Skip to content

[hipblaslt] Replace/Add some tiles from tuned mi350P BBSTN Equality to Origami#7909

Open
yenong-amd wants to merge 4 commits into
developfrom
users/yenong-amd/mi350P_tiles
Open

[hipblaslt] Replace/Add some tiles from tuned mi350P BBSTN Equality to Origami#7909
yenong-amd wants to merge 4 commits into
developfrom
users/yenong-amd/mi350P_tiles

Conversation

@yenong-amd
Copy link
Copy Markdown
Contributor

@yenong-amd yenong-amd commented May 30, 2026

AIHPBLAS-3632

Motivation

Close the gap between Equality and Origami BBSTN library for mi350P.

Technical Details

Replaced small MT 16x16x256, 32x16x256 and 32x32x256 solutions with higher PGR values.
Added MT 384x160x64, 384x128x64, 384x96x64, 384x64x64, 352x192x64.

Test Plan

Benchmark with OOB and customer datasets for mi350 and mi350P.

Test Result

Good uplift for mi350P and no regressions in mi350.

Submission Checklist

@codecov-commenter
Copy link
Copy Markdown

Codecov Report

✅ All modified and coverable lines are covered by tests.

❌ Your project status has failed because the head coverage (77.83%) 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    #7909      +/-   ##
===========================================
- Coverage    62.06%   62.06%   -0.00%     
===========================================
  Files         2085     2085              
  Lines       357573   357623      +50     
  Branches     54060    54071      +11     
===========================================
+ Hits        221914   221928      +14     
- Misses      116871   116904      +33     
- Partials     18788    18791       +3     
Flag Coverage Δ *Carryforward flag
TensileLite 27.29% <ø> (ø) Carriedforward from 0dda48e
hipBLAS 90.65% <ø> (ø) Carriedforward from 0dda48e
hipBLASLt 41.23% <ø> (-0.05%) ⬇️
hipCUB 82.21% <ø> (ø) Carriedforward from 0dda48e
hipDNN 86.61% <ø> (ø) Carriedforward from 0dda48e
hipFFT 50.00% <ø> (ø) Carriedforward from 0dda48e
hipRAND 76.12% <ø> (ø) Carriedforward from 0dda48e
hipSOLVER 69.24% <ø> (ø) Carriedforward from 0dda48e
hipSPARSE 85.42% <ø> (ø) Carriedforward from 0dda48e
rocBLAS 48.09% <ø> (ø) Carriedforward from 0dda48e
rocFFT 52.07% <ø> (ø) Carriedforward from 0dda48e
rocRAND 57.04% <ø> (ø) Carriedforward from 0dda48e
rocSOLVER 77.83% <ø> (ø) Carriedforward from 0dda48e
rocSPARSE 72.68% <ø> (ø) Carriedforward from 0dda48e

*This pull request uses carry forward flags. Click here to find out more.
see 6 files 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.

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.

2 participants