A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 14, 23:45 UTC
[Misc] Removed undefined cmake variables MOE_PERMUTE_ARCHS (#21262)
Signed-off-by: Yang Chen <yangche@fb.com>
Yang Chen committed
Jul 24, 2025 at 22:54 UTC
85d051f0268b8d835123148096c670955203ae4d
1 file changed
+8
-11
CMakeLists.txt
+8
-11
@@ -768,6 +768,14 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
768
list(APPEND VLLM_MOE_EXT_SRC "csrc/moe/moe_wna16.cu")
769
endif()
770
771
+if(VLLM_GPU_LANG STREQUAL "CUDA")
772
+ set(MOE_PERMUTE_SRC
773
+ "csrc/moe/permute_unpermute_kernels/moe_permute_unpermute_kernel.cu"
774
+ "csrc/moe/moe_permute_unpermute_op.cu")
775
+
776
+ list(APPEND VLLM_MOE_EXT_SRC "${MOE_PERMUTE_SRC}")
777
+endif()
778
+
779
set_gencode_flags_for_srcs(
780
SRCS "${VLLM_MOE_EXT_SRC}"
781
CUDA_ARCHS "${CUDA_ARCHS}")
@@ -836,17 +844,6 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
844
endif()
845
endif()
846
839
-if(VLLM_GPU_LANG STREQUAL "CUDA")
840
- set(MOE_PERMUTE_SRC
841
- "csrc/moe/permute_unpermute_kernels/moe_permute_unpermute_kernel.cu"
842
- "csrc/moe/moe_permute_unpermute_op.cu")
843
-
844
- set_gencode_flags_for_srcs(
845
- SRCS "${MOE_PERMUTE_SRC}"
846
- CUDA_ARCHS "${CUDA_ARCHS}")
847
-
848
- list(APPEND VLLM_MOE_EXT_SRC "${MOE_PERMUTE_SRC}")
849
-endif()
847
message(STATUS "Enabling moe extension.")
848
define_gpu_extension_target(
849
_moe_C