Skip to content

fix(rocprim): check if wave size is supported in warp_sort benchmark - #4886

Merged
stanleytsang-amd merged 1 commit into
developfrom
users/mynameistrez/check-warp-size-in-warp-sort-benchmark
Feb 26, 2026
Merged

stanleytsang-amd merged 1 commit into
developfrom
users/mynameistrez/check-warp-size-in-warp-sort-benchmark

Conversation

@MyNameIsTrez

@MyNameIsTrez MyNameIsTrez commented Feb 25, 2026

Copy link
Copy Markdown
Contributor

Motivation

rocPRIM its benchmark_warp_sort.cpp had compilation and runtime errors, which have now been fixed.

Technical Details

Sinus Rollins from Stream HPC had originally fixed the runtime error in a GitLab merge request, but since he has been sick for the past couple of days, I have created this pull request for him and updated his commit.

PR Temporarily disable benchmark_warp_sort cases was merged yesterday, and it accidentally introduced the compilation error when attempting to fix the very same runtime error. Its fix was Windows-specific, so failed to compile on Linux.

Test Plan

I first reproduced the Linux compilation and runtime error on our Navi GPU (RX 9070 XT, so gfx1201):

ninja benchmark_warp_sort && benchmark/benchmark_warp_sort

Test Result

I confirmed that the compilation and runtime error is fixed by this PR on the same Navi GPU. All 40/40 benchmark specializations run without any runtime errors.

On our non-Navi GPU (MI210, so gfx90a), all 184/184 benchmark specializations run without any runtime errors.

Submission Checklist

@MyNameIsTrez
MyNameIsTrez requested a review from a team as a code owner February 25, 2026 16:47
@MyNameIsTrez MyNameIsTrez added the organization: streamhpc contributors from streamhpc label Feb 25, 2026
@MyNameIsTrez
MyNameIsTrez force-pushed the users/mynameistrez/check-warp-size-in-warp-sort-benchmark branch from cfc1ee7 to 633a1ca Compare February 25, 2026 17:10
@MyNameIsTrez
MyNameIsTrez requested a review from a team as a code owner February 25, 2026 17:10
@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 (76.83%) 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    #4886      +/-   ##
===========================================
- Coverage    65.36%   63.05%   -2.31%     
===========================================
  Files         1718     1870     +152     
  Lines       267205   292897   +25692     
  Branches     37045    37612     +567     
===========================================
+ Hits        174633   184664   +10031     
- Misses       77066    92350   +15284     
- Partials     15506    15883     +377     
Flag Coverage Δ *Carryforward flag
hipBLAS 90.67% <ø> (ø) Carriedforward from 309c8d2
hipBLASLt 43.55% <ø> (ø) Carriedforward from 309c8d2
hipCUB 82.21% <ø> (ø) Carriedforward from 309c8d2
hipDNN 80.74% <ø> (ø) Carriedforward from 309c8d2
hipFFT 55.93% <ø> (ø) Carriedforward from 309c8d2
hipRAND 76.12% <ø> (ø) Carriedforward from 309c8d2
hipSOLVER 68.81% <ø> (ø) Carriedforward from 309c8d2
hipSPARSE 84.70% <ø> (ø) Carriedforward from 309c8d2
rocBLAS 47.97% <ø> (ø) Carriedforward from 309c8d2
rocFFT 47.75% <ø> (ø) Carriedforward from 309c8d2
rocPRIM 39.04% <ø> (?)
rocRAND 57.06% <ø> (ø) Carriedforward from 309c8d2
rocSOLVER 76.83% <ø> (ø) Carriedforward from 309c8d2
rocSPARSE 71.53% <ø> (ø) Carriedforward from 309c8d2

*This pull request uses carry forward flags. Click here to find out more.
see 152 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.

