Skip to content

Instantly share code, notes, and snippets.

@vanzod
Created April 22, 2022 15:01
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/ce6cfc5b823bfe5d71f4e1c8097a1e43 to your computer and use it in GitHub Desktop.
Save vanzod/ce6cfc5b823bfe5d71f4e1c8097a1e43 to your computer and use it in GitHub Desktop.
UCX 1.11.2 debug log for failing osu_scatter with forced sysV
This file has been truncated, but you can view the full file.
[1650638541.466186] [ndv4:68756:0] debug.c:1198 UCX DEBUG using signal stack 0x2b1397c2a000 size 141824
[1650638541.466330] [ndv4:68756:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650638541.466349] [ndv4:68756:0] 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 0x2b1397a98000
[1650638541.466373] [ndv4:68756:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650638541.466382] [ndv4:68756:0] 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
[1650638541.466389] [ndv4:68756:0] module.c:253 UCX DEBUG loading modules for ucs
[1650638541.469621] [ndv4:68756:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650638541.469653] [ndv4:68756:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650638541.469692] [ndv4:68756:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650638541.469695] [ndv4:68756:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650638541.469704] [ndv4:68756:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650638541.469711] [ndv4:68756:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650638541.469714] [ndv4:68756:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650638541.469720] [ndv4:68756:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650638541.469722] [ndv4:68756:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650638541.469725] [ndv4:68756:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650638541.469727] [ndv4:68756:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650638541.469730] [ndv4:68756:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650638541.469740] [ndv4:68756:0] module.c:253 UCX DEBUG loading modules for uct
[1650638541.478932] [ndv4:68756:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650638541.479298] [ndv4:68756:0] ucp_context.c:1117 UCX DEBUG closing md posix because it has no selected transport resources
[1650638541.479321] [ndv4:68756:0] ucp_context.c:1117 UCX DEBUG closing md self because it has no selected transport resources
[1650638541.479553] [ndv4:68756:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650638541.479564] [ndv4:68756:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650638541.479575] [ndv4:68756:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650638541.479585] [ndv4:68756:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650638541.479594] [ndv4:68756:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650638541.479604] [ndv4:68756:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650638541.479615] [ndv4:68756:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650638541.479655] [ndv4:68756:0] ucp_context.c:1117 UCX DEBUG closing md tcp because it has no selected transport resources
[1650638541.480021] [ndv4:68756:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650638541.487568] [ndv4:68756:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650638541.487935] [ndv4:68756:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650638541.536026] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0xccc470 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650638541.536156] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650638541.536738] [ndv4:68756:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650638541.541945] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638541.541969] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638541.542003] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638541.542084] [ndv4:68756:0] module.c:253 UCX DEBUG loading modules for ucm
[1650638541.542112] [ndv4:68756:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650638541.543539] [ndv4:68756:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638541.543552] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638541.543806] [ndv4:68756:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638541.543871] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638541.543878] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638541.546928] [ndv4:68756:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650638541.546935] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638541.549626] [ndv4:68756:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638541.549632] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638541.552269] [ndv4:68756:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638541.552275] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638541.554560] [ndv4:68756:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638541.554566] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638541.557493] [ndv4:68756:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638541.557499] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638541.557835] [ndv4:68756:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638541.569796] [ndv4:68756:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650638541.570313] [ndv4:68756:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650638541.571539] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0xca5040 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650638541.571551] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650638541.571554] [ndv4:68756:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650638541.572075] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638541.572084] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638541.572089] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638541.572131] [ndv4:68756:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650638541.573356] [ndv4:68756:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638541.573361] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638541.573717] [ndv4:68756:0] ib_md.c:1604 UCX DEBUG mlx5_ib1: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638541.573892] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638541.573897] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638541.576205] [ndv4:68756:0] topo.c:99 UCX DEBUG bus id 0x102000000 doesn't exist. sys_dev = 1
[1650638541.576264] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638541.579826] [ndv4:68756:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638541.579833] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638541.587690] [ndv4:68756:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638541.587697] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638541.591428] [ndv4:68756:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638541.591434] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638541.597662] [ndv4:68756:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638541.597669] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638541.597898] [ndv4:68756:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638541.606792] [ndv4:68756:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib2 vendor_id: 0x15b3 device_id: 4124
[1650638541.607351] [ndv4:68756:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib2: disable ODP because it's not supported for DevX QP
[1650638541.610325] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0xcc4c50 [id=61 ref 1] uct_ib_async_event_handler() to hash
[1650638541.610337] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 61 events 0x1 mode thread_spinlock
[1650638541.610341] [ndv4:68756:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib2' (InfiniBand channel adapter) with 1 ports
[1650638541.610515] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638541.610523] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638541.610529] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638541.610572] [ndv4:68756:0] ib_md.c:1319 UCX DEBUG mlx5_ib2: using registration cache
[1650638541.614064] [ndv4:68756:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638541.614079] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638541.614531] [ndv4:68756:0] ib_md.c:1604 UCX DEBUG mlx5_ib2: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638541.615058] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638541.615066] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638541.619588] [ndv4:68756:0] topo.c:99 UCX DEBUG bus id 0x103000000 doesn't exist. sys_dev = 2
[1650638541.619596] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638541.623172] [ndv4:68756:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638541.623185] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638541.626271] [ndv4:68756:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638541.626277] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638541.628544] [ndv4:68756:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638541.628551] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638541.632437] [ndv4:68756:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638541.632444] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638541.632705] [ndv4:68756:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638541.639879] [ndv4:68756:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib3 vendor_id: 0x15b3 device_id: 4124
[1650638541.640401] [ndv4:68756:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib3: disable ODP because it's not supported for DevX QP
[1650638541.641545] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0xcc4b40 [id=63 ref 1] uct_ib_async_event_handler() to hash
[1650638541.641584] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 63 events 0x1 mode thread_spinlock
[1650638541.641589] [ndv4:68756:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib3' (InfiniBand channel adapter) with 1 ports
[1650638541.641892] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638541.641903] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638541.641910] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638541.641959] [ndv4:68756:0] ib_md.c:1319 UCX DEBUG mlx5_ib3: using registration cache
[1650638541.643679] [ndv4:68756:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638541.643688] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638541.643935] [ndv4:68756:0] ib_md.c:1604 UCX DEBUG mlx5_ib3: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638541.644041] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638541.644047] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638541.647292] [ndv4:68756:0] topo.c:99 UCX DEBUG bus id 0x104000000 doesn't exist. sys_dev = 3
[1650638541.647300] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638541.649251] [ndv4:68756:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638541.649257] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638541.650435] [ndv4:68756:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638541.650442] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638541.653604] [ndv4:68756:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638541.653611] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638541.656310] [ndv4:68756:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638541.656317] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638541.656596] [ndv4:68756:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638541.665110] [ndv4:68756:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib4 vendor_id: 0x15b3 device_id: 4124
[1650638541.665545] [ndv4:68756:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib4: disable ODP because it's not supported for DevX QP
[1650638541.668526] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x13ae360 [id=65 ref 1] uct_ib_async_event_handler() to hash
[1650638541.668556] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 65 events 0x1 mode thread_spinlock
[1650638541.668560] [ndv4:68756:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib4' (InfiniBand channel adapter) with 1 ports
[1650638541.668790] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638541.668799] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638541.668805] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638541.668847] [ndv4:68756:0] ib_md.c:1319 UCX DEBUG mlx5_ib4: using registration cache
[1650638541.670164] [ndv4:68756:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638541.670172] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638541.670515] [ndv4:68756:0] ib_md.c:1604 UCX DEBUG mlx5_ib4: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638541.670830] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638541.670837] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638541.672512] [ndv4:68756:0] topo.c:99 UCX DEBUG bus id 0x105000000 doesn't exist. sys_dev = 4
[1650638541.672519] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638541.674099] [ndv4:68756:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650638541.674106] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638541.677269] [ndv4:68756:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650638541.677275] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638541.687049] [ndv4:68756:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650638541.687057] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638541.689678] [ndv4:68756:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650638541.689686] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638541.689926] [ndv4:68756:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638541.699545] [ndv4:68756:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib5 vendor_id: 0x15b3 device_id: 4124
[1650638541.699921] [ndv4:68756:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib5: disable ODP because it's not supported for DevX QP
[1650638541.702288] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0xcc4880 [id=67 ref 1] uct_ib_async_event_handler() to hash
[1650638541.702327] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 67 events 0x1 mode thread_spinlock
[1650638541.702331] [ndv4:68756:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib5' (InfiniBand channel adapter) with 1 ports
[1650638541.702773] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638541.702783] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638541.702789] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638541.702837] [ndv4:68756:0] ib_md.c:1319 UCX DEBUG mlx5_ib5: using registration cache
[1650638541.704709] [ndv4:68756:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638541.704717] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638541.704995] [ndv4:68756:0] ib_md.c:1604 UCX DEBUG mlx5_ib5: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638541.705010] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638541.705015] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638541.709012] [ndv4:68756:0] topo.c:99 UCX DEBUG bus id 0x106000000 doesn't exist. sys_dev = 5
[1650638541.709020] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638541.713662] [ndv4:68756:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650638541.713670] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638541.718645] [ndv4:68756:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650638541.718652] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638541.723005] [ndv4:68756:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650638541.723012] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638541.726838] [ndv4:68756:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650638541.726859] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638541.727132] [ndv4:68756:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638541.740267] [ndv4:68756:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib6 vendor_id: 0x15b3 device_id: 4124
[1650638541.740712] [ndv4:68756:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib6: disable ODP because it's not supported for DevX QP
[1650638541.742973] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0xcc61a0 [id=69 ref 1] uct_ib_async_event_handler() to hash
[1650638541.743018] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 69 events 0x1 mode thread_spinlock
[1650638541.743023] [ndv4:68756:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib6' (InfiniBand channel adapter) with 1 ports
[1650638541.743677] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638541.743691] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638541.743698] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638541.743757] [ndv4:68756:0] ib_md.c:1319 UCX DEBUG mlx5_ib6: using registration cache
[1650638541.744879] [ndv4:68756:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638541.744888] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638541.745484] [ndv4:68756:0] ib_md.c:1604 UCX DEBUG mlx5_ib6: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638541.745587] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638541.745593] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638541.752159] [ndv4:68756:0] topo.c:99 UCX DEBUG bus id 0x107000000 doesn't exist. sys_dev = 6
[1650638541.752168] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650638541.757649] [ndv4:68756:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650638541.757656] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650638541.763532] [ndv4:68756:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650638541.763540] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650638541.769533] [ndv4:68756:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650638541.769540] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650638541.775539] [ndv4:68756:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650638541.775545] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650638541.775774] [ndv4:68756:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638541.785158] [ndv4:68756:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib7 vendor_id: 0x15b3 device_id: 4124
[1650638541.785545] [ndv4:68756:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib7: disable ODP because it's not supported for DevX QP
[1650638541.787775] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x13afa40 [id=71 ref 1] uct_ib_async_event_handler() to hash
[1650638541.787804] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 71 events 0x1 mode thread_spinlock
[1650638541.787807] [ndv4:68756:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib7' (InfiniBand channel adapter) with 1 ports
[1650638541.788637] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638541.788647] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638541.788652] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638541.788695] [ndv4:68756:0] ib_md.c:1319 UCX DEBUG mlx5_ib7: using registration cache
[1650638541.791159] [ndv4:68756:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638541.791167] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638541.791397] [ndv4:68756:0] ib_md.c:1604 UCX DEBUG mlx5_ib7: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638541.792002] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638541.792008] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638541.823645] [ndv4:68756:0] topo.c:99 UCX DEBUG bus id 0x108000000 doesn't exist. sys_dev = 7
[1650638541.823654] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650638541.827698] [ndv4:68756:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650638541.827705] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650638541.855391] [ndv4:68756:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650638541.855399] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650638541.857956] [ndv4:68756:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650638541.857961] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650638541.876086] [ndv4:68756:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650638541.876094] [ndv4:68756:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650638541.876186] [ndv4:68756:0] ucp_context.c:1556 UCX DEBUG created ucp context 0xcb1f10 0xcb1f10 [10 mds 42 tls] features 0x1 tl bitmap 0x3ffffffffff 0x0
[1650638541.972950] [ndv4:68756:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 8447 bytes with hugetlb
[1650638541.973057] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool mm_recv_desc: align 64, maxelems 4294967295, elemsize 8368
[1650638541.974408] [ndv4:68756:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 4292720 bytes with hugetlb
[1650638541.974423] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool mm_recv_desc: allocated chunk 0x2b139ef2f018 of 4296680 bytes with 512 elements
[1650638541.974998] [ndv4:68756:0] mm_iface.c:600 UCX DEBUG created mm iface 0x142e220 FIFO id 0xd002b va 0x2b1397ff4000 size 12288 (128 x 64 elems)
[1650638541.975050] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[0]=0x142e220 using sysv/memory on worker 0x1d084d0
[1650638541.981692] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638541.981704] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638541.981997] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638541.982002] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638541.993010] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib0:1
[1650638542.002683] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650638542.002720] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638542.002723] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638542.002776] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638542.003837] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638542.003845] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650638542.004354] [ndv4:68756:0] ib_iface.c:994 UCX DEBUG iface=0x1439430: created RC QP 0x2121 on mlx5_ib0:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650638542.006517] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[1]=0x1439430 using rc_verbs/mlx5_ib0:1 on worker 0x1d084d0
[1650638542.006705] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638542.006712] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638542.006928] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638542.006933] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638542.007510] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib0:1
[1650638542.007784] [ndv4:68756:0] ib_device.c:1394 UCX DEBUG max IB CQE size is 128
[1650638542.009142] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638542.009152] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638542.009155] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638542.009229] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638542.009752] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x202f010 of 8176 bytes with 127 elements
[1650638542.009972] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638542.009993] [ndv4:68756:0] 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> }
[1650638542.010046] [ndv4:68756:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650638542.010052] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638542.020121] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x139cb00 [id=78 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638542.020161] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 78 events 0x1 mode thread_spinlock
[1650638542.020804] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[2]=0x1446fb0 using rc_mlx5/mlx5_ib0:1 on worker 0x1d084d0
[1650638542.021162] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638542.021170] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638542.021847] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638542.021853] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638542.023118] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib0:1
[1650638542.036074] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638542.036087] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638542.036091] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638542.036109] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638542.036672] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650638542.036681] [ndv4:68756:0] 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> }
[1650638542.036726] [ndv4:68756:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650638542.036730] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638542.036739] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x139c860 [id=80 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638542.036763] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 80 events 0x1 mode thread_spinlock
[1650638542.037415] [ndv4:68756:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650638542.047046] [ndv4:68756:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x2031050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0x2144
[1650638542.047394] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638542.047400] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638542.047423] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b13a1b49008 of 151544 bytes with 1052 elements
[1650638542.051105] [ndv4:68756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b139f400000..0x2b13a1a00000 on mlx5_ib0 lkey 0x80800 rkey 0x80800 access 0xf flags 0x3e4
[1650638542.051131] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b139f400018 of 39845864 bytes with 4752 elements
[1650638542.051306] [ndv4:68756:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x2031050
[1650638542.051340] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[3]=0x2031050 using dc_mlx5/mlx5_ib0:1 on worker 0x1d084d0
[1650638542.051558] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638542.051567] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638542.051728] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638542.051733] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638542.052570] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib0:1
[1650638542.054059] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638542.054535] [ndv4:68756:0] ib_iface.c:994 UCX DEBUG iface=0x14445f0: created UD QP 0x212a on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638542.055559] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638542.055948] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638542.055955] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638542.056066] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638542.056071] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638542.056571] [ndv4:68756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b13a1b6e000..0x2b13a1bf3000 on mlx5_ib0 lkey 0x80900 rkey 0x80900 access 0xf flags 0x3e4
[1650638542.056577] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b13a1b6e018 of 544744 bytes with 128 elements
[1650638542.056581] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638542.057614] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x14445f0: adding gid fe80::15:5dff:fd34:1b to hash on device mlx5_ib0 port 1 index 0)
[1650638542.058552] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x14445f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650638542.059154] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x14445f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650638542.060177] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x14445f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650638542.061151] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x14445f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650638542.061204] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x14445f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650638542.061702] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x14445f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650638542.069592] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x14445f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650638542.069972] [ndv4:68756:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638542.079823] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x13a09c0 [id=81 ref 1] uct_ud_iface_async_handler() to hash
[1650638542.079867] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 81 events 0x5 mode thread_spinlock
[1650638542.080370] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[4]=0x14445f0 using ud_verbs/mlx5_ib0:1 on worker 0x1d084d0
[1650638542.080412] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638542.080420] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638542.080664] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638542.080669] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638542.081190] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib0:1
[1650638542.090539] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638542.090958] [ndv4:68756:0] ib_iface.c:994 UCX DEBUG iface=0x144f920: created UD QP 0x212b on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638542.090976] [ndv4:68756:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650638542.091780] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638542.092235] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638542.092241] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638542.092363] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638542.092367] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638542.092905] [ndv4:68756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b13a1bf3000..0x2b13a1c78000 on mlx5_ib0 lkey 0x80a00 rkey 0x80a00 access 0xf flags 0x3e4
[1650638542.092912] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b13a1bf3018 of 544744 bytes with 128 elements
[1650638542.092918] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638542.093508] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x144f920: adding gid fe80::15:5dff:fd34:1b to hash on device mlx5_ib0 port 1 index 0)
[1650638542.094465] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x144f920: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650638542.095625] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x144f920: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650638542.096290] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x144f920: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650638542.096879] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x144f920: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650638542.097867] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x144f920: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650638542.098806] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x144f920: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650638542.099084] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x144f920: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650638542.099092] [ndv4:68756:0] 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> }
[1650638542.099125] [ndv4:68756:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638542.099130] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x13a0680 [id=82 ref 1] uct_ud_iface_async_handler() to hash
[1650638542.099153] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 82 events 0x5 mode thread_spinlock
[1650638542.099166] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[5]=0x144f920 using ud_mlx5/mlx5_ib0:1 on worker 0x1d084d0
[1650638542.099535] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638542.099541] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638542.099808] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638542.099815] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638542.107732] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib1:1
[1650638542.116296] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650638542.116305] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638542.116308] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638542.116344] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638542.117347] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638542.117355] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650638542.117910] [ndv4:68756:0] ib_iface.c:994 UCX DEBUG iface=0x25d8050: created RC QP 0x2077 on mlx5_ib1:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650638542.118368] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[6]=0x25d8050 using rc_verbs/mlx5_ib1:1 on worker 0x1d084d0
[1650638542.118582] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638542.118588] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638542.118720] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638542.118725] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638542.119597] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib1:1
[1650638542.121629] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638542.121639] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638542.121642] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638542.121677] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638542.122160] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x2531010 of 8176 bytes with 127 elements
[1650638542.122456] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638542.122464] [ndv4:68756:0] 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> }
[1650638542.122503] [ndv4:68756:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650638542.122507] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638542.122520] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x139f910 [id=85 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638542.122548] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 85 events 0x1 mode thread_spinlock
[1650638542.122558] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[7]=0x245e020 using rc_mlx5/mlx5_ib1:1 on worker 0x1d084d0
[1650638542.122758] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638542.122764] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638542.122936] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638542.122941] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638542.123477] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib1:1
[1650638542.125956] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638542.125971] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638542.125974] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638542.126031] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638542.126673] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650638542.126680] [ndv4:68756:0] 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> }
[1650638542.126713] [ndv4:68756:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650638542.126716] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638542.126723] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0xccde70 [id=87 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638542.126753] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 87 events 0x1 mode thread_spinlock
[1650638542.127715] [ndv4:68756:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650638542.144166] [ndv4:68756:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x277d010: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0x209a
[1650638542.153058] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638542.153067] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638542.153077] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b13a4479008 of 151544 bytes with 1052 elements
[1650638542.157126] [ndv4:68756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b13a1e00000..0x2b13a4400000 on mlx5_ib1 lkey 0x80400 rkey 0x80400 access 0xf flags 0x3e4
[1650638542.157148] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b13a1e00018 of 39845864 bytes with 4752 elements
[1650638542.157304] [ndv4:68756:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x277d010
[1650638542.157340] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[8]=0x277d010 using dc_mlx5/mlx5_ib1:1 on worker 0x1d084d0
[1650638542.157856] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638542.157870] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638542.158043] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638542.158049] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638542.159813] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib1:1
[1650638542.161848] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638542.162183] [ndv4:68756:0] ib_iface.c:994 UCX DEBUG iface=0x1443710: created UD QP 0x20b9 on mlx5_ib1:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638542.162955] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638542.170450] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638542.170458] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638542.170568] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638542.170574] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638542.171376] [ndv4:68756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b13a449e000..0x2b13a4523000 on mlx5_ib1 lkey 0x80500 rkey 0x80500 access 0xf flags 0x3e4
[1650638542.171382] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b13a449e018 of 544744 bytes with 128 elements
[1650638542.171386] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638542.172154] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x1443710: adding gid fe80::15:5dff:fd34:1c to hash on device mlx5_ib1 port 1 index 0)
[1650638542.173326] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x1443710: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 1)
[1650638542.174087] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x1443710: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 2)
[1650638542.175003] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x1443710: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 3)
[1650638542.175694] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x1443710: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 4)
[1650638542.176419] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x1443710: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 5)
[1650638542.176824] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x1443710: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 6)
[1650638542.178000] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x1443710: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 7)
[1650638542.178569] [ndv4:68756:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638542.178577] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x1450fb0 [id=88 ref 1] uct_ud_iface_async_handler() to hash
[1650638542.178611] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 88 events 0x5 mode thread_spinlock
[1650638542.178620] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[9]=0x1443710 using ud_verbs/mlx5_ib1:1 on worker 0x1d084d0
[1650638542.178801] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638542.178819] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638542.178862] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638542.178866] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638542.179830] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib1:1
[1650638542.181637] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638542.182181] [ndv4:68756:0] ib_iface.c:994 UCX DEBUG iface=0x1d286c0: created UD QP 0x20da on mlx5_ib1:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638542.182188] [ndv4:68756:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650638542.183081] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638542.183494] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638542.183502] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638542.183528] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638542.183544] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638542.184151] [ndv4:68756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b13a4523000..0x2b13a45a8000 on mlx5_ib1 lkey 0x80600 rkey 0x80600 access 0xf flags 0x3e4
[1650638542.184158] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b13a4523018 of 544744 bytes with 128 elements
[1650638542.184162] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638542.184200] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x1d286c0: adding gid fe80::15:5dff:fd34:1c to hash on device mlx5_ib1 port 1 index 0)
[1650638542.194058] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x1d286c0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 1)
[1650638542.194681] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x1d286c0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 2)
[1650638542.194845] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x1d286c0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 3)
[1650638542.195150] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x1d286c0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 4)
[1650638542.195572] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x1d286c0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 5)
[1650638542.196443] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x1d286c0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 6)
[1650638542.196773] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x1d286c0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 7)
[1650638542.196780] [ndv4:68756:0] 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> }
[1650638542.196816] [ndv4:68756:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638542.196820] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x142fb90 [id=89 ref 1] uct_ud_iface_async_handler() to hash
[1650638542.196853] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 89 events 0x5 mode thread_spinlock
[1650638542.196866] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[10]=0x1d286c0 using ud_mlx5/mlx5_ib1:1 on worker 0x1d084d0
[1650638542.196878] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638542.196884] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638542.196943] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638542.196948] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638542.197798] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib2:1
[1650638542.199958] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650638542.199978] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638542.199982] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638542.200035] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638542.201310] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638542.201317] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650638542.201805] [ndv4:68756:0] ib_iface.c:994 UCX DEBUG iface=0x2b70020: created RC QP 0x20bd on mlx5_ib2:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650638542.202158] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[11]=0x2b70020 using rc_verbs/mlx5_ib2:1 on worker 0x1d084d0
[1650638542.202445] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638542.202456] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638542.202725] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638542.202739] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638542.203641] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib2:1
[1650638542.206445] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638542.206461] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638542.206464] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638542.206516] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638542.206986] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x2e91010 of 8176 bytes with 127 elements
[1650638542.207280] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638542.207286] [ndv4:68756:0] 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> }
[1650638542.207356] [ndv4:68756:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib2 length=2048) failed: Invalid argument
[1650638542.207360] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638542.207370] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x2466cc0 [id=92 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638542.207398] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 92 events 0x1 mode thread_spinlock
[1650638542.207409] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[12]=0x2cbd030 using rc_mlx5/mlx5_ib2:1 on worker 0x1d084d0
[1650638542.207470] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638542.207475] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638542.207658] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638542.207664] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638542.217072] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib2:1
[1650638542.219152] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638542.219170] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638542.219173] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638542.219252] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638542.219818] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650638542.219829] [ndv4:68756:0] 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> }
[1650638542.219883] [ndv4:68756:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib2 length=2048) failed: Invalid argument
[1650638542.219887] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638542.219897] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x2cc5f00 [id=94 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638542.219921] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 94 events 0x1 mode thread_spinlock
[1650638542.220651] [ndv4:68756:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650638542.229903] [ndv4:68756:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x2e93050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0x2100
[1650638542.237007] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638542.237030] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638542.237041] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b13a6da9008 of 151544 bytes with 1052 elements
[1650638542.241110] [ndv4:68756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b13a4600000..0x2b13a6c00000 on mlx5_ib2 lkey 0x80400 rkey 0x80400 access 0xf flags 0x3e4
[1650638542.241133] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b13a4600018 of 39845864 bytes with 4752 elements
[1650638542.241332] [ndv4:68756:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x2e93050
[1650638542.241363] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[13]=0x2e93050 using dc_mlx5/mlx5_ib2:1 on worker 0x1d084d0
[1650638542.241603] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638542.241620] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638542.241819] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638542.241824] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638542.248757] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib2:1
[1650638542.258538] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638542.258966] [ndv4:68756:0] ib_iface.c:994 UCX DEBUG iface=0x21fc500: created UD QP 0x20e7 on mlx5_ib2:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638542.259778] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638542.260071] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638542.260089] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638542.260605] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638542.260624] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638542.261322] [ndv4:68756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b13a6dce000..0x2b13a6e53000 on mlx5_ib2 lkey 0x80500 rkey 0x80500 access 0xf flags 0x3e4
[1650638542.261327] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b13a6dce018 of 544744 bytes with 128 elements
[1650638542.261332] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638542.262110] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x21fc500: adding gid fe80::15:5dff:fd34:1d to hash on device mlx5_ib2 port 1 index 0)
[1650638542.263028] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x21fc500: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 1)
[1650638542.264868] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x21fc500: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 2)
[1650638542.264896] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x21fc500: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 3)
[1650638542.272459] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x21fc500: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 4)
[1650638542.272809] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x21fc500: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 5)
[1650638542.273550] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x21fc500: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 6)
[1650638542.273909] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x21fc500: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 7)
[1650638542.274326] [ndv4:68756:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638542.274334] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x14371b0 [id=95 ref 1] uct_ud_iface_async_handler() to hash
[1650638542.274365] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 95 events 0x5 mode thread_spinlock
[1650638542.274379] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[14]=0x21fc500 using ud_verbs/mlx5_ib2:1 on worker 0x1d084d0
[1650638542.274700] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638542.274706] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638542.274772] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638542.274776] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638542.275444] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib2:1
[1650638542.277412] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638542.278022] [ndv4:68756:0] ib_iface.c:994 UCX DEBUG iface=0x142c5a0: created UD QP 0x20e8 on mlx5_ib2:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638542.278031] [ndv4:68756:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650638542.278771] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638542.286671] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638542.286707] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638542.286966] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638542.286971] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638542.287421] [ndv4:68756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b13a6e53000..0x2b13a6ed8000 on mlx5_ib2 lkey 0x80600 rkey 0x80600 access 0xf flags 0x3e4
[1650638542.287433] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b13a6e53018 of 544744 bytes with 128 elements
[1650638542.287438] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638542.288501] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x142c5a0: adding gid fe80::15:5dff:fd34:1d to hash on device mlx5_ib2 port 1 index 0)
[1650638542.294741] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x142c5a0: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 1)
[1650638542.295288] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x142c5a0: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 2)
[1650638542.296130] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x142c5a0: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 3)
[1650638542.296431] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x142c5a0: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 4)
[1650638542.296873] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x142c5a0: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 5)
[1650638542.297111] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x142c5a0: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 6)
[1650638542.297488] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x142c5a0: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 7)
[1650638542.297496] [ndv4:68756:0] 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> }
[1650638542.297533] [ndv4:68756:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638542.297538] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x2b78c40 [id=96 ref 1] uct_ud_iface_async_handler() to hash
[1650638542.297564] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 96 events 0x5 mode thread_spinlock
[1650638542.297578] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[15]=0x142c5a0 using ud_mlx5/mlx5_ib2:1 on worker 0x1d084d0
[1650638542.297747] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638542.297754] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638542.297922] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638542.297928] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638542.308064] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib3:1
[1650638542.318199] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650638542.318319] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638542.318324] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638542.318385] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638542.319437] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638542.319447] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650638542.319931] [ndv4:68756:0] ib_iface.c:994 UCX DEBUG iface=0x328d0a0: created RC QP 0x20cf on mlx5_ib3:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650638542.320328] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[16]=0x328d0a0 using rc_verbs/mlx5_ib3:1 on worker 0x1d084d0
[1650638542.320551] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638542.320557] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638542.320897] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638542.320905] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638542.321788] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib3:1
[1650638542.333137] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638542.333152] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638542.333155] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638542.333260] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638542.333945] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x35ae010 of 8176 bytes with 127 elements
[1650638542.334175] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638542.334184] [ndv4:68756:0] 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> }
[1650638542.334275] [ndv4:68756:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib3 length=2048) failed: Invalid argument
[1650638542.334279] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638542.334290] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x1451c70 [id=99 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638542.334324] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 99 events 0x1 mode thread_spinlock
[1650638542.334348] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[17]=0x33da030 using rc_mlx5/mlx5_ib3:1 on worker 0x1d084d0
[1650638542.334762] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638542.334769] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638542.335406] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638542.335414] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638542.335841] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib3:1
[1650638542.344977] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638542.344996] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638542.344999] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638542.345056] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638542.345519] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650638542.345531] [ndv4:68756:0] 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> }
[1650638542.345570] [ndv4:68756:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib3 length=2048) failed: Invalid argument
[1650638542.345574] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638542.345583] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x306faa0 [id=101 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638542.345611] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 101 events 0x1 mode thread_spinlock
[1650638542.346275] [ndv4:68756:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650638542.355552] [ndv4:68756:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x35b0050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0x20fa
[1650638542.355812] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638542.355819] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638542.355828] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b13a96d9008 of 151544 bytes with 1052 elements
[1650638542.359873] [ndv4:68756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b13a7000000..0x2b13a9600000 on mlx5_ib3 lkey 0x80400 rkey 0x80400 access 0xf flags 0x3e4
[1650638542.359895] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b13a7000018 of 39845864 bytes with 4752 elements
[1650638542.360034] [ndv4:68756:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x35b0050
[1650638542.360070] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[18]=0x35b0050 using dc_mlx5/mlx5_ib3:1 on worker 0x1d084d0
[1650638542.360507] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638542.360519] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638542.360765] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638542.360771] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638542.368445] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib3:1
[1650638542.379290] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638542.379636] [ndv4:68756:0] ib_iface.c:994 UCX DEBUG iface=0x378c060: created UD QP 0x2108 on mlx5_ib3:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638542.380339] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638542.381154] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638542.381162] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638542.381285] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638542.381291] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638542.381832] [ndv4:68756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b13a96fe000..0x2b13a9783000 on mlx5_ib3 lkey 0x80500 rkey 0x80500 access 0xf flags 0x3e4
[1650638542.381839] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b13a96fe018 of 544744 bytes with 128 elements
[1650638542.381844] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638542.382520] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x378c060: adding gid fe80::15:5dff:fd34:1e to hash on device mlx5_ib3 port 1 index 0)
[1650638542.403796] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x378c060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 1)
[1650638542.404704] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x378c060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 2)
[1650638542.406464] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x378c060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 3)
[1650638542.406642] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x378c060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 4)
[1650638542.407418] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x378c060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 5)
[1650638542.424963] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x378c060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 6)
[1650638542.425320] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x378c060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 7)
[1650638542.425710] [ndv4:68756:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638542.425720] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x33e2f50 [id=102 ref 1] uct_ud_iface_async_handler() to hash
[1650638542.425756] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 102 events 0x5 mode thread_spinlock
[1650638542.425767] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[19]=0x378c060 using ud_verbs/mlx5_ib3:1 on worker 0x1d084d0
[1650638542.425905] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638542.425912] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638542.426196] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638542.426201] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638542.427059] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib3:1
[1650638542.428322] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638542.428931] [ndv4:68756:0] ib_iface.c:994 UCX DEBUG iface=0x38aa050: created UD QP 0x2109 on mlx5_ib3:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638542.428941] [ndv4:68756:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650638542.429942] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638542.430091] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638542.430097] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638542.430193] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638542.430198] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638542.430846] [ndv4:68756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b13a9783000..0x2b13a9808000 on mlx5_ib3 lkey 0x80600 rkey 0x80600 access 0xf flags 0x3e4
[1650638542.430854] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b13a9783018 of 544744 bytes with 128 elements
[1650638542.430859] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638542.431464] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x38aa050: adding gid fe80::15:5dff:fd34:1e to hash on device mlx5_ib3 port 1 index 0)
[1650638542.440778] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x38aa050: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 1)
[1650638542.441295] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x38aa050: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 2)
[1650638542.442015] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x38aa050: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 3)
[1650638542.442300] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x38aa050: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 4)
[1650638542.443418] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x38aa050: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 5)
[1650638542.451493] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x38aa050: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 6)
[1650638542.466187] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x38aa050: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 7)
[1650638542.466197] [ndv4:68756:0] 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> }
[1650638542.466259] [ndv4:68756:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638542.466265] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x2959ad0 [id=103 ref 1] uct_ud_iface_async_handler() to hash
[1650638542.466300] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 103 events 0x5 mode thread_spinlock
[1650638542.466311] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[20]=0x38aa050 using ud_mlx5/mlx5_ib3:1 on worker 0x1d084d0
[1650638542.466721] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638542.466727] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638542.473973] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638542.473996] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638542.481427] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib4:1
[1650638542.483576] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650638542.483593] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638542.483597] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638542.483650] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638542.484688] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638542.484699] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650638542.485294] [ndv4:68756:0] ib_iface.c:994 UCX DEBUG iface=0x39aa0a0: created RC QP 0x2119 on mlx5_ib4:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650638542.485769] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[21]=0x39aa0a0 using rc_verbs/mlx5_ib4:1 on worker 0x1d084d0
[1650638542.486342] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638542.486348] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638542.492335] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638542.492346] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638542.493752] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib4:1
[1650638542.496120] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638542.496134] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638542.496138] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638542.496188] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638542.496966] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x3ccb010 of 8176 bytes with 127 elements
[1650638542.497243] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638542.497251] [ndv4:68756:0] 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> }
[1650638542.497297] [ndv4:68756:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib4 length=2048) failed: Invalid argument
[1650638542.497301] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638542.497312] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x33e2d20 [id=106 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638542.497349] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 106 events 0x1 mode thread_spinlock
[1650638542.497362] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[22]=0x3af7030 using rc_mlx5/mlx5_ib4:1 on worker 0x1d084d0
[1650638542.497376] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638542.497382] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638542.497707] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638542.497724] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638542.504953] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib4:1
[1650638542.506802] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638542.506825] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638542.506828] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638542.506885] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638542.507310] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650638542.507319] [ndv4:68756:0] 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> }
[1650638542.507358] [ndv4:68756:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib4 length=2048) failed: Invalid argument
[1650638542.507362] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638542.507373] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x3affe80 [id=108 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638542.507402] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 108 events 0x1 mode thread_spinlock
[1650638542.508021] [ndv4:68756:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650638542.518492] [ndv4:68756:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x3ccd050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0x213e
[1650638542.519702] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638542.519713] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638542.519723] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b13ac00a008 of 151544 bytes with 1052 elements
[1650638542.523876] [ndv4:68756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b13a9a00000..0x2b13ac000000 on mlx5_ib4 lkey 0x80400 rkey 0x80400 access 0xf flags 0x3e4
[1650638542.523901] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b13a9a00018 of 39845864 bytes with 4752 elements
[1650638542.524041] [ndv4:68756:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x3ccd050
[1650638542.524078] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[23]=0x3ccd050 using dc_mlx5/mlx5_ib4:1 on worker 0x1d084d0
[1650638542.524200] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638542.524267] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638542.524568] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638542.524573] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638542.525534] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib4:1
[1650638542.535811] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638542.536143] [ndv4:68756:0] ib_iface.c:994 UCX DEBUG iface=0x3ea9060: created UD QP 0x215c on mlx5_ib4:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638542.536856] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638542.545547] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638542.545556] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638542.545604] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638542.545617] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638542.546166] [ndv4:68756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b13ac02f000..0x2b13ac0b4000 on mlx5_ib4 lkey 0x80500 rkey 0x80500 access 0xf flags 0x3e4
[1650638542.546178] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b13ac02f018 of 544744 bytes with 128 elements
[1650638542.546183] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638542.547464] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x3ea9060: adding gid fe80::15:5dff:fd34:1f to hash on device mlx5_ib4 port 1 index 0)
[1650638542.548974] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x3ea9060: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 1)
[1650638542.549541] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x3ea9060: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 2)
[1650638542.550106] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x3ea9060: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 3)
[1650638542.550511] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x3ea9060: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 4)
[1650638542.551034] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x3ea9060: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 5)
[1650638542.551307] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x3ea9060: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 6)
[1650638542.551522] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x3ea9060: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 7)
[1650638542.551877] [ndv4:68756:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638542.551887] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x3985e10 [id=109 ref 1] uct_ud_iface_async_handler() to hash
[1650638542.551902] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 109 events 0x5 mode thread_spinlock
[1650638542.551914] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[24]=0x3ea9060 using ud_verbs/mlx5_ib4:1 on worker 0x1d084d0
[1650638542.552065] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638542.552072] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638542.552455] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638542.552462] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638542.552884] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib4:1
[1650638542.554348] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638542.554642] [ndv4:68756:0] ib_iface.c:994 UCX DEBUG iface=0x3fc7460: created UD QP 0x2176 on mlx5_ib4:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638542.554653] [ndv4:68756:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650638542.555513] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638542.565438] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638542.565450] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638542.565696] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638542.565701] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638542.567303] [ndv4:68756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b13ac0b4000..0x2b13ac139000 on mlx5_ib4 lkey 0x80600 rkey 0x80600 access 0xf flags 0x3e4
[1650638542.567312] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b13ac0b4018 of 544744 bytes with 128 elements
[1650638542.567317] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638542.568445] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x3fc7460: adding gid fe80::15:5dff:fd34:1f to hash on device mlx5_ib4 port 1 index 0)
[1650638542.576343] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x3fc7460: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 1)
[1650638542.577303] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x3fc7460: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 2)
[1650638542.578302] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x3fc7460: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 3)
[1650638542.591343] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x3fc7460: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 4)
[1650638542.591634] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x3fc7460: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 5)
[1650638542.592623] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x3fc7460: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 6)
[1650638542.602050] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x3fc7460: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 7)
[1650638542.602060] [ndv4:68756:0] 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> }
[1650638542.602100] [ndv4:68756:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638542.602105] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x3fc7fb0 [id=110 ref 1] uct_ud_iface_async_handler() to hash
[1650638542.602149] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 110 events 0x5 mode thread_spinlock
[1650638542.602159] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[25]=0x3fc7460 using ud_mlx5/mlx5_ib4:1 on worker 0x1d084d0
[1650638542.602438] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638542.602448] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638542.602571] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638542.602577] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638542.604056] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib5:1
[1650638542.606339] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650638542.606356] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638542.606359] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638542.606413] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638542.607407] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638542.607413] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650638542.607943] [ndv4:68756:0] ib_iface.c:994 UCX DEBUG iface=0x40c70a0: created RC QP 0x1ccb on mlx5_ib5:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650638542.608330] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[26]=0x40c70a0 using rc_verbs/mlx5_ib5:1 on worker 0x1d084d0
[1650638542.608493] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638542.608500] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638542.608671] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638542.608676] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638542.609009] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib5:1
[1650638542.610124] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638542.610138] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638542.610141] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638542.610189] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638542.610824] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x43e8010 of 8176 bytes with 127 elements
[1650638542.611094] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638542.611102] [ndv4:68756:0] 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> }
[1650638542.611137] [ndv4:68756:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib5 length=2048) failed: Invalid argument
[1650638542.611140] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638542.611151] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x40cff50 [id=113 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638542.611172] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 113 events 0x1 mode thread_spinlock
[1650638542.611180] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[27]=0x4214030 using rc_mlx5/mlx5_ib5:1 on worker 0x1d084d0
[1650638542.611336] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638542.611347] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638542.611521] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638542.611527] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638542.624650] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib5:1
[1650638542.626187] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638542.626205] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638542.626291] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638542.626348] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638542.626831] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650638542.626838] [ndv4:68756:0] 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> }
[1650638542.626872] [ndv4:68756:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib5 length=2048) failed: Invalid argument
[1650638542.626876] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638542.626883] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x43f6fc0 [id=115 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638542.626907] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 115 events 0x1 mode thread_spinlock
[1650638542.627550] [ndv4:68756:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650638542.638291] [ndv4:68756:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x43ea050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0x1cee
[1650638542.638705] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638542.638714] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638542.638726] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b13ae93c008 of 151544 bytes with 1052 elements
[1650638542.643153] [ndv4:68756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b13ac200000..0x2b13ae800000 on mlx5_ib5 lkey 0x80400 rkey 0x80400 access 0xf flags 0x3e4
[1650638542.643180] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b13ac200018 of 39845864 bytes with 4752 elements
[1650638542.643363] [ndv4:68756:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x43ea050
[1650638542.643397] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[28]=0x43ea050 using dc_mlx5/mlx5_ib5:1 on worker 0x1d084d0
[1650638542.643423] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638542.643435] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638542.643664] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638542.643669] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638542.644374] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib5:1
[1650638542.645720] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638542.646131] [ndv4:68756:0] ib_iface.c:994 UCX DEBUG iface=0x31ba3f0: created UD QP 0x1cd4 on mlx5_ib5:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638542.646862] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638542.646942] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638542.646950] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638542.647365] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638542.647382] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638542.647854] [ndv4:68756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b13ae961000..0x2b13ae9e6000 on mlx5_ib5 lkey 0x80500 rkey 0x80500 access 0xf flags 0x3e4
[1650638542.647861] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b13ae961018 of 544744 bytes with 128 elements
[1650638542.647866] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638542.647913] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x31ba3f0: adding gid fe80::15:5dff:fd34:20 to hash on device mlx5_ib5 port 1 index 0)
[1650638542.648755] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x31ba3f0: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 1)
[1650638542.657111] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x31ba3f0: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 2)
[1650638542.657198] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x31ba3f0: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 3)
[1650638542.658364] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x31ba3f0: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 4)
[1650638542.659676] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x31ba3f0: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 5)
[1650638542.660016] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x31ba3f0: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 6)
[1650638542.669389] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x31ba3f0: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 7)
[1650638542.669752] [ndv4:68756:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638542.669764] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x2b4b720 [id=116 ref 1] uct_ud_iface_async_handler() to hash
[1650638542.669796] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 116 events 0x5 mode thread_spinlock
[1650638542.669807] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[29]=0x31ba3f0 using ud_verbs/mlx5_ib5:1 on worker 0x1d084d0
[1650638542.669822] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638542.669828] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638542.670137] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638542.670143] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638542.671296] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib5:1
[1650638542.673066] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638542.673492] [ndv4:68756:0] ib_iface.c:994 UCX DEBUG iface=0x46e4050: created UD QP 0x1cd5 on mlx5_ib5:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638542.673500] [ndv4:68756:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650638542.674166] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638542.674655] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638542.674662] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638542.674783] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638542.674789] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638542.675184] [ndv4:68756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b13ae9e6000..0x2b13aea6b000 on mlx5_ib5 lkey 0x80600 rkey 0x80600 access 0xf flags 0x3e4
[1650638542.675191] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b13ae9e6018 of 544744 bytes with 128 elements
[1650638542.675195] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638542.683541] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x46e4050: adding gid fe80::15:5dff:fd34:20 to hash on device mlx5_ib5 port 1 index 0)
[1650638542.683855] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x46e4050: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 1)
[1650638542.684929] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x46e4050: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 2)
[1650638542.685283] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x46e4050: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 3)
[1650638542.686300] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x46e4050: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 4)
[1650638542.695638] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x46e4050: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 5)
[1650638542.696188] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x46e4050: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 6)
[1650638542.697813] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x46e4050: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 7)
[1650638542.697821] [ndv4:68756:0] 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> }
[1650638542.697858] [ndv4:68756:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638542.697863] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x40a22e0 [id=117 ref 1] uct_ud_iface_async_handler() to hash
[1650638542.697889] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 117 events 0x5 mode thread_spinlock
[1650638542.697901] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[30]=0x46e4050 using ud_mlx5/mlx5_ib5:1 on worker 0x1d084d0
[1650638542.698013] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638542.698019] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638542.698304] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638542.698311] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638542.698489] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib6:1
[1650638542.700280] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650638542.700300] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638542.700303] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638542.700354] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638542.701358] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638542.701369] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650638542.701917] [ndv4:68756:0] ib_iface.c:994 UCX DEBUG iface=0x47e40a0: created RC QP 0x1c26 on mlx5_ib6:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650638542.702304] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[31]=0x47e40a0 using rc_verbs/mlx5_ib6:1 on worker 0x1d084d0
[1650638542.702457] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638542.702463] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638542.702590] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638542.702596] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638542.711871] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib6:1
[1650638542.713187] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638542.713201] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638542.713204] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638542.713290] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638542.713799] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x4b05010 of 8176 bytes with 127 elements
[1650638542.714053] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638542.714063] [ndv4:68756:0] 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> }
[1650638542.714098] [ndv4:68756:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib6 length=2048) failed: Invalid argument
[1650638542.714102] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638542.714112] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x47bf2f0 [id=120 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638542.714138] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 120 events 0x1 mode thread_spinlock
[1650638542.714148] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[32]=0x4931030 using rc_mlx5/mlx5_ib6:1 on worker 0x1d084d0
[1650638542.714295] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638542.714301] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638542.714486] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638542.714491] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638542.715153] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib6:1
[1650638542.716516] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638542.716534] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638542.716537] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638542.716592] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638542.716978] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650638542.716986] [ndv4:68756:0] 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> }
[1650638542.717019] [ndv4:68756:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib6 length=2048) failed: Invalid argument
[1650638542.717023] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638542.717029] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x4b13fc0 [id=122 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638542.717061] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 122 events 0x1 mode thread_spinlock
[1650638544.271600] [ndv4:68756:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650638544.295584] [ndv4:68756:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x4b07050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0x21ab
[1650638544.297628] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638544.298195] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638544.298301] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b13b126e008 of 151544 bytes with 1052 elements
[1650638544.302339] [ndv4:68756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b13aec00000..0x2b13b1200000 on mlx5_ib6 lkey 0x80400 rkey 0x80400 access 0xf flags 0x3e4
[1650638544.302362] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b13aec00018 of 39845864 bytes with 4752 elements
[1650638544.302501] [ndv4:68756:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x4b07050
[1650638544.302533] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[33]=0x4b07050 using dc_mlx5/mlx5_ib6:1 on worker 0x1d084d0
[1650638544.302840] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638544.303181] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638544.303675] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638544.303750] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638544.305410] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib6:1
[1650638544.306624] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638544.307151] [ndv4:68756:0] ib_iface.c:994 UCX DEBUG iface=0x4ce3060: created UD QP 0x21f8 on mlx5_ib6:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638544.307846] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638544.308440] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638544.308772] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638544.309068] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638544.309321] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638544.309810] [ndv4:68756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b13b1293000..0x2b13b1318000 on mlx5_ib6 lkey 0x80500 rkey 0x80500 access 0xf flags 0x3e4
[1650638544.309817] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b13b1293018 of 544744 bytes with 128 elements
[1650638544.309822] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638544.310639] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x4ce3060: adding gid fe80::15:5dff:fd34:21 to hash on device mlx5_ib6 port 1 index 0)
[1650638544.311146] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x4ce3060: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 1)
[1650638544.311679] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x4ce3060: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 2)
[1650638544.311996] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x4ce3060: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 3)
[1650638544.321086] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x4ce3060: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 4)
[1650638544.321841] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x4ce3060: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 5)
[1650638544.336504] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x4ce3060: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 6)
[1650638544.336998] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x4ce3060: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 7)
[1650638544.337620] [ndv4:68756:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638544.337630] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x38d7530 [id=123 ref 1] uct_ud_iface_async_handler() to hash
[1650638544.337664] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 123 events 0x5 mode thread_spinlock
[1650638544.337676] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[34]=0x4ce3060 using ud_verbs/mlx5_ib6:1 on worker 0x1d084d0
[1650638544.338068] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638544.338348] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638544.338972] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638544.339494] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638544.339825] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib6:1
[1650638544.341074] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638544.341870] [ndv4:68756:0] ib_iface.c:994 UCX DEBUG iface=0x3ff47b0: created UD QP 0x2234 on mlx5_ib6:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638544.341878] [ndv4:68756:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650638544.342652] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638544.350400] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638544.350645] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638544.358891] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638544.359164] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638544.359743] [ndv4:68756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b13b1318000..0x2b13b139d000 on mlx5_ib6 lkey 0x80600 rkey 0x80600 access 0xf flags 0x3e4
[1650638544.359751] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b13b1318018 of 544744 bytes with 128 elements
[1650638544.359757] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638544.360038] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x3ff47b0: adding gid fe80::15:5dff:fd34:21 to hash on device mlx5_ib6 port 1 index 0)
[1650638544.360652] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x3ff47b0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 1)
[1650638544.367844] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x3ff47b0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 2)
[1650638544.368812] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x3ff47b0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 3)
[1650638544.368913] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x3ff47b0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 4)
[1650638544.369193] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x3ff47b0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 5)
[1650638544.370123] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x3ff47b0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 6)
[1650638544.380775] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x3ff47b0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 7)
[1650638544.380786] [ndv4:68756:0] 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> }
[1650638544.380888] [ndv4:68756:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638544.380895] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x38d72e0 [id=124 ref 1] uct_ud_iface_async_handler() to hash
[1650638544.380926] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 124 events 0x5 mode thread_spinlock
[1650638544.380948] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[35]=0x3ff47b0 using ud_mlx5/mlx5_ib6:1 on worker 0x1d084d0
[1650638544.388181] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638544.388281] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638544.388553] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638544.389290] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638544.389905] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib7:1
[1650638544.391627] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650638544.391644] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638544.391647] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638544.391696] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638544.392885] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638544.392891] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650638544.393431] [ndv4:68756:0] ib_iface.c:994 UCX DEBUG iface=0x4f010a0: created RC QP 0x21c1 on mlx5_ib7:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650638544.394733] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[36]=0x4f010a0 using rc_verbs/mlx5_ib7:1 on worker 0x1d084d0
[1650638544.403722] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638544.403760] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638544.404299] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638544.404523] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638544.405267] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib7:1
[1650638544.416590] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638544.416606] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638544.416609] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638544.416658] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638544.417281] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x5222010 of 8176 bytes with 127 elements
[1650638544.417546] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638544.417553] [ndv4:68756:0] 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> }
[1650638544.417592] [ndv4:68756:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib7 length=2048) failed: Invalid argument
[1650638544.417596] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638544.417607] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x4e01830 [id=127 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638544.417635] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 127 events 0x1 mode thread_spinlock
[1650638544.417646] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[37]=0x504e030 using rc_mlx5/mlx5_ib7:1 on worker 0x1d084d0
[1650638544.417868] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638544.418343] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638544.418800] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638544.419199] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638544.744299] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib7:1
[1650638544.746379] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638544.746414] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638544.746418] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638544.746478] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638544.747134] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650638544.747159] [ndv4:68756:0] 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> }
[1650638544.747256] [ndv4:68756:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib7 length=2048) failed: Invalid argument
[1650638544.747260] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638544.747276] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x4edc6c0 [id=129 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638544.747299] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 129 events 0x1 mode thread_spinlock
[1650638544.753840] [ndv4:69188:0] debug.c:1198 UCX DEBUG using signal stack 0x2ae4caddb000 size 141824
[1650638544.754475] [ndv4:69188:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650638544.754498] [ndv4:69188:0] 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 0x2ae4cac2f000
[1650638544.755638] [ndv4:69188:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650638544.755651] [ndv4:69188:0] 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
[1650638544.755658] [ndv4:69188:0] module.c:253 UCX DEBUG loading modules for ucs
[1650638544.771042] [ndv4:69188:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650638544.771071] [ndv4:69188:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650638544.771117] [ndv4:69188:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650638544.771121] [ndv4:69188:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650638544.771129] [ndv4:69188:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650638544.771137] [ndv4:69188:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650638544.771140] [ndv4:69188:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650638544.771146] [ndv4:69188:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650638544.771149] [ndv4:69188:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650638544.771151] [ndv4:69188:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650638544.771153] [ndv4:69188:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650638544.771155] [ndv4:69188:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650638544.771169] [ndv4:69188:0] module.c:253 UCX DEBUG loading modules for uct
[1650638544.799847] [ndv4:69188:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650638544.802391] [ndv4:69188:0] ucp_context.c:1117 UCX DEBUG closing md posix because it has no selected transport resources
[1650638544.802417] [ndv4:69188:0] ucp_context.c:1117 UCX DEBUG closing md self because it has no selected transport resources
[1650638544.802984] [ndv4:69188:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650638544.802997] [ndv4:69188:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650638544.803007] [ndv4:69188:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650638544.803018] [ndv4:69188:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650638544.803028] [ndv4:69188:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650638544.803039] [ndv4:69188:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650638544.803050] [ndv4:69188:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650638544.803091] [ndv4:69188:0] ucp_context.c:1117 UCX DEBUG closing md tcp because it has no selected transport resources
[1650638544.808704] [ndv4:69188:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650638544.831114] [ndv4:69188:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650638544.832428] [ndv4:69188:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650638544.844010] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0xc4b340 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650638544.844116] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650638544.844652] [ndv4:69188:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650638544.850581] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638544.851328] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638544.851358] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638544.851428] [ndv4:69188:0] module.c:253 UCX DEBUG loading modules for ucm
[1650638544.851451] [ndv4:69188:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650638544.864571] [ndv4:69188:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638544.864589] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638544.864861] [ndv4:69188:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638544.865337] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638544.865457] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638544.876244] [ndv4:69188:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650638544.876250] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638544.881982] [ndv4:69188:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638544.881989] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638544.888987] [ndv4:69188:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638544.889003] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638544.897270] [ndv4:69188:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638544.897275] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638544.901963] [ndv4:69188:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638544.901969] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638544.904237] [ndv4:69188:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638544.924590] [ndv4:69188:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650638544.925059] [ndv4:69188:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650638544.927852] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0xc553e0 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650638544.927880] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650638544.927883] [ndv4:69188:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650638544.928192] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638544.928251] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638544.928256] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638544.928299] [ndv4:69188:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650638544.930446] [ndv4:69188:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638544.930452] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638544.930721] [ndv4:69188:0] ib_md.c:1604 UCX DEBUG mlx5_ib1: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638544.930955] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638544.931054] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638544.947557] [ndv4:69188:0] topo.c:99 UCX DEBUG bus id 0x102000000 doesn't exist. sys_dev = 1
[1650638544.947564] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638544.955998] [ndv4:69188:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638544.956004] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638544.963780] [ndv4:69188:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638544.963786] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638544.978201] [ndv4:69188:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638544.978237] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638544.984559] [ndv4:69188:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638544.984565] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638544.987870] [ndv4:69188:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638545.013591] [ndv4:69188:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib2 vendor_id: 0x15b3 device_id: 4124
[1650638545.014078] [ndv4:69188:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib2: disable ODP because it's not supported for DevX QP
[1650638545.015904] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0xc270c0 [id=61 ref 1] uct_ib_async_event_handler() to hash
[1650638545.015933] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 61 events 0x1 mode thread_spinlock
[1650638545.015937] [ndv4:69188:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib2' (InfiniBand channel adapter) with 1 ports
[1650638545.016346] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638545.016864] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638545.016869] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638545.016913] [ndv4:69188:0] ib_md.c:1319 UCX DEBUG mlx5_ib2: using registration cache
[1650638545.018661] [ndv4:69188:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638545.018669] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638545.018972] [ndv4:69188:0] ib_md.c:1604 UCX DEBUG mlx5_ib2: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638545.019624] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638545.019699] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638545.045191] [ndv4:69188:0] topo.c:99 UCX DEBUG bus id 0x103000000 doesn't exist. sys_dev = 2
[1650638545.045198] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638545.060407] [ndv4:69188:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638545.060416] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638545.086354] [ndv4:69188:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638545.086364] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638545.111935] [ndv4:69188:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638545.111946] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638545.118193] [ndv4:69188:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638545.118200] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638545.120265] [ndv4:69188:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638545.147526] [ndv4:69188:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib3 vendor_id: 0x15b3 device_id: 4124
[1650638545.148189] [ndv4:69188:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib3: disable ODP because it's not supported for DevX QP
[1650638545.149863] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x1336070 [id=63 ref 1] uct_ib_async_event_handler() to hash
[1650638545.149889] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 63 events 0x1 mode thread_spinlock
[1650638545.149892] [ndv4:69188:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib3' (InfiniBand channel adapter) with 1 ports
[1650638545.150683] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638545.151007] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638545.151013] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638545.151057] [ndv4:69188:0] ib_md.c:1319 UCX DEBUG mlx5_ib3: using registration cache
[1650638545.153271] [ndv4:69188:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638545.153278] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638545.153577] [ndv4:69188:0] ib_md.c:1604 UCX DEBUG mlx5_ib3: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638545.154024] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638545.154287] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638545.163013] [ndv4:69188:0] topo.c:99 UCX DEBUG bus id 0x104000000 doesn't exist. sys_dev = 3
[1650638545.163020] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638545.172546] [ndv4:69188:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638545.172554] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638545.179134] [ndv4:69188:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638545.179141] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638545.186042] [ndv4:69188:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638545.186048] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638545.196284] [ndv4:69188:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638545.196291] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638545.196639] [ndv4:69188:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638545.225416] [ndv4:69188:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib4 vendor_id: 0x15b3 device_id: 4124
[1650638545.225976] [ndv4:69188:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib4: disable ODP because it's not supported for DevX QP
[1650638545.228567] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0xc534c0 [id=65 ref 1] uct_ib_async_event_handler() to hash
[1650638545.228593] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 65 events 0x1 mode thread_spinlock
[1650638545.228597] [ndv4:69188:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib4' (InfiniBand channel adapter) with 1 ports
[1650638545.229094] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638545.229527] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638545.229536] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638545.229578] [ndv4:69188:0] ib_md.c:1319 UCX DEBUG mlx5_ib4: using registration cache
[1650638545.231260] [ndv4:69188:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638545.231267] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638545.231603] [ndv4:69188:0] ib_md.c:1604 UCX DEBUG mlx5_ib4: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638545.232447] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638545.233083] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638545.238936] [ndv4:69188:0] topo.c:99 UCX DEBUG bus id 0x105000000 doesn't exist. sys_dev = 4
[1650638545.238943] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638545.257856] [ndv4:69188:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650638545.257865] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638545.285554] [ndv4:69188:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650638545.285563] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638545.301876] [ndv4:69188:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650638545.301885] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638545.309508] [ndv4:69188:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650638545.309515] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638545.313507] [ndv4:69188:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638545.368979] [ndv4:69188:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib5 vendor_id: 0x15b3 device_id: 4124
[1650638545.369642] [ndv4:69188:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib5: disable ODP because it's not supported for DevX QP
[1650638545.371964] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0xc4d780 [id=67 ref 1] uct_ib_async_event_handler() to hash
[1650638545.371995] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 67 events 0x1 mode thread_spinlock
[1650638545.371999] [ndv4:69188:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib5' (InfiniBand channel adapter) with 1 ports
[1650638545.372728] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638545.372798] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638545.372803] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638545.372847] [ndv4:69188:0] ib_md.c:1319 UCX DEBUG mlx5_ib5: using registration cache
[1650638545.374875] [ndv4:69188:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638545.374882] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638545.375185] [ndv4:69188:0] ib_md.c:1604 UCX DEBUG mlx5_ib5: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638545.375234] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638545.375239] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638545.383967] [ndv4:69188:0] topo.c:99 UCX DEBUG bus id 0x106000000 doesn't exist. sys_dev = 5
[1650638545.383975] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638545.390525] [ndv4:69188:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650638545.390532] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638545.406364] [ndv4:69188:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650638545.406371] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638545.412415] [ndv4:69188:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650638545.412421] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638545.429510] [ndv4:69188:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650638545.429518] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638545.432088] [ndv4:69188:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638545.441744] [ndv4:69188:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib6 vendor_id: 0x15b3 device_id: 4124
[1650638545.442157] [ndv4:69188:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib6: disable ODP because it's not supported for DevX QP
[1650638545.453091] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x13376d0 [id=69 ref 1] uct_ib_async_event_handler() to hash
[1650638545.453121] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 69 events 0x1 mode thread_spinlock
[1650638545.453125] [ndv4:69188:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib6' (InfiniBand channel adapter) with 1 ports
[1650638545.453686] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638545.454163] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638545.454168] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638545.454287] [ndv4:69188:0] ib_md.c:1319 UCX DEBUG mlx5_ib6: using registration cache
[1650638545.455870] [ndv4:69188:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638545.455891] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638545.456202] [ndv4:69188:0] ib_md.c:1604 UCX DEBUG mlx5_ib6: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638545.457504] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638545.457642] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638545.470284] [ndv4:69188:0] topo.c:99 UCX DEBUG bus id 0x107000000 doesn't exist. sys_dev = 6
[1650638545.470294] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650638545.483110] [ndv4:69188:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650638545.483119] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650638545.493286] [ndv4:69188:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650638545.493293] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650638545.504532] [ndv4:69188:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650638545.504539] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650638545.515904] [ndv4:69188:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650638545.515912] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650638545.516294] [ndv4:69188:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638545.537748] [ndv4:69188:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib7 vendor_id: 0x15b3 device_id: 4124
[1650638545.538246] [ndv4:69188:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib7: disable ODP because it's not supported for DevX QP
[1650638545.539067] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0xc41a90 [id=71 ref 1] uct_ib_async_event_handler() to hash
[1650638545.539095] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 71 events 0x1 mode thread_spinlock
[1650638545.539098] [ndv4:69188:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib7' (InfiniBand channel adapter) with 1 ports
[1650638545.539415] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638545.539426] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638545.539432] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638545.539486] [ndv4:69188:0] ib_md.c:1319 UCX DEBUG mlx5_ib7: using registration cache
[1650638545.540951] [ndv4:69188:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638545.540958] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638545.541164] [ndv4:69188:0] ib_md.c:1604 UCX DEBUG mlx5_ib7: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638545.541356] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638545.541362] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638545.550597] [ndv4:69188:0] topo.c:99 UCX DEBUG bus id 0x108000000 doesn't exist. sys_dev = 7
[1650638545.550603] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650638545.562827] [ndv4:69188:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650638545.562834] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650638545.573036] [ndv4:69188:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650638545.573043] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650638545.583123] [ndv4:69188:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650638545.583130] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650638545.612110] [ndv4:69188:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650638545.612118] [ndv4:69188:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650638545.612268] [ndv4:69188:0] ucp_context.c:1556 UCX DEBUG created ucp context 0xc391b0 0xc391b0 [10 mds 42 tls] features 0x1 tl bitmap 0x3ffffffffff 0x0
[1650638545.671264] [ndv4:69188:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 8447 bytes with hugetlb
[1650638545.671364] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool mm_recv_desc: align 64, maxelems 4294967295, elemsize 8368
[1650638545.671390] [ndv4:69188:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 4292720 bytes with hugetlb
[1650638545.671400] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool mm_recv_desc: allocated chunk 0x2ae4d2454018 of 4296680 bytes with 512 elements
[1650638545.671955] [ndv4:69188:0] mm_iface.c:600 UCX DEBUG created mm iface 0x13b51e0 FIFO id 0xd0037 va 0x2ae4cbc81000 size 12288 (128 x 64 elems)
[1650638545.672003] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[0]=0x13b51e0 using sysv/memory on worker 0x1c8f6f0
[1650638545.672195] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638545.672285] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638545.672354] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638545.672361] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638545.680944] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib0:1
[1650638545.683311] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650638545.683350] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638545.683354] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638545.683407] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638545.684647] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638545.684656] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650638545.686329] [ndv4:69188:0] ib_iface.c:994 UCX DEBUG iface=0x13c0480: created RC QP 0x30d4 on mlx5_ib0:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650638545.688722] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[1]=0x13c0480 using rc_verbs/mlx5_ib0:1 on worker 0x1c8f6f0
[1650638545.688797] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638545.688803] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638545.688905] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638545.688910] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638545.689315] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib0:1
[1650638545.689610] [ndv4:69188:0] ib_device.c:1394 UCX DEBUG max IB CQE size is 128
[1650638545.690693] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638545.690702] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638545.690704] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638545.690754] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638545.691336] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x1fb6010 of 8176 bytes with 127 elements
[1650638545.691630] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638545.691653] [ndv4:69188:0] 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> }
[1650638545.691703] [ndv4:69188:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650638545.691707] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638545.700936] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x131c590 [id=78 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638545.700969] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 78 events 0x1 mode thread_spinlock
[1650638545.701621] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[2]=0x13ce010 using rc_mlx5/mlx5_ib0:1 on worker 0x1c8f6f0
[1650638545.701791] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638545.701797] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638545.701973] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638545.701978] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638545.711285] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib0:1
[1650638545.721002] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638545.721013] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638545.721015] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638545.721029] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638545.721498] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650638545.721507] [ndv4:69188:0] 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> }
[1650638545.721540] [ndv4:69188:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650638545.721543] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638545.721550] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x131c8a0 [id=80 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638545.721571] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 80 events 0x1 mode thread_spinlock
[1650638545.722386] [ndv4:69188:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650638545.730066] [ndv4:69188:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x1fb8050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0x30f7
[1650638545.730375] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638545.730381] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638545.730400] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2ae4cbc86008 of 151544 bytes with 1052 elements
[1650638545.734388] [ndv4:69188:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae4d2a00000..0x2ae4d5000000 on mlx5_ib0 lkey 0x80b00 rkey 0x80b00 access 0xf flags 0x3e4
[1650638545.734413] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2ae4d2a00018 of 39845864 bytes with 4752 elements
[1650638545.734552] [ndv4:69188:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x1fb8050
[1650638545.734583] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[3]=0x1fb8050 using dc_mlx5/mlx5_ib0:1 on worker 0x1c8f6f0
[1650638545.734665] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638545.734673] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638545.734794] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638545.734798] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638545.735188] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib0:1
[1650638545.736048] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638545.736444] [ndv4:69188:0] ib_iface.c:994 UCX DEBUG iface=0x13cb650: created UD QP 0x30dd on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638545.737021] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638545.737103] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638545.737110] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638545.737140] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638545.737145] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638545.737569] [ndv4:69188:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae4cbcab000..0x2ae4cbd30000 on mlx5_ib0 lkey 0x80c00 rkey 0x80c00 access 0xf flags 0x3e4
[1650638545.737577] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae4cbcab018 of 544744 bytes with 128 elements
[1650638545.737581] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638545.737775] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x13cb650: adding gid fe80::15:5dff:fd34:1b to hash on device mlx5_ib0 port 1 index 0)
[1650638545.738075] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x13cb650: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650638545.738297] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x13cb650: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650638545.748820] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x13cb650: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650638545.748942] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x13cb650: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650638545.748958] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x13cb650: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650638545.749194] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x13cb650: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650638545.749334] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x13cb650: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650638545.749652] [ndv4:69188:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638545.757568] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x13bcd40 [id=81 ref 1] uct_ud_iface_async_handler() to hash
[1650638545.757603] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 81 events 0x5 mode thread_spinlock
[1650638545.758144] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[4]=0x13cb650 using ud_verbs/mlx5_ib0:1 on worker 0x1c8f6f0
[1650638545.758380] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638545.758389] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638545.758607] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638545.758613] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638545.759121] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib0:1
[1650638545.760038] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638545.760329] [ndv4:69188:0] ib_iface.c:994 UCX DEBUG iface=0x1caf570: created UD QP 0x30de on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638545.760348] [ndv4:69188:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650638545.760893] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638545.768528] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638545.768538] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638545.768619] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638545.768624] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638545.769057] [ndv4:69188:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae4cbd30000..0x2ae4cbdb5000 on mlx5_ib0 lkey 0x80d00 rkey 0x80d00 access 0xf flags 0x3e4
[1650638545.769064] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae4cbd30018 of 544744 bytes with 128 elements
[1650638545.769068] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638545.769392] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x1caf570: adding gid fe80::15:5dff:fd34:1b to hash on device mlx5_ib0 port 1 index 0)
[1650638545.769523] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x1caf570: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650638545.769650] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x1caf570: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650638545.769851] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x1caf570: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650638545.770074] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x1caf570: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650638545.770426] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x1caf570: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650638545.770774] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x1caf570: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650638545.770969] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x1caf570: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650638545.770977] [ndv4:69188:0] 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> }
[1650638545.771006] [ndv4:69188:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638545.771012] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x1322b60 [id=82 ref 1] uct_ud_iface_async_handler() to hash
[1650638545.771035] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 82 events 0x5 mode thread_spinlock
[1650638545.771048] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[5]=0x1caf570 using ud_mlx5/mlx5_ib0:1 on worker 0x1c8f6f0
[1650638545.771306] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638545.771312] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638545.771421] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638545.771428] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638545.771994] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib1:1
[1650638545.784988] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650638545.784997] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638545.784999] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638545.785013] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638545.786168] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638545.786174] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650638545.787604] [ndv4:69188:0] ib_iface.c:994 UCX DEBUG iface=0x255f050: created RC QP 0x302a on mlx5_ib1:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650638545.788655] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[6]=0x255f050 using rc_verbs/mlx5_ib1:1 on worker 0x1c8f6f0
[1650638545.788766] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638545.788772] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638545.789095] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638545.789100] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638545.789914] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib1:1
[1650638545.791300] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638545.791310] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638545.791313] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638545.791326] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638545.791823] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x24b8010 of 8176 bytes with 127 elements
[1650638545.792084] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638545.792091] [ndv4:69188:0] 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> }
[1650638545.792126] [ndv4:69188:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650638545.792130] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638545.792142] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x1325ba0 [id=85 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638545.792161] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 85 events 0x1 mode thread_spinlock
[1650638545.792170] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[7]=0x23e5020 using rc_mlx5/mlx5_ib1:1 on worker 0x1c8f6f0
[1650638545.792259] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638545.792266] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638545.792361] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638545.792366] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638545.792586] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib1:1
[1650638545.793628] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638545.793644] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638545.793647] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638545.793702] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638545.794064] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650638545.794071] [ndv4:69188:0] 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> }
[1650638545.794103] [ndv4:69188:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650638545.794107] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638545.794115] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x13ba760 [id=87 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638545.794135] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 87 events 0x1 mode thread_spinlock
[1650638545.794822] [ndv4:69188:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650638545.802747] [ndv4:69188:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x2704010: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0x304d
[1650638545.803289] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638545.803295] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638545.803303] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2ae4cbdb7008 of 151544 bytes with 1052 elements
[1650638545.807320] [ndv4:69188:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae4d5200000..0x2ae4d7800000 on mlx5_ib1 lkey 0x80700 rkey 0x80700 access 0xf flags 0x3e4
[1650638545.807342] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2ae4d5200018 of 39845864 bytes with 4752 elements
[1650638545.807477] [ndv4:69188:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x2704010
[1650638545.807512] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[8]=0x2704010 using dc_mlx5/mlx5_ib1:1 on worker 0x1c8f6f0
[1650638545.808007] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638545.808015] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638545.808090] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638545.808094] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638545.809196] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib1:1
[1650638545.810408] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638545.810790] [ndv4:69188:0] ib_iface.c:994 UCX DEBUG iface=0x13d6be0: created UD QP 0x3033 on mlx5_ib1:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638545.811370] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638545.811858] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638545.811865] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638545.812071] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638545.812075] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638545.812580] [ndv4:69188:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae4cbddc000..0x2ae4cbe61000 on mlx5_ib1 lkey 0x80800 rkey 0x80800 access 0xf flags 0x3e4
[1650638545.812587] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae4cbddc018 of 544744 bytes with 128 elements
[1650638545.812591] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638545.812921] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x13d6be0: adding gid fe80::15:5dff:fd34:1c to hash on device mlx5_ib1 port 1 index 0)
[1650638545.813546] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x13d6be0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 1)
[1650638545.813966] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x13d6be0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 2)
[1650638545.815259] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x13d6be0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 3)
[1650638545.815472] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x13d6be0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 4)
[1650638545.815606] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x13d6be0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 5)
[1650638545.815621] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x13d6be0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 6)
[1650638545.823696] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x13d6be0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 7)
[1650638545.824070] [ndv4:69188:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638545.824078] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x13d8010 [id=88 ref 1] uct_ud_iface_async_handler() to hash
[1650638545.824105] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 88 events 0x5 mode thread_spinlock
[1650638545.824115] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[9]=0x13d6be0 using ud_verbs/mlx5_ib1:1 on worker 0x1c8f6f0
[1650638545.824325] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638545.824331] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638545.824447] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638545.824451] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638545.825051] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib1:1
[1650638545.826102] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638545.826433] [ndv4:69188:0] ib_iface.c:994 UCX DEBUG iface=0x24ba050: created UD QP 0x3034 on mlx5_ib1:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638545.826441] [ndv4:69188:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650638545.826953] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638545.827036] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638545.827041] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638545.827100] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638545.827105] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638545.827510] [ndv4:69188:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae4cbe61000..0x2ae4cbee6000 on mlx5_ib1 lkey 0x80900 rkey 0x80900 access 0xf flags 0x3e4
[1650638545.827516] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae4cbe61018 of 544744 bytes with 128 elements
[1650638545.827521] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638545.827561] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x24ba050: adding gid fe80::15:5dff:fd34:1c to hash on device mlx5_ib1 port 1 index 0)
[1650638545.827586] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x24ba050: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 1)
[1650638545.827606] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x24ba050: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 2)
[1650638545.827621] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x24ba050: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 3)
[1650638545.827643] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x24ba050: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 4)
[1650638545.827661] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x24ba050: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 5)
[1650638545.827685] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x24ba050: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 6)
[1650638545.827703] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x24ba050: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 7)
[1650638545.827709] [ndv4:69188:0] 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> }
[1650638545.827738] [ndv4:69188:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638545.827743] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x23edc40 [id=89 ref 1] uct_ud_iface_async_handler() to hash
[1650638545.827763] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 89 events 0x5 mode thread_spinlock
[1650638545.827773] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[10]=0x24ba050 using ud_mlx5/mlx5_ib1:1 on worker 0x1c8f6f0
[1650638545.827786] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638545.827791] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638545.827852] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638545.827856] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638545.828116] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib2:1
[1650638545.829770] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650638545.829787] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638545.829790] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638545.829838] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638545.831103] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638545.831109] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650638545.832469] [ndv4:69188:0] ib_iface.c:994 UCX DEBUG iface=0x2af7020: created RC QP 0x3036 on mlx5_ib2:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650638545.833500] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[11]=0x2af7020 using rc_verbs/mlx5_ib2:1 on worker 0x1c8f6f0
[1650638545.833642] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638545.833648] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638545.833865] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638545.833870] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638545.834503] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib2:1
[1650638545.836172] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638545.836185] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638545.836188] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638545.836284] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638545.836802] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x2e18010 of 8176 bytes with 127 elements
[1650638545.837058] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638545.837065] [ndv4:69188:0] 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> }
[1650638545.837099] [ndv4:69188:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib2 length=2048) failed: Invalid argument
[1650638545.837110] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638545.837122] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x2570350 [id=92 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638545.837144] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 92 events 0x1 mode thread_spinlock
[1650638545.837154] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[12]=0x2c44030 using rc_mlx5/mlx5_ib2:1 on worker 0x1c8f6f0
[1650638545.837333] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638545.837339] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638545.845509] [ndv4:69197:0] debug.c:1198 UCX DEBUG using signal stack 0x2afd7ddfc000 size 141824
[1650638545.845599] [ndv4:69197:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650638545.845619] [ndv4:69197:0] 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 0x2afd7dc50000
[1650638545.845637] [ndv4:69197:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650638545.845646] [ndv4:69197:0] 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
[1650638545.845652] [ndv4:69197:0] module.c:253 UCX DEBUG loading modules for ucs
[1650638545.847905] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638545.847917] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638545.848301] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib2:1
[1650638545.848419] [ndv4:69197:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650638545.848441] [ndv4:69197:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650638545.848480] [ndv4:69197:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650638545.848483] [ndv4:69197:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650638545.848490] [ndv4:69197:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650638545.848497] [ndv4:69197:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650638545.848500] [ndv4:69197:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650638545.848504] [ndv4:69197:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650638545.848506] [ndv4:69197:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650638545.848508] [ndv4:69197:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650638545.848511] [ndv4:69197:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650638545.848513] [ndv4:69197:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650638545.848522] [ndv4:69197:0] module.c:253 UCX DEBUG loading modules for uct
[1650638545.850013] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638545.850038] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638545.850042] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638545.850105] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638545.850671] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650638545.850737] [ndv4:69188:0] 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> }
[1650638545.850776] [ndv4:69188:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib2 length=2048) failed: Invalid argument
[1650638545.850780] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638545.850794] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x2c4cfa0 [id=94 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638545.850818] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 94 events 0x1 mode thread_spinlock
[1650638545.851523] [ndv4:69188:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650638545.854491] [ndv4:69197:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650638545.854837] [ndv4:69197:0] ucp_context.c:1117 UCX DEBUG closing md posix because it has no selected transport resources
[1650638545.854859] [ndv4:69197:0] ucp_context.c:1117 UCX DEBUG closing md self because it has no selected transport resources
[1650638545.855117] [ndv4:69197:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650638545.855132] [ndv4:69197:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650638545.855143] [ndv4:69197:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650638545.855153] [ndv4:69197:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650638545.855163] [ndv4:69197:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650638545.855173] [ndv4:69197:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650638545.855184] [ndv4:69197:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650638545.855291] [ndv4:69197:0] ucp_context.c:1117 UCX DEBUG closing md tcp because it has no selected transport resources
[1650638545.855657] [ndv4:69197:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650638545.859155] [ndv4:69188:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x2e1a050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0x3059
[1650638545.859454] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638545.859462] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638545.859476] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2ae4cbee8008 of 151544 bytes with 1052 elements
[1650638545.864391] [ndv4:69188:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae4d7a00000..0x2ae4da000000 on mlx5_ib2 lkey 0x80700 rkey 0x80700 access 0xf flags 0x3e4
[1650638545.864421] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2ae4d7a00018 of 39845864 bytes with 4752 elements
[1650638545.864564] [ndv4:69188:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x2e1a050
[1650638545.864629] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[13]=0x2e1a050 using dc_mlx5/mlx5_ib2:1 on worker 0x1c8f6f0
[1650638545.864660] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638545.864671] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638545.864767] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638545.864772] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638545.865302] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib2:1
[1650638545.875777] [ndv4:69197:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650638545.876175] [ndv4:69197:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650638545.876325] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638545.876712] [ndv4:69188:0] ib_iface.c:994 UCX DEBUG iface=0x13ca770: created UD QP 0x303f on mlx5_ib2:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638545.878039] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638545.878414] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638545.878420] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638545.878527] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638545.878531] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638545.879033] [ndv4:69188:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae4cbf0d000..0x2ae4cbf92000 on mlx5_ib2 lkey 0x80800 rkey 0x80800 access 0xf flags 0x3e4
[1650638545.879040] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae4cbf0d018 of 544744 bytes with 128 elements
[1650638545.879045] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638545.879401] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x13ca770: adding gid fe80::15:5dff:fd34:1d to hash on device mlx5_ib2 port 1 index 0)
[1650638545.879759] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x13ca770: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 1)
[1650638545.880252] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x13ca770: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 2)
[1650638545.884083] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x2029320 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650638545.884200] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650638545.884370] [ndv4:69197:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650638545.885164] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638545.885177] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638545.885222] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638545.885288] [ndv4:69197:0] module.c:253 UCX DEBUG loading modules for ucm
[1650638545.885310] [ndv4:69197:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650638545.889153] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x13ca770: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 3)
[1650638545.889751] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x13ca770: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 4)
[1650638545.890107] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x13ca770: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 5)
[1650638545.890401] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x13ca770: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 6)
[1650638545.897868] [ndv4:69197:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638545.897885] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638545.898154] [ndv4:69197:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638545.898308] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638545.898314] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638545.905923] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x13ca770: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 7)
[1650638545.906323] [ndv4:69188:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638545.906333] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x13b6700 [id=95 ref 1] uct_ud_iface_async_handler() to hash
[1650638545.906362] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 95 events 0x5 mode thread_spinlock
[1650638545.906376] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[14]=0x13ca770 using ud_verbs/mlx5_ib2:1 on worker 0x1c8f6f0
[1650638545.906510] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638545.906516] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638545.906608] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638545.906612] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638545.906999] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib2:1
[1650638545.907959] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638545.908347] [ndv4:69188:0] ib_iface.c:994 UCX DEBUG iface=0x13b3590: created UD QP 0x3040 on mlx5_ib2:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638545.908356] [ndv4:69188:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650638545.908939] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638545.909356] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638545.909362] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638545.909498] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638545.909503] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638545.909980] [ndv4:69197:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650638545.909988] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638545.909885] [ndv4:69188:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae4da070000..0x2ae4da0f5000 on mlx5_ib2 lkey 0x80900 rkey 0x80900 access 0xf flags 0x3e4
[1650638545.909891] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae4da070018 of 544744 bytes with 128 elements
[1650638545.909895] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638545.910621] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x13b3590: adding gid fe80::15:5dff:fd34:1d to hash on device mlx5_ib2 port 1 index 0)
[1650638545.911579] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x13b3590: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 1)
[1650638545.911997] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x13b3590: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 2)
[1650638545.912279] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x13b3590: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 3)
[1650638545.912723] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x13b3590: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 4)
[1650638545.912995] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x13b3590: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 5)
[1650638545.913359] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x13b3590: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 6)
[1650638545.913665] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x13b3590: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 7)
[1650638545.913671] [ndv4:69188:0] 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> }
[1650638545.913701] [ndv4:69188:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638545.913706] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x13d8ee0 [id=96 ref 1] uct_ud_iface_async_handler() to hash
[1650638545.913728] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 96 events 0x5 mode thread_spinlock
[1650638545.913739] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[15]=0x13b3590 using ud_mlx5/mlx5_ib2:1 on worker 0x1c8f6f0
[1650638545.913934] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638545.913939] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638545.914055] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638545.914060] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638545.912579] [ndv4:69197:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638545.912586] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638545.915144] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib3:1
[1650638545.917363] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650638545.917380] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638545.917383] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638545.917432] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638545.918681] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638545.918688] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650638545.920041] [ndv4:69188:0] ib_iface.c:994 UCX DEBUG iface=0x32140a0: created RC QP 0x3028 on mlx5_ib3:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650638545.921025] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[16]=0x32140a0 using rc_verbs/mlx5_ib3:1 on worker 0x1c8f6f0
[1650638545.921221] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638545.921228] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638545.921412] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638545.921416] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638545.921968] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib3:1
[1650638545.923272] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638545.923285] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638545.923288] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638545.923340] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638545.923866] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x3535010 of 8176 bytes with 127 elements
[1650638545.924148] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638545.924155] [ndv4:69188:0] 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> }
[1650638545.924192] [ndv4:69188:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib3 length=2048) failed: Invalid argument
[1650638545.924196] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638545.924257] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x13c9f00 [id=99 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638545.924278] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 99 events 0x1 mode thread_spinlock
[1650638545.924300] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[17]=0x3361030 using rc_mlx5/mlx5_ib3:1 on worker 0x1c8f6f0
[1650638545.924329] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638545.924334] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638545.924427] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638545.924431] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638545.924666] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib3:1
[1650638545.925794] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638545.925812] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638545.925815] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638545.925870] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638545.926279] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650638545.926286] [ndv4:69188:0] 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> }
[1650638545.926318] [ndv4:69188:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib3 length=2048) failed: Invalid argument
[1650638545.926322] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638545.926329] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x28e0f80 [id=101 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638545.926352] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 101 events 0x1 mode thread_spinlock
[1650638545.927026] [ndv4:69188:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650638545.931750] [ndv4:69197:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638545.931758] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638545.934669] [ndv4:69188:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x3537050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0x304b
[1650638545.934927] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638545.934934] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638545.934944] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2ae4cbf94008 of 151544 bytes with 1052 elements
[1650638545.938953] [ndv4:69188:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae4da200000..0x2ae4dc800000 on mlx5_ib3 lkey 0x80700 rkey 0x80700 access 0xf flags 0x3e4
[1650638545.938974] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2ae4da200018 of 39845864 bytes with 4752 elements
[1650638545.939116] [ndv4:69188:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x3537050
[1650638545.939153] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[18]=0x3537050 using dc_mlx5/mlx5_ib3:1 on worker 0x1c8f6f0
[1650638545.939235] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638545.939244] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638545.939314] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638545.939318] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638545.949152] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib3:1
[1650638545.950526] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638545.950911] [ndv4:69188:0] ib_iface.c:994 UCX DEBUG iface=0x3713060: created UD QP 0x3031 on mlx5_ib3:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638545.951422] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638545.951790] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638545.951797] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638545.951880] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638545.951885] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638545.952258] [ndv4:69197:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638545.952266] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638545.952307] [ndv4:69188:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae4dc8f6000..0x2ae4dc97b000 on mlx5_ib3 lkey 0x80800 rkey 0x80800 access 0xf flags 0x3e4
[1650638545.952315] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae4dc8f6018 of 544744 bytes with 128 elements
[1650638545.952320] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638545.952746] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3713060: adding gid fe80::15:5dff:fd34:1e to hash on device mlx5_ib3 port 1 index 0)
[1650638545.953165] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3713060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 1)
[1650638545.953508] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3713060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 2)
[1650638545.953766] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3713060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 3)
[1650638545.954120] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3713060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 4)
[1650638545.954523] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3713060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 5)
[1650638545.954770] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3713060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 6)
[1650638545.955018] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3713060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 7)
[1650638545.955315] [ndv4:69188:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638545.955324] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0xc3f3e0 [id=102 ref 1] uct_ud_iface_async_handler() to hash
[1650638545.955353] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 102 events 0x5 mode thread_spinlock
[1650638545.955363] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[19]=0x3713060 using ud_verbs/mlx5_ib3:1 on worker 0x1c8f6f0
[1650638545.955476] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638545.955482] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638545.955563] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638545.955568] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638545.955761] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib3:1
[1650638545.956720] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638545.957033] [ndv4:69188:0] ib_iface.c:994 UCX DEBUG iface=0x3831050: created UD QP 0x3032 on mlx5_ib3:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638545.957040] [ndv4:69188:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650638545.957547] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638545.957673] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638545.957678] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638545.957689] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638545.957693] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638545.958027] [ndv4:69188:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae4dc97b000..0x2ae4dca00000 on mlx5_ib3 lkey 0x80900 rkey 0x80900 access 0xf flags 0x3e4
[1650638545.958033] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae4dc97b018 of 544744 bytes with 128 elements
[1650638545.958037] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638545.959320] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3831050: adding gid fe80::15:5dff:fd34:1e to hash on device mlx5_ib3 port 1 index 0)
[1650638545.959779] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3831050: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 1)
[1650638545.960446] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3831050: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 2)
[1650638545.960503] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3831050: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 3)
[1650638545.961028] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3831050: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 4)
[1650638545.961296] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3831050: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 5)
[1650638545.961556] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3831050: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 6)
[1650638545.961923] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3831050: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 7)
[1650638545.961929] [ndv4:69188:0] 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> }
[1650638545.961957] [ndv4:69188:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638545.961961] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x28e0b40 [id=103 ref 1] uct_ud_iface_async_handler() to hash
[1650638545.961982] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 103 events 0x5 mode thread_spinlock
[1650638545.961994] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[20]=0x3831050 using ud_mlx5/mlx5_ib3:1 on worker 0x1c8f6f0
[1650638545.962436] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638545.962443] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638545.962798] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638545.962803] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638545.963801] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib4:1
[1650638545.965529] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650638545.965547] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638545.965550] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638545.965601] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638545.966736] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638545.966742] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650638545.967717] [ndv4:69188:0] ib_iface.c:994 UCX DEBUG iface=0x39310a0: created RC QP 0x3046 on mlx5_ib4:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650638545.968704] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[21]=0x39310a0 using rc_verbs/mlx5_ib4:1 on worker 0x1c8f6f0
[1650638545.968869] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638545.968875] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638545.969052] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638545.969057] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638545.969623] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib4:1
[1650638545.970850] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638545.970864] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638545.970867] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638545.970916] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638545.971456] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x3c52010 of 8176 bytes with 127 elements
[1650638545.971714] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638545.971720] [ndv4:69188:0] 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> }
[1650638545.971756] [ndv4:69188:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib4 length=2048) failed: Invalid argument
[1650638545.971760] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638545.971771] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x2ff6930 [id=106 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638545.971792] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 106 events 0x1 mode thread_spinlock
[1650638545.971801] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[22]=0x3a7e030 using rc_mlx5/mlx5_ib4:1 on worker 0x1c8f6f0
[1650638545.971890] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638545.971895] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638545.972859] [ndv4:69197:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638545.972866] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638545.973098] [ndv4:69197:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638545.980077] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638545.980085] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638545.980767] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib4:1
[1650638545.981926] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638545.981942] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638545.981945] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638545.982004] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638545.982443] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650638545.982451] [ndv4:69188:0] 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> }
[1650638545.982483] [ndv4:69188:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib4 length=2048) failed: Invalid argument
[1650638545.982486] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638545.982496] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x3a86f60 [id=108 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638545.982517] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 108 events 0x1 mode thread_spinlock
[1650638545.983228] [ndv4:69188:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650638545.983949] [ndv4:69197:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650638545.984340] [ndv4:69197:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650638545.986192] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x20333c0 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650638545.986310] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650638545.986315] [ndv4:69197:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650638545.986679] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638545.986688] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638545.986696] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638545.986756] [ndv4:69197:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650638545.990875] [ndv4:69188:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x3c54050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0x3059
[1650638545.995428] [ndv4:69197:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638545.995436] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638545.995702] [ndv4:69197:0] ib_md.c:1604 UCX DEBUG mlx5_ib1: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638545.995789] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638545.995795] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638546.000730] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638546.000743] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638546.000755] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2ae4cbfbb008 of 151544 bytes with 1052 elements
[1650638546.004607] [ndv4:69188:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae4dcc00000..0x2ae4df200000 on mlx5_ib4 lkey 0x80900 rkey 0x80900 access 0xf flags 0x3e4
[1650638546.004631] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2ae4dcc00018 of 39845864 bytes with 4752 elements
[1650638546.004774] [ndv4:69188:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x3c54050
[1650638546.004813] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[23]=0x3c54050 using dc_mlx5/mlx5_ib4:1 on worker 0x1c8f6f0
[1650638546.004915] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638546.004924] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638546.005028] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638546.005032] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638546.005757] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib4:1
[1650638546.006858] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638546.007249] [ndv4:69188:0] ib_iface.c:994 UCX DEBUG iface=0x3e30060: created UD QP 0x304f on mlx5_ib4:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638546.007781] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638546.015510] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638546.015518] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638546.015689] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638546.015694] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638546.016520] [ndv4:69197:0] topo.c:99 UCX DEBUG bus id 0x102000000 doesn't exist. sys_dev = 1
[1650638546.016528] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638546.016074] [ndv4:69188:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae4df201000..0x2ae4df286000 on mlx5_ib4 lkey 0x80a00 rkey 0x80a00 access 0xf flags 0x3e4
[1650638546.016080] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae4df201018 of 544744 bytes with 128 elements
[1650638546.016085] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638546.016523] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3e30060: adding gid fe80::15:5dff:fd34:1f to hash on device mlx5_ib4 port 1 index 0)
[1650638546.016759] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3e30060: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 1)
[1650638546.016776] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3e30060: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 2)
[1650638546.016791] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3e30060: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 3)
[1650638546.017145] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3e30060: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 4)
[1650638546.017557] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3e30060: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 5)
[1650638546.017677] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3e30060: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 6)
[1650638546.018010] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3e30060: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 7)
[1650638546.018321] [ndv4:69188:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638546.018330] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x3114fc0 [id=109 ref 1] uct_ud_iface_async_handler() to hash
[1650638546.018358] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 109 events 0x5 mode thread_spinlock
[1650638546.018368] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[24]=0x3e30060 using ud_verbs/mlx5_ib4:1 on worker 0x1c8f6f0
[1650638546.018411] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638546.018416] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638546.027738] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638546.027748] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638546.028180] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib4:1
[1650638546.029319] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638546.029659] [ndv4:69188:0] ib_iface.c:994 UCX DEBUG iface=0x3f4e460: created UD QP 0x3050 on mlx5_ib4:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638546.029666] [ndv4:69188:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650638546.030203] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638546.030347] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638546.030354] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638546.030366] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638546.030370] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638546.030702] [ndv4:69188:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae4df286000..0x2ae4df30b000 on mlx5_ib4 lkey 0x80b00 rkey 0x80b00 access 0xf flags 0x3e4
[1650638546.030708] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae4df286018 of 544744 bytes with 128 elements
[1650638546.030712] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638546.031069] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3f4e460: adding gid fe80::15:5dff:fd34:1f to hash on device mlx5_ib4 port 1 index 0)
[1650638546.041071] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3f4e460: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 1)
[1650638546.041240] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3f4e460: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 2)
[1650638546.041376] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3f4e460: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 3)
[1650638546.041527] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3f4e460: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 4)
[1650638546.041679] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3f4e460: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 5)
[1650638546.041922] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3f4e460: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 6)
[1650638546.041967] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3f4e460: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 7)
[1650638546.041972] [ndv4:69188:0] 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> }
[1650638546.042002] [ndv4:69188:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638546.042006] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x390ce30 [id=110 ref 1] uct_ud_iface_async_handler() to hash
[1650638546.042026] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 110 events 0x5 mode thread_spinlock
[1650638546.042034] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[25]=0x3f4e460 using ud_mlx5/mlx5_ib4:1 on worker 0x1c8f6f0
[1650638546.042091] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638546.042096] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638546.042187] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638546.042192] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638546.042876] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib5:1
[1650638546.044674] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650638546.044691] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638546.044694] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638546.044745] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638546.045937] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638546.045943] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650638546.047524] [ndv4:69197:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638546.047532] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638546.047611] [ndv4:69188:0] ib_iface.c:994 UCX DEBUG iface=0x404e0a0: created RC QP 0x2b9d on mlx5_ib5:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650638546.048565] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[26]=0x404e0a0 using rc_verbs/mlx5_ib5:1 on worker 0x1c8f6f0
[1650638546.048766] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638546.048771] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638546.048929] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638546.048934] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638546.049510] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib5:1
[1650638546.050883] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638546.050896] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638546.050899] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638546.050950] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638546.051480] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x436f010 of 8176 bytes with 127 elements
[1650638546.051738] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638546.051746] [ndv4:69188:0] 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> }
[1650638546.051780] [ndv4:69188:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib5 length=2048) failed: Invalid argument
[1650638546.051783] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638546.051795] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x3f4e3e0 [id=113 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638546.051815] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 113 events 0x1 mode thread_spinlock
[1650638546.051825] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[27]=0x419b030 using rc_mlx5/mlx5_ib5:1 on worker 0x1c8f6f0
[1650638546.052080] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638546.052086] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638546.052422] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638546.052428] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638546.052996] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib5:1
[1650638546.054416] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638546.054434] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638546.054437] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638546.054492] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638546.054856] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650638546.054863] [ndv4:69188:0] 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> }
[1650638546.054903] [ndv4:69188:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib5 length=2048) failed: Invalid argument
[1650638546.054907] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638546.054913] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x2ad29a0 [id=115 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638546.054932] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 115 events 0x1 mode thread_spinlock
[1650638546.055612] [ndv4:69188:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650638546.063252] [ndv4:69188:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x4371050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0x2bc0
[1650638546.063582] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638546.063588] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638546.063596] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2ae4e1b0c008 of 151544 bytes with 1052 elements
[1650638546.067612] [ndv4:69188:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae4df400000..0x2ae4e1a00000 on mlx5_ib5 lkey 0x80700 rkey 0x80700 access 0xf flags 0x3e4
[1650638546.067636] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2ae4df400018 of 39845864 bytes with 4752 elements
[1650638546.067771] [ndv4:69188:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x4371050
[1650638546.067805] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[28]=0x4371050 using dc_mlx5/mlx5_ib5:1 on worker 0x1c8f6f0
[1650638546.068076] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638546.068086] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638546.068365] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638546.068370] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638546.069121] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib5:1
[1650638546.070021] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638546.070370] [ndv4:69188:0] ib_iface.c:994 UCX DEBUG iface=0x3141480: created UD QP 0x2ba6 on mlx5_ib5:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638546.070937] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638546.074953] [ndv4:69197:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638546.074960] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638546.079781] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638546.079790] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638546.079931] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638546.079937] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638546.080362] [ndv4:69188:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae4e1b31000..0x2ae4e1bb6000 on mlx5_ib5 lkey 0x80800 rkey 0x80800 access 0xf flags 0x3e4
[1650638546.080369] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae4e1b31018 of 544744 bytes with 128 elements
[1650638546.080373] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638546.080415] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3141480: adding gid fe80::15:5dff:fd34:20 to hash on device mlx5_ib5 port 1 index 0)
[1650638546.080433] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3141480: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 1)
[1650638546.080448] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3141480: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 2)
[1650638546.089801] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3141480: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 3)
[1650638546.090352] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3141480: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 4)
[1650638546.090908] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3141480: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 5)
[1650638546.092440] [ndv4:69197:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638546.092447] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638546.093711] [ndv4:69197:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638546.093717] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638546.093997] [ndv4:69197:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638546.100279] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3141480: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 6)
[1650638546.100699] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3141480: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 7)
[1650638546.101030] [ndv4:69188:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638546.101038] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x2ad27d0 [id=116 ref 1] uct_ud_iface_async_handler() to hash
[1650638546.101064] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 116 events 0x5 mode thread_spinlock
[1650638546.101075] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[29]=0x3141480 using ud_verbs/mlx5_ib5:1 on worker 0x1c8f6f0
[1650638546.101256] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638546.101262] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638546.101425] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638546.101430] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638546.101944] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib5:1
[1650638546.103060] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638546.103477] [ndv4:69188:0] ib_iface.c:994 UCX DEBUG iface=0x466b050: created UD QP 0x2ba7 on mlx5_ib5:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638546.103485] [ndv4:69188:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650638546.104025] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638546.104270] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638546.104275] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638546.104318] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638546.104322] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638546.104844] [ndv4:69188:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae4e1bb6000..0x2ae4e1c3b000 on mlx5_ib5 lkey 0x80900 rkey 0x80900 access 0xf flags 0x3e4
[1650638546.104851] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae4e1bb6018 of 544744 bytes with 128 elements
[1650638546.104855] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638546.105516] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x466b050: adding gid fe80::15:5dff:fd34:20 to hash on device mlx5_ib5 port 1 index 0)
[1650638546.105725] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x466b050: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 1)
[1650638546.105888] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x466b050: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 2)
[1650638546.106358] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x466b050: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 3)
[1650638546.106804] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x466b050: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 4)
[1650638546.107088] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x466b050: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 5)
[1650638546.112383] [ndv4:69197:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib2 vendor_id: 0x15b3 device_id: 4124
[1650638546.112736] [ndv4:69197:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib2: disable ODP because it's not supported for DevX QP
[1650638546.114138] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x2005060 [id=61 ref 1] uct_ib_async_event_handler() to hash
[1650638546.114162] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 61 events 0x1 mode thread_spinlock
[1650638546.114167] [ndv4:69197:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib2' (InfiniBand channel adapter) with 1 ports
[1650638546.114466] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638546.114478] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638546.114486] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638546.114550] [ndv4:69197:0] ib_md.c:1319 UCX DEBUG mlx5_ib2: using registration cache
[1650638546.115179] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x466b050: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 6)
[1650638546.115882] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x466b050: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 7)
[1650638546.115888] [ndv4:69188:0] 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> }
[1650638546.115916] [ndv4:69188:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638546.115921] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x40293a0 [id=117 ref 1] uct_ud_iface_async_handler() to hash
[1650638546.115941] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 117 events 0x5 mode thread_spinlock
[1650638546.115951] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[30]=0x466b050 using ud_mlx5/mlx5_ib5:1 on worker 0x1c8f6f0
[1650638546.116037] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638546.116043] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638546.116168] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638546.116172] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638546.116510] [ndv4:69197:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638546.116519] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638546.116793] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib6:1
[1650638546.116779] [ndv4:69197:0] ib_md.c:1604 UCX DEBUG mlx5_ib2: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638546.116988] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638546.116994] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638546.118844] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650638546.118862] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638546.118865] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638546.118914] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638546.120154] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638546.120160] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650638546.121319] [ndv4:69188:0] ib_iface.c:994 UCX DEBUG iface=0x476b0a0: created RC QP 0x2b0c on mlx5_ib6:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650638546.122351] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[31]=0x476b0a0 using rc_verbs/mlx5_ib6:1 on worker 0x1c8f6f0
[1650638546.122389] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638546.122396] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638546.122486] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638546.122490] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638546.122831] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib6:1
[1650638546.124010] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638546.124025] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638546.124028] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638546.124078] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638546.124617] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x4a8c010 of 8176 bytes with 127 elements
[1650638546.124868] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638546.124875] [ndv4:69188:0] 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> }
[1650638546.124910] [ndv4:69188:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib6 length=2048) failed: Invalid argument
[1650638546.124914] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638546.124925] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x47462d0 [id=120 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638546.124946] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 120 events 0x1 mode thread_spinlock
[1650638546.124956] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[32]=0x48b8030 using rc_mlx5/mlx5_ib6:1 on worker 0x1c8f6f0
[1650638546.125065] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638546.125071] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638546.125202] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638546.125227] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638546.126247] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib6:1
[1650638546.127934] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638546.127957] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638546.127962] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638546.128016] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638546.128424] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650638546.128435] [ndv4:69188:0] 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> }
[1650638546.128472] [ndv4:69188:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib6 length=2048) failed: Invalid argument
[1650638546.128476] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638546.128487] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x48c0f60 [id=122 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638546.128514] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 122 events 0x1 mode thread_spinlock
[1650638546.129198] [ndv4:69188:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650638546.136976] [ndv4:69188:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x4a8e050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0x2b13
[1650638546.137295] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638546.137303] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638546.137315] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2ae4e443c008 of 151544 bytes with 1052 elements
[1650638546.138748] [ndv4:69197:0] topo.c:99 UCX DEBUG bus id 0x103000000 doesn't exist. sys_dev = 2
[1650638546.138757] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638546.141139] [ndv4:69188:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae4e1e00000..0x2ae4e4400000 on mlx5_ib6 lkey 0x80b00 rkey 0x80b00 access 0xf flags 0x3e4
[1650638546.141159] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2ae4e1e00018 of 39845864 bytes with 4752 elements
[1650638546.141361] [ndv4:69188:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x4a8e050
[1650638546.141397] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[33]=0x4a8e050 using dc_mlx5/mlx5_ib6:1 on worker 0x1c8f6f0
[1650638546.141486] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638546.141495] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638546.141559] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638546.141563] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638546.141939] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib6:1
[1650638546.142904] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638546.143309] [ndv4:69188:0] ib_iface.c:994 UCX DEBUG iface=0x4c6a060: created UD QP 0x2b18 on mlx5_ib6:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638546.143847] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638546.151872] [ndv4:69197:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638546.151880] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638546.155343] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638546.155351] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638546.155424] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638546.155429] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638546.155824] [ndv4:69188:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae4e4461000..0x2ae4e44e6000 on mlx5_ib6 lkey 0x80c00 rkey 0x80c00 access 0xf flags 0x3e4
[1650638546.155830] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae4e4461018 of 544744 bytes with 128 elements
[1650638546.155834] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638546.156297] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x4c6a060: adding gid fe80::15:5dff:fd34:21 to hash on device mlx5_ib6 port 1 index 0)
[1650638546.156692] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x4c6a060: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 1)
[1650638546.156935] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x4c6a060: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 2)
[1650638546.156996] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x4c6a060: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 3)
[1650638546.160738] [ndv4:69197:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638546.160746] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638546.161803] [ndv4:69197:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638546.161807] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638546.164174] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x4c6a060: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 4)
[1650638546.164944] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x4c6a060: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 5)
[1650638546.169893] [ndv4:69197:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638546.169901] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638546.170282] [ndv4:69197:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638546.180669] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x4c6a060: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 6)
[1650638546.180977] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x4c6a060: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 7)
[1650638546.181334] [ndv4:69188:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638546.181343] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x4c6af20 [id=123 ref 1] uct_ud_iface_async_handler() to hash
[1650638546.181369] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 123 events 0x5 mode thread_spinlock
[1650638546.181379] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[34]=0x4c6a060 using ud_verbs/mlx5_ib6:1 on worker 0x1c8f6f0
[1650638546.181513] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638546.181519] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638546.181646] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638546.181651] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638546.182299] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib6:1
[1650638546.183698] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638546.184579] [ndv4:69188:0] ib_iface.c:994 UCX DEBUG iface=0x3f7b3a0: created UD QP 0x2b19 on mlx5_ib6:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638546.184586] [ndv4:69188:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650638546.185237] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638546.185436] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638546.185441] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638546.185456] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638546.185460] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638546.185876] [ndv4:69188:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae4e44e6000..0x2ae4e456b000 on mlx5_ib6 lkey 0x80d00 rkey 0x80d00 access 0xf flags 0x3e4
[1650638546.185882] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae4e44e6018 of 544744 bytes with 128 elements
[1650638546.185886] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638546.186238] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3f7b3a0: adding gid fe80::15:5dff:fd34:21 to hash on device mlx5_ib6 port 1 index 0)
[1650638546.186487] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3f7b3a0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 1)
[1650638546.186767] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3f7b3a0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 2)
[1650638546.186825] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3f7b3a0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 3)
[1650638546.187166] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3f7b3a0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 4)
[1650638546.187541] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3f7b3a0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 5)
[1650638546.187822] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3f7b3a0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 6)
[1650638546.187836] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x3f7b3a0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 7)
[1650638546.187842] [ndv4:69188:0] 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> }
[1650638546.187871] [ndv4:69188:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638546.187875] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x47460e0 [id=124 ref 1] uct_ud_iface_async_handler() to hash
[1650638546.187893] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 124 events 0x5 mode thread_spinlock
[1650638546.187901] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[35]=0x3f7b3a0 using ud_mlx5/mlx5_ib6:1 on worker 0x1c8f6f0
[1650638546.188027] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638546.188032] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638546.188155] [ndv4:69197:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib3 vendor_id: 0x15b3 device_id: 4124
[1650638546.188568] [ndv4:69197:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib3: disable ODP because it's not supported for DevX QP
[1650638546.188380] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638546.188388] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638546.189975] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x2714020 [id=63 ref 1] uct_ib_async_event_handler() to hash
[1650638546.190001] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 63 events 0x1 mode thread_spinlock
[1650638546.190005] [ndv4:69197:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib3' (InfiniBand channel adapter) with 1 ports
[1650638546.196453] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib7:1
[1650638546.197440] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638546.197452] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638546.197460] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638546.197523] [ndv4:69197:0] ib_md.c:1319 UCX DEBUG mlx5_ib3: using registration cache
[1650638546.198290] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650638546.198307] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638546.198310] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638546.198361] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638546.198939] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638546.198945] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650638546.200103] [ndv4:69188:0] ib_iface.c:994 UCX DEBUG iface=0x4e880a0: created RC QP 0x2af4 on mlx5_ib7:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650638546.200980] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[36]=0x4e880a0 using rc_verbs/mlx5_ib7:1 on worker 0x1c8f6f0
[1650638546.201095] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638546.201101] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638546.201198] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638546.201202] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638546.201454] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib7:1
[1650638546.209750] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638546.209765] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638546.209768] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638546.209819] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638546.210287] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x51a9010 of 8176 bytes with 127 elements
[1650638546.210541] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638546.210548] [ndv4:69188:0] 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> }
[1650638546.210618] [ndv4:69188:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib7 length=2048) failed: Invalid argument
[1650638546.210623] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638546.210635] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x4d888c0 [id=127 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638546.210654] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 127 events 0x1 mode thread_spinlock
[1650638546.210667] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[37]=0x4fd5030 using rc_mlx5/mlx5_ib7:1 on worker 0x1c8f6f0
[1650638546.210899] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638546.210905] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638546.211200] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638546.211221] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638546.211688] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib7:1
[1650638546.213164] [ndv4:69197:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638546.213173] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638546.213540] [ndv4:69197:0] ib_md.c:1604 UCX DEBUG mlx5_ib3: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638546.213651] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638546.213658] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638546.221071] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638546.221088] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638546.221091] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638546.221147] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638546.221559] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650638546.221567] [ndv4:69188:0] 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> }
[1650638546.221636] [ndv4:69188:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib7 length=2048) failed: Invalid argument
[1650638546.221640] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638546.221648] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x4e63760 [id=129 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638546.221666] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 129 events 0x1 mode thread_spinlock
[1650638546.224117] [ndv4:69197:0] topo.c:99 UCX DEBUG bus id 0x104000000 doesn't exist. sys_dev = 3
[1650638546.224125] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638546.234025] [ndv4:69197:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638546.234032] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638546.267757] [ndv4:69197:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638546.267764] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638546.289829] [ndv4:69197:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638546.289836] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638546.291563] [ndv4:69197:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638546.291569] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638546.291867] [ndv4:69197:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638546.309070] [ndv4:68756:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650638546.313956] [ndv4:69188:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650638546.315996] [ndv4:69197:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib4 vendor_id: 0x15b3 device_id: 4124
[1650638546.316463] [ndv4:69197:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib4: disable ODP because it's not supported for DevX QP
[1650638546.323644] [ndv4:68756:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x5224050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0x2afb
[1650638546.324256] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638546.324265] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638546.324283] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b13b3ba0008 of 151544 bytes with 1052 elements
[1650638546.322417] [ndv4:69188:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x51ab050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0x2afc
[1650638546.322563] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638546.322570] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638546.322580] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2ae4e6d6c008 of 151544 bytes with 1052 elements
[1650638546.326551] [ndv4:69188:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae4e4600000..0x2ae4e6c00000 on mlx5_ib7 lkey 0x80400 rkey 0x80400 access 0xf flags 0x3e4
[1650638546.326574] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2ae4e4600018 of 39845864 bytes with 4752 elements
[1650638546.326708] [ndv4:69188:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x51ab050
[1650638546.326747] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[38]=0x51ab050 using dc_mlx5/mlx5_ib7:1 on worker 0x1c8f6f0
[1650638546.326971] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638546.326982] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638546.327156] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638546.327160] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638546.327787] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib7:1
[1650638546.327919] [ndv4:68756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b13b1400000..0x2b13b3a00000 on mlx5_ib7 lkey 0x80500 rkey 0x80500 access 0xf flags 0x3e4
[1650638546.327942] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b13b1400018 of 39845864 bytes with 4752 elements
[1650638546.328082] [ndv4:68756:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x5224050
[1650638546.328157] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[38]=0x5224050 using dc_mlx5/mlx5_ib7:1 on worker 0x1d084d0
[1650638546.328302] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638546.328314] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638546.328428] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638546.328433] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638546.332316] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x27136d0 [id=65 ref 1] uct_ib_async_event_handler() to hash
[1650638546.332338] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 65 events 0x1 mode thread_spinlock
[1650638546.332342] [ndv4:69197:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib4' (InfiniBand channel adapter) with 1 ports
[1650638546.332648] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638546.332656] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638546.332662] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638546.332712] [ndv4:69197:0] ib_md.c:1319 UCX DEBUG mlx5_ib4: using registration cache
[1650638546.333944] [ndv4:69197:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638546.333950] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638546.334278] [ndv4:69197:0] ib_md.c:1604 UCX DEBUG mlx5_ib4: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638546.334318] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638546.334324] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638546.335600] [ndv4:69197:0] topo.c:99 UCX DEBUG bus id 0x105000000 doesn't exist. sys_dev = 4
[1650638546.335604] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638546.343743] [ndv4:69197:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650638546.343752] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638546.343760] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638546.344189] [ndv4:69188:0] ib_iface.c:994 UCX DEBUG iface=0x4698570: created UD QP 0x2b09 on mlx5_ib7:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638546.344771] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638546.345035] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638546.345042] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638546.345413] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638546.345419] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638546.345903] [ndv4:69188:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae4e6d91000..0x2ae4e6e16000 on mlx5_ib7 lkey 0x80600 rkey 0x80600 access 0xf flags 0x3e4
[1650638546.345909] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae4e6d91018 of 544744 bytes with 128 elements
[1650638546.345913] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638546.346497] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x4698570: adding gid fe80::15:5dff:fd34:22 to hash on device mlx5_ib7 port 1 index 0)
[1650638546.347103] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x4698570: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 1)
[1650638546.345593] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib7:1
[1650638546.346958] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638546.347313] [ndv4:68756:0] ib_iface.c:994 UCX DEBUG iface=0x47114e0: created UD QP 0x2b0a on mlx5_ib7:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638546.346949] [ndv4:69197:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650638546.346955] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638546.348049] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x4698570: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 2)
[1650638546.349191] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x4698570: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 3)
[1650638546.347762] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638546.348138] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638546.348147] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638546.348336] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638546.348341] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638546.348701] [ndv4:68756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b13b3bc5000..0x2b13b3c4a000 on mlx5_ib7 lkey 0x80700 rkey 0x80700 access 0xf flags 0x3e4
[1650638546.348707] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b13b3bc5018 of 544744 bytes with 128 elements
[1650638546.348712] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638546.349464] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x4698570: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 4)
[1650638546.349507] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x4698570: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 5)
[1650638546.349738] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x4698570: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 6)
[1650638546.349907] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x4698570: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 7)
[1650638546.350191] [ndv4:69188:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638546.350200] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x4698fc0 [id=130 ref 1] uct_ud_iface_async_handler() to hash
[1650638546.350269] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 130 events 0x5 mode thread_spinlock
[1650638546.350283] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[39]=0x4698570 using ud_verbs/mlx5_ib7:1 on worker 0x1c8f6f0
[1650638546.350385] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638546.350390] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638546.350461] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638546.350466] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638546.350615] [ndv4:69188:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib7:1
[1650638546.351625] [ndv4:69188:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638546.351943] [ndv4:69188:0] ib_iface.c:994 UCX DEBUG iface=0x54a5460: created UD QP 0x2b0b on mlx5_ib7:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638546.351952] [ndv4:69188:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650638546.352436] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638546.352579] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638546.352585] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638546.352598] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638546.352602] [ndv4:69188:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638546.352935] [ndv4:69188:0] ib_md.c:812 UCX DEBUG registered memory 0x2ae4e6e16000..0x2ae4e6e9b000 on mlx5_ib7 lkey 0x80800 rkey 0x80800 access 0xf flags 0x3e4
[1650638546.352942] [ndv4:69188:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ae4e6e16018 of 544744 bytes with 128 elements
[1650638546.352946] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638546.353302] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x54a5460: adding gid fe80::15:5dff:fd34:22 to hash on device mlx5_ib7 port 1 index 0)
[1650638546.353422] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x54a5460: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 1)
[1650638546.357005] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x47114e0: adding gid fe80::15:5dff:fd34:22 to hash on device mlx5_ib7 port 1 index 0)
[1650638546.357439] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x47114e0: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 1)
[1650638546.357977] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x47114e0: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 2)
[1650638546.361920] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x54a5460: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 2)
[1650638546.365423] [ndv4:69197:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650638546.365432] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638546.365905] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x47114e0: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 3)
[1650638546.366550] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x47114e0: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 4)
[1650638546.367061] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x47114e0: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 5)
[1650638546.367341] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x47114e0: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 6)
[1650638546.372593] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x54a5460: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 3)
[1650638546.372830] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x54a5460: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 4)
[1650638546.372971] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x54a5460: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 5)
[1650638546.373118] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x54a5460: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 6)
[1650638546.373149] [ndv4:69188:0] ud_iface.c:393 UCX DEBUG iface 0x54a5460: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 7)
[1650638546.373155] [ndv4:69188:0] 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> }
[1650638546.373185] [ndv4:69188:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638546.373189] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x54a5f80 [id=131 ref 1] uct_ud_iface_async_handler() to hash
[1650638546.373270] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 131 events 0x5 mode thread_spinlock
[1650638546.373281] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[40]=0x54a5460 using ud_mlx5/mlx5_ib7:1 on worker 0x1c8f6f0
[1650638546.373384] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool uct_scopy_iface_tx_mp: align 64, maxelems 4294967295, elemsize 736
[1650638546.373442] [ndv4:69188:0] ucp_worker.c:1159 UCX DEBUG created interface[41]=0x4db5600 using cma/memory on worker 0x1c8f6f0
[1650638546.373448] [ndv4:69188:0] ucp_worker.c:982 UCX DEBUG selected scalable tl bitmap: 0x3ffffffffff 0x0 (42 tls)
[1650638546.378434] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x47114e0: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 7)
[1650638546.378662] [ndv4:68756:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638546.378674] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x4711dd0 [id=130 ref 1] uct_ud_iface_async_handler() to hash
[1650638546.378703] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 130 events 0x5 mode thread_spinlock
[1650638546.379159] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[39]=0x47114e0 using ud_verbs/mlx5_ib7:1 on worker 0x1d084d0
[1650638546.379260] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638546.379267] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638546.379327] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638546.379332] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638546.379594] [ndv4:68756:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib7:1
[1650638546.380485] [ndv4:68756:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638546.380815] [ndv4:68756:0] ib_iface.c:994 UCX DEBUG iface=0x551e460: created UD QP 0x2b0c on mlx5_ib7:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638546.380823] [ndv4:68756:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650638546.381388] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638546.381516] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638546.381522] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638546.381596] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638546.381601] [ndv4:68756:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638546.381962] [ndv4:68756:0] ib_md.c:812 UCX DEBUG registered memory 0x2b13b3c4a000..0x2b13b3ccf000 on mlx5_ib7 lkey 0x80900 rkey 0x80900 access 0xf flags 0x3e4
[1650638546.381968] [ndv4:68756:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b13b3c4a018 of 544744 bytes with 128 elements
[1650638546.381972] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638546.382318] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x551e460: adding gid fe80::15:5dff:fd34:22 to hash on device mlx5_ib7 port 1 index 0)
[1650638546.382642] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x551e460: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 1)
[1650638546.382899] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x551e460: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 2)
[1650638546.392828] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x551e460: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 3)
[1650638546.393161] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x551e460: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 4)
[1650638546.395898] [ndv4:69197:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650638546.395907] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638546.396164] [ndv4:69197:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638546.400794] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x551e460: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 5)
[1650638546.401140] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x551e460: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 6)
[1650638546.401927] [ndv4:68756:0] ud_iface.c:393 UCX DEBUG iface 0x551e460: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 7)
[1650638546.401938] [ndv4:68756:0] 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> }
[1650638546.401972] [ndv4:68756:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638546.401976] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x4edc300 [id=131 ref 1] uct_ud_iface_async_handler() to hash
[1650638546.401998] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 131 events 0x5 mode thread_spinlock
[1650638546.402009] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[40]=0x551e460 using ud_mlx5/mlx5_ib7:1 on worker 0x1d084d0
[1650638546.402100] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool uct_scopy_iface_tx_mp: align 64, maxelems 4294967295, elemsize 736
[1650638546.402149] [ndv4:68756:0] ucp_worker.c:1159 UCX DEBUG created interface[41]=0x4e2e600 using cma/memory on worker 0x1d084d0
[1650638546.402155] [ndv4:68756:0] ucp_worker.c:982 UCX DEBUG selected scalable tl bitmap: 0x3ffffffffff 0x0 (42 tls)
[1650638546.405799] [ndv4:69197:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib5 vendor_id: 0x15b3 device_id: 4124
[1650638546.406155] [ndv4:69197:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib5: disable ODP because it's not supported for DevX QP
[1650638546.413832] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x201bf70 [id=67 ref 1] uct_ib_async_event_handler() to hash
[1650638546.413853] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 67 events 0x1 mode thread_spinlock
[1650638546.413857] [ndv4:69197:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib5' (InfiniBand channel adapter) with 1 ports
[1650638546.414067] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638546.414079] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638546.414085] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638546.414132] [ndv4:69197:0] ib_md.c:1319 UCX DEBUG mlx5_ib5: using registration cache
[1650638546.414922] [ndv4:69197:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638546.414928] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638546.415198] [ndv4:69197:0] ib_md.c:1604 UCX DEBUG mlx5_ib5: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638546.415435] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638546.415445] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638546.424969] [ndv4:69197:0] topo.c:99 UCX DEBUG bus id 0x106000000 doesn't exist. sys_dev = 5
[1650638546.425003] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638546.466183] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x5387fb0 [id=75 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.466263] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 75 events 0x0 mode thread_spinlock
[1650638546.466359] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x4711f90 [id=75 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.466405] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 75 events 0x0 mode thread_spinlock
[1650638546.466602] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x145c290 [id=76 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.466644] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 76 events 0x0 mode thread_spinlock
[1650638546.466761] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x145c2d0 [id=77 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.466780] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 77 events 0x0 mode thread_spinlock
[1650638546.466806] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x145c310 [id=79 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.466825] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 79 events 0x0 mode thread_spinlock
[1650638546.466879] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x145c350 [id=83 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.466895] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 83 events 0x0 mode thread_spinlock
[1650638546.466911] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x145c390 [id=84 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.466938] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 84 events 0x0 mode thread_spinlock
[1650638546.466953] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x145c3d0 [id=86 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.466972] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 86 events 0x0 mode thread_spinlock
[1650638546.467008] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x145c410 [id=90 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.467024] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 90 events 0x0 mode thread_spinlock
[1650638546.467040] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x145c450 [id=91 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.467058] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 91 events 0x0 mode thread_spinlock
[1650638546.467078] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x142dd20 [id=93 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.467096] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 93 events 0x0 mode thread_spinlock
[1650638546.467140] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x4e2ef60 [id=97 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.467161] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 97 events 0x0 mode thread_spinlock
[1650638546.467181] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x4e2efa0 [id=98 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.467201] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 98 events 0x0 mode thread_spinlock
[1650638546.467274] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x3affee0 [id=100 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.467294] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 100 events 0x0 mode thread_spinlock
[1650638546.467331] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x3afff20 [id=104 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.467348] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 104 events 0x0 mode thread_spinlock
[1650638546.467365] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x3afff60 [id=105 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.467379] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 105 events 0x0 mode thread_spinlock
[1650638546.467393] [ndv4:68756:0] async.c:228 UCX DEBUG ad[1650638546.466532] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x4e63370 [id=76 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.466558] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 76 events 0x0 mode thread_spinlock
[1650638546.466617] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x13b6a70 [id=77 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.466634] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 77 events 0x0 mode thread_spinlock
[1650638546.466651] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x13b6ab0 [id=79 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.466671] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 79 events 0x0 mode thread_spinlock
[1650638546.466717] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x13b6af0 [id=83 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.466733] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 83 events 0x0 mode thread_spinlock
[1650638546.466747] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x13b6b30 [id=84 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.466762] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 84 events 0x0 mode thread_spinlock
[1650638546.466776] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x13b6b70 [id=86 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.466793] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 86 events 0x0 mode thread_spinlock
[1650638546.466827] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x1cafdc0 [id=90 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.466844] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 90 events 0x0 mode thread_spinlock
[1650638546.466859] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x1cafe00 [id=91 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.466873] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 91 events 0x0 mode thread_spinlock
[1650638546.466888] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x1cafe40 [id=93 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.466903] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 93 events 0x0 mode thread_spinlock
[1650638546.466937] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x4db5f60 [id=97 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.466951] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 97 events 0x0 mode thread_spinlock
[1650638546.466966] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x4db5fa0 [id=98 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.466979] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 98 events 0x0 mode thread_spinlock
[1650638546.466993] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0xc51820 [id=100 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.467006] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 100 events 0x0 mode thread_spinlock
[1650638546.467041] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0xc51860 [id=104 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.467056] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 104 events 0x0 mode thread_spinlock
[1650638546.467070] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0xc518a0 [id=105 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.467083] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 105 events 0x0 mode thread_spinlock
[1650638546.467097] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0xc518e0 [id=107 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.467109] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 107 events 0x0 mode thread_spinlock
[1650638546.467143] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x1cafe80 [id=111 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.467158] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 111 events 0x0 mode thread_spinlock
[1650638546.467581] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x1cafec0 [id=112 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.467597] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 112 events 0x0 mode thread_spinlock
[1650638546.467612] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x1caff00 [id=114 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.467626] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 114 events 0x0 mode thread_spinlock
[1650638546.467666] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x1caff40 [id=118 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.467682] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 118 events 0x0 mode thread_spinlock
[1650638546.467697] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x1caff80 [id=119 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.467710] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 119 events 0x0 mode thread_spinlock
[1650638546.467725] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x4fdde10 [id=121 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.467739] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 121 events 0x0 mode thread_spinlock
[1650638546.467774] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x4fdde50 [id=125 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.467787] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 125 events 0x0 mode thread_spinlock
[1650638546.467800] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x4fdde90 [id=126 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.467815] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 126 events 0x0 mode thread_spinlock
[1650638546.467828] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x4fdded0 [id=128 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.467844] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 128 events 0x0 mode thread_spinlock
ded async handler 0x3afffa0 [id=107 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.467408] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 107 events 0x0 mode thread_spinlock
[1650638546.467444] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x142dd90 [id=111 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.467460] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 111 events 0x0 mode thread_spinlock
[1650638546.467556] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x142ddd0 [id=112 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.467572] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 112 events 0x0 mode thread_spinlock
[1650638546.467589] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x142de10 [id=114 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.467608] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 114 events 0x0 mode thread_spinlock
[1650638546.467647] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x142de50 [id=118 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.467663] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 118 events 0x0 mode thread_spinlock
[1650638546.467679] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x142de90 [id=119 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.467694] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 119 events 0x0 mode thread_spinlock
[1650638546.467709] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x142ded0 [id=121 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.467723] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 121 events 0x0 mode thread_spinlock
[1650638546.467760] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x5056e10 [id=125 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.467775] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 125 events 0x0 mode thread_spinlock
[1650638546.467790] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x5056e50 [id=126 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.467806] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 126 events 0x0 mode thread_spinlock
[1650638546.467821] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x5056e90 [id=128 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638546.467836] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 128 events 0x0 mode thread_spinlock
[1650638546.470784] [ndv4:69197:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650638546.470794] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638546.472824] [ndv4:69188:0] async.c:228 UCX DEBUG added async handler 0x31147e0 [id=132 ref 1] uct_rdmacm_cm_event_handler() to hash
[1650638546.472847] [ndv4:69188:0] async.c:506 UCX DEBUG listening to async event fd 132 events 0x1 mode thread_spinlock
[1650638546.472932] [ndv4:69188:0] rdmacm_cm.c:922 UCX DEBUG created rdmacm_cm 0x5580050 with event_channel 0x4db5fe0 (fd=132)
[1650638546.472964] [ndv4:69188:0] tcp_sockcm.c:186 UCX DEBUG created tcp_sockcm 0x4a6a7e0
[1650638546.473138] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool ucp_requests: align 64, maxelems 4294967295, elemsize 440
[1650638546.473144] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool ucp_rkeys: align 64, maxelems 4294967295, elemsize 168
[1650638546.473148] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool ucp_am_bufs: align 64, maxelems 4294967295, elemsize 8344
[1650638546.473151] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool ucp_reg_bufs: align 64, maxelems 4294967295, elemsize 8208
[1650638546.473154] [ndv4:69188:0] mpool.c:88 UCX DEBUG mpool ucp_rndv_frags: align 512, maxelems 4294967295, elemsize 524304
[1650638546.473062] [ndv4:68756:0] async.c:228 UCX DEBUG added async handler 0x318d7e0 [id=132 ref 1] uct_rdmacm_cm_event_handler() to hash
[1650638546.473087] [ndv4:68756:0] async.c:506 UCX DEBUG listening to async event fd 132 events 0x1 mode thread_spinlock
[1650638546.473109] [ndv4:68756:0] rdmacm_cm.c:922 UCX DEBUG created rdmacm_cm 0x55f9050 with event_channel 0x4e2efe0 (fd=132)
[1650638546.473145] [ndv4:68756:0] tcp_sockcm.c:186 UCX DEBUG created tcp_sockcm 0x4ae37e0
[1650638546.473165] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool ucp_requests: align 64, maxelems 4294967295, elemsize 440
[1650638546.473168] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool ucp_rkeys: align 64, maxelems 4294967295, elemsize 168
[1650638546.473172] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool ucp_am_bufs: align 64, maxelems 4294967295, elemsize 8344
[1650638546.473175] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool ucp_reg_bufs: align 64, maxelems 4294967295, elemsize 8208
[1650638546.473177] [ndv4:68756:0] mpool.c:88 UCX DEBUG mpool ucp_rndv_frags: align 512, maxelems 4294967295, elemsize 524304
[1650638546.473423] [ndv4:69188:0] parser.c:1893 UCX INFO UCX_* env variables: UCX_TLS=sysv,cma,ib UCX_POSIX_USE_PROC_LINK=n UCX_LOG_LEVEL=debug
[1650638546.473444] [ndv4:68756:0] parser.c:1893 UCX INFO UCX_* env variables: UCX_TLS=sysv,cma,ib UCX_POSIX_USE_PROC_LINK=n UCX_LOG_LEVEL=debug
[1650638546.480843] [ndv4:69197:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650638546.480850] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638546.499582] [ndv4:69197:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650638546.499590] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638546.501321] [ndv4:69197:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650638546.501327] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638546.501604] [ndv4:69197:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638546.527308] [ndv4:69197:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib6 vendor_id: 0x15b3 device_id: 4124
[1650638546.527664] [ndv4:69197:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib6: disable ODP because it's not supported for DevX QP
[1650638546.528127] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x2714ec0 [id=69 ref 1] uct_ib_async_event_handler() to hash
[1650638546.528149] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 69 events 0x1 mode thread_spinlock
[1650638546.528152] [ndv4:69197:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib6' (InfiniBand channel adapter) with 1 ports
[1650638546.528388] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638546.528398] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638546.528403] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638546.528452] [ndv4:69197:0] ib_md.c:1319 UCX DEBUG mlx5_ib6: using registration cache
[1650638546.529048] [ndv4:69197:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638546.529053] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638546.529320] [ndv4:69197:0] ib_md.c:1604 UCX DEBUG mlx5_ib6: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638546.529380] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638546.529386] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638546.539397] [ndv4:69197:0] topo.c:99 UCX DEBUG bus id 0x107000000 doesn't exist. sys_dev = 6
[1650638546.539404] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650638546.540304] [ndv4:69197:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650638546.540308] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650638546.549308] [ndv4:69197:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650638546.549315] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650638546.567594] [ndv4:69197:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650638546.567601] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650638546.576885] [ndv4:69197:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650638546.576891] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650638546.577114] [ndv4:69197:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638546.586384] [ndv4:69197:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib7 vendor_id: 0x15b3 device_id: 4124
[1650638546.586718] [ndv4:69197:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib7: disable ODP because it's not supported for DevX QP
[1650638546.587430] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x202f7b0 [id=71 ref 1] uct_ib_async_event_handler() to hash
[1650638546.587464] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 71 events 0x1 mode thread_spinlock
[1650638546.587469] [ndv4:69197:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib7' (InfiniBand channel adapter) with 1 ports
[1650638546.587771] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638546.587782] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638546.587795] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638546.587874] [ndv4:69197:0] ib_md.c:1319 UCX DEBUG mlx5_ib7: using registration cache
[1650638546.597542] [ndv4:69197:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638546.597553] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638546.597814] [ndv4:69197:0] ib_md.c:1604 UCX DEBUG mlx5_ib7: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638546.597985] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638546.597990] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638546.607306] [ndv4:69197:0] topo.c:99 UCX DEBUG bus id 0x108000000 doesn't exist. sys_dev = 7
[1650638546.607314] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650638546.607882] [ndv4:69197:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650638546.607885] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650638546.636403] [ndv4:69197:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650638546.636410] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650638546.648221] [ndv4:69197:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650638546.648229] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650638546.657508] [ndv4:68758:0] debug.c:1198 UCX DEBUG using signal stack 0x2ad1de44a000 size 141824
[1650638546.657583] [ndv4:68758:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650638546.657600] [ndv4:68758:0] 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 0x2ad1de29e000
[1650638546.657622] [ndv4:68758:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650638546.657630] [ndv4:68758:0] 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
[1650638546.657636] [ndv4:68758:0] module.c:253 UCX DEBUG loading modules for ucs
[1650638546.659943] [ndv4:68758:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650638546.659964] [ndv4:68758:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650638546.659997] [ndv4:68758:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650638546.660000] [ndv4:68758:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650638546.660006] [ndv4:68758:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650638546.660014] [ndv4:68758:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650638546.660016] [ndv4:68758:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650638546.660021] [ndv4:68758:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650638546.660023] [ndv4:68758:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650638546.660025] [ndv4:68758:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650638546.660028] [ndv4:68758:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650638546.660031] [ndv4:68758:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650638546.660039] [ndv4:68758:0] module.c:253 UCX DEBUG loading modules for uct
[1650638546.661888] [ndv4:69197:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650638546.661903] [ndv4:69197:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650638546.662085] [ndv4:69197:0] ucp_context.c:1556 UCX DEBUG created ucp context 0x2017190 0x2017190 [10 mds 42 tls] features 0x1 tl bitmap 0x3ffffffffff 0x0
[1650638546.665502] [ndv4:68758:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650638546.665859] [ndv4:68758:0] ucp_context.c:1117 UCX DEBUG closing md posix because it has no selected transport resources
[1650638546.665882] [ndv4:68758:0] ucp_context.c:1117 UCX DEBUG closing md self because it has no selected transport resources
[1650638546.666108] [ndv4:68758:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650638546.666120] [ndv4:68758:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650638546.666130] [ndv4:68758:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650638546.666140] [ndv4:68758:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650638546.666150] [ndv4:68758:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650638546.666161] [ndv4:68758:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650638546.666171] [ndv4:68758:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650638546.666234] [ndv4:68758:0] ucp_context.c:1117 UCX DEBUG closing md tcp because it has no selected transport resources
[1650638546.666589] [ndv4:68758:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650638546.698808] [ndv4:68758:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650638546.699432] [ndv4:68758:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650638546.707387] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x1a25680 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650638546.707490] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650638546.707784] [ndv4:68758:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650638546.708068] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638546.708079] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638546.708105] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638546.708170] [ndv4:68758:0] module.c:253 UCX DEBUG loading modules for ucm
[1650638546.708192] [ndv4:68758:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650638546.708867] [ndv4:68758:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638546.708877] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638546.709187] [ndv4:68758:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638546.709284] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638546.709290] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638546.729378] [ndv4:69197:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 8447 bytes with hugetlb
[1650638546.729487] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool mm_recv_desc: align 64, maxelems 4294967295, elemsize 8368
[1650638546.729517] [ndv4:69197:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 4292720 bytes with hugetlb
[1650638546.729527] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool mm_recv_desc: allocated chunk 0x2afd85462018 of 4296680 bytes with 512 elements
[1650638546.730099] [ndv4:69197:0] mm_iface.c:600 UCX DEBUG created mm iface 0x27931a0 FIFO id 0xd8006 va 0x2afd7fc8d000 size 12288 (128 x 64 elems)
[1650638546.730155] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[0]=0x27931a0 using sysv/memory on worker 0x306d6f0
[1650638546.730655] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638546.730666] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638546.730854] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638546.730860] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638546.731764] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib0:1
[1650638546.731241] [ndv4:68758:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650638546.731249] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638546.733583] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650638546.733641] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638546.733646] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638546.733709] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638546.734407] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638546.734419] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650638546.735836] [ndv4:69197:0] ib_iface.c:994 UCX DEBUG iface=0x279e440: created RC QP 0x30df on mlx5_ib0:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650638546.738660] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[1]=0x279e440 using rc_verbs/mlx5_ib0:1 on worker 0x306d6f0
[1650638546.738746] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638546.738753] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638546.738959] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638546.738964] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638546.739523] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib0:1
[1650638546.739579] [ndv4:69197:0] ib_device.c:1394 UCX DEBUG max IB CQE size is 128
[1650638546.740568] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638546.740580] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638546.740583] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638546.740634] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638546.741012] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x3394010 of 8176 bytes with 127 elements
[1650638546.741293] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638546.741329] [ndv4:69197:0] 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> }
[1650638546.741381] [ndv4:69197:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650638546.741387] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638546.750642] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x2703ae0 [id=78 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638546.750671] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 78 events 0x1 mode thread_spinlock
[1650638546.751240] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[2]=0x27abfd0 using rc_mlx5/mlx5_ib0:1 on worker 0x306d6f0
[1650638546.751264] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638546.751271] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638546.751448] [ndv4:68758:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638546.751458] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638546.752321] [ndv4:68758:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638546.752325] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638546.753088] [ndv4:68758:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638546.753092] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638546.763575] [ndv4:68758:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638546.763582] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638546.763824] [ndv4:68758:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638546.768005] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638546.768016] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638546.768824] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib0:1
[1650638546.771914] [ndv4:68758:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650638546.772286] [ndv4:68758:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650638546.772583] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x19f9080 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650638546.772605] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650638546.772608] [ndv4:68758:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650638546.772794] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638546.772803] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638546.772809] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638546.772849] [ndv4:68758:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650638546.773287] [ndv4:68758:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638546.773292] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638546.773572] [ndv4:68758:0] ib_md.c:1604 UCX DEBUG mlx5_ib1: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638546.773587] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638546.773592] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638546.773728] [ndv4:68758:0] topo.c:99 UCX DEBUG bus id 0x102000000 doesn't exist. sys_dev = 1
[1650638546.773732] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638546.773850] [ndv4:68758:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638546.773854] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638546.773953] [ndv4:68758:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638546.773956] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638546.774063] [ndv4:68758:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638546.774066] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638546.774166] [ndv4:68758:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638546.774169] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638546.774406] [ndv4:68758:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638546.782928] [ndv4:68758:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib2 vendor_id: 0x15b3 device_id: 4124
[1650638546.783242] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638546.783252] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638546.783255] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638546.783270] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638546.783343] [ndv4:68758:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib2: disable ODP because it's not supported for DevX QP
[1650638546.783651] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650638546.783660] [ndv4:69197:0] 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> }
[1650638546.783693] [ndv4:69197:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650638546.783697] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638546.783707] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x26fa5b0 [id=80 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638546.783730] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 80 events 0x1 mode thread_spinlock
[1650638546.784454] [ndv4:69197:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650638546.791676] [ndv4:69197:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x3396050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0x3105
[1650638546.793878] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x1a1dd50 [id=61 ref 1] uct_ib_async_event_handler() to hash
[1650638546.793903] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 61 events 0x1 mode thread_spinlock
[1650638546.793907] [ndv4:68758:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib2' (InfiniBand channel adapter) with 1 ports
[1650638546.794108] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638546.794117] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638546.794122] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638546.794162] [ndv4:68758:0] ib_md.c:1319 UCX DEBUG mlx5_ib2: using registration cache
[1650638546.794899] [ndv4:68758:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638546.794905] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638546.795187] [ndv4:68758:0] ib_md.c:1604 UCX DEBUG mlx5_ib2: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638546.795261] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638546.795269] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638546.800871] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638546.800881] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638546.800906] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2afd7fc92008 of 151544 bytes with 1052 elements
[1650638546.804918] [ndv4:69197:0] ib_md.c:812 UCX DEBUG registered memory 0x2afd85a00000..0x2afd88000000 on mlx5_ib0 lkey 0x80600 rkey 0x80600 access 0xf flags 0x3e4
[1650638546.804963] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2afd85a00018 of 39845864 bytes with 4752 elements
[1650638546.805286] [ndv4:69197:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x3396050
[1650638546.805338] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[3]=0x3396050 using dc_mlx5/mlx5_ib0:1 on worker 0x306d6f0
[1650638546.805406] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638546.805416] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638546.805526] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638546.805531] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638546.806071] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib0:1
[1650638546.805538] [ndv4:68758:0] topo.c:99 UCX DEBUG bus id 0x103000000 doesn't exist. sys_dev = 2
[1650638546.805556] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638546.806340] [ndv4:68758:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638546.806345] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638546.806602] [ndv4:68758:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638546.806605] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638546.807296] [ndv4:68758:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638546.807301] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638546.807449] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638546.807816] [ndv4:69197:0] ib_iface.c:994 UCX DEBUG iface=0x27a9610: created UD QP 0x30e8 on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638546.808346] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638546.808422] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638546.808431] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638546.808523] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638546.808528] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638546.808867] [ndv4:69197:0] ib_md.c:812 UCX DEBUG registered memory 0x2afd7fcb7000..0x2afd7fd3c000 on mlx5_ib0 lkey 0x80700 rkey 0x80700 access 0xf flags 0x3e4
[1650638546.808881] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2afd7fcb7018 of 544744 bytes with 128 elements
[1650638546.808894] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638546.809539] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x27a9610: adding gid fe80::15:5dff:fd34:1b to hash on device mlx5_ib0 port 1 index 0)
[1650638546.809940] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x27a9610: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650638546.810315] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x27a9610: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650638546.810429] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x27a9610: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650638546.810721] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x27a9610: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650638546.811071] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x27a9610: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650638546.818345] [ndv4:68758:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638546.818355] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638546.818653] [ndv4:68758:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638546.821377] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x27a9610: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650638546.821576] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x27a9610: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650638546.821946] [ndv4:69197:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638546.829309] [ndv4:68758:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib3 vendor_id: 0x15b3 device_id: 4124
[1650638546.829685] [ndv4:68758:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib3: disable ODP because it's not supported for DevX QP
[1650638546.829995] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x1a1dc40 [id=63 ref 1] uct_ib_async_event_handler() to hash
[1650638546.830025] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 63 events 0x1 mode thread_spinlock
[1650638546.830029] [ndv4:68758:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib3' (InfiniBand channel adapter) with 1 ports
[1650638546.830243] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638546.830256] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638546.830263] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638546.830317] [ndv4:68758:0] ib_md.c:1319 UCX DEBUG mlx5_ib3: using registration cache
[1650638546.830797] [ndv4:68758:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638546.830803] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638546.830184] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x279ad00 [id=81 ref 1] uct_ud_iface_async_handler() to hash
[1650638546.830283] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 81 events 0x5 mode thread_spinlock
[1650638546.830417] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[4]=0x27a9610 using ud_verbs/mlx5_ib0:1 on worker 0x306d6f0
[1650638546.830453] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638546.830461] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638546.830534] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638546.830538] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638546.830782] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib0:1
[1650638546.831065] [ndv4:68758:0] ib_md.c:1604 UCX DEBUG mlx5_ib3: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638546.831095] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638546.831102] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638546.831375] [ndv4:68758:0] topo.c:99 UCX DEBUG bus id 0x104000000 doesn't exist. sys_dev = 3
[1650638546.831380] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638546.843926] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638546.844302] [ndv4:69197:0] ib_iface.c:994 UCX DEBUG iface=0x308d570: created UD QP 0x30e9 on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638546.844318] [ndv4:69197:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650638546.844416] [ndv4:68758:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638546.844423] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638546.844794] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638546.845019] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638546.845026] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638546.845102] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638546.845107] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638546.845528] [ndv4:69197:0] ib_md.c:812 UCX DEBUG registered memory 0x2afd7fd3c000..0x2afd7fdc1000 on mlx5_ib0 lkey 0x80e00 rkey 0x80e00 access 0xf flags 0x3e4
[1650638546.845539] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2afd7fd3c018 of 544744 bytes with 128 elements
[1650638546.845553] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638546.845712] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x308d570: adding gid fe80::15:5dff:fd34:1b to hash on device mlx5_ib0 port 1 index 0)
[1650638546.857858] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x308d570: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650638546.858046] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x308d570: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650638546.858118] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x308d570: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650638546.858269] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x308d570: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650638546.858375] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x308d570: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650638546.858530] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x308d570: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650638546.858742] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x308d570: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650638546.858760] [ndv4:69197:0] 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> }
[1650638546.859050] [ndv4:69197:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638546.859069] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x2703420 [id=82 ref 1] uct_ud_iface_async_handler() to hash
[1650638546.859103] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 82 events 0x5 mode thread_spinlock
[1650638546.859131] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[5]=0x308d570 using ud_mlx5/mlx5_ib0:1 on worker 0x306d6f0
[1650638546.859294] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638546.859300] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638546.859437] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638546.859442] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638546.869672] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib1:1
[1650638546.867826] [ndv4:68758:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638546.867836] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638546.871056] [ndv4:68758:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638546.871062] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638546.876994] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650638546.877033] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638546.877038] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638546.877110] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638546.877796] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638546.877804] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650638546.878986] [ndv4:69197:0] ib_iface.c:994 UCX DEBUG iface=0x393d050: created RC QP 0x3036 on mlx5_ib1:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650638546.880068] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[6]=0x393d050 using rc_verbs/mlx5_ib1:1 on worker 0x306d6f0
[1650638546.880176] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638546.880183] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638546.880404] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638546.880410] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638546.889464] [ndv4:68758:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638546.889472] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638546.889698] [ndv4:68758:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638546.891097] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib1:1
[1650638546.892648] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638546.892659] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638546.892662] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638546.892697] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638546.893079] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x3896010 of 8176 bytes with 127 elements
[1650638546.893352] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638546.893361] [ndv4:69197:0] 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> }
[1650638546.893398] [ndv4:69197:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650638546.893403] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638546.893417] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x2703ee0 [id=85 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638546.893443] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 85 events 0x1 mode thread_spinlock
[1650638546.893452] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[7]=0x37c3020 using rc_mlx5/mlx5_ib1:1 on worker 0x306d6f0
[1650638546.893470] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638546.893476] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638546.893720] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638546.893725] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638546.894498] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib1:1
[1650638546.904060] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638546.904076] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638546.904079] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638546.904134] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638546.904541] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650638546.904549] [ndv4:69197:0] 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> }
[1650638546.904580] [ndv4:69197:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650638546.904584] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638546.904590] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x2798720 [id=87 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638546.904609] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 87 events 0x1 mode thread_spinlock
[1650638546.905326] [ndv4:69197:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650638546.906945] [ndv4:68758:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib4 vendor_id: 0x15b3 device_id: 4124
[1650638546.907438] [ndv4:68758:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib4: disable ODP because it's not supported for DevX QP
[1650638546.912183] [ndv4:69197:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x3ae2010: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0x305c
[1650638546.912663] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638546.912670] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638546.912680] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2afd7fdc3008 of 151544 bytes with 1052 elements
[1650638546.916516] [ndv4:69197:0] ib_md.c:812 UCX DEBUG registered memory 0x2afd88200000..0x2afd8a800000 on mlx5_ib1 lkey 0x80c00 rkey 0x80c00 access 0xf flags 0x3e4
[1650638546.916530] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2afd88200018 of 39845864 bytes with 4752 elements
[1650638546.916710] [ndv4:69197:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x3ae2010
[1650638546.916728] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[8]=0x3ae2010 using dc_mlx5/mlx5_ib1:1 on worker 0x306d6f0
[1650638546.916879] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638546.916886] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638546.916947] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638546.916952] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638546.917050] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x21076a0 [id=65 ref 1] uct_ib_async_event_handler() to hash
[1650638546.917076] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 65 events 0x1 mode thread_spinlock
[1650638546.917080] [ndv4:68758:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib4' (InfiniBand channel adapter) with 1 ports
[1650638546.917388] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638546.917397] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638546.917403] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638546.917444] [ndv4:68758:0] ib_md.c:1319 UCX DEBUG mlx5_ib4: using registration cache
[1650638546.918445] [ndv4:68758:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638546.918451] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638546.917138] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib1:1
[1650638546.918366] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638546.918703] [ndv4:69197:0] ib_iface.c:994 UCX DEBUG iface=0x39e1020: created UD QP 0x303f on mlx5_ib1:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638546.919152] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638546.920499] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638546.920509] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638546.920827] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638546.920832] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638546.918779] [ndv4:68758:0] ib_md.c:1604 UCX DEBUG mlx5_ib4: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638546.919005] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638546.919011] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638546.921150] [ndv4:69197:0] ib_md.c:812 UCX DEBUG registered memory 0x2afd7fde8000..0x2afd7fe6d000 on mlx5_ib1 lkey 0x80d00 rkey 0x80d00 access 0xf flags 0x3e4
[1650638546.921159] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2afd7fde8018 of 544744 bytes with 128 elements
[1650638546.921167] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638546.921792] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x39e1020: adding gid fe80::15:5dff:fd34:1c to hash on device mlx5_ib1 port 1 index 0)
[1650638546.922106] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x39e1020: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 1)
[1650638546.922684] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x39e1020: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 2)
[1650638546.929044] [ndv4:68758:0] topo.c:99 UCX DEBUG bus id 0x105000000 doesn't exist. sys_dev = 4
[1650638546.929052] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638546.931376] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x39e1020: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 3)
[1650638546.931623] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x39e1020: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 4)
[1650638546.931642] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x39e1020: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 5)
[1650638546.931657] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x39e1020: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 6)
[1650638546.931731] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x39e1020: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 7)
[1650638546.932040] [ndv4:69197:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638546.932052] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x279c1c0 [id=88 ref 1] uct_ud_iface_async_handler() to hash
[1650638546.932082] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 88 events 0x5 mode thread_spinlock
[1650638546.932106] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[9]=0x39e1020 using ud_verbs/mlx5_ib1:1 on worker 0x306d6f0
[1650638546.932154] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638546.932160] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638546.932295] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638546.932300] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638546.932607] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib1:1
[1650638546.933426] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638546.933805] [ndv4:69197:0] ib_iface.c:994 UCX DEBUG iface=0x38983c0: created UD QP 0x3040 on mlx5_ib1:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638546.933814] [ndv4:69197:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650638546.934286] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638546.934368] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638546.934373] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638546.934387] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638546.934391] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638546.934696] [ndv4:69197:0] ib_md.c:812 UCX DEBUG registered memory 0x2afd7fe6d000..0x2afd7fef2000 on mlx5_ib1 lkey 0x80e00 rkey 0x80e00 access 0xf flags 0x3e4
[1650638546.934707] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2afd7fe6d018 of 544744 bytes with 128 elements
[1650638546.934716] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638546.934988] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x38983c0: adding gid fe80::15:5dff:fd34:1c to hash on device mlx5_ib1 port 1 index 0)
[1650638546.935057] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x38983c0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 1)
[1650638546.935073] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x38983c0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 2)
[1650638546.935420] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x38983c0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 3)
[1650638546.935624] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x38983c0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 4)
[1650638546.943025] [ndv4:68758:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650638546.943032] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638546.944712] [ndv4:68758:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650638546.944718] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638546.945443] [ndv4:68758:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650638546.945447] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638546.946001] [ndv4:68758:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650638546.946004] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638546.946308] [ndv4:68758:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638546.946528] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x38983c0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 5)
[1650638546.947054] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x38983c0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 6)
[1650638546.947270] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x38983c0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 7)
[1650638546.947275] [ndv4:69197:0] 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> }
[1650638546.947417] [ndv4:69197:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638546.947422] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x3945b90 [id=89 ref 1] uct_ud_iface_async_handler() to hash
[1650638546.947443] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 89 events 0x5 mode thread_spinlock
[1650638546.947454] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[10]=0x38983c0 using ud_mlx5/mlx5_ib1:1 on worker 0x306d6f0
[1650638546.947580] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638546.947586] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638546.947787] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638546.947791] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638546.955278] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib2:1
[1650638546.956783] [ndv4:68758:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib5 vendor_id: 0x15b3 device_id: 4124
[1650638546.957185] [ndv4:68758:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib5: disable ODP because it's not supported for DevX QP
[1650638546.957709] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x1a254b0 [id=67 ref 1] uct_ib_async_event_handler() to hash
[1650638546.957733] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 67 events 0x1 mode thread_spinlock
[1650638546.957736] [ndv4:68758:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib5' (InfiniBand channel adapter) with 1 ports
[1650638546.964739] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650638546.964756] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638546.964760] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638546.964810] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638546.965457] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638546.965472] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650638546.966857] [ndv4:69197:0] ib_iface.c:994 UCX DEBUG iface=0x3ed5020: created RC QP 0x3041 on mlx5_ib2:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650638546.968019] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[11]=0x3ed5020 using rc_verbs/mlx5_ib2:1 on worker 0x306d6f0
[1650638546.968181] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638546.968188] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638546.968658] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638546.968665] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638546.966759] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638546.966770] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638546.966775] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638546.966818] [ndv4:68758:0] ib_md.c:1319 UCX DEBUG mlx5_ib5: using registration cache
[1650638546.968478] [ndv4:68758:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638546.968485] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638546.969432] [ndv4:68758:0] ib_md.c:1604 UCX DEBUG mlx5_ib5: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638546.969776] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638546.969783] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638546.969225] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib2:1
[1650638546.970292] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638546.970312] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638546.970315] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638546.970364] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638546.970778] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x41f6010 of 8176 bytes with 127 elements
[1650638546.971024] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638546.971033] [ndv4:69197:0] 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> }
[1650638546.971068] [ndv4:69197:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib2 length=2048) failed: Invalid argument
[1650638546.971072] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638546.971085] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x37cbc80 [id=92 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638546.971106] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 92 events 0x1 mode thread_spinlock
[1650638546.971117] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[12]=0x4022030 using rc_mlx5/mlx5_ib2:1 on worker 0x306d6f0
[1650638546.971306] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638546.971312] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638546.971555] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638546.971560] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638546.971898] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib2:1
[1650638546.980866] [ndv4:68758:0] topo.c:99 UCX DEBUG bus id 0x106000000 doesn't exist. sys_dev = 5
[1650638546.980874] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638546.981538] [ndv4:68758:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650638546.981543] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638546.982768] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638546.982785] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638546.982788] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638546.982842] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638546.983247] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650638546.983255] [ndv4:69197:0] 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> }
[1650638546.983288] [ndv4:69197:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib2 length=2048) failed: Invalid argument
[1650638546.983292] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638546.983299] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x27962a0 [id=94 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638546.983319] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 94 events 0x1 mode thread_spinlock
[1650638546.984001] [ndv4:69197:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650638546.990929] [ndv4:69197:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x41f8050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0x3067
[1650638546.991362] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638546.991369] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638546.991380] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2afd7fef4008 of 151544 bytes with 1052 elements
[1650638546.992163] [ndv4:68758:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650638546.992171] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638546.993744] [ndv4:68758:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650638546.993749] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638546.995373] [ndv4:69197:0] ib_md.c:812 UCX DEBUG registered memory 0x2afd8aa00000..0x2afd8d000000 on mlx5_ib2 lkey 0x80a00 rkey 0x80a00 access 0xf flags 0x3e4
[1650638546.995402] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2afd8aa00018 of 39845864 bytes with 4752 elements
[1650638546.995572] [ndv4:69197:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x41f8050
[1650638546.995609] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[13]=0x41f8050 using dc_mlx5/mlx5_ib2:1 on worker 0x306d6f0
[1650638546.995864] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638546.995876] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638546.996100] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638546.996106] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638546.997297] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib2:1
[1650638547.005544] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638547.005855] [ndv4:69197:0] ib_iface.c:994 UCX DEBUG iface=0x27a8bf0: created UD QP 0x304a on mlx5_ib2:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638547.006415] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638547.006638] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638547.006648] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638547.010970] [ndv4:68758:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650638547.010977] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638547.011196] [ndv4:68758:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638547.015294] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638547.015303] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638547.015691] [ndv4:69197:0] ib_md.c:812 UCX DEBUG registered memory 0x2afd7ff19000..0x2afd7ff9e000 on mlx5_ib2 lkey 0x80b00 rkey 0x80b00 access 0xf flags 0x3e4
[1650638547.015698] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2afd7ff19018 of 544744 bytes with 128 elements
[1650638547.015703] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638547.016519] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x27a8bf0: adding gid fe80::15:5dff:fd34:1d to hash on device mlx5_ib2 port 1 index 0)
[1650638547.017172] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x27a8bf0: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 1)
[1650638547.018121] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x27a8bf0: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 2)
[1650638547.018343] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x27a8bf0: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 3)
[1650638547.018754] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x27a8bf0: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 4)
[1650638547.019607] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x27a8bf0: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 5)
[1650638547.020120] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x27a8bf0: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 6)
[1650638547.020427] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x27a8bf0: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 7)
[1650638547.020716] [ndv4:69197:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638547.020726] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x279a6e0 [id=95 ref 1] uct_ud_iface_async_handler() to hash
[1650638547.020753] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 95 events 0x5 mode thread_spinlock
[1650638547.020777] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[14]=0x27a8bf0 using ud_verbs/mlx5_ib2:1 on worker 0x306d6f0
[1650638547.020802] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638547.020807] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638547.021118] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638547.021124] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638547.021883] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib2:1
[1650638547.034573] [ndv4:68758:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib6 vendor_id: 0x15b3 device_id: 4124
[1650638547.035002] [ndv4:68758:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib6: disable ODP because it's not supported for DevX QP
[1650638547.036548] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x2107da0 [id=69 ref 1] uct_ib_async_event_handler() to hash
[1650638547.036570] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 69 events 0x1 mode thread_spinlock
[1650638547.036573] [ndv4:68758:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib6' (InfiniBand channel adapter) with 1 ports
[1650638547.036792] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638547.036802] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638547.036807] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638547.036847] [ndv4:68758:0] ib_md.c:1319 UCX DEBUG mlx5_ib6: using registration cache
[1650638547.038394] [ndv4:68758:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638547.038400] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638547.038682] [ndv4:68758:0] ib_md.c:1604 UCX DEBUG mlx5_ib6: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638547.038785] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638547.038790] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638547.040349] [ndv4:68758:0] topo.c:99 UCX DEBUG bus id 0x107000000 doesn't exist. sys_dev = 6
[1650638547.040354] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650638547.044960] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638547.045328] [ndv4:69197:0] ib_iface.c:994 UCX DEBUG iface=0x2791550: created UD QP 0x304b on mlx5_ib2:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638547.045337] [ndv4:69197:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650638547.045885] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638547.046073] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638547.046080] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638547.046135] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638547.046140] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638547.046539] [ndv4:69197:0] ib_md.c:812 UCX DEBUG registered memory 0x2afd8d07e000..0x2afd8d103000 on mlx5_ib2 lkey 0x80c00 rkey 0x80c00 access 0xf flags 0x3e4
[1650638547.046550] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2afd8d07e018 of 544744 bytes with 128 elements
[1650638547.046559] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638547.047100] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x2791550: adding gid fe80::15:5dff:fd34:1d to hash on device mlx5_ib2 port 1 index 0)
[1650638547.053036] [ndv4:68758:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650638547.053042] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650638547.060417] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x2791550: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 1)
[1650638547.060457] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x2791550: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 2)
[1650638547.060475] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x2791550: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 3)
[1650638547.060544] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x2791550: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 4)
[1650638547.060561] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x2791550: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 5)
[1650638547.066171] [ndv4:68758:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650638547.066177] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650638547.069993] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x2791550: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 6)
[1650638547.070117] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x2791550: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 7)
[1650638547.070126] [ndv4:69197:0] 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> }
[1650638547.070276] [ndv4:69197:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638547.070285] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x2792340 [id=96 ref 1] uct_ud_iface_async_handler() to hash
[1650638547.070311] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 96 events 0x5 mode thread_spinlock
[1650638547.070326] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[15]=0x2791550 using ud_mlx5/mlx5_ib2:1 on worker 0x306d6f0
[1650638547.070451] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638547.070457] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638547.070655] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638547.070660] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638547.077245] [ndv4:68758:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650638547.077251] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650638547.080703] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib3:1
[1650638547.082533] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650638547.082572] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.082577] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.082665] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.083311] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638547.083318] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650638547.084767] [ndv4:69197:0] ib_iface.c:994 UCX DEBUG iface=0x45f20a0: created RC QP 0x3033 on mlx5_ib3:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650638547.085888] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[16]=0x45f20a0 using rc_verbs/mlx5_ib3:1 on worker 0x306d6f0
[1650638547.085960] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638547.085966] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638547.086067] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638547.086072] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638547.090334] [ndv4:68758:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650638547.090341] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650638547.090565] [ndv4:68758:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638547.095497] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib3:1
[1650638547.097360] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638547.097375] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.097378] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.097428] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.097888] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x4913010 of 8176 bytes with 127 elements
[1650638547.098139] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638547.098148] [ndv4:69197:0] 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> }
[1650638547.098183] [ndv4:69197:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib3 length=2048) failed: Invalid argument
[1650638547.098187] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638547.098200] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x27b6ca0 [id=99 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638547.098270] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 99 events 0x1 mode thread_spinlock
[1650638547.098295] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[17]=0x473f030 using rc_mlx5/mlx5_ib3:1 on worker 0x306d6f0
[1650638547.098435] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638547.098442] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638547.098658] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638547.098663] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638547.098863] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib3:1
[1650638547.099861] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638547.099875] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.099878] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.099932] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.100278] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650638547.100284] [ndv4:69197:0] 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> }
[1650638547.100316] [ndv4:69197:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib3 length=2048) failed: Invalid argument
[1650638547.100320] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638547.100329] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x43d4f40 [id=101 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638547.100351] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 101 events 0x1 mode thread_spinlock
[1650638547.101039] [ndv4:69197:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650638547.103549] [ndv4:68758:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib7 vendor_id: 0x15b3 device_id: 4124
[1650638547.103983] [ndv4:68758:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib7: disable ODP because it's not supported for DevX QP
[1650638547.104516] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x1a21010 [id=71 ref 1] uct_ib_async_event_handler() to hash
[1650638547.104548] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 71 events 0x1 mode thread_spinlock
[1650638547.104551] [ndv4:68758:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib7' (InfiniBand channel adapter) with 1 ports
[1650638547.104827] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638547.104837] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638547.104843] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638547.104887] [ndv4:68758:0] ib_md.c:1319 UCX DEBUG mlx5_ib7: using registration cache
[1650638547.107942] [ndv4:69197:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x4915050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0x3059
[1650638547.108305] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638547.108312] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638547.108322] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2afd7ffa0008 of 151544 bytes with 1052 elements
[1650638547.112335] [ndv4:69197:0] ib_md.c:812 UCX DEBUG registered memory 0x2afd8d200000..0x2afd8f800000 on mlx5_ib3 lkey 0x80a00 rkey 0x80a00 access 0xf flags 0x3e4
[1650638547.112349] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2afd8d200018 of 39845864 bytes with 4752 elements
[1650638547.112519] [ndv4:69197:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x4915050
[1650638547.112539] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[18]=0x4915050 using dc_mlx5/mlx5_ib3:1 on worker 0x306d6f0
[1650638547.112640] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638547.112647] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638547.112748] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638547.112754] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638547.113228] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib3:1
[1650638547.123991] [ndv4:68758:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638547.124001] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638547.124313] [ndv4:68758:0] ib_md.c:1604 UCX DEBUG mlx5_ib7: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638547.125571] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638547.125939] [ndv4:69197:0] ib_iface.c:994 UCX DEBUG iface=0x4af1060: created UD QP 0x303c on mlx5_ib3:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638547.126516] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638547.126719] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638547.126728] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638547.126837] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638547.126842] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638547.127180] [ndv4:69197:0] ib_md.c:812 UCX DEBUG registered memory 0x2afd8f904000..0x2afd8f989000 on mlx5_ib3 lkey 0x80b00 rkey 0x80b00 access 0xf flags 0x3e4
[1650638547.127186] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2afd8f904018 of 544744 bytes with 128 elements
[1650638547.127191] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638547.127611] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x4af1060: adding gid fe80::15:5dff:fd34:1e to hash on device mlx5_ib3 port 1 index 0)
[1650638547.134941] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638547.134954] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638547.136928] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x4af1060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 1)
[1650638547.137114] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x4af1060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 2)
[1650638547.137525] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x4af1060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 3)
[1650638547.137571] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x4af1060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 4)
[1650638547.138009] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x4af1060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 5)
[1650638547.138616] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x4af1060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 6)
[1650638547.143917] [ndv4:68758:0] topo.c:99 UCX DEBUG bus id 0x108000000 doesn't exist. sys_dev = 7
[1650638547.143925] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650638547.146051] [ndv4:68758:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650638547.146060] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650638547.146083] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x4af1060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 7)
[1650638547.146361] [ndv4:69197:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638547.146369] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x3cbec00 [id=102 ref 1] uct_ud_iface_async_handler() to hash
[1650638547.146398] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 102 events 0x5 mode thread_spinlock
[1650638547.146421] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[19]=0x4af1060 using ud_verbs/mlx5_ib3:1 on worker 0x306d6f0
[1650638547.146569] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638547.146576] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638547.146764] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638547.146768] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638547.147413] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib3:1
[1650638547.148651] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638547.149066] [ndv4:69197:0] ib_iface.c:994 UCX DEBUG iface=0x4c0f050: created UD QP 0x303d on mlx5_ib3:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638547.149073] [ndv4:69197:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650638547.149583] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638547.149911] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638547.149917] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638547.150014] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638547.150019] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638547.150406] [ndv4:69197:0] ib_md.c:812 UCX DEBUG registered memory 0x2afd8f989000..0x2afd8fa0e000 on mlx5_ib3 lkey 0x80c00 rkey 0x80c00 access 0xf flags 0x3e4
[1650638547.150413] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2afd8f989018 of 544744 bytes with 128 elements
[1650638547.150418] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638547.155575] [ndv4:68758:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650638547.155583] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650638547.157773] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x4c0f050: adding gid fe80::15:5dff:fd34:1e to hash on device mlx5_ib3 port 1 index 0)
[1650638547.158076] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x4c0f050: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 1)
[1650638547.158352] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x4c0f050: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 2)
[1650638547.158699] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x4c0f050: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 3)
[1650638547.165988] [ndv4:68758:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650638547.165996] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650638547.167726] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x4c0f050: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 4)
[1650638547.167875] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x4c0f050: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 5)
[1650638547.176911] [ndv4:68758:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650638547.176919] [ndv4:68758:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650638547.177008] [ndv4:68758:0] ucp_context.c:1556 UCX DEBUG created ucp context 0x1a0b0c0 0x1a0b0c0 [10 mds 42 tls] features 0x1 tl bitmap 0x3ffffffffff 0x0
[1650638547.177442] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x4c0f050: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 6)
[1650638547.185868] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x4c0f050: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 7)
[1650638547.185877] [ndv4:69197:0] 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> }
[1650638547.185912] [ndv4:69197:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638547.185917] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x3cbeb40 [id=103 ref 1] uct_ud_iface_async_handler() to hash
[1650638547.185937] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 103 events 0x5 mode thread_spinlock
[1650638547.185947] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[20]=0x4c0f050 using ud_mlx5/mlx5_ib3:1 on worker 0x306d6f0
[1650638547.185990] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638547.185996] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638547.186098] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638547.186103] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638547.187194] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib4:1
[1650638547.201310] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650638547.201331] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.201334] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.201385] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.201995] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638547.202002] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650638547.203295] [ndv4:69197:0] ib_iface.c:994 UCX DEBUG iface=0x4d0f0a0: created RC QP 0x3052 on mlx5_ib4:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650638547.204423] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[21]=0x4d0f0a0 using rc_verbs/mlx5_ib4:1 on worker 0x306d6f0
[1650638547.204710] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638547.204717] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638547.204888] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638547.204893] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638547.205779] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib4:1
[1650638547.207881] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638547.207894] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.207896] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.207947] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.208429] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x5030010 of 8176 bytes with 127 elements
[1650638547.208659] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638547.208666] [ndv4:69197:0] 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> }
[1650638547.208702] [ndv4:69197:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib4 length=2048) failed: Invalid argument
[1650638547.208706] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638547.208718] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x43d4630 [id=106 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638547.208746] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 106 events 0x1 mode thread_spinlock
[1650638547.208756] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[22]=0x4e5c030 using rc_mlx5/mlx5_ib4:1 on worker 0x306d6f0
[1650638547.208883] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638547.208889] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638547.208960] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638547.208966] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638547.219556] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib4:1
[1650638547.220918] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638547.220934] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.220937] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.220992] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.221360] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650638547.221367] [ndv4:69197:0] 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> }
[1650638547.221402] [ndv4:69197:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib4 length=2048) failed: Invalid argument
[1650638547.221406] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638547.221416] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x4e64f60 [id=108 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638547.221438] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 108 events 0x1 mode thread_spinlock
[1650638547.222117] [ndv4:69197:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650638547.229088] [ndv4:69197:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x5032050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0x3068
[1650638547.229358] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638547.229365] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638547.229374] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2afd7ffc7008 of 151544 bytes with 1052 elements
[1650638547.233219] [ndv4:69197:0] ib_md.c:812 UCX DEBUG registered memory 0x2afd8fc00000..0x2afd92200000 on mlx5_ib4 lkey 0x80e00 rkey 0x80e00 access 0xf flags 0x3e4
[1650638547.233234] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2afd8fc00018 of 39845864 bytes with 4752 elements
[1650638547.233404] [ndv4:69197:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x5032050
[1650638547.233420] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[23]=0x5032050 using dc_mlx5/mlx5_ib4:1 on worker 0x306d6f0
[1650638547.233495] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638547.233502] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638547.233557] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638547.233562] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638547.233917] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib4:1
[1650638547.234031] [ndv4:68758:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 8447 bytes with hugetlb
[1650638547.234124] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool mm_recv_desc: align 64, maxelems 4294967295, elemsize 8368
[1650638547.234153] [ndv4:68758:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 4292720 bytes with hugetlb
[1650638547.234163] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool mm_recv_desc: allocated chunk 0x2ad1e5841018 of 4296680 bytes with 512 elements
[1650638547.234742] [ndv4:68758:0] mm_iface.c:600 UCX DEBUG created mm iface 0x2187280 FIFO id 0xd800f va 0x2ad1dfeff000 size 12288 (128 x 64 elems)
[1650638547.234792] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[0]=0x2187280 using sysv/memory on worker 0x2a61550
[1650638547.234842] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638547.234850] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638547.234965] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638547.234970] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638547.235493] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib0:1
[1650638547.236970] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650638547.237014] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.237017] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.237068] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.237640] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638547.237655] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650638547.238775] [ndv4:68758:0] ib_iface.c:994 UCX DEBUG iface=0x2192490: created RC QP 0x30ea on mlx5_ib0:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650638547.247009] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638547.247383] [ndv4:69197:0] ib_iface.c:994 UCX DEBUG iface=0x520e060: created UD QP 0x305e on mlx5_ib4:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638547.247969] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638547.248322] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638547.248333] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638547.248371] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638547.248376] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638547.248742] [ndv4:69197:0] ib_md.c:812 UCX DEBUG registered memory 0x2afd9220f000..0x2afd92294000 on mlx5_ib4 lkey 0x80f00 rkey 0x80f00 access 0xf flags 0x3e4
[1650638547.248753] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2afd9220f018 of 544744 bytes with 128 elements
[1650638547.248761] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638547.250685] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[1]=0x2192490 using rc_verbs/mlx5_ib0:1 on worker 0x2a61550
[1650638547.250996] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638547.251004] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638547.251272] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638547.251278] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638547.252084] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib0:1
[1650638547.252610] [ndv4:68758:0] ib_device.c:1394 UCX DEBUG max IB CQE size is 128
[1650638547.253699] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638547.253709] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.253712] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.253761] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.254309] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x2d88010 of 8176 bytes with 127 elements
[1650638547.254599] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638547.254624] [ndv4:68758:0] 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> }
[1650638547.254672] [ndv4:68758:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650638547.254676] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638547.264366] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x20f5b60 [id=78 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638547.264398] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 78 events 0x1 mode thread_spinlock
[1650638547.264779] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x520e060: adding gid fe80::15:5dff:fd34:1f to hash on device mlx5_ib4 port 1 index 0)
[1650638547.264857] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[2]=0x21a0010 using rc_mlx5/mlx5_ib0:1 on worker 0x2a61550
[1650638547.264917] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638547.264924] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638547.265022] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638547.265028] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638547.265540] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib0:1
[1650638547.273463] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x520e060: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 1)
[1650638547.273712] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x520e060: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 2)
[1650638547.273817] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x520e060: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 3)
[1650638547.274028] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x520e060: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 4)
[1650638547.274286] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x520e060: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 5)
[1650638547.274575] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x520e060: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 6)
[1650638547.274625] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x520e060: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 7)
[1650638547.274873] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638547.274887] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.274891] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.274905] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.275303] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650638547.275314] [ndv4:68758:0] 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> }
[1650638547.275348] [ndv4:68758:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650638547.275352] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638547.275361] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x20f4b60 [id=80 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638547.275382] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 80 events 0x1 mode thread_spinlock
[1650638547.274989] [ndv4:69197:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638547.275003] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x44f2fc0 [id=109 ref 1] uct_ud_iface_async_handler() to hash
[1650638547.275033] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 109 events 0x5 mode thread_spinlock
[1650638547.275065] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[24]=0x520e060 using ud_verbs/mlx5_ib4:1 on worker 0x306d6f0
[1650638547.275176] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638547.275183] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638547.275327] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638547.275332] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638547.275546] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib4:1
[1650638547.276103] [ndv4:68758:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650638547.276358] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638547.276716] [ndv4:69197:0] ib_iface.c:994 UCX DEBUG iface=0x532c460: created UD QP 0x305f on mlx5_ib4:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638547.276727] [ndv4:69197:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650638547.277271] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638547.277556] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638547.277563] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638547.277652] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638547.277657] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638547.278012] [ndv4:69197:0] ib_md.c:812 UCX DEBUG registered memory 0x2afd92294000..0x2afd92319000 on mlx5_ib4 lkey 0x81000 rkey 0x81000 access 0xf flags 0x3e4
[1650638547.278021] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2afd92294018 of 544744 bytes with 128 elements
[1650638547.278029] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638547.278382] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x532c460: adding gid fe80::15:5dff:fd34:1f to hash on device mlx5_ib4 port 1 index 0)
[1650638547.278711] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x532c460: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 1)
[1650638547.278929] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x532c460: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 2)
[1650638547.279175] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x532c460: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 3)
[1650638547.279429] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x532c460: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 4)
[1650638547.279613] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x532c460: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 5)
[1650638547.279743] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x532c460: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 6)
[1650638547.279833] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x532c460: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 7)
[1650638547.279843] [ndv4:69197:0] 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> }
[1650638547.279932] [ndv4:69197:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638547.279938] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x4ceae30 [id=110 ref 1] uct_ud_iface_async_handler() to hash
[1650638547.279965] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 110 events 0x5 mode thread_spinlock
[1650638547.279976] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[25]=0x532c460 using ud_mlx5/mlx5_ib4:1 on worker 0x306d6f0
[1650638547.280042] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638547.280049] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638547.280149] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638547.280154] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638547.280884] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib5:1
[1650638547.282391] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650638547.282413] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.282416] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.282472] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.283103] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638547.283109] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650638547.284525] [ndv4:69197:0] ib_iface.c:994 UCX DEBUG iface=0x542c0a0: created RC QP 0x2ba8 on mlx5_ib5:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650638547.285637] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[26]=0x542c0a0 using rc_verbs/mlx5_ib5:1 on worker 0x306d6f0
[1650638547.285677] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638547.285684] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638547.285900] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638547.285905] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638547.284183] [ndv4:68758:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x2d8a050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0x3113
[1650638547.284353] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638547.284361] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638547.284388] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2ad1dff04008 of 151544 bytes with 1052 elements
[1650638547.288324] [ndv4:68758:0] ib_md.c:812 UCX DEBUG registered memory 0x2ad1e5e00000..0x2ad1e8400000 on mlx5_ib0 lkey 0x80f00 rkey 0x80f00 access 0xf flags 0x3e4
[1650638547.288347] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2ad1e5e00018 of 39845864 bytes with 4752 elements
[1650638547.288491] [ndv4:68758:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x2d8a050
[1650638547.288528] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[3]=0x2d8a050 using dc_mlx5/mlx5_ib0:1 on worker 0x2a61550
[1650638547.288965] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638547.288976] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638547.289434] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638547.289439] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638547.290096] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib0:1
[1650638547.291170] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638547.291613] [ndv4:68758:0] ib_iface.c:994 UCX DEBUG iface=0x219d650: created UD QP 0x30f3 on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638547.292233] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638547.292378] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638547.292384] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638547.292551] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638547.292556] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638547.292956] [ndv4:68758:0] ib_md.c:812 UCX DEBUG registered memory 0x2ad1dff29000..0x2ad1dffae000 on mlx5_ib0 lkey 0x81000 rkey 0x81000 access 0xf flags 0x3e4
[1650638547.292962] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ad1dff29018 of 544744 bytes with 128 elements
[1650638547.292966] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638547.296494] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib5:1
[1650638547.298025] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638547.298043] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.298046] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.298101] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.298567] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x574d010 of 8176 bytes with 127 elements
[1650638547.298846] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638547.298852] [ndv4:69197:0] 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> }
[1650638547.298894] [ndv4:69197:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib5 length=2048) failed: Invalid argument
[1650638547.298898] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638547.298909] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x532c3e0 [id=113 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638547.298933] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 113 events 0x1 mode thread_spinlock
[1650638547.298944] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[27]=0x5579030 using rc_mlx5/mlx5_ib5:1 on worker 0x306d6f0
[1650638547.298977] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638547.298985] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638547.299089] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638547.299094] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638547.299385] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib5:1
[1650638547.300513] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638547.300531] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.300534] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.300590] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.300926] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650638547.300933] [ndv4:69197:0] 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> }
[1650638547.300968] [ndv4:69197:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib5 length=2048) failed: Invalid argument
[1650638547.300972] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638547.300981] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x3eb09a0 [id=115 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638547.301005] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 115 events 0x1 mode thread_spinlock
[1650638547.301645] [ndv4:69197:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650638547.304135] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x219d650: adding gid fe80::15:5dff:fd34:1b to hash on device mlx5_ib0 port 1 index 0)
[1650638547.304243] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x219d650: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650638547.304484] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x219d650: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650638547.304655] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x219d650: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650638547.304870] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x219d650: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650638547.305406] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x219d650: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650638547.305919] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x219d650: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650638547.308600] [ndv4:69197:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x574f050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0x2bce
[1650638547.308841] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638547.308849] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638547.308860] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2afd94b1a008 of 151544 bytes with 1052 elements
[1650638547.312786] [ndv4:69197:0] ib_md.c:812 UCX DEBUG registered memory 0x2afd92400000..0x2afd94a00000 on mlx5_ib5 lkey 0x80a00 rkey 0x80a00 access 0xf flags 0x3e4
[1650638547.312801] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2afd92400018 of 39845864 bytes with 4752 elements
[1650638547.312969] [ndv4:69197:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x574f050
[1650638547.312988] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[28]=0x574f050 using dc_mlx5/mlx5_ib5:1 on worker 0x306d6f0
[1650638547.313010] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638547.313017] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638547.313071] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638547.313076] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638547.313334] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib5:1
[1650638547.314183] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638547.314526] [ndv4:69197:0] ib_iface.c:994 UCX DEBUG iface=0x451f480: created UD QP 0x2bb1 on mlx5_ib5:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638547.314995] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638547.315055] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638547.315064] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638547.315085] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638547.315090] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638547.315436] [ndv4:69197:0] ib_md.c:812 UCX DEBUG registered memory 0x2afd94b3f000..0x2afd94bc4000 on mlx5_ib5 lkey 0x80b00 rkey 0x80b00 access 0xf flags 0x3e4
[1650638547.315445] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2afd94b3f018 of 544744 bytes with 128 elements
[1650638547.315452] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638547.315511] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x451f480: adding gid fe80::15:5dff:fd34:20 to hash on device mlx5_ib5 port 1 index 0)
[1650638547.315542] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x451f480: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 1)
[1650638547.315568] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x451f480: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 2)
[1650638547.315590] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x451f480: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 3)
[1650638547.315616] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x451f480: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 4)
[1650638547.315641] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x451f480: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 5)
[1650638547.315670] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x451f480: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 6)
[1650638547.315695] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x451f480: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 7)
[1650638547.315957] [ndv4:69197:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638547.315967] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x3eb07d0 [id=116 ref 1] uct_ud_iface_async_handler() to hash
[1650638547.315998] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 116 events 0x5 mode thread_spinlock
[1650638547.316024] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[29]=0x451f480 using ud_verbs/mlx5_ib5:1 on worker 0x306d6f0
[1650638547.316046] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638547.316052] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638547.316125] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638547.316129] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638547.316304] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib5:1
[1650638547.316921] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638547.317275] [ndv4:69197:0] ib_iface.c:994 UCX DEBUG iface=0x5a49050: created UD QP 0x2bb2 on mlx5_ib5:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638547.317283] [ndv4:69197:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650638547.317671] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638547.317754] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638547.317761] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638547.317791] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638547.317795] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638547.318064] [ndv4:69197:0] ib_md.c:812 UCX DEBUG registered memory 0x2afd94bc4000..0x2afd94c49000 on mlx5_ib5 lkey 0x80c00 rkey 0x80c00 access 0xf flags 0x3e4
[1650638547.318072] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2afd94bc4018 of 544744 bytes with 128 elements
[1650638547.318080] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638547.318152] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x5a49050: adding gid fe80::15:5dff:fd34:20 to hash on device mlx5_ib5 port 1 index 0)
[1650638547.324713] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x219d650: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650638547.325075] [ndv4:68758:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638547.332622] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x5a49050: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 1)
[1650638547.332937] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x20f6a70 [id=81 ref 1] uct_ud_iface_async_handler() to hash
[1650638547.332974] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 81 events 0x5 mode thread_spinlock
[1650638547.333119] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[4]=0x219d650 using ud_verbs/mlx5_ib0:1 on worker 0x2a61550
[1650638547.333146] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638547.333153] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638547.333416] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638547.333421] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638547.334063] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib0:1
[1650638547.334986] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638547.335391] [ndv4:68758:0] ib_iface.c:994 UCX DEBUG iface=0x21a8980: created UD QP 0x30f5 on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638547.335409] [ndv4:68758:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650638547.336023] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638547.336140] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638547.336146] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638547.336185] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638547.336189] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638547.336665] [ndv4:68758:0] ib_md.c:812 UCX DEBUG registered memory 0x2ad1e845b000..0x2ad1e84e0000 on mlx5_ib0 lkey 0x81100 rkey 0x81100 access 0xf flags 0x3e4
[1650638547.336671] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ad1e845b018 of 544744 bytes with 128 elements
[1650638547.336675] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638547.337245] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x21a8980: adding gid fe80::15:5dff:fd34:1b to hash on device mlx5_ib0 port 1 index 0)
[1650638547.337523] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x21a8980: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650638547.337856] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x21a8980: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650638547.338006] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x21a8980: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650638547.338026] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x21a8980: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650638547.338040] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x21a8980: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650638547.338272] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x21a8980: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650638547.338501] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x21a8980: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650638547.338511] [ndv4:68758:0] 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> }
[1650638547.338542] [ndv4:68758:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638547.338547] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x20f7ba0 [id=82 ref 1] uct_ud_iface_async_handler() to hash
[1650638547.338568] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 82 events 0x5 mode thread_spinlock
[1650638547.338582] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[5]=0x21a8980 using ud_mlx5/mlx5_ib0:1 on worker 0x2a61550
[1650638547.338664] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638547.338670] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638547.338791] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638547.338796] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638547.341902] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x5a49050: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 2)
[1650638547.351658] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x5a49050: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 3)
[1650638547.351874] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x5a49050: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 4)
[1650638547.352316] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x5a49050: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 5)
[1650638547.357289] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib1:1
[1650638547.358368] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650638547.358378] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.358381] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.358397] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.358877] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638547.358882] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650638547.359814] [ndv4:68758:0] ib_iface.c:994 UCX DEBUG iface=0x3331050: created RC QP 0x3042 on mlx5_ib1:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650638547.361018] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[6]=0x3331050 using rc_verbs/mlx5_ib1:1 on worker 0x2a61550
[1650638547.361097] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638547.361102] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638547.361182] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638547.361187] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638547.361544] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib1:1
[1650638547.362937] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638547.362945] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.362948] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.362964] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.363489] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x328a010 of 8176 bytes with 127 elements
[1650638547.363741] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638547.363747] [ndv4:68758:0] 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> }
[1650638547.363785] [ndv4:68758:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650638547.363789] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638547.363800] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x20f85e0 [id=85 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638547.363821] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 85 events 0x1 mode thread_spinlock
[1650638547.363830] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[7]=0x31b7020 using rc_mlx5/mlx5_ib1:1 on worker 0x2a61550
[1650638547.363904] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638547.363909] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638547.364071] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638547.364076] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638547.364611] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib1:1
[1650638547.364677] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x5a49050: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 6)
[1650638547.364913] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x5a49050: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 7)
[1650638547.364931] [ndv4:69197:0] 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> }
[1650638547.365026] [ndv4:69197:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638547.365034] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x54073a0 [id=117 ref 1] uct_ud_iface_async_handler() to hash
[1650638547.365062] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 117 events 0x5 mode thread_spinlock
[1650638547.365084] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[30]=0x5a49050 using ud_mlx5/mlx5_ib5:1 on worker 0x306d6f0
[1650638547.365413] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638547.365420] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638547.365481] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638547.365486] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638547.366121] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638547.366135] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.366138] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.366194] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.366614] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650638547.366620] [ndv4:68758:0] 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> }
[1650638547.366656] [ndv4:68758:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650638547.366660] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638547.366666] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x1a27080 [id=87 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638547.366687] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 87 events 0x1 mode thread_spinlock
[1650638547.367294] [ndv4:68758:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650638547.366279] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib6:1
[1650638547.368347] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650638547.368367] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.368370] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.368420] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.369042] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638547.369049] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650638547.370518] [ndv4:69197:0] ib_iface.c:994 UCX DEBUG iface=0x5b490a0: created RC QP 0x2b1a on mlx5_ib6:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650638547.371647] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[31]=0x5b490a0 using rc_verbs/mlx5_ib6:1 on worker 0x306d6f0
[1650638547.371771] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638547.371780] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638547.371909] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638547.371914] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638547.372064] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib6:1
[1650638547.373277] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638547.373297] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.373300] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.373352] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.373746] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x5e6a010 of 8176 bytes with 127 elements
[1650638547.374147] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638547.374156] [ndv4:69197:0] 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> }
[1650638547.374304] [ndv4:69197:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib6 length=2048) failed: Invalid argument
[1650638547.374311] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638547.374324] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x5b242d0 [id=120 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638547.374347] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 120 events 0x1 mode thread_spinlock
[1650638547.374363] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[32]=0x5c96030 using rc_mlx5/mlx5_ib6:1 on worker 0x306d6f0
[1650638547.374511] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638547.374517] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638547.374717] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638547.374722] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638547.375472] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib6:1
[1650638547.374928] [ndv4:68758:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x34d6010: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0x306b
[1650638547.375033] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638547.375039] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638547.375048] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2ad1dffb0008 of 151544 bytes with 1052 elements
[1650638547.377671] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638547.377692] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.377696] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.377755] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.378167] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650638547.378176] [ndv4:69197:0] 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> }
[1650638547.378275] [ndv4:69197:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib6 length=2048) failed: Invalid argument
[1650638547.378281] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638547.378292] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x5c9ef60 [id=122 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638547.378318] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 122 events 0x1 mode thread_spinlock
[1650638547.379040] [ndv4:69197:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650638547.378995] [ndv4:68758:0] ib_md.c:812 UCX DEBUG registered memory 0x2ad1e8600000..0x2ad1eac00000 on mlx5_ib1 lkey 0x81100 rkey 0x81100 access 0xf flags 0x3e4
[1650638547.379013] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2ad1e8600018 of 39845864 bytes with 4752 elements
[1650638547.379156] [ndv4:68758:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x34d6010
[1650638547.379190] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[8]=0x34d6010 using dc_mlx5/mlx5_ib1:1 on worker 0x2a61550
[1650638547.379382] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638547.379391] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638547.379588] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638547.379593] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638547.386200] [ndv4:69197:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x5e6c050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0x2b21
[1650638547.386387] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638547.386395] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638547.386405] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2afd9744a008 of 151544 bytes with 1052 elements
[1650638547.390001] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib1:1
[1650638547.390374] [ndv4:69197:0] ib_md.c:812 UCX DEBUG registered memory 0x2afd94e00000..0x2afd97400000 on mlx5_ib6 lkey 0x80900 rkey 0x80900 access 0xf flags 0x3e4
[1650638547.390394] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2afd94e00018 of 39845864 bytes with 4752 elements
[1650638547.390567] [ndv4:69197:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x5e6c050
[1650638547.390597] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[33]=0x5e6c050 using dc_mlx5/mlx5_ib6:1 on worker 0x306d6f0
[1650638547.391171] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638547.391666] [ndv4:68758:0] ib_iface.c:994 UCX DEBUG iface=0x219c9d0: created UD QP 0x304e on mlx5_ib1:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638547.392288] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638547.399588] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638547.399603] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638547.399728] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638547.399732] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638547.400160] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib6:1
[1650638547.402749] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638547.403165] [ndv4:69197:0] ib_iface.c:994 UCX DEBUG iface=0x6048060: created UD QP 0x2b26 on mlx5_ib6:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638547.403898] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638547.404321] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638547.404332] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638547.404375] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638547.404380] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638547.404741] [ndv4:69197:0] ib_md.c:812 UCX DEBUG registered memory 0x2afd9746f000..0x2afd974f4000 on mlx5_ib6 lkey 0x80a00 rkey 0x80a00 access 0xf flags 0x3e4
[1650638547.404750] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2afd9746f018 of 544744 bytes with 128 elements
[1650638547.404759] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638547.404876] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x6048060: adding gid fe80::15:5dff:fd34:21 to hash on device mlx5_ib6 port 1 index 0)
[1650638547.406083] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x6048060: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 1)
[1650638547.410336] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638547.410346] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638547.410496] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638547.410501] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638547.410968] [ndv4:68758:0] ib_md.c:812 UCX DEBUG registered memory 0x2ad1eace1000..0x2ad1ead66000 on mlx5_ib1 lkey 0x81200 rkey 0x81200 access 0xf flags 0x3e4
[1650638547.410975] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ad1eace1018 of 544744 bytes with 128 elements
[1650638547.410979] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638547.411479] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x219c9d0: adding gid fe80::15:5dff:fd34:1c to hash on device mlx5_ib1 port 1 index 0)
[1650638547.411697] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x219c9d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 1)
[1650638547.411782] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x219c9d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 2)
[1650638547.411866] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x219c9d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 3)
[1650638547.411900] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x219c9d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 4)
[1650638547.411935] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x219c9d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 5)
[1650638547.411962] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x219c9d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 6)
[1650638547.411988] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x219c9d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 7)
[1650638547.412319] [ndv4:68758:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638547.412328] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x219c5f0 [id=88 ref 1] uct_ud_iface_async_handler() to hash
[1650638547.412353] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 88 events 0x5 mode thread_spinlock
[1650638547.412364] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[9]=0x219c9d0 using ud_verbs/mlx5_ib1:1 on worker 0x2a61550
[1650638547.412379] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638547.412385] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638547.412429] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638547.412434] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638547.412622] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib1:1
[1650638547.413396] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638547.413790] [ndv4:68758:0] ib_iface.c:994 UCX DEBUG iface=0x2a813d0: created UD QP 0x304f on mlx5_ib1:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638547.413798] [ndv4:68758:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650638547.414450] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638547.414483] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638547.414489] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638547.414511] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638547.414515] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638547.414839] [ndv4:68758:0] ib_md.c:812 UCX DEBUG registered memory 0x2ad1ead66000..0x2ad1eadeb000 on mlx5_ib1 lkey 0x81300 rkey 0x81300 access 0xf flags 0x3e4
[1650638547.414846] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ad1ead66018 of 544744 bytes with 128 elements
[1650638547.414850] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638547.414940] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x2a813d0: adding gid fe80::15:5dff:fd34:1c to hash on device mlx5_ib1 port 1 index 0)
[1650638547.415008] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x2a813d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 1)
[1650638547.415039] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x2a813d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 2)
[1650638547.415062] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x2a813d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 3)
[1650638547.415087] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x2a813d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 4)
[1650638547.415110] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x2a813d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 5)
[1650638547.415134] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x2a813d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 6)
[1650638547.415159] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x2a813d0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 7)
[1650638547.415165] [ndv4:68758:0] 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> }
[1650638547.415195] [ndv4:68758:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638547.415199] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x2a81ed0 [id=89 ref 1] uct_ud_iface_async_handler() to hash
[1650638547.415262] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 89 events 0x5 mode thread_spinlock
[1650638547.415271] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[10]=0x2a813d0 using ud_mlx5/mlx5_ib1:1 on worker 0x2a61550
[1650638547.415283] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638547.415287] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638547.415341] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638547.415344] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638547.415755] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib2:1
[1650638547.416984] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650638547.417000] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.417003] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.417051] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.417558] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638547.417564] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650638547.418585] [ndv4:68758:0] ib_iface.c:994 UCX DEBUG iface=0x38c9020: created RC QP 0x304c on mlx5_ib2:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650638547.419537] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x6048060: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 2)
[1650638547.419564] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x6048060: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 3)
[1650638547.419708] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x6048060: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 4)
[1650638547.420012] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x6048060: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 5)
[1650638547.420285] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x6048060: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 6)
[1650638547.419794] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[11]=0x38c9020 using rc_verbs/mlx5_ib2:1 on worker 0x2a61550
[1650638547.419848] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638547.419854] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638547.420013] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638547.420019] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638547.420738] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib2:1
[1650638547.431119] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x6048060: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 7)
[1650638547.431452] [ndv4:69197:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638547.431466] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x6048f20 [id=123 ref 1] uct_ud_iface_async_handler() to hash
[1650638547.431496] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 123 events 0x5 mode thread_spinlock
[1650638547.431522] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[34]=0x6048060 using ud_verbs/mlx5_ib6:1 on worker 0x306d6f0
[1650638547.431729] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638547.431736] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638547.431869] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638547.431874] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638547.432764] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib6:1
[1650638547.442733] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638547.442753] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.442757] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.442808] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.444535] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x3bea010 of 8176 bytes with 127 elements
[1650638547.444539] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638547.444899] [ndv4:69197:0] ib_iface.c:994 UCX DEBUG iface=0x53593a0: created UD QP 0x2b27 on mlx5_ib6:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638547.444908] [ndv4:69197:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650638547.444849] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638547.444860] [ndv4:68758:0] 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> }
[1650638547.444949] [ndv4:68758:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib2 length=2048) failed: Invalid argument
[1650638547.444954] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638547.444969] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x3339b90 [id=92 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638547.444992] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 92 events 0x1 mode thread_spinlock
[1650638547.445013] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[12]=0x3a16030 using rc_mlx5/mlx5_ib2:1 on worker 0x2a61550
[1650638547.445113] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638547.445120] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638547.445268] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638547.445274] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638547.445432] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638547.445590] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638547.445599] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638547.445672] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638547.445677] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638547.446174] [ndv4:69197:0] ib_md.c:812 UCX DEBUG registered memory 0x2afd974f4000..0x2afd97579000 on mlx5_ib6 lkey 0x80e00 rkey 0x80e00 access 0xf flags 0x3e4
[1650638547.446182] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2afd974f4018 of 544744 bytes with 128 elements
[1650638547.446186] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638547.446343] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x53593a0: adding gid fe80::15:5dff:fd34:21 to hash on device mlx5_ib6 port 1 index 0)
[1650638547.446571] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x53593a0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 1)
[1650638547.446645] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x53593a0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 2)
[1650638547.455748] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib2:1
[1650638547.456525] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x53593a0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 3)
[1650638547.456868] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x53593a0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 4)
[1650638547.457119] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x53593a0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 5)
[1650638547.457433] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x53593a0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 6)
[1650638547.457453] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x53593a0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 7)
[1650638547.457459] [ndv4:69197:0] 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> }
[1650638547.457489] [ndv4:69197:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638547.457494] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x5b240e0 [id=124 ref 1] uct_ud_iface_async_handler() to hash
[1650638547.457515] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 124 events 0x5 mode thread_spinlock
[1650638547.457525] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[35]=0x53593a0 using ud_mlx5/mlx5_ib6:1 on worker 0x306d6f0
[1650638547.457626] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638547.457633] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638547.457692] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638547.457697] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638547.458365] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib7:1
[1650638547.465611] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638547.465636] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.465640] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.465703] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.466159] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650638547.466169] [ndv4:68758:0] 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> }
[1650638547.466235] [ndv4:68758:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib2 length=2048) failed: Invalid argument
[1650638547.466240] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638547.466252] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x3a1efc0 [id=94 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638547.466277] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 94 events 0x1 mode thread_spinlock
[1650638547.467035] [ndv4:68758:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650638547.467042] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650638547.467066] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.467070] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.467125] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.467956] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638547.467964] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650638547.469444] [ndv4:69197:0] ib_iface.c:994 UCX DEBUG iface=0x62660a0: created RC QP 0x2b0d on mlx5_ib7:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650638547.470668] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[36]=0x62660a0 using rc_verbs/mlx5_ib7:1 on worker 0x306d6f0
[1650638547.470786] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638547.470793] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638547.471011] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638547.471017] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638547.471403] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib7:1
[1650638547.472533] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638547.472546] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.472550] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.472597] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.473018] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x6587010 of 8176 bytes with 127 elements
[1650638547.473334] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638547.473344] [ndv4:69197:0] 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> }
[1650638547.473384] [ndv4:69197:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib7 length=2048) failed: Invalid argument
[1650638547.473388] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638547.473401] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x61668c0 [id=127 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638547.473430] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 127 events 0x1 mode thread_spinlock
[1650638547.473440] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[37]=0x63b3030 using rc_mlx5/mlx5_ib7:1 on worker 0x306d6f0
[1650638547.473468] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638547.473475] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638547.473545] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638547.473550] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638547.473756] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib7:1
[1650638547.475130] [ndv4:68758:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x3bec050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0x3075
[1650638547.475439] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638547.475449] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638547.475459] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2ad1dffd7008 of 151544 bytes with 1052 elements
[1650638547.474980] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638547.474995] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.474998] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.475056] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.475436] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650638547.475443] [ndv4:69197:0] 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> }
[1650638547.475476] [ndv4:69197:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib7 length=2048) failed: Invalid argument
[1650638547.475480] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638547.475487] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x6241760 [id=129 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638547.475510] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 129 events 0x1 mode thread_spinlock
[1650638547.476146] [ndv4:69197:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650638547.479245] [ndv4:68758:0] ib_md.c:812 UCX DEBUG registered memory 0x2ad1eae00000..0x2ad1ed400000 on mlx5_ib2 lkey 0x80d00 rkey 0x80d00 access 0xf flags 0x3e4
[1650638547.479262] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2ad1eae00018 of 39845864 bytes with 4752 elements
[1650638547.479398] [ndv4:68758:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x3bec050
[1650638547.479431] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[13]=0x3bec050 using dc_mlx5/mlx5_ib2:1 on worker 0x2a61550
[1650638547.479482] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638547.479492] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638547.479582] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638547.479588] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638547.483385] [ndv4:69197:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x6589050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0x2b14
[1650638547.483671] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638547.483680] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638547.483689] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2afd99d7a008 of 151544 bytes with 1052 elements
[1650638547.487563] [ndv4:69197:0] ib_md.c:812 UCX DEBUG registered memory 0x2afd97600000..0x2afd99c00000 on mlx5_ib7 lkey 0x80a00 rkey 0x80a00 access 0xf flags 0x3e4
[1650638547.487578] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2afd97600018 of 39845864 bytes with 4752 elements
[1650638547.487743] [ndv4:69197:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x6589050
[1650638547.487765] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[38]=0x6589050 using dc_mlx5/mlx5_ib7:1 on worker 0x306d6f0
[1650638547.487810] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638547.487819] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638547.487868] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638547.487874] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638547.488301] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib7:1
[1650638547.489157] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638547.489504] [ndv4:69197:0] ib_iface.c:994 UCX DEBUG iface=0x5a76570: created UD QP 0x2b19 on mlx5_ib7:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638547.489976] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638547.490190] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638547.490201] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638547.490277] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638547.490282] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638547.490624] [ndv4:69197:0] ib_md.c:812 UCX DEBUG registered memory 0x2afd99d9f000..0x2afd99e24000 on mlx5_ib7 lkey 0x80b00 rkey 0x80b00 access 0xf flags 0x3e4
[1650638547.490635] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2afd99d9f018 of 544744 bytes with 128 elements
[1650638547.490643] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638547.490767] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x5a76570: adding gid fe80::15:5dff:fd34:22 to hash on device mlx5_ib7 port 1 index 0)
[1650638547.490983] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x5a76570: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 1)
[1650638547.491007] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x5a76570: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 2)
[1650638547.491023] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x5a76570: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 3)
[1650638547.491255] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x5a76570: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 4)
[1650638547.491274] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x5a76570: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 5)
[1650638547.491531] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x5a76570: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 6)
[1650638547.491552] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x5a76570: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 7)
[1650638547.490952] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib2:1
[1650638547.492057] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638547.492435] [ndv4:68758:0] ib_iface.c:994 UCX DEBUG iface=0x2f55500: created UD QP 0x3055 on mlx5_ib2:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638547.491799] [ndv4:69197:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638547.491810] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x5a76fc0 [id=130 ref 1] uct_ud_iface_async_handler() to hash
[1650638547.491844] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 130 events 0x5 mode thread_spinlock
[1650638547.491875] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[39]=0x5a76570 using ud_verbs/mlx5_ib7:1 on worker 0x306d6f0
[1650638547.492200] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638547.492253] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638547.492381] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638547.492387] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638547.492857] [ndv4:69197:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib7:1
[1650638547.493724] [ndv4:69197:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638547.493015] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638547.493450] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638547.493458] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638547.493621] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638547.493626] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638547.494029] [ndv4:69197:0] ib_iface.c:994 UCX DEBUG iface=0x6883460: created UD QP 0x2b1a on mlx5_ib7:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638547.494038] [ndv4:69197:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650638547.494521] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638547.494968] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638547.494975] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638547.494994] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638547.494999] [ndv4:69197:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638547.494140] [ndv4:68758:0] ib_md.c:812 UCX DEBUG registered memory 0x2ad1ed5ec000..0x2ad1ed671000 on mlx5_ib2 lkey 0x80e00 rkey 0x80e00 access 0xf flags 0x3e4
[1650638547.494147] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ad1ed5ec018 of 544744 bytes with 128 elements
[1650638547.494151] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638547.495006] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x2f55500: adding gid fe80::15:5dff:fd34:1d to hash on device mlx5_ib2 port 1 index 0)
[1650638547.496157] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x2f55500: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 1)
[1650638547.496838] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x2f55500: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 2)
[1650638547.497528] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x2f55500: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 3)
[1650638547.497821] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x2f55500: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 4)
[1650638547.498165] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x2f55500: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 5)
[1650638547.495323] [ndv4:69197:0] ib_md.c:812 UCX DEBUG registered memory 0x2afd99e24000..0x2afd99ea9000 on mlx5_ib7 lkey 0x80c00 rkey 0x80c00 access 0xf flags 0x3e4
[1650638547.495333] [ndv4:69197:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2afd99e24018 of 544744 bytes with 128 elements
[1650638547.495341] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638547.496315] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x6883460: adding gid fe80::15:5dff:fd34:22 to hash on device mlx5_ib7 port 1 index 0)
[1650638547.497062] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x6883460: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 1)
[1650638547.498646] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x2f55500: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 6)
[1650638547.499008] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x2f55500: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 7)
[1650638547.499324] [ndv4:68758:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638547.499333] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x21901c0 [id=95 ref 1] uct_ud_iface_async_handler() to hash
[1650638547.499363] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 95 events 0x5 mode thread_spinlock
[1650638547.499373] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[14]=0x2f55500 using ud_verbs/mlx5_ib2:1 on worker 0x2a61550
[1650638547.499517] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638547.499524] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638547.499645] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638547.499651] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638547.500173] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib2:1
[1650638547.501302] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638547.501664] [ndv4:68758:0] ib_iface.c:994 UCX DEBUG iface=0x2185600: created UD QP 0x3057 on mlx5_ib2:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638547.501672] [ndv4:68758:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650638547.502276] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638547.502524] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638547.502530] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638547.510688] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638547.510698] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638547.511227] [ndv4:68758:0] ib_md.c:812 UCX DEBUG registered memory 0x2ad1ed671000..0x2ad1ed6f6000 on mlx5_ib2 lkey 0x80f00 rkey 0x80f00 access 0xf flags 0x3e4
[1650638547.511233] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ad1ed671018 of 544744 bytes with 128 elements
[1650638547.511237] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638547.511275] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x2185600: adding gid fe80::15:5dff:fd34:1d to hash on device mlx5_ib2 port 1 index 0)
[1650638547.511739] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x2185600: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 1)
[1650638547.512101] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x2185600: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 2)
[1650638547.512119] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x2185600: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 3)
[1650638547.512767] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x2185600: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 4)
[1650638547.513167] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x2185600: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 5)
[1650638547.513599] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x2185600: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 6)
[1650638547.513860] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x2185600: adding gid fe80:: to hash on device mlx5_ib2 port 1 index 7)
[1650638547.513867] [ndv4:68758:0] 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> }
[1650638547.513896] [ndv4:68758:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638547.513901] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x328c960 [id=96 ref 1] uct_ud_iface_async_handler() to hash
[1650638547.513924] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 96 events 0x5 mode thread_spinlock
[1650638547.513938] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[15]=0x2185600 using ud_mlx5/mlx5_ib2:1 on worker 0x2a61550
[1650638547.514042] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638547.514047] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638547.514145] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638547.514151] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638547.514706] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x6883460: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 2)
[1650638547.515070] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x6883460: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 3)
[1650638547.515397] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x6883460: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 4)
[1650638547.525012] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib3:1
[1650638547.525993] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x6883460: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 5)
[1650638547.526029] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x6883460: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 6)
[1650638547.526131] [ndv4:69197:0] ud_iface.c:393 UCX DEBUG iface 0x6883460: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 7)
[1650638547.526139] [ndv4:69197:0] 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> }
[1650638547.526273] [ndv4:69197:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638547.526280] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x6883f80 [id=131 ref 1] uct_ud_iface_async_handler() to hash
[1650638547.526302] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 131 events 0x5 mode thread_spinlock
[1650638547.526314] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[40]=0x6883460 using ud_mlx5/mlx5_ib7:1 on worker 0x306d6f0
[1650638547.526378] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool uct_scopy_iface_tx_mp: align 64, maxelems 4294967295, elemsize 736
[1650638547.526440] [ndv4:69197:0] ucp_worker.c:1159 UCX DEBUG created interface[41]=0x6193600 using cma/memory on worker 0x306d6f0
[1650638547.526447] [ndv4:69197:0] ucp_worker.c:982 UCX DEBUG selected scalable tl bitmap: 0x3ffffffffff 0x0 (42 tls)
[1650638547.527533] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650638547.527553] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.527556] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.527606] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.529467] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638547.529474] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650638547.530733] [ndv4:68758:0] ib_iface.c:994 UCX DEBUG iface=0x3fe60a0: created RC QP 0x303e on mlx5_ib3:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650638547.532003] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[16]=0x3fe60a0 using rc_verbs/mlx5_ib3:1 on worker 0x2a61550
[1650638547.532039] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638547.532045] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638547.532118] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638547.532123] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638547.532555] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib3:1
[1650638547.533959] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638547.533971] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.533974] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.534022] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.534527] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x4307010 of 8176 bytes with 127 elements
[1650638547.534820] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638547.534827] [ndv4:68758:0] 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> }
[1650638547.534862] [ndv4:68758:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib3 length=2048) failed: Invalid argument
[1650638547.534865] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638547.534876] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x21aaf80 [id=99 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638547.534905] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 99 events 0x1 mode thread_spinlock
[1650638547.534925] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[17]=0x4133030 using rc_mlx5/mlx5_ib3:1 on worker 0x2a61550
[1650638547.534971] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638547.534976] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638547.535172] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638547.535177] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638547.535473] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib3:1
[1650638547.536771] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638547.536788] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.536791] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.536846] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.537329] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650638547.537336] [ndv4:68758:0] 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> }
[1650638547.537368] [ndv4:68758:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib3 length=2048) failed: Invalid argument
[1650638547.537372] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638547.537380] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x413bfc0 [id=101 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638547.537403] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 101 events 0x1 mode thread_spinlock
[1650638547.538143] [ndv4:68758:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650638547.539080] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x6765fb0 [id=75 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638547.539112] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 75 events 0x0 mode thread_spinlock
[1650638547.539182] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x6241370 [id=76 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638547.539198] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 76 events 0x0 mode thread_spinlock
[1650638547.539425] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x2794a30 [id=77 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638547.539446] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 77 events 0x0 mode thread_spinlock
[1650638547.539465] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x2794a70 [id=79 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638547.539511] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 79 events 0x0 mode thread_spinlock
[1650638547.539562] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x2794ab0 [id=83 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638547.539579] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 83 events 0x0 mode thread_spinlock
[1650638547.539595] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x2794af0 [id=84 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638547.539600] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 84 events 0x0 mode thread_spinlock
[1650638547.539614] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x2794b30 [id=86 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638547.539619] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 86 events 0x0 mode thread_spinlock
[1650638547.539653] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x308ddc0 [id=90 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638547.539658] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 90 events 0x0 mode thread_spinlock
[1650638547.539674] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x308de00 [id=91 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638547.539678] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 91 events 0x0 mode thread_spinlock
[1650638547.539695] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x308de40 [id=93 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638547.539700] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 93 events 0x0 mode thread_spinlock
[1650638547.539734] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x6193f60 [id=97 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638547.539758] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 97 events 0x0 mode thread_spinlock
[1650638547.539774] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x6193fa0 [id=98 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638547.539792] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 98 events 0x0 mode thread_spinlock
[1650638547.539808] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x202f820 [id=100 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638547.539826] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 100 events 0x0 mode thread_spinlock
[1650638547.539862] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x202f860 [id=104 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638547.539880] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 104 events 0x0 mode thread_spinlock
[1650638547.539895] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x202f8a0 [id=105 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638547.539931] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 105 events 0x0 mode thread_spinlock
[1650638547.539945] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x202f8e0 [id=107 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638547.540006] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 107 events 0x0 mode thread_spinlock
[1650638547.540041] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x308de80 [id=111 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638547.540067] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 111 events 0x0 mode thread_spinlock
[1650638547.540084] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x308dec0 [id=112 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638547.540113] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 112 events 0x0 mode thread_spinlock
[1650638547.540265] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x308df00 [id=114 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638547.540294] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 114 events 0x0 mode thread_spinlock
[1650638547.540331] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x308df40 [id=118 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638547.540348] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 118 events 0x0 mode thread_spinlock
[1650638547.540364] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x308df80 [id=119 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638547.540379] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 119 events 0x0 mode thread_spinlock
[1650638547.540393] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x63bbe10 [id=121 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638547.540410] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 121 events 0x0 mode thread_spinlock
[1650638547.540447] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x63bbe50 [id=125 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638547.540464] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 125 events 0x0 mode thread_spinlock
[1650638547.540480] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x63bbe90 [id=126 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638547.540501] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 126 events 0x0 mode thread_spinlock
[1650638547.540515] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x63bbed0 [id=128 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638547.540531] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 128 events 0x0 mode thread_spinlock
[1650638547.545376] [ndv4:69197:0] async.c:228 UCX DEBUG added async handler 0x44f27e0 [id=132 ref 1] uct_rdmacm_cm_event_handler() to hash
[1650638547.545402] [ndv4:69197:0] async.c:506 UCX DEBUG listening to async event fd 132 events 0x1 mode thread_spinlock
[1650638547.545504] [ndv4:69197:0] rdmacm_cm.c:922 UCX DEBUG created rdmacm_cm 0x695e050 with event_channel 0x6193fe0 (fd=132)
[1650638547.545540] [ndv4:69197:0] tcp_sockcm.c:186 UCX DEBUG created tcp_sockcm 0x5e487e0
[1650638547.545558] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool ucp_requests: align 64, maxelems 4294967295, elemsize 440
[1650638547.545561] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool ucp_rkeys: align 64, maxelems 4294967295, elemsize 168
[1650638547.545566] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool ucp_am_bufs: align 64, maxelems 4294967295, elemsize 8344
[1650638547.545569] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool ucp_reg_bufs: align 64, maxelems 4294967295, elemsize 8208
[1650638547.545571] [ndv4:69197:0] mpool.c:88 UCX DEBUG mpool ucp_rndv_frags: align 512, maxelems 4294967295, elemsize 524304
[1650638547.545820] [ndv4:69197:0] parser.c:1893 UCX INFO UCX_* env variables: UCX_TLS=sysv,cma,ib UCX_POSIX_USE_PROC_LINK=n UCX_LOG_LEVEL=debug
[1650638547.546243] [ndv4:68758:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x4309050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0x3067
[1650638547.546674] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638547.546681] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638547.546690] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2ad1efef7008 of 151544 bytes with 1052 elements
[1650638547.550612] [ndv4:68758:0] ib_md.c:812 UCX DEBUG registered memory 0x2ad1ed800000..0x2ad1efe00000 on mlx5_ib3 lkey 0x80d00 rkey 0x80d00 access 0xf flags 0x3e4
[1650638547.550631] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2ad1ed800018 of 39845864 bytes with 4752 elements
[1650638547.550768] [ndv4:68758:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x4309050
[1650638547.550800] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[18]=0x4309050 using dc_mlx5/mlx5_ib3:1 on worker 0x2a61550
[1650638547.550952] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638547.550960] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638547.551079] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638547.551093] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638547.551639] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib3:1
[1650638547.552815] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638547.554508] [ndv4:68758:0] ib_iface.c:994 UCX DEBUG iface=0x44e5060: created UD QP 0x3047 on mlx5_ib3:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638547.555421] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638547.555660] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638547.555668] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638547.555826] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638547.555831] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638547.556320] [ndv4:68758:0] ib_md.c:812 UCX DEBUG registered memory 0x2ad1eff1c000..0x2ad1effa1000 on mlx5_ib3 lkey 0x80e00 rkey 0x80e00 access 0xf flags 0x3e4
[1650638547.556326] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ad1eff1c018 of 544744 bytes with 128 elements
[1650638547.556331] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638547.556513] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x44e5060: adding gid fe80::15:5dff:fd34:1e to hash on device mlx5_ib3 port 1 index 0)
[1650638547.556952] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x44e5060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 1)
[1650638547.567037] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x44e5060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 2)
[1650638547.567247] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x44e5060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 3)
[1650638547.567266] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x44e5060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 4)
[1650638547.567281] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x44e5060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 5)
[1650638547.587752] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x44e5060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 6)
[1650638547.587790] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x44e5060: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 7)
[1650638547.588769] [ndv4:68758:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638547.588778] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x44e5e60 [id=102 ref 1] uct_ud_iface_async_handler() to hash
[1650638547.588806] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 102 events 0x5 mode thread_spinlock
[1650638547.588816] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[19]=0x44e5060 using ud_verbs/mlx5_ib3:1 on worker 0x2a61550
[1650638547.588852] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638547.588858] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638547.588916] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638547.588922] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638547.591348] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib3:1
[1650638547.600535] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638547.601417] [ndv4:68758:0] ib_iface.c:994 UCX DEBUG iface=0x4603050: created UD QP 0x3049 on mlx5_ib3:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638547.601429] [ndv4:68758:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650638547.602320] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638547.602399] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638547.602406] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638547.602440] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638547.602445] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638547.602913] [ndv4:68758:0] ib_md.c:812 UCX DEBUG registered memory 0x2ad1effa1000..0x2ad1f0026000 on mlx5_ib3 lkey 0x80f00 rkey 0x80f00 access 0xf flags 0x3e4
[1650638547.602919] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ad1effa1018 of 544744 bytes with 128 elements
[1650638547.602924] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638547.602972] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x4603050: adding gid fe80::15:5dff:fd34:1e to hash on device mlx5_ib3 port 1 index 0)
[1650638547.602998] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x4603050: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 1)
[1650638547.603015] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x4603050: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 2)
[1650638547.603040] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x4603050: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 3)
[1650638547.603062] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x4603050: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 4)
[1650638547.603084] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x4603050: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 5)
[1650638547.603102] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x4603050: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 6)
[1650638547.603126] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x4603050: adding gid fe80:: to hash on device mlx5_ib3 port 1 index 7)
[1650638547.603137] [ndv4:68758:0] 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> }
[1650638547.603176] [ndv4:68758:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638547.603184] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x44e5f30 [id=103 ref 1] uct_ud_iface_async_handler() to hash
[1650638547.603241] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 103 events 0x5 mode thread_spinlock
[1650638547.603386] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[20]=0x4603050 using ud_mlx5/mlx5_ib3:1 on worker 0x2a61550
[1650638547.603445] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638547.603451] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638547.603558] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638547.603563] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638547.603644] [ndv4:68766:0] debug.c:1198 UCX DEBUG using signal stack 0x2b8e797ee000 size 141824
[1650638547.603724] [ndv4:68766:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650638547.603745] [ndv4:68766:0] 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 0x2b8e7965c000
[1650638547.603767] [ndv4:68766:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650638547.603777] [ndv4:68766:0] 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
[1650638547.603783] [ndv4:68766:0] module.c:253 UCX DEBUG loading modules for ucs
[1650638547.604000] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib4:1
[1650638547.605493] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650638547.605511] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.605515] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.605566] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.606110] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638547.606116] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650638547.607572] [ndv4:68758:0] ib_iface.c:994 UCX DEBUG iface=0x47030a0: created RC QP 0x3060 on mlx5_ib4:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650638547.608140] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[21]=0x47030a0 using rc_verbs/mlx5_ib4:1 on worker 0x2a61550
[1650638547.608157] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638547.608163] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638547.608306] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638547.608312] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638547.608072] [ndv4:68766:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650638547.608091] [ndv4:68766:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650638547.608125] [ndv4:68766:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650638547.608128] [ndv4:68766:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650638547.608135] [ndv4:68766:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650638547.608141] [ndv4:68766:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650638547.608145] [ndv4:68766:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650638547.608149] [ndv4:68766:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650638547.608151] [ndv4:68766:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650638547.608154] [ndv4:68766:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650638547.608156] [ndv4:68766:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650638547.608158] [ndv4:68766:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650638547.608166] [ndv4:68766:0] module.c:253 UCX DEBUG loading modules for uct
[1650638547.608981] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib4:1
[1650638547.610307] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638547.610326] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.610329] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.610382] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.611035] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x4a24010 of 8176 bytes with 127 elements
[1650638547.611307] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638547.611318] [ndv4:68758:0] 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> }
[1650638547.611356] [ndv4:68758:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib4 length=2048) failed: Invalid argument
[1650638547.611360] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638547.611372] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x46defc0 [id=106 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638547.611392] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 106 events 0x1 mode thread_spinlock
[1650638547.611403] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[22]=0x4850030 using rc_mlx5/mlx5_ib4:1 on worker 0x2a61550
[1650638547.611525] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638547.611531] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638547.611642] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638547.611646] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638547.611841] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib4:1
[1650638547.612867] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638547.612882] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.612885] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.612939] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.613467] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650638547.613474] [ndv4:68758:0] 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> }
[1650638547.613506] [ndv4:68758:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib4 length=2048) failed: Invalid argument
[1650638547.613510] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638547.613519] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x3fc1bf0 [id=108 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638547.613540] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 108 events 0x1 mode thread_spinlock
[1650638547.614182] [ndv4:68758:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650638547.614157] [ndv4:68766:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650638547.614532] [ndv4:68766:0] ucp_context.c:1117 UCX DEBUG closing md posix because it has no selected transport resources
[1650638547.614554] [ndv4:68766:0] ucp_context.c:1117 UCX DEBUG closing md self because it has no selected transport resources
[1650638547.614674] [ndv4:68766:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650638547.614687] [ndv4:68766:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650638547.614698] [ndv4:68766:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650638547.614709] [ndv4:68766:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650638547.614720] [ndv4:68766:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650638547.614731] [ndv4:68766:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650638547.614742] [ndv4:68766:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650638547.614776] [ndv4:68766:0] ucp_context.c:1117 UCX DEBUG closing md tcp because it has no selected transport resources
[1650638547.615139] [ndv4:68766:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650638547.622112] [ndv4:68758:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x4a26050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0x3076
[1650638547.622181] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638547.622188] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638547.622198] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2ad1f2827008 of 151544 bytes with 1052 elements
[1650638547.624181] [ndv4:68766:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650638547.624620] [ndv4:68766:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650638547.626524] [ndv4:68758:0] ib_md.c:812 UCX DEBUG registered memory 0x2ad1f0200000..0x2ad1f2800000 on mlx5_ib4 lkey 0x80c00 rkey 0x80c00 access 0xf flags 0x3e4
[1650638547.626549] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2ad1f0200018 of 39845864 bytes with 4752 elements
[1650638547.626901] [ndv4:68758:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x4a26050
[1650638547.626938] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[23]=0x4a26050 using dc_mlx5/mlx5_ib4:1 on worker 0x2a61550
[1650638547.627017] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638547.627027] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638547.627112] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638547.627116] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638547.627554] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib4:1
[1650638547.628549] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638547.628912] [ndv4:68758:0] ib_iface.c:994 UCX DEBUG iface=0x4c02060: created UD QP 0x306c on mlx5_ib4:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638547.629468] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638547.632151] [ndv4:68766:0] async.c:228 UCX DEBUG added async handler 0x1c15610 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650638547.632287] [ndv4:68766:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650638547.632404] [ndv4:68766:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650638547.632655] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638547.632665] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638547.632689] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638547.632753] [ndv4:68766:0] module.c:253 UCX DEBUG loading modules for ucm
[1650638547.632774] [ndv4:68766:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650638547.642808] [ndv4:68766:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638547.642821] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638547.643126] [ndv4:68766:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638547.643358] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638547.643365] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638547.646915] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638547.646925] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638547.654516] [ndv4:68766:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650638547.654524] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638547.656094] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638547.656103] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638547.656646] [ndv4:68758:0] ib_md.c:812 UCX DEBUG registered memory 0x2ad1f284c000..0x2ad1f28d1000 on mlx5_ib4 lkey 0x80d00 rkey 0x80d00 access 0xf flags 0x3e4
[1650638547.656653] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ad1f284c018 of 544744 bytes with 128 elements
[1650638547.656658] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638547.657119] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x4c02060: adding gid fe80::15:5dff:fd34:1f to hash on device mlx5_ib4 port 1 index 0)
[1650638547.657397] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x4c02060: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 1)
[1650638547.657605] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x4c02060: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 2)
[1650638547.664669] [ndv4:68766:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638547.664676] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638547.665497] [ndv4:68766:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638547.665501] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638547.666487] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x4c02060: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 3)
[1650638547.666845] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x4c02060: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 4)
[1650638547.680782] [ndv4:68766:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638547.680788] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638547.681457] [ndv4:68766:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638547.681461] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638547.681686] [ndv4:68766:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638547.681764] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x4c02060: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 5)
[1650638547.682102] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x4c02060: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 6)
[1650638547.682265] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x4c02060: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 7)
[1650638547.682621] [ndv4:68758:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638547.682630] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x3ee6fc0 [id=109 ref 1] uct_ud_iface_async_handler() to hash
[1650638547.682652] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 109 events 0x5 mode thread_spinlock
[1650638547.682665] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[24]=0x4c02060 using ud_verbs/mlx5_ib4:1 on worker 0x2a61550
[1650638547.682676] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638547.682680] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638547.682726] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638547.682731] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638547.682902] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib4:1
[1650638547.683977] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638547.684417] [ndv4:68758:0] ib_iface.c:994 UCX DEBUG iface=0x4d20460: created UD QP 0x306d on mlx5_ib4:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638547.684426] [ndv4:68758:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650638547.685137] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638547.685200] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638547.685205] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638547.685244] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638547.685249] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638547.685633] [ndv4:68758:0] ib_md.c:812 UCX DEBUG registered memory 0x2ad1f28d1000..0x2ad1f2956000 on mlx5_ib4 lkey 0x81100 rkey 0x81100 access 0xf flags 0x3e4
[1650638547.685639] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ad1f28d1018 of 544744 bytes with 128 elements
[1650638547.685643] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638547.685763] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x4d20460: adding gid fe80::15:5dff:fd34:1f to hash on device mlx5_ib4 port 1 index 0)
[1650638547.699371] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x4d20460: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 1)
[1650638547.699477] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x4d20460: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 2)
[1650638547.699596] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x4d20460: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 3)
[1650638547.699739] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x4d20460: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 4)
[1650638547.700250] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x4d20460: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 5)
[1650638547.704682] [ndv4:68766:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650638547.706533] [ndv4:68766:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650638547.706797] [ndv4:68766:0] async.c:228 UCX DEBUG added async handler 0x1be8ff0 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650638547.706820] [ndv4:68766:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650638547.706823] [ndv4:68766:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650638547.707032] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638547.707040] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638547.707044] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638547.707084] [ndv4:68766:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650638547.712935] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x4d20460: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 6)
[1650638547.713607] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x4d20460: adding gid fe80:: to hash on device mlx5_ib4 port 1 index 7)
[1650638547.713616] [ndv4:68758:0] 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> }
[1650638547.713646] [ndv4:68758:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638547.713650] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x3ee6ec0 [id=110 ref 1] uct_ud_iface_async_handler() to hash
[1650638547.713670] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 110 events 0x5 mode thread_spinlock
[1650638547.713679] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[25]=0x4d20460 using ud_mlx5/mlx5_ib4:1 on worker 0x2a61550
[1650638547.713748] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638547.713754] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638547.713870] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638547.713875] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638547.714448] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib5:1
[1650638547.714655] [ndv4:68766:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638547.714662] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638547.714937] [ndv4:68766:0] ib_md.c:1604 UCX DEBUG mlx5_ib1: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638547.714985] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638547.714992] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638547.715628] [ndv4:68766:0] topo.c:99 UCX DEBUG bus id 0x102000000 doesn't exist. sys_dev = 1
[1650638547.715633] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638547.715993] [ndv4:68766:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638547.715998] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638547.716397] [ndv4:68766:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638547.716401] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638547.715792] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650638547.715807] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.715810] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.715858] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.716393] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638547.716399] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650638547.717477] [ndv4:68758:0] ib_iface.c:994 UCX DEBUG iface=0x4e200a0: created RC QP 0x2bb4 on mlx5_ib5:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650638547.718491] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[26]=0x4e200a0 using rc_verbs/mlx5_ib5:1 on worker 0x2a61550
[1650638547.718626] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638547.718632] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638547.755223] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638547.755231] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638547.755895] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib5:1
[1650638547.757718] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638547.757731] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.757734] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.757783] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.758308] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x5141010 of 8176 bytes with 127 elements
[1650638547.758607] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638547.758614] [ndv4:68758:0] 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> }
[1650638547.758650] [ndv4:68758:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib5 length=2048) failed: Invalid argument
[1650638547.758654] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638547.758665] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x4e28db0 [id=113 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638547.758685] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 113 events 0x1 mode thread_spinlock
[1650638547.758694] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[27]=0x4f6d030 using rc_mlx5/mlx5_ib5:1 on worker 0x2a61550
[1650638547.758807] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638547.758812] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638547.759031] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638547.759036] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638547.760076] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib5:1
[1650638547.761478] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638547.761492] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.761495] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.761552] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.761970] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650638547.761976] [ndv4:68758:0] 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> }
[1650638547.762008] [ndv4:68758:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib5 length=2048) failed: Invalid argument
[1650638547.762012] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638547.762019] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x38a4970 [id=115 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638547.762040] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 115 events 0x1 mode thread_spinlock
[1650638547.762753] [ndv4:68758:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650638547.766032] [ndv4:68766:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638547.766039] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638547.766578] [ndv4:68766:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638547.766582] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638547.766797] [ndv4:68766:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638547.770564] [ndv4:68758:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x5143050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0x2bdd
[1650638547.770848] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638547.770854] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638547.770862] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2ad1f5159008 of 151544 bytes with 1052 elements
[1650638547.774784] [ndv4:68758:0] ib_md.c:812 UCX DEBUG registered memory 0x2ad1f2a00000..0x2ad1f5000000 on mlx5_ib5 lkey 0x80f00 rkey 0x80f00 access 0xf flags 0x3e4
[1650638547.774804] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2ad1f2a00018 of 39845864 bytes with 4752 elements
[1650638547.774940] [ndv4:68758:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x5143050
[1650638547.774970] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[28]=0x5143050 using dc_mlx5/mlx5_ib5:1 on worker 0x2a61550
[1650638547.775462] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638547.775472] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638547.775643] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638547.775648] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638547.786600] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib5:1
[1650638547.786803] [ndv4:68766:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib2 vendor_id: 0x15b3 device_id: 4124
[1650638547.787279] [ndv4:68766:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib2: disable ODP because it's not supported for DevX QP
[1650638547.787729] [ndv4:68766:0] async.c:228 UCX DEBUG added async handler 0x1c0dce0 [id=61 ref 1] uct_ib_async_event_handler() to hash
[1650638547.787753] [ndv4:68766:0] async.c:506 UCX DEBUG listening to async event fd 61 events 0x1 mode thread_spinlock
[1650638547.787757] [ndv4:68766:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib2' (InfiniBand channel adapter) with 1 ports
[1650638547.787776] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638547.788163] [ndv4:68758:0] ib_iface.c:994 UCX DEBUG iface=0x3f13480: created UD QP 0x2bbe on mlx5_ib5:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638547.788802] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638547.788933] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638547.788940] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638547.788984] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638547.788989] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638547.788055] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638547.788063] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638547.788069] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638547.788110] [ndv4:68766:0] ib_md.c:1319 UCX DEBUG mlx5_ib2: using registration cache
[1650638547.789571] [ndv4:68766:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638547.789578] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638547.789848] [ndv4:68766:0] ib_md.c:1604 UCX DEBUG mlx5_ib2: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638547.789975] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638547.789981] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638547.790979] [ndv4:68766:0] topo.c:99 UCX DEBUG bus id 0x103000000 doesn't exist. sys_dev = 2
[1650638547.790983] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638547.789344] [ndv4:68758:0] ib_md.c:812 UCX DEBUG registered memory 0x2ad1f517e000..0x2ad1f5203000 on mlx5_ib5 lkey 0x81000 rkey 0x81000 access 0xf flags 0x3e4
[1650638547.789350] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ad1f517e018 of 544744 bytes with 128 elements
[1650638547.789355] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638547.789386] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x3f13480: adding gid fe80::15:5dff:fd34:20 to hash on device mlx5_ib5 port 1 index 0)
[1650638547.789677] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x3f13480: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 1)
[1650638547.789987] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x3f13480: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 2)
[1650638547.804134] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x3f13480: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 3)
[1650638547.804440] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x3f13480: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 4)
[1650638547.804999] [ndv4:68766:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638547.805006] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638547.812660] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x3f13480: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 5)
[1650638547.812765] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x3f13480: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 6)
[1650638547.812780] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x3f13480: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 7)
[1650638547.813126] [ndv4:68758:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638547.813135] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x531f630 [id=116 ref 1] uct_ud_iface_async_handler() to hash
[1650638547.813160] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 116 events 0x5 mode thread_spinlock
[1650638547.813170] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[29]=0x3f13480 using ud_verbs/mlx5_ib5:1 on worker 0x2a61550
[1650638547.813339] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638547.813346] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638547.813541] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638547.813546] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638547.814149] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib5:1
[1650638547.814663] [ndv4:68766:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638547.814670] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638547.815712] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638547.816077] [ndv4:68758:0] ib_iface.c:994 UCX DEBUG iface=0x543d050: created UD QP 0x2bc1 on mlx5_ib5:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638547.816084] [ndv4:68758:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650638547.816672] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638547.816903] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638547.816910] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638547.817011] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638547.817016] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638547.817431] [ndv4:68758:0] ib_md.c:812 UCX DEBUG registered memory 0x2ad1f5203000..0x2ad1f5288000 on mlx5_ib5 lkey 0x81100 rkey 0x81100 access 0xf flags 0x3e4
[1650638547.817443] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ad1f5203018 of 544744 bytes with 128 elements
[1650638547.817448] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638547.818386] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x543d050: adding gid fe80::15:5dff:fd34:20 to hash on device mlx5_ib5 port 1 index 0)
[1650638547.818715] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x543d050: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 1)
[1650638547.818977] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x543d050: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 2)
[1650638547.817343] [ndv4:68766:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638547.817350] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638547.819310] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x543d050: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 3)
[1650638547.819329] [ndv4:68766:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638547.819335] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638547.819550] [ndv4:68766:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638547.829932] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x543d050: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 4)
[1650638547.830740] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x543d050: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 5)
[1650638547.831608] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x543d050: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 6)
[1650638547.832169] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x543d050: adding gid fe80:: to hash on device mlx5_ib5 port 1 index 7)
[1650638547.832175] [ndv4:68758:0] 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> }
[1650638547.832224] [ndv4:68758:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638547.832229] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x4dfb3a0 [id=117 ref 1] uct_ud_iface_async_handler() to hash
[1650638547.832281] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 117 events 0x5 mode thread_spinlock
[1650638547.832291] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[30]=0x543d050 using ud_mlx5/mlx5_ib5:1 on worker 0x2a61550
[1650638547.832529] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638547.832536] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638547.832689] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638547.832694] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638547.841382] [ndv4:68766:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib3 vendor_id: 0x15b3 device_id: 4124
[1650638547.841772] [ndv4:68766:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib3: disable ODP because it's not supported for DevX QP
[1650638547.841242] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib6:1
[1650638547.843302] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650638547.843322] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.843326] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.843376] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.843932] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638547.843939] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650638547.845244] [ndv4:68758:0] ib_iface.c:994 UCX DEBUG iface=0x553d0a0: created RC QP 0x2b28 on mlx5_ib6:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650638547.846300] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[31]=0x553d0a0 using rc_verbs/mlx5_ib6:1 on worker 0x2a61550
[1650638547.846522] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638547.846528] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638547.846590] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638547.846594] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638547.847451] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib6:1
[1650638547.849194] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638547.849229] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.849232] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.849282] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.849807] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x585e010 of 8176 bytes with 127 elements
[1650638547.850090] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638547.850098] [ndv4:68758:0] 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> }
[1650638547.850132] [ndv4:68758:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib6 length=2048) failed: Invalid argument
[1650638547.850137] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638547.850148] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x55182d0 [id=120 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638547.850168] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 120 events 0x1 mode thread_spinlock
[1650638547.850178] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[32]=0x568a030 using rc_mlx5/mlx5_ib6:1 on worker 0x2a61550
[1650638547.850382] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638547.850389] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638547.850523] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638547.850528] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638547.850735] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib6:1
[1650638547.851623] [ndv4:68766:0] async.c:228 UCX DEBUG added async handler 0x1c15730 [id=63 ref 1] uct_ib_async_event_handler() to hash
[1650638547.851646] [ndv4:68766:0] async.c:506 UCX DEBUG listening to async event fd 63 events 0x1 mode thread_spinlock
[1650638547.851649] [ndv4:68766:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib3' (InfiniBand channel adapter) with 1 ports
[1650638547.851706] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638547.851721] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.851725] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.851778] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.852133] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650638547.852140] [ndv4:68758:0] 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> }
[1650638547.852171] [ndv4:68758:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib6 length=2048) failed: Invalid argument
[1650638547.852176] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638547.852183] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x5692f60 [id=122 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638547.852201] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 122 events 0x1 mode thread_spinlock
[1650638547.852884] [ndv4:68758:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650638547.852005] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638547.852013] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638547.852018] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638547.852061] [ndv4:68766:0] ib_md.c:1319 UCX DEBUG mlx5_ib3: using registration cache
[1650638547.860761] [ndv4:68758:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x5860050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0x2b2f
[1650638547.860815] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638547.860821] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638547.860830] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2ad1f7a8b008 of 151544 bytes with 1052 elements
[1650638547.864768] [ndv4:68758:0] ib_md.c:812 UCX DEBUG registered memory 0x2ad1f5400000..0x2ad1f7a00000 on mlx5_ib6 lkey 0x80f00 rkey 0x80f00 access 0xf flags 0x3e4
[1650638547.864786] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2ad1f5400018 of 39845864 bytes with 4752 elements
[1650638547.864923] [ndv4:68758:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x5860050
[1650638547.864953] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[33]=0x5860050 using dc_mlx5/mlx5_ib6:1 on worker 0x2a61550
[1650638547.864995] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638547.865003] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638547.865070] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638547.865074] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638547.865326] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib6:1
[1650638547.866135] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638547.866533] [ndv4:68758:0] ib_iface.c:994 UCX DEBUG iface=0x5a3c060: created UD QP 0x2b34 on mlx5_ib6:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638547.867091] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638547.867147] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638547.867154] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638547.867175] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638547.867180] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638547.867580] [ndv4:68758:0] ib_md.c:812 UCX DEBUG registered memory 0x2ad1f7ab0000..0x2ad1f7b35000 on mlx5_ib6 lkey 0x81000 rkey 0x81000 access 0xf flags 0x3e4
[1650638547.867586] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ad1f7ab0018 of 544744 bytes with 128 elements
[1650638547.867591] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638547.867648] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x5a3c060: adding gid fe80::15:5dff:fd34:21 to hash on device mlx5_ib6 port 1 index 0)
[1650638547.867701] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x5a3c060: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 1)
[1650638547.867747] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x5a3c060: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 2)
[1650638547.867775] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x5a3c060: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 3)
[1650638547.867794] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x5a3c060: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 4)
[1650638547.867818] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x5a3c060: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 5)
[1650638547.867844] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x5a3c060: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 6)
[1650638547.867873] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x5a3c060: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 7)
[1650638547.868129] [ndv4:68758:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638547.868139] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x5a3cf20 [id=123 ref 1] uct_ud_iface_async_handler() to hash
[1650638547.868167] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 123 events 0x5 mode thread_spinlock
[1650638547.868179] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[34]=0x5a3c060 using ud_verbs/mlx5_ib6:1 on worker 0x2a61550
[1650638547.868193] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638547.868199] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638547.868262] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638547.868267] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638547.868415] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib6:1
[1650638547.869133] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638547.869476] [ndv4:68758:0] ib_iface.c:994 UCX DEBUG iface=0x4d4d3a0: created UD QP 0x2b35 on mlx5_ib6:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638547.869484] [ndv4:68758:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650638547.869983] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638547.870017] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638547.870023] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638547.870037] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638547.870042] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638547.870378] [ndv4:68758:0] ib_md.c:812 UCX DEBUG registered memory 0x2ad1f7b35000..0x2ad1f7bba000 on mlx5_ib6 lkey 0x81100 rkey 0x81100 access 0xf flags 0x3e4
[1650638547.870385] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ad1f7b35018 of 544744 bytes with 128 elements
[1650638547.870389] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638547.870617] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x4d4d3a0: adding gid fe80::15:5dff:fd34:21 to hash on device mlx5_ib6 port 1 index 0)
[1650638547.870814] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x4d4d3a0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 1)
[1650638547.870944] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x4d4d3a0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 2)
[1650638547.882488] [ndv4:68766:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638547.882499] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638547.883008] [ndv4:68766:0] ib_md.c:1604 UCX DEBUG mlx5_ib3: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638547.884461] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x4d4d3a0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 3)
[1650638547.885118] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x4d4d3a0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 4)
[1650638547.885622] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x4d4d3a0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 5)
[1650638547.885979] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x4d4d3a0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 6)
[1650638547.886751] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x4d4d3a0: adding gid fe80:: to hash on device mlx5_ib6 port 1 index 7)
[1650638547.886759] [ndv4:68758:0] 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> }
[1650638547.886791] [ndv4:68758:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638547.886795] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x55180e0 [id=124 ref 1] uct_ud_iface_async_handler() to hash
[1650638547.886816] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 124 events 0x5 mode thread_spinlock
[1650638547.887052] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[35]=0x4d4d3a0 using ud_mlx5/mlx5_ib6:1 on worker 0x2a61550
[1650638547.887111] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638547.887117] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638547.887355] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638547.887362] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638547.887935] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib7:1
[1650638547.890018] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650638547.890040] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.890043] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.890108] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.890768] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638547.890774] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650638547.892071] [ndv4:68758:0] ib_iface.c:994 UCX DEBUG iface=0x5c5a0a0: created RC QP 0x2b1b on mlx5_ib7:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650638547.892076] [ndv4:69236:0] debug.c:1198 UCX DEBUG using signal stack 0x2abade93e000 size 141824
[1650638547.892156] [ndv4:69236:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650638547.892179] [ndv4:69236:0] 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 0x2abade792000
[1650638547.892196] [ndv4:69236:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650638547.892290] [ndv4:69236:0] 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
[1650638547.892299] [ndv4:69236:0] module.c:253 UCX DEBUG loading modules for ucs
[1650638547.893334] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638547.893347] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638547.893529] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[36]=0x5c5a0a0 using rc_verbs/mlx5_ib7:1 on worker 0x2a61550
[1650638547.893575] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638547.893581] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638547.893824] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638547.893830] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638547.894748] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib7:1
[1650638547.894834] [ndv4:69236:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650638547.894853] [ndv4:69236:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650638547.894887] [ndv4:69236:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650638547.894891] [ndv4:69236:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650638547.894898] [ndv4:69236:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650638547.894904] [ndv4:69236:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650638547.894907] [ndv4:69236:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650638547.894911] [ndv4:69236:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650638547.894914] [ndv4:69236:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650638547.894916] [ndv4:69236:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650638547.894918] [ndv4:69236:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650638547.894921] [ndv4:69236:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650638547.894928] [ndv4:69236:0] module.c:253 UCX DEBUG loading modules for uct
[1650638547.895703] [ndv4:68766:0] topo.c:99 UCX DEBUG bus id 0x104000000 doesn't exist. sys_dev = 3
[1650638547.895713] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638547.897011] [ndv4:68766:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638547.897017] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638547.896852] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638547.896872] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.896875] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.896925] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.897500] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x5f7b010 of 8176 bytes with 127 elements
[1650638547.897784] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638547.897802] [ndv4:68758:0] 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> }
[1650638547.897846] [ndv4:68758:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib7 length=2048) failed: Invalid argument
[1650638547.897850] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638547.897867] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x5b5a8c0 [id=127 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638547.897891] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 127 events 0x1 mode thread_spinlock
[1650638547.897903] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[37]=0x5da7030 using rc_mlx5/mlx5_ib7:1 on worker 0x2a61550
[1650638547.900722] [ndv4:69236:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650638547.901076] [ndv4:69236:0] ucp_context.c:1117 UCX DEBUG closing md posix because it has no selected transport resources
[1650638547.901096] [ndv4:69236:0] ucp_context.c:1117 UCX DEBUG closing md self because it has no selected transport resources
[1650638547.905921] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638547.905931] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638547.906196] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638547.906201] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638547.907038] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib7:1
[1650638547.907337] [ndv4:69236:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650638547.907358] [ndv4:69236:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650638547.907370] [ndv4:69236:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650638547.907380] [ndv4:69236:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650638547.907391] [ndv4:69236:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650638547.907401] [ndv4:69236:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650638547.907413] [ndv4:69236:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650638547.907447] [ndv4:69236:0] ucp_context.c:1117 UCX DEBUG closing md tcp because it has no selected transport resources
[1650638547.907827] [ndv4:69236:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650638547.910080] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638547.910098] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638547.910101] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638547.910182] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638547.910709] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650638547.910717] [ndv4:68758:0] 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> }
[1650638547.910744] [ndv4:68758:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib7 length=2048) failed: Invalid argument
[1650638547.910748] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638547.910756] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x5c35760 [id=129 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638547.910777] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 129 events 0x1 mode thread_spinlock
[1650638547.911451] [ndv4:68758:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650638547.919725] [ndv4:68758:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x5f7d050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0x2b22
[1650638547.919793] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638547.919800] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638547.919812] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2ad1fa3bd008 of 151544 bytes with 1052 elements
[1650638547.923766] [ndv4:68758:0] ib_md.c:812 UCX DEBUG registered memory 0x2ad1f7c00000..0x2ad1fa200000 on mlx5_ib7 lkey 0x80d00 rkey 0x80d00 access 0xf flags 0x3e4
[1650638547.923785] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2ad1f7c00018 of 39845864 bytes with 4752 elements
[1650638547.923923] [ndv4:68758:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x5f7d050
[1650638547.923957] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[38]=0x5f7d050 using dc_mlx5/mlx5_ib7:1 on worker 0x2a61550
[1650638547.924033] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638547.924042] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638547.924335] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638547.924340] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638547.924740] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib7:1
[1650638547.925998] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638547.926415] [ndv4:68758:0] ib_iface.c:994 UCX DEBUG iface=0x546a570: created UD QP 0x2b27 on mlx5_ib7:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638547.926971] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638547.927344] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638547.927352] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638547.927453] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638547.927458] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638547.927824] [ndv4:68758:0] ib_md.c:812 UCX DEBUG registered memory 0x2ad1fa3e2000..0x2ad1fa467000 on mlx5_ib7 lkey 0x80e00 rkey 0x80e00 access 0xf flags 0x3e4
[1650638547.927830] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ad1fa3e2018 of 544744 bytes with 128 elements
[1650638547.927835] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638547.928202] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x546a570: adding gid fe80::15:5dff:fd34:22 to hash on device mlx5_ib7 port 1 index 0)
[1650638547.928446] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x546a570: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 1)
[1650638547.929284] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x546a570: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 2)
[1650638547.930052] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x546a570: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 3)
[1650638547.930595] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x546a570: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 4)
[1650638547.930322] [ndv4:68766:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638547.930329] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638547.932916] [ndv4:68766:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638547.932922] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638547.934009] [ndv4:69236:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650638547.934493] [ndv4:69236:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650638547.934267] [ndv4:68766:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638547.934272] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638547.934538] [ndv4:68766:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638547.940482] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x546a570: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 5)
[1650638547.942423] [ndv4:69236:0] async.c:228 UCX DEBUG added async handler 0x14a7610 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650638547.942526] [ndv4:69236:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650638547.942977] [ndv4:69236:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650638547.943331] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638547.943346] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638547.943393] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638547.943479] [ndv4:69236:0] module.c:253 UCX DEBUG loading modules for ucm
[1650638547.943512] [ndv4:69236:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650638547.946220] [ndv4:69236:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638547.946233] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638547.946553] [ndv4:69236:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638547.946857] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638547.946863] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638547.947932] [ndv4:69236:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650638547.947939] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638547.954172] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x546a570: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 6)
[1650638547.955230] [ndv4:68766:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib4 vendor_id: 0x15b3 device_id: 4124
[1650638547.955715] [ndv4:68766:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib4: disable ODP because it's not supported for DevX QP
[1650638547.956090] [ndv4:68766:0] async.c:228 UCX DEBUG added async handler 0x22f7610 [id=65 ref 1] uct_ib_async_event_handler() to hash
[1650638547.956115] [ndv4:68766:0] async.c:506 UCX DEBUG listening to async event fd 65 events 0x1 mode thread_spinlock
[1650638547.956119] [ndv4:68766:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib4' (InfiniBand channel adapter) with 1 ports
[1650638547.957529] [ndv4:69236:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638547.957536] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638547.957867] [ndv4:69236:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638547.957871] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638547.958376] [ndv4:69236:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638547.958381] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638547.958454] [ndv4:69236:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638547.958457] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638547.958767] [ndv4:69236:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638547.956413] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638547.956422] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638547.956433] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638547.956490] [ndv4:68766:0] ib_md.c:1319 UCX DEBUG mlx5_ib4: using registration cache
[1650638547.963283] [ndv4:68766:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638547.963290] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638547.963580] [ndv4:68766:0] ib_md.c:1604 UCX DEBUG mlx5_ib4: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638547.963595] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638547.963601] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638547.963740] [ndv4:68766:0] topo.c:99 UCX DEBUG bus id 0x105000000 doesn't exist. sys_dev = 4
[1650638547.963744] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638547.963849] [ndv4:68766:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650638547.963853] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638547.963970] [ndv4:68766:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650638547.963973] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638547.964080] [ndv4:68766:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650638547.964083] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638547.964188] [ndv4:68766:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650638547.964191] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638547.964433] [ndv4:68766:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638547.966658] [ndv4:69236:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650638547.967003] [ndv4:69236:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650638547.969341] [ndv4:69236:0] async.c:228 UCX DEBUG added async handler 0x147aff0 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650638547.969363] [ndv4:69236:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650638547.969367] [ndv4:69236:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650638547.969680] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638547.969689] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638547.969694] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638547.969735] [ndv4:69236:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650638547.972622] [ndv4:69236:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638547.972630] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638547.972898] [ndv4:69236:0] ib_md.c:1604 UCX DEBUG mlx5_ib1: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638547.972925] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638547.972930] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638547.973017] [ndv4:69236:0] topo.c:99 UCX DEBUG bus id 0x102000000 doesn't exist. sys_dev = 1
[1650638547.973021] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638547.974294] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x546a570: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 7)
[1650638547.974674] [ndv4:68758:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638547.974685] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x546afc0 [id=130 ref 1] uct_ud_iface_async_handler() to hash
[1650638547.974710] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 130 events 0x5 mode thread_spinlock
[1650638547.974732] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[39]=0x546a570 using ud_verbs/mlx5_ib7:1 on worker 0x2a61550
[1650638547.974754] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638547.974760] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638547.974938] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638547.974944] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638547.975504] [ndv4:68758:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib7:1
[1650638547.976314] [ndv4:69236:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638547.976321] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638547.976389] [ndv4:69236:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638547.976392] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638547.977729] [ndv4:68758:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638547.978234] [ndv4:68758:0] ib_iface.c:994 UCX DEBUG iface=0x6277460: created UD QP 0x2b28 on mlx5_ib7:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638547.978244] [ndv4:68758:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650638547.979052] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638547.979431] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638547.979437] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638547.979448] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638547.979452] [ndv4:68758:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638547.979863] [ndv4:68758:0] ib_md.c:812 UCX DEBUG registered memory 0x2ad1fa467000..0x2ad1fa4ec000 on mlx5_ib7 lkey 0x80f00 rkey 0x80f00 access 0xf flags 0x3e4
[1650638547.979869] [ndv4:68758:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2ad1fa467018 of 544744 bytes with 128 elements
[1650638547.979873] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638547.980557] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x6277460: adding gid fe80::15:5dff:fd34:22 to hash on device mlx5_ib7 port 1 index 0)
[1650638547.981002] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x6277460: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 1)
[1650638547.990029] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x6277460: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 2)
[1650638547.996035] [ndv4:69236:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638547.996045] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638547.996975] [ndv4:68755:0] debug.c:1198 UCX DEBUG using signal stack 0x2af633c84000 size 141824
[1650638547.997057] [ndv4:68755:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650638547.997078] [ndv4:68755:0] 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 0x2af633af2000
[1650638547.997097] [ndv4:68755:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650638547.997106] [ndv4:68755:0] 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
[1650638547.997111] [ndv4:68755:0] module.c:253 UCX DEBUG loading modules for ucs
[1650638547.998897] [ndv4:69236:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638547.998904] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638547.999406] [ndv4:69236:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638547.999644] [ndv4:68755:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650638547.999666] [ndv4:68755:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650638547.999700] [ndv4:68755:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650638547.999703] [ndv4:68755:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650638547.999709] [ndv4:68755:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650638547.999716] [ndv4:68755:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650638547.999719] [ndv4:68755:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650638547.999723] [ndv4:68755:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650638547.999726] [ndv4:68755:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650638547.999728] [ndv4:68755:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650638547.999731] [ndv4:68755:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650638547.999733] [ndv4:68755:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650638547.999741] [ndv4:68755:0] module.c:253 UCX DEBUG loading modules for uct
[1650638548.005750] [ndv4:68755:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650638548.006153] [ndv4:68755:0] ucp_context.c:1117 UCX DEBUG closing md posix because it has no selected transport resources
[1650638548.006178] [ndv4:68755:0] ucp_context.c:1117 UCX DEBUG closing md self because it has no selected transport resources
[1650638548.006486] [ndv4:68755:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650638548.006499] [ndv4:68755:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650638548.006510] [ndv4:68755:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650638548.006520] [ndv4:68755:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650638548.006556] [ndv4:68755:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650638548.006566] [ndv4:68755:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650638548.006577] [ndv4:68755:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650638548.006613] [ndv4:68755:0] ucp_context.c:1117 UCX DEBUG closing md tcp because it has no selected transport resources
[1650638548.007058] [ndv4:68755:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650638548.008692] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x6277460: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 3)
[1650638548.008937] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x6277460: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 4)
[1650638548.012546] [ndv4:68766:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib5 vendor_id: 0x15b3 device_id: 4124
[1650638548.013060] [ndv4:68766:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib5: disable ODP because it's not supported for DevX QP
[1650638548.013552] [ndv4:68766:0] async.c:228 UCX DEBUG added async handler 0x1c11420 [id=67 ref 1] uct_ib_async_event_handler() to hash
[1650638548.013579] [ndv4:68766:0] async.c:506 UCX DEBUG listening to async event fd 67 events 0x1 mode thread_spinlock
[1650638548.013582] [ndv4:68766:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib5' (InfiniBand channel adapter) with 1 ports
[1650638548.013895] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638548.013906] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638548.013921] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.013991] [ndv4:68766:0] ib_md.c:1319 UCX DEBUG mlx5_ib5: using registration cache
[1650638548.015052] [ndv4:68766:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.015058] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.014691] [ndv4:68755:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650638548.015035] [ndv4:68755:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650638548.015359] [ndv4:68766:0] ib_md.c:1604 UCX DEBUG mlx5_ib5: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.015487] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638548.015493] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638548.017022] [ndv4:68766:0] topo.c:99 UCX DEBUG bus id 0x106000000 doesn't exist. sys_dev = 5
[1650638548.017032] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638548.023048] [ndv4:68755:0] async.c:228 UCX DEBUG added async handler 0xcaa2c0 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650638548.023152] [ndv4:68755:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650638548.023530] [ndv4:68755:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650638548.023749] [ndv4:68755:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.023764] [ndv4:68755:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.023790] [ndv4:68755:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.023848] [ndv4:68755:0] module.c:253 UCX DEBUG loading modules for ucm
[1650638548.023870] [ndv4:68755:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650638548.026401] [ndv4:68755:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.026415] [ndv4:68755:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.026647] [ndv4:68755:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.026668] [ndv4:68755:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.026674] [ndv4:68755:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.026862] [ndv4:68755:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650638548.026867] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.026991] [ndv4:68755:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.026995] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.027168] [ndv4:68755:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.027174] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.027258] [ndv4:68755:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.027262] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.027436] [ndv4:68755:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.027439] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.027725] [ndv4:68755:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.030142] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x6277460: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 5)
[1650638548.030336] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x6277460: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 6)
[1650638548.030553] [ndv4:68758:0] ud_iface.c:393 UCX DEBUG iface 0x6277460: adding gid fe80:: to hash on device mlx5_ib7 port 1 index 7)
[1650638548.030578] [ndv4:68758:0] 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> }
[1650638548.030625] [ndv4:68758:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638548.030634] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x6277f80 [id=131 ref 1] uct_ud_iface_async_handler() to hash
[1650638548.030660] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 131 events 0x5 mode thread_spinlock
[1650638548.030709] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[40]=0x6277460 using ud_mlx5/mlx5_ib7:1 on worker 0x2a61550
[1650638548.030814] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool uct_scopy_iface_tx_mp: align 64, maxelems 4294967295, elemsize 736
[1650638548.030871] [ndv4:68758:0] ucp_worker.c:1159 UCX DEBUG created interface[41]=0x5b87600 using cma/memory on worker 0x2a61550
[1650638548.030877] [ndv4:68758:0] ucp_worker.c:982 UCX DEBUG selected scalable tl bitmap: 0x3ffffffffff 0x0 (42 tls)
[1650638548.035842] [ndv4:68755:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650638548.036174] [ndv4:68755:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650638548.036767] [ndv4:69236:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib2 vendor_id: 0x15b3 device_id: 4124
[1650638548.037295] [ndv4:69236:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib2: disable ODP because it's not supported for DevX QP
[1650638548.037705] [ndv4:69236:0] async.c:228 UCX DEBUG added async handler 0x149fce0 [id=61 ref 1] uct_ib_async_event_handler() to hash
[1650638548.037730] [ndv4:69236:0] async.c:506 UCX DEBUG listening to async event fd 61 events 0x1 mode thread_spinlock
[1650638548.037733] [ndv4:69236:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib2' (InfiniBand channel adapter) with 1 ports
[1650638548.037947] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638548.037962] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638548.037974] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.038029] [ndv4:69236:0] ib_md.c:1319 UCX DEBUG mlx5_ib2: using registration cache
[1650638548.039511] [ndv4:69236:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.039518] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.039776] [ndv4:69236:0] ib_md.c:1604 UCX DEBUG mlx5_ib2: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.039955] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638548.039965] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638548.040231] [ndv4:69236:0] topo.c:99 UCX DEBUG bus id 0x103000000 doesn't exist. sys_dev = 2
[1650638548.040236] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.040302] [ndv4:69236:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.040305] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.040364] [ndv4:69236:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.040367] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.040424] [ndv4:69236:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.040427] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.040484] [ndv4:69236:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.040487] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.040881] [ndv4:69236:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.043443] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x6159fb0 [id=75 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638548.043471] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 75 events 0x0 mode thread_spinlock
[1650638548.043603] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x5c35370 [id=76 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638548.043627] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 76 events 0x0 mode thread_spinlock
[1650638548.043689] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x2188510 [id=77 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638548.043709] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 77 events 0x0 mode thread_spinlock
[1650638548.043751] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x2188550 [id=79 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638548.043767] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 79 events 0x0 mode thread_spinlock
[1650638548.043815] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x2188590 [id=83 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638548.043832] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 83 events 0x0 mode thread_spinlock
[1650638548.043847] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x21885d0 [id=84 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638548.043861] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 84 events 0x0 mode thread_spinlock
[1650638548.043876] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x2188610 [id=86 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638548.043887] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 86 events 0x0 mode thread_spinlock
[1650638548.043922] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x219ad00 [id=90 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638548.043939] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 90 events 0x0 mode thread_spinlock
[1650638548.043954] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x219ad40 [id=91 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638548.043971] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 91 events 0x0 mode thread_spinlock
[1650638548.043987] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x219ad80 [id=93 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638548.044006] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 93 events 0x0 mode thread_spinlock
[1650638548.044040] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x5b87f60 [id=97 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638548.044060] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 97 events 0x0 mode thread_spinlock
[1650638548.044074] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x5b87fa0 [id=98 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638548.044088] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 98 events 0x0 mode thread_spinlock
[1650638548.044101] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x3fee910 [id=100 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638548.044115] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 100 events 0x0 mode thread_spinlock
[1650638548.044149] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x3fee950 [id=104 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638548.044165] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 104 events 0x0 mode thread_spinlock
[1650638548.044181] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x3fee990 [id=105 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638548.044195] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 105 events 0x0 mode thread_spinlock
[1650638548.044224] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x3fee9d0 [id=107 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638548.044267] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 107 events 0x0 mode thread_spinlock
[1650638548.044302] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x219adc0 [id=111 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638548.044321] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 111 events 0x0 mode thread_spinlock
[1650638548.044413] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x219ae00 [id=112 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638548.044437] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 112 events 0x0 mode thread_spinlock
[1650638548.044452] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x219ae40 [id=114 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638548.044467] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 114 events 0x0 mode thread_spinlock
[1650638548.044501] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x219ae80 [id=118 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638548.044519] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 118 events 0x0 mode thread_spinlock
[1650638548.044533] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x219aec0 [id=119 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638548.044556] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 119 events 0x0 mode thread_spinlock
[1650638548.044569] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x5dafe10 [id=121 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638548.044588] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 121 events 0x0 mode thread_spinlock
[1650638548.044622] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x5dafe50 [id=125 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638548.044643] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 125 events 0x0 mode thread_spinlock
[1650638548.044657] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x5dafe90 [id=126 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638548.044670] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 126 events 0x0 mode thread_spinlock
[1650638548.044683] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x5dafed0 [id=128 ref 1] ucp_worker_iface_async_fd_event() to hash
[1650638548.044702] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 128 events 0x0 mode thread_spinlock
[1650638548.048129] [ndv4:68755:0] async.c:228 UCX DEBUG added async handler 0xcaa440 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650638548.048157] [ndv4:68755:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650638548.048161] [ndv4:68755:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650638548.048506] [ndv4:68755:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638548.048517] [ndv4:68755:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638548.048525] [ndv4:68755:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.048582] [ndv4:68755:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650638548.049335] [ndv4:68758:0] async.c:228 UCX DEBUG added async handler 0x3ee67e0 [id=132 ref 1] uct_rdmacm_cm_event_handler() to hash
[1650638548.049359] [ndv4:68758:0] async.c:506 UCX DEBUG listening to async event fd 132 events 0x1 mode thread_spinlock
[1650638548.049451] [ndv4:68758:0] rdmacm_cm.c:922 UCX DEBUG created rdmacm_cm 0x6352050 with event_channel 0x5b87fe0 (fd=132)
[1650638548.049482] [ndv4:68758:0] tcp_sockcm.c:186 UCX DEBUG created tcp_sockcm 0x583c7e0
[1650638548.049498] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool ucp_requests: align 64, maxelems 4294967295, elemsize 440
[1650638548.049500] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool ucp_rkeys: align 64, maxelems 4294967295, elemsize 168
[1650638548.049504] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool ucp_am_bufs: align 64, maxelems 4294967295, elemsize 8344
[1650638548.049507] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool ucp_reg_bufs: align 64, maxelems 4294967295, elemsize 8208
[1650638548.049510] [ndv4:68758:0] mpool.c:88 UCX DEBUG mpool ucp_rndv_frags: align 512, maxelems 4294967295, elemsize 524304
[1650638548.049734] [ndv4:68758:0] parser.c:1893 UCX INFO UCX_* env variables: UCX_TLS=sysv,cma,ib UCX_POSIX_USE_PROC_LINK=n UCX_LOG_LEVEL=debug
[1650638548.048622] [ndv4:68766:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650638548.048629] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638548.050636] [ndv4:68755:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.050643] [ndv4:68755:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.050843] [ndv4:68755:0] ib_md.c:1604 UCX DEBUG mlx5_ib1: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.050914] [ndv4:68755:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638548.050920] [ndv4:68755:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638548.051272] [ndv4:68755:0] topo.c:99 UCX DEBUG bus id 0x102000000 doesn't exist. sys_dev = 1
[1650638548.051278] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.051884] [ndv4:68755:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638548.051888] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.052662] [ndv4:68755:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638548.052668] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.053203] [ndv4:68755:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638548.053225] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.063634] [ndv4:69236:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib3 vendor_id: 0x15b3 device_id: 4124
[1650638548.066953] [ndv4:69236:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib3: disable ODP because it's not supported for DevX QP
[1650638548.068452] [ndv4:69236:0] async.c:228 UCX DEBUG added async handler 0x14a7730 [id=63 ref 1] uct_ib_async_event_handler() to hash
[1650638548.068474] [ndv4:69236:0] async.c:506 UCX DEBUG listening to async event fd 63 events 0x1 mode thread_spinlock
[1650638548.068477] [ndv4:69236:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib3' (InfiniBand channel adapter) with 1 ports
[1650638548.068517] [ndv4:68755:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638548.068525] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.068763] [ndv4:68755:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.070288] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638548.070297] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638548.070302] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.070346] [ndv4:69236:0] ib_md.c:1319 UCX DEBUG mlx5_ib3: using registration cache
[1650638548.074687] [ndv4:68755:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib2 vendor_id: 0x15b3 device_id: 4124
[1650638548.075024] [ndv4:68755:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib2: disable ODP because it's not supported for DevX QP
[1650638548.075621] [ndv4:68755:0] async.c:228 UCX DEBUG added async handler 0xcaad60 [id=61 ref 1] uct_ib_async_event_handler() to hash
[1650638548.075644] [ndv4:68755:0] async.c:506 UCX DEBUG listening to async event fd 61 events 0x1 mode thread_spinlock
[1650638548.075647] [ndv4:68755:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib2' (InfiniBand channel adapter) with 1 ports
[1650638548.075817] [ndv4:68755:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638548.075828] [ndv4:68755:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638548.075836] [ndv4:68755:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.075924] [ndv4:68755:0] ib_md.c:1319 UCX DEBUG mlx5_ib2: using registration cache
[1650638548.075957] [ndv4:68766:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650638548.075965] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638548.076128] [ndv4:68766:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650638548.076133] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638548.076452] [ndv4:68766:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650638548.076456] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638548.076728] [ndv4:68766:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.080441] [ndv4:69236:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.080449] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.081403] [ndv4:69236:0] ib_md.c:1604 UCX DEBUG mlx5_ib3: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.081488] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638548.081495] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638548.090421] [ndv4:68754:0] debug.c:1198 UCX DEBUG using signal stack 0x2afb59707000 size 141824
[1650638548.090505] [ndv4:68754:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650638548.090525] [ndv4:68754:0] 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 0x2afb5955b000
[1650638548.090548] [ndv4:68754:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650638548.090557] [ndv4:68754:0] 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
[1650638548.090563] [ndv4:68754:0] module.c:253 UCX DEBUG loading modules for ucs
[1650638548.091830] [ndv4:68755:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.091846] [ndv4:68755:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.092103] [ndv4:68755:0] ib_md.c:1604 UCX DEBUG mlx5_ib2: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.092138] [ndv4:68755:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638548.092144] [ndv4:68755:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638548.091617] [ndv4:69236:0] topo.c:99 UCX DEBUG bus id 0x104000000 doesn't exist. sys_dev = 3
[1650638548.091628] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638548.093593] [ndv4:68754:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650638548.093612] [ndv4:68754:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650638548.093650] [ndv4:68754:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650638548.093653] [ndv4:68754:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650638548.093659] [ndv4:68754:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650638548.093666] [ndv4:68754:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650638548.093669] [ndv4:68754:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650638548.093673] [ndv4:68754:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650638548.093675] [ndv4:68754:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650638548.093678] [ndv4:68754:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650638548.093680] [ndv4:68754:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650638548.093682] [ndv4:68754:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650638548.093691] [ndv4:68754:0] module.c:253 UCX DEBUG loading modules for uct
[1650638548.095197] [ndv4:69236:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638548.095241] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638548.096194] [ndv4:68766:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib6 vendor_id: 0x15b3 device_id: 4124
[1650638548.096907] [ndv4:68766:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib6: disable ODP because it's not supported for DevX QP
[1650638548.099983] [ndv4:68754:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650638548.100380] [ndv4:68754:0] ucp_context.c:1117 UCX DEBUG closing md posix because it has no selected transport resources
[1650638548.100402] [ndv4:68754:0] ucp_context.c:1117 UCX DEBUG closing md self because it has no selected transport resources
[1650638548.100756] [ndv4:68754:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650638548.100769] [ndv4:68754:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650638548.100780] [ndv4:68754:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650638548.100790] [ndv4:68754:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650638548.100802] [ndv4:68754:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650638548.100820] [ndv4:68754:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650638548.100832] [ndv4:68754:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650638548.100886] [ndv4:68754:0] ucp_context.c:1117 UCX DEBUG closing md tcp because it has no selected transport resources
[1650638548.101317] [ndv4:68754:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650638548.108171] [ndv4:68754:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650638548.108560] [ndv4:68754:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650638548.115971] [ndv4:68754:0] async.c:228 UCX DEBUG added async handler 0x212e660 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650638548.116073] [ndv4:68754:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650638548.116508] [ndv4:68754:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650638548.116905] [ndv4:68754:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.116916] [ndv4:68754:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.116942] [ndv4:68754:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.117003] [ndv4:68754:0] module.c:253 UCX DEBUG loading modules for ucm
[1650638548.117025] [ndv4:68754:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650638548.117992] [ndv4:69236:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638548.118004] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638548.120749] [ndv4:68766:0] async.c:228 UCX DEBUG added async handler 0x1bfff60 [id=69 ref 1] uct_ib_async_event_handler() to hash
[1650638548.120776] [ndv4:68766:0] async.c:506 UCX DEBUG listening to async event fd 69 events 0x1 mode thread_spinlock
[1650638548.120782] [ndv4:68766:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib6' (InfiniBand channel adapter) with 1 ports
[1650638548.121110] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638548.121119] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638548.121134] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.121197] [ndv4:68766:0] ib_md.c:1319 UCX DEBUG mlx5_ib6: using registration cache
[1650638548.121003] [ndv4:69236:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638548.121011] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638548.121648] [ndv4:68755:0] topo.c:99 UCX DEBUG bus id 0x103000000 doesn't exist. sys_dev = 2
[1650638548.121660] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.123616] [ndv4:68754:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.123633] [ndv4:68754:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.123900] [ndv4:68754:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.124030] [ndv4:68754:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.124036] [ndv4:68754:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.132599] [ndv4:68766:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.132615] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.132916] [ndv4:68766:0] ib_md.c:1604 UCX DEBUG mlx5_ib6: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.133034] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638548.133041] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638548.133632] [ndv4:68766:0] topo.c:99 UCX DEBUG bus id 0x107000000 doesn't exist. sys_dev = 6
[1650638548.133637] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650638548.141382] [ndv4:69236:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638548.141393] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638548.141717] [ndv4:69236:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.142412] [ndv4:68755:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.142420] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.142794] [ndv4:68754:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650638548.142815] [ndv4:68754:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.147109] [ndv4:68754:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.147116] [ndv4:68754:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.148148] [ndv4:68754:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.148157] [ndv4:68754:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.152478] [ndv4:70439:0] debug.c:1198 UCX DEBUG using signal stack 0x2ab4a4343000 size 141824
[1650638548.152558] [ndv4:70439:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650638548.152579] [ndv4:70439:0] 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 0x2ab4a4197000
[1650638548.152603] [ndv4:70439:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650638548.152613] [ndv4:70439:0] 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
[1650638548.152620] [ndv4:70439:0] module.c:253 UCX DEBUG loading modules for ucs
[1650638548.154229] [ndv4:68766:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650638548.154239] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650638548.154493] [ndv4:68766:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650638548.154497] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650638548.154891] [ndv4:68766:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650638548.154894] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650638548.155170] [ndv4:68766:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650638548.155173] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650638548.155696] [ndv4:68766:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.155512] [ndv4:70439:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650638548.155532] [ndv4:70439:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650638548.155570] [ndv4:70439:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650638548.155574] [ndv4:70439:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650638548.155581] [ndv4:70439:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650638548.155588] [ndv4:70439:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650638548.155591] [ndv4:70439:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650638548.155596] [ndv4:70439:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650638548.155599] [ndv4:70439:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650638548.155601] [ndv4:70439:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650638548.155604] [ndv4:70439:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650638548.155606] [ndv4:70439:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650638548.155614] [ndv4:70439:0] module.c:253 UCX DEBUG loading modules for uct
[1650638548.163347] [ndv4:70439:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650638548.164075] [ndv4:70439:0] ucp_context.c:1117 UCX DEBUG closing md posix because it has no selected transport resources
[1650638548.164106] [ndv4:70439:0] ucp_context.c:1117 UCX DEBUG closing md self because it has no selected transport resources
[1650638548.164245] [ndv4:70439:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650638548.164257] [ndv4:70439:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650638548.164268] [ndv4:70439:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650638548.164278] [ndv4:70439:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650638548.164289] [ndv4:70439:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650638548.164300] [ndv4:70439:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650638548.164311] [ndv4:70439:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650638548.164346] [ndv4:70439:0] ucp_context.c:1117 UCX DEBUG closing md tcp because it has no selected transport resources
[1650638548.164730] [ndv4:70439:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650638548.164388] [ndv4:68766:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib7 vendor_id: 0x15b3 device_id: 4124
[1650638548.164764] [ndv4:68766:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib7: disable ODP because it's not supported for DevX QP
[1650638548.164700] [ndv4:70055:0] debug.c:1198 UCX DEBUG using signal stack 0x2b312a516000 size 141824
[1650638548.164779] [ndv4:70055:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650638548.164799] [ndv4:70055:0] 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 0x2b312a36a000
[1650638548.164821] [ndv4:70055:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650638548.164830] [ndv4:70055:0] 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
[1650638548.164836] [ndv4:70055:0] module.c:253 UCX DEBUG loading modules for ucs
[1650638548.165076] [ndv4:68766:0] async.c:228 UCX DEBUG added async handler 0x1c0ff90 [id=71 ref 1] uct_ib_async_event_handler() to hash
[1650638548.165113] [ndv4:68766:0] async.c:506 UCX DEBUG listening to async event fd 71 events 0x1 mode thread_spinlock
[1650638548.165116] [ndv4:68766:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib7' (InfiniBand channel adapter) with 1 ports
[1650638548.165347] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638548.165354] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638548.165369] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.165428] [ndv4:68766:0] ib_md.c:1319 UCX DEBUG mlx5_ib7: using registration cache
[1650638548.167404] [ndv4:70055:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650638548.167422] [ndv4:70055:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650638548.167461] [ndv4:70055:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650638548.167465] [ndv4:70055:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650638548.167472] [ndv4:70055:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650638548.167479] [ndv4:70055:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650638548.167481] [ndv4:70055:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650638548.167485] [ndv4:70055:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650638548.167488] [ndv4:70055:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650638548.167491] [ndv4:70055:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650638548.167493] [ndv4:70055:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650638548.167495] [ndv4:70055:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650638548.167503] [ndv4:70055:0] module.c:253 UCX DEBUG loading modules for uct
[1650638548.170664] [ndv4:68766:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.170677] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.171501] [ndv4:68754:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.171511] [ndv4:68754:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.170956] [ndv4:68766:0] ib_md.c:1604 UCX DEBUG mlx5_ib7: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.170980] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638548.170985] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638548.173146] [ndv4:70055:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650638548.173508] [ndv4:70055:0] ucp_context.c:1117 UCX DEBUG closing md posix because it has no selected transport resources
[1650638548.173529] [ndv4:70055:0] ucp_context.c:1117 UCX DEBUG closing md self because it has no selected transport resources
[1650638548.173885] [ndv4:70055:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650638548.173898] [ndv4:70055:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650638548.173909] [ndv4:70055:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650638548.173919] [ndv4:70055:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650638548.173930] [ndv4:70055:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650638548.173940] [ndv4:70055:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650638548.173951] [ndv4:70055:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650638548.173990] [ndv4:70055:0] ucp_context.c:1117 UCX DEBUG closing md tcp because it has no selected transport resources
