mirror of
https://git.datalinker.icu/vllm-project/vllm.git
synced 2025-12-11 04:34:54 +08:00
5985e3427d
Inspired by #5146, this PR improves FP8 quantize kernel by vectorizing data transfer to better utilize memory bandwidth. Microbenchmark shows that this improved kernel can achieve 1.0x-1.5x speedup (especially when hidden size is large). In details, we applied 3 optimizations: - Use inverted scale so that most divisions are changed to multiplications. - Unroll the loop by 4 times to improve ILP. - Use vectorized 4 to transfer data between HBM and SRAM.