Skip to content

Upstream sync 26/N: merge 5fa8ca971a [ROCm][CI] Move ROCm AITER quantization tests (#40938) (conflict) - #1273

Merged
eble-amd merged 5 commits into
gfx11from
rogarcia.merge-upstream-122
Sep 10, 2026
Merged

eble-amd merged 5 commits into
gfx11from
rogarcia.merge-upstream-122

Conversation

@roberteg16

@roberteg16 roberteg16 commented Sep 7, 2026 •

Copy link
Copy Markdown

Context

Twenty-sixth step. Stacked on #1272. Conflict-only step.

Merged upstream changes:

  1. 5fa8ca971a [ROCm][CI] Move ROCm AITER quantization tests vllm-project/vllm#40938 — [ROCm][CI] Move ROCm AITER quantization tests

One conflict, and it is a rename/location conflict rather than a content one — the only instance of this shape in the campaign:

CONFLICT (file location): tests/rocm/aiter/test_fused_qk_norm_mrope_kvcache.py
added in rogarcia.merge-upstream-121 inside a directory that was renamed in
5fa8ca971a, suggesting it should perhaps be moved to
tests/kernels/quantization/test_fused_qk_norm_mrope_kvcache.py
ours:     tests/rocm/aiter/test_fused_qk_norm_mrope_kvcache.py
theirs:   <directory renamed away; this file does not exist upstream>
resolved: tests/kernels/quantization/test_fused_qk_norm_mrope_kvcache.py
          (same content, new location — accept git's relocation)

Upstream dissolves tests/rocm/aiter/ into the standard tests/kernels/ tree. The fork owns one extra file there; git relocated it but left it staged AU (added by us, unmerged) rather than deciding alone.

Audit

Verified it is a pure move: diffed the file at the new path against the fork's copy at batch 25 — byte-identical, no content merge occurred — and confirmed tests/rocm/ is now empty rather than a stray half-populated tree.

The part that actually matters, because a misplaced test does not fail, it simply stops running: no reference to tests/rocm or rocm/aiter survives anywhere under .buildkite/, and kernels.yaml runs the whole directory (pytest -v -s kernels/quantization, sharded), so the relocated file is picked up by path. Leaving it at the old location would have quietly dropped the fork's qk-norm/mrope/kvcache coverage from every CI run.

py_compile passes on the relocated file.

Merge commit only — do not squash or rebase.

AI assistance was used to prepare this merge.

Test plan

  • Relocation proven content-identical; old directory confirmed drained
  • CI confirmed to collect the new path by directory, not by filename
  • Build + 6-model benchmark sweep vs the Strix Halo dashboard (queued)
    (N/A: no relevant program code was touched)

AndreasKaratzas and others added 4 commits August 18, 2026 14:50
Signed-off-by: Andreas Karatzas <akaratza@amd.com>
Signed-off-by: Andreas Karatzas <Andreas.Karatzas@amd.com>
Co-authored-by: OpenAI Codex <codex@openai.com>
…tion tests (vllm-project#40938) (conflict)

Conflict-only step. One conflict, and it is a *rename/location* conflict rather
than a content one - the only instance of this shape so far in the campaign:

  CONFLICT (file location): tests/rocm/aiter/test_fused_qk_norm_mrope_kvcache.py
  added in rogarcia.merge-upstream-121 inside a directory that was renamed in
  5fa8ca9, suggesting it should perhaps be moved to
  tests/kernels/quantization/test_fused_qk_norm_mrope_kvcache.py

Upstream dissolves tests/rocm/aiter/ into the standard tests/kernels/ tree
(test_aiter_hipb_mm_linear_kernel.py, test_quant_op_schema.py,
test_grouped_quant.py all move to tests/kernels/quantization/, and
test_mla_fp8_support_check.py becomes
tests/kernels/attention/test_rocm_aiter_mla_fp8_support.py). The fork owns one
additional file in that directory, and git relocated it but left it staged AU
(added by us, unmerged) for confirmation rather than deciding on its own.

  ours:     tests/rocm/aiter/test_fused_qk_norm_mrope_kvcache.py
  theirs:   <directory renamed away; file does not exist upstream>
  resolved: tests/kernels/quantization/test_fused_qk_norm_mrope_kvcache.py
            (same content, new location - accept git's relocation)

Accepted the relocation. Verified it is a pure move: diffed the file at the new
path against the fork's copy at batch 25 - byte-identical, no content merge
happened - and confirmed tests/rocm/ is now empty rather than left as a stray
half-populated directory.

The part worth checking, since a misplaced test does not fail but simply stops
running: CI still covers it. No reference to tests/rocm or rocm/aiter survives
anywhere under .buildkite/, and kernels.yaml runs the whole directory
(pytest -v -s kernels/quantization, sharded), so the relocated file is picked up
by path rather than by name. Leaving it at the old location would have quietly
dropped the fork's qk-norm/mrope/kvcache coverage from every CI run.

py_compile passes on the relocated file.
@roberteg16
roberteg16 requested a review from eble-amd September 9, 2026 13:41
@eble-amd

eble-amd commented Sep 9, 2026

Copy link
Copy Markdown

.github/workflows/build-rocm-wheels.yml needs to be updated to cope with the rename. @roberteg16, would you prefer that I try to push a fix to your PR, create a new PR of my own, or what?

@roberteg16

Copy link
Copy Markdown
Author

.github/workflows/build-rocm-wheels.yml needs to be updated to cope with the rename. @roberteg16, would you prefer that I try to push a fix to your PR, create a new PR of my own, or what?

Feel free to push commits on the PRs

Signed-off-by: Dan Eble <Dan.Eble@amd.com>
@eble-amd
eble-amd changed the base branch from rogarcia.merge-upstream-121 to gfx11 September 10, 2026 11:53
@eble-amd
eble-amd marked this pull request as ready for review September 10, 2026 14:15
@eble-amd
eble-amd merged commit 554a3d4 into gfx11 Sep 10, 2026
8 of 9 checks passed
@roberteg16
roberteg16 deleted the rogarcia.merge-upstream-122 branch October 12, 2026 12:57
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants