Skip to content

Bump rocm-systems from 0decb2c to 6276d4d#3307

Merged
amd-justchen merged 1 commit intomainfrom
dependabot/submodules/main/rocm-systems-6276d4d
Feb 16, 2026
Merged

Bump rocm-systems from 0decb2c to 6276d4d#3307
amd-justchen merged 1 commit intomainfrom
dependabot/submodules/main/rocm-systems-6276d4d

Conversation

@dependabot
Copy link
Contributor

@dependabot dependabot bot commented on behalf of github Feb 9, 2026

Bumps rocm-systems from 0decb2c to 6276d4d.

Commits
  • 6276d4d Fix FindTBB version lookup logic (#3113)
  • 8cc3468 rocrtst: Enable rocrtstPerf.Memory_Async_Copy emu (#3080)
  • f162895 rocr-runtime: fix segfault when queue allocation fails (#2850)
  • 8d84709 fix shutdown ordering race condition, prevent use-after-free crash (#3004)
  • 9c1de5d Disable Direct Reduce Scatter if PXN is disabled (#3077)
  • 330dec2 Fix merging perfetto files from cached data when multiple mpi rangs are avail...
  • e4c0801 [rocprofiler-systems] Fix for Perfetto flow events (#3111)
  • 291173f Add workload specific presets for easier and faster profiling (#2592)
  • 85db4e4 [ROCM] Enhance amd-smi node to display baseboard temp (#2943)
  • 67a57d9 [SWDEV-552020/SWDEV-563971] Fail memory partition tests if ASIC supports memo...
  • Additional commits viewable in compare view

Dependabot will resolve any conflicts with this PR as long as you don't alter it yourself. You can also trigger a rebase manually by commenting @dependabot rebase.


Dependabot commands and options

You can trigger Dependabot actions by commenting on this PR:

  • @dependabot rebase will rebase this PR
  • @dependabot recreate will recreate this PR, overwriting any edits that have been made to it
  • @dependabot show <dependency name> ignore conditions will show all of the ignore conditions of the specified dependency
  • @dependabot ignore this major version will close this PR and stop Dependabot creating any more for this major version (unless you reopen the PR or upgrade to it yourself)
  • @dependabot ignore this minor version will close this PR and stop Dependabot creating any more for this minor version (unless you reopen the PR or upgrade to it yourself)
  • @dependabot ignore this dependency will close this PR and stop Dependabot creating any more for this dependency (unless you reopen the PR or upgrade to it yourself)

@dependabot dependabot bot added dependencies Pull requests that update a dependency file submodules Pull requests that update submodules code labels Feb 9, 2026
@dependabot dependabot bot requested a review from a team as a code owner February 9, 2026 11:35
@dependabot dependabot bot added dependencies Pull requests that update a dependency file submodules Pull requests that update submodules code labels Feb 9, 2026
Copy link
Contributor

@geomin12 geomin12 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

lgtm assuming CI passes

FYI, we need to retry failed jobs, as the GitHub issues earlier caused failures

@amd-justchen
Copy link
Contributor

Will also take a look since on rotation this week!

@amd-justchen
Copy link
Contributor

Looked at https://github.com/ROCm/TheRock/actions/runs/21823463045 and saw some failed tests... Re-running those failed ones.

@dependabot @github
Copy link
Contributor Author

dependabot bot commented on behalf of github Feb 11, 2026

A newer version of rocm-systems exists, but since this PR has been edited by someone other than Dependabot I haven't updated it. You'll get a PR for the updated version as normal once this PR is merged.

@chiranjeevipattigidi
Copy link
Contributor

found problematic commit for rocrtst failures: ROCm/rocm-systems@de8012a

tested locally by reverting above commit, able to build rocrtst

((.venv) ) [root@ff4ffd00d823 rocm-systems]# git log --oneline
c9d9cbce7f (HEAD) Revert "rocrtst: Disabled NUMA async copy build (#3044)"
[rocrtst] 1 warning generated.
[rocrtst] [481/484] Building CXX object gtest/CMakeFiles/google-test-frwk64.dir/src/gtest.cpp.o
[rocrtst] [482/484] Linking CXX static library gtest/libgoogle-test-frwk64.a
[rocrtst] [483/484] Building CXX object CMakeFiles/rocrtst64.dir/therock/src/rocm-systems/projects/rocr-runtime/rocrtst/suites/functional/counted_queues.cc.o
[rocrtst] [484/484] Linking CXX executable rocrtst64
[rocrtst SUCCEEDED in 2 seconds]
[12/13] Stage installing sub-project rocrtst```

Bumps [rocm-systems](https://github.com/ROCm/rocm-systems) from `0decb2c` to `6276d4d`.
- [Release notes](https://github.com/ROCm/rocm-systems/releases)
- [Commits](ROCm/rocm-systems@0decb2c...6276d4d)

---
updated-dependencies:
- dependency-name: rocm-systems
  dependency-version: 6276d4d7ab8350531e84a24d3db65b9f98d85eb6
  dependency-type: direct:production
...

Signed-off-by: dependabot[bot] <support@github.com>
@jharryma jharryma force-pushed the dependabot/submodules/main/rocm-systems-6276d4d branch from 84a3b89 to 1251645 Compare February 11, 2026 21:16
@amd-justchen
Copy link
Contributor

amd-justchen commented Feb 12, 2026

Quick notes on 11 failures so far:

  1. File exists: '../rocrtst64' -> 'build/bin/gfx1013/rocrtst64' file exists gfx94x dcpu mi325? (infra flake?) on Runner name: 'linux-mi325-1gpu-ossci-rocm-2765n-runner-zdv7s'
    https://github.com/ROCm/TheRock/actions/runs/21923355781/job/63322853654#step:6:886

    ++ Flattening 'blas_lib_gfx94X-dcgpu.tar.xz' to 'blas_lib_gfx94X-dcgpu'
    Traceback (most recent call last):
      File "/__w/TheRock/TheRock/./build_tools/install_rocm_from_artifacts.py", line 751, in <module>
        main(sys.argv[1:])
      File "/__w/TheRock/TheRock/./build_tools/install_rocm_from_artifacts.py", line 747, in main
        run(args)
      File "/__w/TheRock/TheRock/./build_tools/install_rocm_from_artifacts.py", line 542, in run
        retrieve_artifacts_by_run_id(args)
      File "/__w/TheRock/TheRock/./build_tools/install_rocm_from_artifacts.py", line 423, in retrieve_artifacts_by_run_id
        fetch_artifacts_main(argv)
      File "/__w/TheRock/TheRock/build_tools/fetch_artifacts.py", line 345, in main
        run(args)
      File "/__w/TheRock/TheRock/build_tools/fetch_artifacts.py", line 249, in run
        [f.result() for f in extract_futures]
         ^^^^^^^^^^
      File "/__w/_tool/Python/3.12.12/x64/lib/python3.12/concurrent/futures/_base.py", line 449, in result
        return self.__get_result()
               ^^^^^^^^^^^^^^^^^^^
      File "/__w/_tool/Python/3.12.12/x64/lib/python3.12/concurrent/futures/_base.py", line 401, in __get_result
        raise self._exception
      File "/__w/_tool/Python/3.12.12/x64/lib/python3.12/concurrent/futures/thread.py", line 59, in run
        result = self.fn(*self.args, **self.kwargs)
                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
      File "/__w/TheRock/TheRock/build_tools/fetch_artifacts.py", line 152, in extract_artifact
        flattener(archive_file)
      File "/__w/TheRock/TheRock/build_tools/_therock_utils/artifacts.py", line 257, in __call__
        dest_path.symlink_to(member.linkname)
      File "/__w/_tool/Python/3.12.12/x64/lib/python3.12/pathlib.py", line 1386, in symlink_to
        os.symlink(target, self, target_is_directory)
    FileExistsError: [Errno 17] File exists: '../rocrtst64' -> 'build/bin/gfx1013/rocrtst64'
    

    succeeded here: https://github.com/ROCm/TheRock/actions/runs/21923355781/job/63423255344#step:6:857

