vllm/csrc/punica/bgmv/bgmv_fp16_fp16_fp16.cu

#include "bgmv_config.h"
#include "bgmv_impl.cuh"

FOR_BGMV_WIDE_NARROW(INST_BGMV_TWOSIDE, nv_half, nv_half, nv_half)
FOR_INST_BGMV_WIDE_NARROW(INST_BGMV_ONESIDE, nv_half, nv_half, nv_half)
Speed up Punica compilation (#2632) 2024-01-28 09:46:56 +08:00			`#include "bgmv_config.h"`
			`#include "bgmv_impl.cuh"`

			`FOR_BGMV_WIDE_NARROW(INST_BGMV_TWOSIDE, nv_half, nv_half, nv_half)`
[Kernel] Full Tensor Parallelism for LoRA Layers (#3524) Co-authored-by: Antoni Baum <antoni.baum@protonmail.com> 2024-04-27 15:03:48 +08:00			`FOR_INST_BGMV_WIDE_NARROW(INST_BGMV_ONESIDE, nv_half, nv_half, nv_half)`