Skip to content

Int cts tests for gfx1250 - #41

Merged
kuhar merged 2 commits into
developfrom
users/lua1235/gfx1250-int-cts
Aug 20, 2026
Merged

Int cts tests for gfx1250#41
kuhar merged 2 commits into
developfrom
users/lua1235/gfx1250-int-cts

Conversation

@lua1235

@lua1235 lua1235 commented Aug 18, 2026 •
edited
Loading

Copy link
Copy Markdown
Contributor

Motivation

Add cts tests for gfx1250 integer intrinsics

Technical Details

  • Integer bitfield extraction
  • Packed-byte interpolation (lerp)
  • SAD/MSAD and packed SAD operations
  • Bit/byte alignment and byte permutation
  • 8-bit and 4-bit integer dot products, including signedness and clamp modes
  • Packed integer conversions and normalized float-to-integer packing
  • Saturating sub-byte packing
  • Packed 16-bit add/min/max
  • Packed 4/6/8-bit lookup permutations
  • M0-relative moves and relative swap
  • Dense integer WMMA
  • Sparse integer SWMMAC
  • Lane reads and first-lane reads
  • Ballot and population-count/mbcnt
  • DS permutation and swizzle
  • DPP lane movement
  • Fixed and variable permlane operations
  • Signed and unsigned integer comparisons
  • Wave-level integer reductions: add, min/max, AND/OR/XOR

Test Plan

Run tests against rocjitsu origin/develop

Test Result

12/22 pass. Most are due to known errors, and fixes are already in flight

Submission Checklist

@lua1235
lua1235 requested a review from a team as a code owner August 18, 2026 20:57
kuhar
kuhar reviewed Aug 18, 2026
View reviewed changes

@kuhar kuhar left a comment

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

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

We should give the two gfx1250 matrix tests target-specific fragment packers before treating them as scalar-oracle tests. At corpus/cts/test_cases/general/int_isa/matrix_test_parts.hpp:227, A/B are packed as contiguous K blocks selected by lane >> 4; the CDNA5 8-bit layout instead interleaves K across lane halves and VGPR slots, matching the in-tree Wmma16x16x64Layout. The sparse path used by corpus/cts/test_cases/gfx1250/int_isa/swmmac_i32_16x16x128_iu8_test.cpp:20 likewise reuses the RDNA4 packing instead of the existing gfx1250 compressed-A, dense-B, and selector mappings. The linked-image gate only checks opcode presence, so it can pass while the runtime compares a differently staged matrix. Please split the packing strategy by target or reuse the gfx1250 layout helpers.

Comment thread corpus/cts/CMakeLists.txt Resolved
Comment thread corpus/cts/int_isa_gfx1250_coverage.json Resolved
Comment thread tests/test_int_isa_coverage.py Resolved
@lua1235
lua1235 force-pushed the users/lua1235/gfx1250-int-cts branch from 082ff38 to 372d1a5 Compare August 19, 2026 18:16
@kuhar
kuhar merged commit d96cc27 into develop Aug 20, 2026
@Groverkss Groverkss mentioned this pull request Oct 6, 2026
Merged
1 task done
kuhar pushed a commit to ROCm/rocm-systems that referenced this pull request Oct 7, 2026
@Groverkss
rocjitsu: bump test corpus for optimized CTS builds (#12932)
0c3808a 
## Motivation

Rocjitsu CI still pins a corpus revision that builds most CTS tests
without optimization. Pick up ROCm/rocjitsu-test-corpus#53 so CI uses
Release builds with assertions enabled while preserving required integer
instruction coverage.

## Technical Details

Bump the HIP/DBT corpus pin from
`cc781dd88cabd1638d1beeb4539c521ccf27de60` to
`387502e864cce3f0490a967e9a185b4922a5e7de` in both the automatic
checkout fallback and the manual-dispatch default. Update the pin's date
comment.

