zhouyuan commented on code in PR #9333:
URL: https://github.com/apache/incubator-gluten/pull/9333#discussion_r2050324397
##########
ep/build-velox/src/build_velox.sh:
##########
@@ -128,7 +128,8 @@ function compile {
COMPILE_OPTION="$COMPILE_OPTION -DVELOX_ENABLE_BENCHMARKS=ON"
fi
if [ $ENABLE_GPU == "ON" ]; then
- COMPILE_OPTION="$COMPILE_OPTION -DVELOX_ENABLE_GPU=ON
-DVELOX_ENABLE_CUDF=ON"
+ # the cuda default options are for Centos9 image from Meta
+ COMPILE_OPTION="$COMPILE_OPTION -DVELOX_ENABLE_GPU=ON
-DVELOX_ENABLE_CUDF=ON -DCMAKE_CUDA_ARCHITECTURES=70
-DCMAKE_CUDA_COMPILER=/usr/local/cuda-12.8/bin/nvcc"
Review Comment:
In my tests it will report `cuda identifier is undefined in device code` if
not setting these two variables. Note the GHA runner does not have NV GPU
installed.
`CUDA_ARCHITECTURES`is able to read from `CMAKE_CUDA_ARCHITECTURES`
https://cmake.org/cmake/help/v3.28/prop_tgt/CUDA_ARCHITECTURES.html
however it seems there's no similar variable like `CUDA_COMPILERS`
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
