Commit 3974984f authored by Stephan Hageboeck's avatar Stephan Hageboeck
Browse files

Update default cuda architecture.

- Cuda 11 dropped support for -arch=sm_30. The default was therefore updated to sm_35.
- Furthermore, it was impossible to select virtual architectures such as compute_70,
  since sm_ is always prepended to the CMake variable.
  Now, one specifies
    -DCUDA_ARCH=sm_35
  instead of
    -DCUDA_ARCH=35
parent 4208d939
Pipeline #2278799 failed with stage
in 9 minutes and 24 seconds
......@@ -135,8 +135,8 @@ if(CUDA)
set(VECGEOM_CUDA_VOLUME_SPECIALIZATION ON)
endif()
set(CUDA_ARCH 30 CACHE STRING "CUDA device architecture.")
set(CUDA_ARCH "-arch=sm_${CUDA_ARCH}")
set(CUDA_ARCH sm_35 CACHE STRING "CUDA device architecture string that's passed to nvcc -arch=..., e.g. sm_30 compute_70.")
set(CUDA_ARCH "-arch=${CUDA_ARCH}")
set(VecCore_COMPONENTS ${VecCore_COMPONENTS} "CUDA")
endif()
......@@ -174,7 +174,7 @@ endif()
option(USE_NAVINDEX "Use navigation index table and states" OFF)
if(CUDA)
### CUDA=ON requires NAVINDEX
message(STATUS CUDA=ON implies USE_NAVINDEX=ON)
message(STATUS "CUDA=ON implies USE_NAVINDEX=ON")
set(USE_NAVINDEX ON)
endif()
if(USE_NAVINDEX)
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment