permute/unpermute kernel for moe optimization (#14568)

Signed-off-by: Caleb_Du <Caleb_Du@zju.edu.cn>
This commit is contained in:
Caleb_Du
2025-05-02 11:31:55 -07:00
committed by GitHub
parent 0f87d8f7b2
commit 3e887d2e0c
19 changed files with 1474 additions and 28 deletions
+13 -1
View File
@@ -15,7 +15,6 @@ project(vllm_extensions LANGUAGES CXX)
# CUDA by default, can be overridden by using -DVLLM_TARGET_DEVICE=... (used by setup.py)
set(VLLM_TARGET_DEVICE "cuda" CACHE STRING "Target device backend for vLLM")
message(STATUS "Build type: ${CMAKE_BUILD_TYPE}")
message(STATUS "Target device: ${VLLM_TARGET_DEVICE}")
@@ -682,6 +681,17 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
endif()
endif()
if(VLLM_GPU_LANG STREQUAL "CUDA")
set(MOE_PERMUTE_SRC
"csrc/moe/permute_unpermute_kernels/moe_permute_unpermute_kernel.cu"
"csrc/moe/moe_permute_unpermute_op.cu")
set_gencode_flags_for_srcs(
SRCS "${MARLIN_PERMUTE_SRC}"
CUDA_ARCHS "${MOE_PERMUTE_ARCHS}")
list(APPEND VLLM_MOE_EXT_SRC "${MOE_PERMUTE_SRC}")
endif()
message(STATUS "Enabling moe extension.")
define_gpu_extension_target(
_moe_C
@@ -690,6 +700,8 @@ define_gpu_extension_target(
SOURCES ${VLLM_MOE_EXT_SRC}
COMPILE_FLAGS ${VLLM_GPU_FLAGS}
ARCHITECTURES ${VLLM_GPU_ARCHES}
INCLUDE_DIRECTORIES ${CUTLASS_INCLUDE_DIR}
INCLUDE_DIRECTORIES ${CUTLASS_TOOLS_UTIL_INCLUDE_DIR}
USE_SABI 3
WITH_SOABI)