KernelBench hard · RTX PRO 6000

FP8 GEMM Kimi K3 (256k)

manually audited: clean

Genuine persistent-TMA fp8 Triton GEMM for SM100: tl.make_tensor_descriptor loads/stores (zero-filled OOB gives exact tile-tail predication), fp8 x fp8 tl.dot lowering to tcgen05 MMA with fp32 accumulate, per-output-channel dequant scale in the fp32 epilogue, persistent grid with grouped tile order and warp specialization. Odd-K (4127) handled by a fused pad kernel copying x and w into 128-aligned buffers (zero padding contributes exactly 0 to the dot). Per-shape configs are pinned from a deterministic sweep. CUDA graphs keyed on (data_ptr, strides, shape) hide Python launch overhead; replay was empirically verified to recompute on live buffer contents, not return stale outputs. No torch._scaled_mm, no output caching, no grader tampering.

20260719_030522_kinetic-claude_kinetic-0715_01_fp8_gemm