Skip to content

Fix stale TopK availability checks - #5210

Merged
coderfeli merged 1 commit into
ROCm:mainfrom
vorapolsiloai:fix/topk-flydsl-availability-import
Sep 2, 2026
Merged

coderfeli merged 1 commit into
ROCm:mainfrom
vorapolsiloai:fix/topk-flydsl-availability-import

Conversation

@vorapolsiloai

Copy link
Copy Markdown
Contributor

Summary

  • remove stale references to the deleted aiter.ops.flydsl.utils module
  • restore FlyDSL decode TopK dispatch on supported gfx942/gfx950 systems
  • fix op_tests/test_topk_per_row.py collection in the standard-test suite

Context

#5116 removed the optional FlyDSL availability helper and made FlyDSL imports unconditional. #5011 landed later with two references to the removed helper:

  • the test imports it eagerly and fails collection with ModuleNotFoundError
  • the runtime dispatch catches the same import failure and silently disables the FlyDSL TopK path

This was observed in the MI35X standard-test failure on #4761. The failure occurs before any GPU kernel executes.

Test plan

  • PYTHONPYCACHEPREFIX=/tmp/aiter-topk-fix-pycache python3 -m py_compile aiter/ops/topk.py op_tests/test_topk_per_row.py
  • git diff --check
  • verify no Python references to aiter.ops.flydsl.utils or is_flydsl_available remain
  • upstream CI

Made with Cursor

FlyDSL is required after ROCm#5116, so the removed helper breaks test collection and silently disables decode dispatch.

Co-authored-by: Cursor <cursoragent@cursor.com>
@vorapolsiloai
vorapolsiloai requested a review from a team September 2, 2026 11:21
@github-actions

github-actions Bot commented Sep 2, 2026

Copy link
Copy Markdown
Contributor

🏷️ CI Guide

Runs automatically on every PR:

  • ✅ Pre-checks (submodule verification, code formatting)
  • ✅ Aiter op tests (gfx942 + gfx950)
  • ✅ Triton tests on MI35X (only when aiter/ops/triton/** or related paths are changed)

Extended tests (opt-in via labels):

Label Tests
ci:gfx1250-ffm-triton Run the five-shard gfx1250 FFM Triton test suite
ci:triton-300x Run an additional Triton test job on MI300X in PRs; main branch always runs both MI35X and MI300X
multigpu Aiter multi-GPU tests on the 8-GPU runner
ci:sglang SGLang integration tests: DeepSeek-R1-MXFP4 accuracy, Qwen 3.5 accuracy
ci:atom ATOM benchmark: DeepSeek-R1-0528, GPT-OSS-120B
ci:atom_full ATOM accuracy suite for PR and main models from ATOM models_accuracy.json
ci:vllm vLLM benchmark: GPT-OSS-120B, DeepSeek-R1-0528, Kimi-K2.5
ci:all All standard extended tests (excludes ci:atom_full)

Only add ci:atom_full for FlyDSL or Triton upgrades.
Add labels via the sidebar or gh pr edit 5210 --add-label <label>

PR title tags & labels:
Component tags ([Triton/Gluon], [HIP], [CK], [ASM], ...) are added to the PR title and as PR labels automatically from the changed files and re-synced on every push — change-type tags like [fix]/[Perf], op tags like [MLA], and human labels (ci:*) are left untouched. Add the no-auto-title label to opt this PR out.

@github-actions github-actions Bot changed the title [FlyDSL] Fix stale TopK availability checks Fix stale TopK availability checks Sep 2, 2026
@zufayu
zufayu requested a review from junhaha666 September 2, 2026 12:22
@coderfeli
coderfeli merged commit f53e872 into ROCm:main Sep 2, 2026
46 of 49 checks passed
Fyzyukk pushed a commit to Fyzyukk/aiter that referenced this pull request Sep 3, 2026
FlyDSL is required after ROCm#5116, so the removed helper breaks test collection and silently disables decode dispatch.

Co-authored-by: Cursor <cursoragent@cursor.com>
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.

2 participants