@stanleytsang-amd stanleytsang-amd left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

Apologies for the oversight on my part. Approved, pending CI checks

@stanleytsang-amd
stanleytsang-amd merged commit e413e22 into develop Feb 26, 2026
22 of 24 checks passed
@stanleytsang-amd
stanleytsang-amd deleted the users/mynameistrez/check-warp-size-in-warp-sort-benchmark branch February 26, 2026 18:10
assistant-librarian Bot pushed a commit to ROCm/rocPRIM that referenced this pull request Feb 26, 2026
fix(rocprim): check if wave size is supported in warp_sort
 benchmark (#4886)

## Motivation

rocPRIM its `benchmark_warp_sort.cpp` had compilation and runtime
errors, which have now been fixed.

## Technical Details

Sinus Rollins from Stream HPC had originally fixed the runtime error in
a GitLab merge request, but since he has been sick for the past couple
of days, I have created this pull request for him and updated his
commit.

PR [Temporarily disable benchmark_warp_sort
cases](ROCm/rocm-libraries#4779) was merged
yesterday, and it accidentally introduced the compilation error when
attempting to fix the very same runtime error. Its fix was
Windows-specific, so failed to compile on Linux.

## Test Plan

I first reproduced the Linux compilation and runtime error on our Navi
GPU (RX 9070 XT, so gfx1201):
```sh
ninja benchmark_warp_sort && benchmark/benchmark_warp_sort
```

## Test Result

I confirmed that the compilation and runtime error is fixed by this PR
on the same Navi GPU. All 40/40 benchmark specializations run without
any runtime errors.

On our non-Navi GPU (MI210, so gfx90a), all 184/184 benchmark
specializations run without any runtime errors.

## Submission Checklist

- [x] Look over the contributing guidelines at
https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests.
jovanau pushed a commit to jovanau/rocm-libraries that referenced this pull request Mar 19, 2026
…OCm#4886)

## Motivation

rocPRIM its `benchmark_warp_sort.cpp` had compilation and runtime
errors, which have now been fixed.

## Technical Details

Sinus Rollins from Stream HPC had originally fixed the runtime error in
a GitLab merge request, but since he has been sick for the past couple
of days, I have created this pull request for him and updated his
commit.

PR [Temporarily disable benchmark_warp_sort
cases](ROCm#4779) was merged
yesterday, and it accidentally introduced the compilation error when
attempting to fix the very same runtime error. Its fix was
Windows-specific, so failed to compile on Linux.

## Test Plan

I first reproduced the Linux compilation and runtime error on our Navi
GPU (RX 9070 XT, so gfx1201):
```sh
ninja benchmark_warp_sort && benchmark/benchmark_warp_sort
```

## Test Result

I confirmed that the compilation and runtime error is fixed by this PR
on the same Navi GPU. All 40/40 benchmark specializations run without
any runtime errors.

On our non-Navi GPU (MI210, so gfx90a), all 184/184 benchmark
specializations run without any runtime errors.

## Submission Checklist

- [x] Look over the contributing guidelines at
https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests.
stanleytsang-amd pushed a commit that referenced this pull request Sep 18, 2026
…4886)

## Motivation

rocPRIM its `benchmark_warp_sort.cpp` had compilation and runtime
errors, which have now been fixed.

## Technical Details

Sinus Rollins from Stream HPC had originally fixed the runtime error in
a GitLab merge request, but since he has been sick for the past couple
of days, I have created this pull request for him and updated his
commit.

PR [Temporarily disable benchmark_warp_sort
cases](#4779) was merged
yesterday, and it accidentally introduced the compilation error when
attempting to fix the very same runtime error. Its fix was
Windows-specific, so failed to compile on Linux.

## Test Plan

I first reproduced the Linux compilation and runtime error on our Navi
GPU (RX 9070 XT, so gfx1201):
```sh
ninja benchmark_warp_sort && benchmark/benchmark_warp_sort
```

## Test Result

I confirmed that the compilation and runtime error is fixed by this PR
on the same Navi GPU. All 40/40 benchmark specializations run without
any runtime errors.

On our non-Navi GPU (MI210, so gfx90a), all 184/184 benchmark
specializations run without any runtime errors.

## Submission Checklist

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