CI: remove dep_gen workaround and disable dump_passes debug flag#675
Conversation
Revert the PR hw-native-sys#656 dep_gen CI workaround now that it is redundant, and turn off the leftover dump_passes debug flag across model entrypoints. - ci.yml: drop the `# ci: no-dep-gen` directive handling (marker grep + depgen_args) from the a2a3 loop - qwen3-14b decode_layer.py: remove the deprecated `--no-dep-gen` CLI option and its stale references; enable_dep_gen stays force-disabled in-kernel (full-occupancy syncall vs dep_gen, pypto#1931) - dump_passes=True -> False in v3_2 decode_back/decode_front/prefill_back, qwen3-32b decode/decode_4d/prefill_draft, and qwen3-14b decode_layer
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Pro Run ID: 📒 Files selected for processing (8)
📝 WalkthroughWalkthroughThis PR removes the CI workflow's conditional ChangesCompile config and CI toggle cleanup
Estimated code review effort: 1 (Trivial) | ~5 minutes Possibly related PRs
Poem
🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
There was a problem hiding this comment.
Code Review
This pull request disables the dumping of compiler passes (setting dump_passes to False) across several DeepSeek and Qwen3 model configuration files. Additionally, in models/qwen3/14b/decode_layer.py, the deprecated --no-dep-gen command-line argument, its associated CI marker, and related comments have been removed. As there are no review comments, I have no feedback to provide.
Important
The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.
Summary
# ci: no-dep-gendirective handling (marker grep +depgen_args) from the a2a3 test loop — added by perf(qwen3-14b decode): fuse rope+attn+softmax into one fa_fused kernel #656 as a workaround, now redundant.decode_layer.py: remove the deprecated--no-dep-genCLI option and its stale references.enable_dep_genstays force-disabled in-kernel (full-occupancypl.system.syncallis incompatible with dep_gen, pypto#1931), so the flag was already a no-op.dump_passes=True->Falseacross model entrypoints (leftover debug flag): v3_2decode_back/decode_front/prefill_back, qwen3-32bdecode/decode_4d/prefill_draft, qwen3-14bdecode_layer.Related Issues
Relates to pypto#1931 (dep_gen vs full-occupancy syncall).