AI/rocm-systems
2
0
派生 0
代码 工单 合并请求 工作流 软件包 项目 发布 百科 活动
1,581 提交 3 分支 0 Git 标签
2b9394f08aef38e9cb682048d2bfcd727129cf60
提交图

1464 次代码提交

作者 SHA1 备注 提交日期
qiwei_ji 2b9394f08a Check nvlink_node instead of xgmi_node in xml.cc (#1407) 
It seems like here wants to check xgmi_node instead. If checks node for "nvlink", it will verify the link_info everytime.
If checks node for "xgmi", when get yes answer, it won't need check vsmi topo interface.

[ROCm/rccl commit: f2ee8d9132]
 2025-01-06 17:09:27 -08:00
Xeonacid f3e883805c Define wc_store_fence for riscv (#1475) 
[ROCm/rccl commit: c6c7b6db98]
 2025-01-06 16:59:14 -08:00
dependabot[bot] e61194239d Bump jinja2 from 3.1.4 to 3.1.5 in /docs/sphinx (#1473) 
Bumps [jinja2](https://github.com/pallets/jinja) from 3.1.4 to 3.1.5.
- [Release notes](https://github.com/pallets/jinja/releases)
- [Changelog](https://github.com/pallets/jinja/blob/main/CHANGES.rst)
- [Commits](https://github.com/pallets/jinja/compare/3.1.4...3.1.5)

---
updated-dependencies:
- dependency-name: jinja2
  dependency-type: indirect
...

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

[ROCm/rccl commit: e77c3c2975]
 2025-01-06 17:09:32 -07:00
corey-derochie-amd 17becdb7f8 [SWDEV-497665] Blocked cudaMemcpyAsync race condition by synchronizing (#1447) 
* Switched calls to `cudaMemcpyAsync` to be `cudaMemcpy` in `ncclTransportP2pSetup` to avoid race condition with `cudaIpcOpenMemHandle` inside p2p `connect`. See `ncclP2pImportShareableBuffer`.

* Moved synchronize outside of the loop, as it isn't necessary to sync between every iteration of the loop.

[ROCm/rccl commit: c158d3a9b4]
 2025-01-03 13:06:47 -07:00
Mustafa Abduljabbar dd3fc22531 Add macro for channel mask offset (#1467) 
[ROCm/rccl commit: a9d6e7661c]
 2025-01-03 08:41:41 -05:00
mberenjk 300f954185 Initializing all ranks to the same value to avoid failure of UT AllR… (#1459) 
* Initializing all ranks to the same value to avoid failure of  UT AllReduce for FP8 type

Co-authored-by: Marzieh Berenjkoub <mberenjk@amd.com>

[ROCm/rccl commit: 39483c55f8]
 2025-01-02 11:39:02 -06:00
Nilesh M Negi f1bada26ef [BUILD] Fix ASAN build if GPU targets has xnack+ (#1474) 
Signed-off-by: nileshnegi <Nilesh.Negi@amd.com>

[ROCm/rccl commit: fd03b5b6a5]
 2024-12-26 12:13:36 -06:00
dependabot[bot] beece84733 Bump rocm-docs-core from 1.9.2 to 1.12.0 in /docs/sphinx (#1466) 
Bumps [rocm-docs-core](https://github.com/ROCm/rocm-docs-core) from 1.9.2 to 1.12.0.
- [Release notes](https://github.com/ROCm/rocm-docs-core/releases)
- [Changelog](https://github.com/ROCm/rocm-docs-core/blob/develop/CHANGELOG.md)
- [Commits](https://github.com/ROCm/rocm-docs-core/compare/v1.9.2...v1.12.0)

---
updated-dependencies:
- dependency-name: rocm-docs-core
  dependency-type: direct:production
  update-type: version-update:semver-minor
...

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

[ROCm/rccl commit: 648a58dd27]
 2024-12-18 10:04:07 -07:00
Jeffrey Novotny 7c220660ba Change kernel reference to use new terminology (#1462) 
[ROCm/rccl commit: 2934bf6fc6]
 2024-12-16 13:34:18 -05:00
Mustafa Abduljabbar 4588da1888 Remove unneeded highestTransportType (#1461) 
[ROCm/rccl commit: e6b179d627]
 2024-12-16 13:28:47 -05:00
Ziyue Yang 987dfc3b5d Fix MSCCL algorithm loading order (#1460) 
[ROCm/rccl commit: 83c5eb7378]
 2024-12-16 07:41:17 -08:00
akolliasAMD c65d4ab18f changed the CMake option from AMDGPU_TARGETS to GPU_TARGETS (#1440) 
[ROCm/rccl commit: 45c1c1a781]
 2024-12-12 12:09:30 -07:00
Shilei Tian 791f45733a Improve the handling of CMake deduplication (#1450) 
Certain CMake functions deduplicates arguments by default. For example, if we
have two `target_link_options` with both `-Xoffload-linker -opt-A` and then
`-Xoffload-linker -opt-B`, the final link command would be `-Xoffload-linker
-opt-A -opt-B`, which is not what we want.

[ROCm/rccl commit: 7386fac64a]
 2024-12-11 13:48:18 -08:00
Shilei Tian 699980206e Check -parallel-jobs before use (#1451) 
`-parallel-jobs` is not always available, such as upstream LLVM.

[ROCm/rccl commit: 8e9fcf111a]
 2024-12-11 11:40:49 -06:00
Hujingbo bea86f6248 increase p2p channels for Intel platform (#1448) 
Co-authored-by: hujingbo <hujingbo@kuaishou.com>

[ROCm/rccl commit: ad4c36dc34]
 2024-12-10 07:33:37 -08:00
Jeffrey Novotny d7498b88a5 Refactor how to docs and formatting fixes (#1444) 
[ROCm/rccl commit: 9aa5b9f02e]
 2024-12-10 08:47:24 -05:00
Jeffrey Novotny 531476dacf Add RCCL debugging guide (#1420) 
* Add RCCL debugging guide

* Changes from external review

* More edits from internal review

* Additional edits

* Minor correction

* More changes after external review

* Integrate index and ToC changes with incoming merge changes

* Integrate feedback from management review

* Minor edits from the internal review

[ROCm/rccl commit: 6d34fb7632]
 2024-12-06 13:25:58 -05:00
Nusrat Islam b19a809788 ext-src: tune TP=8 case on MI308 CPX mode (#1446) 
Tune the number of blocks for hierarchical mscclpp allreduce.

[ROCm/rccl commit: 42b6831a39]
 2024-12-06 08:16:39 -06:00
Benjamin Kitor fe806d5427 Add Topologies for 16-GPU gfx942 SuperNode (#1417) 
* Add Topologies for 16-GPU gfx942 SuperNode

- Add GigaIO topologies to tools/topo_expl for dev and testing
- Add GigaIO Columba 16 GPU romeModel and adjust topology
  matching algorithm in rome_models for 16 GPU system
- Fix bug which failed to match Rome Model when using subsets
  of system resources (i.e. ROCR_VISIBLE_DEVICES is set)
- Fixes for topo_expl

* Fix bug w/ 1H16P

[ROCm/rccl commit: a05329bd0d]
 2024-12-03 13:12:03 -08:00
Jeffrey Novotny 0f4558ea59 Modify cmake instruction in build from source (#1445) 
[ROCm/rccl commit: 28594b26b3]
 2024-12-03 11:26:02 -05:00
dependabot[bot] 43ff8386a8 Bump rocm-docs-core from 1.8.3 to 1.9.2 in /docs/sphinx (#1441) 
Bumps [rocm-docs-core](https://github.com/ROCm/rocm-docs-core) from 1.8.3 to 1.9.2.
- [Release notes](https://github.com/ROCm/rocm-docs-core/releases)
- [Changelog](https://github.com/ROCm/rocm-docs-core/blob/develop/CHANGELOG.md)
- [Commits](https://github.com/ROCm/rocm-docs-core/compare/v1.8.3...v1.9.2)

---
updated-dependencies:
- dependency-name: rocm-docs-core
  dependency-type: direct:production
  update-type: version-update:semver-minor
...

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

[ROCm/rccl commit: 1f789d6836]
 2024-11-29 15:21:43 -07:00
Jeffrey Novotny 1d1e17b3c9 Refactor RCCL install guide into several pages (#1427) 
* Refactor RCCL install guide into several pages

* Changes from code review and new docker guide

* Add missing entries to ToC

* Minor fixes

* Fix help strings

* Edits after review and remove extra white space

[ROCm/rccl commit: bf7c130631]
 2024-11-27 15:34:26 -05:00
Jeffrey Novotny cc9209f770 Update rccl changelog for 6.3.1 (#1433) 
* Update rccl changelog for 6.3.1

* Fix version number

* Correct RCCL release version

* Added details to 6.3.0 changelog

---------

Co-authored-by: corey-derochie-amd <161367113+corey-derochie-amd@users.noreply.github.com>

[ROCm/rccl commit: e42f10a361]
 2024-11-26 08:46:37 -05:00
gilbertlee-amd be0865f335 Adding RCCL_MODEL_REVERSAL_DISABLE env var to disable model reversal (#1431) 
* Adding RCCL_MODEL_REVERSAL_DISABLE env var to disable model reversal

[ROCm/rccl commit: 000575867c]
 2024-11-25 11:24:54 -07:00
Bertan Dogancay df51937f09 Fix typo in ncclGetKernelIndex macro (#1424) 
[ROCm/rccl commit: dfe4a3ed81]
 2024-11-18 10:40:05 -05:00
corey-derochie-amd 6d61a4e21f Added latest users to CODEOWNERS. (#1422) 
[ROCm/rccl commit: 4336a0f3a3]
 2024-11-14 16:55:18 -07:00
Bertan Dogancay 9041572e23 Template generic kernel for unroll factor (#1419) 
* Template generic kernel for unroll factor

[ROCm/rccl commit: cb175fb0b3]
 2024-11-12 18:27:29 -05:00
Jeffrey Novotny 9898395fbe Refactor landing page and move some info to What is RCCL (#1415) 
[ROCm/rccl commit: 2d07f18696]
 2024-11-12 13:15:27 -05:00
akolliasAMD 891689aab9 removing unused gfx targets (#1411) 
[ROCm/rccl commit: 2284101624]
 2024-11-06 08:50:08 -07:00
darren-amd 539cc81748 Merge pull request #1406 from ROCm/darren-amd-remove-computeColl-declaration 
remove undefined computeColl declaration

[ROCm/rccl commit: 52d5f4cde2]
 2024-11-06 10:43:35 -05:00
gilbertlee-amd 71809e8bc0 Updating RCCL Replayer README (#1408) 
[ROCm/rccl commit: cb1027de97]
 2024-11-05 08:06:11 -07:00
darren-amd 61258b9a9f remove undefined computeColl declaration 
[ROCm/rccl commit: ebf0417e90]
 2024-11-04 13:42:01 -05:00
saurabhAMD 69d976532b GPU allocation for CPX Unit Tests using PCI bus id (#1403) 
* mapping devices wrt pci

* Gpu allocation by using pci mapping

* Passing gpuPriorityOrder in as an argument rather than making the functions non-static.

* Removing redundant testBed instance calling

[ROCm/rccl commit: 69b2b712ab]
 2024-11-04 10:51:00 -06:00
corey-derochie-amd ad1384bea1 Hide or fix all build warnings (#1331) 
* Changing C-strings to be const.

* Changed variable-length arrays to std::vector to avoid warnings. VLA is a compiler extension.

* Changed `#define` inside functions into `constexpr int` to preserve scoping and avoid macro redefinition warnings.

* Disabled warnings for modifying `CMAKE_CXX_FLAGS` caused by `check_symbol_exists`, which temporarily modifies the flag to do a compile check.

* Fixed VLA in rccl UT.

[ROCm/rccl commit: 1c45962273]
 2024-11-04 09:46:42 -07:00
Abhishek Kulkarni 595cda2ab9 GDR enablement logic fix for kernel 6.4.0+ (#1378) 
[ROCm/rccl commit: 6178556853]
 2024-11-03 01:20:07 -05:00
Bertan Dogancay 251df02d42 Increase MAX_STACK_SIZE for UT (#1398) 
[ROCm/rccl commit: 984f1e4343]
 2024-11-01 13:07:45 -04:00
corey-derochie-amd 8444e5fe7f Set minimum ROCm version for MSCCLPP to 6.2 (#1401) 
* Added ROCm version check around setting `ENABLE_MSCCLPP` flag.

[ROCm/rccl commit: 6db2644766]
 2024-10-30 16:48:54 -06:00
Avinash da3887bafb Memory leak fixes in hostside functions (#1388) 
memory leak fixes for parseRome4P2H and ncclTopoAddGPU

[ROCm/rccl commit: d6006f0425]
 2024-10-30 14:25:56 -05:00
Tim e346e19065 Adjustment for UT Sendrecv (#1400) 
Enabled UT sendrecv to same rank and refactor UBR call

[ROCm/rccl commit: fd9924cfe7]
 2024-10-30 15:13:53 -04:00
Nusrat Islam e1c20e7f24 ext-src: Improved allreduce performance in cpx mode for MI308 (#1393) 
To get the improved performance for TP=4, the user needs to use
RCCL_MSCCL_FORCE_ENABLE=1 and MSCCLPP_READ_ALLRED=1. For TP=8, the
user should use MSCCLPP_HIERARCHICAL_ALLRED=1.

[ROCm/rccl commit: 0fb3b5eba9]
 2024-10-30 08:30:15 -05:00
corey-derochie-amd af1e36a7ee Remove MSCCL switch case fall-through by adding break statement. (#1342) 
[ROCm/rccl commit: ea20af698e]
 2024-10-29 15:47:59 -06:00
corey-derochie-amd f9d38d8858 6.2 final documentation fixes updated for 6.3 (#1252) (#1399) 
* Update CHANGELOG.md

* Update NOTICES.txt

* [DOCS] Note on using less than 8 MI300 GPUs



* Update README.md



---------

Signed-off-by: nileshnegi <Nilesh.Negi@amd.com>
Co-authored-by: nileshnegi <Nilesh.Negi@amd.com>

[ROCm/rccl commit: 8ac63e7e70]
 2024-10-29 15:23:45 -06:00
gilbertlee-amd 02bf3a3bf8 Adding support for odd nodes for model_87 (#1309) 
[ROCm/rccl commit: 0cbce2a757]
 2024-10-24 08:38:12 -06:00
corey-derochie-amd 1c700083b2 Update CHANGELOG to match release branches 6.2 and 6.3 (#1391) 
* [CHANGELOG] Add Known issues for ROCm 6.2.1

Signed-off-by: nileshnegi <Nilesh.Negi@amd.com>

* Updated 6.2.1 known issues to match the content in develop.

* Updated CHANGELOG for ROCm 6.3 release. (#1380)

* Updated CHANGELOG for ROCm 6.3 release.

* Update CHANGELOG to new format.

Co-authored-by: Jeffrey Novotny <jnovotny@amd.com>

---------

Co-authored-by: Jeffrey Novotny <jnovotny@amd.com>

---------

Signed-off-by: nileshnegi <Nilesh.Negi@amd.com>
Co-authored-by: nileshnegi <Nilesh.Negi@amd.com>
Co-authored-by: Jeffrey Novotny <jnovotny@amd.com>

[ROCm/rccl commit: 6ed513e1b9]
 2024-10-23 13:49:40 -06:00
Arm Patinyasakdikul 928414ac06 Increased maximum number of XML nodes to support CPX mode. (#1386) 
[ROCm/rccl commit: 29f87c7191]
 2024-10-23 11:15:11 -05:00
Wenkai Du 075381ee2e Fix topology discovery in container with subset of GPUs (#1384) 
* Fix topology discovery in container with subset of GPUs

* Move links counting out of loop

[ROCm/rccl commit: e0780ba4d4]
 2024-10-22 13:50:23 -07:00
Bertan Dogancay fcb0b2da3f [Replayer] Add validation (#1387) 
* Add validation to rccl_replayer

[ROCm/rccl commit: cfecce790f]
 2024-10-22 10:41:08 -04:00
dependabot[bot] 64aead445c Bump rocm-docs-core from 1.8.2 to 1.8.3 in /docs/sphinx (#1385) 
Bumps [rocm-docs-core](https://github.com/ROCm/rocm-docs-core) from 1.8.2 to 1.8.3.
- [Release notes](https://github.com/ROCm/rocm-docs-core/releases)
- [Changelog](https://github.com/ROCm/rocm-docs-core/blob/develop/CHANGELOG.md)
- [Commits](https://github.com/ROCm/rocm-docs-core/compare/v1.8.2...v1.8.3)

---
updated-dependencies:
- dependency-name: rocm-docs-core
  dependency-type: direct:production
  update-type: version-update:semver-patch
...

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

[ROCm/rccl commit: 4685d3c546]
 2024-10-21 10:05:58 -06:00
Bertan Dogancay 57710c1183 Dynamically select unroll factor to build for when targeting local arch (#1371) 
* Dynamically select unroll factor to build for when targeting local arch only

[ROCm/rccl commit: 373f113524]
 2024-10-21 10:53:11 -04:00
Wenkai Du 5ee84e0353 Increase CQ size to 3*MAX_REQUESTS (#1374) 
* Increase CQ size to 3*MAX_REQUESTS

Suggested by Rukhsana Ansari <rukhsana.ansari@broadcom.com>

* Reword comments based on feedback from Rukhsana

[ROCm/rccl commit: 7c077db307]
 2024-10-18 11:01:03 -07:00