Граф коммитов

688 Коммитов

Автор SHA1 Сообщение Дата
gilbertlee-amd aa917c3fc8 [TransferBench] Adding ability to specify suffix for numBytes (#435) 2021-10-08 16:36:19 -06:00
gilbertlee-amd a6368bac99 Updating licensing / attribution for documentation (#432) 2021-10-08 13:17:24 -06:00
gilbertlee-amd e506d14d18 [TransferBench] Fixing advanced config, adding new all-1-hop sample test (#433)
* [TransferBench] Fixing advanced config, adding new all-1-hop sample test
2021-10-07 15:57:21 -06:00
Wenkai Du e0053311c0 Add another Rome model (#431) 2021-10-06 08:17:12 -07:00
Wenkai Du 29c729d8b6 Trim NICs when all GPUs are connected by XGMI (#430)
* Trim NICs when all GPUs are connected by XGMI

* Only enable clique with maximum of 2 hops
2021-10-05 18:27:43 -07:00
Wenkai Du 51a1cf428e Merge pull request #428 from ROCmSoftwarePlatform/2.10.3
Sync up with NCCL 2.10.3
2021-09-17 08:23:43 -07:00
Wenkai Du 5ae3f3f954 Remove extra L1 cache invalidate and restore __ATOMIC_SEQ_CST atomics (#426) 2021-09-14 18:30:16 -07:00
Wenkai Du 020484bf40 Use relaxed atomics and add sleep and wakeup in barrier loop (#425)
* Use relaxed atomics and add sleep and wakeup in barrier loop

* atomicAdd in ROCm 4.3 only support unsigned long long

* Switch to atomicAdd and atomicExch in more places

* Restore LOAD/STORE define to __ATOMIC_SEQ_CST

* Restore atomic for sizes FIFO
2021-09-13 17:03:49 -07:00
Wenkai Du ef432e48e1 Update tuning table (#424) 2021-09-13 08:39:01 -07:00
Wenkai Du a2421f8b4a Merge pull request #423 from wenkaidu/prim-test
rccl-prim-test: support 8p1h and 16p1h testing
2021-09-08 17:01:19 -07:00
Wenkai Du adb8d63352 Improve barrier implementation 2021-09-08 16:14:32 -05:00
Wenkai Du 31bd4236f1 Remove atomic from profiling 2021-09-08 14:20:32 -05:00
Wenkai Du 7558b5e2bf rccl-prim-test: enable 8p1h and 16p1h test 2021-09-08 11:51:26 -05:00
Wenkai Du b22d097524 Revert "rccl-prim-test: add all-to-all benchmark (#185)"
This reverts commit ebc823e603.
2021-09-07 16:41:46 -05:00
gilbertlee-amd 51d64894ff [TransferBench] ConfigFile parsing fixes, adding additional info (#422)
* [TransferBench] Adding GPU to NUMA distance detection, parsing fixes, config file generation fix

* [TransferBench] Fixing up NUMA node detection by filtering pools
2021-09-07 15:28:16 -06:00
Wenkai Du 5c8380ff5b Implement NIC identification and remapping (#420)
* Add 1H16P GPU model

* Implement NIC identification and remapping

* Revert "Sort IB devices based on device name (#413)"

This reverts commit 2d0ed8dff6.

* Fix permute and check order

* Correction on IB speed reporting

* Revert "Allow user to link layer with RCCL_IB_HCA_SKIP_LINK_LAYER (#361)"

This reverts commit caf5c9992a.
2021-08-24 09:42:04 -07:00
Wenkai Du 5f15ed6e3e Add gfx908 VM model (#418) 2021-08-10 08:55:11 -07:00
gilbertlee-amd 1ed272e5f0 [TransferBench] Removing dependency on hip_fp16 header, fixing swapped output CSV header (#416) 2021-08-04 10:53:41 -06:00
Wenkai Du 2d0ed8dff6 Sort IB devices based on device name (#413) 2021-08-03 15:32:41 -07:00
Gilbert Lee 68ec3f84e6 [TransferBench] Update to 2.10.3 2021-08-02 05:53:20 -05:00
Wenkai Du 5b72727670 Merge remote-tracking branch 'origin/develop' into 2.10.3 2021-09-15 10:33:25 -07:00
Wenkai Du 3667d308ab Merge remote-tracking branch 'origin/develop' into 2.10.3 2021-09-13 17:32:48 -07:00
Wenkai Du 8ee2b7932a Merge remote-tracking branch 'origin/develop' into 2.10.3 2021-09-13 15:51:53 -07:00
Wenkai Du 21969d7f89 Merge remote-tracking branch 'nccl/master' into 2.10.3 2021-09-01 16:34:44 -07:00
John Bachan 5f2f2f670f Fix to https://github.com/NVIDIA/nccl/issues/560
ncclGroup's containing operations of mixed datatype, element, or collective
would induce crash.
2021-08-31 15:50:05 -07:00
Gilbert Lee ae13d2a354 Modifying ReduceOrCopyMulti to accept number of preOp source to support clique-based kernels 2021-08-11 11:00:34 -05:00
Wenkai Du a4929465c5 Improve clique kernel performance by increasing unroll 2021-08-26 18:06:09 -07:00
Wenkai Du 574f0aca53 Fix typo that affects clique kernels 2021-08-26 11:10:58 -07:00
Wenkai Du 1faff323b4 Unit Test: support ncclAvg 2021-08-25 14:15:54 -07:00
Wenkai Du 60ca7484c0 Fix kernel data trace 2021-08-24 14:02:53 -07:00
Wenkai Du d5f93649ff Merge remote-tracking branch 'origin/develop' into 2.10.3 2021-08-24 09:49:47 -07:00
Wenkai Du 707c687090 Use noinline for kernel functions 2021-08-06 09:15:04 -07:00
Wenkai Du 01d3b20a66 Fix incorrect network proxy received bytes reporting 2021-08-05 17:45:48 -07:00
Wenkai Du babbd1047b Merge branch 'develop' into 2.10.3 2021-08-04 09:45:22 -07:00
Wenkai Du bf2339f93e Merge remote-tracking branch 'nccl/master' into 2.10.3 2021-07-30 16:23:14 -07:00
Wenkai Du 3e27227562 XGMI connection is always prioritized over NET regardless of hops (#412) 2021-07-29 11:12:42 -07:00
Eiden Yoshida 229ca88ee6 Add basic rtest.xml (#411) 2021-07-28 11:53:03 -06:00
Wenkai Du 818cdb16a8 Query XGMI links from xml and adjust gfx906 channel usage (#410) 2021-07-27 17:32:41 -07:00
Liam Wrubleski e579de7ec2 Update changelog with packaging information (#409) 2021-07-27 10:38:36 -06:00
Wenkai Du 135d47d125 topo_expl: fix build after switching to rocm-smi-lib (#405)
* topo_expl: fix build after switching to rocm-smi-lib

* Use minimal of 4 channels for gfx908
2021-07-27 08:30:08 -07:00
Eiden Yoshida 56801656f3 Extend test stage timeout to 24 hours (#408) 2021-07-26 15:29:21 -06:00
Liam Wrubleski 97d9cf40e7 Setup runtime and development packages (#407)
* changes to enable devel package

* Update rocm-cmake version & build
2021-07-26 15:06:17 -06:00
Wenkai Du dfc62d5fbb Skipping unnecessary functions in Doxygen by marking as internal (#353) (#406)
(cherry picked from commit 1c982d819d9c7fe0310b80f9a25808e54c71137e)

Co-authored-by: saadrahim <44449863+saadrahim@users.noreply.github.com>
2021-07-24 11:04:27 -07:00
Hubert Lu c3e2e7cb5d Merge pull request #404 from hubertlu-tw/coll_trace
Enhancement of RCCL logging information for topology-aware optimization
2021-07-22 13:50:28 -07:00
Lu bd6dbca8fb Add more info to RCCL logging for topo-aware optim. 2021-07-22 09:52:39 -07:00
Wenkai Du 215904ee8e Fix unit tests static build (#403) 2021-07-09 09:35:32 -07:00
Ke Wen 7e51592129 2.10.3-1
Add support for bfloat16.
Add ncclAvg reduction operation.
Improve performance for aggregated operations.
Improve performance for tree.
Improve network error reporting.
Add NCCL_NET parameter to force a specific network.
Add NCCL_IB_QPS_PER_CONNECTION parameter to split IB traffic onto multiple queue pairs.
Fix topology detection error in WSL2.
Fix proxy memory elements affinity (improve alltoall performance).
Fix graph search on cubemesh topologies.
Fix hang in cubemesh during NVB connections.
2021-07-08 14:30:14 -07:00
Wenkai Du 56155ff5b6 Use rocm_smi_lib for getting topology information (#402)
* Use rocm_smi_lib for getting topology information

* Add rocm-smi-lib dependency to RCCL package
2021-07-08 13:23:11 -07:00
Eiden Yoshida 5c3e7d8b67 Fix static builds (#393) 2021-06-23 09:19:48 -06:00
gilbertlee-amd 2b0b608270 [TransferBench] Fixing a typo in TransferBench usage example (#401) 2021-06-22 17:08:57 -06:00