[1650638548.172685] [ndv4:68766:0] topo.c:99 UCX DEBUG bus id 0x108000000 doesn't exist. sys_dev = 7
[1650638548.172691] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650638548.173595] [ndv4:68766:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650638548.173601] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650638548.174417] [ndv4:70055:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650638548.177183] [ndv4:68755:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.177198] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.177855] [ndv4:68755:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.177860] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.180630] [ndv4:70055:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650638548.180985] [ndv4:70055:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650638548.184785] [ndv4:68754:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.184794] [ndv4:68754:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.185055] [ndv4:68754:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.185019] [ndv4:70439:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650638548.185436] [ndv4:70439:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650638548.187920] [ndv4:68766:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650638548.187930] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650638548.188377] [ndv4:70055:0] async.c:228 UCX DEBUG added async handler 0x2358640 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650638548.188491] [ndv4:70055:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650638548.188510] [ndv4:70055:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650638548.189973] [ndv4:68766:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650638548.189979] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650638548.189284] [ndv4:70055:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.189294] [ndv4:70055:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.189318] [ndv4:70055:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.189379] [ndv4:70055:0] module.c:253 UCX DEBUG loading modules for ucm
[1650638548.189399] [ndv4:70055:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650638548.189653] [ndv4:68755:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.189664] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.189981] [ndv4:68755:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.190885] [ndv4:68766:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650638548.190891] [ndv4:68766:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650638548.191007] [ndv4:68766:0] ucp_context.c:1556 UCX DEBUG created ucp context 0x1bfb130 0x1bfb130 [10 mds 42 tls] features 0x1 tl bitmap 0x3ffffffffff 0x0
[1650638548.195105] [ndv4:69236:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib4 vendor_id: 0x15b3 device_id: 4124
[1650638548.195747] [ndv4:69236:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib4: disable ODP because it's not supported for DevX QP
[1650638548.196446] [ndv4:70055:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.196463] [ndv4:70055:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.196750] [ndv4:70055:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.196867] [ndv4:70055:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.196874] [ndv4:70055:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.196603] [ndv4:69236:0] async.c:228 UCX DEBUG added async handler 0x1b89610 [id=65 ref 1] uct_ib_async_event_handler() to hash
[1650638548.196653] [ndv4:69236:0] async.c:506 UCX DEBUG listening to async event fd 65 events 0x1 mode thread_spinlock
[1650638548.196658] [ndv4:69236:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib4' (InfiniBand channel adapter) with 1 ports
[1650638548.197135] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638548.197146] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638548.197165] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.197293] [ndv4:69236:0] ib_md.c:1319 UCX DEBUG mlx5_ib4: using registration cache
[1650638548.199998] [ndv4:70055:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650638548.200008] [ndv4:70055:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.203041] [ndv4:68754:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650638548.203547] [ndv4:68754:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650638548.205602] [ndv4:68754:0] async.c:228 UCX DEBUG added async handler 0x21020b0 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650638548.205643] [ndv4:68754:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650638548.205647] [ndv4:68754:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650638548.206072] [ndv4:68754:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638548.206083] [ndv4:68754:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638548.206099] [ndv4:68754:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.206372] [ndv4:68754:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650638548.206900] [ndv4:70439:0] async.c:228 UCX DEBUG added async handler 0x1c41340 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650638548.207012] [ndv4:70439:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650638548.207038] [ndv4:70439:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650638548.208200] [ndv4:68755:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib3 vendor_id: 0x15b3 device_id: 4124
[1650638548.208559] [ndv4:68755:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib3: disable ODP because it's not supported for DevX QP
[1650638548.210587] [ndv4:68755:0] async.c:228 UCX DEBUG added async handler 0xcaa590 [id=63 ref 1] uct_ib_async_event_handler() to hash
[1650638548.210627] [ndv4:68755:0] async.c:506 UCX DEBUG listening to async event fd 63 events 0x1 mode thread_spinlock
[1650638548.210633] [ndv4:68755:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib3' (InfiniBand channel adapter) with 1 ports
[1650638548.211349] [ndv4:68755:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638548.211360] [ndv4:68755:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638548.211375] [ndv4:68755:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.211462] [ndv4:68755:0] ib_md.c:1319 UCX DEBUG mlx5_ib3: using registration cache
[1650638548.207387] [ndv4:70439:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.207397] [ndv4:70439:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.207425] [ndv4:70439:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.207495] [ndv4:70439:0] module.c:253 UCX DEBUG loading modules for ucm
[1650638548.207517] [ndv4:70439:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650638548.213144] [ndv4:70439:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.213163] [ndv4:70439:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.213501] [ndv4:70439:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.213703] [ndv4:70439:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.213710] [ndv4:70439:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.213171] [ndv4:69236:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.213241] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.213607] [ndv4:69236:0] ib_md.c:1604 UCX DEBUG mlx5_ib4: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.213836] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638548.213844] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638548.209187] [ndv4:69075:0] debug.c:1198 UCX DEBUG using signal stack 0x2b4f90d00000 size 141824
[1650638548.209297] [ndv4:69075:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650638548.209318] [ndv4:69075:0] 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 0x2b4f90b54000
[1650638548.209342] [ndv4:69075:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650638548.209351] [ndv4:69075:0] 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
[1650638548.209357] [ndv4:69075:0] module.c:253 UCX DEBUG loading modules for ucs
[1650638548.211906] [ndv4:69075:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650638548.211926] [ndv4:69075:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650638548.211958] [ndv4:69075:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650638548.211961] [ndv4:69075:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650638548.211968] [ndv4:69075:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650638548.211975] [ndv4:69075:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650638548.211978] [ndv4:69075:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650638548.211982] [ndv4:69075:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650638548.211985] [ndv4:69075:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650638548.211987] [ndv4:69075:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650638548.211989] [ndv4:69075:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650638548.211991] [ndv4:69075:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650638548.212000] [ndv4:69075:0] module.c:253 UCX DEBUG loading modules for uct
[1650638548.215227] [ndv4:70439:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650638548.215236] [ndv4:70439:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.215290] [ndv4:69236:0] topo.c:99 UCX DEBUG bus id 0x105000000 doesn't exist. sys_dev = 4
[1650638548.215297] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638548.214600] [ndv4:68755:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.214613] [ndv4:68755:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.215602] [ndv4:68755:0] ib_md.c:1604 UCX DEBUG mlx5_ib3: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.215638] [ndv4:68755:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638548.215645] [ndv4:68755:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638548.215809] [ndv4:68755:0] topo.c:99 UCX DEBUG bus id 0x104000000 doesn't exist. sys_dev = 3
[1650638548.215814] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638548.215947] [ndv4:68755:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638548.215951] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638548.216094] [ndv4:68755:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638548.216097] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638548.215517] [ndv4:69236:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650638548.215525] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638548.215883] [ndv4:69236:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650638548.215890] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638548.216070] [ndv4:69236:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650638548.216074] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638548.216275] [ndv4:69236:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650638548.216279] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638548.216589] [ndv4:69236:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.216302] [ndv4:68755:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638548.216308] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638548.216371] [ndv4:68755:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638548.216374] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638548.216703] [ndv4:68755:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.218040] [ndv4:69075:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650638548.218410] [ndv4:69075:0] ucp_context.c:1117 UCX DEBUG closing md posix because it has no selected transport resources
[1650638548.218431] [ndv4:69075:0] ucp_context.c:1117 UCX DEBUG closing md self because it has no selected transport resources
[1650638548.218580] [ndv4:69075:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650638548.218593] [ndv4:69075:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650638548.218604] [ndv4:69075:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650638548.218615] [ndv4:69075:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650638548.218625] [ndv4:69075:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650638548.218636] [ndv4:69075:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650638548.218646] [ndv4:69075:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650638548.218680] [ndv4:69075:0] ucp_context.c:1117 UCX DEBUG closing md tcp because it has no selected transport resources
[1650638548.219036] [ndv4:69075:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650638548.223150] [ndv4:68755:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib4 vendor_id: 0x15b3 device_id: 4124
[1650638548.223543] [ndv4:68755:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib4: disable ODP because it's not supported for DevX QP
[1650638548.223802] [ndv4:68755:0] async.c:228 UCX DEBUG added async handler 0x1395cf0 [id=65 ref 1] uct_ib_async_event_handler() to hash
[1650638548.223829] [ndv4:68755:0] async.c:506 UCX DEBUG listening to async event fd 65 events 0x1 mode thread_spinlock
[1650638548.223834] [ndv4:68755:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib4' (InfiniBand channel adapter) with 1 ports
[1650638548.223985] [ndv4:68755:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638548.223995] [ndv4:68755:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638548.224002] [ndv4:68755:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.224053] [ndv4:68755:0] ib_md.c:1319 UCX DEBUG mlx5_ib4: using registration cache
[1650638548.227294] [ndv4:69236:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib5 vendor_id: 0x15b3 device_id: 4124
[1650638548.227700] [ndv4:69236:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib5: disable ODP because it's not supported for DevX QP
[1650638548.227996] [ndv4:69236:0] async.c:228 UCX DEBUG added async handler 0x14a3420 [id=67 ref 1] uct_ib_async_event_handler() to hash
[1650638548.228032] [ndv4:69236:0] async.c:506 UCX DEBUG listening to async event fd 67 events 0x1 mode thread_spinlock
[1650638548.228035] [ndv4:69236:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib5' (InfiniBand channel adapter) with 1 ports
[1650638548.228292] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638548.228303] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638548.228318] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.228416] [ndv4:69236:0] ib_md.c:1319 UCX DEBUG mlx5_ib5: using registration cache
[1650638548.238127] [ndv4:68764:0] debug.c:1198 UCX DEBUG using signal stack 0x2ab717dec000 size 141824
[1650638548.238276] [ndv4:68764:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650638548.238299] [ndv4:68764:0] 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 0x2ab717c40000
[1650638548.238320] [ndv4:68764:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650638548.238330] [ndv4:68764:0] 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
[1650638548.238336] [ndv4:68764:0] module.c:253 UCX DEBUG loading modules for ucs
[1650638548.241097] [ndv4:68764:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650638548.241116] [ndv4:68764:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650638548.241150] [ndv4:68764:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650638548.241154] [ndv4:68764:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650638548.241162] [ndv4:68764:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650638548.241170] [ndv4:68764:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650638548.241173] [ndv4:68764:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650638548.241177] [ndv4:68764:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650638548.241180] [ndv4:68764:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650638548.241182] [ndv4:68764:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650638548.241185] [ndv4:68764:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650638548.241187] [ndv4:68764:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650638548.241196] [ndv4:68764:0] module.c:253 UCX DEBUG loading modules for uct
[1650638548.247758] [ndv4:68764:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650638548.248102] [ndv4:68764:0] ucp_context.c:1117 UCX DEBUG closing md posix because it has no selected transport resources
[1650638548.248124] [ndv4:68764:0] ucp_context.c:1117 UCX DEBUG closing md self because it has no selected transport resources
[1650638548.248280] [ndv4:68764:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650638548.248293] [ndv4:68764:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650638548.248304] [ndv4:68764:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650638548.248316] [ndv4:68764:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650638548.248327] [ndv4:68764:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650638548.248339] [ndv4:68764:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650638548.248351] [ndv4:68764:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650638548.248389] [ndv4:68764:0] ucp_context.c:1117 UCX DEBUG closing md tcp because it has no selected transport resources
[1650638548.248776] [ndv4:68764:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650638548.238960] [ndv4:70462:0] debug.c:1198 UCX DEBUG using signal stack 0x2b87e10ae000 size 141824
[1650638548.239040] [ndv4:70462:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650638548.239061] [ndv4:70462:0] 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 0x2b87e0f1c000
[1650638548.239079] [ndv4:70462:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650638548.239088] [ndv4:70462:0] 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
[1650638548.239094] [ndv4:70462:0] module.c:253 UCX DEBUG loading modules for ucs
[1650638548.241846] [ndv4:70462:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650638548.241864] [ndv4:70462:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650638548.241899] [ndv4:70462:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650638548.241902] [ndv4:70462:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650638548.241909] [ndv4:70462:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650638548.241916] [ndv4:70462:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650638548.241918] [ndv4:70462:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650638548.241923] [ndv4:70462:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650638548.241925] [ndv4:70462:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650638548.241927] [ndv4:70462:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650638548.241930] [ndv4:70462:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650638548.241932] [ndv4:70462:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650638548.241940] [ndv4:70462:0] module.c:253 UCX DEBUG loading modules for uct
[1650638548.247768] [ndv4:70462:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650638548.248120] [ndv4:70462:0] ucp_context.c:1117 UCX DEBUG closing md posix because it has no selected transport resources
[1650638548.248141] [ndv4:70462:0] ucp_context.c:1117 UCX DEBUG closing md self because it has no selected transport resources
[1650638548.248293] [ndv4:70462:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650638548.248306] [ndv4:70462:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650638548.248317] [ndv4:70462:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650638548.248328] [ndv4:70462:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650638548.248340] [ndv4:70462:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650638548.248352] [ndv4:70462:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650638548.248365] [ndv4:70462:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650638548.248400] [ndv4:70462:0] ucp_context.c:1117 UCX DEBUG closing md tcp because it has no selected transport resources
[1650638548.248798] [ndv4:70462:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650638548.255416] [ndv4:68764:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650638548.255405] [ndv4:70462:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650638548.255827] [ndv4:68764:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650638548.255816] [ndv4:70462:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650638548.263713] [ndv4:68764:0] async.c:228 UCX DEBUG added async handler 0x25402c0 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650638548.263808] [ndv4:68764:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650638548.263621] [ndv4:70462:0] async.c:228 UCX DEBUG added async handler 0xa96680 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650638548.263720] [ndv4:70462:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650638548.264139] [ndv4:68764:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650638548.264120] [ndv4:70462:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650638548.264351] [ndv4:68764:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.264361] [ndv4:68764:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.264384] [ndv4:68764:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.264447] [ndv4:68764:0] module.c:253 UCX DEBUG loading modules for ucm
[1650638548.264468] [ndv4:68764:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650638548.264343] [ndv4:70462:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.264352] [ndv4:70462:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.264377] [ndv4:70462:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.264442] [ndv4:70462:0] module.c:253 UCX DEBUG loading modules for ucm
[1650638548.264464] [ndv4:70462:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650638548.269595] [ndv4:69236:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.269624] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.269967] [ndv4:69236:0] ib_md.c:1604 UCX DEBUG mlx5_ib5: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.270446] [ndv4:70462:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.271175] [ndv4:70462:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.271251] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638548.271264] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638548.271673] [ndv4:69236:0] topo.c:99 UCX DEBUG bus id 0x106000000 doesn't exist. sys_dev = 5
[1650638548.271682] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638548.271787] [ndv4:69236:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650638548.271792] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638548.271472] [ndv4:70462:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.271491] [ndv4:70462:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.271497] [ndv4:70462:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.271522] [ndv4:68764:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.271546] [ndv4:68764:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.271817] [ndv4:68764:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.272069] [ndv4:68764:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.272078] [ndv4:68764:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.272096] [ndv4:69236:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650638548.272104] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638548.272278] [ndv4:69236:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650638548.272283] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638548.272354] [ndv4:69236:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650638548.272358] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638548.274448] [ndv4:68764:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650638548.274461] [ndv4:68764:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.274436] [ndv4:70462:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650638548.274451] [ndv4:70462:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.274712] [ndv4:70462:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.274716] [ndv4:70462:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.274826] [ndv4:70462:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.274830] [ndv4:70462:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.274888] [ndv4:70462:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.274891] [ndv4:70462:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.274956] [ndv4:70462:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.274960] [ndv4:70462:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.275198] [ndv4:70462:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.274784] [ndv4:69236:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.278851] [ndv4:70038:0] debug.c:1198 UCX DEBUG using signal stack 0x2aec0006d000 size 141824
[1650638548.278929] [ndv4:70038:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650638548.278948] [ndv4:70038:0] 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 0x2aebffec1000
[1650638548.278968] [ndv4:70038:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650638548.278978] [ndv4:70038:0] 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
[1650638548.278984] [ndv4:70038:0] module.c:253 UCX DEBUG loading modules for ucs
[1650638548.281574] [ndv4:70038:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650638548.281593] [ndv4:70038:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650638548.281626] [ndv4:70038:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650638548.281629] [ndv4:70038:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650638548.281636] [ndv4:70038:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650638548.281642] [ndv4:70038:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650638548.281645] [ndv4:70038:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650638548.281649] [ndv4:70038:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650638548.281651] [ndv4:70038:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650638548.281654] [ndv4:70038:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650638548.281656] [ndv4:70038:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650638548.281658] [ndv4:70038:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650638548.281665] [ndv4:70038:0] module.c:253 UCX DEBUG loading modules for uct
[1650638548.282504] [ndv4:69236:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib6 vendor_id: 0x15b3 device_id: 4124
[1650638548.282492] [ndv4:70462:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650638548.282878] [ndv4:70462:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650638548.282910] [ndv4:69236:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib6: disable ODP because it's not supported for DevX QP
[1650638548.283202] [ndv4:69236:0] async.c:228 UCX DEBUG added async handler 0x1491f60 [id=69 ref 1] uct_ib_async_event_handler() to hash
[1650638548.283281] [ndv4:69236:0] async.c:506 UCX DEBUG listening to async event fd 69 events 0x1 mode thread_spinlock
[1650638548.283285] [ndv4:69236:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib6' (InfiniBand channel adapter) with 1 ports
[1650638548.283135] [ndv4:70462:0] async.c:228 UCX DEBUG added async handler 0xa6a080 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650638548.283161] [ndv4:70462:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650638548.283164] [ndv4:70462:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650638548.283391] [ndv4:70462:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638548.283399] [ndv4:70462:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638548.283409] [ndv4:70462:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.283464] [ndv4:70462:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650638548.287348] [ndv4:70038:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650638548.287680] [ndv4:70038:0] ucp_context.c:1117 UCX DEBUG closing md posix because it has no selected transport resources
[1650638548.287701] [ndv4:70038:0] ucp_context.c:1117 UCX DEBUG closing md self because it has no selected transport resources
[1650638548.287803] [ndv4:70038:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650638548.287814] [ndv4:70038:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650638548.287825] [ndv4:70038:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650638548.287835] [ndv4:70038:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650638548.287847] [ndv4:70038:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650638548.287857] [ndv4:70038:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650638548.287869] [ndv4:70038:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650638548.287903] [ndv4:70038:0] ucp_context.c:1117 UCX DEBUG closing md tcp because it has no selected transport resources
[1650638548.288319] [ndv4:70038:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650638548.283470] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638548.283482] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638548.283502] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.283588] [ndv4:69236:0] ib_md.c:1319 UCX DEBUG mlx5_ib6: using registration cache
[1650638548.287756] [ndv4:69236:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.287764] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.287982] [ndv4:69236:0] ib_md.c:1604 UCX DEBUG mlx5_ib6: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.287997] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638548.288002] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638548.288331] [ndv4:69236:0] topo.c:99 UCX DEBUG bus id 0x107000000 doesn't exist. sys_dev = 6
[1650638548.288340] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650638548.288404] [ndv4:69236:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650638548.288408] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650638548.288496] [ndv4:70462:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.288504] [ndv4:70462:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.288804] [ndv4:70462:0] ib_md.c:1604 UCX DEBUG mlx5_ib1: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.288819] [ndv4:70462:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638548.288824] [ndv4:70462:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638548.288915] [ndv4:70462:0] topo.c:99 UCX DEBUG bus id 0x102000000 doesn't exist. sys_dev = 1
[1650638548.288918] [ndv4:70462:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.288980] [ndv4:70462:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638548.288983] [ndv4:70462:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.289041] [ndv4:70462:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638548.289044] [ndv4:70462:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.289104] [ndv4:70462:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638548.289107] [ndv4:70462:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.289166] [ndv4:70462:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638548.289168] [ndv4:70462:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.289415] [ndv4:70462:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.288591] [ndv4:69236:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650638548.288599] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650638548.288663] [ndv4:69236:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650638548.288667] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650638548.288729] [ndv4:69236:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650638548.288732] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650638548.288987] [ndv4:69236:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.294796] [ndv4:70038:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650638548.295166] [ndv4:70038:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650638548.297200] [ndv4:69236:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib7 vendor_id: 0x15b3 device_id: 4124
[1650638548.297180] [ndv4:70462:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib2 vendor_id: 0x15b3 device_id: 4124
[1650638548.297607] [ndv4:70462:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib2: disable ODP because it's not supported for DevX QP
[1650638548.297629] [ndv4:69236:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib7: disable ODP because it's not supported for DevX QP
[1650638548.297877] [ndv4:70462:0] async.c:228 UCX DEBUG added async handler 0xa8ed50 [id=61 ref 1] uct_ib_async_event_handler() to hash
[1650638548.297900] [ndv4:70462:0] async.c:506 UCX DEBUG listening to async event fd 61 events 0x1 mode thread_spinlock
[1650638548.297904] [ndv4:70462:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib2' (InfiniBand channel adapter) with 1 ports
[1650638548.297901] [ndv4:69236:0] async.c:228 UCX DEBUG added async handler 0x14a1f90 [id=71 ref 1] uct_ib_async_event_handler() to hash
[1650638548.297929] [ndv4:69236:0] async.c:506 UCX DEBUG listening to async event fd 71 events 0x1 mode thread_spinlock
[1650638548.297933] [ndv4:69236:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib7' (InfiniBand channel adapter) with 1 ports
[1650638548.298099] [ndv4:70462:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638548.298108] [ndv4:70462:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638548.298114] [ndv4:70462:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.298154] [ndv4:70462:0] ib_md.c:1319 UCX DEBUG mlx5_ib2: using registration cache
[1650638548.306951] [ndv4:70462:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.306958] [ndv4:70462:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.307179] [ndv4:70462:0] ib_md.c:1604 UCX DEBUG mlx5_ib2: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.307202] [ndv4:70462:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638548.307222] [ndv4:70462:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638548.307530] [ndv4:70462:0] topo.c:99 UCX DEBUG bus id 0x103000000 doesn't exist. sys_dev = 2
[1650638548.307534] [ndv4:70462:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.307685] [ndv4:70462:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.307689] [ndv4:70462:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.307853] [ndv4:70462:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.307856] [ndv4:70462:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.308007] [ndv4:70462:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.308009] [ndv4:70462:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.308153] [ndv4:70462:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.308156] [ndv4:70462:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.298121] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638548.298132] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638548.298139] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.298190] [ndv4:69236:0] ib_md.c:1319 UCX DEBUG mlx5_ib7: using registration cache
[1650638548.305467] [ndv4:69236:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.305475] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.305720] [ndv4:69236:0] ib_md.c:1604 UCX DEBUG mlx5_ib7: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.305736] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638548.305741] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638548.306172] [ndv4:69236:0] topo.c:99 UCX DEBUG bus id 0x108000000 doesn't exist. sys_dev = 7
[1650638548.306179] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650638548.306354] [ndv4:69236:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650638548.306358] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650638548.306613] [ndv4:69236:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650638548.306619] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650638548.306805] [ndv4:69236:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650638548.306809] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650638548.306969] [ndv4:69236:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650638548.306973] [ndv4:69236:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650638548.307098] [ndv4:69236:0] ucp_context.c:1556 UCX DEBUG created ucp context 0x148d130 0x148d130 [10 mds 42 tls] features 0x1 tl bitmap 0x3ffffffffff 0x0
[1650638548.302544] [ndv4:70038:0] async.c:228 UCX DEBUG added async handler 0x14f52c0 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650638548.302633] [ndv4:70038:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650638548.302725] [ndv4:70038:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650638548.302912] [ndv4:70038:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.302919] [ndv4:70038:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.302944] [ndv4:70038:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.303003] [ndv4:70038:0] module.c:253 UCX DEBUG loading modules for ucm
[1650638548.303023] [ndv4:70038:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650638548.308447] [ndv4:70462:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.310293] [ndv4:70038:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.310306] [ndv4:70038:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.310539] [ndv4:70038:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.310555] [ndv4:70038:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.310560] [ndv4:70038:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.310721] [ndv4:70038:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650638548.310726] [ndv4:70038:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.310892] [ndv4:70038:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.310896] [ndv4:70038:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.311010] [ndv4:70038:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.311013] [ndv4:70038:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.311073] [ndv4:70038:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.311075] [ndv4:70038:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.311134] [ndv4:70038:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.311136] [ndv4:70038:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.311393] [ndv4:70038:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.311689] [ndv4:70055:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.311708] [ndv4:70055:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.312123] [ndv4:70055:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.312128] [ndv4:70055:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.312650] [ndv4:70055:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.312654] [ndv4:70055:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.312940] [ndv4:70055:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.312943] [ndv4:70055:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.313291] [ndv4:70055:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.318046] [ndv4:68754:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.318066] [ndv4:68754:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.318316] [ndv4:68754:0] ib_md.c:1604 UCX DEBUG mlx5_ib1: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.318504] [ndv4:68754:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638548.318511] [ndv4:68754:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638548.319104] [ndv4:69075:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650638548.319575] [ndv4:69075:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650638548.320136] [ndv4:68755:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.320158] [ndv4:68755:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.320724] [ndv4:68755:0] ib_md.c:1604 UCX DEBUG mlx5_ib4: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.320807] [ndv4:68755:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638548.320815] [ndv4:68755:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638548.321011] [ndv4:68754:0] topo.c:99 UCX DEBUG bus id 0x102000000 doesn't exist. sys_dev = 1
[1650638548.321019] [ndv4:68754:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.322701] [ndv4:68764:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.322709] [ndv4:68764:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.325425] [ndv4:68764:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.325431] [ndv4:68764:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.326471] [ndv4:70439:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.326482] [ndv4:70439:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.327331] [ndv4:70055:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650638548.327679] [ndv4:70055:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650638548.328020] [ndv4:70055:0] async.c:228 UCX DEBUG added async handler 0x232bfd0 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650638548.328046] [ndv4:70055:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650638548.328050] [ndv4:70055:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650638548.328336] [ndv4:70055:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638548.328346] [ndv4:70055:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638548.328356] [ndv4:70055:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.328410] [ndv4:70055:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650638548.330267] [ndv4:70462:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib3 vendor_id: 0x15b3 device_id: 4124
[1650638548.331198] [ndv4:70462:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib3: disable ODP because it's not supported for DevX QP
[1650638548.332560] [ndv4:70055:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.332574] [ndv4:70055:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.332851] [ndv4:70055:0] ib_md.c:1604 UCX DEBUG mlx5_ib1: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.332932] [ndv4:70055:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638548.332937] [ndv4:70055:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638548.333576] [ndv4:68755:0] topo.c:99 UCX DEBUG bus id 0x105000000 doesn't exist. sys_dev = 4
[1650638548.333585] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638548.335555] [ndv4:68755:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650638548.335562] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638548.335372] [ndv4:70055:0] topo.c:99 UCX DEBUG bus id 0x102000000 doesn't exist. sys_dev = 1
[1650638548.335379] [ndv4:70055:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.336159] [ndv4:70055:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638548.336165] [ndv4:70055:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.336303] [ndv4:70055:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638548.336310] [ndv4:70055:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.337027] [ndv4:70055:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638548.337031] [ndv4:70055:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.336825] [ndv4:68755:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650638548.336832] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638548.338317] [ndv4:69075:0] async.c:228 UCX DEBUG added async handler 0xad95f0 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650638548.338419] [ndv4:69075:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650638548.338516] [ndv4:69075:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650638548.338670] [ndv4:70038:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650638548.339354] [ndv4:70439:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.339362] [ndv4:70439:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.339410] [ndv4:68755:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650638548.339416] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638548.339379] [ndv4:70038:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650638548.339274] [ndv4:69075:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.339284] [ndv4:69075:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.339316] [ndv4:69075:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.339387] [ndv4:69075:0] module.c:253 UCX DEBUG loading modules for ucm
[1650638548.339409] [ndv4:69075:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650638548.340440] [ndv4:70055:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638548.340446] [ndv4:70055:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.340680] [ndv4:70055:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.340589] [ndv4:70462:0] async.c:228 UCX DEBUG added async handler 0xa8ec40 [id=63 ref 1] uct_ib_async_event_handler() to hash
[1650638548.340616] [ndv4:70462:0] async.c:506 UCX DEBUG listening to async event fd 63 events 0x1 mode thread_spinlock
[1650638548.340620] [ndv4:70462:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib3' (InfiniBand channel adapter) with 1 ports
[1650638548.341092] [ndv4:68755:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650638548.341098] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638548.341505] [ndv4:68755:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.341044] [ndv4:70462:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638548.341054] [ndv4:70462:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638548.341059] [ndv4:70462:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.341107] [ndv4:70462:0] ib_md.c:1319 UCX DEBUG mlx5_ib3: using registration cache
[1650638548.341698] [ndv4:68754:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638548.341710] [ndv4:68754:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.341880] [ndv4:68754:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638548.341884] [ndv4:68754:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.344029] [ndv4:68764:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.344040] [ndv4:68764:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.345805] [ndv4:68764:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.345811] [ndv4:68764:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.346046] [ndv4:68764:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.347932] [ndv4:68766:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 8447 bytes with hugetlb
[1650638548.348021] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool mm_recv_desc: align 64, maxelems 4294967295, elemsize 8368
[1650638548.348047] [ndv4:68766:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 4292720 bytes with hugetlb
[1650638548.348058] [ndv4:68766:0] mpool.c:205 UCX DEBUG mpool mm_recv_desc: allocated chunk 0x2b8e80c25018 of 4296680 bytes with 512 elements
[1650638548.348686] [ndv4:68766:0] mm_iface.c:600 UCX DEBUG created mm iface 0x23771f0 FIFO id 0xe000c va 0x2b8e7beb3000 size 12288 (128 x 64 elems)
[1650638548.348739] [ndv4:68766:0] ucp_worker.c:1159 UCX DEBUG created interface[0]=0x23771f0 using sysv/memory on worker 0x2c514d0
[1650638548.348896] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.348904] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.349138] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.349143] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.350078] [ndv4:70038:0] async.c:228 UCX DEBUG added async handler 0x14f5440 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650638548.350099] [ndv4:70038:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650638548.350103] [ndv4:70038:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650638548.350862] [ndv4:70462:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.350871] [ndv4:70462:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.351149] [ndv4:70462:0] ib_md.c:1604 UCX DEBUG mlx5_ib3: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.351505] [ndv4:70462:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638548.351512] [ndv4:70462:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638548.350354] [ndv4:70038:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638548.350363] [ndv4:70038:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638548.350368] [ndv4:70038:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.350411] [ndv4:70038:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650638548.350176] [ndv4:68766:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib0:1
[1650638548.351842] [ndv4:68766:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650638548.351884] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638548.351888] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638548.351942] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638548.350991] [ndv4:69075:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.351007] [ndv4:69075:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.351307] [ndv4:69075:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.351691] [ndv4:69075:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.351698] [ndv4:69075:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.352574] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638548.352584] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650638548.353934] [ndv4:68766:0] ib_iface.c:994 UCX DEBUG iface=0x2382400: created RC QP 0x30f9 on mlx5_ib0:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650638548.354760] [ndv4:68754:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638548.354773] [ndv4:68754:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.355524] [ndv4:68764:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650638548.355911] [ndv4:68766:0] ucp_worker.c:1159 UCX DEBUG created interface[1]=0x2382400 using rc_verbs/mlx5_ib0:1 on worker 0x2c514d0
[1650638548.356191] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.356199] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.355888] [ndv4:68764:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650638548.356534] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.356541] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.357318] [ndv4:68766:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib0:1
[1650638548.357356] [ndv4:68766:0] ib_device.c:1394 UCX DEBUG max IB CQE size is 128
[1650638548.357002] [ndv4:68754:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638548.357009] [ndv4:68754:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.357434] [ndv4:68754:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.358377] [ndv4:68766:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638548.358386] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638548.358389] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638548.358441] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638548.358020] [ndv4:70055:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib2 vendor_id: 0x15b3 device_id: 4124
[1650638548.358513] [ndv4:70055:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib2: disable ODP because it's not supported for DevX QP
[1650638548.358301] [ndv4:68764:0] async.c:228 UCX DEBUG added async handler 0x2540440 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650638548.358337] [ndv4:68764:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650638548.358340] [ndv4:68764:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650638548.358628] [ndv4:68764:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638548.358638] [ndv4:68764:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638548.358646] [ndv4:68764:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.358691] [ndv4:68764:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650638548.358990] [ndv4:68766:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x2f78010 of 8176 bytes with 127 elements
[1650638548.359314] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638548.359343] [ndv4:68766:0] 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> }
[1650638548.359393] [ndv4:68766:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650638548.359399] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638548.360323] [ndv4:68755:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib5 vendor_id: 0x15b3 device_id: 4124
[1650638548.360428] [ndv4:70055:0] async.c:228 UCX DEBUG added async handler 0x2350d10 [id=61 ref 1] uct_ib_async_event_handler() to hash
[1650638548.360458] [ndv4:70055:0] async.c:506 UCX DEBUG listening to async event fd 61 events 0x1 mode thread_spinlock
[1650638548.360467] [ndv4:70055:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib2' (InfiniBand channel adapter) with 1 ports
[1650638548.361818] [ndv4:70439:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.361843] [ndv4:70439:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.361914] [ndv4:68755:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib5: disable ODP because it's not supported for DevX QP
[1650638548.362845] [ndv4:70462:0] topo.c:99 UCX DEBUG bus id 0x104000000 doesn't exist. sys_dev = 3
[1650638548.362855] [ndv4:70462:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638548.363610] [ndv4:70462:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638548.363616] [ndv4:70462:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638548.363711] [ndv4:68755:0] async.c:228 UCX DEBUG added async handler 0xc9cfa0 [id=67 ref 1] uct_ib_async_event_handler() to hash
[1650638548.363745] [ndv4:68755:0] async.c:506 UCX DEBUG listening to async event fd 67 events 0x1 mode thread_spinlock
[1650638548.363751] [ndv4:68755:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib5' (InfiniBand channel adapter) with 1 ports
[1650638548.364437] [ndv4:68764:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.364460] [ndv4:68764:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.364721] [ndv4:68764:0] ib_md.c:1604 UCX DEBUG mlx5_ib1: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.364860] [ndv4:68764:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638548.364869] [ndv4:68764:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638548.363981] [ndv4:68755:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638548.363994] [ndv4:68755:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638548.364002] [ndv4:68755:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.364080] [ndv4:68755:0] ib_md.c:1319 UCX DEBUG mlx5_ib5: using registration cache
[1650638548.366276] [ndv4:70462:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638548.366283] [ndv4:70462:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638548.367291] [ndv4:68764:0] topo.c:99 UCX DEBUG bus id 0x102000000 doesn't exist. sys_dev = 1
[1650638548.367298] [ndv4:68764:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.367654] [ndv4:68764:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638548.367657] [ndv4:68764:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.368130] [ndv4:68764:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638548.368133] [ndv4:68764:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.368600] [ndv4:68755:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.368608] [ndv4:68755:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.368867] [ndv4:68755:0] ib_md.c:1604 UCX DEBUG mlx5_ib5: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.368889] [ndv4:68755:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638548.368895] [ndv4:68755:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638548.369651] [ndv4:70038:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.369665] [ndv4:70038:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.369845] [ndv4:70038:0] ib_md.c:1604 UCX DEBUG mlx5_ib1: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.369991] [ndv4:70038:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638548.369998] [ndv4:70038:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638548.371130] [ndv4:70038:0] topo.c:99 UCX DEBUG bus id 0x102000000 doesn't exist. sys_dev = 1
[1650638548.371135] [ndv4:70038:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.373738] [ndv4:70038:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638548.373744] [ndv4:70038:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.373816] [ndv4:70038:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638548.373819] [ndv4:70038:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.378620] [ndv4:70038:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638548.378625] [ndv4:70038:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.378694] [ndv4:70038:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638548.378697] [ndv4:70038:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.378938] [ndv4:70038:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.369788] [ndv4:70055:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638548.369801] [ndv4:70055:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638548.369808] [ndv4:70055:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.369853] [ndv4:70055:0] ib_md.c:1319 UCX DEBUG mlx5_ib2: using registration cache
[1650638548.379702] [ndv4:68764:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638548.379709] [ndv4:68764:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.377200] [ndv4:69236:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 8447 bytes with hugetlb
[1650638548.377329] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool mm_recv_desc: align 64, maxelems 4294967295, elemsize 8368
[1650638548.377357] [ndv4:69236:0] mm_sysv.c:94 UCX DEBUG mm failed to allocate 4292720 bytes with hugetlb
[1650638548.377368] [ndv4:69236:0] mpool.c:205 UCX DEBUG mpool mm_recv_desc: allocated chunk 0x2abae5c64018 of 4296680 bytes with 512 elements
[1650638548.378033] [ndv4:69236:0] mm_iface.c:600 UCX DEBUG created mm iface 0x1c091f0 FIFO id 0xe0014 va 0x2abadffd6000 size 12288 (128 x 64 elems)
[1650638548.378103] [ndv4:69236:0] ucp_worker.c:1159 UCX DEBUG created interface[0]=0x1c091f0 using sysv/memory on worker 0x24e34d0
[1650638548.378276] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.378284] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.378925] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.378930] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.379947] [ndv4:69236:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib0:1
[1650638548.381377] [ndv4:69236:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650638548.381424] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638548.381427] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638548.381482] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638548.381867] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638548.381961] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650638548.371608] [ndv4:70439:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.371616] [ndv4:70439:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.371896] [ndv4:70439:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.381079] [ndv4:70439:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650638548.381462] [ndv4:70439:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650638548.378782] [ndv4:68754:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib2 vendor_id: 0x15b3 device_id: 4124
[1650638548.379289] [ndv4:68754:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib2: disable ODP because it's not supported for DevX QP
[1650638548.371305] [ndv4:68766:0] async.c:228 UCX DEBUG added async handler 0x22e5820 [id=78 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638548.371342] [ndv4:68766:0] async.c:506 UCX DEBUG listening to async event fd 78 events 0x1 mode thread_spinlock
[1650638548.371413] [ndv4:68766:0] ucp_worker.c:1159 UCX DEBUG created interface[2]=0x238ff80 using rc_mlx5/mlx5_ib0:1 on worker 0x2c514d0
[1650638548.371699] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.371706] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.371788] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.371793] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.371968] [ndv4:68766:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib0:1
[1650638548.374545] [ndv4:68766:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638548.374555] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638548.374558] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638548.374572] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638548.375005] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650638548.375014] [ndv4:68766:0] 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> }
[1650638548.375048] [ndv4:68766:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650638548.375052] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638548.375060] [ndv4:68766:0] async.c:228 UCX DEBUG added async handler 0x22e7440 [id=80 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638548.375079] [ndv4:68766:0] async.c:506 UCX DEBUG listening to async event fd 80 events 0x1 mode thread_spinlock
[1650638548.375913] [ndv4:68766:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650638548.382606] [ndv4:69236:0] ib_iface.c:994 UCX DEBUG iface=0x1c14400: created RC QP 0x30ff on mlx5_ib0:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650638548.385252] [ndv4:70055:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.385266] [ndv4:70055:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.385494] [ndv4:70055:0] ib_md.c:1604 UCX DEBUG mlx5_ib2: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.382713] [ndv4:69075:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650638548.382723] [ndv4:69075:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.382822] [ndv4:70439:0] async.c:228 UCX DEBUG added async handler 0x1c4b3e0 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650638548.382850] [ndv4:70439:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650638548.382854] [ndv4:70439:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650638548.383288] [ndv4:70439:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638548.383297] [ndv4:70439:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638548.383306] [ndv4:70439:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.383364] [ndv4:70439:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650638548.385278] [ndv4:68766:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x2f7a050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0x3122
[1650638548.385869] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.385878] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.385903] [ndv4:68766:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b8e7beb8008 of 151544 bytes with 1052 elements
[1650638548.385697] [ndv4:70055:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638548.385706] [ndv4:70055:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638548.390663] [ndv4:70462:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638548.390753] [ndv4:70462:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638548.390992] [ndv4:70462:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638548.390995] [ndv4:70462:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638548.390645] [ndv4:68766:0] ib_md.c:812 UCX DEBUG registered memory 0x2b8e81200000..0x2b8e83800000 on mlx5_ib0 lkey 0x81400 rkey 0x81400 access 0xf flags 0x3e4
[1650638548.390763] [ndv4:68766:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b8e81200018 of 39845864 bytes with 4752 elements
[1650638548.390907] [ndv4:68766:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x2f7a050
[1650638548.391123] [ndv4:68766:0] ucp_worker.c:1159 UCX DEBUG created interface[3]=0x2f7a050 using dc_mlx5/mlx5_ib0:1 on worker 0x2c514d0
[1650638548.391392] [ndv4:70462:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.391439] [ndv4:68946:0] debug.c:1198 UCX DEBUG using signal stack 0x2ad22f670000 size 141824
[1650638548.391517] [ndv4:68946:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650638548.391537] [ndv4:68946:0] 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 0x2ad22f4c4000
[1650638548.391555] [ndv4:68946:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650638548.391564] [ndv4:68946:0] 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
[1650638548.391570] [ndv4:68946:0] module.c:253 UCX DEBUG loading modules for ucs
[1650638548.391550] [ndv4:70439:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.391570] [ndv4:70439:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.391827] [ndv4:70439:0] ib_md.c:1604 UCX DEBUG mlx5_ib1: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.391864] [ndv4:70439:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638548.391873] [ndv4:70439:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638548.393327] [ndv4:68754:0] async.c:228 UCX DEBUG added async handler 0x2126c20 [id=61 ref 1] uct_ib_async_event_handler() to hash
[1650638548.393372] [ndv4:68754:0] async.c:506 UCX DEBUG listening to async event fd 61 events 0x1 mode thread_spinlock
[1650638548.393378] [ndv4:68754:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib2' (InfiniBand channel adapter) with 1 ports
[1650638548.393734] [ndv4:68754:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638548.393743] [ndv4:68754:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638548.393755] [ndv4:68754:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.393848] [ndv4:68754:0] ib_md.c:1319 UCX DEBUG mlx5_ib2: using registration cache
[1650638548.393459] [ndv4:68755:0] topo.c:99 UCX DEBUG bus id 0x106000000 doesn't exist. sys_dev = 5
[1650638548.393471] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638548.394403] [ndv4:68946:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650638548.394422] [ndv4:68946:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650638548.394460] [ndv4:68946:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650638548.394463] [ndv4:68946:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650638548.394470] [ndv4:68946:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650638548.394477] [ndv4:68946:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650638548.394480] [ndv4:68946:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650638548.394485] [ndv4:68946:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650638548.394487] [ndv4:68946:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650638548.394490] [ndv4:68946:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650638548.394492] [ndv4:68946:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650638548.394495] [ndv4:68946:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650638548.394503] [ndv4:68946:0] module.c:253 UCX DEBUG loading modules for uct
[1650638548.397304] [ndv4:69236:0] ucp_worker.c:1159 UCX DEBUG created interface[1]=0x1c14400 using rc_verbs/mlx5_ib0:1 on worker 0x24e34d0
[1650638548.397855] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.397864] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.397240] [ndv4:69075:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.397253] [ndv4:69075:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.398377] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.398386] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.399363] [ndv4:69236:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib0:1
[1650638548.400177] [ndv4:70055:0] topo.c:99 UCX DEBUG bus id 0x103000000 doesn't exist. sys_dev = 2
[1650638548.400192] [ndv4:70055:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.400573] [ndv4:69236:0] ib_device.c:1394 UCX DEBUG max IB CQE size is 128
[1650638548.400708] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.400722] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.401038] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.401046] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.400793] [ndv4:68754:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.400806] [ndv4:68754:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.401129] [ndv4:68754:0] ib_md.c:1604 UCX DEBUG mlx5_ib2: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.401154] [ndv4:68754:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638548.401159] [ndv4:68754:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638548.402088] [ndv4:70038:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib2 vendor_id: 0x15b3 device_id: 4124
[1650638548.402655] [ndv4:70038:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib2: disable ODP because it's not supported for DevX QP
[1650638548.402250] [ndv4:68946:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650638548.402597] [ndv4:68946:0] ucp_context.c:1117 UCX DEBUG closing md posix because it has no selected transport resources
[1650638548.402620] [ndv4:68946:0] ucp_context.c:1117 UCX DEBUG closing md self because it has no selected transport resources
[1650638548.401774] [ndv4:69236:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638548.401790] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638548.401794] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638548.401849] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638548.402310] [ndv4:69236:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x280a010 of 8176 bytes with 127 elements
[1650638548.402554] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638548.402578] [ndv4:69236:0] 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> }
[1650638548.402626] [ndv4:69236:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650638548.402634] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638548.401381] [ndv4:70462:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib4 vendor_id: 0x15b3 device_id: 4124
[1650638548.401882] [ndv4:70462:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib4: disable ODP because it's not supported for DevX QP
[1650638548.402319] [ndv4:70462:0] async.c:228 UCX DEBUG added async handler 0x11786a0 [id=65 ref 1] uct_ib_async_event_handler() to hash
[1650638548.402344] [ndv4:70462:0] async.c:506 UCX DEBUG listening to async event fd 65 events 0x1 mode thread_spinlock
[1650638548.402348] [ndv4:70462:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib4' (InfiniBand channel adapter) with 1 ports
[1650638548.402600] [ndv4:70462:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638548.402608] [ndv4:70462:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638548.402617] [ndv4:70462:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.402671] [ndv4:70462:0] ib_md.c:1319 UCX DEBUG mlx5_ib4: using registration cache
[1650638548.402669] [ndv4:68766:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib0:1
[1650638548.402914] [ndv4:70439:0] topo.c:99 UCX DEBUG bus id 0x102000000 doesn't exist. sys_dev = 1
[1650638548.402922] [ndv4:70439:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.403329] [ndv4:70055:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.403336] [ndv4:70055:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.403851] [ndv4:70055:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.403855] [ndv4:70055:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.404122] [ndv4:68946:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650638548.404150] [ndv4:68946:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650638548.404161] [ndv4:68946:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650638548.404176] [ndv4:68946:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650638548.404000] [ndv4:68755:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650638548.404010] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638548.404495] [ndv4:68946:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650638548.404510] [ndv4:68946:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650638548.404522] [ndv4:68946:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650638548.404731] [ndv4:68946:0] ucp_context.c:1117 UCX DEBUG closing md tcp because it has no selected transport resources
[1650638548.404851] [ndv4:70038:0] async.c:228 UCX DEBUG added async handler 0x14f5d60 [id=61 ref 1] uct_ib_async_event_handler() to hash
[1650638548.404887] [ndv4:70038:0] async.c:506 UCX DEBUG listening to async event fd 61 events 0x1 mode thread_spinlock
[1650638548.405043] [ndv4:70439:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638548.405050] [ndv4:70439:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.405055] [ndv4:70038:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib2' (InfiniBand channel adapter) with 1 ports
[1650638548.405080] [ndv4:68766:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638548.406047] [ndv4:70038:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638548.406062] [ndv4:70038:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638548.406080] [ndv4:70038:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.406169] [ndv4:70038:0] ib_md.c:1319 UCX DEBUG mlx5_ib2: using registration cache
[1650638548.406689] [ndv4:70439:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638548.406698] [ndv4:70439:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.406958] [ndv4:70439:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638548.406961] [ndv4:70439:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.406066] [ndv4:68946:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650638548.405801] [ndv4:68766:0] ib_iface.c:994 UCX DEBUG iface=0x238d5c0: created UD QP 0x3106 on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638548.406812] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638548.407376] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.407384] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.407827] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.407832] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.405969] [ndv4:70462:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.406053] [ndv4:70462:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.406317] [ndv4:70462:0] ib_md.c:1604 UCX DEBUG mlx5_ib4: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.406350] [ndv4:70462:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638548.406356] [ndv4:70462:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638548.407694] [ndv4:68754:0] topo.c:99 UCX DEBUG bus id 0x103000000 doesn't exist. sys_dev = 2
[1650638548.407704] [ndv4:68754:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.408276] [ndv4:68766:0] ib_md.c:812 UCX DEBUG registered memory 0x2b8e7bedd000..0x2b8e7bf62000 on mlx5_ib0 lkey 0x81500 rkey 0x81500 access 0xf flags 0x3e4
[1650638548.408283] [ndv4:68766:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b8e7bedd018 of 544744 bytes with 128 elements
[1650638548.408288] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638548.408961] [ndv4:68766:0] ud_iface.c:393 UCX DEBUG iface 0x238d5c0: adding gid fe80::15:5dff:fd34:1b to hash on device mlx5_ib0 port 1 index 0)
[1650638548.409436] [ndv4:68755:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650638548.409448] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638548.409881] [ndv4:68766:0] ud_iface.c:393 UCX DEBUG iface 0x238d5c0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650638548.410395] [ndv4:70038:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.410425] [ndv4:70038:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.410640] [ndv4:70038:0] ib_md.c:1604 UCX DEBUG mlx5_ib2: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.410752] [ndv4:70038:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638548.410761] [ndv4:70038:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638548.410504] [ndv4:69048:0] debug.c:1198 UCX DEBUG using signal stack 0x2b565b69f000 size 141824
[1650638548.410598] [ndv4:69048:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650638548.410617] [ndv4:69048:0] 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 0x2b565b4f3000
[1650638548.410641] [ndv4:69048:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650638548.410649] [ndv4:69048:0] 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
[1650638548.410656] [ndv4:69048:0] module.c:253 UCX DEBUG loading modules for ucs
[1650638548.410973] [ndv4:68754:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.410981] [ndv4:68754:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.411282] [ndv4:68764:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638548.411294] [ndv4:68764:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.411616] [ndv4:68764:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.412248] [ndv4:70038:0] topo.c:99 UCX DEBUG bus id 0x103000000 doesn't exist. sys_dev = 2
[1650638548.412254] [ndv4:70038:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.414055] [ndv4:68946:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650638548.413282] [ndv4:70055:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.413295] [ndv4:70055:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.413605] [ndv4:70055:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.413609] [ndv4:70055:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.413897] [ndv4:70055:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.413602] [ndv4:69048:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650638548.413630] [ndv4:69048:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650638548.413669] [ndv4:69048:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650638548.413672] [ndv4:69048:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650638548.413681] [ndv4:69048:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650638548.413689] [ndv4:69048:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650638548.413692] [ndv4:69048:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650638548.413698] [ndv4:69048:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650638548.413700] [ndv4:69048:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650638548.413702] [ndv4:69048:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650638548.413705] [ndv4:69048:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650638548.413707] [ndv4:69048:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650638548.413717] [ndv4:69048:0] module.c:253 UCX DEBUG loading modules for uct
[1650638548.414958] [ndv4:68946:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650638548.415845] [ndv4:69236:0] async.c:228 UCX DEBUG added async handler 0x1b77820 [id=78 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638548.415875] [ndv4:69236:0] async.c:506 UCX DEBUG listening to async event fd 78 events 0x1 mode thread_spinlock
[1650638548.415947] [ndv4:69236:0] ucp_worker.c:1159 UCX DEBUG created interface[2]=0x1c21f80 using rc_mlx5/mlx5_ib0:1 on worker 0x24e34d0
[1650638548.416003] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.416010] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.418034] [ndv4:69075:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.418048] [ndv4:69075:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.419988] [ndv4:68764:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib2 vendor_id: 0x15b3 device_id: 4124
[1650638548.420406] [ndv4:68764:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib2: disable ODP because it's not supported for DevX QP
[1650638548.421175] [ndv4:68764:0] async.c:228 UCX DEBUG added async handler 0x2540d60 [id=61 ref 1] uct_ib_async_event_handler() to hash
[1650638548.421276] [ndv4:68764:0] async.c:506 UCX DEBUG listening to async event fd 61 events 0x1 mode thread_spinlock
[1650638548.421281] [ndv4:68764:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib2' (InfiniBand channel adapter) with 1 ports
[1650638548.422517] [ndv4:68766:0] ud_iface.c:393 UCX DEBUG iface 0x238d5c0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650638548.421655] [ndv4:68764:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638548.421666] [ndv4:68764:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638548.421676] [ndv4:68764:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.421739] [ndv4:68764:0] ib_md.c:1319 UCX DEBUG mlx5_ib2: using registration cache
[1650638548.422094] [ndv4:69048:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650638548.422794] [ndv4:69048:0] ucp_context.c:1117 UCX DEBUG closing md posix because it has no selected transport resources
[1650638548.422827] [ndv4:69048:0] ucp_context.c:1117 UCX DEBUG closing md self because it has no selected transport resources
[1650638548.423657] [ndv4:69048:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650638548.423670] [ndv4:69048:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650638548.423680] [ndv4:69048:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650638548.423691] [ndv4:69048:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650638548.423703] [ndv4:69048:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650638548.423715] [ndv4:69048:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650638548.423725] [ndv4:69048:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650638548.423760] [ndv4:69048:0] ucp_context.c:1117 UCX DEBUG closing md tcp because it has no selected transport resources
[1650638548.424151] [ndv4:69048:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650638548.423492] [ndv4:68766:0] ud_iface.c:393 UCX DEBUG iface 0x238d5c0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650638548.423521] [ndv4:68766:0] ud_iface.c:393 UCX DEBUG iface 0x238d5c0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650638548.423536] [ndv4:68766:0] ud_iface.c:393 UCX DEBUG iface 0x238d5c0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650638548.423550] [ndv4:68766:0] ud_iface.c:393 UCX DEBUG iface 0x238d5c0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650638548.423565] [ndv4:68766:0] ud_iface.c:393 UCX DEBUG iface 0x238d5c0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650638548.423979] [ndv4:68766:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638548.423738] [ndv4:68755:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650638548.423751] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638548.424826] [ndv4:68764:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.424841] [ndv4:68764:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.425034] [ndv4:68764:0] ib_md.c:1604 UCX DEBUG mlx5_ib2: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.425135] [ndv4:68764:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638548.425151] [ndv4:68764:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638548.425966] [ndv4:68946:0] async.c:228 UCX DEBUG added async handler 0x22b9390 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650638548.426090] [ndv4:68946:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650638548.426355] [ndv4:68762:0] debug.c:1198 UCX DEBUG using signal stack 0x2b6485377000 size 141824
[1650638548.426431] [ndv4:68762:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650638548.426453] [ndv4:68762:0] 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 0x2b64851e5000
[1650638548.426476] [ndv4:68762:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650638548.426486] [ndv4:68762:0] 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
[1650638548.426493] [ndv4:68762:0] module.c:253 UCX DEBUG loading modules for ucs
[1650638548.426601] [ndv4:68946:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650638548.426905] [ndv4:68946:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.426916] [ndv4:68946:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.426955] [ndv4:68946:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.427033] [ndv4:68946:0] module.c:253 UCX DEBUG loading modules for ucm
[1650638548.427059] [ndv4:68946:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650638548.427383] [ndv4:68754:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.427396] [ndv4:68754:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.428612] [ndv4:68946:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.428627] [ndv4:68946:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.428855] [ndv4:68946:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.429392] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.429406] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.429611] [ndv4:68762:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650638548.429628] [ndv4:68762:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650638548.429663] [ndv4:68762:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650638548.429667] [ndv4:68762:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650638548.429674] [ndv4:68762:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650638548.429681] [ndv4:68762:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650638548.429684] [ndv4:68762:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650638548.429688] [ndv4:68762:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650638548.429691] [ndv4:68762:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650638548.429694] [ndv4:68762:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650638548.429696] [ndv4:68762:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650638548.429699] [ndv4:68762:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650638548.429705] [ndv4:68762:0] module.c:253 UCX DEBUG loading modules for uct
[1650638548.430614] [ndv4:69236:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib0:1
[1650638548.431195] [ndv4:70439:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638548.431231] [ndv4:70439:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.431571] [ndv4:70439:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.431724] [ndv4:69236:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638548.431745] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638548.431748] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638548.431768] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638548.432086] [ndv4:70462:0] topo.c:99 UCX DEBUG bus id 0x105000000 doesn't exist. sys_dev = 4
[1650638548.432099] [ndv4:70462:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638548.432423] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650638548.432442] [ndv4:69236:0] 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> }
[1650638548.432480] [ndv4:69236:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib0 length=2048) failed: Invalid argument
[1650638548.432485] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638548.432498] [ndv4:69236:0] async.c:228 UCX DEBUG added async handler 0x1b79440 [id=80 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638548.432521] [ndv4:69236:0] async.c:506 UCX DEBUG listening to async event fd 80 events 0x1 mode thread_spinlock
[1650638548.432837] [ndv4:69048:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650638548.433098] [ndv4:69236:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650638548.433316] [ndv4:69048:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650638548.433736] [ndv4:68766:0] async.c:228 UCX DEBUG added async handler 0x22e68f0 [id=81 ref 1] uct_ud_iface_async_handler() to hash
[1650638548.433773] [ndv4:68766:0] async.c:506 UCX DEBUG listening to async event fd 81 events 0x5 mode thread_spinlock
[1650638548.433838] [ndv4:68766:0] ucp_worker.c:1159 UCX DEBUG created interface[4]=0x238d5c0 using ud_verbs/mlx5_ib0:1 on worker 0x2c514d0
[1650638548.433949] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.433956] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.434006] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.434011] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.434178] [ndv4:68766:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib0:1
[1650638548.434596] [ndv4:70055:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib3 vendor_id: 0x15b3 device_id: 4124
[1650638548.434888] [ndv4:68766:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638548.435136] [ndv4:70055:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib3: disable ODP because it's not supported for DevX QP
[1650638548.435341] [ndv4:68766:0] ib_iface.c:994 UCX DEBUG iface=0x23988f0: created UD QP 0x3108 on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638548.435356] [ndv4:68766:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650638548.436548] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638548.436697] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.436703] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.437014] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.437019] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.436493] [ndv4:68762:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650638548.436768] [ndv4:68762:0] ucp_context.c:1117 UCX DEBUG closing md posix because it has no selected transport resources
[1650638548.436788] [ndv4:68762:0] ucp_context.c:1117 UCX DEBUG closing md self because it has no selected transport resources
[1650638548.436860] [ndv4:68762:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650638548.436872] [ndv4:68762:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650638548.436882] [ndv4:68762:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650638548.436891] [ndv4:68762:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650638548.436901] [ndv4:68762:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650638548.436912] [ndv4:68762:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650638548.436922] [ndv4:68762:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650638548.436956] [ndv4:68762:0] ucp_context.c:1117 UCX DEBUG closing md tcp because it has no selected transport resources
[1650638548.437486] [ndv4:68766:0] ib_md.c:812 UCX DEBUG registered memory 0x2b8e7bf62000..0x2b8e7bfe7000 on mlx5_ib0 lkey 0x81600 rkey 0x81600 access 0xf flags 0x3e4
[1650638548.437492] [ndv4:68766:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b8e7bf62018 of 544744 bytes with 128 elements
[1650638548.437497] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638548.437770] [ndv4:68766:0] ud_iface.c:393 UCX DEBUG iface 0x23988f0: adding gid fe80::15:5dff:fd34:1b to hash on device mlx5_ib0 port 1 index 0)
[1650638548.438189] [ndv4:68766:0] ud_iface.c:393 UCX DEBUG iface 0x23988f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650638548.438279] [ndv4:68766:0] ud_iface.c:393 UCX DEBUG iface 0x23988f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650638548.437976] [ndv4:70038:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.437993] [ndv4:70038:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.438466] [ndv4:70038:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.438472] [ndv4:70038:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.439099] [ndv4:70038:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.439103] [ndv4:70038:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.437391] [ndv4:68762:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650638548.438823] [ndv4:70439:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib2 vendor_id: 0x15b3 device_id: 4124
[1650638548.439192] [ndv4:70439:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib2: disable ODP because it's not supported for DevX QP
[1650638548.438456] [ndv4:68946:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.438469] [ndv4:68946:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.439374] [ndv4:68946:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650638548.439379] [ndv4:68946:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.439685] [ndv4:68946:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.439689] [ndv4:68946:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.440948] [ndv4:68946:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.440954] [ndv4:68946:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.441608] [ndv4:69236:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x280c050: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0x312f
[1650638548.442128] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.442135] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.442165] [ndv4:69236:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2abadffdb008 of 151544 bytes with 1052 elements
[1650638548.441381] [ndv4:68946:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.441388] [ndv4:68946:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.442413] [ndv4:68946:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.442418] [ndv4:68946:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.442691] [ndv4:68946:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.442547] [ndv4:69048:0] async.c:228 UCX DEBUG added async handler 0x2541340 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650638548.442644] [ndv4:69048:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650638548.442659] [ndv4:69048:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650638548.443321] [ndv4:70462:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650638548.443331] [ndv4:70462:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638548.443934] [ndv4:70462:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650638548.443938] [ndv4:70462:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638548.444877] [ndv4:70462:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650638548.444882] [ndv4:70462:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638548.443858] [ndv4:68762:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650638548.444257] [ndv4:68762:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650638548.443053] [ndv4:69048:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.443063] [ndv4:69048:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.443087] [ndv4:69048:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.443153] [ndv4:69048:0] module.c:253 UCX DEBUG loading modules for ucm
[1650638548.443176] [ndv4:69048:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650638548.447505] [ndv4:68766:0] ud_iface.c:393 UCX DEBUG iface 0x23988f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650638548.447710] [ndv4:68766:0] ud_iface.c:393 UCX DEBUG iface 0x23988f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650638548.447872] [ndv4:68766:0] ud_iface.c:393 UCX DEBUG iface 0x23988f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650638548.448275] [ndv4:68766:0] ud_iface.c:393 UCX DEBUG iface 0x23988f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650638548.448598] [ndv4:68766:0] ud_iface.c:393 UCX DEBUG iface 0x23988f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650638548.448613] [ndv4:68766:0] 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> }
[1650638548.448652] [ndv4:68766:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638548.448659] [ndv4:68766:0] async.c:228 UCX DEBUG added async handler 0x22e7a90 [id=82 ref 1] uct_ud_iface_async_handler() to hash
[1650638548.448686] [ndv4:68766:0] async.c:506 UCX DEBUG listening to async event fd 82 events 0x5 mode thread_spinlock
[1650638548.448707] [ndv4:68766:0] ucp_worker.c:1159 UCX DEBUG created interface[5]=0x23988f0 using ud_mlx5/mlx5_ib0:1 on worker 0x2c514d0
[1650638548.448726] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638548.448732] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638548.448790] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638548.448794] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638548.446123] [ndv4:69236:0] ib_md.c:812 UCX DEBUG registered memory 0x2abae6200000..0x2abae8800000 on mlx5_ib0 lkey 0x81700 rkey 0x81700 access 0xf flags 0x3e4
[1650638548.446144] [ndv4:69236:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2abae6200018 of 39845864 bytes with 4752 elements
[1650638548.446306] [ndv4:69236:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x280c050
[1650638548.446369] [ndv4:69236:0] ucp_worker.c:1159 UCX DEBUG created interface[3]=0x280c050 using dc_mlx5/mlx5_ib0:1 on worker 0x24e34d0
[1650638548.447036] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.447045] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.447184] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.447189] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.447997] [ndv4:69236:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib0:1
[1650638548.449071] [ndv4:69236:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638548.447983] [ndv4:70462:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650638548.447989] [ndv4:70462:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638548.448275] [ndv4:70462:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.448404] [ndv4:69048:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.448417] [ndv4:69048:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.448607] [ndv4:69048:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.448625] [ndv4:69048:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.448630] [ndv4:69048:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.446116] [ndv4:70055:0] async.c:228 UCX DEBUG added async handler 0x2350420 [id=63 ref 1] uct_ib_async_event_handler() to hash
[1650638548.446148] [ndv4:70055:0] async.c:506 UCX DEBUG listening to async event fd 63 events 0x1 mode thread_spinlock
[1650638548.446153] [ndv4:70055:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib3' (InfiniBand channel adapter) with 1 ports
[1650638548.446369] [ndv4:70055:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638548.446378] [ndv4:70055:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638548.446392] [ndv4:70055:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.446450] [ndv4:70055:0] ib_md.c:1319 UCX DEBUG mlx5_ib3: using registration cache
[1650638548.445096] [ndv4:68764:0] topo.c:99 UCX DEBUG bus id 0x103000000 doesn't exist. sys_dev = 2
[1650638548.445107] [ndv4:68764:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.447700] [ndv4:68764:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.447707] [ndv4:68764:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.449493] [ndv4:69236:0] ib_iface.c:994 UCX DEBUG iface=0x1c1f5c0: created UD QP 0x3111 on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638548.450182] [ndv4:68766:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib1:1
[1650638548.450017] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638548.450052] [ndv4:70439:0] async.c:228 UCX DEBUG added async handler 0x1c1d0c0 [id=61 ref 1] uct_ib_async_event_handler() to hash
[1650638548.450088] [ndv4:70439:0] async.c:506 UCX DEBUG listening to async event fd 61 events 0x1 mode thread_spinlock
[1650638548.450092] [ndv4:70439:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib2' (InfiniBand channel adapter) with 1 ports
[1650638548.450350] [ndv4:70439:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638548.450358] [ndv4:70439:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638548.450366] [ndv4:70439:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.450420] [ndv4:70439:0] ib_md.c:1319 UCX DEBUG mlx5_ib2: using registration cache
[1650638548.451502] [ndv4:68766:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650638548.451511] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638548.451515] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638548.451529] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638548.451160] [ndv4:69075:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.451178] [ndv4:69075:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.452076] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638548.452083] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650638548.452244] [ndv4:69075:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.452251] [ndv4:69075:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.452523] [ndv4:69075:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.452507] [ndv4:70055:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.452518] [ndv4:70055:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.452592] [ndv4:68946:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650638548.453196] [ndv4:68946:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650638548.452777] [ndv4:70055:0] ib_md.c:1604 UCX DEBUG mlx5_ib3: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.452970] [ndv4:70055:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638548.452976] [ndv4:70055:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638548.453071] [ndv4:70055:0] topo.c:99 UCX DEBUG bus id 0x104000000 doesn't exist. sys_dev = 3
[1650638548.453075] [ndv4:70055:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638548.453469] [ndv4:68766:0] ib_iface.c:994 UCX DEBUG iface=0x3521050: created RC QP 0x3051 on mlx5_ib1:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650638548.452804] [ndv4:68755:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650638548.452815] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638548.453175] [ndv4:68755:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.456624] [ndv4:70462:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib5 vendor_id: 0x15b3 device_id: 4124
[1650638548.456978] [ndv4:70462:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib5: disable ODP because it's not supported for DevX QP
[1650638548.455229] [ndv4:70439:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.455236] [ndv4:70439:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.455450] [ndv4:70439:0] ib_md.c:1604 UCX DEBUG mlx5_ib2: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.455587] [ndv4:70439:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638548.455593] [ndv4:70439:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638548.456686] [ndv4:68764:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.456695] [ndv4:68764:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.456767] [ndv4:68764:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.456770] [ndv4:68764:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.454484] [ndv4:68766:0] ucp_worker.c:1159 UCX DEBUG created interface[6]=0x3521050 using rc_verbs/mlx5_ib1:1 on worker 0x2c514d0
[1650638548.454641] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638548.454648] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638548.454846] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638548.454852] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638548.454517] [ndv4:68946:0] async.c:228 UCX DEBUG added async handler 0x22a21b0 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650638548.454548] [ndv4:68946:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650638548.454552] [ndv4:68946:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650638548.454713] [ndv4:68946:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638548.454722] [ndv4:68946:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638548.454728] [ndv4:68946:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.454790] [ndv4:68946:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650638548.456282] [ndv4:68754:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.456290] [ndv4:68754:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.456991] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.457001] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.457425] [ndv4:70462:0] async.c:228 UCX DEBUG added async handler 0xa964b0 [id=67 ref 1] uct_ib_async_event_handler() to hash
[1650638548.457456] [ndv4:70462:0] async.c:506 UCX DEBUG listening to async event fd 67 events 0x1 mode thread_spinlock
[1650638548.457460] [ndv4:70462:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib5' (InfiniBand channel adapter) with 1 ports
[1650638548.457750] [ndv4:70462:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638548.457759] [ndv4:70462:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638548.457769] [ndv4:70462:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.457822] [ndv4:70462:0] ib_md.c:1319 UCX DEBUG mlx5_ib5: using registration cache
[1650638548.457690] [ndv4:68764:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.457696] [ndv4:68764:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.457964] [ndv4:68764:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.459827] [ndv4:68946:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.459841] [ndv4:68946:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.460139] [ndv4:68946:0] ib_md.c:1604 UCX DEBUG mlx5_ib1: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.460364] [ndv4:68946:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638548.460371] [ndv4:68946:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638548.458754] [ndv4:68754:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.458761] [ndv4:68754:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.459020] [ndv4:68754:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.458825] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.458833] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.459827] [ndv4:69236:0] ib_md.c:812 UCX DEBUG registered memory 0x2abae887e000..0x2abae8903000 on mlx5_ib0 lkey 0x81800 rkey 0x81800 access 0xf flags 0x3e4
[1650638548.459837] [ndv4:69236:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2abae887e018 of 544744 bytes with 128 elements
[1650638548.459847] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638548.460384] [ndv4:69236:0] ud_iface.c:393 UCX DEBUG iface 0x1c1f5c0: adding gid fe80::15:5dff:fd34:1b to hash on device mlx5_ib0 port 1 index 0)
[1650638548.458810] [ndv4:69048:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650638548.458819] [ndv4:69048:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.460762] [ndv4:69048:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.460771] [ndv4:69048:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.461362] [ndv4:68755:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib6 vendor_id: 0x15b3 device_id: 4124
[1650638548.461707] [ndv4:68755:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib6: disable ODP because it's not supported for DevX QP
[1650638548.462571] [ndv4:68755:0] async.c:228 UCX DEBUG added async handler 0x1395d70 [id=69 ref 1] uct_ib_async_event_handler() to hash
[1650638548.462604] [ndv4:68755:0] async.c:506 UCX DEBUG listening to async event fd 69 events 0x1 mode thread_spinlock
[1650638548.462608] [ndv4:68755:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib6' (InfiniBand channel adapter) with 1 ports
[1650638548.463018] [ndv4:68755:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638548.463029] [ndv4:68755:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638548.463040] [ndv4:68755:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.463103] [ndv4:68755:0] ib_md.c:1319 UCX DEBUG mlx5_ib6: using registration cache
[1650638548.462448] [ndv4:68766:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib1:1
[1650638548.463706] [ndv4:69048:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.463713] [ndv4:69048:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.465956] [ndv4:68762:0] async.c:228 UCX DEBUG added async handler 0x1f04640 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650638548.466035] [ndv4:68762:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650638548.466051] [ndv4:68762:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650638548.463821] [ndv4:70439:0] topo.c:99 UCX DEBUG bus id 0x103000000 doesn't exist. sys_dev = 2
[1650638548.463830] [ndv4:70439:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.465062] [ndv4:70038:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.465072] [ndv4:70038:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.465378] [ndv4:70038:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.467465] [ndv4:68764:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib3 vendor_id: 0x15b3 device_id: 4124
[1650638548.467872] [ndv4:68764:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib3: disable ODP because it's not supported for DevX QP
[1650638548.467786] [ndv4:69236:0] ud_iface.c:393 UCX DEBUG iface 0x1c1f5c0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650638548.468441] [ndv4:68755:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.468449] [ndv4:68755:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.468693] [ndv4:68755:0] ib_md.c:1604 UCX DEBUG mlx5_ib6: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.468714] [ndv4:68755:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638548.468719] [ndv4:68755:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638548.468977] [ndv4:70462:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.468989] [ndv4:70462:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.469269] [ndv4:70462:0] ib_md.c:1604 UCX DEBUG mlx5_ib5: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.469521] [ndv4:70462:0] ib_md.c:296 UCX DEBUG mlx5_ib5: cuda GPUDirect RDMA is enabled
[1650638548.469527] [ndv4:70462:0] ib_md.c:296 UCX DEBUG mlx5_ib5: rocm GPUDirect RDMA is disabled
[1650638548.469554] [ndv4:68766:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638548.469570] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638548.469573] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638548.469591] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638548.470004] [ndv4:68766:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x347a010 of 8176 bytes with 127 elements
[1650638548.470304] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638548.470314] [ndv4:68766:0] 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> }
[1650638548.470346] [ndv4:68766:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650638548.470350] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638548.470362] [ndv4:68766:0] async.c:228 UCX DEBUG added async handler 0x22e6960 [id=85 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638548.470386] [ndv4:68766:0] async.c:506 UCX DEBUG listening to async event fd 85 events 0x1 mode thread_spinlock
[1650638548.470398] [ndv4:68766:0] ucp_worker.c:1159 UCX DEBUG created interface[7]=0x33a7020 using rc_mlx5/mlx5_ib1:1 on worker 0x2c514d0
[1650638548.470681] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638548.470687] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638548.470765] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638548.470770] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638548.469935] [ndv4:68764:0] async.c:228 UCX DEBUG added async handler 0x2540590 [id=63 ref 1] uct_ib_async_event_handler() to hash
[1650638548.469972] [ndv4:68764:0] async.c:506 UCX DEBUG listening to async event fd 63 events 0x1 mode thread_spinlock
[1650638548.469975] [ndv4:68764:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib3' (InfiniBand channel adapter) with 1 ports
[1650638548.470323] [ndv4:68764:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638548.470332] [ndv4:68764:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638548.470338] [ndv4:68764:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.470394] [ndv4:68764:0] ib_md.c:1319 UCX DEBUG mlx5_ib3: using registration cache
[1650638548.472310] [ndv4:68755:0] topo.c:99 UCX DEBUG bus id 0x107000000 doesn't exist. sys_dev = 6
[1650638548.472317] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650638548.471969] [ndv4:68766:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib1:1
[1650638548.473285] [ndv4:68766:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638548.473298] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638548.473301] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638548.473354] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638548.473542] [ndv4:68762:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.473555] [ndv4:68762:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.473580] [ndv4:68762:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.473639] [ndv4:68762:0] module.c:253 UCX DEBUG loading modules for ucm
[1650638548.473658] [ndv4:68762:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650638548.473776] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650638548.473784] [ndv4:68766:0] 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> }
[1650638548.473816] [ndv4:68766:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650638548.473820] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638548.473828] [ndv4:68766:0] async.c:228 UCX DEBUG added async handler 0x3529f90 [id=87 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638548.473848] [ndv4:68766:0] async.c:506 UCX DEBUG listening to async event fd 87 events 0x1 mode thread_spinlock
[1650638548.474583] [ndv4:68766:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650638548.475234] [ndv4:70439:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.475243] [ndv4:70439:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.474129] [ndv4:69075:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650638548.474644] [ndv4:69075:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650638548.475811] [ndv4:69236:0] ud_iface.c:393 UCX DEBUG iface 0x1c1f5c0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650638548.476657] [ndv4:69236:0] ud_iface.c:393 UCX DEBUG iface 0x1c1f5c0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650638548.475995] [ndv4:70055:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638548.476004] [ndv4:70055:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638548.474037] [ndv4:68754:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib3 vendor_id: 0x15b3 device_id: 4124
[1650638548.474413] [ndv4:68754:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib3: disable ODP because it's not supported for DevX QP
[1650638548.476009] [ndv4:68754:0] async.c:228 UCX DEBUG added async handler 0x2126b10 [id=63 ref 1] uct_ib_async_event_handler() to hash
[1650638548.476038] [ndv4:68754:0] async.c:506 UCX DEBUG listening to async event fd 63 events 0x1 mode thread_spinlock
[1650638548.476042] [ndv4:68754:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib3' (InfiniBand channel adapter) with 1 ports
[1650638548.476187] [ndv4:68754:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638548.476195] [ndv4:68754:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638548.476202] [ndv4:68754:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.476276] [ndv4:68754:0] ib_md.c:1319 UCX DEBUG mlx5_ib3: using registration cache
[1650638548.473930] [ndv4:68755:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650638548.473937] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650638548.476555] [ndv4:68755:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650638548.476561] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650638548.476951] [ndv4:68946:0] topo.c:99 UCX DEBUG bus id 0x102000000 doesn't exist. sys_dev = 1
[1650638548.476960] [ndv4:68946:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.479086] [ndv4:68762:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.479104] [ndv4:68762:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.479349] [ndv4:68762:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.479819] [ndv4:68762:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.479826] [ndv4:68762:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.479079] [ndv4:70439:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.479085] [ndv4:70439:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.479179] [ndv4:70055:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638548.479186] [ndv4:70055:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638548.480510] [ndv4:69048:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.480519] [ndv4:69048:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.481606] [ndv4:68754:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.481613] [ndv4:68754:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.481852] [ndv4:68754:0] ib_md.c:1604 UCX DEBUG mlx5_ib3: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.482281] [ndv4:68754:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638548.482287] [ndv4:68754:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638548.482362] [ndv4:69075:0] async.c:228 UCX DEBUG added async handler 0xaacf80 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650638548.482391] [ndv4:69075:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650638548.482395] [ndv4:69075:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650638548.482731] [ndv4:69075:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638548.482740] [ndv4:69075:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638548.482751] [ndv4:69075:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.482807] [ndv4:69075:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650638548.482263] [ndv4:69236:0] ud_iface.c:393 UCX DEBUG iface 0x1c1f5c0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650638548.483992] [ndv4:69236:0] ud_iface.c:393 UCX DEBUG iface 0x1c1f5c0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650638548.484873] [ndv4:69236:0] ud_iface.c:393 UCX DEBUG iface 0x1c1f5c0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650638548.482787] [ndv4:68766:0] dc_mlx5.c:1327 UCX DEBUG dc iface 0x36c6010: using 'dcs_quota' policy with 8 dcis and 4608 cqes, dct 0x307a
[1650638548.483010] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638548.483016] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638548.483026] [ndv4:68766:0] mpool.c:205 UCX DEBUG mpool rcache_mp: allocated chunk 0x2b8e86040008 of 151544 bytes with 1052 elements
[1650638548.484150] [ndv4:68946:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638548.484158] [ndv4:68946:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.484259] [ndv4:68946:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638548.484264] [ndv4:68946:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.485750] [ndv4:68946:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638548.485756] [ndv4:68946:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.485817] [ndv4:68946:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638548.485820] [ndv4:68946:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.486141] [ndv4:68946:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.484476] [ndv4:68754:0] topo.c:99 UCX DEBUG bus id 0x104000000 doesn't exist. sys_dev = 3
[1650638548.484487] [ndv4:68754:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638548.485241] [ndv4:70462:0] topo.c:99 UCX DEBUG bus id 0x106000000 doesn't exist. sys_dev = 5
[1650638548.485250] [ndv4:70462:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638548.485547] [ndv4:70439:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.485565] [ndv4:70439:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.486046] [ndv4:69236:0] ud_iface.c:393 UCX DEBUG iface 0x1c1f5c0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650638548.486397] [ndv4:70462:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650638548.486403] [ndv4:70462:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638548.486575] [ndv4:70462:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650638548.486579] [ndv4:70462:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638548.486582] [ndv4:69236:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638548.486968] [ndv4:68766:0] ib_md.c:812 UCX DEBUG registered memory 0x2b8e83a00000..0x2b8e86000000 on mlx5_ib1 lkey 0x81400 rkey 0x81400 access 0xf flags 0x3e4
[1650638548.486984] [ndv4:68766:0] mpool.c:205 UCX DEBUG mpool rc_recv_desc: allocated chunk 0x2b8e83a00018 of 39845864 bytes with 4752 elements
[1650638548.487121] [ndv4:68766:0] dc_mlx5.c:1346 UCX DEBUG created dc iface 0x36c6010
[1650638548.487154] [ndv4:68766:0] ucp_worker.c:1159 UCX DEBUG created interface[8]=0x36c6010 using dc_mlx5/mlx5_ib1:1 on worker 0x2c514d0
[1650638548.486708] [ndv4:68764:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.486722] [ndv4:68764:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.486910] [ndv4:68764:0] ib_md.c:1604 UCX DEBUG mlx5_ib3: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.486941] [ndv4:68764:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638548.486948] [ndv4:68764:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638548.487343] [ndv4:68764:0] topo.c:99 UCX DEBUG bus id 0x104000000 doesn't exist. sys_dev = 3
[1650638548.487349] [ndv4:68764:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638548.488626] [ndv4:68764:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638548.488634] [ndv4:68764:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638548.489283] [ndv4:68764:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638548.489289] [ndv4:68764:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638548.487800] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638548.487826] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638548.488161] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638548.488168] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638548.489483] [ndv4:68766:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib1:1
[1650638548.488599] [ndv4:70038:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib3 vendor_id: 0x15b3 device_id: 4124
[1650638548.489012] [ndv4:70038:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib3: disable ODP because it's not supported for DevX QP
[1650638548.490173] [ndv4:70038:0] async.c:228 UCX DEBUG added async handler 0x14f5590 [id=63 ref 1] uct_ib_async_event_handler() to hash
[1650638548.490227] [ndv4:70038:0] async.c:506 UCX DEBUG listening to async event fd 63 events 0x1 mode thread_spinlock
[1650638548.490230] [ndv4:70038:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib3' (InfiniBand channel adapter) with 1 ports
[1650638548.490499] [ndv4:70038:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638548.490508] [ndv4:70038:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638548.490517] [ndv4:70038:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.490568] [ndv4:70038:0] ib_md.c:1319 UCX DEBUG mlx5_ib3: using registration cache
[1650638548.490448] [ndv4:68762:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650638548.490458] [ndv4:68762:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.490844] [ndv4:69048:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.490853] [ndv4:69048:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.491098] [ndv4:69048:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.491023] [ndv4:68764:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638548.491029] [ndv4:68764:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638548.491279] [ndv4:68764:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638548.491284] [ndv4:68764:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638548.491516] [ndv4:68764:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.492499] [ndv4:68946:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib2 vendor_id: 0x15b3 device_id: 4124
[1650638548.492862] [ndv4:68946:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib2: disable ODP because it's not supported for DevX QP
[1650638548.493125] [ndv4:68946:0] async.c:228 UCX DEBUG added async handler 0x22b1a60 [id=61 ref 1] uct_ib_async_event_handler() to hash
[1650638548.493150] [ndv4:68946:0] async.c:506 UCX DEBUG listening to async event fd 61 events 0x1 mode thread_spinlock
[1650638548.493154] [ndv4:68946:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib2' (InfiniBand channel adapter) with 1 ports
[1650638548.497031] [ndv4:69236:0] async.c:228 UCX DEBUG added async handler 0x1b788f0 [id=81 ref 1] uct_ud_iface_async_handler() to hash
[1650638548.497075] [ndv4:69236:0] async.c:506 UCX DEBUG listening to async event fd 81 events 0x5 mode thread_spinlock
[1650638548.497361] [ndv4:69236:0] ucp_worker.c:1159 UCX DEBUG created interface[4]=0x1c1f5c0 using ud_verbs/mlx5_ib0:1 on worker 0x24e34d0
[1650638548.497391] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.497399] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.497460] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.497464] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.497674] [ndv4:69236:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib0:1
[1650638548.498493] [ndv4:69236:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638548.498971] [ndv4:69236:0] ib_iface.c:994 UCX DEBUG iface=0x1c2a8f0: created UD QP 0x3114 on mlx5_ib0:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638548.498987] [ndv4:69236:0] ib_mlx5.c:568 UCX DEBUG tx wq 65536 bytes [bb=64, nwqe=1024] mmio_mode bf_post
[1650638548.499682] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638548.499726] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.499731] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.499744] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.499748] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.500229] [ndv4:69236:0] ib_md.c:812 UCX DEBUG registered memory 0x2abae8903000..0x2abae8988000 on mlx5_ib0 lkey 0x81900 rkey 0x81900 access 0xf flags 0x3e4
[1650638548.500235] [ndv4:69236:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2abae8903018 of 544744 bytes with 128 elements
[1650638548.500240] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638548.500281] [ndv4:69236:0] ud_iface.c:393 UCX DEBUG iface 0x1c2a8f0: adding gid fe80::15:5dff:fd34:1b to hash on device mlx5_ib0 port 1 index 0)
[1650638548.500315] [ndv4:69236:0] ud_iface.c:393 UCX DEBUG iface 0x1c2a8f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 1)
[1650638548.500346] [ndv4:69236:0] ud_iface.c:393 UCX DEBUG iface 0x1c2a8f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 2)
[1650638548.500364] [ndv4:69236:0] ud_iface.c:393 UCX DEBUG iface 0x1c2a8f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 3)
[1650638548.500378] [ndv4:69236:0] ud_iface.c:393 UCX DEBUG iface 0x1c2a8f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 4)
[1650638548.500393] [ndv4:69236:0] ud_iface.c:393 UCX DEBUG iface 0x1c2a8f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 5)
[1650638548.500418] [ndv4:69236:0] ud_iface.c:393 UCX DEBUG iface 0x1c2a8f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 6)
[1650638548.500431] [ndv4:69236:0] ud_iface.c:393 UCX DEBUG iface 0x1c2a8f0: adding gid fe80:: to hash on device mlx5_ib0 port 1 index 7)
[1650638548.500443] [ndv4:69236:0] 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> }
[1650638548.500537] [ndv4:69236:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638548.500547] [ndv4:69236:0] async.c:228 UCX DEBUG added async handler 0x1b7[1650638548.493321] [ndv4:68946:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638548.493331] [ndv4:68946:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638548.493336] [ndv4:68946:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.493379] [ndv4:68946:0] ib_md.c:1319 UCX DEBUG mlx5_ib2: using registration cache
[1650638548.498489] [ndv4:68764:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib4 vendor_id: 0x15b3 device_id: 4124
[1650638548.498865] [ndv4:68764:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib4: disable ODP because it's not supported for DevX QP
[1650638548.499117] [ndv4:68764:0] async.c:228 UCX DEBUG added async handler 0x2c2bcf0 [id=65 ref 1] uct_ib_async_event_handler() to hash
[1650638548.499145] [ndv4:68764:0] async.c:506 UCX DEBUG listening to async event fd 65 events 0x1 mode thread_spinlock
[1650638548.499149] [ndv4:68764:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib4' (InfiniBand channel adapter) with 1 ports
[1650638548.499314] [ndv4:68764:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638548.499321] [ndv4:68764:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638548.499334] [ndv4:68764:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.499392] [ndv4:68764:0] ib_md.c:1319 UCX DEBUG mlx5_ib4: using registration cache
[1650638548.569385] [ndv4:68764:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.569414] [ndv4:68764:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.569753] [ndv4:68764:0] ib_md.c:1604 UCX DEBUG mlx5_ib4: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.570062] [ndv4:68764:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638548.570070] [ndv4:68764:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638548.499967] [ndv4:69919:0] debug.c:1198 UCX DEBUG using signal stack 0x2b889d013000 size 141824
[1650638548.500045] [ndv4:69919:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650638548.500064] [ndv4:69919:0] 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 0x2b889ce67000
[1650638548.500086] [ndv4:69919:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650638548.500094] [ndv4:69919:0] 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
[1650638548.500100] [ndv4:69919:0] module.c:253 UCX DEBUG loading modules for ucs
[1650638548.502483] [ndv4:69919:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650638548.502502] [ndv4:69919:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650638548.502534] [ndv4:69919:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650638548.502537] [ndv4:69919:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650638548.502543] [ndv4:69919:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650638548.502549] [ndv4:69919:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650638548.502552] [ndv4:69919:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650638548.502556] [ndv4:69919:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650638548.502559] [ndv4:69919:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650638548.502561] [ndv4:69919:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650638548.502563] [ndv4:69919:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650638548.502565] [ndv4:69919:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650638548.502573] [ndv4:69919:0] module.c:253 UCX DEBUG loading modules for uct
[1650638548.507732] [ndv4:69919:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650638548.508085] [ndv4:69919:0] ucp_context.c:1117 UCX DEBUG closing md posix because it has no selected transport resources
[1650638548.508109] [ndv4:69919:0] ucp_context.c:1117 UCX DEBUG closing md self because it has no selected transport resources
[1650638548.508229] [ndv4:69919:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650638548.508243] [ndv4:69919:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650638548.508253] [ndv4:69919:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650638548.508264] [ndv4:69919:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650638548.508274] [ndv4:69919:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650638548.508285] [ndv4:69919:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650638548.508296] [ndv4:69919:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650638548.508331] [ndv4:69919:0] ucp_context.c:1117 UCX DEBUG closing md tcp because it has no selected transport resources
[1650638548.508692] [ndv4:69919:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650638548.514540] [ndv4:69919:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650638548.514840] [ndv4:69919:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650638548.522394] [ndv4:69919:0] asy[1650638548.525155] [ndv4:70448:0] debug.c:1198 UCX DEBUG using signal stack 0x2b091fc2a000 size 141824
[1650638548.525269] [ndv4:70448:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650638548.525291] [ndv4:70448:0] 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 0x2b091fa98000
[1650638548.525312] [ndv4:70448:0] init.c:115 UCX DEBUG cmd line: osu_scatter
[1650638548.525321] [ndv4:70448:0] 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
[1650638548.525327] [ndv4:70448:0] module.c:253 UCX DEBUG loading modules for ucs
[1650638548.527637] [ndv4:70448:0] cpu.c:231 UCX DEBUG CPU does not support invariant TSC, using fallback timer
[1650638548.527658] [ndv4:70448:0] time.c:22 UCX DEBUG measured arch clock speed: 1000000.00 Hz
[1650638548.527693] [ndv4:70448:0] ucp_context.c:1361 UCX DEBUG estimated number of endpoints is 96
[1650638548.527696] [ndv4:70448:0] ucp_context.c:1368 UCX DEBUG estimated number of endpoints per node is 96
[1650638548.527703] [ndv4:70448:0] ucp_context.c:1375 UCX DEBUG estimated bcopy bandwidth is 5251268608.000000
[1650638548.527710] [ndv4:70448:0] ucp_context.c:1427 UCX DEBUG allocation method[0] is md 'sysv'
[1650638548.527712] [ndv4:70448:0] ucp_context.c:1427 UCX DEBUG allocation method[1] is md 'posix'
[1650638548.527717] [ndv4:70448:0] ucp_context.c:1439 UCX DEBUG allocation method[2] is 'huge'
[1650638548.527719] [ndv4:70448:0] ucp_context.c:1439 UCX DEBUG allocation method[3] is 'thp'
[1650638548.527721] [ndv4:70448:0] ucp_context.c:1427 UCX DEBUG allocation method[4] is md '*'
[1650638548.527724] [ndv4:70448:0] ucp_context.c:1439 UCX DEBUG allocation method[5] is 'mmap'
[1650638548.527726] [ndv4:70448:0] ucp_context.c:1439 UCX DEBUG allocation method[6] is 'heap'
[1650638548.527734] [ndv4:70448:0] module.c:253 UCX DEBUG loading modules for uct
[1650638548.532720] [ndv4:70448:0] module.c:253 UCX DEBUG loading modules for uct_ib
[1650638548.533059] [ndv4:70448:0] ucp_context.c:1117 UCX DEBUG closing md posix because it has no selected transport resources
[1650638548.533079] [ndv4:70448:0] ucp_context.c:1117 UCX DEBUG closing md self because it has no selected transport resources
[1650638548.533244] [ndv4:70448:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib4) failed: Cannot assign requested address
[1650638548.533261] [ndv4:70448:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib3) failed: Cannot assign requested address
[1650638548.533274] [ndv4:70448:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib2) failed: Cannot assign requested address
[1650638548.533286] [ndv4:70448:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib7) failed: Cannot assign requested address
[1650638548.533298] [ndv4:70448:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib6) failed: Cannot assign requested address
[1650638548.533310] [ndv4:70448:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib5) failed: Cannot assign requested address
[1650638548.533322] [ndv4:70448:0] sock.c:88 UCX DEBUG ioctl(req=35093, ifr_name=ib1) failed: Cannot assign requested address
[1650638548.533355] [ndv4:70448:0] ucp_context.c:1117 UCX DEBUG closing md tcp because it has no selected transport resources
[1650638548.533709] [ndv4:70448:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: Success, continuing, but fork may be unsafe.
[1650638548.572886] [ndv4:70448:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib0 vendor_id: 0x15b3 device_id: 4124
[1650638548.573229] [ndv4:70448:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib0: disable ODP because it's not supported for DevX QP
[1650638548.580187] [ndv4:70448:0] asy9a90 [id=82 ref 1] uct_ud_iface_async_handler() to hash
[1650638548.500567] [ndv4:69236:0] async.c:506 UCX DEBUG listening to async event fd 82 events 0x5 mode thread_spinlock
[1650638548.500581] [ndv4:69236:0] ucp_worker.c:1159 UCX DEBUG created interface[5]=0x1c2a8f0 using ud_mlx5/mlx5_ib0:1 on worker 0x24e34d0
[1650638548.500593] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638548.500599] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638548.500656] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638548.500660] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638548.500832] [ndv4:69236:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib1:1
[1650638548.501496] [ndv4:69236:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 91 data_sz 8256
[1650638548.501505] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638548.501508] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638548.501525] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638548.501861] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638548.501866] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool rc_verbs_short_desc: align 64, maxelems 4294967295, elemsize 192
[1650638548.502498] [ndv4:69236:0] ib_iface.c:994 UCX DEBUG iface=0x2db3050: created RC QP 0x305d on mlx5_ib1:1 TX wr:409 sge:4 inl:124 resp:64 RX wr:0 sge:0 resp:64
[1650638548.503626] [ndv4:69236:0] ucp_worker.c:1159 UCX DEBUG created interface[6]=0x2db3050 using rc_verbs/mlx5_ib1:1 on worker 0x24e34d0
[1650638548.503646] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638548.503651] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638548.503719] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638548.503723] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638548.503901] [ndv4:69236:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib1:1
[1650638548.504988] [ndv4:69236:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638548.504998] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638548.505001] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638548.505017] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638548.505427] [ndv4:69236:0] mpool.c:205 UCX DEBUG mpool devx dbrec: allocated chunk 0x2d0c010 of 8176 bytes with 127 elements
[1650638548.505632] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 64
[1650638548.505641] [ndv4:69236:0] 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> }
[1650638548.505678] [ndv4:69236:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650638548.505682] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638548.505694] [ndv4:69236:0] async.c:228 UCX DEBUG added async handler 0x1b78960 [id=85 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638548.505717] [ndv4:69236:0] async.c:506 UCX DEBUG listening [1650638548.513875] [ndv4:68946:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.513896] [ndv4:68946:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.514072] [ndv4:68946:0] ib_md.c:1604 UCX DEBUG mlx5_ib2: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.514103] [ndv4:68946:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638548.514109] [ndv4:68946:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638548.514200] [ndv4:68946:0] topo.c:99 UCX DEBUG bus id 0x103000000 doesn't exist. sys_dev = 2
[1650638548.514244] [ndv4:68946:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.514307] [ndv4:68946:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.514311] [ndv4:68946:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.514459] [ndv4:68946:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.514462] [ndv4:68946:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.514559] [ndv4:68946:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.514562] [ndv4:68946:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.514621] [ndv4:68946:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.514623] [ndv4:68946:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.514873] [ndv4:68946:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.520880] [ndv4:68946:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib3 vendor_id: 0x15b3 device_id: 4124
[1650638548.521260] [ndv4:68946:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib3: disable ODP because it's not supported for DevX QP
[1650638548.521871] [ndv4:68946:0] async.c:228 UCX DEBUG added async handler 0x22ba680 [id=63 ref 1] uct_ib_async_event_handler() to hash
[1650638548.521904] [ndv4:68946:0] async.c:506 UCX DEBUG listening to async event fd 63 events 0x1 mode thread_spinlock
[1650638548.521907] [ndv4:68946:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib3' (InfiniBand channel adapter) with 1 ports
[1650638548.522041] [ndv4:68946:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638548.522049] [ndv4:68946:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638548.522063] [ndv4:68946:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.522227] [ndv4:68946:0] ib_md.c:1319 UCX DEBUG mlx5_ib3: using registration cache
[1650638548.498009] [ndv4:69048:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650638548.498390] [ndv4:69048:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650638548.498651] [ndv4:69048:0] async.c:228 UCX DEBUG added async handler 0x254b3e0 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650638548.498676] [ndv4:69048:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650638548.498680] [ndv4:69048:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650638548.498836] [ndv4:69048:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638548.498843] [ndv4:69048:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638548.498855] [ndv4:69048:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.498915] [ndv4:69048:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650638548.499703] [ndv4:69048:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.499710] [ndv4:69048:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.499916] [ndv4:69048:0] ib_md.c:1604 UCX DEBUG mlx5_ib1: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.499937] [ndv4:69048:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638548.499942] [ndv4:69048:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638548.500028] [ndv4:69048:0] topo.c:99 UCX DEBUG bus id 0x102000000 doesn't exist. sys_dev = 1
[1650638548.500032] [ndv4:69048:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.500092] [ndv4:69048:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638548.500095] [ndv4:69048:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.500154] [ndv4:69048:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638548.500156] [ndv4:69048:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.500232] [ndv4:69048:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638548.500236] [ndv4:69048:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.500344] [ndv4:69048:0] topo.c:91 UCX DEBUG bus id 0x102000000 exists. sys_dev = 1
[1650638548.500347] [ndv4:69048:0] ib_device.c:1124 UCX DEBUG mlx5_ib1 bus id 258:0:0.0 sys_dev 1
[1650638548.500594] [ndv4:69048:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.506978] [ndv4:69048:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib2 vendor_id: 0x15b3 device_id: 4124
[1650638548.507382] [ndv4:69048:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib2: disable ODP because it's not supported for DevX QP
[1650638548.507622] [ndv4:69048:0] async.c:228 UCX DEBUG added async handler 0x251d0c0 [id=61 ref 1] uct_ib_async_event_handler() to hash
[1650638548.507645] [ndv4:69048:0] async.c:506 UCX DEBUG listening to async event fd 61 events 0x1 mode thread_spinlock
[1650638548.507648] [ndv4:69048:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib2' (InfiniBand channel adapter) with 1 ports
[1650638548.507875] [ndv4:69048:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638548.507883] [ndv4:69048:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638548.507887] [ndv4:69048:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.507929] [ndv4:69048:0] ib_md.c:1319 UCX DEBUG mlx5_ib2: usinto async event fd 85 events 0x1 mode thread_spinlock
[1650638548.505726] [ndv4:69236:0] ucp_worker.c:1159 UCX DEBUG created interface[7]=0x2c39020 using rc_mlx5/mlx5_ib1:1 on worker 0x24e34d0
[1650638548.505739] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638548.505743] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638548.505812] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638548.505816] [ndv4:69236:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638548.505968] [ndv4:69236:0] ib_iface.c:857 UCX DEBUG using pkey[0] 0x8019 on mlx5_ib1:1
[1650638548.507019] [ndv4:69236:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 90 data_sz 8256
[1650638548.507038] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool rc_recv_desc: align 64, maxelems 4294967295, elemsize 8356
[1650638548.507042] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool rc_send_desc: align 64, maxelems 4294967295, elemsize 8320
[1650638548.507094] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool send-ops-mpool: align 64, maxelems 4294967295, elemsize 48
[1650638548.586545] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool pending-ops: align 1, maxelems 4294967295, elemsize 112
[1650638548.586595] [ndv4:69236:0] 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> }
[1650638548.586642] [ndv4:69236:0] rc_mlx5_common.c:727 UCX DEBUG ibv_alloc_dm(dev=mlx5_ib1 length=2048) failed: Invalid argument
[1650638548.586651] [ndv4:69236:0] mpool.c:88 UCX DEBUG mpool rc_mlx5_atomic_desc: align 64, maxelems 4294967295, elemsize 72
[1650638548.586665] [ndv4:69236:0] async.c:228 UCX DEBUG added async handler 0x2dbbf90 [id=87 ref 1] uct_rc_mlx5_devx_iface_event_handler() to hash
[1650638548.586693] [ndv4:69236:0] async.c:506 UCX DEBUG listening to async event fd 87 events 0x1 mode thread_spinlock
[1650638548.587458] [ndv4:69236:0] dc_mlx5.c:823 UCX DEBUG creating dci pool 0 with 8 QPs
[1650638548.573643] [ndv4:69075:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.573679] [ndv4:69075:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.573987] [ndv4:69075:0] ib_md.c:1604 UCX DEBUG mlx5_ib1: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.574043] [ndv4:69075:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638548.574056] [ndv4:69075:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638548.570689] [ndv4:68755:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650638548.570702] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650638548.570912] [ndv4:68755:0] topo.c:91 UCX DEBUG bus id 0x107000000 exists. sys_dev = 6
[1650638548.570916] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib6 bus id 263:0:0.0 sys_dev 6
[1650638548.571196] [ndv4:68755:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.579372] [ndv4:68755:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib7 vendor_id: 0x15b3 device_id: 4124
[1650638548.579758] [ndv4:68755:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib7: disable ODP because it's not supported for DevX QP
[1650638548.580016] [ndv4:68755:0] async.c:228 UCX DEBUG added async handler 0xcaadb0 [id=71 ref 1] uct_ib_async_event_handler() to hash
[1650638548.580048] [ndv4:68755:0] async.c:506 UCX DEBUG listening to async event fd 71 events 0x1 mode thread_spinlock
[1650638548.580052] [ndv4:68755:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib7' (InfiniBand channel adapter) with 1 ports
[1650638548.580246] [ndv4:68755:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638548.580258] [ndv4:68755:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638548.580271] [ndv4:68755:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.580341] [ndv4:68755:0] ib_md.c:1319 UCX DEBUG mlx5_ib7: using registration cache
[1650638548.588731] [ndv4:68755:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.588834] [ndv4:68755:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
nc.c:228 UCX DEBUG added async handler 0x11dc340 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650638548.522486] [ndv4:69919:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650638548.522575] [ndv4:69919:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650638548.522776] [ndv4:69919:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.522784] [ndv4:69919:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.522813] [ndv4:69919:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.522900] [ndv4:69919:0] module.c:253 UCX DEBUG loading modules for ucm
[1650638548.522925] [ndv4:69919:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650638548.566126] [ndv4:69919:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.566146] [ndv4:69919:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.566377] [ndv4:69919:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.566401] [ndv4:69919:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.566407] [ndv4:69919:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.567049] [ndv4:69919:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650638548.567054] [ndv4:69919:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.567978] [ndv4:69919:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.567983] [ndv4:69919:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.569007] [ndv4:69919:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.569013] [ndv4:69919:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.569280] [ndv4:69919:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.569285] [ndv4:69919:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.569489] [ndv4:69919:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.569493] [ndv4:69919:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.569786] [ndv4:69919:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.574068] [ndv4:70462:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650638548.574081] [ndv4:70462:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638548.575038] [ndv4:70462:0] topo.c:91 UCX DEBUG bus id 0x106000000 exists. sys_dev = 5
[1650638548.575044] [ndv4:70462:0] ib_device.c:1124 UCX DEBUG mlx5_ib5 bus id 262:0:0.0 sys_dev 5
[1650638548.575350] [ndv4:70462:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.582784] [ndv4:70462:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib6 vendor_id: 0x15b3 device_id: 4124
[1650638548.583140] [ndv4:70462:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib6: disable ODP because it's not supported for DevX QP
[1650638548.583413] [ndv4:70462:0] async.c:228 UCX DEBUG added async handler 0x1178da0 [id=69 ref 1] uct_ib_async_event_handler() to hash
[1650638548.583437] [ndv4:70462:0] async.c:506 UCX DEBUG listening to async event fd 69 events 0x1 mode thread_spinlock
[1650638548.583441] [ndv4:70462:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib6' (InfiniBand channel adapter) with 1 ports
[1650638548.583619] [ndv4:70462:0] ib_md.c:296 UCX DEBUG mlx5_ib6: cuda GPUDirect RDMA is enabled
[1650638548.583628] [ndv4:70462:0] ib_md.c:296 UCX DEBUG mlx5_ib6: rocm GPUDirect RDMA is disabled
[1650638548.583640] [ndv4:70462:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.583695] [ndv4:70462:0] ib_md.c:1319 UCX DEBUG mlx5_ib6: using registration cache
[1650638548.565841] [ndv4:70055:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638548.565852] [ndv4:70055:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638548.566358] [ndv4:70055:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638548.566362] [ndv4:70055:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638548.566620] [ndv4:70055:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
nc.c:228 UCX DEBUG added async handler 0x136e690 [id=54 ref 1] uct_ib_async_event_handler() to hash
[1650638548.580292] [ndv4:70448:0] async.c:506 UCX DEBUG listening to async event fd 54 events 0x1 mode thread_spinlock
[1650638548.580782] [ndv4:70448:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib0' (InfiniBand channel adapter) with 1 ports
[1650638548.580991] [ndv4:70448:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.580999] [ndv4:70448:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.581021] [ndv4:70448:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.581081] [ndv4:70448:0] module.c:253 UCX DEBUG loading modules for ucm
[1650638548.581102] [ndv4:70448:0] ib_md.c:1319 UCX DEBUG mlx5_ib0: using registration cache
[1650638548.582800] [ndv4:70448:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.582813] [ndv4:70448:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.582976] [ndv4:70448:0] ib_md.c:1604 UCX DEBUG mlx5_ib0: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.582998] [ndv4:70448:0] ib_md.c:296 UCX DEBUG mlx5_ib0: cuda GPUDirect RDMA is enabled
[1650638548.583004] [ndv4:70448:0] ib_md.c:296 UCX DEBUG mlx5_ib0: rocm GPUDirect RDMA is disabled
[1650638548.583221] [ndv4:70448:0] topo.c:99 UCX DEBUG bus id 0x101000000 doesn't exist. sys_dev = 0
[1650638548.583226] [ndv4:70448:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.583335] [ndv4:70448:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.583339] [ndv4:70448:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.583556] [ndv4:70448:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.583561] [ndv4:70448:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.583690] [ndv4:70448:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.583693] [ndv4:70448:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.583855] [ndv4:70448:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.583858] [ndv4:70448:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.584085] [ndv4:70448:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
g registration cache
[1650638548.512587] [ndv4:69048:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.512593] [ndv4:69048:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.512830] [ndv4:69048:0] ib_md.c:1604 UCX DEBUG mlx5_ib2: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.512841] [ndv4:69048:0] ib_md.c:296 UCX DEBUG mlx5_ib2: cuda GPUDirect RDMA is enabled
[1650638548.512859] [ndv4:69048:0] ib_md.c:296 UCX DEBUG mlx5_ib2: rocm GPUDirect RDMA is disabled
[1650638548.512935] [ndv4:69048:0] topo.c:99 UCX DEBUG bus id 0x103000000 doesn't exist. sys_dev = 2
[1650638548.512938] [ndv4:69048:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.512998] [ndv4:69048:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.513001] [ndv4:69048:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.513059] [ndv4:69048:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.513062] [ndv4:69048:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.513119] [ndv4:69048:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.513122] [ndv4:69048:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.513179] [ndv4:69048:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.513181] [ndv4:69048:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.513425] [ndv4:69048:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.520414] [ndv4:69048:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib3 vendor_id: 0x15b3 device_id: 4124
[1650638548.520761] [ndv4:69048:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib3: disable ODP because it's not supported for DevX QP
[1650638548.521007] [ndv4:69048:0] async.c:228 UCX DEBUG added async handler 0x2c2c070 [id=63 ref 1] uct_ib_async_event_handler() to hash
[1650638548.521029] [ndv4:69048:0] async.c:506 UCX DEBUG listening to async event fd 63 events 0x1 mode thread_spinlock
[1650638548.521033] [ndv4:69048:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib3' (InfiniBand channel adapter) with 1 ports
[1650638548.521194] [ndv4:69048:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638548.521201] [ndv4:69048:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638548.521224] [ndv4:69048:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.521264] [ndv4:69048:0] ib_md.c:1319 UCX DEBUG mlx5_ib3: using registration cache
[1650638548.523417] [ndv4:69048:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.523430] [ndv4:69048:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.523624] [ndv4:69048:0] ib_md.c:1604 UCX DEBUG mlx5_ib3: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.523651] [ndv4:69048:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638548.523656] [ndv4:69048:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638548.523750] [ndv4:69048:0] topo.c:99 UCX DEBUG bus id 0x104000000 doesn't exist. sys_dev = 3
[1650638548.523754] [ndv4:69048:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638548.523837] [ndv4:69048:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638548.523840] [ndv4:69048:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[16506[1650638548.568825] [ndv4:70439:0] topo.c:91 UCX DEBUG bus id 0x103000000 exists. sys_dev = 2
[1650638548.568838] [ndv4:70439:0] ib_device.c:1124 UCX DEBUG mlx5_ib2 bus id 259:0:0.0 sys_dev 2
[1650638548.569194] [ndv4:70439:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.576781] [ndv4:70439:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib3 vendor_id: 0x15b3 device_id: 4124
[1650638548.577132] [ndv4:70439:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib3: disable ODP because it's not supported for DevX QP
[1650638548.577428] [ndv4:70439:0] async.c:228 UCX DEBUG added async handler 0x232c070 [id=63 ref 1] uct_ib_async_event_handler() to hash
[1650638548.577458] [ndv4:70439:0] async.c:506 UCX DEBUG listening to async event fd 63 events 0x1 mode thread_spinlock
[1650638548.577462] [ndv4:70439:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib3' (InfiniBand channel adapter) with 1 ports
[1650638548.577607] [ndv4:70439:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638548.577615] [ndv4:70439:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638548.577625] [ndv4:70439:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.577689] [ndv4:70439:0] ib_md.c:1319 UCX DEBUG mlx5_ib3: using registration cache
[1650638548.587010] [ndv4:68766:0] ib_iface.c:1469 UCX DEBUG created uct_ib_iface_t headroom_ofs 12 payload_ofs 92 hdr_ofs 44 data_sz 4096
[1650638548.587879] [ndv4:68766:0] ib_iface.c:994 UCX DEBUG iface=0x238c6e0: created UD QP 0x305e on mlx5_ib1:1 TX wr:341 sge:5 inl:124 resp:0 RX wr:4096 sge:1 resp:0
[1650638548.589172] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool ud_recv_skb: align 64, maxelems 4294967295, elemsize 4196
[1650638548.589291] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638548.589298] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638548.589521] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638548.589527] [ndv4:68766:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638548.590173] [ndv4:68766:0] ib_md.c:812 UCX DEBUG registered memory 0x2b8e86065000..0x2b8e860ea000 on mlx5_ib1 lkey 0x81500 rkey 0x81500 access 0xf flags 0x3e4
[1650638548.590180] [ndv4:68766:0] mpool.c:205 UCX DEBUG mpool ud_recv_skb: allocated chunk 0x2b8e86065018 of 544744 bytes with 128 elements
[1650638548.590184] [ndv4:68766:0] mpool.c:88 UCX DEBUG mpool ud_tx_skb: align 64, maxelems 4294967295, elemsize 4168
[1650638548.590379] [ndv4:68766:0] ud_iface.c:393 UCX DEBUG iface 0x238c6e0: adding gid fe80::15:5dff:fd34:1c to hash on device mlx5_ib1 port 1 index 0)
[1650638548.590520] [ndv4:68766:0] ud_iface.c:393 UCX DEBUG iface 0x238c6e0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 1)
[1650638548.590609] [ndv4:68766:0] ud_iface.c:393 UCX DEBUG iface 0x238c6e0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 2)
[1650638548.590696] [ndv4:68766:0] ud_iface.c:393 UCX DEBUG iface 0x238c6e0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 3)
[1650638548.590809] [ndv4:68766:0] ud_iface.c:393 UCX DEBUG iface 0x238c6e0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 4)
[1650638548.590889] [ndv4:68766:0] ud_iface.c:393 UCX DEBUG iface 0x238c6e0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 5)
[1650638548.590935] [ndv4:68766:0] ud_iface.c:393 UCX DEBUG iface 0x238c6e0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 6)
[1650638548.590991] [ndv4:68766:0] ud_iface.c:393 UCX DEBUG iface 0x238c6e0: adding gid fe80:: to hash on device mlx5_ib1 port 1 index 7)
[1650638548.586647] [ndv4:68946:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.586664] [ndv4:68946:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.586905] [ndv4:68946:0] ib_md.c:1604 UCX DEBUG mlx5_ib3: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.587124] [ndv4:68946:0] ib_md.c:296 UCX DEBUG mlx5_ib3: cuda GPUDirect RDMA is enabled
[1650638548.587134] [ndv4:68946:0] ib_md.c:296 UCX DEBUG mlx5_ib3: rocm GPUDirect RDMA is disabled
[1650638548.587490] [ndv4:68946:0] topo.c:99 UCX DEBUG bus id 0x104000000 doesn't exist. sys_dev = 3
[1650638548.587499] [ndv4:68946:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638548.587679] [ndv4:68946:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638548.587683] [ndv4:68946:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638548.587988] [ndv4:68946:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638548.587993] [ndv4:68946:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638548.588126] [ndv4:68946:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638548.588129] [ndv4:68946:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638548.588275] [ndv4:68946:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638548.588278] [ndv4:68946:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638548.589062] [ndv4:68946:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
38548.524022] [ndv4:69048:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638548.524027] [ndv4:69048:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638548.524106] [ndv4:69048:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638548.524109] [ndv4:69048:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638548.524176] [ndv4:69048:0] topo.c:91 UCX DEBUG bus id 0x104000000 exists. sys_dev = 3
[1650638548.524179] [ndv4:69048:0] ib_device.c:1124 UCX DEBUG mlx5_ib3 bus id 260:0:0.0 sys_dev 3
[1650638548.524612] [ndv4:69048:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.531061] [ndv4:69048:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib4 vendor_id: 0x15b3 device_id: 4124
[1650638548.531406] [ndv4:69048:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib4: disable ODP because it's not supported for DevX QP
[1650638548.531659] [ndv4:69048:0] async.c:228 UCX DEBUG added async handler 0x25494c0 [id=65 ref 1] uct_ib_async_event_handler() to hash
[1650638548.531683] [ndv4:69048:0] async.c:506 UCX DEBUG listening to async event fd 65 events 0x1 mode thread_spinlock
[1650638548.531686] [ndv4:69048:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib4' (InfiniBand channel adapter) with 1 ports
[1650638548.531830] [ndv4:69048:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638548.531838] [ndv4:69048:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638548.531844] [ndv4:69048:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.589547] [ndv4:69048:0] ib_md.c:1319 UCX DEBUG mlx5_ib4: using registration cache
[1650638548.589070] [ndv4:68755:0] ib_md.c:1604 UCX DEBUG mlx5_ib7: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.589110] [ndv4:68755:0] ib_md.c:296 UCX DEBUG mlx5_ib7: cuda GPUDirect RDMA is enabled
[1650638548.589117] [ndv4:68755:0] ib_md.c:296 UCX DEBUG mlx5_ib7: rocm GPUDirect RDMA is disabled
[1650638548.589362] [ndv4:68755:0] topo.c:99 UCX DEBUG bus id 0x108000000 doesn't exist. sys_dev = 7
[1650638548.589367] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650638548.589552] [ndv4:68755:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650638548.589556] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650638548.589729] [ndv4:68755:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650638548.589733] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650638548.590355] [ndv4:68755:0] topo.c:91 UCX DEBUG bus id 0x108000000 exists. sys_dev = 7
[1650638548.590366] [ndv4:68755:0] ib_device.c:1124 UCX DEBUG mlx5_ib7 bus id 264:0:0.0 sys_dev 7
[1650638548.591038] [ndv4:68762:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.591059] [ndv4:68762:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.591152] [ndv4:68762:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.591155] [ndv4:68762:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.591276] [ndv4:68762:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.591285] [ndv4:68762:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.591364] [ndv4:68762:0] topo.c:91 UCX DEBUG bus id 0x101000000 exists. sys_dev = 0
[1650638548.591368] [ndv4:68762:0] ib_device.c:1124 UCX DEBUG mlx5_ib0 bus id 257:0:0.0 sys_dev 0
[1650638548.591706] [ndv4:68762:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.591068] [ndv4:70448:0] ib_device.c:542 UCX DEBUG PF: mlx5_ib1 vendor_id: 0x15b3 device_id: 4124
[1650638548.591489] [ndv4:70448:0] ib_mlx5dv_md.c:489 UCX DEBUG mlx5_ib1: disable ODP because it's not supported for DevX QP
[1650638548.591735] [ndv4:70448:0] async.c:228 UCX DEBUG added async handler 0x1366fa0 [id=59 ref 1] uct_ib_async_event_handler() to hash
[1650638548.591769] [ndv4:70448:0] async.c:506 UCX DEBUG listening to async event fd 59 events 0x1 mode thread_spinlock
[1650638548.591772] [ndv4:70448:0] ib_device.c:655 UCX DEBUG initialized device 'mlx5_ib1' (InfiniBand channel adapter) with 1 ports
[1650638548.591579] [ndv4:69048:0] ib_md.c:1499 UCX DEBUG incorrect format of current_link_speed file: expected: <double> GT/s, actual: Unknown speed
[1650638548.591596] [ndv4:69048:0] mpool.c:88 UCX DEBUG mpool devx dbrec: align 64, maxelems 4294967295, elemsize 40
[1650638548.591803] [ndv4:69048:0] ib_md.c:1604 UCX DEBUG mlx5_ib4: md open by 'uct_ib_mlx5_devx_md_ops' is successful
[1650638548.591829] [ndv4:69048:0] ib_md.c:296 UCX DEBUG mlx5_ib4: cuda GPUDirect RDMA is enabled
[1650638548.591835] [ndv4:69048:0] ib_md.c:296 UCX DEBUG mlx5_ib4: rocm GPUDirect RDMA is disabled
[1650638548.591928] [ndv4:69048:0] topo.c:99 UCX DEBUG bus id 0x105000000 doesn't exist. sys_dev = 4
[1650638548.591932] [ndv4:69048:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638548.592046] [ndv4:69048:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650638548.592050] [ndv4:69048:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638548.592108] [ndv4:69048:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650638548.592111] [ndv4:69048:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638548.592200] [ndv4:69048:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650638548.592235] [ndv4:69048:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638548.592299] [ndv4:69048:0] topo.c:91 UCX DEBUG bus id 0x105000000 exists. sys_dev = 4
[1650638548.592302] [ndv4:69048:0] ib_device.c:1124 UCX DEBUG mlx5_ib4 bus id 261:0:0.0 sys_dev 4
[1650638548.592539] [ndv4:69048:0] ib_md.c:1592 UCX DEBUG ibv_fork_init() failed: No such file or directory, continuing, but fork may be unsafe.
[1650638548.591997] [ndv4:70448:0] ib_md.c:296 UCX DEBUG mlx5_ib1: cuda GPUDirect RDMA is enabled
[1650638548.592007] [ndv4:70448:0] ib_md.c:296 UCX DEBUG mlx5_ib1: rocm GPUDirect RDMA is disabled
[1650638548.592022] [ndv4:70448:0] mpool.c:88 UCX DEBUG mpool rcache_mp: align 8, maxelems 4294967295, elemsize 144
[1650638548.592113] [ndv4:70448:0] ib_md.c:1319 UCX DEBUG mlx5_ib1: using registration cache
[1650638548.591320] [ndv4:68766:0] timer_wheel.c:40 UCX DEBUG high res timer created log=12 resolution=4096.000000 usec wanted: 2500.000000 usec
[1650638548.591335] [ndv4:68766:0] async.c:228 UCX DEBUG added async handler 0x237c350 [id=88 ref 1] uct_ud_iface_async_handler() to hash
[1650638548.591367] [ndv4:68766:0] async.c:506 UCX DEBUG listening to async event fd 88 events 0x5 mode thread_spinlock
[1650638548.591400] [ndv4:68766:0] ucp_worker.c:1159 UCX DEBUG created interface[9]=0x238c6e0 using ud_verbs/mlx5_ib1:1
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment