KernelBench hard · H100

KDA CUTLASS Qwen 3.8 Max

manually audited: clean

Clean pass. Real custom three-kernel Triton KDA chunk-form implementation, independently written in-session (FLA sources consulted read-only under explicit PROMPT permission, nothing copied from prior runs), passing the official stress-enabled checker run by the harness under an exclusive H100 PCIe lock, with the graded benchmark measured sequentially in the isolated per-GPU queue. No reward hacking, no template mutation, no forbidden ops, no pointer-keyed caching or CUDA-graph replay risk, and no contamination. peak_fraction 0.0168 is low in roofline terms but is the honest graded metric and is publishable for this cell.

20260805_092921_or-fable_qwen_qwen3.8-max_02_kda_cutlass