Skip to content

fix(tensilelite): Accept .dat.zlib in --use-cache existence check#8892

Merged
archana-ramalingam merged 2 commits into
developfrom
users/aramalin/fix-dat-zlib-existence-check
Jul 1, 2026
Merged

fix(tensilelite): Accept .dat.zlib in --use-cache existence check#8892
archana-ramalingam merged 2 commits into
developfrom
users/aramalin/fix-dat-zlib-existence-check

Conversation

@archana-ramalingam

@archana-ramalingam archana-ramalingam commented Jun 29, 2026

Copy link
Copy Markdown
Contributor

Motivation

PR #8294 changed writeMsgPack() to compress library catalogs with zlib, producing .dat.zlib files instead of .dat. The C++ runtime's fileToMsgObject() was updated to probe for .zlib automatically, but the Python --use-cache existence check in BenchmarkProblems.py was missed — it still checks only for .dat, causing all --use-cache benchmark tests to FATAL when LibraryFormat=msgpack.

Technical Details

Widen the os.path.isfile() existence check to accept both .dat and .dat.zlib on disk. The path passed to the C++ client stays as .dat, matching the C++ probe convention from PR #8294 where fileToMsgObject() tries filename + ".zlib" before the base file.

Test Plan

  • Emulation CI: common tests pass with LibraryFormat=msgpack
  • No change to tensileLibraryFile() or writeClientConfigIni — C++ client probe convention preserved
  • Jenkins tox -e py3 (yaml format, unaffected path)

Submission Checklist

JIRA ID: AIHPBLAS-3508

@therock-pr-bot

therock-pr-bot Bot commented Jun 29, 2026

Copy link
Copy Markdown

✅ All Checks Passed — Ready for Review

Check Status Details
🌿 Branch Name ✅ Pass
📝 PR Title/Description ✅ Pass
Forbidden Files ✅ Pass
🧪 Unit Test ✅ Pass
🔎 pre-commit ✅ Pass
🚫 Draft PR 🔜 To Be Enabled
🚩 Feature Flag 🔜 To Be Enabled
📊 Code Coverage 🔜 To Be Enabled
🤖 therock-pr-bot ✅ Pass

🎉 All checks passed! This PR is ready for review.

📖 Need help? See the Policy FAQ for details on every check and how to fix failures.

@therock-pr-bot

therock-pr-bot Bot commented Jun 29, 2026

Copy link
Copy Markdown

🎉 All checks passed! This PR is ready for review.

@codecov-commenter

codecov-commenter commented Jun 29, 2026

Copy link
Copy Markdown

Codecov Report

❌ Patch coverage is 0% with 1 line in your changes missing coverage. Please review.

Files with missing lines Patch % Lines
...hipblaslt/tensilelite/Tensile/BenchmarkProblems.py 0.00% 1 Missing ⚠️

❌ Your project status has failed because the head coverage (76.92%) is below the target coverage (80.00%). You can increase the head coverage or adjust the target coverage.

Additional details and impacted files
@@             Coverage Diff             @@
##           develop    #8892      +/-   ##
===========================================
- Coverage    71.27%   71.27%   -0.00%     
===========================================
  Files         2613     2613              
  Lines       409506   409509       +3     
  Branches     61210    61210              
===========================================
  Hits        291868   291868              
- Misses       96183    96186       +3     
  Partials     21455    21455              
Flag Coverage Δ *Carryforward flag
TensileLite 76.66% <0.00%> (ø)
hipBLAS 90.81% <ø> (ø) Carriedforward from edcb934
hipBLASLt 41.35% <ø> (-0.01%) ⬇️
hipCUB 82.68% <ø> (ø) Carriedforward from edcb934
hipDNN 86.53% <ø> (ø) Carriedforward from edcb934
hipFFT 50.17% <ø> (ø) Carriedforward from edcb934
hipRAND 76.12% <ø> (ø) Carriedforward from edcb934
hipSOLVER 69.18% <ø> (ø) Carriedforward from edcb934
hipSPARSE 86.55% <ø> (ø) Carriedforward from edcb934
rocBLAS 48.08% <ø> (ø) Carriedforward from edcb934
rocFFT 46.30% <ø> (ø) Carriedforward from edcb934
rocRAND 57.07% <ø> (ø) Carriedforward from edcb934
rocSOLVER 76.92% <ø> (ø) Carriedforward from edcb934
rocSPARSE 72.37% <ø> (ø) Carriedforward from edcb934
rocThrust 91.36% <ø> (ø) Carriedforward from edcb934

*This pull request uses carry forward flags. Click here to find out more.

Files with missing lines Coverage Δ
...hipblaslt/tensilelite/Tensile/BenchmarkProblems.py 73.11% <0.00%> (ø)

... and 1 file with indirect coverage changes

🚀 New features to boost your workflow:
  • ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
  • 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

@archana-ramalingam archana-ramalingam force-pushed the users/aramalin/fix-dat-zlib-existence-check branch from 35c8f5e to 3787c94 Compare June 29, 2026 22:28
…HPBLAS-3508]

PR #8294 changed writeMsgPack() to produce .dat.zlib files but missed
updating the Python existence check in BenchmarkProblems.py. The C++
client already probes for .zlib automatically (fileToMsgObject tries
filename + ".zlib" first), so the fix widens only the Python check to
accept both .dat and .dat.zlib on disk.

Without this fix, all --use-cache benchmark tests FATAL with
"cache.yaml refers to a library file that no longer exists" when
LibraryFormat=msgpack.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
@archana-ramalingam archana-ramalingam force-pushed the users/aramalin/fix-dat-zlib-existence-check branch from 65c402b to 0f87fd5 Compare June 30, 2026 07:20

@talumbau talumbau left a comment

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

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

thanks for the quick fix!

@archana-ramalingam archana-ramalingam merged commit 23563fb into develop Jul 1, 2026
61 of 65 checks passed
@archana-ramalingam archana-ramalingam deleted the users/aramalin/fix-dat-zlib-existence-check branch July 1, 2026 00:36
JP-Fernando pushed a commit that referenced this pull request Jul 1, 2026
)

## Motivation

PR #8294 changed `writeMsgPack()` to compress library catalogs with
zlib, producing `.dat.zlib` files instead of `.dat`. The C++ runtime's
`fileToMsgObject()` was updated to probe for `.zlib` automatically, but
the Python `--use-cache` existence check in `BenchmarkProblems.py` was
missed — it still checks only for `.dat`, causing all `--use-cache`
benchmark tests to FATAL when `LibraryFormat=msgpack`.

## Technical Details

Widen the `os.path.isfile()` existence check to accept both `.dat` and
`.dat.zlib` on disk. The path passed to the C++ client stays as `.dat`,
matching the C++ probe convention from PR #8294 where
`fileToMsgObject()` tries `filename + ".zlib"` before the base file.

## Test Plan

- [x] Emulation CI: common tests pass with `LibraryFormat=msgpack`
- [x] No change to `tensileLibraryFile()` or `writeClientConfigIni` —
C++ client probe convention preserved
- [x] Jenkins `tox -e py3` (yaml format, unaffected path)

## Submission Checklist

- [x] Look over the contributing guidelines at
https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests.

JIRA ID: AIHPBLAS-3508

Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants