refactored implicit gemm v1r3

This commit is contained in:
Chao Liu
2019-07-29 15:01:01 -05:00
parent 9ba3b49158
commit efd419ecbe
6 changed files with 94 additions and 115 deletions

View File

@@ -13,11 +13,11 @@ cmake
-D CMAKE_BUILD_TYPE=Release \
-D CMAKE_VERBOSE_MAKEFILE:BOOL=ON \
-D DEVICE_BACKEND=NVIDIA \
-D BOOST_ROOT="/package/install/boost_1.67.0" \
-D CUDA_COMMON_INCLUDE_DIR="/home/chao/code/test_feature/cuda_common/cuda_10.0_common/inc" \
-D CMAKE_CUDA_FLAGS="-ccbin clang++ -m64 -Xcompiler -fopenmp -lineinfo --source-in-ptx -keep -Xptxas -v -gencode=arch=compute_61,code=sm_61 -Xptxas -v -Xptxas -v -maxrregcount=128" \
-D CUDA_COMMON_INCLUDE_DIR="/package/install/cuda/10.1/NVIDIA_CUDA-10.1_Samples/common/inc" \
-D CMAKE_CUDA_FLAGS="-ccbin clang++ -m64 -Xcompiler -fopenmp -lineinfo --source-in-ptx -keep -Xptxas -v -gencode=arch=compute_61,code=sm_61" \
${MY_PROJECT_SOURCE}
#-D BOOST_ROOT="/package/install/boost_1.67.0" \
#-D CMAKE_CUDA_COMPILER="/package/install/cuda_10.0/bin/nvcc" \
#-D CMAKE_CUDA_FLAGS="-ccbin clang++ -m64 -Xcompiler -fopenmp -lineinfo --source-in-ptx -keep -Xptxas -v -gencode=arch=compute_61,code=sm_61" \