b943523660
NCCL_TESTS_VERSION 2.18.4
...
Signed-off-by: David Addison <daddison@nvidia.com >
2026-06-05 12:45:35 -07:00
39669b6fe9
Merge remote-tracking branch 'github/master'
2026-06-05 12:36:52 -07:00
f67c258e8b
Merge branch 'codex/fix-memory-report-option' into 'master'
...
Correctly document --memory option
See merge request nccl/nccl-tests-internal!37
2026-06-05 10:30:30 -07:00
7407701f98
Document --memory option
...
Update help and README to use the existing --memory long option
instead of documenting the unsupported --memory_report spelling.
Fixes NVIDIA/nccl-tests#386
Signed-off-by: David Addison <daddison@nvidia.com >
2026-06-05 10:25:47 -07:00
632ad398cf
Merge pull request #383 from jeffreyyjp/jeffreyyjp-patch-1
...
Correct documentation to say the example shows a scanning range of up to 8GiB not 32GiB
2026-05-20 09:05:09 -07:00
ba9026a4d6
Correct scanning range from 32GiB to 8GiB
...
Updated the scanning range in the README for clarity.
2026-05-20 22:08:49 +08:00
b7a579c0bb
Merge remote-tracking branch 'github/master'
...
v2.18.3
2026-04-13 13:39:18 -07:00
f727aa2a54
NCCL_TESTS_VERSION 2.18.3
...
Signed-off-by: David Addison <daddison@nvidia.com >
2026-04-13 13:37:21 -07:00
5dc0670e56
Add maxP2pPeers comm config for sendrecv
...
Add optional testEngine.initCommConfig, invoked from initComms
after the shared ncclConfig_t setup.
sendrecv registers SendRecvInitCommConfig to set maxP2pPeers=2
Signed-off-by: David Addison <daddison@nvidia.com >
2026-04-13 11:19:09 -07:00
5c07824337
Merge branch 'daddison/maxP2pPeers-sendrecv-config' into 'master'
...
Add maxP2pPeers comm config for sendrecv
See merge request nccl/nccl-tests-internal!34
2026-04-13 11:17:51 -07:00
e87919a603
Add maxP2pPeers comm config for sendrecv
...
Add optional testEngine.initCommConfig, invoked from initComms
after the shared ncclConfig_t setup.
sendrecv registers SendRecvInitCommConfig to set maxP2pPeers=2
Signed-off-by: David Addison <daddison@nvidia.com >
2026-03-30 18:45:08 -07:00
4f95d4ae1b
Merge remote-tracking branch 'github/master'
2026-03-11 15:45:30 -07:00
af1dcac92a
NCCL_TESTS_VERSION 2.18.2
2026-03-11 15:35:13 -07:00
eb0d3d2a00
Display unalign setting in output
2026-03-11 15:05:54 -07:00
e02c20b898
NCCL_TESTS_VERSION 2.18.1
2026-03-11 09:55:31 -07:00
c1af7df1f3
Update -z option description in README.md
2026-03-11 09:54:53 -07:00
ba52a70492
Allow blocking collectives without MPI_Barrier in timing loop
2026-03-11 09:36:53 -07:00
8d26b23319
Allocate buffers during thread initialization
...
Signed-off-by: Theofilos Ioannis Manitaras <tmanitaras@nvidia.com >
2026-03-11 09:36:38 -07:00
a4879fca93
Merge branch 'daddison/blocking-coll-no-barrier' into 'master'
...
Allow blocking collectives without MPI_Barrier in timing loop
See merge request nccl/nccl-tests-internal!28
2026-03-11 09:26:08 -07:00
c78b8bf693
Allow blocking collectives without MPI_Barrier in timing loop
2026-03-11 09:26:08 -07:00
abde154277
Merge branch 'tmanitaras/fix_buffer_allocation_parallel_init' into 'master'
...
Allocate buffers during thread initialization
See merge request nccl/nccl-tests-internal!31
2026-03-11 09:17:18 -07:00
265873726b
Allocate buffers during thread initialization
...
Signed-off-by: Theofilos Ioannis Manitaras <tmanitaras@nvidia.com >
2026-03-10 18:33:50 +01:00
97bb1ad3ad
Merge remote-tracking branch 'github/master'
...
NCCL_TESTS_VERSION 2.18.0
2026-03-09 13:17:40 -07:00
dd0bafd178
NCCL_TESTS_VERSION 2.18.0
2026-03-06 17:55:12 -08:00
115fb09377
Add new unalign flag to README.md and update help text
2026-03-06 17:53:29 -08:00
e986a6156c
Add -u <index> to force unaligned buffer addresses
2026-03-06 17:39:25 -08:00
42b9e47067
Merge remote-tracking branch 'github/master'
...
NCCL_TESTS_VERSION 2.17.10
2026-03-05 16:24:48 -08:00
c379e19a71
NCCL_TESTS_VERSION 2.17.10
2026-03-05 15:35:13 -08:00
db221defdb
Request GIN to be explicitly enabled in all to all test
...
Based on the changes in NCCL v2.29.3, update the alltoall test to
either provide a ginConnectionType or set ginForceEnable to true.
Signed-off-by: Ahsan Pervaiz <apervaiz@nvidia.com >
2026-03-05 15:34:19 -08:00
cd9cad3de4
Fix syntax error in the GIN fix MR
2026-03-05 15:29:29 -08:00
12860f2680
Merge branch 'apervaiz/enable-gin-explicitly' into 'master'
...
Request GIN to be explicitly enabled in all to all test
See merge request nccl/nccl-tests-internal!30
2026-03-05 14:36:49 -08:00
26be02998a
Request GIN to be explicitly enabled in all to all test
...
Based on the changes in NCCL v2.29.3, update the alltoall test to
either provide a ginConnectionType or set ginForceEnable to true.
Signed-off-by: Ahsan Pervaiz <apervaiz@nvidia.com >
2026-03-03 14:29:00 -08:00
ae98985f55
Fix Clang compilation errors with VLA initialization
...
Signed-off-by: David Addison <daddison@nvidia.com >
2026-02-09 10:38:44 -08:00
cfa8a9674b
Merge remote-tracking branch 'github/master'
2026-02-05 11:53:39 -08:00
9938d5a657
Fix compilation issues with latest NCCL release headers
...
Add --extended-lambda to NVCUFLAGS
2026-02-04 16:43:20 -08:00
2535da805b
NCCL_TESTS_VERSION 2.17.9
2026-02-03 11:04:48 -08:00
85ca91d1b1
Fix: corrected typos in the JSON output
...
Signed-off-by: David Addison <daddison@nvidia.com >
2026-02-03 11:03:35 -08:00
88d7e33207
Add -M memory report option to README.md
2026-01-15 13:32:55 -08:00
36a853984f
Merge branch 'apervaiz/v2.29-perftest-p2p-connectivity-validation' into 'master'
...
Skip tests that require LSA but are executed on multiple nodes
See merge request nccl/nccl-tests-internal!25
2026-01-14 10:51:47 -08:00
9590b4f074
Merge branch 'shsnyder/dev-api-readme' into 'master'
...
update README to account for dev API -D,-V options
See merge request nccl/nccl-tests-internal!26
2026-01-14 06:38:18 -08:00
475f8805f9
Error out tests that require LSA but are executed on multiple nodes
...
Compare the total number of ranks in the COMM and the number of ranks
in the LSA team to determine if all ranks are LSAable. If they are
not, report an invalid usage error on tests that require P2P
connectivity and print an error message.
Signed-off-by: Ahsan Pervaiz <apervaiz@nvidia.com >
2026-01-13 14:29:16 -08:00
67d7367734
update README to account for dev API -D,-V options
2026-01-13 13:45:02 -06:00
056136d232
Merge branch 'missing-devcommdestroy' into 'master'
...
add missing call to ncclDevCommDestroy
See merge request nccl/nccl-tests-internal!22
2026-01-13 11:17:51 -08:00
1d317d53ba
Pass COMM to getDevCommRequirements function
...
Pass the target ncclComm_t to the getDevCommmRequirements function and
use that to retrieve properties.
This change is required to add a condition in getDevCommRequirements
that validates whether a test can be run on a cluster given its configuration.
Signed-off-by: Ahsan Pervaiz <apervaiz@nvidia.com >
2026-01-12 16:32:14 -08:00
ca7625d41a
Merge branch 'v2.29_check-devapi-support' into 'master'
...
check deviceAPI support
See merge request nccl/nccl-tests-internal!23
2026-01-08 11:49:16 -08:00
cd773dbe2f
Merge remote-tracking branch 'github/master'
...
NCCL_TESTS_VERSION 2.17.8
2026-01-06 15:11:27 -08:00
81463c58d0
NCCL_TESTS_VERSION 2.17.8
2026-01-06 15:00:17 -08:00
7278698c1b
Clarified use of Mebibytes and Gibibytes for sizes
2026-01-06 14:59:17 -08:00
064cbede79
NCCL_TESTS_VERSION 2.17.7
2025-12-31 02:18:39 +00:00
2656c58421
NCCL_TESTS_VERSION 2.17.7
2025-12-30 20:18:25 +00:00
David Addison