The pinned revision includes the wave32 `mbcnt` and gfx1250 comparison
optimization overrides and the strengthened `mbcnt_hi` coverage
requirement. The bump also includes intervening corpus changes,
including the newer gfx1250 ISA coverage and runtime CTS additions; see
the [corpus commit
range](ROCm/rocjitsu-test-corpus@cc781dd...387502e).

Temporarily exclude all 37 gfx1250 CTS cases introduced by
ROCm/rocjitsu-test-corpus#41 (22 integer cases) and
ROCm/rocjitsu-test-corpus#42 (15 memory cases), including their coverage
checks. The existing skip configuration applies to both the initial run
and failed-test retries.

Reason: These tests are really slow and haven't been verified for their
usefulness. Someone needs to investigate these before they are enabled

## Issue Tracking

ISSUE ID: ROCm/rocjitsu-test-corpus#53

## Test Plan

- Parse the workflow YAML and verify the dispatch default and automatic
checkout pin both select the merged corpus revision.
- Run the repository pre-commit checks on the changed workflow and `git
diff --check`.
- Exercise the updated pin through this PR's rocjitsu corpus workflow.

## Test Result

Local YAML/pin validation, applicable pre-commit checks, and whitespace
checks passed. A targeted pytest run confirms all 37 integer/memory
cases are skipped before build or execution (37 skipped, 43 deselected).

## Submission Checklist

- [x] Look over the contributing guidelines at
https://github.com/ROCm/rocm-systems/blob/develop/CONTRIBUTING.md.
anujshuk-amd pushed a commit to ROCm/rocm-systems that referenced this pull request Oct 12, 2026
@Groverkss @anujshuk-amd
rocjitsu: bump test corpus for optimized CTS builds (#12932)
1358653 
## Motivation

Rocjitsu CI still pins a corpus revision that builds most CTS tests
without optimization. Pick up ROCm/rocjitsu-test-corpus#53 so CI uses
Release builds with assertions enabled while preserving required integer
instruction coverage.

## Technical Details

Bump the HIP/DBT corpus pin from
`cc781dd88cabd1638d1beeb4539c521ccf27de60` to
`387502e864cce3f0490a967e9a185b4922a5e7de` in both the automatic
checkout fallback and the manual-dispatch default. Update the pin's date
comment.

The pinned revision includes the wave32 `mbcnt` and gfx1250 comparison
optimization overrides and the strengthened `mbcnt_hi` coverage
requirement. The bump also includes intervening corpus changes,
including the newer gfx1250 ISA coverage and runtime CTS additions; see
the [corpus commit
range](ROCm/rocjitsu-test-corpus@cc781dd...387502e).

Temporarily exclude all 37 gfx1250 CTS cases introduced by
ROCm/rocjitsu-test-corpus#41 (22 integer cases) and
ROCm/rocjitsu-test-corpus#42 (15 memory cases), including their coverage
checks. The existing skip configuration applies to both the initial run
and failed-test retries.

Reason: These tests are really slow and haven't been verified for their
usefulness. Someone needs to investigate these before they are enabled

## Issue Tracking

ISSUE ID: ROCm/rocjitsu-test-corpus#53

## Test Plan

- Parse the workflow YAML and verify the dispatch default and automatic
checkout pin both select the merged corpus revision.
- Run the repository pre-commit checks on the changed workflow and `git
diff --check`.
- Exercise the updated pin through this PR's rocjitsu corpus workflow.

## Test Result

Local YAML/pin validation, applicable pre-commit checks, and whitespace
checks passed. A targeted pytest run confirms all 37 integer/memory
cases are skipped before build or execution (37 skipped, 43 deselected).

## Submission Checklist

- [x] Look over the contributing guidelines at
https://github.com/ROCm/rocm-systems/blob/develop/CONTRIBUTING.md.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@kuhar kuhar kuhar approved these changes

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

2 participants

@lua1235 @kuhar