  2. 6x windows 1151 no space issue (infra)

  3. docker issue on linux 1151 lib wheels (infra)

  4. test lib+devel wheel could not download gfx120x, timeout (infra)

  5. linux rocm lib(only) wheels 1151 infra issue cannot start container (infra)

  6. lib+devel wheels for windows on gfx1151 rocm_sdk test failed

  7. hipcub actual failure? unspecified launch failure testing on 1151

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

  • lib+devel wheels for windows on gfx1151 rocm_sdk test failed

https://github.com/ROCm/TheRock/actions/runs/21923355781/job/63423253017?pr=3307#step:6:37

testCLIPathBin (rocm_sdk.tests.devel_test.ROCmDevelTest.testCLIPathBin) ... ++ Exec [B:\actions-runner\_work\TheRock\TheRock]$ 'B:\actions-runner\_work\TheRock\TheRock\.venv\Scripts\python.exe' -P -m rocm_sdk path --bin
Traceback (most recent call last):
  File "<frozen runpy>", line 198, in _run_module_as_main
  File "<frozen runpy>", line 88, in _run_code
  File "B:\actions-runner\_work\TheRock\TheRock\.venv\Lib\site-packages\rocm_sdk\__main__.py", line 154, in <module>
    main()
  File "B:\actions-runner\_work\TheRock\TheRock\.venv\Lib\site-packages\rocm_sdk\__main__.py", line 150, in main
    args.func(args)
  File "B:\actions-runner\_work\TheRock\TheRock\.venv\Lib\site-packages\rocm_sdk\__main__.py", line 16, in _do_path
    root_path = _devel.get_devel_root()
                ^^^^^^^^^^^^^^^^^^^^^^^
  File "B:\actions-runner\_work\TheRock\TheRock\.venv\Lib\site-packages\rocm_sdk\_devel.py", line 63, in get_devel_root
    _expand_devel_contents(rocm_sdk_devel_path, site_lib_path)
  File "B:\actions-runner\_work\TheRock\TheRock\.venv\Lib\site-packages\rocm_sdk\_devel.py", line 154, in _expand_devel_contents
    _lock_and_expand(
  File "B:\actions-runner\_work\TheRock\TheRock\.venv\Lib\site-packages\rocm_sdk\_devel.py", line 208, in _lock_and_expand
    dest_path.hardlink_to(hardlink_target)
  File "B:\actions-runner\_work\_tool\Python\3.12.10\x64\Lib\pathlib.py", line 1396, in hardlink_to
    os.link(target, self)
FileNotFoundError: [WinError 3] The system cannot find the path specified: 'B:\\actions-runner\\_work\\TheRock\\TheRock\\.venv\\Lib\\site-packages\\_rocm_sdk_devel\\bin\\hipblaslt\\library\\..\\..\\..\\..\\_rocm_sdk_libraries_gfx1151\\bin\\hipblaslt\\library\\TensileLibrary_BB_BB_HA_Bias_Aux_SAV_UA_Type_BB_HPA_Contraction_l_Ailk_Bjlk_Cijk_Dijk_gfx1151.co' -> 'B:\\actions-runner\\_work\\TheRock\\TheRock\\.venv\\Lib\\site-packages\\_rocm_sdk_devel\\bin\\hipblaslt\\library\\TensileLibrary_BB_BB_HA_Bias_Aux_SAV_UA_Type_BB_HPA_Contraction_l_Ailk_Bjlk_Cijk_Dijk_gfx1151.co'
ERROR
testCLIPathCMake (rocm_sdk.tests.devel_test.ROCmDevelTest.testCLIPathCMake) ... ++ Exec [B:\actions-runner\_work\TheRock\TheRock]$ 'B:\actions-runner\_work\TheRock\TheRock\.venv\Scripts\python.exe' -P -m rocm_sdk path --cmake
FAIL

This looks like a file path length issue? FileNotFoundError: [WinError 3] The system cannot find the path specified: 'B:\\actions-runner\\_work\\TheRock\\TheRock\\.venv\\Lib\\site-packages\\_rocm_sdk_devel\\bin\\hipblaslt\\library\\..\\..\\..\\..\\_rocm_sdk_libraries_gfx1151\\bin\\hipblaslt\\library\\TensileLibrary_BB_BB_HA_Bias_Aux_SAV_UA_Type_BB_HPA_Contraction_l_Ailk_Bjlk_Cijk_Dijk_gfx1151.co' -> 'B:\\actions-runner\\_work\\TheRock\\TheRock\\.venv\\Lib\\site-packages\\_rocm_sdk_devel\\bin\\hipblaslt\\library\\TensileLibrary_BB_BB_HA_Bias_Aux_SAV_UA_Type_BB_HPA_Contraction_l_Ailk_Bjlk_Cijk_Dijk_gfx1151.co'

I wouldn't expect a change to rocm-systems to affect those files from rocm-libraries... I wonder if the tests are flaky or dependent on the runner? Is long path support enabled on that test runner?

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Good call, will check for that on

Runner name: 'windows-strix-halo-gpu-rocm-2'
Runner group name: 'default'
Machine name: 'CHRN-SI-112'

'B:\\actions-runner\\_work\\TheRock\\TheRock\\.venv\\Lib\\site-packages\\_rocm_sdk_devel\\bin\\hipblaslt\\library\\..\\..\\..\\..\\_rocm_sdk_libraries_gfx1151\\bin\\hipblaslt\\library\\TensileLibrary_BB_BB_HA_Bias_Aux_SAV_UA_Type_BB_HPA_Contraction_l_Ailk_Bjlk_Cijk_Dijk_gfx1151.co' seems to be exactly 260 characters, though it gets resolved to the latter path to 198

Copy link
Contributor

@amd-justchen amd-justchen Feb 16, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

HKLM\SYSTEM\CurrentControlSet\Control\FileSystem\LongPathsEnabled was currently 0, so will now set to 1

(May be toggled via Windows Search > MAX_PATH > Enable long paths > Toggle: On)

@amd-justchen amd-justchen self-assigned this Feb 14, 2026
@amd-justchen
Copy link
Contributor

Noticed a force push, so dependabot no longer is managing, but it did create a new bump PR here #3411 (for later commits?)

@amd-justchen
Copy link
Contributor

amd-justchen commented Feb 16, 2026

Will go ahead and merge this after some standup discussions and speaking with @geomin12 about whether these failures are blocking for merging.

1. File exists: '../rocrtst64' -> 'build/bin/gfx1013/rocrtst64' file exists gfx94x dcpu mix324? (infra flake?)
2. 6x windows 1151 no space issue (infra)

3. docker issue on linux 1151 lib wheels (infra)

4. test lib+devel wheel could not download gfx120x, timeout (infra)

5. linux rocm lib(only) wheels 1151 infra issue cannot start container (infra)

6. lib+devel wheels for windows on gfx1151 rocm_sdk test failed

7. hipcub actual failure? `unspecified launch failure` testing on 1151
  • it seems these are all typical INFRA issues with these details:
  1. gfx120x - lab machines timing out (shark40 and shark42)
  1. gfx1151 miopen 3/4 shard

For try 4 and 5, the failures could be related to MIOpen test sizes. ROCm/rocm-libraries#3956 (comment)

@ScottTodd
Copy link
Member

1. File exists: '../rocrtst64' -> 'build/bin/gfx1013/rocrtst64' file exists gfx94x dcpu mix324? (infra flake?)

That is #3438 . There's a real code issue there, but it isn't a regression.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

dependencies Pull requests that update a dependency file submodules Pull requests that update submodules code

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants