ViperEkura 74e694921c perf: speed up fp8 gemm tiles and scheduling ...

- K tile 32->64 (new default): fewer barriers, more MMA per stage; generalize tile_at swizzle and load_operand_tile accordingly
- 64x128 small-M CTA for m<=64 (2x at 64x4096x4096)
- L2 rasterization for crosswise-A layouts (+6..21%)
- micro-bench: NT 4096^3 +35%; linear fwd 1.24-1.76x, bwd 1.71-2.27x vs bf16
- add csrc/tests/fp8_test.cu (single MMA demo + GEMM layouts x K-tiles vs CPU reference)

2026-08-24 18:29:55 +08:00

..

common.h

refactor: harden param PODs and CUTLASS-style fp8 layout tags

2026-08-24 15:27:52 +08:00

gemm.cuh

perf: speed up fp8 gemm tiles and scheduling

2026-08-24 18:29:55 +08:00

ops.cu

refactor: harden param PODs and CUTLASS-style fp8 layout tags

2026-08-24 15:27:52 +08:00