ViperEkura 2c0b5d0b5e perf: enable MMA decode path for G=1 full attention ...

- inline decode_use_mma() into dispatch_decode()
- drop G>1 guard, MMA works correctly for G>=1
- decode is memory-bound, tensor cores + cp.async still win at G=1

2026-07-11 11:45:13 +08:00

..

kernels

perf: enable MMA decode path for G=1 full attention

2026-07-11 11:45:13 +08:00

tests

perf: increase decode split-K parallelism for short sequences

2026-07-11 11:26:42 +08:00

__init__.py

feat: add optional CUDA kernel system (csrc/) + fused GQA decode attention

2026-07-06 12:09:58 +08:00

build.py

perf: increase decode split-K parallelism for short sequences

2026-07-11 11:26:42 +08:00