AI/rocm-systems
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
1,465 Commits 3 Branches 0 Tags
9fdecceefbece5cf97ea495b3f7148ee57865ddb
Commit Graph

166 Commits

Author SHA1 Message Date
Nilesh M Negi 35f4a405f0 [BUILD] Update gfxTargets for ASAN build (#1242) 
Signed-off-by: nileshnegi <Nilesh.Negi@amd.com>

[ROCm/rccl commit: 4f31ab85ea]
 2024-08-06 10:53:51 -05:00
Nilesh M Negi 713ed3341d [BUILD] Disable MSCCLPP build by default (#1283) 
Signed-off-by: nileshnegi <Nilesh.Negi@amd.com>

[ROCm/rccl commit: cb2e0615d7]
 2024-08-02 23:17:51 -05:00
Wenkai Du 27b7998d13 Restore number of parallel linking jobs (#1278) 
* Restore number of parallel linking jobs

* Dynamically adjust number of linker jobs with limit of 16 jobs max

* Fix typo

* Add cgroup v1 support

[ROCm/rccl commit: ca5341d419]
 2024-07-30 08:04:14 -07:00
corey-derochie-amd f2b2372056 Only initialize MSCCL++ when runtime-enabled. (#1266) 
[ROCm/rccl commit: b31b4082dd]
 2024-07-22 00:41:31 -06:00
Wenkai Du 54e4899607 Template unroll for RCCL kernels (#1250) 
* Template unroll for RCCL kernels

* Adding unroll template arg during CMake hipification

* Reduce linking parallel jobs to avoid OOM in CI

* Workaround issues with UT tests

SWDEV-469533: register spill fix is needed for mainline build
LWPCOMMLIBS-369: cannot enable 112 channels with 80 CUs
Use -parallel-jobs=8 for linking

* CI: do not use -j 16 when building

* CI: use -j 8 when building

* Only reduce parallel linking job for CI extended

* Restore original jenkins command. Change parallel linking jobs in cmake

* Disable MSCCLPP

---------

Co-authored-by: gilbertlee-amd <gilbert.lee@amd.com>

[ROCm/rccl commit: 89349f2ce4]
 2024-07-19 08:15:59 -07:00
corey-derochie-amd b8542c2477 Integrated RCCL with MSCCL++ for small message sizes (#1231) 
[ROCm/rccl commit: 6dc47eecd7]
 2024-07-12 15:32:58 -06:00
Rahul Vaidya f60367f1c3 Improved version reporting in NCCL_DEBUG=VERSION (#1232) 
* Improved version reporting in NCCL_DEBUG=VERSION.

Signed-off-by: rahulvaidya20 <ravaidya@amd.com>

* Version reporting changes

Signed-off-by: rahulvaidya20 <ravaidya@amd.com>

* Versioning changes: Initialized char arrays to null and fixed typo.

---------

Signed-off-by: rahulvaidya20 <ravaidya@amd.com>

[ROCm/rccl commit: c755b9cf93]
 2024-07-12 08:14:29 -05:00
akolliasAMD 942555fd21 gfx12 initial enablement (#1219) 
[ROCm/rccl commit: 63e4d76e23]
 2024-07-10 13:32:09 -06:00
akolliasAMD 1934d0f377 fixed typo on BFD linkage (#1192) 
[ROCm/rccl commit: 6475da2ed9]
 2024-06-03 10:05:47 -06:00
Nilesh M Negi 7ca67f1cb9 [BUILD] Update install.sh for RCCL build (#1191) 
Signed-off-by: nileshnegi <Nilesh.Negi@amd.com>

[ROCm/rccl commit: 5aaf7121d9]
 2024-05-31 17:58:34 -05:00
Edgar Gabriel c31ac0d17b add alternative to rocm_smi_lib 
[ROCm/rccl commit: 9ad913bfa8]
 2024-05-14 13:51:41 -07:00
Wenkai Du dd5d66d01a Use rocm-smi thread only mutex when available (#1169) 
[ROCm/rccl commit: a64aab5f63]
 2024-05-08 14:32:24 -07:00
Wenkai Du 63d8de3bd3 Add compiler warning for uninitialized variable and fix (#1163) 
* Add compiler warning for uninitialized variable and fix

* Add -Wsometimes-uninitialized

* Convert warning to error

[ROCm/rccl commit: b18784d8b8]
 2024-05-08 07:00:25 -07:00
BertanDogancay 36f9492cda Merge remote-tracking branch 'nccl/master' into develop 
[ROCm/rccl commit: e1a835910e]
 2024-04-23 13:34:00 -07:00
Bertan Dogancay 8ddb74e3b1 Add unique files to source list (#1144) 
[ROCm/rccl commit: 3caad91f32]
 2024-04-15 09:46:53 -06:00
mberenjk da835cff9c replacing rccl_bfloat16 with hip_bfloat16 (#1126) 
Co-authored-by: mberenjk <mberenjk@amd.com>

[ROCm/rccl commit: 428837ffe4]
 2024-04-11 11:30:37 -05:00
arvindcheru 2c0284885a Update Depends with correct HIP Runtime package name (#1130) 
[ROCm/rccl commit: c1b8eab8e1]
 2024-04-09 19:27:07 -04:00
arvindcheru a285fda3a1 Static Build update - Moved all cmake install() to rocm-cmake APIs, static build update (#1123) 
[ROCm/rccl commit: c0a51dc84b]
 2024-03-26 11:11:09 -04:00
corey-derochie-amd 62a6a07d49 Replaced ROCmSoftwarePlatform and RadeonOpenCompute links with ROCm links. (#1125) 
[ROCm/rccl commit: 503a472a25]
 2024-03-25 16:29:13 -06:00
Wenkai Du 43bbee4dcc Remove hipEventDisableSystemFence (#1122) 
There is no indication that disabling system fence has any latency improvement.
Removing it per recommendation from HIP.

[ROCm/rccl commit: 5976f757dd]
 2024-03-25 08:01:57 -07:00
Nilesh M Negi f93831cf6a BUILD: Enable RCCL static build (#1114) 
Signed-off-by: nileshnegi <Nilesh.Negi@amd.com>

[ROCm/rccl commit: 53fad75001]
 2024-03-15 12:18:18 -05:00
Andy li e373bd44bf Enable fp8 support (#1101) 
* initial checkin

* resolve cr comments

* resolve the build issue

* fix the data correctless issue

* update fp8 header file and update the unit test for fp8 support

* remove fp16 from fp8 headers

* fix ut issue and catch up the latest code from develop

* udate according to cr comments

* update ut according to cr comments

* update num floats for each SumPostDiv from 4 to 6

* update fp8 header file name

* fix the typo

[ROCm/rccl commit: 6777e65c1d]
 2024-03-08 15:17:53 -08:00
Wenkai Du e5aedb153e Add support for using contiguous for GPU direct RDMA (#1096) 
Enabled by env var RCCL_NET_CONTIGUOUS_MEM=1

[ROCm/rccl commit: cbd955627e]
 2024-02-29 10:06:43 -08:00
Bertan Dogancay cee279fd99 Implement ROCTX (#1094) 
* Implement roctx

[ROCm/rccl commit: b617aecc31]
 2024-02-27 15:46:15 -07:00
BertanDogancay 7842411fb3 Merge remote-tracking branch 'rccl/develop' into 2.19.4 
[ROCm/rccl commit: 76f83f95ab]
 2024-02-15 13:37:14 -08:00
Bertan Dogancay bee47d9e91 Add stack size UT (#1081) 
* Add stack size UT

[ROCm/rccl commit: dc2d486ba0]
 2024-02-12 17:56:15 -07:00
Wenkai Du d5f5091e5d Merge remote-tracking branch 'rccl/develop' into 2.19.4 
[ROCm/rccl commit: d999d9ad21]
 2024-02-09 11:31:03 -06:00
Bertan Dogancay 45ed3ef4e7 Nvtx support (#1076) 
* NVTX support

[ROCm/rccl commit: 8a442faa12]
 2024-02-08 14:08:24 -07:00
Wenkai Du e319d0a49d Merge remote-tracking branch 'rccl/develop' into HEAD 
[ROCm/rccl commit: e64324a64a]
 2024-02-01 12:17:09 -06:00
Nilesh M Negi f23716de80 Enable kernarg preloading for ROCm 6.1 (#1068) 
Signed-off-by: nileshnegi <Nilesh.Negi@amd.com>

[ROCm/rccl commit: 2458f158b1]
 2024-02-01 12:14:04 -06:00
Wenkai Du 99564b560c Remove enhcompat.cc 
[ROCm/rccl commit: 6afabf0d0b]
 2024-01-24 17:13:30 -08:00
BertanDogancay 404d398bac Merge remote-tracking branch 'nccl/v2.19' into develop 
[ROCm/rccl commit: 81ddf9de89]
 2024-01-24 15:25:33 -08:00
Wenkai Du 8b8179a689 Use new HIP graph API compatible with CUDA 11030 (#991) 
* Use new HIP graph API compatible with CUDA 11030

* Update dependency to ROCm 6.1

* Fix single stream use case

[ROCm/rccl commit: 7e25d5bc55]
 2024-01-21 19:00:50 -08:00
Bertan Dogancay 4be87126fe Turn IFC off (#1053) 
[ROCm/rccl commit: 5f365a9957]
 2024-01-18 15:29:36 -07:00
Bertan Dogancay 11674674fc [DEV] Configure functions in RCCL (#986) 
* configure functions in rccl

[ROCm/rccl commit: 28d9b170c9]
 2024-01-18 15:07:16 -07:00
Wenkai Du 64cf812da0 Re-enable L128 on gfx90a of compiler supports it (#1036) 
[ROCm/rccl commit: 5851ae5974]
 2024-01-10 08:01:11 -08:00
Nilesh M Negi c1acf97c05 Remove FORCE from AMDGPU_TARGETS and add support in install script (#989) 
Signed-off-by: nileshnegi <Nilesh.Negi@amd.com>

[ROCm/rccl commit: 414884c6cb]
 2024-01-09 13:29:47 -06:00
akolliasAMD 9cb0f98e73 CMake does not allow for capital letters been used in package names (#1020) 
[ROCm/rccl commit: a924454f0f]
 2023-12-15 12:39:17 -07:00
Bertan Dogancay fe5a902f97 correct package name (#1012) 
[ROCm/rccl commit: fca459baaf]
 2023-12-11 09:40:29 -07:00
Bertan Dogancay ae0bdad45c IFC mix build (#998) 
[ROCm/rccl commit: 7c0f49a878]
 2023-12-02 18:49:52 -07:00
Bertan Dogancay 5efe13655d Renaming unit-tests package (#987) 
[ROCm/rccl commit: 20b02af19b]
 2023-11-29 15:05:32 -07:00
Bertan Dogancay 9e8eb41337 Check to support older ROCm versions (#963) 
[ROCm/rccl commit: 198f14923b]
 2023-11-15 12:36:31 -07:00
Bertan Dogancay 1eef637d8c Revert "Remove hip::device (#954)" (#956) 
This reverts commit 2d04f5390d.

[ROCm/rccl commit: 8e0258a73d]
 2023-11-07 13:40:41 -07:00
Bertan Dogancay 2d04f5390d Remove hip::device (#954) 
[ROCm/rccl commit: 7edb486154]
 2023-11-03 19:31:00 -06:00
Nilesh M Negi 706750597c remove gcnArch support (#920) 
Signed-off-by: nileshnegi <Nilesh.Negi@amd.com>

[ROCm/rccl commit: f22df90e5c]
 2023-10-26 12:09:15 -05:00
searlmc1 212453b2fb Remove quotes causing asan build breakage 
The quotes around "-fsanitize=address -shared-libasan" cause the BUILD_ADDRESS_SANITIZER build to fail; remove the quotes

[ROCm/rccl commit: f59de10524]
 2023-10-19 16:13:39 -07:00
Wen-Heng (Jack) Chung dfa0d98f9e Change MSCCL kernel signature to allow kernel arguments be preloaded via SGPR (#911) 
* Adding a script that will download/compile/run TransferBench/RCCL/UCX/RCCL-tests/RCCL-Unittests/hip-mpi-testsuite (#895)

Co-authored-by: Pedram Alizadeh <pmohamma@banff-pla-r27-05.pla.dcgpu>

* Only build gfx941

* demo

* fine tune malloc

* Fix merge errors

* Fix merge errors

* Disable parallel build

* Adopt --amdgpu-kernarg-preload-count

* Revert "Adding a script that will download/compile/run TransferBench/RCCL/UCX/RCCL-tests/RCCL-Unittests/hip-mpi-testsuite (#895)"

This reverts commit f5e252dddf02a41b4d1bc512f306f45f97166304.

* Revert CMake changes.

* NPKIT changes.

* Remove some license declarations.

* Address code review feedbacks on msccl_kernel_impl.h

* Update CMakeLists.txt

* Add CMake logic to check the existence of --amdgpu-kernarg-preload-count

* Fix NPKIT trace logic.

---------

Co-authored-by: Pedram Alizadeh <pmohamma@amd.com>
Co-authored-by: Pedram Alizadeh <pmohamma@banff-pla-r27-05.pla.dcgpu>
Co-authored-by: Ziyue Yang <ziyyang@microsoft.com>

[ROCm/rccl commit: 7ee5c1c28b]
 2023-10-12 20:17:08 -05:00
Edgar Gabriel e6c3e9fd8e turn bfd compilation off by default 
revert the logic to ensure that we are not accidentally creating
a dependency on the bfd libraries when deploying rccl binaries.


[ROCm/rccl commit: 88a55cef83]
 2023-09-29 20:25:33 +00:00
Cen Zhao d3c20a1210 Update install.sh to take "--static" option (#894) 
* Update install.sh to take "--static" option

* Fix static build errors

---------

Co-authored-by: BertanDogancay <bertan.dogancay@gmail.com>

[ROCm/rccl commit: fb57a438d7]
 2023-09-27 12:45:21 -04:00
Audrey MP 2e3d45a53a Gcn arch name (#886) 
We use CMake to determine if we're compiling against a version of ROCm that supports gcnArchName and handles architecture checking appropriately. It includes a few helper functions as drop ins for the functionality we used gcnArch for before; sometimes to enable flags, and sometimes to set frequencies.

[ROCm/rccl commit: e58ec78d35]
 2023-09-12 15:34:40 -04:00