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 11, 15:00 UTC
[Bugfix] Fix CMakeLists Environment Variable (#21804)
Signed-off-by: wu-kan <github@wu-kan.com> Signed-off-by: 吴坎 <github@wu-kan.cn> Signed-off-by: wu-kan <github@wu-kan.cn>
吴坎 committed
Dec 12, 2025 at 18:54 UTC
91401c7a266450e332e88c3b569e93aeecca9a89
1 file changed
+3
-3
CMakeLists.txt
+3
-3
@@ -384,7 +384,7 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
384
OR NOT $CACHE{MARLIN_GEN_SCRIPT_HASH_AND_ARCH} STREQUAL ${MARLIN_GEN_SCRIPT_HASH_AND_ARCH})
385
execute_process(
386
COMMAND ${CMAKE_COMMAND} -E env
387
- PYTHONPATH=$PYTHONPATH
387
+ PYTHONPATH=$ENV{PYTHONPATH}
388
${Python_EXECUTABLE} ${MARLIN_GEN_SCRIPT} ${CUDA_ARCHS_STR}
389
RESULT_VARIABLE marlin_generation_result
390
OUTPUT_VARIABLE marlin_generation_result
@@ -822,7 +822,7 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
822
OR NOT $CACHE{MACHETE_GEN_SCRIPT_HASH} STREQUAL ${MACHETE_GEN_SCRIPT_HASH})
823
execute_process(
824
COMMAND ${CMAKE_COMMAND} -E env
825
- PYTHONPATH=${CMAKE_CURRENT_SOURCE_DIR}/csrc/cutlass_extensions/:${CUTLASS_DIR}/python/:${VLLM_PYTHON_PATH}:$PYTHONPATH
825
+ PYTHONPATH=${CMAKE_CURRENT_SOURCE_DIR}/csrc/cutlass_extensions/:${CUTLASS_DIR}/python/:${VLLM_PYTHON_PATH}:$ENV{PYTHONPATH}
826
${Python_EXECUTABLE} ${MACHETE_GEN_SCRIPT}
827
RESULT_VARIABLE machete_generation_result
828
OUTPUT_VARIABLE machete_generation_output
@@ -1004,7 +1004,7 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
1004
OR NOT $CACHE{MOE_MARLIN_GEN_SCRIPT_HASH_AND_ARCH} STREQUAL ${MOE_MARLIN_GEN_SCRIPT_HASH_AND_ARCH})
1005
execute_process(
1006
COMMAND ${CMAKE_COMMAND} -E env
1007
- PYTHONPATH=$PYTHONPATH
1007
+ PYTHONPATH=$ENV{PYTHONPATH}
1008
${Python_EXECUTABLE} ${MOE_MARLIN_GEN_SCRIPT} ${CUDA_ARCHS_STR}
1009
RESULT_VARIABLE moe_marlin_generation_result
1010
OUTPUT_VARIABLE moe_marlin_generation_output