inference-kernel-whitebox-debug
White-box debug a custom CUDA/CUTLASS kernel producing a WRONG numeric result (over-amplification, NaN, coherence) after black-box bisection is EXHAUSTED: every external lever varied (grid, pipeline depth, layout, dtype), the result deterministic + lever-independent, yet the structure reads correct. Two tracks: (A) an in-kernel operand + accumulator trace (`if constexpr`-gated to survive megakernel pipelining), (B) a standalone reproducer .cu instantiating the EXACT kernel template + TMA descriptors with controlled inputs (all-ones then real-dump) vs a host GEMM. Localizes the defect to operand-load vs MMA vs epilogue vs descriptor. Escalation tier above `inference-kernel-ncu-profile` when H/P are fine but the output is numerically wrong. Triggers on "white-box kernel", "kernel over-amplifies / NaN", "standalone reproducer", "in-kernel trace", "operand dump", "cutlass UMMA bug", or any combination of "white-box / reproducer / in-kernel / operand-trace" with "kernel / cutlass / UMMA / megakernel / vllm".
ソース情報
- リポジトリ
- cfregly/gpu-perf-tune
- ソースの最終更新活動
- 2026年8月16日 20:38
- 検出された SKILL.md の言語
- 英語
- スター
- 4
- フォーク
- 0
インストール方法
デフォルトでは、最初にソースを確認する Prompt が選択されています。直接コマンドに切り替えるか、ローカルコピーをダウンロードすることもできます。
ソースファイルを確認
インストールを決める前に、SKILL.md と SkillsMP に表示されている付属ファイルをお読みください。
SKILL.md を表示中
| name | inference-kernel-whitebox-debug |
|---|---|
| description | White-box debug a custom CUDA/CUTLASS kernel producing a WRONG numeric result (over-amplification, NaN, coherence) after black-box bisection is EXHAUSTED: every external lever varied (grid, pipeline depth, layout, dtype), the result deterministic + lever-independent, yet the structure reads correct. Two tracks: (A) an in-kernel operand + accumulator trace (`if constexpr`-gated to survive megakernel pipelining), (B) a standalone reproducer .cu instantiating the EXACT kernel template + TMA descriptors with controlled inputs (all-ones then real-dump) vs a host GEMM. Localizes the defect to operand-load vs MMA vs epilogue vs descriptor. Escalation tier above `inference-kernel-ncu-profile` when H/P are fine but the output is numerically wrong. Triggers on "white-box kernel", "kernel over-amplifies / NaN", "standalone reproducer", "in-kernel trace", "operand dump", "cutlass UMMA bug", or any combination of "white-box / reproducer / in-kernel / operand-trace" with "kernel / cutlass / UMMA / megakernel / vllm". |