.c:506 UCX DEBUG listening to async event fd 99 events 0x1 mode thread_spinlock
[1650465630.150901] [ndv4:55516:0] ucp_worker.c:1159 UCX DEBUG created interface[17]=0x326b030 using rc_mlx5/mlx5_ib2:1 on worker 0x22b38d0
[1650465630.151123] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.151132] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.151254] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.151263] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.151649] [ndv4:55516:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465630.152927] [ndv4:55516:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.152949] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.152953] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.153014] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.153367] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465630.153376] [ndv4:55516:0] 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> }
[1650465630.153411] [ndv4:55516:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib2 length=2048) failed: Invalid argument
[1650465630.153416] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.153424] [ndv4:55516:0] async.c:228 UCX DEBUG added async handler 0x27abef0 [id=101 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.153444] [ndv4:55516:0] async.c:506 UCX DEBUG listening to async event fd 101 events 0x1 mode thread_spinlock
[1650465630.153923] [ndv4:55516:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465630.161454] [ndv4:55516:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x3441050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xdb18
[1650465630.161759] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.161767] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.161777] [ndv4:55516:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2ae24bfa1008 of 151544 bytes with 1052 elements
[1650465630.165488] [ndv4:55516:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae255e00000..0x2ae258400000 on mlx5_ib2 lkey 0x81500 rkey 0x81500 access 0xf flags 0x3e4
[1650465630.165503] [ndv4:55516:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2ae255e00018 of 39845864 bytes with 4752 elements
[1650465630.165689] [ndv4:55516:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x3441050
[1650465630.165717] [ndv4:55516:0] ucp_worker.c:1159 UCX DEBUG created interface[18]=0x3441050 using dc_mlx5/mlx5_ib2:1 on worker 0x22b38d0
[1650465630.165873] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[16504[1650465630.119225] [ndv4:55152:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465630.120438] [ndv4:55152:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.120457] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.120461] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.120519] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.121071] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465630.121083] [ndv4:55152:0] 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> }
[1650465630.121125] [ndv4:55152:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib4 length=2048) failed: Invalid argument
[1650465630.121129] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.121143] [ndv4:55152:0] async.c:228 UCX DEBUG added async handler 0x5374fc0 [id=115 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.121174] [ndv4:55152:0] async.c:506 UCX DEBUG listening to async event fd 115 events 0x1 mode thread_spinlock
[1650465630.121900] [ndv4:55152:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465630.130294] [ndv4:55152:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x5542050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xdae6
[1650465630.130720] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.130729] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.130740] [ndv4:55152:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b3b97d6a008 of 151544 bytes with 1052 elements
[1650465630.134536] [ndv4:55152:0] ib_md.c:812 UCX DEBUG registered memory 0x2b3ba6200000..0x2b3ba8800000 on mlx5_ib4 lkey 0x80900 rkey 0x80900 access 0xf flags 0x3e4
[1650465630.134567] [ndv4:55152:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b3ba6200018 of 39845864 bytes with 4752 elements
[1650465630.134763] [ndv4:55152:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x5542050
[1650465630.134800] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[28]=0x5542050 using dc_mlx5/mlx5_ib4:1 on worker 0x357a8c0
[1650465630.135066] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.135077] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.135137] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.135142] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.153262] [ndv4:55152:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465630.154664] [ndv4:55152:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.155195] [ndv4:55152:0] ib_iface.c:994 UCX DEBUG iface=0x571e220: created UD QP 0xdae9 on mlx5_ib4:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465630.155878] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465630.156010] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.156017] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.156099] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[165046563[1650465630.121476] [ndv4:55039:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465630.121483] [ndv4:55039:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465630.121738] [ndv4:55039:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.128161] [ndv4:55039:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650465630.128480] [ndv4:55039:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650465630.129224] [ndv4:55039:0] async.c:228 UCX DEBUG added async handler 0x19d5f50 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650465630.129251] [ndv4:55039:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650465630.129256] [ndv4:55039:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650465630.129417] [ndv4:55039:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.129425] [ndv4:55039:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.129430] [ndv4:55039:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.129482] [ndv4:55039:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650465630.130689] [ndv4:55039:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.130696] [ndv4:55039:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.130899] [ndv4:55039:0] ib_md.c:1604 UCX DEBUG mlx5_ib1: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.131002] [ndv4:55039:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.131008] [ndv4:55039:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.131778] [ndv4:55039:0] topo.c:99 UCX DEBUG bus id 0x102000000 doesn't exist. sys_dev = 1
[1650465630.131783] [ndv4:55039:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465630.143938] [ndv4:55039:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465630.143944] [ndv4:55039:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465630.146970] [ndv4:55039:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465630.146976] [ndv4:55039:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465630.148708] [ndv4:55039:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465630.148713] [ndv4:55039:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465630.150484] [ndv4:55039:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650465630.150490] [ndv4:55039:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650465630.150742] [ndv4:55039:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.158650] [ndv4:55039:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib2 vendor_id: 0x15b3 device_id: 4124
[1650465630.159015] [ndv4:55039:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib2: disable ODP because it's not supported for DevX QP
[1650465630.171221] [ndv4:55039:0] async.c:228 UCX DEBUG added async handler 0x19ddce0 [id=61 ref 1] uct_ib_async_event_handler() to hash
[1650465630.171265] [ndv4:55039:0] async.c:506 UCX DEBUG listening to async event fd 61 events 0x1 mode thread_spinlock
[1650465630.171272] [ndv4:55039:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib2' (InfiniBand channel adapter) with 1 ports
[1650465630.171931] [ndv4:55039:0] ib_md.c:296 UCX D[1650465630.119657] [ndv4:54715:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465630.119665] [ndv4:54715:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465630.119824] [ndv4:54715:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465630.119828] [ndv4:54715:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465630.120311] [ndv4:54715:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465630.120315] [ndv4:54715:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465630.120526] [ndv4:54715:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.126435] [ndv4:54715:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib3 vendor_id: 0x15b3 device_id: 4124
[1650465630.126760] [ndv4:54715:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib3: disable ODP because it's not supported for DevX QP
[1650465630.127231] [ndv4:54715:0] async.c:228 UCX DEBUG added async handler 0x20063a0 [id=63 ref 1] uct_ib_async_event_handler() to hash
[1650465630.127256] [ndv4:54715:0] async.c:506 UCX DEBUG listening to async event fd 63 events 0x1 mode thread_spinlock
[1650465630.127260] [ndv4:54715:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib3' (InfiniBand channel adapter) with 1 ports
[1650465630.127731] [ndv4:54715:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.127742] [ndv4:54715:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.127747] [ndv4:54715:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.127788] [ndv4:54715:0] ib_md.c:1319 UCX DEBUG mlx5_ib3: using registration cache
[1650465630.128769] [ndv4:54715:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.128774] [ndv4:54715:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.129009] [ndv4:54715:0] ib_md.c:1604 UCX DEBUG mlx5_ib3: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.129092] [ndv4:54715:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.129098] [ndv4:54715:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.151033] [ndv4:54715:0] topo.c:99 UCX DEBUG bus id 0x104000000 doesn't exist. sys_dev = 3
[1650465630.151041] [ndv4:54715:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465630.115317] [ndv4:55324:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465630.115325] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465630.117406] [ndv4:55324:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465630.117412] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465630.117739] [ndv4:55324:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.126006] [ndv4:55324:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib4 vendor_id: 0x15b3 device_id: 4124
[1650465630.126381] [ndv4:55324:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib4: disable ODP because it's not supported for DevX QP
[1650465630.126893] [ndv4:55324:0] async.c:228 UCX DEBUG added async handler 0x1661f90 [id=65 ref 1] uct_ib_async_event_handler() to hash
[1650465630.126922] [ndv4:55324:0] async.c:506 UCX DEBUG listening to async event fd 65 events 0x1 mode thread_spinlock
[1650465630.126926] [ndv4:55324:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib4' (InfiniBand channel adapter) with 1 ports
[1650465630.127116] [ndv4:55324:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.127124] [ndv4:55324:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.127129] [ndv4:55324:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.127181] [ndv4:55324:0] ib_md.c:1319 UCX DEBUG mlx5_ib4: using registration cache
[1650465630.127800] [ndv4:55324:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.127805] [ndv4:55324:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.127979] [ndv4:55324:0] ib_md.c:1604 UCX DEBUG mlx5_ib4: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.127999] [ndv4:55324:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.128005] [ndv4:55324:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.141645] [ndv4:55324:0] topo.c:99 UCX DEBUG bus id 0x105000000 doesn't exist. sys_dev = 4
[1650465630.141652] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465630.152393] [ndv4:55324:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465630.152399] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465630.154948] [ndv4:55324:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465630.154954] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465630.157562] [ndv4:55324:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465630.157568] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465630.158921] [ndv4:55324:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465630.158927] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465630.159163] [ndv4:55324:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.117509] [ndv4:54861:0] topo.c:99 UCX DEBUG bus id 0x104000000 doesn't exist. sys_dev = 3
[1650465630.117520] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465630.129747] [ndv4:54861:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465630.129764] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465630.150708] [ndv4:54861:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465630.150718] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465630.151603] [ndv4:54861:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465630.151608] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465630.152237] [ndv4:54861:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465630.152241] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465630.152523] [ndv4:54861:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.170735] [ndv4:54861:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib4 vendor_id: 0x15b3 device_id: 4124
[1650465630.171185] [ndv4:54861:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib4: disable ODP because it's not supported for DevX QP
[1650465630.172070] [ndv4:54861:0] async.c:228 UCX DEBUG added async handler 0x1fa4d30 [id=65 ref 1] uct_ib_async_event_handler() to hash
[1650465630.172100] [ndv4:54861:0] async.c:506 UCX DEBUG listening to async event fd 65 events 0x1 mode thread_spinlock
[1650465630.172104] [ndv4:54861:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib4' (InfiniBand channel adapter) with 1 ports
[1650465630.172418] [ndv4:54861:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.172427] [ndv4:54861:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.172435] [ndv4:54861:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.172489] [ndv4:54861:0] ib_md.c:1319 UCX DEBUG mlx5_ib4: using registration cache
[1650465630.174338] [ndv4:54861:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.174345] [ndv4:54861:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.174632] [ndv4:54861:0] ib_md.c:1604 UCX DEBUG mlx5_ib4: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.174765] [ndv4:54861:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.174772] [ndv4:54861:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.176322] [ndv4:54861:0] topo.c:99 UCX DEBUG bus id 0x105000000 doesn't exist. sys_dev = 4
[1650465630.176328] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465630.114842] [ndv4:55512:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x3457050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xdb0b
[1650465630.115188] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.115196] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.115340] [ndv4:55512:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2aea7c1e9008 of 151544 bytes with 1052 elements
[1650465630.119007] [ndv4:55512:0] ib_md.c:812 UCX DEBUG registered memory 0x2aea79a00000..0x2aea7c000000 on mlx5_ib2 lkey 0x81300 rkey 0x81300 access 0xf flags 0x3e4
[1650465630.119026] [ndv4:55512:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2aea79a00018 of 39845864 bytes with 4752 elements
[1650465630.119164] [ndv4:55512:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x3457050
[1650465630.119200] [ndv4:55512:0] ucp_worker.c:1159 UCX DEBUG created interface[18]=0x3457050 using dc_mlx5/mlx5_ib2:1 on worker 0x22c98d0
[1650465630.119241] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.119252] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.119320] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.119326] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.135156] [ndv4:55512:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465630.136388] [ndv4:55512:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.136766] [ndv4:55512:0] ib_iface.c:994 UCX DEBUG iface=0x3633060: created UD QP 0xdb02 on mlx5_ib2:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465630.137380] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465630.137696] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.137703] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.137810] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.137815] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.138380] [ndv4:55512:0] ib_md.c:812 UCX DEBUG registered memory 0x2aea7c20e000..0x2aea7c293000 on mlx5_ib2 lkey 0x81400 rkey 0x81400 access 0xf flags 0x3e4
[1650465630.138386] [ndv4:55512:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2aea7c20e018 of 544744 bytes with 128 elements
[1650465630.138391] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465630.138913] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x3633060: adding gid fe80::15:5dff:fd33:fffd to hash on device mlx5_ib2 port 1 index 0)
[1650465630.139464] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x3633060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 1)
[1650465630.139486] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x3633060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 2)
[1650465630.140370] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x3633060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 3)
[1650465630.141095] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x3633060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 4)
[1650465630.141833] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x3633060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 5)
[1650465630.142275] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x3633060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 6)
[1650465630.163136] [ndv4:55512:0] [1650465630.116983] [ndv4:54853:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465630.116992] [ndv4:54853:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465630.117352] [ndv4:54853:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.147418] [ndv4:54853:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib3 vendor_id: 0x15b3 device_id: 4124
[1650465630.147799] [ndv4:54853:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib3: disable ODP because it's not supported for DevX QP
[1650465630.159497] [ndv4:54853:0] async.c:228 UCX DEBUG added async handler 0x138c3a0 [id=63 ref 1] uct_ib_async_event_handler() to hash
[1650465630.159525] [ndv4:54853:0] async.c:506 UCX DEBUG listening to async event fd 63 events 0x1 mode thread_spinlock
[1650465630.159529] [ndv4:54853:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib3' (InfiniBand channel adapter) with 1 ports
[1650465630.159684] [ndv4:54853:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.159692] [ndv4:54853:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.159698] [ndv4:54853:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.159763] [ndv4:54853:0] ib_md.c:1319 UCX DEBUG mlx5_ib3: using registration cache
[1650465630.160592] [ndv4:54853:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.160597] [ndv4:54853:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.160792] [ndv4:54853:0] ib_md.c:1604 UCX DEBUG mlx5_ib3: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.160903] [ndv4:54853:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.160909] [ndv4:54853:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.173499] [ndv4:54853:0] topo.c:99 UCX DEBUG bus id 0x104000000 doesn't exist. sys_dev = 3
[1650465630.173514] [ndv4:54853:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465630.120868] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x4a40460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 2)
[1650465630.121098] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x4a40460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 3)
[1650465630.121200] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x4a40460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 4)
[1650465630.121214] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x4a40460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 5)
[1650465630.121327] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x4a40460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 6)
[1650465630.121342] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x4a40460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 7)
[1650465630.121347] [ndv4:54756:0] 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> }
[1650465630.121436] [ndv4:54756:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.121441] [ndv4:54756:0] async.c:228 UCX DEBUG added async handler 0x25d3fb0 [id=110 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.121464] [ndv4:54756:0] async.c:506 UCX DEBUG listening to async event fd 110 events 0x5 mode thread_spinlock
[1650465630.121474] [ndv4:54756:0] ucp_worker.c:1159 UCX DEBUG created interface[25]=0x4a40460 using ud_mlx5/mlx5_ib3:1 on worker 0x2e9a8d0
[1650465630.121647] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.121653] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.121779] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.121784] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.122144] [ndv4:54756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465630.123063] [ndv4:54756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465630.123082] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.123085] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.123141] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.123459] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.123464] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465630.124010] [ndv4:54756:0] ib_iface.c:994 UCX DEBUG iface=0x4b400a0: created RC QP 0xdad2 on mlx5_ib4:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465630.124601] [ndv4:54756:0] ucp_worker.c:1159 UCX DEBUG created interface[26]=0x4b400a0 using rc_verbs/mlx5_ib4:1 on worker 0x2e9a8d0
[1650465630.124686] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.124692] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.124842] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.124851] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.125249] [ndv4:54756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465630.126492] [ndv4:54756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.126512] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.126517] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.126658] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.126945] [ndv4:54756:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x4e61010 of 8176 bytes with 127 elements
[1650465630.127130] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.127138] [ndv4:54756:0] 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> }
[1650465630.127264] [ndv4:54756:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib4 length=2048) failed: Invalid argument
[1650465630.127269] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.127281] [ndv4:54756:0] async.c:228 UCX DEBUG added async handler 0x3c06970 [id=113 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.127304] [ndv4:54756:0] async.c:506 UCX DEBUG listening to async event fd 113 events 0x1 mode thread_spinlock
[1650465630.127318] [ndv4:54756:0] ucp_worker.c:1159 UCX DEBUG created interface[27]=0x4c8d030 using rc_mlx5/mlx5_ib4:1 on worker 0x2e9a8d0
[1650465630.127389] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.127396] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.127479] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.127485] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.127727] [ndv4:54756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465630.128870] [ndv4:54756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.128886] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.128889] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.128971] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.129283] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465630.129290] [ndv4:54756:0] 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> }
[1650465630.129323] [ndv4:54756:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib4 length=2048) failed: Invalid argument
[1650465630.129327] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.129334] [ndv4:54756:0] async.c:228 UCX DEBUG added async handler 0x4e6ffc0 [id=115 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.129372] [ndv4:54756:0] async.c:506 UCX DEBUG listening to async event fd 115 events 0x1 mode thread_spinlock
[1650465630.129662] [ndv4:54756:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465630.135245] [ndv4:54756:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x4e63050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xdaf0
[1650465630.135383] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.135390] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.135398] [ndv4:54756:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b365630.165880] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.166002] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.166008] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
0.156104] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.156494] [ndv4:55152:0] ib_md.c:812 UCX DEBUG registered memory 0x2b3b97d8f000..0x2b3b97e14000 on mlx5_ib4 lkey 0x80b00 rkey 0x80b00 access 0xf flags 0x3e4
[1650465630.156502] [ndv4:55152:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b3b97d8f018 of 544744 bytes with 128 elements
[1650465630.156508] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465630.157305] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x571e220: adding gid fe80::15:5dff:fd33:ffff to hash on device mlx5_ib4 port 1 index 0)
[1650465630.157796] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x571e220: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 1)
[1650465630.158198] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x571e220: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 2)
[1650465630.158455] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x571e220: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 3)
[1650465630.168295] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x571e220: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 4)
EBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.171941] [ndv4:55039:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.171952] [ndv4:55039:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.172063] [ndv4:55039:0] ib_md.c:1319 UCX DEBUG mlx5_ib2: using registration cache
[1650465630.173833] [ndv4:55039:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.173840] [ndv4:55039:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.174023] [ndv4:55039:0] ib_md.c:1604 UCX DEBUG mlx5_ib2: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.174195] [ndv4:55039:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.174201] [ndv4:55039:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.174802] [ndv4:55039:0] topo.c:99 UCX DEBUG bus id 0x103000000 doesn't exist. sys_dev = 2
[1650465630.174808] [ndv4:55039:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
ud_iface.c:393 UCX DEBUG iface 0x3633060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 7)
[1650465630.163396] [ndv4:55512:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.163404] [ndv4:55512:0] async.c:228 UCX DEBUG added async handler 0x2c441f0 [id=102 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.163434] [ndv4:55512:0] async.c:506 UCX DEBUG listening to async event fd 102 events 0x5 mode thread_spinlock
[1650465630.163450] [ndv4:55512:0] ucp_worker.c:1159 UCX DEBUG created interface[19]=0x3633060 using ud_verbs/mlx5_ib2:1 on worker 0x22c98d0
[1650465630.163630] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.163637] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.163690] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.163695] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.164011] [ndv4:55512:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465630.165098] [ndv4:55512:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.165455] [ndv4:55512:0] ib_iface.c:994 UCX DEBUG iface=0x3751050: created UD QP 0xdb0e on mlx5_ib2:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465630.165462] [ndv4:55512:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465630.165984] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465630.166157] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.166163] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.166200] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.166206] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.166539] [ndv4:55512:0] ib_md.c:812 UCX DEBUG registered memory 0x2aea7c293000..0x2aea7c318000 on mlx5_ib2 lkey 0x81600 rkey 0x81600 access 0xf flags 0x3e4
[1650465630.166545] [ndv4:55512:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2aea7c293018 of 544744 bytes with 128 elements
[1650465630.166550] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465630.167494] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x3751050: adding gid fe80::15:5dff:fd33:fffd to hash on device mlx5_ib2 port 1 index 0)
[1650465630.167516] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x3751050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 1)
[1650465630.167897] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x3751050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 2)
[1650465630.168270] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x3751050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 3)
[1650465630.168761] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x3751050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 4)
[1650465630.168812] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x3751050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 5)
[1650465630.168827] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x3751050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 6)
[1650465630.169285] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x3751050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 7)
[1650465630.169293] [ndv4:55512:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib2:1, SLs with AR support = { <none> }, SLs without Amlx5_ib7:1
[1650465630.149344] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465630.149361] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.149364] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.149416] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.150423] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.150428] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465630.150972] [ndv4:54965:0] ib_iface.c:994 UCX DEBUG iface=0x55d80a0: created RC QP 0xda4c on mlx5_ib7:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465630.151302] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[41]=0x55d80a0 using rc_verbs/mlx5_ib7:1 on worker 0x23dc840
[1650465630.151429] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465630.151435] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465630.151496] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465630.151500] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465630.151826] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib7:1
[1650465630.153363] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.153378] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.153381] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.153431] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.153889] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x58f9010 of 8176 bytes with 127 elements
[1650465630.154108] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.154114] [ndv4:54965:0] 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> }
[1650465630.154148] [ndv4:54965:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib7 length=2048) failed: Invalid argument
[1650465630.154152] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.154162] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x5505960 [id=134 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.154182] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 134 events 0x1 mode thread_spinlock
[1650465630.154192] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[42]=0x5725030 using rc_mlx5/mlx5_ib7:1 on worker 0x23dc840
[1650465630.154310] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465630.154316] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465630.154467] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465630.154473] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465630.154795] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib7:1
[1650465630.155774] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs [1650465630.174478] [ndv4:55691:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465630.174485] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465630.175144] [ndv4:55691:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465630.175148] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465630.175276] [ndv4:55691:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465630.175279] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465630.176593] [ndv4:55691:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465630.176597] [ndv4:55691:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465630.176723] [ndv4:55691:0] ucp_context.c:1556 UCX DEBUG created ucp context 0x23a8c60 0x23a8c60 [13 mds 47 tls] features 0x1 tl bitmap 0x7fffffffffff 0x0
7:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
b53fd9008 of 151544 bytes with 1052 elements
[1650465630.139036] [ndv4:54756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b3b62c00000..0x2b3b65200000 on mlx5_ib4 lkey 0x80a00 rkey 0x80a00 access 0xf flags 0x3e4
[1650465630.139053] [ndv4:54756:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b3b62c00018 of 39845864 bytes with 4752 elements
[1650465630.139191] [ndv4:54756:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x4e63050
[1650465630.139228] [ndv4:54756:0] ucp_worker.c:1159 UCX DEBUG created interface[28]=0x4e63050 using dc_mlx5/mlx5_ib4:1 on worker 0x2e9a8d0
[1650465630.139392] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.139402] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.139721] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.139727] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.149116] [ndv4:54756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465630.161202] [ndv4:54756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.161476] [ndv4:54756:0] ib_iface.c:994 UCX DEBUG iface=0x503f220: created UD QP 0xdaeb on mlx5_ib4:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
R support = { <none> }
[1650465630.169324] [ndv4:55512:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.169328] [ndv4:55512:0] async.c:228 UCX DEBUG added async handler 0x2bb0cb0 [id=103 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.169349] [ndv4:55512:0] async.c:506 UCX DEBUG listening to async event fd 103 events 0x5 mode thread_spinlock
[1650465630.169359] [ndv4:55512:0] ucp_worker.c:1159 UCX DEBUG created interface[20]=0x3751050 using ud_mlx5/mlx5_ib2:1 on worker 0x22c98d0
[1650465630.169552] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.169557] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.169750] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.169756] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.170361] [ndv4:55512:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
92 hdr_ofs 90 data_sz 8256
[1650465630.155788] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.155791] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.155846] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.156197] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465630.156202] [ndv4:54965:0] 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> }
[1650465630.156236] [ndv4:54965:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib7 length=2048) failed: Invalid argument
[1650465630.156239] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.176948] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x2c1ed50 [id=136 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.176971] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 136 events 0x1 mode thread_spinlock
[1650465630.177522] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465630.177823] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.177831] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.178021] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.178027] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.178428] [ndv4:54756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b3b6524b000..0x2b3b652d0000 on mlx5_ib4 lkey 0x80e00 rkey 0x80e00 access 0xf flags 0x3e4
[1650465630.178443] [ndv4:54756:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b3b6524b018 of 544744 bytes with 128 elements
[1650465630.178452] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465630.179338] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x503f220: adding gid fe80::15:5dff:fd33:ffff to hash on device mlx5_ib4 port 1 index 0)
[1650465630.180146] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x503f220: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 1)
[1650465630.180691] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x503f220: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 2)
[1650465630.180915] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x503f220: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 3)
[1650465630.181662] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x503f220: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 4)
[1650465630.181888] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x503f220: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 5)
[1650465630.177565] [ndv4:54965:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465630.177284] [ndv4:55516:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465630.178170] [ndv4:55516:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.178501] [ndv4:55516:0] ib_iface.c:994 UCX DEBUG iface=0x361d060: created UD QP 0xdb0f on mlx5_ib2:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465630.178982] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465630.179086] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.179092] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.179108] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.179112] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.179483] [ndv4:55516:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae258458000..0x2ae2584dd000 on mlx5_ib2 lkey 0x81700 rkey 0x81700 access 0xf flags 0x3e4
[1650465630.179493] [ndv4:55516:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae258458018 of 544744 bytes with 128 elements
[1650465630.179501] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465630.180223] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x361d060: adding gid fe80::15:5dff:fd33:fffd to hash on device mlx5_ib2 port 1 index 0)
[1650465630.180496] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x361d060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 1)
[1650465630.180876] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x361d060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 2)
[1650465630.180894] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x361d060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 3)
[1650465630.180910] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x361d060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 4)
[1650465630.181253] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x361d060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 5)
[1650465630.181665] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x361d060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 6)
[1650465630.181902] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x361d060: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 7)
[1650465630.177352] [ndv4:54861:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465630.177359] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465630.178570] [ndv4:54861:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465630.178646] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465630.180506] [ndv4:54717:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib4 vendor_id: 0x15b3 device_id: 4124
[1650465630.180856] [ndv4:54717:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib4: disable ODP because it's not supported for DevX QP
[1650465630.181928] [ndv4:54717:0] async.c:228 UCX DEBUG added async handler 0x2acabf0 [id=65 ref 1] uct_ib_async_event_handler() to hash
[1650465630.181959] [ndv4:54717:0] async.c:506 UCX DEBUG listening to async event fd 65 events 0x1 mode thread_spinlock
[1650465630.181962] [ndv4:54717:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib4' (InfiniBand channel adapter) with 1 ports
[1650465630.182116] [ndv4:54717:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.182125] [ndv4:54717:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.182130] [ndv4:54717:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.182176] [ndv4:54717:0] ib_md.c:1319 UCX DEBUG mlx5_ib4: using registration cache
[1650465630.180471] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x571e220: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 5)
[1650465630.180842] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x571e220: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 6)
[1650465630.180863] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x571e220: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 7)
[1650465630.181203] [ndv4:55152:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.181211] [ndv4:55152:0] async.c:228 UCX DEBUG added async handler 0x4a2f3a0 [id=116 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.181234] [ndv4:55152:0] async.c:506 UCX DEBUG listening to async event fd 116 events 0x5 mode thread_spinlock
[1650465630.181246] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[29]=0x571e220 using ud_verbs/mlx5_ib4:1 on worker 0x357a8c0
[1650465630.181263] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.181269] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.181422] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.181427] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.182436] [ndv4:55152:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465630.184518] [ndv4:55152:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.180322] [ndv4:55324:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib5 vendor_id: 0x15b3 device_id: 4124
[1650465630.180776] [ndv4:55324:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib5: disable ODP because it's not supported for DevX QP
[1650465630.181300] [ndv4:55324:0] async.c:228 UCX DEBUG added async handler 0x1663f40 [id=67 ref 1] uct_ib_async_event_handler() to hash
[1650465630.181323] [ndv4:55324:0] async.c:506 UCX DEBUG listening to async event fd 67 events 0x1 mode thread_spinlock
[1650465630.181327] [ndv4:55324:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib5' (InfiniBand channel adapter) with 1 ports
[1650465630.181699] [ndv4:55324:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.181708] [ndv4:55324:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.181713] [ndv4:55324:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.181762] [ndv4:55324:0] ib_md.c:1319 UCX DEBUG mlx5_ib5: using registration cache
[1650465630.182075] [ndv4:55516:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.182083] [ndv4:55516:0] async.c:228 UCX DEBUG added async handler 0x2c2e1f0 [id=102 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.182107] [ndv4:55516:0] async.c:506 UCX DEBUG listening to async event fd 102 events 0x5 mode thread_spinlock
[1650465630.182128] [ndv4:55516:0] ucp_worker.c:1159 UCX DEBUG created interface[19]=0x361d060 using ud_verbs/mlx5_ib2:1 on worker 0x22b38d0
[1650465630.182324] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.182329] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.182493] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.182498] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.183099] [ndv4:55516:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib2:1
[1650465630.183958] [ndv4:55516:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.184215] [ndv4:55516:0] ib_iface.c:994 UCX DEBUG iface=0x373b050: created UD QP 0xdb10 on mlx5_ib2:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465630.184222] [ndv4:55516:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465630.184640] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465630.184676] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.184681] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.184707] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650465630.184712] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650465630.184959] [ndv4:55152:0] ib_iface.c:994 UCX DEBUG iface=0x583c050: created UD QP 0xdaec on mlx5_ib4:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465630.184965] [ndv4:55152:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465630.185643] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465630.185669] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.185674] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.185685] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.185688] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.185011] [ndv4:55516:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae2584dd000..0x2ae258562000 on mlx5_ib2 lkey 0x81800 rkey 0x81800 access 0xf flags 0x3e4
[1650465630.185021] [ndv4:55516:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae2584dd018 of 544744 bytes with 128 elements
[1650465630.185029] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465630.185085] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x373b050: adding gid fe80::15:5dff:fd33:fffd to hash on device mlx5_ib2 port 1 index 0)
[1650465630.185117] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x373b050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 1)
[1650465630.185167] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x373b050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 2)
[1650465630.185198] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x373b050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 3)
[1650465630.185222] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x373b050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 4)
[1650465630.185246] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x373b050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 5)
[1650465630.185276] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x373b050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 6)
[1650465630.185294] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x373b050: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 7)
[1650465630.185302] [ndv4:55516:0] 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> }
[1650465630.185396] [ndv4:55516:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.185401] [ndv4:55516:0] async.c:228 UCX DEBUG added async handler 0x2b9acb0 [id=103 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.185425] [ndv4:55516:0] async.c:506 UCX DEBUG listening to async event fd 103 events 0x5 mode thread_spinlock
[1650465630.185437] [ndv4:55516:0] ucp_worker.c:1159 UCX DEBUG created interface[20]=0x373b050 using ud_mlx5/mlx5_ib2:1 on worker 0x22b38d0
[1650465630.185450] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.185456] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.185511] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.185516] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.185700] [ndv4:55516:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465630.184525] [ndv4:54717:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.184531] [ndv4:54717:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.184690] [ndv4:54717:0] ib_md.c:1604 UCX DEBUG mlx5_ib4: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.184712] [ndv4:54717:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.184718] [ndv4:54717:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.184875] [ndv4:54717:0] topo.c:99 UCX DEBUG bus id 0x105000000 doesn't exist. sys_dev = 4
[1650465630.184879] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465630.184982] [ndv4:54717:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465630.184986] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465630.185083] [ndv4:54717:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465630.185086] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465630.185223] [ndv4:54717:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465630.185227] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465630.185333] [ndv4:54717:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465630.185337] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465630.185553] [ndv4:54717:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.184994] [ndv4:54965:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x58fb050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xda6f
[1650465630.185071] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465630.185078] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465630.185087] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b4b396d9008 of 151544 bytes with 1052 elements
[1650465630.186639] [ndv4:55516:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465630.186658] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.186662] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.186717] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.186097] [ndv4:55152:0] ib_md.c:812 UCX DEBUG registered memory 0x2b3b97e14000..0x2b3b97e99000 on mlx5_ib4 lkey 0x80f00 rkey 0x80f00 access 0xf flags 0x3e4
[1650465630.186103] [ndv4:55152:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b3b97e14018 of 544744 bytes with 128 elements
[1650465630.186108] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465630.186142] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x583c050: adding gid fe80::15:5dff:fd33:ffff to hash on device mlx5_ib4 port 1 index 0)
[1650465630.186158] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x583c050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 1)
[1650465630.186172] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x583c050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 2)
[1650465630.186187] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x583c050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 3)
[1650465630.186201] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x583c050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 4)
[1650465630.186215] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x583c050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 5)
[1650465630.186229] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x583c050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 6)
[1650465630.186242] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x583c050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 7)
[1650465630.186246] [ndv4:55152:0] 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> }
[1650465630.186275] [ndv4:55152:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.186280] [ndv4:55152:0] async.c:228 UCX DEBUG added async handler 0x5227c00 [id=117 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.186300] [ndv4:55152:0] async.c:506 UCX DEBUG listening to async event fd 117 events 0x5 mode thread_spinlock
[1650465630.186309] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[30]=0x583c050 using ud_mlx5/mlx5_ib4:1 on worker 0x357a8c0
[1650465630.186319] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.186323] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.186377] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.186381] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.186564] [ndv4:55152:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib5:1
[1650465630.187198] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.187204] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465630.187494] [ndv4:55152:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465630.187511] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.187514] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.187567] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.188183] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.188190] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465630.188110] [ndv4:55516:0] ib_iface.c:994 UCX DEBUG iface=0x383b0a0: created RC QP 0xda7b on mlx5_ib3:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465630.187997] [ndv4:55512:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465630.188018] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.188021] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.188073] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.189204] [ndv4:55516:0] ucp_worker.c:1159 UCX DEBUG created interface[21]=0x383b0a0 using rc_verbs/mlx5_ib3:1 on worker 0x22b38d0
[1650465630.189233] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.189240] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.189327] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.189334] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.189661] [ndv4:55516:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465630.188754] [ndv4:54965:0] ib_md.c:812 UCX DEBUG registered memory 0x2b4b37000000..0x2b4b39600000 on mlx5_ib7 lkey 0x80800 rkey 0x80800 access 0xf flags 0x3e4
[1650465630.188770] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b4b37000018 of 39845864 bytes with 4752 elements
[1650465630.188911] [ndv4:54965:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x58fb050
[1650465630.188943] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[43]=0x58fb050 using dc_mlx5/mlx5_ib7:1 on worker 0x23dc840
[1650465630.189103] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465630.189114] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465630.189253] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465630.189259] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465630.189568] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib7:1
[1650465630.189475] [ndv4:55152:0] ib_iface.c:994 UCX DEBUG iface=0x593c0a0: created RC QP 0xda60 on mlx5_ib5:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465630.189543] [ndv4:54715:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465630.189552] [ndv4:54715:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465630.189736] [ndv4:54715:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465630.189740] [ndv4:54715:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465630.190033] [ndv4:55039:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465630.190061] [ndv4:55039:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465630.190133] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.190144] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465630.190768] [ndv4:55512:0] ib_iface.c:994 UCX DEBUG iface=0x38510a0: created RC QP 0xda7e on mlx5_ib3:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465630.190435] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[31]=0x593c0a0 using rc_verbs/mlx5_ib5:1 on worker 0x357a8c0
[1650465630.190536] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.190544] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.190707] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.190714] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.190529] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.190880] [ndv4:54965:0] ib_iface.c:994 UCX DEBUG iface=0x5ad7060: created UD QP 0xda55 on mlx5_ib7:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465630.191130] [ndv4:55516:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.191147] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.191150] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.191208] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.191541] [ndv4:55516:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x3b5c010 of 8176 bytes with 127 elements
[1650465630.191769] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.191776] [ndv4:55516:0] 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> }
[1650465630.191816] [ndv4:55516:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib3 length=2048) failed: Invalid argument
[1650465630.191820] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.191831] [ndv4:55516:0] async.c:228 UCX DEBUG added async handler 0x19d7ef0 [id=106 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.191860] [ndv4:55516:0] async.c:506 UCX DEBUG listening to async event fd 106 events 0x1 mode thread_spinlock
[1650465630.191871] [ndv4:55516:0] ucp_worker.c:1159 UCX DEBUG created interface[22]=0x3988030 using rc_mlx5/mlx5_ib3:1 on worker 0x22b38d0
[1650465630.192199] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.192206] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.192309] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.192314] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.191137] [ndv4:55512:0] ucp_worker.c:1159 UCX DEBUG created interface[21]=0x38510a0 using rc_verbs/mlx5_ib3:1 on worker 0x22c98d0
[1650465630.191418] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.191425] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.191487] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.191491] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.191716] [ndv4:55512:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465630.191065] [ndv4:55152:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib5:1
[1650465630.191409] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465630.191873] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465630.191882] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465630.192108] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465630.192113] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465630.192557] [ndv4:54965:0] ib_md.c:812 UCX DEBUG registered memory 0x2b4b396fe000..0x2b4b39783000 on mlx5_ib7 lkey 0x80900 rkey 0x80900 access 0xf flags 0x3e4
[1650465630.192564] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b4b396fe018 of 544744 bytes with 128 elements
[1650465630.192568] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465630.192913] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x5ad7060: adding gid fe80::15:5dff:fd34:2 to hash on device mlx5_ib7 port 1 index 0)
[1650465630.192853] [ndv4:55516:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465630.193453] [ndv4:55512:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.193468] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.193471] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.193521] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.193260] [ndv4:54861:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465630.193267] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465630.193513] [ndv4:54861:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465630.193517] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465630.193435] [ndv4:54717:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib5 vendor_id: 0x15b3 device_id: 4124
[1650465630.193111] [ndv4:55152:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.193128] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.193131] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.193184] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.193672] [ndv4:55152:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x5c5d010 of 8176 bytes with 127 elements
[1650465630.193809] [ndv4:54861:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.193787] [ndv4:54717:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib5: disable ODP because it's not supported for DevX QP
[1650465630.194110] [ndv4:55516:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.194126] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.194129] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.194210] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.194505] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465630.194512] [ndv4:55516:0] 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> }
[1650465630.194548] [ndv4:55516:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib3 length=2048) failed: Invalid argument
[1650465630.194552] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.194559] [ndv4:55516:0] async.c:228 UCX DEBUG added async handler 0x195ab00 [id=108 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.194614] [ndv4:55516:0] async.c:506 UCX DEBUG listening to async event fd 108 events 0x1 mode thread_spinlock
[1650465630.195136] [ndv4:55516:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465630.194181] [ndv4:55432:0] mm_posix.c:518 UCX DEBUG allocated posix shared memory at 0x2aee3be67000 length 12288
[1650465630.194248] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool mm_recv_desc: align 64, maxelems 4294967295, elemsize 8368
[1650465630.195325] [ndv4:55432:0] mm_posix.c:305 UCX DEBUG shared memory mmap(addr=(nil), length=6291456, flags= HUGETLB, fd=76) failed: Invalid argument
[1650465630.195334] [ndv4:55432:0] mm_posix.c:518 UCX DEBUG allocated posix shared memory at 0x2aee40ec7000 length 4296704
[1650465630.195339] [ndv4:55432:0] mpool.c:205 UCX DEBUG mpool mm_recv_desc: allocated chunk 0x2aee40ec7018 of 4296680 bytes with 512 elements
[1650465630.194320] [ndv4:54717:0] async.c:228 UCX DEBUG added async handler 0x23e5710 [id=67 ref 1] uct_ib_async_event_handler() to hash
[1650465630.194346] [ndv4:54717:0] async.c:506 UCX DEBUG listening to async event fd 67 events 0x1 mode thread_spinlock
[1650465630.194349] [ndv4:54717:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib5' (InfiniBand channel adapter) with 1 ports
[1650465630.194500] [ndv4:54717:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.194507] [ndv4:54717:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.194512] [ndv4:54717:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.194567] [ndv4:54717:0] ib_md.c:1319 UCX DEBUG mlx5_ib5: using registration cache
[1650465630.193973] [ndv4:55512:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x3b72010 of 8176 bytes with 127 elements
[1650465630.194183] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.194194] [ndv4:55512:0] 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> }
[1650465630.194272] [ndv4:55512:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib3 length=2048) failed: Invalid argument
[1650465630.194277] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.194287] [ndv4:55512:0] async.c:228 UCX DEBUG added async handler 0x19edef0 [id=106 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.194311] [ndv4:55512:0] async.c:506 UCX DEBUG listening to async event fd 106 events 0x1 mode thread_spinlock
[1650465630.194322] [ndv4:55512:0] ucp_worker.c:1159 UCX DEBUG created interface[22]=0x399e030 using rc_mlx5/mlx5_ib3:1 on worker 0x22c98d0
[1650465630.194349] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.194355] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.194452] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.194456] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.194685] [ndv4:55512:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465630.193938] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.193948] [ndv4:55152:0] 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> }
[1650465630.194021] [ndv4:55152:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib5 length=2048) failed: Invalid argument
[1650465630.194025] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.194038] [ndv4:55152:0] async.c:228 UCX DEBUG added async handler 0x4a2fe50 [id=120 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.194062] [ndv4:55152:0] async.c:506 UCX DEBUG listening to async event fd 120 events 0x1 mode thread_spinlock
[1650465630.194074] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[32]=0x5a89030 using rc_mlx5/mlx5_ib5:1 on worker 0x357a8c0
[1650465630.194153] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.194160] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.194260] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.194265] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.195724] [ndv4:55432:0] mm_iface.c:600 UCX DEBUG created mm iface 0x289ede0 FIFO id 0x400000005232a315 va 0x2aee3be67000 size 12288 (128 x 64 elems)
[1650465630.195774] [ndv4:55432:0] ucp_worker.c:1159 UCX DEBUG created interface[0]=0x289ede0 using posix/memory on worker 0x3178950
[1650465630.195799] [ndv4:55432:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 8447 bytes with hugetlb
[1650465630.195836] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool mm_recv_desc: align 64, maxelems 4294967295, elemsize 8368
[1650465630.195850] [ndv4:55432:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 4292720 bytes with hugetlb
[1650465630.195859] [ndv4:55432:0] mpool.c:205 UCX DEBUG mpool mm_recv_desc: allocated chunk 0x2aee412e0018 of 4296680 bytes with 512 elements
[1650465630.196418] [ndv4:55432:0] mm_iface.c:600 UCX DEBUG created mm iface 0x289f3b0 FIFO id 0x72001f va 0x2aee3be6a000 size 12288 (128 x 64 elems)
[1650465630.196428] [ndv4:55432:0] ucp_worker.c:1159 UCX DEBUG created interface[1]=0x289f3b0 using sysv/memory on worker 0x3178950
[1650465630.196441] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool self_msg_desc: align 64, maxelems 4294967295, elemsize 8200
[1650465630.196444] [ndv4:55432:0] self.c:220 UCX DEBUG created self iface id 0xe074c57977df98cc send_size 8192
[1650465630.196450] [ndv4:55432:0] ucp_worker.c:1159 UCX DEBUG created interface[2]=0x28b2f90 using self/memory0 on worker 0x3178950
[1650465630.196476] [ndv4:55432:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465630.196482] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465630.196486] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465630.197202] [ndv4:55324:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.197219] [ndv4:55324:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.197447] [ndv4:55324:0] ib_md.c:1604 UCX DEBUG mlx5_ib5: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.195676] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x503f220: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 6)
[1650465630.195989] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x503f220: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 7)
[1650465630.196161] [ndv4:54756:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.196174] [ndv4:54756:0] async.c:228 UCX DEBUG added async handler 0x43503a0 [id=116 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.196200] [ndv4:54756:0] async.c:506 UCX DEBUG listening to async event fd 116 events 0x5 mode thread_spinlock
[1650465630.196220] [ndv4:54756:0] ucp_worker.c:1159 UCX DEBUG created interface[29]=0x503f220 using ud_verbs/mlx5_ib4:1 on worker 0x2e9a8d0
[1650465630.196244] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.196250] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.196303] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.196308] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.196809] [ndv4:54756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465630.195843] [ndv4:55512:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.195859] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.195863] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.195923] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.196291] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465630.196297] [ndv4:55512:0] 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> }
[1650465630.196363] [ndv4:55512:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib3 length=2048) failed: Invalid argument
[1650465630.196367] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.196375] [ndv4:55512:0] async.c:228 UCX DEBUG added async handler 0x1970b00 [id=108 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.196394] [ndv4:55512:0] async.c:506 UCX DEBUG listening to async event fd 108 events 0x1 mode thread_spinlock
[1650465630.196694] [ndv4:55512:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465630.199072] [ndv4:55432:0] async.c:228 UCX DEBUG added async handler 0x28c0fb0 [id=78 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465630.199109] [ndv4:55432:0] async.c:506 UCX DEBUG listening to async event fd 78 events 0x5 mode thread_spinlock
[1650465630.199127] [ndv4:55432:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x28b38f0: listening for connections (fd=78) on 10.5.0.5:49327
[1650465630.199266] [ndv4:55432:0] ucp_worker.c:1159 UCX DEBUG created interface[3]=0x28b38f0 using tcp/eth0 on worker 0x3178950
[1650465630.199288] [ndv4:55432:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465630.199292] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465630.199295] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465630.199335] [ndv4:55432:0] async.c:228 UCX DEBUG added async handler 0x280f370 [id=80 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465630.199359] [ndv4:55432:0] async.c:506 UCX DEBUG listening to async event fd 80 events 0x5 mode thread_spinlock
[1650465630.199363] [ndv4:55432:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x28b3f70: listening for connections (fd=80) on 127.0.0.1:43614
[1650465630.199379] [ndv4:55432:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465630.199384] [ndv4:55432:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465630.199444] [ndv4:55432:0] ucp_worker.c:1159 UCX DEBUG created interface[4]=0x28b3f70 using tcp/lo on worker 0x3178950
[1650465630.199461] [ndv4:55432:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465630.199464] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465630.199467] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465630.199508] [ndv4:55432:0] async.c:228 UCX DEBUG added async handler 0x280f0a0 [id=82 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465630.199531] [ndv4:55432:0] async.c:506 UCX DEBUG listening to async event fd 82 events 0x5 mode thread_spinlock
[1650465630.199535] [ndv4:55432:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x289b600: listening for connections (fd=82) on 172.16.1.242:42240
[1650465630.199805] [ndv4:55432:0] ucp_worker.c:1159 UCX DEBUG created interface[5]=0x289b600 using tcp/ib0 on worker 0x3178950
[1650465630.200198] [ndv4:54853:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465630.200211] [ndv4:54853:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465630.201278] [ndv4:54853:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465630.201283] [ndv4:54853:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465630.201008] [ndv4:54715:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465630.201017] [ndv4:54715:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465630.202363] [ndv4:54861:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib5 vendor_id: 0x15b3 device_id: 4124
[1650465630.202759] [ndv4:54861:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib5: disable ODP because it's not supported for DevX QP
[1650465630.201940] [ndv4:55516:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x3b5e050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xda92
[1650465630.202175] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.202182] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.202191] [ndv4:55516:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2ae24bfc8008 of 151544 bytes with 1052 elements
[1650465630.203322] [ndv4:55512:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x3b74050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xda94
[1650465630.203448] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.203455] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.203466] [ndv4:55512:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2aea7eb19008 of 151544 bytes with 1052 elements
[1650465630.205218] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x5ad7060: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 1)
[1650465630.205297] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x5ad7060: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 2)
[1650465630.205398] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x5ad7060: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 3)
[1650465630.205502] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x5ad7060: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 4)
[1650465630.205780] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x5ad7060: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 5)
[1650465630.206152] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x5ad7060: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 6)
[1650465630.206419] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x5ad7060: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 7)
[1650465630.206699] [ndv4:54965:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.206712] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x2cc3510 [id=137 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.206737] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 137 events 0x5 mode thread_spinlock
[1650465630.206750] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[44]=0x5ad7060 using ud_verbs/mlx5_ib7:1 on worker 0x23dc840
[1650465630.206765] [ndv4:54717:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.206787] [ndv4:54717:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.206997] [ndv4:54717:0] ib_md.c:1604 UCX DEBUG mlx5_ib5: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.207094] [ndv4:54717:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.207106] [ndv4:54717:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.208394] [ndv4:54717:0] topo.c:99 UCX DEBUG bus id 0x106000000 doesn't exist. sys_dev = 5
[1650465630.208406] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465630.206690] [ndv4:55516:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae258600000..0x2ae25ac00000 on mlx5_ib3 lkey 0x81100 rkey 0x81100 access 0xf flags 0x3e4
[1650465630.206709] [ndv4:55516:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2ae258600018 of 39845864 bytes with 4752 elements
[1650465630.206869] [ndv4:55516:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x3b5e050
[1650465630.206905] [ndv4:55516:0] ucp_worker.c:1159 UCX DEBUG created interface[23]=0x3b5e050 using dc_mlx5/mlx5_ib3:1 on worker 0x22b38d0
[1650465630.207625] [ndv4:55512:0] ib_md.c:812 UCX DEBUG registered memory 0x2aea7c400000..0x2aea7ea00000 on mlx5_ib3 lkey 0x81200 rkey 0x81200 access 0xf flags 0x3e4
[1650465630.207651] [ndv4:55512:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2aea7c400018 of 39845864 bytes with 4752 elements
[1650465630.207790] [ndv4:55512:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x3b74050
[1650465630.207824] [ndv4:55512:0] ucp_worker.c:1159 UCX DEBUG created interface[23]=0x3b74050 using dc_mlx5/mlx5_ib3:1 on worker 0x22c98d0
[1650465630.207935] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.207947] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.208059] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.208063] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.208391] [ndv4:55512:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465630.209255] [ndv4:55512:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.209615] [ndv4:55512:0] ib_iface.c:994 UCX DEBUG iface=0x3d50060: created UD QP 0xda95 on mlx5_ib3:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465630.209714] [ndv4:55324:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.209726] [ndv4:55324:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.210440] [ndv4:55324:0] topo.c:99 UCX DEBUG bus id 0x106000000 doesn't exist. sys_dev = 5
[1650465630.210445] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465630.210108] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465630.210186] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.210193] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.210226] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.210230] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.210596] [ndv4:55512:0] ib_md.c:812 UCX DEBUG registered memory 0x2aea7eb3e000..0x2aea7ebc3000 on mlx5_ib3 lkey 0x81300 rkey 0x81300 access 0xf flags 0x3e4
[1650465630.210603] [ndv4:55512:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2aea7eb3e018 of 544744 bytes with 128 elements
[1650465630.210609] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465630.210925] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x3d50060: adding gid fe80::15:5dff:fd33:fffe to hash on device mlx5_ib3 port 1 index 0)
[1650465630.211142] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x3d50060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 1)
[1650465630.211308] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x3d50060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 2)
[1650465630.211494] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x3d50060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 3)
[1650465630.211832] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x3d50060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 4)
[1650465630.212147] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x3d50060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 5)
[1650465630.212018] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465630.212031] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465630.212174] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465630.212180] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465630.210416] [ndv4:54756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.210826] [ndv4:54756:0] ib_iface.c:994 UCX DEBUG iface=0x515d050: created UD QP 0xdaed on mlx5_ib4:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465630.210834] [ndv4:54756:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465630.211345] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465630.211635] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.211646] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.211687] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.211692] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.212056] [ndv4:54756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b3b652d0000..0x2b3b65355000 on mlx5_ib4 lkey 0x81000 rkey 0x81000 access 0xf flags 0x3e4
[1650465630.212064] [ndv4:54756:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b3b652d0018 of 544744 bytes with 128 elements
[1650465630.212072] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465630.212114] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x515d050: adding gid fe80::15:5dff:fd33:ffff to hash on device mlx5_ib4 port 1 index 0)
[1650465630.212815] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x515d050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 1)
[1650465630.213069] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x515d050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 2)
[1650465630.213317] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x515d050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 3)
[1650465630.213472] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x515d050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 4)
[1650465630.213903] [ndv4:54715:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465630.213912] [ndv4:54715:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465630.214129] [ndv4:54715:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.214939] [ndv4:55039:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465630.214954] [ndv4:55039:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465630.215556] [ndv4:55039:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465630.215560] [ndv4:55039:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465630.216668] [ndv4:55039:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650465630.216674] [ndv4:55039:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650465630.217029] [ndv4:55039:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.215738] [ndv4:55691:0] mm_posix.c:518 UCX DEBUG allocated posix shared memory at 0x2b871f8df000 length 12288
[1650465630.215798] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool mm_recv_desc: align 64, maxelems 4294967295, elemsize 8368
[1650465630.216852] [ndv4:55691:0] mm_posix.c:305 UCX DEBUG shared memory mmap(addr=(nil), length=6291456, flags= HUGETLB, fd=76) failed: Invalid argument
[1650465630.216860] [ndv4:55691:0] mm_posix.c:518 UCX DEBUG allocated posix shared memory at 0x2b871f8e2000 length 4296704
[1650465630.216865] [ndv4:55691:0] mpool.c:205 UCX DEBUG mpool mm_recv_desc: allocated chunk 0x2b871f8e2018 of 4296680 bytes with 512 elements
[1650465630.217126] [ndv4:55691:0] mm_iface.c:600 UCX DEBUG created mm iface 0x2b28a90 FIFO id 0x400000005555a826 va 0x2b871f8df000 size 12288 (128 x 64 elems)
[1650465630.217171] [ndv4:55691:0] ucp_worker.c:1159 UCX DEBUG created interface[0]=0x2b28a90 using posix/memory on worker 0x33ff770
[1650465630.217197] [ndv4:55691:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 8447 bytes with hugetlb
[1650465630.217231] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool mm_recv_desc: align 64, maxelems 4294967295, elemsize 8368
[1650465630.217246] [ndv4:55691:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 4292720 bytes with hugetlb
[1650465630.217255] [ndv4:55691:0] mpool.c:205 UCX DEBUG mpool mm_recv_desc: allocated chunk 0x2b8724000018 of 4296680 bytes with 512 elements
[1650465630.217872] [ndv4:55691:0] mm_iface.c:600 UCX DEBUG created mm iface 0x2b3a7e0 FIFO id 0x720021 va 0x2b871fcfb000 size 12288 (128 x 64 elems)
[1650465630.217880] [ndv4:55691:0] ucp_worker.c:1159 UCX DEBUG created interface[1]=0x2b3a7e0 using sysv/memory on worker 0x33ff770
[1650465630.217892] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool self_msg_desc: align 64, maxelems 4294967295, elemsize 8200
[1650465630.217895] [ndv4:55691:0] self.c:220 UCX DEBUG created self iface id 0x442a197005aa060c send_size 8192
[1650465630.217901] [ndv4:55691:0] ucp_worker.c:1159 UCX DEBUG created interface[2]=0x2b39e80 using self/memory0 on worker 0x33ff770
[1650465630.217924] [ndv4:55691:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465630.217929] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465630.217932] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465630.218533] [ndv4:55152:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib5:1
[1650465630.217065] [ndv4:54861:0] async.c:228 UCX DEBUG added async handler 0x18bff90 [id=67 ref 1] uct_ib_async_event_handler() to hash
[1650465630.217090] [ndv4:54861:0] async.c:506 UCX DEBUG listening to async event fd 67 events 0x1 mode thread_spinlock
[1650465630.217093] [ndv4:54861:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib5' (InfiniBand channel adapter) with 1 ports
[1650465630.217305] [ndv4:54861:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.217314] [ndv4:54861:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.217320] [ndv4:54861:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.217368] [ndv4:54861:0] ib_md.c:1319 UCX DEBUG mlx5_ib5: using registration cache
[1650465630.219019] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465630.219031] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465630.219323] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465630.219328] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465630.219440] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.219462] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.219699] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.219705] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.220137] [ndv4:55152:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.220159] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.220163] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.220229] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.220105] [ndv4:55516:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465630.220502] [ndv4:55691:0] async.c:228 UCX DEBUG added async handler 0x2b481a0 [id=78 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465630.220536] [ndv4:55691:0] async.c:506 UCX DEBUG listening to async event fd 78 events 0x5 mode thread_spinlock
[1650465630.220553] [ndv4:55691:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x2b3adb0: listening for connections (fd=78) on 10.5.0.5:54662
[1650465630.220740] [ndv4:55691:0] ucp_worker.c:1159 UCX DEBUG created interface[3]=0x2b3adb0 using tcp/eth0 on worker 0x33ff770
[1650465630.220761] [ndv4:55691:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465630.220765] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465630.220768] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465630.220805] [ndv4:55691:0] async.c:228 UCX DEBUG added async handler 0x2b47b60 [id=80 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465630.220835] [ndv4:55691:0] async.c:506 UCX DEBUG listening to async event fd 80 events 0x5 mode thread_spinlock
[1650465630.220838] [ndv4:55691:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x2b22530: listening for connections (fd=80) on 127.0.0.1:45852
[1650465630.220858] [ndv4:55691:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465630.220863] [ndv4:55691:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465630.220922] [ndv4:55691:0] ucp_worker.c:1159 UCX DEBUG created interface[4]=0x2b22530 using tcp/lo on worker 0x33ff770
[1650465630.220938] [ndv4:55691:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465630.220941] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465630.220943] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465630.220979] [ndv4:55691:0] async.c:228 UCX DEBUG added async handler 0x2a94a20 [id=82 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465630.220999] [ndv4:55691:0] async.c:506 UCX DEBUG listening to async event fd 82 events 0x5 mode thread_spinlock
[1650465630.221003] [ndv4:55691:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x2b22bb0: listening for connections (fd=82) on 172.16.1.242:52533
[1650465630.220684] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465630.220695] [ndv4:55152:0] 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> }
[1650465630.220734] [ndv4:55152:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib5 length=2048) failed: Invalid argument
[1650465630.220738] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.220752] [ndv4:55152:0] async.c:228 UCX DEBUG added async handler 0x5a91f80 [id=122 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.220781] [ndv4:55152:0] async.c:506 UCX DEBUG listening to async event fd 122 events 0x1 mode thread_spinlock
[1650465630.221256] [ndv4:55691:0] ucp_worker.c:1159 UCX DEBUG created interface[5]=0x2b22bb0 using tcp/ib0 on worker 0x33ff770
[1650465630.221324] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465630.221332] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465630.221452] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465630.221458] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465630.221427] [ndv4:55152:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465630.221145] [ndv4:55516:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.221678] [ndv4:55516:0] ib_iface.c:994 UCX DEBUG iface=0x3d3a060: created UD QP 0xda96 on mlx5_ib3:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465630.221946] [ndv4:54715:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib4 vendor_id: 0x15b3 device_id: 4124
[1650465630.222225] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465630.222405] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.222412] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.222474] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.222479] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.222280] [ndv4:54715:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib4: disable ODP because it's not supported for DevX QP
[1650465630.222900] [ndv4:55516:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae25ad63000..0x2ae25ade8000 on mlx5_ib3 lkey 0x81400 rkey 0x81400 access 0xf flags 0x3e4
[1650465630.222908] [ndv4:55516:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae25ad63018 of 544744 bytes with 128 elements
[1650465630.222913] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465630.223162] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x3d3a060: adding gid fe80::15:5dff:fd33:fffe to hash on device mlx5_ib3 port 1 index 0)
[1650465630.223309] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x3d3a060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 1)
[1650465630.223502] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x3d3a060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 2)
[1650465630.223915] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x3d3a060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 3)
[1650465630.224133] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x3d3a060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 4)
[1650465630.224522] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x3d3a060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 5)
[1650465630.222842] [ndv4:54717:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465630.222856] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465630.224995] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x3d3a060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 6)
[1650465630.225527] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x3d3a060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 7)
[1650465630.225789] [ndv4:55516:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.225800] [ndv4:55516:0] async.c:228 UCX DEBUG added async handler 0x3d3ae60 [id=109 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.225826] [ndv4:55516:0] async.c:506 UCX DEBUG listening to async event fd 109 events 0x5 mode thread_spinlock
[1650465630.225843] [ndv4:55516:0] ucp_worker.c:1159 UCX DEBUG created interface[24]=0x3d3a060 using ud_verbs/mlx5_ib3:1 on worker 0x22b38d0
[1650465630.226084] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.226091] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.226234] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.226240] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.225191] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x3d50060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 6)
[1650465630.225751] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x3d50060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 7)
[1650465630.225894] [ndv4:55512:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.225904] [ndv4:55512:0] async.c:228 UCX DEBUG added async handler 0x3d50e60 [id=109 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.225930] [ndv4:55512:0] async.c:506 UCX DEBUG listening to async event fd 109 events 0x5 mode thread_spinlock
[1650465630.225943] [ndv4:55512:0] ucp_worker.c:1159 UCX DEBUG created interface[24]=0x3d50060 using ud_verbs/mlx5_ib3:1 on worker 0x22c98d0
[1650465630.225960] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.225966] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.226008] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.226012] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.226366] [ndv4:55512:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465630.225696] [ndv4:55432:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465630.227381] [ndv4:55432:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465630.227419] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.227422] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.227441] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.227351] [ndv4:55512:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.227682] [ndv4:55512:0] ib_iface.c:994 UCX DEBUG iface=0x3e6e460: created UD QP 0xda97 on mlx5_ib3:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465630.227690] [ndv4:55512:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465630.227009] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x515d050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 5)
[1650465630.227109] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x515d050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 6)
[1650465630.227124] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x515d050: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 7)
[1650465630.227130] [ndv4:54756:0] 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> }
[1650465630.227210] [ndv4:54756:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.227216] [ndv4:54756:0] async.c:228 UCX DEBUG added async handler 0x4b1b340 [id=117 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.227243] [ndv4:54756:0] async.c:506 UCX DEBUG listening to async event fd 117 events 0x5 mode thread_spinlock
[1650465630.227260] [ndv4:54756:0] ucp_worker.c:1159 UCX DEBUG created interface[30]=0x515d050 using ud_mlx5/mlx5_ib4:1 on worker 0x2e9a8d0
[1650465630.227284] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.227290] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.227463] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.227469] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.227755] [ndv4:54756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib5:1
[1650465630.227931] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.227940] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465630.227974] [ndv4:54853:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465630.227987] [ndv4:54853:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465630.228171] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465630.228254] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.228261] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.228292] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.228297] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.228685] [ndv4:55512:0] ib_md.c:812 UCX DEBUG registered memory 0x2aea7ebc3000..0x2aea7ec48000 on mlx5_ib3 lkey 0x81500 rkey 0x81500 access 0xf flags 0x3e4
[1650465630.228691] [ndv4:55512:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2aea7ebc3018 of 544744 bytes with 128 elements
[1650465630.228695] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465630.228734] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x3e6e460: adding gid fe80::15:5dff:fd33:fffe to hash on device mlx5_ib3 port 1 index 0)
[1650465630.228967] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x3e6e460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 1)
[1650465630.229036] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x3e6e460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 2)
[1650465630.229162] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x3e6e460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 3)
[1650465630.229329] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x3e6e460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 4)
[1650465630.229477] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x3e6e460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 5)
[1650465630.229677] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x3e6e460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 6)
[1650465630.229909] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x3e6e460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 7)
[1650465630.229914] [ndv4:55512:0] 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> }
[1650465630.229970] [ndv4:55512:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.229976] [ndv4:55512:0] async.c:228 UCX DEBUG added async handler 0x382ca10 [id=110 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.229998] [ndv4:55512:0] async.c:506 UCX DEBUG listening to async event fd 110 events 0x5 mode thread_spinlock
[1650465630.230011] [ndv4:55512:0] ucp_worker.c:1159 UCX DEBUG created interface[25]=0x3e6e460 using ud_mlx5/mlx5_ib3:1 on worker 0x22c98d0
[1650465630.230061] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.230067] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.230187] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.230193] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.228815] [ndv4:54756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465630.228835] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.228839] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.228890] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.229205] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.229212] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465630.229655] [ndv4:54756:0] ib_iface.c:994 UCX DEBUG iface=0x525d0a0: created RC QP 0xda6c on mlx5_ib5:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465630.230189] [ndv4:54756:0] ucp_worker.c:1159 UCX DEBUG created interface[31]=0x525d0a0 using rc_verbs/mlx5_ib5:1 on worker 0x2e9a8d0
[1650465630.230346] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.230352] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.230495] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.230500] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.229241] [ndv4:55152:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x5c5f050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xda76
[1650465630.229531] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.229539] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.229548] [ndv4:55152:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b3b97e9b008 of 151544 bytes with 1052 elements
[1650465630.229153] [ndv4:55432:0] ib_iface.c:994 UCX DEBUG iface=0x28a9740: created RC QP 0xf4b8 on mlx5_ib0:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465630.231397] [ndv4:55432:0] ucp_worker.c:1159 UCX DEBUG created interface[6]=0x28a9740 using rc_verbs/mlx5_ib0:1 on worker 0x3178950
[1650465630.231481] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465630.231487] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465630.230989] [ndv4:54756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib5:1
[1650465630.230782] [ndv4:55512:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465630.231984] [ndv4:54965:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib7:1
[1650465630.233530] [ndv4:55152:0] ib_md.c:812 UCX DEBUG registered memory 0x2b3ba8a00000..0x2b3bab000000 on mlx5_ib5 lkey 0x80700 rkey 0x80700 access 0xf flags 0x3e4
[1650465630.233544] [ndv4:55152:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b3ba8a00018 of 39845864 bytes with 4752 elements
[1650465630.233750] [ndv4:55152:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x5c5f050
[1650465630.233785] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[33]=0x5c5f050 using dc_mlx5/mlx5_ib5:1 on worker 0x357a8c0
[1650465630.233845] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.233854] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.233916] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.233921] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.233826] [ndv4:55691:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465630.232258] [ndv4:55512:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465630.232281] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.232285] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.232341] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.233904] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.233913] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465630.231812] [ndv4:54861:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.231820] [ndv4:54861:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.232070] [ndv4:54861:0] ib_md.c:1604 UCX DEBUG mlx5_ib5: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.232166] [ndv4:54861:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.232173] [ndv4:54861:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.232426] [ndv4:54756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.232439] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.232442] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.232493] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.232835] [ndv4:54756:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x557e010 of 8176 bytes with 127 elements
[1650465630.233018] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.233026] [ndv4:54756:0] 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> }
[1650465630.233062] [ndv4:54756:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib5 length=2048) failed: Invalid argument
[1650465630.233067] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.233079] [ndv4:54756:0] async.c:228 UCX DEBUG added async handler 0x4350e50 [id=120 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.233105] [ndv4:54756:0] async.c:506 UCX DEBUG listening to async event fd 120 events 0x1 mode thread_spinlock
[1650465630.233116] [ndv4:54756:0] ucp_worker.c:1159 UCX DEBUG created interface[32]=0x53aa030 using rc_mlx5/mlx5_ib5:1 on worker 0x2e9a8d0
[1650465630.233162] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.233168] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.233249] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.233254] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.233588] [ndv4:54756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib5:1
[1650465630.234201] [ndv4:55152:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib5:1
[1650465630.234263] [ndv4:54715:0] async.c:228 UCX DEBUG added async handler 0x200e280 [id=65 ref 1] uct_ib_async_event_handler() to hash
[1650465630.234289] [ndv4:54715:0] async.c:506 UCX DEBUG listening to async event fd 65 events 0x1 mode thread_spinlock
[1650465630.234292] [ndv4:54715:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib4' (InfiniBand channel adapter) with 1 ports
[1650465630.234542] [ndv4:54756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.234559] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.234563] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.234661] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.234991] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465630.234997] [ndv4:54756:0] 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> }
[1650465630.235029] [ndv4:54756:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib5 length=2048) failed: Invalid argument
[1650465630.235032] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.235039] [ndv4:54756:0] async.c:228 UCX DEBUG added async handler 0x53b2f80 [id=122 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.235079] [ndv4:54756:0] async.c:506 UCX DEBUG listening to async event fd 122 events 0x1 mode thread_spinlock
[1650465630.234506] [ndv4:54715:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.234514] [ndv4:54715:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.234519] [ndv4:54715:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.234559] [ndv4:54715:0] ib_md.c:1319 UCX DEBUG mlx5_ib4: using registration cache
[1650465630.235236] [ndv4:54715:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.235241] [ndv4:54715:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.234554] [ndv4:55512:0] ib_iface.c:994 UCX DEBUG iface=0x3f6e0a0: created RC QP 0xdaee on mlx5_ib4:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465630.234675] [ndv4:55691:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465630.234718] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.234723] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.234755] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.235073] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.235084] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465630.235079] [ndv4:55152:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.235382] [ndv4:55324:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465630.235390] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465630.235856] [ndv4:55324:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465630.235861] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465630.235407] [ndv4:54756:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465630.235557] [ndv4:55691:0] ib_iface.c:994 UCX DEBUG iface=0x2b30630: created RC QP 0xf4b9 on mlx5_ib0:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465630.235563] [ndv4:55039:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib3 vendor_id: 0x15b3 device_id: 4124
[1650465630.236129] [ndv4:55039:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib3: disable ODP because it's not supported for DevX QP
[1650465630.235375] [ndv4:55152:0] ib_iface.c:994 UCX DEBUG iface=0x5e3b060: created UD QP 0xda6d on mlx5_ib5:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465630.235968] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465630.236014] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.236020] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.236055] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.236059] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.236365] [ndv4:55152:0] ib_md.c:812 UCX DEBUG registered memory 0x2b3b97ec0000..0x2b3b97f45000 on mlx5_ib5 lkey 0x80800 rkey 0x80800 access 0xf flags 0x3e4
[1650465630.236372] [ndv4:55152:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b3b97ec0018 of 544744 bytes with 128 elements
[1650465630.236376] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465630.236409] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x5e3b060: adding gid fe80::15:5dff:fd34:0 to hash on device mlx5_ib5 port 1 index 0)
[1650465630.236424] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x5e3b060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 1)
[1650465630.236439] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x5e3b060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 2)
[1650465630.236454] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x5e3b060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 3)
[1650465630.235612] [ndv4:54715:0] ib_md.c:1604 UCX DEBUG mlx5_ib4: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.235647] [ndv4:54715:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.235653] [ndv4:54715:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.235781] [ndv4:54715:0] topo.c:99 UCX DEBUG bus id 0x105000000 doesn't exist. sys_dev = 4
[1650465630.235785] [ndv4:54715:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465630.236116] [ndv4:54715:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465630.236120] [ndv4:54715:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465630.235648] [ndv4:55512:0] ucp_worker.c:1159 UCX DEBUG created interface[26]=0x3f6e0a0 using rc_verbs/mlx5_ib4:1 on worker 0x22c98d0
[1650465630.235804] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.235811] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.235877] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.235882] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.236084] [ndv4:55512:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465630.236812] [ndv4:55039:0] async.c:228 UCX DEBUG added async handler 0x19d52b0 [id=63 ref 1] uct_ib_async_event_handler() to hash
[1650465630.236837] [ndv4:55039:0] async.c:506 UCX DEBUG listening to async event fd 63 events 0x1 mode thread_spinlock
[1650465630.236841] [ndv4:55039:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib3' (InfiniBand channel adapter) with 1 ports
[1650465630.236961] [ndv4:55039:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.236970] [ndv4:55039:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.236975] [ndv4:55039:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.237031] [ndv4:55039:0] ib_md.c:1319 UCX DEBUG mlx5_ib3: using registration cache
[1650465630.237655] [ndv4:55039:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.237661] [ndv4:55039:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.236370] [ndv4:55691:0] ucp_worker.c:1159 UCX DEBUG created interface[6]=0x2b30630 using rc_verbs/mlx5_ib0:1 on worker 0x33ff770
[1650465630.236480] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465630.236487] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465630.236624] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465630.236630] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465630.236917] [ndv4:55691:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465630.236952] [ndv4:55691:0] ib_device.c:1394 UCX DEBUG max IB CQE size is 128
[1650465630.237541] [ndv4:55512:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.237556] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.237559] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.237623] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.238022] [ndv4:55512:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x428f010 of 8176 bytes with 127 elements
[1650465630.238277] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.238284] [ndv4:55512:0] 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> }
[1650465630.238324] [ndv4:55512:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib4 length=2048) failed: Invalid argument
[1650465630.238327] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.238341] [ndv4:55512:0] async.c:228 UCX DEBUG added async handler 0x3061690 [id=113 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.238370] [ndv4:55512:0] async.c:506 UCX DEBUG listening to async event fd 113 events 0x1 mode thread_spinlock
[1650465630.238381] [ndv4:55512:0] ucp_worker.c:1159 UCX DEBUG created interface[27]=0x40bb030 using rc_mlx5/mlx5_ib4:1 on worker 0x22c98d0
[1650465630.238400] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.238406] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.238479] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.238484] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.238882] [ndv4:55512:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465630.237923] [ndv4:55691:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.237936] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.237939] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.237990] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.238434] [ndv4:55691:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x34c9010 of 8176 bytes with 127 elements
[1650465630.238677] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.238695] [ndv4:55691:0] 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> }
[1650465630.238742] [ndv4:55691:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650465630.238747] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.237898] [ndv4:55039:0] ib_md.c:1604 UCX DEBUG mlx5_ib3: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.237984] [ndv4:55039:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.237992] [ndv4:55039:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.238468] [ndv4:55039:0] topo.c:99 UCX DEBUG bus id 0x104000000 doesn't exist. sys_dev = 3
[1650465630.238473] [ndv4:55039:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465630.239125] [ndv4:55039:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465630.239130] [ndv4:55039:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465630.239315] [ndv4:55039:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465630.239319] [ndv4:55039:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465630.238875] [ndv4:55516:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib3:1
[1650465630.240058] [ndv4:55516:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.240472] [ndv4:55516:0] ib_iface.c:994 UCX DEBUG iface=0x3e58460: created UD QP 0xda98 on mlx5_ib3:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465630.240479] [ndv4:55516:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465630.240048] [ndv4:55512:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.240065] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.240068] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.240121] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.240478] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465630.240483] [ndv4:55512:0] 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> }
[1650465630.240516] [ndv4:55512:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib4 length=2048) failed: Invalid argument
[1650465630.240519] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.240527] [ndv4:55512:0] async.c:228 UCX DEBUG added async handler 0x40c3f10 [id=115 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.240546] [ndv4:55512:0] async.c:506 UCX DEBUG listening to async event fd 115 events 0x1 mode thread_spinlock
[1650465630.240974] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465630.241129] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.241136] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.241215] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650465630.241220] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650465630.241596] [ndv4:55516:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae25ade8000..0x2ae25ae6d000 on mlx5_ib3 lkey 0x81600 rkey 0x81600 access 0xf flags 0x3e4
[1650465630.241611] [ndv4:55516:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae25ade8018 of 544744 bytes with 128 elements
[1650465630.241621] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465630.241661] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x3e58460: adding gid fe80::15:5dff:fd33:fffe to hash on device mlx5_ib3 port 1 index 0)
[1650465630.242026] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x3e58460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 1)
[1650465630.242195] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x3e58460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 2)
[1650465630.241165] [ndv4:54756:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x5580050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xda82
[1650465630.241326] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.241332] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.241342] [ndv4:54756:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b3b67b56008 of 151544 bytes with 1052 elements
[1650465630.241099] [ndv4:55512:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465630.240898] [ndv4:54853:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465630.240913] [ndv4:54853:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465630.241185] [ndv4:54853:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.240780] [ndv4:55039:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465630.240788] [ndv4:55039:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465630.241366] [ndv4:55039:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650465630.241371] [ndv4:55039:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650465630.241746] [ndv4:55039:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.243504] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465630.243516] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465630.244933] [ndv4:54756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b3b65400000..0x2b3b67a00000 on mlx5_ib5 lkey 0x80900 rkey 0x80900 access 0xf flags 0x3e4
[1650465630.244951] [ndv4:54756:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b3b65400018 of 39845864 bytes with 4752 elements
[1650465630.245116] [ndv4:54756:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x5580050
[1650465630.245143] [ndv4:54756:0] ucp_worker.c:1159 UCX DEBUG created interface[33]=0x5580050 using dc_mlx5/mlx5_ib5:1 on worker 0x2e9a8d0
[1650465630.245287] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.245295] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.245460] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.245466] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.246392] [ndv4:54756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib5:1
[1650465630.248209] [ndv4:54756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.248494] [ndv4:54756:0] ib_iface.c:994 UCX DEBUG iface=0x575c060: created UD QP 0xda79 on mlx5_ib5:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465630.245760] [ndv4:55691:0] async.c:228 UCX DEBUG added async handler 0x2a95fc0 [id=85 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.245787] [ndv4:55691:0] async.c:506 UCX DEBUG listening to async event fd 85 events 0x1 mode thread_spinlock
[1650465630.245803] [ndv4:55691:0] ucp_worker.c:1159 UCX DEBUG created interface[7]=0x2b3e080 using rc_mlx5/mlx5_ib0:1 on worker 0x33ff770
[1650465630.245951] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465630.245958] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465630.246109] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465630.246115] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465630.246944] [ndv4:55691:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465630.247652] [ndv4:55691:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.247661] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.247664] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.247678] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.247970] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465630.247976] [ndv4:55691:0] 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> }
[1650465630.248008] [ndv4:55691:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650465630.248012] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.248021] [ndv4:55691:0] async.c:228 UCX DEBUG added async handler 0x2a8aef0 [id=87 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.248040] [ndv4:55691:0] async.c:506 UCX DEBUG listening to async event fd 87 events 0x1 mode thread_spinlock
[1650465630.248639] [ndv4:55691:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465630.244194] [ndv4:55432:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465630.244846] [ndv4:55432:0] ib_device.c:1394 UCX DEBUG max IB CQE size is 128
[1650465630.245851] [ndv4:55432:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.245859] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.245862] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.245917] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.246281] [ndv4:55432:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x3242010 of 8176 bytes with 127 elements
[1650465630.246487] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.246514] [ndv4:55432:0] 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> }
[1650465630.246564] [ndv4:55432:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650465630.246569] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.245438] [ndv4:54965:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.245917] [ndv4:54965:0] ib_iface.c:994 UCX DEBUG iface=0x31477b0: created UD QP 0xda56 on mlx5_ib7:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465630.245924] [ndv4:54965:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465630.247343] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465630.247772] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465630.247780] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465630.247841] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465630.247846] [ndv4:54965:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465630.248235] [ndv4:54965:0] ib_md.c:812 UCX DEBUG registered memory 0x2b4b39783000..0x2b4b39808000 on mlx5_ib7 lkey 0x80a00 rkey 0x80a00 access 0xf flags 0x3e4
[1650465630.248240] [ndv4:54965:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b4b39783018 of 544744 bytes with 128 elements
[1650465630.248245] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465630.248491] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x31477b0: adding gid fe80::15:5dff:fd34:2 to hash on device mlx5_ib7 port 1 index 0)
[1650465630.248772] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x31477b0: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 1)
[1650465630.248893] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x31477b0: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 2)
[1650465630.248901] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465630.249027] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.249037] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.249081] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.249085] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.248866] [ndv4:55039:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib4 vendor_id: 0x15b3 device_id: 4124
[1650465630.249290] [ndv4:55324:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465630.249297] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465630.249411] [ndv4:54756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b3b67b7b000..0x2b3b67c00000 on mlx5_ib5 lkey 0x80a00 rkey 0x80a00 access 0xf flags 0x3e4
[1650465630.249420] [ndv4:54756:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b3b67b7b018 of 544744 bytes with 128 elements
[1650465630.249428] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465630.249561] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x575c060: adding gid fe80::15:5dff:fd34:0 to hash on device mlx5_ib5 port 1 index 0)
[1650465630.250065] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x575c060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 1)
[1650465630.250445] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x575c060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 2)
[1650465630.250869] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x575c060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 3)
[1650465630.251101] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x575c060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 4)
[1650465630.251274] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x575c060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 5)
[1650465630.251476] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x575c060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 6)
[1650465630.251686] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x575c060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 7)
[1650465630.249325] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x5e3b060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 4)
[1650465630.249779] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x5e3b060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 5)
[1650465630.250229] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x5e3b060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 6)
[1650465630.249850] [ndv4:55512:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x4291050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xdb03
[1650465630.250142] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.250150] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.250159] [ndv4:55512:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2aea81449008 of 151544 bytes with 1052 elements
[1650465630.249461] [ndv4:54715:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465630.249469] [ndv4:54715:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465630.249636] [ndv4:54715:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465630.249640] [ndv4:54715:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465630.251757] [ndv4:55039:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib4: disable ODP because it's not supported for DevX QP
[1650465630.251950] [ndv4:54756:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.251961] [ndv4:54756:0] async.c:228 UCX DEBUG added async handler 0x52389b0 [id=123 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.251989] [ndv4:54756:0] async.c:506 UCX DEBUG listening to async event fd 123 events 0x5 mode thread_spinlock
[1650465630.252013] [ndv4:54756:0] ucp_worker.c:1159 UCX DEBUG created interface[34]=0x575c060 using ud_verbs/mlx5_ib5:1 on worker 0x2e9a8d0
[1650465630.252105] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.252112] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.252206] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.252210] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.253468] [ndv4:55432:0] async.c:228 UCX DEBUG added async handler 0x27fedf0 [id=85 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.253497] [ndv4:55432:0] async.c:506 UCX DEBUG listening to async event fd 85 events 0x1 mode thread_spinlock
[1650465630.253516] [ndv4:55432:0] ucp_worker.c:1159 UCX DEBUG created interface[7]=0x28b7190 using rc_mlx5/mlx5_ib0:1 on worker 0x3178950
[1650465630.253641] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465630.253650] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465630.253728] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465630.253734] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465630.253978] [ndv4:55432:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465630.254922] [ndv4:55432:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.254931] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.254934] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.254948] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.253789] [ndv4:55512:0] ib_md.c:812 UCX DEBUG registered memory 0x2aea7ee00000..0x2aea81400000 on mlx5_ib4 lkey 0x81100 rkey 0x81100 access 0xf flags 0x3e4
[1650465630.253806] [ndv4:55512:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2aea7ee00018 of 39845864 bytes with 4752 elements
[1650465630.253945] [ndv4:55512:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x4291050
[1650465630.253984] [ndv4:55512:0] ucp_worker.c:1159 UCX DEBUG created interface[28]=0x4291050 using dc_mlx5/mlx5_ib4:1 on worker 0x22c98d0
[1650465630.254123] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.254135] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.253985] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x3e58460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 3)
[1650465630.254137] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x3e58460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 4)
[1650465630.254306] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x3e58460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 5)
[1650465630.254419] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x3e58460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 6)
[1650465630.254608] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x3e58460: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 7)
[1650465630.254614] [ndv4:55516:0] 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> }
[1650465630.254708] [ndv4:55516:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.254715] [ndv4:55516:0] async.c:228 UCX DEBUG added async handler 0x3816a10 [id=110 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.254740] [ndv4:55516:0] async.c:506 UCX DEBUG listening to async event fd 110 events 0x5 mode thread_spinlock
[1650465630.254759] [ndv4:55516:0] ucp_worker.c:1159 UCX DEBUG created interface[25]=0x3e58460 using ud_mlx5/mlx5_ib3:1 on worker 0x22b38d0
[1650465630.254793] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.254799] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.254918] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.254923] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.255146] [ndv4:55516:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465630.255313] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465630.255320] [ndv4:55432:0] 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> }
[1650465630.255352] [ndv4:55432:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650465630.255356] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.255367] [ndv4:55432:0] async.c:228 UCX DEBUG added async handler 0x280db70 [id=87 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.255389] [ndv4:55432:0] async.c:506 UCX DEBUG listening to async event fd 87 events 0x1 mode thread_spinlock
[1650465630.255649] [ndv4:55691:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x3733010: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xf4e3
[1650465630.255822] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465630.255855] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465630.255875] [ndv4:55691:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b871fd00008 of 151544 bytes with 1052 elements
[1650465630.255717] [ndv4:55432:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465630.262408] [ndv4:55432:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x34ac010: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xf4eb
[1650465630.262610] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465630.262616] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465630.262636] [ndv4:55432:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2aee3be6f008 of 151544 bytes with 1052 elements
[1650465630.266343] [ndv4:55432:0] ib_md.c:812 UCX DEBUG registered memory 0x2aee41800000..0x2aee43e00000 on mlx5_ib0 lkey 0x81d00 rkey 0x81d00 access 0xf flags 0x3e4
[1650465630.266361] [ndv4:55432:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2aee41800018 of 39845864 bytes with 4752 elements
[1650465630.266500] [ndv4:55432:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x34ac010
[1650465630.266534] [ndv4:55432:0] ucp_worker.c:1159 UCX DEBUG created interface[8]=0x34ac010 using dc_mlx5/mlx5_ib0:1 on worker 0x3178950
[1650465630.266766] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465630.266776] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465630.266918] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465630.266923] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465630.267420] [ndv4:55432:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465630.268260] [ndv4:55432:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.268626] [ndv4:55432:0] ib_iface.c:994 UCX DEBUG iface=0x28bfb00: created UD QP 0xf4cf on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465630.269126] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465630.269510] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465630.269517] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465630.269810] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465630.269816] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465630.270146] [ndv4:55432:0] ib_md.c:812 UCX DEBUG registered memory 0x2aee3be94000..0x2aee3bf19000 on mlx5_ib0 lkey 0x81e00 rkey 0x81e00 access 0xf flags 0x3e4
[1650465630.270151] [ndv4:55432:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2aee3be94018 of 544744 bytes with 128 elements
[1650465630.270155] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465630.271218] [ndv4:55432:0] ud_iface.c:393 UCX DEBUG iface 0x28bfb00: adding gid fe80::15:5dff:fd33:fffb to hash on device mlx5_ib0 port 1 index 0)
[1650465630.271833] [ndv4:55432:0] ud_iface.c:393 UCX DEBUG iface 0x28bfb00: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650465630.272268] [ndv4:55432:0] ud_iface.c:393 UCX DEBUG iface 0x28bfb00: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650465630.283461] [ndv4:55432:0] ud_iface.c:393 UCX DEBUG iface 0x28bfb00: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650465630.283604] [ndv4:55432:0] ud_iface.c:393 UCX DEBUG iface 0x28bfb00: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650465630.295321] [ndv4:55432:0] ud_iface.c:393 UCX DEBUG iface 0x28bfb00: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650465630.295564] [ndv4:55432:0] ud_iface.c:393 UCX DEBUG iface 0x28bfb0[1650465630.256048] [ndv4:55516:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465630.256069] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.256074] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.256128] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.256471] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.256476] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465630.256989] [ndv4:55516:0] ib_iface.c:994 UCX DEBUG iface=0x3f580a0: created RC QP 0xdaf9 on mlx5_ib4:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465630.257509] [ndv4:55516:0] ucp_worker.c:1159 UCX DEBUG created interface[26]=0x3f580a0 using rc_verbs/mlx5_ib4:1 on worker 0x22b38d0
[1650465630.257623] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.257631] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.257706] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.257710] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.257839] [ndv4:55516:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465630.259062] [ndv4:55516:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.259078] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.259081] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.259135] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.259448] [ndv4:55516:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x4279010 of 8176 bytes with 127 elements
[1650465630.259642] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.259650] [ndv4:55516:0] 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> }
[1650465630.259688] [ndv4:55516:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib4 length=2048) failed: Invalid argument
[1650465630.259693] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.259705] [ndv4:55516:0] async.c:228 UCX DEBUG added async handler 0x304b690 [id=113 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.259732] [ndv4:55516:0] async.c:506 UCX DEBUG listening to async event fd 113 events 0x1 mode thread_spinlock
[1650465630.259742] [ndv4:55516:0] ucp_worker.c:1159 UCX DEBUG created interface[27]=0x40a5030 using rc_mlx5/mlx5_ib4:1 on worker 0x22b38d0
[1650465630.259873] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.259879] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.259950] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.259954] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.260135] [ndv4:55516:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465630.261171] [ndv4:55516:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs [1650465630.262191] [ndv4:55324:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465630.262198] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465630.262485] [ndv4:55324:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.270201] [ndv4:55324:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib6 vendor_id: 0x15b3 device_id: 4124
[1650465630.270565] [ndv4:55324:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib6: disable ODP because it's not supported for DevX QP
[1650465630.272334] [ndv4:55324:0] async.c:228 UCX DEBUG added async handler 0x165dcb0 [id=69 ref 1] uct_ib_async_event_handler() to hash
[1650465630.272356] [ndv4:55324:0] async.c:506 UCX DEBUG listening to async event fd 69 events 0x1 mode thread_spinlock
[1650465630.272359] [ndv4:55324:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib6' (InfiniBand channel adapter) with 1 ports
[1650465630.272696] [ndv4:55324:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.272705] [ndv4:55324:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.272710] [ndv4:55324:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.272760] [ndv4:55324:0] ib_md.c:1319 UCX DEBUG mlx5_ib6: using registration cache
[1650465630.273803] [ndv4:55324:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.273809] [ndv4:55324:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.274047] [ndv4:55324:0] ib_md.c:1604 UCX DEBUG mlx5_ib6: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.286736] [ndv4:55324:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.286748] [ndv4:55324:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.287247] [ndv4:55324:0] topo.c:99 UCX DEBUG bus id 0x107000000 doesn't exist. sys_dev = 6
[1650465630.287252] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465630.299455] [ndv4:55324:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465630.299462] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465630.299817] [ndv4:55324:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465630.299822] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465630.300545] [ndv4:55324:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465630.300549] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465630.301481] [ndv4:55324:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465630.301486] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465630.301766] [ndv4:55324:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.327638] [ndv4:55324:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib7 vendor_id: 0x15b3 device_id: 4124
[1650465630.328012] [ndv4:55324:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib7: disable ODP because it's not supported for DevX QP
[1650465630.329135] [ndv4:55324:0] async.c:228 UCX DEBUG added async handler 0x165f7e0 [id=71 ref 1] uct_ib_async_event_handler() to hash
[1650465630.329158] [ndv4:55324:0] async.c:506 UCX DEBUG listening to async event fd 71 events 0x1 mode thread_spinlock
[1650465630.329161] [ndv4:55324:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib7' (InfiniBand channel adapter) with 1 ports
[1650465630.329357] [ndv4:55324:0] ib_md.c:296 UCX D[1650465630.258933] [ndv4:54717:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465630.258959] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465630.259769] [ndv4:54717:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465630.259774] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465630.260067] [ndv4:54717:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465630.260071] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465630.260440] [ndv4:54717:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.270040] [ndv4:54717:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib6 vendor_id: 0x15b3 device_id: 4124
[1650465630.270383] [ndv4:54717:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib6: disable ODP because it's not supported for DevX QP
[1650465630.272129] [ndv4:54717:0] async.c:228 UCX DEBUG added async handler 0x23e5560 [id=69 ref 1] uct_ib_async_event_handler() to hash
[1650465630.272154] [ndv4:54717:0] async.c:506 UCX DEBUG listening to async event fd 69 events 0x1 mode thread_spinlock
[1650465630.272159] [ndv4:54717:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib6' (InfiniBand channel adapter) with 1 ports
[1650465630.272561] [ndv4:54717:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.272570] [ndv4:54717:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.272594] [ndv4:54717:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.272640] [ndv4:54717:0] ib_md.c:1319 UCX DEBUG mlx5_ib6: using registration cache
[1650465630.274043] [ndv4:54717:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.274049] [ndv4:54717:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.274195] [ndv4:54717:0] ib_md.c:1604 UCX DEBUG mlx5_ib6: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.274259] [ndv4:54717:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.274265] [ndv4:54717:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.274801] [ndv4:54717:0] topo.c:99 UCX DEBUG bus id 0x107000000 doesn't exist. sys_dev = 6
[1650465630.274806] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465630.275201] [ndv4:54717:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465630.275205] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465630.275785] [ndv4:54717:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465630.275789] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465630.276334] [ndv4:54717:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465630.276337] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465630.277238] [ndv4:54717:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465630.277244] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465630.277518] [ndv4:54717:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.286409] [ndv4:54717:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib7 vendor_id: 0x15b3 device_id: 4124
[1650465630.286746] [ndv4:54717:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib7: disable ODP because it's not supported for DevX QP
[1650465630.298564] [ndv4:54717:0] async.c:228 UCX[1650465630.265298] [ndv4:54756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib5:1
[1650465630.266159] [ndv4:54756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.266501] [ndv4:54756:0] ib_iface.c:994 UCX DEBUG iface=0x587a460: created UD QP 0xda7a on mlx5_ib5:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465630.266506] [ndv4:54756:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465630.267036] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465630.267196] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.267203] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.267278] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.267283] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.268010] [ndv4:54756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b3b67c00000..0x2b3b67c85000 on mlx5_ib5 lkey 0x80b00 rkey 0x80b00 access 0xf flags 0x3e4
[1650465630.268017] [ndv4:54756:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b3b67c00018 of 544744 bytes with 128 elements
[1650465630.268022] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465630.268051] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x587a460: adding gid fe80::15:5dff:fd34:0 to hash on device mlx5_ib5 port 1 index 0)
[1650465630.268067] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x587a460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 1)
[1650465630.268309] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x587a460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 2)
[1650465630.268712] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x587a460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 3)
[1650465630.269344] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x587a460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 4)
[1650465630.270456] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x587a460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 5)
[1650465630.271401] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x587a460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 6)
[1650465630.281963] [ndv4:54756:0] ud_iface.c:393 UCX DEBUG iface 0x587a460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 7)
[1650465630.281979] [ndv4:54756:0] 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> }
[1650465630.282064] [ndv4:54756:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.282073] [ndv4:54756:0] async.c:228 UCX DEBUG added async handler 0x575cf20 [id=124 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.282110] [ndv4:54756:0] async.c:506 UCX DEBUG listening to async event fd 124 events 0x5 mode thread_spinlock
[1650465630.282139] [ndv4:54756:0] ucp_worker.c:1159 UCX DEBUG created interface[35]=0x587a460 using ud_mlx5/mlx5_ib5:1 on worker 0x2e9a8d0
[1650465630.282210] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.282219] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.282319] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.282324] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.282801] [ndv4:54756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on [1650465630.261068] [ndv4:54853:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib4 vendor_id: 0x15b3 device_id: 4124
[1650465630.261496] [ndv4:54853:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib4: disable ODP because it's not supported for DevX QP
[1650465630.261899] [ndv4:54853:0] async.c:228 UCX DEBUG added async handler 0x1394280 [id=65 ref 1] uct_ib_async_event_handler() to hash
[1650465630.261926] [ndv4:54853:0] async.c:506 UCX DEBUG listening to async event fd 65 events 0x1 mode thread_spinlock
[1650465630.261930] [ndv4:54853:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib4' (InfiniBand channel adapter) with 1 ports
[1650465630.272037] [ndv4:54853:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.272055] [ndv4:54853:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.272064] [ndv4:54853:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.272128] [ndv4:54853:0] ib_md.c:1319 UCX DEBUG mlx5_ib4: using registration cache
[1650465630.274174] [ndv4:54853:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.274180] [ndv4:54853:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.274400] [ndv4:54853:0] ib_md.c:1604 UCX DEBUG mlx5_ib4: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.274494] [ndv4:54853:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.274500] [ndv4:54853:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.288365] [ndv4:54853:0] topo.c:99 UCX DEBUG bus id 0x105000000 doesn't exist. sys_dev = 4
[1650465630.288372] [ndv4:54853:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465630.288908] [ndv4:54853:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465630.288913] [ndv4:54853:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465630.288976] [ndv4:54853:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465630.288979] [ndv4:54853:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465630.289618] [ndv4:54853:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465630.289621] [ndv4:54853:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465630.300950] [ndv4:54853:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465630.300957] [ndv4:54853:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465630.301185] [ndv4:54853:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.317485] [ndv4:54853:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib5 vendor_id: 0x15b3 device_id: 4124
[1650465630.317903] [ndv4:54853:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib5: disable ODP because it's not supported for DevX QP
[1650465630.318466] [ndv4:54853:0] async.c:228 UCX DEBUG added async handler 0x1392f90 [id=67 ref 1] uct_ib_async_event_handler() to hash
[1650465630.318491] [ndv4:54853:0] async.c:506 UCX DEBUG listening to async event fd 67 events 0x1 mode thread_spinlock
[1650465630.318495] [ndv4:54853:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib5' (InfiniBand channel adapter) with 1 ports
[1650465630.318722] [ndv4:54853:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.318731] [ndv4:54853:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.318738] [ndv4:54853:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.318802] [ndv4:54853:0] ib_md.c:1319 UCX DEBUG mlx5_ib5: usin[1650465630.260402] [ndv4:55691:0] ib_md.c:812 UCX DEBUG registered memory 0x2b8724600000..0x2b8726c00000 on mlx5_ib0 lkey 0x81b00 rkey 0x81b00 access 0xf flags 0x3e4
[1650465630.260426] [ndv4:55691:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b8724600018 of 39845864 bytes with 4752 elements
[1650465630.260656] [ndv4:55691:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x3733010
[1650465630.260691] [ndv4:55691:0] ucp_worker.c:1159 UCX DEBUG created interface[8]=0x3733010 using dc_mlx5/mlx5_ib0:1 on worker 0x33ff770
[1650465630.260779] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465630.260790] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465630.260891] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465630.260895] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465630.261283] [ndv4:55691:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465630.262029] [ndv4:55691:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.262357] [ndv4:55691:0] ib_iface.c:994 UCX DEBUG iface=0x341f5f0: created UD QP 0xf4ce on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465630.262859] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465630.262949] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465630.262957] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465630.263011] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465630.263016] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465630.263344] [ndv4:55691:0] ib_md.c:812 UCX DEBUG registered memory 0x2b871fd25000..0x2b871fdaa000 on mlx5_ib0 lkey 0x81c00 rkey 0x81c00 access 0xf flags 0x3e4
[1650465630.263352] [ndv4:55691:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b871fd25018 of 544744 bytes with 128 elements
[1650465630.263362] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465630.263421] [ndv4:55691:0] ud_iface.c:393 UCX DEBUG iface 0x341f5f0: adding gid fe80::15:5dff:fd33:fffb to hash on device mlx5_ib0 port 1 index 0)
[1650465630.263597] [ndv4:55691:0] ud_iface.c:393 UCX DEBUG iface 0x341f5f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650465630.263696] [ndv4:55691:0] ud_iface.c:393 UCX DEBUG iface 0x341f5f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650465630.263852] [ndv4:55691:0] ud_iface.c:393 UCX DEBUG iface 0x341f5f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650465630.263975] [ndv4:55691:0] ud_iface.c:393 UCX DEBUG iface 0x341f5f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650465630.275343] [ndv4:55691:0] ud_iface.c:393 UCX DEBUG iface 0x341f5f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650465630.275442] [ndv4:55691:0] ud_iface.c:393 UCX DEBUG iface 0x341f5f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650465630.275464] [ndv4:55691:0] ud_iface.c:393 UCX DEBUG iface 0x341f5f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650465630.275745] [ndv4:55691:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.281534] [ndv4:55691:0] async.c:228 UCX DEBUG added async handler 0x2b2e420 [id=88 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.281567] [ndv4:55691:0] async.c:506 UCX DEBUG listening to async event fd 88 events 0x5 m[1650465630.264245] [ndv4:54715:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465630.264254] [ndv4:54715:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465630.264470] [ndv4:54715:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.273815] [ndv4:54715:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib5 vendor_id: 0x15b3 device_id: 4124
[1650465630.274173] [ndv4:54715:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib5: disable ODP because it's not supported for DevX QP
[1650465630.274794] [ndv4:54715:0] async.c:228 UCX DEBUG added async handler 0x200cf90 [id=67 ref 1] uct_ib_async_event_handler() to hash
[1650465630.274817] [ndv4:54715:0] async.c:506 UCX DEBUG listening to async event fd 67 events 0x1 mode thread_spinlock
[1650465630.274820] [ndv4:54715:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib5' (InfiniBand channel adapter) with 1 ports
[1650465630.287671] [ndv4:54715:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.287683] [ndv4:54715:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.287688] [ndv4:54715:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.287733] [ndv4:54715:0] ib_md.c:1319 UCX DEBUG mlx5_ib5: using registration cache
[1650465630.288948] [ndv4:54715:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.288953] [ndv4:54715:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.289156] [ndv4:54715:0] ib_md.c:1604 UCX DEBUG mlx5_ib5: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.289246] [ndv4:54715:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.289252] [ndv4:54715:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.289858] [ndv4:54715:0] topo.c:99 UCX DEBUG bus id 0x106000000 doesn't exist. sys_dev = 5
[1650465630.289863] [ndv4:54715:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465630.301420] [ndv4:54715:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465630.301428] [ndv4:54715:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465630.313487] [ndv4:54715:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465630.313494] [ndv4:54715:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465630.339279] [ndv4:54715:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465630.339287] [ndv4:54715:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465630.353227] [ndv4:54715:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465630.353234] [ndv4:54715:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465630.353452] [ndv4:54715:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.371641] [ndv4:54715:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib6 vendor_id: 0x15b3 device_id: 4124
[1650465630.265772] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.265782] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.267862] [ndv4:55512:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465630.269303] [ndv4:55512:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.269864] [ndv4:55512:0] ib_iface.c:994 UCX DEBUG iface=0x446d1c0: created UD QP 0xdb01 on mlx5_ib4:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465630.271885] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465630.272209] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.272217] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.272348] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.272354] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.272933] [ndv4:55512:0] ib_md.c:812 UCX DEBUG registered memory 0x2aea8146e000..0x2aea814f3000 on mlx5_ib4 lkey 0x81200 rkey 0x81200 access 0xf flags 0x3e4
[1650465630.272939] [ndv4:55512:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2aea8146e018 of 544744 bytes with 128 elements
[1650465630.272944] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465630.284425] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x446d1c0: adding gid fe80::15:5dff:fd33:ffff to hash on device mlx5_ib4 port 1 index 0)
[1650465630.284825] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x446d1c0: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 1)
[1650465630.285174] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x446d1c0: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 2)
[1650465630.285453] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x446d1c0: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 3)
[1650465630.296210] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x446d1c0: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 4)
[1650465630.296648] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x446d1c0: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 5)
[1650465630.296904] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x446d1c0: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 6)
[1650465630.297052] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x446d1c0: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 7)
[1650465630.297686] [ndv4:55512:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.297694] [ndv4:55512:0] async.c:228 UCX DEBUG added async handler 0x3f499f0 [id=116 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.297723] [ndv4:55512:0] async.c:506 UCX DEBUG listening to async event fd 116 events 0x5 mode thread_spinlock
[1650465630.297734] [ndv4:55512:0] ucp_worker.c:1159 UCX DEBUG created interface[29]=0x446d1c0 using ud_verbs/mlx5_ib4:1 on worker 0x22c98d0
[1650465630.297773] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.297780] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.297860] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.297865] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.298139] [ndv4:55512:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465630.299883] [ndv4:55512:0] ib_iface.c:1469 UCX DEBUG created uct_ib[1650465630.261558] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x31477b0: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 3)
[1650465630.261660] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x31477b0: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 4)
[1650465630.261779] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x31477b0: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 5)
[1650465630.261804] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x31477b0: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 6)
[1650465630.262017] [ndv4:54965:0] ud_iface.c:393 UCX DEBUG iface 0x31477b0: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 7)
[1650465630.262023] [ndv4:54965:0] 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> }
[1650465630.262057] [ndv4:54965:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.262061] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x2cc3580 [id=138 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.262084] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 138 events 0x5 mode thread_spinlock
[1650465630.262094] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[45]=0x31477b0 using ud_mlx5/mlx5_ib7:1 on worker 0x23dc840
[1650465630.262155] [ndv4:54965:0] mpool.c:88 UCX DEBUG mpool uct_scopy_iface_tx_mp: align 64, maxelems 4294967295, elemsize 736
[1650465630.262194] [ndv4:54965:0] ucp_worker.c:1159 UCX DEBUG created interface[46]=0x5bf5880 using cma/memory on worker 0x23dc840
[1650465630.262201] [ndv4:54965:0] ucp_worker.c:982 UCX DEBUG selected scalable tl bitmap: 0x7fffffffffff 0x0 (47 tls)
[1650465630.272104] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x55b35c0 [id=75 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465630.272132] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 75 events 0x0 mode thread_spinlock
[1650465630.272179] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x3147600 [id=76 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465630.272195] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 76 events 0x0 mode thread_spinlock
[1650465630.272213] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x3147640 [id=77 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465630.272230] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 77 events 0x0 mode thread_spinlock
[1650465630.273023] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x3147680 [id=79 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465630.273047] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 79 events 0x0 mode thread_spinlock
[1650465630.273065] [ndv4:54965:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465630.273070] [ndv4:54965:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465630.273131] [ndv4:54965:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465630.273136] [ndv4:54965:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465630.273189] [ndv4:54965:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465630.273194] [ndv4:54965:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465630.273245] [ndv4:54965:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465630.273250] [ndv4:54965:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465630.273301] [ndv4:54965:0] sock.c:88 UCX0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650465630.295804] [ndv4:55432:0] ud_iface.c:393 UCX DEBUG iface 0x28bfb00: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650465630.296046] [ndv4:55432:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.301565] [ndv4:55432:0] async.c:228 UCX DEBUG added async handler 0x28a3e10 [id=88 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.301644] [ndv4:55432:0] async.c:506 UCX DEBUG listening to async event fd 88 events 0x5 mode thread_spinlock
[1650465630.301673] [ndv4:55432:0] ucp_worker.c:1159 UCX DEBUG created interface[9]=0x28bfb00 using ud_verbs/mlx5_ib0:1 on worker 0x3178950
[1650465630.301771] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465630.301778] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465630.301914] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465630.301919] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465630.302798] [ndv4:55432:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465630.303656] [ndv4:55432:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.304020] [ndv4:55432:0] ib_iface.c:994 UCX DEBUG iface=0x28c1390: created UD QP 0xf4d1 on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465630.304030] [ndv4:55432:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465630.304560] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465630.304625] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465630.304631] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465630.304724] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465630.304729] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465630.305133] [ndv4:55432:0] ib_md.c:812 UCX DEBUG registered memory 0x2aee3bf19000..0x2aee3bf9e000 on mlx5_ib0 lkey 0x82000 rkey 0x82000 access 0xf flags 0x3e4
[1650465630.305138] [ndv4:55432:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2aee3bf19018 of 544744 bytes with 128 elements
[1650465630.305143] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465630.305337] [ndv4:55432:0] ud_iface.c:393 UCX DEBUG iface 0x28c1390: adding gid fe80::15:5dff:fd33:fffb to hash on device mlx5_ib0 port 1 index 0)
[1650465630.305546] [ndv4:55432:0] ud_iface.c:393 UCX DEBUG iface 0x28c1390: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650465630.305734] [ndv4:55432:0] ud_iface.c:393 UCX DEBUG iface 0x28c1390: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650465630.306009] [ndv4:55432:0] ud_iface.c:393 UCX DEBUG iface 0x28c1390: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650465630.306120] [ndv4:55432:0] ud_iface.c:393 UCX DEBUG iface 0x28c1390: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650465630.306412] [ndv4:55432:0] ud_iface.c:393 UCX DEBUG iface 0x28c1390: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650465630.306721] [ndv4:55432:0] ud_iface.c:393 UCX DEBUG iface 0x28c1390: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650465630.307246] [ndv4:55432:0] ud_iface.c:393 UCX DEBUG iface 0x28c1390: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650465630.307252] [ndv4:55432:0] ib_mlx5.c:858 UCX DEBUG SL=0 (A[1650465630.265194] [ndv4:55039:0] async.c:228 UCX DEBUG added async handler 0x20bebf0 [id=65 ref 1] uct_ib_async_event_handler() to hash
[1650465630.265227] [ndv4:55039:0] async.c:506 UCX DEBUG listening to async event fd 65 events 0x1 mode thread_spinlock
[1650465630.265231] [ndv4:55039:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib4' (InfiniBand channel adapter) with 1 ports
[1650465630.265511] [ndv4:55039:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.265520] [ndv4:55039:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.265526] [ndv4:55039:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.265602] [ndv4:55039:0] ib_md.c:1319 UCX DEBUG mlx5_ib4: using registration cache
[1650465630.267981] [ndv4:55039:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.267986] [ndv4:55039:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.268153] [ndv4:55039:0] ib_md.c:1604 UCX DEBUG mlx5_ib4: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.268280] [ndv4:55039:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.268287] [ndv4:55039:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.270522] [ndv4:55039:0] topo.c:99 UCX DEBUG bus id 0x105000000 doesn't exist. sys_dev = 4
[1650465630.270527] [ndv4:55039:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465630.281643] [ndv4:55039:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465630.281650] [ndv4:55039:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465630.282039] [ndv4:55039:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465630.282043] [ndv4:55039:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465630.294369] [ndv4:55039:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465630.294376] [ndv4:55039:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465630.295817] [ndv4:55039:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650465630.295821] [ndv4:55039:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650465630.296054] [ndv4:55039:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.322013] [ndv4:55039:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib5 vendor_id: 0x15b3 device_id: 4124
[1650465630.322346] [ndv4:55039:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib5: disable ODP because it's not supported for DevX QP
[1650465630.322996] [ndv4:55039:0] async.c:228 UCX DEBUG added async handler 0x19d9710 [id=67 ref 1] uct_ib_async_event_handler() to hash
[1650465630.323021] [ndv4:55039:0] async.c:506 UCX DEBUG listening to async event fd 67 events 0x1 mode thread_spinlock
[1650465630.323024] [ndv4:55039:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib5' (InfiniBand channel adapter) with 1 ports
[1650465630.323179] [ndv4:55039:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.323186] [ndv4:55039:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.323191] [ndv4:55039:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.323243] [ndv4:55039:0] ib_md.c:1319 UCX DEBUG mlx5_ib5: using registration cache
[1650465630.324276] [ndv4:55039:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.324282] [ndv4:55039:0] mpool.c:88 UCX D[1650465630.257746] [ndv4:54861:0] topo.c:99 UCX DEBUG bus id 0x106000000 doesn't exist. sys_dev = 5
[1650465630.257755] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465630.269454] [ndv4:54861:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465630.269461] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465630.283628] [ndv4:54861:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465630.283635] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465630.284656] [ndv4:54861:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465630.284659] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465630.308327] [ndv4:54861:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465630.308334] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465630.308560] [ndv4:54861:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.318344] [ndv4:54861:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib6 vendor_id: 0x15b3 device_id: 4124
[1650465630.318756] [ndv4:54861:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib6: disable ODP because it's not supported for DevX QP
[1650465630.319642] [ndv4:54861:0] async.c:228 UCX DEBUG added async handler 0x18bf720 [id=69 ref 1] uct_ib_async_event_handler() to hash
[1650465630.319666] [ndv4:54861:0] async.c:506 UCX DEBUG listening to async event fd 69 events 0x1 mode thread_spinlock
[1650465630.319671] [ndv4:54861:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib6' (InfiniBand channel adapter) with 1 ports
[1650465630.319906] [ndv4:54861:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.319914] [ndv4:54861:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.319920] [ndv4:54861:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.319986] [ndv4:54861:0] ib_md.c:1319 UCX DEBUG mlx5_ib6: using registration cache
[1650465630.334032] [ndv4:54861:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.334041] [ndv4:54861:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.334260] [ndv4:54861:0] ib_md.c:1604 UCX DEBUG mlx5_ib6: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.334282] [ndv4:54861:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.334288] [ndv4:54861:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.334430] [ndv4:54861:0] topo.c:99 UCX DEBUG bus id 0x107000000 doesn't exist. sys_dev = 6
[1650465630.334435] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465630.334557] [ndv4:54861:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465630.334560] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465630.334727] [ndv4:54861:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465630.334731] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465630.334844] [ndv4:54861:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465630.334847] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465630.334965] [ndv4:54861:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465630.334968] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465630.335258] [ndv4:5486[1650465630.344293] [ndv4:54724:0] debug.c:1198 UCX DEBUG using signal stack 0x2b5dafc84000 size 141824
[1650465630.344370] [ndv4:54724:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465630.344393] [ndv4:54724:0] 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 0x2b5dafaf2000
[1650465630.344414] [ndv4:54724:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650465630.344423] [ndv4:54724:0] 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
[1650465630.344429] [ndv4:54724:0] module.c:253 UCX DEBUG loading modules for ucs
[1650465630.347381] [ndv4:54724:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465630.347404] [ndv4:54724:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650465630.347448] [ndv4:54724:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650465630.347451] [ndv4:54724:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650465630.347460] [ndv4:54724:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650465630.347469] [ndv4:54724:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650465630.347472] [ndv4:54724:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650465630.347479] [ndv4:54724:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650465630.347481] [ndv4:54724:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650465630.347485] [ndv4:54724:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650465630.347487] [ndv4:54724:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650465630.347489] [ndv4:54724:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650465630.347503] [ndv4:54724:0] module.c:253 UCX DEBUG loading modules for uct
[1650465630.348391] [ndv4:54724:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650465630.349080] [ndv4:54724:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650465630.349096] [ndv4:54724:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650465630.349107] [ndv4:54724:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650465630.349117] [ndv4:54724:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650465630.349127] [ndv4:54724:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650465630.349137] [ndv4:54724:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650465630.349148] [ndv4:54724:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650465630.349518] [ndv4:54724:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650465630.367046] [ndv4:54724:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650465630.367463] [ndv4:54724:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650465630.272858] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x5e3b060: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 7)
[1650465630.273463] [ndv4:55152:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.273473] [ndv4:55152:0] async.c:228 UCX DEBUG added async handler 0x59179b0 [id=123 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.273503] [ndv4:55152:0] async.c:506 UCX DEBUG listening to async event fd 123 events 0x5 mode thread_spinlock
[1650465630.273521] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[34]=0x5e3b060 using ud_verbs/mlx5_ib5:1 on worker 0x357a8c0
[1650465630.273688] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.273695] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.273784] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.273789] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.285510] [ndv4:55152:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib5:1
[1650465630.297920] [ndv4:55152:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.298514] [ndv4:55152:0] ib_iface.c:994 UCX DEBUG iface=0x5f59460: created UD QP 0xda7b on mlx5_ib5:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465630.298521] [ndv4:55152:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465630.299247] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465630.299425] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.299433] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.299479] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.299484] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.299952] [ndv4:55152:0] ib_md.c:812 UCX DEBUG registered memory 0x2b3b97f45000..0x2b3b97fca000 on mlx5_ib5 lkey 0x80c00 rkey 0x80c00 access 0xf flags 0x3e4
[1650465630.299958] [ndv4:55152:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b3b97f45018 of 544744 bytes with 128 elements
[1650465630.299962] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465630.299990] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x5f59460: adding gid fe80::15:5dff:fd34:0 to hash on device mlx5_ib5 port 1 index 0)
[1650465630.300007] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x5f59460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 1)
[1650465630.300022] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x5f59460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 2)
[1650465630.300168] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x5f59460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 3)
[1650465630.300558] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x5f59460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 4)
[1650465630.300931] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x5f59460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 5)
[1650465630.300957] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x5f59460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 6)
[1650465630.301498] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x5f59460: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 7)
[1650465630.301505] [ndv4:55152:0] ib_mlx5.c:858 UCX DEBUG SL=0 (AR support - unknown) was selected on mlx5_ib5:1, SLs with AR sup[1650465630.372054] [ndv4:54715:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib6: disable ODP because it's not supported for DevX QP
_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.300300] [ndv4:55512:0] ib_iface.c:994 UCX DEBUG iface=0x377e7b0: created UD QP 0xdb07 on mlx5_ib4:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465630.300309] [ndv4:55512:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465630.302445] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465630.302608] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.302614] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.302771] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.302776] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.303975] [ndv4:55512:0] ib_md.c:812 UCX DEBUG registered memory 0x2aea814f3000..0x2aea81578000 on mlx5_ib4 lkey 0x81400 rkey 0x81400 access 0xf flags 0x3e4
[1650465630.303980] [ndv4:55512:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2aea814f3018 of 544744 bytes with 128 elements
[1650465630.303985] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465630.304210] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x377e7b0: adding gid fe80::15:5dff:fd33:ffff to hash on device mlx5_ib4 port 1 index 0)
[1650465630.304239] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x377e7b0: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 1)
[1650465630.304400] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x377e7b0: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 2)
[1650465630.304521] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x377e7b0: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 3)
[1650465630.304629] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x377e7b0: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 4)
[1650465630.304772] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x377e7b0: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 5)
[1650465630.304809] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x377e7b0: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 6)
[1650465630.304828] [ndv4:55512:0] ud_iface.c:393 UCX DEBUG iface 0x377e7b0: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 7)
[1650465630.304835] [ndv4:55512:0] 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> }
[1650465630.304867] [ndv4:55512:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.304872] [ndv4:55512:0] async.c:228 UCX DEBUG added async handler 0x458b960 [id=117 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.304897] [ndv4:55512:0] async.c:506 UCX DEBUG listening to async event fd 117 events 0x5 mode thread_spinlock
[1650465630.304908] [ndv4:55512:0] ucp_worker.c:1159 UCX DEBUG created interface[30]=0x377e7b0 using ud_mlx5/mlx5_ib4:1 on worker 0x22c98d0
[1650465630.304921] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.304927] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.305002] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.305008] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.305177] [ndv4:55512:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib5:1
[1650465630.306205] [ndv4:55512:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465630.306220] [ndv4:port = { <none> }, SLs without AR support = { <none> }
[1650465630.301541] [ndv4:55152:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.301547] [ndv4:55152:0] async.c:228 UCX DEBUG added async handler 0x5e3bf20 [id=124 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.301619] [ndv4:55152:0] async.c:506 UCX DEBUG listening to async event fd 124 events 0x5 mode thread_spinlock
[1650465630.301637] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[35]=0x5f59460 using ud_mlx5/mlx5_ib5:1 on worker 0x357a8c0
[1650465630.301729] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.301735] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.301903] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.301908] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.302819] [ndv4:55152:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib6:1
[1650465630.303849] [ndv4:55152:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465630.303865] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.303868] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.303915] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.304201] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.304206] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465630.304790] [ndv4:55152:0] ib_iface.c:994 UCX DEBUG iface=0x60590a0: created RC QP 0xdac7 on mlx5_ib6:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465630.305350] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[36]=0x60590a0 using rc_verbs/mlx5_ib6:1 on worker 0x357a8c0
[1650465630.305467] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.305474] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.305541] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.305546] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.305876] [ndv4:55152:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib6:1
[1650465630.308653] [ndv4:55152:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.308672] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.308676] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.308731] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.309127] [ndv4:55152:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x637a010 of 8176 bytes with 127 elements
[1650465630.309319] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.309326] [ndv4:55152:0] 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> }
[1650465630.309362] [ndv4:55152:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib6 length=2048) failed: Invalid argument
[1650465630.309367] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpooR support - unknown) was selected on mlx5_ib0:1, SLs with AR support = { <none> }, SLs without AR support = { <none> }
[1650465630.307287] [ndv4:55432:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.307292] [ndv4:55432:0] async.c:228 UCX DEBUG added async handler 0x2130500 [id=89 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.307313] [ndv4:55432:0] async.c:506 UCX DEBUG listening to async event fd 89 events 0x5 mode thread_spinlock
[1650465630.307328] [ndv4:55432:0] ucp_worker.c:1159 UCX DEBUG created interface[10]=0x28c1390 using ud_mlx5/mlx5_ib0:1 on worker 0x3178950
[1650465630.307498] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.307505] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.315045] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.315054] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.316524] [ndv4:55432:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
EBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.324431] [ndv4:55039:0] ib_md.c:1604 UCX DEBUG mlx5_ib5: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.324472] [ndv4:55039:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.324478] [ndv4:55039:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.325019] [ndv4:55039:0] topo.c:99 UCX DEBUG bus id 0x106000000 doesn't exist. sys_dev = 5
[1650465630.325024] [ndv4:55039:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465630.356146] [ndv4:55039:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465630.356155] [ndv4:55039:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465630.368755] [ndv4:55039:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465630.368762] [ndv4:55039:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465630.369792] [ndv4:55039:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465630.369796] [ndv4:55039:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
1:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.365719] [ndv4:54861:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib7 vendor_id: 0x15b3 device_id: 4124
[1650465630.366124] [ndv4:54861:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib7: disable ODP because it's not supported for DevX QP
DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465630.273305] [ndv4:54965:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465630.273358] [ndv4:54965:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465630.273362] [ndv4:54965:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465630.273414] [ndv4:54965:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465630.273419] [ndv4:54965:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465630.273470] [ndv4:54965:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465630.273475] [ndv4:54965:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465630.273528] [ndv4:54965:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465630.273533] [ndv4:54965:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465630.273645] [ndv4:54965:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465630.273650] [ndv4:54965:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465630.273706] [ndv4:54965:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465630.273711] [ndv4:54965:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465630.273763] [ndv4:54965:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465630.273768] [ndv4:54965:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465630.273821] [ndv4:54965:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465630.273826] [ndv4:54965:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465630.273874] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x31476c0 [id=81 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465630.273895] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 81 events 0x0 mode thread_spinlock
[1650465630.276446] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x3147700 [id=83 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465630.276467] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 83 events 0x0 mode thread_spinlock
[1650465630.276537] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x3147740 [id=84 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465630.276554] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 84 events 0x0 mode thread_spinlock
[1650465630.276587] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x3e63e80 [id=86 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465630.276635] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 86 events 0x0 mode thread_spinlock
[1650465630.276679] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x3e63ec0 [id=90 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465630.276694] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 90 events 0x0 mode thread_spinlock
[1650465630.276710] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x3e63f00 [id=91 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465630.276731] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 91 events 0x0 mode thread_spinlock
[1650465630.276745] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x3e63f40 [id=93 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465630.276761] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 93 e90 data_sz 8256
[1650465630.261186] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.261190] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.261252] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.261558] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465630.261566] [ndv4:55516:0] 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> }
[1650465630.261620] [ndv4:55516:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib4 length=2048) failed: Invalid argument
[1650465630.261624] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.261633] [ndv4:55516:0] async.c:228 UCX DEBUG added async handler 0x40adf10 [id=115 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.261655] [ndv4:55516:0] async.c:506 UCX DEBUG listening to async event fd 115 events 0x1 mode thread_spinlock
[1650465630.262013] [ndv4:55516:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465630.274772] [ndv4:55516:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x427b050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xdb0f
[1650465630.275021] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.275029] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.275038] [ndv4:55516:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2ae25d66e008 of 151544 bytes with 1052 elements
[1650465630.278826] [ndv4:55516:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae25b000000..0x2ae25d600000 on mlx5_ib4 lkey 0x81300 rkey 0x81300 access 0xf flags 0x3e4
[1650465630.278844] [ndv4:55516:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2ae25b000018 of 39845864 bytes with 4752 elements
[1650465630.279017] [ndv4:55516:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x427b050
[1650465630.279048] [ndv4:55516:0] ucp_worker.c:1159 UCX DEBUG created interface[28]=0x427b050 using dc_mlx5/mlx5_ib4:1 on worker 0x22b38d0
[1650465630.279102] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.279109] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.279159] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.279163] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.279458] [ndv4:55516:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
[1650465630.280484] [ndv4:55516:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.280808] [ndv4:55516:0] ib_iface.c:994 UCX DEBUG iface=0x44571c0: created UD QP 0xdb06 on mlx5_ib4:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465630.281229] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465630.306131] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.306145] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.306171] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.306176] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.306824] [ndv4:55516:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae25d693000..0x2ae25d718000 on mlx5_ib4 lkeyEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465630.329366] [ndv4:55324:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465630.329371] [ndv4:55324:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.329415] [ndv4:55324:0] ib_md.c:1319 UCX DEBUG mlx5_ib7: using registration cache
[1650465630.332113] [ndv4:55324:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.332119] [ndv4:55324:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.332304] [ndv4:55324:0] ib_md.c:1604 UCX DEBUG mlx5_ib7: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.332321] [ndv4:55324:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465630.332326] [ndv4:55324:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465630.332470] [ndv4:55324:0] topo.c:99 UCX DEBUG bus id 0x108000000 doesn't exist. sys_dev = 7
[1650465630.332474] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465630.332642] [ndv4:55324:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465630.332646] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465630.332778] [ndv4:55324:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465630.332781] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465630.332889] [ndv4:55324:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465630.332892] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465630.333010] [ndv4:55324:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465630.333013] [ndv4:55324:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465630.333110] [ndv4:55324:0] ucp_context.c:1556 UCX DEBUG created ucp context 0x164ad70 0x164ad70 [13 mds 47 tls] features 0x1 tl bitmap 0x7fffffffffff 0x0
DEBUG added async handler 0x2acb940 [id=71 ref 1] uct_ib_async_event_handler() to hash
[1650465630.298633] [ndv4:54717:0] async.c:506 UCX DEBUG listening to async event fd 71 events 0x1 mode thread_spinlock
[1650465630.298639] [ndv4:54717:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib7' (InfiniBand channel adapter) with 1 ports
[1650465630.298812] [ndv4:54717:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465630.298832] [ndv4:54717:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465630.298839] [ndv4:54717:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.298895] [ndv4:54717:0] ib_md.c:1319 UCX DEBUG mlx5_ib7: using registration cache
[1650465630.302000] [ndv4:54717:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.302007] [ndv4:54717:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.302259] [ndv4:54717:0] ib_md.c:1604 UCX DEBUG mlx5_ib7: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.302434] [ndv4:54717:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465630.302442] [ndv4:54717:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465630.303398] [ndv4:54717:0] topo.c:99 UCX DEBUG bus id 0x108000000 doesn't exist. sys_dev = 7
[1650465630.303403] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465630.303863] [ndv4:54717:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465630.303867] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465630.316246] [ndv4:54717:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465630.316252] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465630.318294] [ndv4:54717:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465630.318299] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465630.320216] [ndv4:54717:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465630.320227] [ndv4:54717:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465630.320362] [ndv4:54717:0] ucp_context.c:1556 UCX DEBUG created ucp context 0x23cec90 0x23cec90 [13 mds 47 tls] features 0x1 tl bitmap 0x7fffffffffff 0x0
[1650465630.362135] [ndv4:54717:0] mm_posix.c:518 UCX DEBUG allocated posix shared memory at 0x2b7cc3f0f000 length 12288
[1650465630.362204] [ndv4:54717:0] mpool.c:88 UCX DEBUG mpool mm_recv_desc: align 64, maxelems 4294967295, elemsize 8368
[1650465630.363217] [ndv4:54717:0] mm_posix.c:305 UCX DEBUG shared memory mmap(addr=(nil), length=6291456, flags= HUGETLB, fd=76) failed: Invalid argument
[1650465630.363225] [ndv4:54717:0] mm_posix.c:518 UCX DEBUG allocated posix shared memory at 0x2b7cc8423000 length 4296704
[1650465630.363230] [ndv4:54717:0] mpool.c:205 UCX DEBUG mpool mm_recv_desc: allocated chunk 0x2b7cc8423018 of 4296680 bytes with 512 elements
[1650465630.363487] [ndv4:54717:0] mm_iface.c:600 UCX DEBUG created mm iface 0x2b4bd10 FIFO id 0x4000000049a0888e va 0x2b7cc3f0f000 size 12288 (128 x 64 elems)
[1650465630.363537] [ndv4:54717:0] ucp_worker.c:1159 UCX DEBUG created interface[0]=0x2b4bd10 using posix/memory on worker 0x34258d0
[1650465630.363562] [ndv4:54717:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 8447 bytes with hugetlb
[1650465630.363626] [ndv4:54717:0] mpool.c:88 UCX DEBUG mpool mm_recv_desc: align 64, maxelems 4294967295, elemsize 8368
[1650465630.363644] [ndv4:54717:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 4292720 bytes with hugetlmlx5_ib6:1
[1650465630.283851] [ndv4:54756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465630.283871] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.283875] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.283927] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.284420] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.284426] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465630.285553] [ndv4:54756:0] ib_iface.c:994 UCX DEBUG iface=0x597a0a0: created RC QP 0xdac6 on mlx5_ib6:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465630.286526] [ndv4:54756:0] ucp_worker.c:1159 UCX DEBUG created interface[36]=0x597a0a0 using rc_verbs/mlx5_ib6:1 on worker 0x2e9a8d0
[1650465630.286547] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.286553] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.286798] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.286805] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.287174] [ndv4:54756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib6:1
[1650465630.288444] [ndv4:54756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.288456] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.288459] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.288511] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.288907] [ndv4:54756:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x5c9b010 of 8176 bytes with 127 elements
[1650465630.289145] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.289152] [ndv4:54756:0] 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> }
[1650465630.289189] [ndv4:54756:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib6 length=2048) failed: Invalid argument
[1650465630.289193] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.289204] [ndv4:54756:0] async.c:228 UCX DEBUG added async handler 0x5acfc40 [id=127 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.289227] [ndv4:54756:0] async.c:506 UCX DEBUG listening to async event fd 127 events 0x1 mode thread_spinlock
[1650465630.289236] [ndv4:54756:0] ucp_worker.c:1159 UCX DEBUG created interface[37]=0x5ac7030 using rc_mlx5/mlx5_ib6:1 on worker 0x2e9a8d0
[1650465630.289313] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.289319] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.289445] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.289450] [ndv4:54756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.300806] [ndv4:54756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib6:1
[1650465630.302532] [ndv4:54756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs g registration cache
[1650465630.321115] [ndv4:54853:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.321122] [ndv4:54853:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.321303] [ndv4:54853:0] ib_md.c:1604 UCX DEBUG mlx5_ib5: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.321366] [ndv4:54853:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.321373] [ndv4:54853:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.339741] [ndv4:54853:0] topo.c:99 UCX DEBUG bus id 0x106000000 doesn't exist. sys_dev = 5
[1650465630.339749] [ndv4:54853:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465630.340720] [ndv4:54853:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465630.340725] [ndv4:54853:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465630.364799] [ndv4:54853:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465630.364807] [ndv4:54853:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
ode thread_spinlock
[1650465630.281613] [ndv4:55691:0] ucp_worker.c:1159 UCX DEBUG created interface[9]=0x341f5f0 using ud_verbs/mlx5_ib0:1 on worker 0x33ff770
[1650465630.281629] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465630.281635] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465630.281681] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465630.281686] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465630.282184] [ndv4:55691:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib0:1
[1650465630.283167] [ndv4:55691:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.283539] [ndv4:55691:0] ib_iface.c:994 UCX DEBUG iface=0x2b469f0: created UD QP 0xf4d0 on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465630.283549] [ndv4:55691:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650465630.284075] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465630.284229] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465630.284258] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465630.284309] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465630.284314] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465630.284694] [ndv4:55691:0] ib_md.c:812 UCX DEBUG registered memory 0x2b871fdaa000..0x2b871fe2f000 on mlx5_ib0 lkey 0x81f00 rkey 0x81f00 access 0xf flags 0x3e4
[1650465630.284703] [ndv4:55691:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b871fdaa018 of 544744 bytes with 128 elements
[1650465630.284713] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465630.284771] [ndv4:55691:0] ud_iface.c:393 UCX DEBUG iface 0x2b469f0: adding gid fe80::15:5dff:fd33:fffb to hash on device mlx5_ib0 port 1 index 0)
[1650465630.285199] [ndv4:55691:0] ud_iface.c:393 UCX DEBUG iface 0x2b469f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650465630.285306] [ndv4:55691:0] ud_iface.c:393 UCX DEBUG iface 0x2b469f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650465630.285455] [ndv4:55691:0] ud_iface.c:393 UCX DEBUG iface 0x2b469f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650465630.285616] [ndv4:55691:0] ud_iface.c:393 UCX DEBUG iface 0x2b469f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650465630.285632] [ndv4:55691:0] ud_iface.c:393 UCX DEBUG iface 0x2b469f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650465630.285646] [ndv4:55691:0] ud_iface.c:393 UCX DEBUG iface 0x2b469f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650465630.285751] [ndv4:55691:0] ud_iface.c:393 UCX DEBUG iface 0x2b469f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650465630.285759] [ndv4:55691:0] 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> }
[1650465630.285856] [ndv4:55691:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.285862] [ndv4:55691:0] async.c:228 UCX DEBUG added async handler 0x2a94b00 [id=89 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.285887] [ndv4:55691:0] async.c:506 UCX DEBUG listening to async event fd 89 events 0x5 mode thread_spinlock
[1650465630.285903] [ndv4:55691:0] ucp_worker.c:1159 UCX DEBUG crevents 0x0 mode thread_spinlock
[1650465630.276796] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x393f3c0 [id=97 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465630.276817] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 97 events 0x0 mode thread_spinlock
[1650465630.276832] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x393f400 [id=98 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465630.276848] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 98 events 0x0 mode thread_spinlock
[1650465630.276862] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x393f440 [id=100 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465630.276877] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 100 events 0x0 mode thread_spinlock
[1650465630.276912] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x393f480 [id=104 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465630.276926] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 104 events 0x0 mode thread_spinlock
[1650465630.276940] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x393f4c0 [id=105 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465630.276953] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 105 events 0x0 mode thread_spinlock
[1650465630.373524] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x393f500 [id=107 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465630.373553] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 107 events 0x0 mode thread_spinlock
[1650465630.373619] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x393f540 [id=111 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465630.373647] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 111 events 0x0 mode thread_spinlock
[1650465630.373668] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x393f580 [id=112 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465630.373717] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 112 events 0x0 mode thread_spinlock
[1650465630.373733] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x5bf5dd0 [id=114 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465630.373771] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 114 events 0x0 mode thread_spinlock
[1650465630.373808] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x5bf5e10 [id=118 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465630.373826] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 118 events 0x0 mode thread_spinlock
[1650465630.373843] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x5bf5e50 [id=119 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465630.373880] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 119 events 0x0 mode thread_spinlock
[1650465630.373895] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x5bf5e90 [id=121 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465630.373929] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 121 events 0x0 mode thread_spinlock
[1650465630.373966] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x5bf5ed0 [id=125 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465630.373985] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 125 events 0x0 mode thread_spinlock
[1650465630.374001] [ndv4:54965:0] async.c:228 UCX DEBUG added async handler 0x5bf5f10 [id=126 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650465630.374018] [ndv4:54965:0] async.c:506 UCX DEBUG listening to async event fd 126 events 0x0 mode thread_spinlock
[1650465630.374033] [n 0x81500 rkey 0x81500 access 0xf flags 0x3e4
[1650465630.306832] [ndv4:55516:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae25d693018 of 544744 bytes with 128 elements
[1650465630.306837] [ndv4:55516:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465630.306986] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x44571c0: adding gid fe80::15:5dff:fd33:ffff to hash on device mlx5_ib4 port 1 index 0)
[1650465630.307152] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x44571c0: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 1)
[1650465630.307616] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x44571c0: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 2)
[1650465630.308115] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x44571c0: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 3)
[1650465630.308595] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x44571c0: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 4)
[1650465630.308772] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x44571c0: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 5)
[1650465630.309322] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x44571c0: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 6)
[1650465630.374136] [ndv4:55516:0] ud_iface.c:393 UCX DEBUG iface 0x44571c0: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 7)
[1650465630.374511] [ndv4:55516:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.374524] [ndv4:55516:0] async.c:228 UCX DEBUG added async handler 0x3f339f0 [id=116 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.374550] [ndv4:55516:0] async.c:506 UCX DEBUG listening to async event fd 116 events 0x5 mode thread_spinlock
[1650465630.374571] [ndv4:55516:0] ucp_worker.c:1159 UCX DEBUG created interface[29]=0x44571c0 using ud_verbs/mlx5_ib4:1 on worker 0x22b38d0
[1650465630.374757] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.374764] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.374912] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650465630.374918] [ndv4:55516:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650465630.375267] [ndv4:55516:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib4:1
b
[1650465630.363654] [ndv4:54717:0] mpool.c:205 UCX DEBUG mpool mm_recv_desc: allocated chunk 0x2b7cc883c018 of 4296680 bytes with 512 elements
[1650465630.364215] [ndv4:54717:0] mm_iface.c:600 UCX DEBUG created mm iface 0x2b4c2e0 FIFO id 0x720035 va 0x2b7cc3f12000 size 12288 (128 x 64 elems)
[1650465630.364223] [ndv4:54717:0] ucp_worker.c:1159 UCX DEBUG created interface[1]=0x2b4c2e0 using sysv/memory on worker 0x34258d0
[1650465630.364236] [ndv4:54717:0] mpool.c:88 UCX DEBUG mpool self_msg_desc: align 64, maxelems 4294967295, elemsize 8200
[1650465630.364239] [ndv4:54717:0] self.c:220 UCX DEBUG created self iface id 0xce7e29e3d1ed3a06 send_size 8192
[1650465630.364245] [ndv4:54717:0] ucp_worker.c:1159 UCX DEBUG created interface[2]=0x2b5fec0 using self/memory0 on worker 0x34258d0
[1650465630.364268] [ndv4:54717:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465630.364274] [ndv4:54717:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465630.364277] [ndv4:54717:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465630.367264] [ndv4:54717:0] async.c:228 UCX DEBUG added async handler 0x2b62000 [id=78 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465630.367296] [ndv4:54717:0] async.c:506 UCX DEBUG listening to async event fd 78 events 0x5 mode thread_spinlock
[1650465630.373550] [ndv4:54717:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x2b60820: listening for connections (fd=78) on 10.5.0.5:33858
[1650465630.373689] [ndv4:54717:0] ucp_worker.c:1159 UCX DEBUG created interface[3]=0x2b60820 using tcp/eth0 on worker 0x34258d0
[1650465630.373711] [ndv4:54717:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465630.373715] [ndv4:54717:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465630.373719] [ndv4:54717:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465630.373758] [ndv4:54717:0] async.c:228 UCX DEBUG added async handler 0x2aae5c0 [id=80 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465630.373783] [ndv4:54717:0] async.c:506 UCX DEBUG listening to async event fd 80 events 0x5 mode thread_spinlock
[1650465630.373787] [ndv4:54717:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x2b60ea0: listening for connections (fd=80) on 127.0.0.1:58703
[1650465630.373804] [ndv4:54717:0] sock.c:88 UCX DEBUG ioctl(req=35142, ifr_name=lo) failed: Operation not supported
[1650465630.373810] [ndv4:54717:0] tcp_net.c:61 UCX DEBUG speed of lo is UNKNOWN, assuming 100 Mbps
[1650465630.373877] [ndv4:54717:0] ucp_worker.c:1159 UCX DEBUG created interface[4]=0x2b60ea0 using tcp/lo on worker 0x34258d0
[1650465630.373895] [ndv4:54717:0] tcp_iface.c:547 UCX DEBUG using TCP port range: 0-0
[1650465630.373898] [ndv4:54717:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_tx_buf_mp: align 64, maxelems 4294967295, elemsize 8205
[1650465630.373901] [ndv4:54717:0] mpool.c:88 UCX DEBUG mpool uct_tcp_iface_rx_buf_mp: align 64, maxelems 4294967295, elemsize 131090
[1650465630.373938] [ndv4:54717:0] async.c:228 UCX DEBUG added async handler 0x2abc960 [id=82 ref 1] uct_tcp_iface_connect_handler() to hash
[1650465630.373967] [ndv4:54717:0] async.c:506 UCX DEBUG listening to async event fd 82 events 0x5 mode thread_spinlock
[1650465630.373971] [ndv4:54717:0] tcp_iface.c:495 UCX DEBUG tcp_iface 0x2b48530: listening for connections (fd=82) on 172.16.1.242:39096
[1650465630.374283] [ndv4:54717:0] ucp_worker.c:1159 UCX DEBUG created interface[5]=0x2b48530 using tcp/ib0 on worker 0x34258d0
[1650465630.374435] [ndv4:54717:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465630.374442] [ndv4:54717:0] 92 hdr_ofs 90 data_sz 8256
[1650465630.302547] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.302550] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.302654] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.303028] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465630.303035] [ndv4:54756:0] 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> }
[1650465630.303067] [ndv4:54756:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib6 length=2048) failed: Invalid argument
[1650465630.303071] [ndv4:54756:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.373565] [ndv4:54756:0] async.c:228 UCX DEBUG added async handler 0x5955360 [id=129 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.373666] [ndv4:54756:0] async.c:506 UCX DEBUG listening to async event fd 129 events 0x1 mode thread_spinlock
[1650465630.374474] [ndv4:54756:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
55512:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.306224] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.306271] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.306682] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.306688] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465630.307259] [ndv4:55512:0] ib_iface.c:994 UCX DEBUG iface=0x468b0a0: created RC QP 0xda7c on mlx5_ib5:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465630.307910] [ndv4:55512:0] ucp_worker.c:1159 UCX DEBUG created interface[31]=0x468b0a0 using rc_verbs/mlx5_ib5:1 on worker 0x22c98d0
[1650465630.322761] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.372951] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.373094] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.373099] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.373313] [ndv4:55512:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib5:1
[1650465630.375064] [ndv4:55512:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.375082] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.375085] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.375137] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.375680] [ndv4:55512:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x49ac010 of 8176 bytes with 127 elements
[1650465630.376114] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.376120] [ndv4:55512:0] 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> }
[1650465630.376151] [ndv4:55512:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib5 length=2048) failed: Invalid argument
[1650465630.376155] [ndv4:55512:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.376166] [ndv4:55512:0] async.c:228 UCX DEBUG added async handler 0x4666fc0 [id=120 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.376187] [ndv4:55512:0] async.c:506 UCX DEBUG listening to async event fd 120 events 0x1 mode thread_spinlock
[1650465630.376197] [ndv4:55512:0] ucp_worker.c:1159 UCX DEBUG created interface[32]=0x47d8030 using rc_mlx5/mlx5_ib5:1 on worker 0x22c98d0
[1650465630.376505] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.376512] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.376811] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650465630.376818] [ndv4:55512:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650465630.373436] [ndv4:54715:0] async.c:228 UCX DEBUG added async handler 0x200cf40 [id=69 ref 1] uct_ib_async_event_handler() to hash
[1650465630.373461] [ndv4:54715:0] async.c:506 UCX DEBUG listening to async event fd 69 events 0x1 mode thread_spinlock
[1650465630.373464] [ndv4:54715:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib6' (InfiniBand channel adapter) with 1 ports
[1650465630.373970] [ndv4:54715:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.373980] [ndv4:54715:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.373986] [ndv4:54715:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.374033] [ndv4:54715:0] ib_md.c:1319 UCX DEBUG mlx5_ib6: using registration cache
[1650465630.375896] [ndv4:54715:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.375903] [ndv4:54715:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.376100] [ndv4:54715:0] ib_md.c:1604 UCX DEBUG mlx5_ib6: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.376387] [ndv4:54715:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.376394] [ndv4:54715:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
ated interface[10]=0x2b469f0 using ud_mlx5/mlx5_ib0:1 on worker 0x33ff770
[1650465630.285945] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.285953] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.286041] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.286045] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.287279] [ndv4:55691:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465630.288807] [ndv4:55691:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465630.288826] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.288830] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.288861] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.289354] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.373599] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465630.375100] [ndv4:55691:0] ib_iface.c:994 UCX DEBUG iface=0x3cd0050: created RC QP 0xdaa9 on mlx5_ib1:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465630.376187] [ndv4:55691:0] ucp_worker.c:1159 UCX DEBUG created interface[11]=0x3cd0050 using rc_verbs/mlx5_ib1:1 on worker 0x33ff770
[1650465630.376434] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.376442] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.376653] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.376661] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.386289] [ndv4:55691:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465630.387546] [ndv4:55691:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.387556] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.387559] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.387630] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.387950] [ndv4:55691:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x3ce3010 of 8176 bytes with 127 elements
[1650465630.388129] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.388136] [ndv4:55691:0] 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> }
[1650465630.388171] [ndv4:55691:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650465630.388175] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.388185] [ndv4:55691:0] async.c:228 UCX DEBUG added async handler 0x2b29590 [id=92 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.388213] [ndv4:55691:0] async.c:506 UCX DEBUG listening to async event fd 92 events 0x1 mode thread_spinlock
[1650465630.388223] [ndv4:55691:0] ucp_worker.c:1159 UCX DEBUG created interface[12]=0x3b51010 using rc_mlx5/mlx5_ib1:1 on worker 0x33ff770
[1650465630.388886] [ndv4:55691:0] ib_md.c:296 Ul rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.309378] [ndv4:55152:0] async.c:228 UCX DEBUG added async handler 0x61aec40 [id=127 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.309399] [ndv4:55152:0] async.c:506 UCX DEBUG listening to async event fd 127 events 0x1 mode thread_spinlock
[1650465630.309410] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[37]=0x61a6030 using rc_mlx5/mlx5_ib6:1 on worker 0x357a8c0
[1650465630.309561] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.309567] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.309696] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.309702] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.310478] [ndv4:55152:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib6:1
[1650465630.312188] [ndv4:55152:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.312203] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.373469] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.373542] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.374267] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465630.374275] [ndv4:55152:0] 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> }
[1650465630.374311] [ndv4:55152:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib6 length=2048) failed: Invalid argument
[1650465630.374316] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.374324] [ndv4:55152:0] async.c:228 UCX DEBUG added async handler 0x6034360 [id=129 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.374347] [ndv4:55152:0] async.c:506 UCX DEBUG listening to async event fd 129 events 0x1 mode thread_spinlock
[1650465630.374666] [ndv4:55152:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465630.383207] [ndv4:55152:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x637c050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xdaf0
[1650465630.383679] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.383690] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.383701] [ndv4:55152:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b3b97fcc008 of 151544 bytes with 1052 elements
[1650465630.387840] [ndv4:55152:0] ib_md.c:812 UCX DEBUG registered memory 0x2b3bab200000..0x2b3bad800000 on mlx5_ib6 lkey 0x81100 rkey 0x81100 access 0xf flags 0x3e4
[1650465630.387860] [ndv4:55152:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b3bab200018 of 39845864 bytes with 4752 elements
[1650465630.388011] [ndv4:55152:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x637c050
[1650465630.388049] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[38]=0x637c050 using dc_mlx5/mlx5_ib6:1 on worker 0x357a8c0
[1650465630.388524] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.388536] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.389517] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.389524] [ndv4:55152:0] ib_md.c:296 UCX DEBUG[1650465630.405356] [ndv4:55674:0] debug.c:1198 UCX DEBUG using signal stack 0x2b83735fe000 size 141824
[1650465630.405433] [ndv4:55674:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465630.405455] [ndv4:55674:0] 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 0x2b8373452000
[1650465630.405478] [ndv4:55674:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650465630.405488] [ndv4:55674:0] 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
[1650465630.405495] [ndv4:55674:0] module.c:253 UCX DEBUG loading modules for ucs
[1650465630.408192] [ndv4:55674:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465630.408211] [ndv4:55674:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650465630.408244] [ndv4:55674:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650465630.408247] [ndv4:55674:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650465630.408254] [ndv4:55674:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650465630.408261] [ndv4:55674:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650465630.408264] [ndv4:55674:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650465630.408269] [ndv4:55674:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650465630.408271] [ndv4:55674:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650465630.408273] [ndv4:55674:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650465630.408276] [ndv4:55674:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650465630.408278] [ndv4:55674:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650465630.408286] [ndv4:55674:0] module.c:253 UCX DEBUG loading modules for uct
[1650465630.409011] [ndv4:55674:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650465630.410117] [ndv4:55674:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650465630.410132] [ndv4:55674:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650465630.410143] [ndv4:55674:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650465630.410179] [ndv4:55674:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650465630.410190] [ndv4:55674:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650465630.410201] [ndv4:55674:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650465630.410210] [ndv4:55674:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650465630.411539] [ndv4:55674:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650465630.431799] [ndv4:55674:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650465630.432297] [ndv4:55674:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650465630.437835] [ndv4:55674:0] async.c:228 UCX DEBUG added async handler 0x1d3a2c0 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650465630.437944] [ndv4:55674:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650465630.437953] [ndv4:55674:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650465630.438209] [[1650465630.455310] [ndv4:55796:0] debug.c:1198 UCX DEBUG using signal stack 0x2b1095d2f000 size 141824
[1650465630.455394] [ndv4:55796:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465630.455417] [ndv4:55796:0] 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 0x2b1095b9d000
[1650465630.455444] [ndv4:55796:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650465630.455453] [ndv4:55796:0] 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
[1650465630.455460] [ndv4:55796:0] module.c:253 UCX DEBUG loading modules for ucs
[1650465630.458173] [ndv4:55796:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650465630.458193] [ndv4:55796:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650465630.458232] [ndv4:55796:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650465630.458235] [ndv4:55796:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650465630.458242] [ndv4:55796:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650465630.458249] [ndv4:55796:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650465630.458251] [ndv4:55796:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650465630.458256] [ndv4:55796:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650465630.458259] [ndv4:55796:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650465630.458261] [ndv4:55796:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650465630.458263] [ndv4:55796:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650465630.458265] [ndv4:55796:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650465630.458273] [ndv4:55796:0] module.c:253 UCX DEBUG loading modules for uct
[1650465630.459040] [ndv4:55796:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650465630.460164] [ndv4:55796:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650465630.460179] [ndv4:55796:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650465630.460192] [ndv4:55796:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650465630.460232] [ndv4:55796:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650465630.460243] [ndv4:55796:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650465630.460253] [ndv4:55796:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650465630.460264] [ndv4:55796:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650465630.460710] [ndv4:55796:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650465630.467802] [ndv4:55796:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650465630.468133] [ndv4:55796:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.390643] [ndv4:55152:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib6:1
[1650465630.392437] [ndv4:55152:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.393969] [ndv4:55152:0] ib_iface.c:994 UCX DEBUG iface=0x6558340: created UD QP 0xdad8 on mlx5_ib6:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465630.394662] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650465630.394699] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.394705] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.394724] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.394729] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.395084] [ndv4:55152:0] ib_md.c:812 UCX DEBUG registered memory 0x2b3bad807000..0x2b3bad88c000 on mlx5_ib6 lkey 0x81200 rkey 0x81200 access 0xf flags 0x3e4
[1650465630.395090] [ndv4:55152:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b3bad807018 of 544744 bytes with 128 elements
[1650465630.395095] [ndv4:55152:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650465630.395542] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x6558340: adding gid fe80::15:5dff:fd34:1 to hash on device mlx5_ib6 port 1 index 0)
[1650465630.407754] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x6558340: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 1)
[1650465630.407796] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x6558340: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 2)
[1650465630.407811] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x6558340: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 3)
[1650465630.408169] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x6558340: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 4)
[1650465630.421657] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x6558340: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 5)
[1650465630.421997] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x6558340: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 6)
[1650465630.422103] [ndv4:55152:0] ud_iface.c:393 UCX DEBUG iface 0x6558340: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 7)
[1650465630.422402] [ndv4:55152:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650465630.422412] [ndv4:55152:0] async.c:228 UCX DEBUG added async handler 0x6558e90 [id=130 ref 1] uct_ud_iface_async_handler() to hash
[1650465630.422448] [ndv4:55152:0] async.c:506 UCX DEBUG listening to async event fd 130 events 0x5 mode thread_spinlock
[1650465630.422463] [ndv4:55152:0] ucp_worker.c:1159 UCX DEBUG created interface[39]=0x6558340 using ud_verbs/mlx5_ib6:1 on worker 0x357a8c0
[1650465630.422520] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.422527] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.422740] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.422746] [ndv4:55152:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.423395] [ndv4:55152:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib6:1
[1650465630.424731] [ndv4:55152:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.425025] [ndv4:55152:0] ib_iface.c:994 UCX DEBUG iface=0x6676460: created UD QP 0xdadb on mlx5_ibCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.388893] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.400818] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.400829] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.401979] [ndv4:55691:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465630.403845] [ndv4:55691:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.403862] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.403866] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.403949] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.404289] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650465630.404296] [ndv4:55691:0] 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> }
[1650465630.404322] [ndv4:55691:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650465630.404326] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.404334] [ndv4:55691:0] async.c:228 UCX DEBUG added async handler 0x2b2cf20 [id=94 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.404354] [ndv4:55691:0] async.c:506 UCX DEBUG listening to async event fd 94 events 0x1 mode thread_spinlock
[1650465630.404663] [ndv4:55691:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650465630.410382] [ndv4:55691:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x3e7b010: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0xdadd
[1650465630.410836] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.410844] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.410853] [ndv4:55691:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b871fe31008 of 151544 bytes with 1052 elements
[1650465630.414670] [ndv4:55691:0] ib_md.c:812 UCX DEBUG registered memory 0x2b8726e00000..0x2b8729400000 on mlx5_ib1 lkey 0x81500 rkey 0x81500 access 0xf flags 0x3e4
[1650465630.414688] [ndv4:55691:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b8726e00018 of 39845864 bytes with 4752 elements
[1650465630.414860] [ndv4:55691:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x3e7b010
[1650465630.414889] [ndv4:55691:0] ucp_worker.c:1159 UCX DEBUG created interface[13]=0x3e7b010 using dc_mlx5/mlx5_ib1:1 on worker 0x33ff770
[1650465630.415444] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.415451] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.415801] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.415807] [ndv4:55691:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.416006] [ndv4:55691:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465630.418121] [ndv4:55691:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650465630.418456] [ndv4:55691:0] ib_iface.c:994 UCX DEBUG iface=0x3d7a380: created UD QP 0xdabe on mlx5_ib1:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650465630.418910] [ndv4:55691:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64,[1650465630.373905] [ndv4:55432:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650465630.373918] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.373923] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.373944] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.374461] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.374466] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650465630.375101] [ndv4:55432:0] ib_iface.c:994 UCX DEBUG iface=0x3a49050: created RC QP 0xdaaa on mlx5_ib1:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650465630.376196] [ndv4:55432:0] ucp_worker.c:1159 UCX DEBUG created interface[11]=0x3a49050 using rc_verbs/mlx5_ib1:1 on worker 0x3178950
[1650465630.376218] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.376224] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.376469] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.376475] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.377078] [ndv4:55432:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465630.380177] [ndv4:55432:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650465630.380198] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650465630.380203] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650465630.380228] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650465630.381817] [ndv4:55432:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x3a5c010 of 8176 bytes with 127 elements
[1650465630.382162] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650465630.382171] [ndv4:55432:0] 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> }
[1650465630.382245] [ndv4:55432:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650465630.382249] [ndv4:55432:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650465630.382262] [ndv4:55432:0] async.c:228 UCX DEBUG added async handler 0x2128910 [id=92 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650465630.382287] [ndv4:55432:0] async.c:506 UCX DEBUG listening to async event fd 92 events 0x1 mode thread_spinlock
[1650465630.382309] [ndv4:55432:0] ucp_worker.c:1159 UCX DEBUG created interface[12]=0x38ca010 using rc_mlx5/mlx5_ib1:1 on worker 0x3178950
[1650465630.382328] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.382334] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.382830] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.382838] [ndv4:55432:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.383007] [ndv4:55432:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8014 on mlx5_ib1:1
[1650465630.394096] [ndv4:55432:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90[1650465630.380129] [ndv4:55039:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650465630.380137] [ndv4:55039:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650465630.380425] [ndv4:55039:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.390503] [ndv4:55039:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib6 vendor_id: 0x15b3 device_id: 4124
[1650465630.390830] [ndv4:55039:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib6: disable ODP because it's not supported for DevX QP
[1650465630.391511] [ndv4:55039:0] async.c:228 UCX DEBUG added async handler 0x19d9560 [id=69 ref 1] uct_ib_async_event_handler() to hash
[1650465630.391537] [ndv4:55039:0] async.c:506 UCX DEBUG listening to async event fd 69 events 0x1 mode thread_spinlock
[1650465630.391541] [ndv4:55039:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib6' (InfiniBand channel adapter) with 1 ports
[1650465630.391984] [ndv4:55039:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.391994] [ndv4:55039:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.391999] [ndv4:55039:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.392048] [ndv4:55039:0] ib_md.c:1319 UCX DEBUG mlx5_ib6: using registration cache
[1650465630.397091] [ndv4:55039:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.397116] [ndv4:55039:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.397307] [ndv4:55039:0] ib_md.c:1604 UCX DEBUG mlx5_ib6: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.397474] [ndv4:55039:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650465630.397485] [ndv4:55039:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650465630.399974] [ndv4:55039:0] topo.c:99 UCX DEBUG bus id 0x107000000 doesn't exist. sys_dev = 6
[1650465630.399981] [ndv4:55039:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465630.400693] [ndv4:55039:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465630.400698] [ndv4:55039:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465630.425671] [ndv4:55039:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465630.425679] [ndv4:55039:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465630.427312] [ndv4:55039:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465630.427318] [ndv4:55039:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465630.428613] [ndv4:55039:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465630.428617] [ndv4:55039:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465630.428953] [ndv4:55039:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.435239] [ndv4:55039:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib7 vendor_id: 0x15b3 device_id: 4124
[1650465630.435568] [ndv4:55039:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib7: disable ODP because it's not supported for DevX QP
[1650465630.435875] [ndv4:55039:0] async.c:228 UCX DEBUG added async handler 0x20bf940 [id=71 ref 1] uct_ib_async_event_handler() to hash
[1650465630.435905] [ndv4:55039:0] async.c:506 UCX DEBUG listening to async event fd 71 events 0x1 mode thread_spinlock
[1650465630.435909] [ndv4:55039:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib7' (InfiniBand channel adapter) with 1 ports
[1650465630.436058] [ndv4:55039:0] ib_md.c:296 UCX D[1650465630.382124] [ndv4:54724:0] async.c:228 UCX DEBUG added async handler 0x27001f0 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650465630.382246] [ndv4:54724:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650465630.382255] [ndv4:54724:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650465630.382767] [ndv4:54724:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465630.382777] [ndv4:54724:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465630.382805] [ndv4:54724:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.382870] [ndv4:54724:0] module.c:253 UCX DEBUG loading modules for ucm
[1650465630.382892] [ndv4:54724:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650465630.385265] [ndv4:54724:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.385274] [ndv4:54724:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.385530] [ndv4:54724:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.385891] [ndv4:54724:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650465630.385898] [ndv4:54724:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650465630.386007] [ndv4:54724:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650465630.386011] [ndv4:54724:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465630.390822] [ndv4:54724:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465630.390828] [ndv4:54724:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465630.391210] [ndv4:54724:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465630.391214] [ndv4:54724:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465630.418782] [ndv4:54724:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465630.418797] [ndv4:54724:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465630.432140] [ndv4:54724:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650465630.432160] [ndv4:54724:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650465630.432523] [ndv4:54724:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.441107] [ndv4:54724:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650465630.441448] [ndv4:54724:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650465630.441850] [ndv4:54724:0] async.c:228 UCX DEBUG added async handler 0x26f8f60 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650465630.441878] [ndv4:54724:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650465630.441882] [ndv4:54724:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650465630.442061] [ndv4:54724:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650465630.442070] [ndv4:54724:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650465630.442077] [ndv4:54724:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.442132] [ndv4:54724:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650465630.451865] [ndv4:54724:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double>[1650465630.384949] [ndv4:54861:0] async.c:228 UCX DEBUG added async handler 0x1fa5cd0 [id=71 ref 1] uct_ib_async_event_handler() to hash
[1650465630.384977] [ndv4:54861:0] async.c:506 UCX DEBUG listening to async event fd 71 events 0x1 mode thread_spinlock
[1650465630.384981] [ndv4:54861:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib7' (InfiniBand channel adapter) with 1 ports
[1650465630.385225] [ndv4:54861:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465630.385234] [ndv4:54861:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465630.385239] [ndv4:54861:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650465630.385290] [ndv4:54861:0] ib_md.c:1319 UCX DEBUG mlx5_ib7: using registration cache
[1650465630.388783] [ndv4:54861:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650465630.388790] [ndv4:54861:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650465630.389035] [ndv4:54861:0] ib_md.c:1604 UCX DEBUG mlx5_ib7: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650465630.389170] [ndv4:54861:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650465630.389176] [ndv4:54861:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650465630.404254] [ndv4:54861:0] topo.c:99 UCX DEBUG bus id 0x108000000 doesn't exist. sys_dev = 7
[1650465630.404268] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465630.405449] [ndv4:54861:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465630.405453] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465630.416878] [ndv4:54861:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465630.416885] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465630.418324] [ndv4:54861:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465630.418328] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465630.418476] [ndv4:54861:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650465630.418480] [ndv4:54861:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650465630.418657] [ndv4:54861:0] ucp_context.c:1556 UCX DEBUG created ucp context 0x18a8dc0 0x18a8dc0 [13 mds 47 tls] features 0x1 tl bitmap 0x7fffffffffff 0x0
[1650465630.385879] [ndv4:54715:0] topo.c:99 UCX DEBUG bus id 0x107000000 doesn't exist. sys_dev = 6
[1650465630.385887] [ndv4:54715:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465630.388464] [ndv4:54715:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465630.388470] [ndv4:54715:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465630.403732] [ndv4:54715:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465630.403741] [ndv4:54715:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465630.417243] [ndv4:54715:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465630.417269] [ndv4:54715:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465630.419181] [ndv4:54715:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650465630.419186] [ndv4:54715:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650465630.419522] [ndv4:54715:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650465630.461300] [ndv4:54715:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib7 vendor_id: 0x15b3 device_id: 4124
[1650465630.461778] [ndv4:54715:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib7: disable ODP because it's not supported for DevX QP
[1650465630.462498] [ndv4:54715:0] async.c:228 UCX DEBUG added async handler 0x200e820 [id=71 ref 1] uct_ib_async_event_handler() to hash
[1650465630.462529] [ndv4:54715:0] async.c:506 UCX
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment