Commit 19dc5672 authored by Robert Maynard's avatar Robert Maynard

Correct CUDA fallback to "fermi" when trying to do auto detection.

The logic would properly fall to fermi, but not set the proper NVCC flags
to actually compile targeting the fermi compute architecture.
parent 5df0cf42
......@@ -107,8 +107,11 @@ if (VTKm_Base_FOUND)
"Falling back to fermi, please manually specify if you want something else.")
set(VTKm_CUDA_Architecture "fermi")
elseif(VTKm_CUDA_Architecture STREQUAL "fermi")
#since when we are native we can fail, and fall back to "fermi" these have
#to happen after, and separately of the native check
if(VTKm_CUDA_Architecture STREQUAL "fermi")
set(CUDA_NVCC_FLAGS "${CUDA_NVCC_FLAGS} --generate-code arch=compute_20,code=compute_20")
elseif(VTKm_CUDA_Architecture STREQUAL "kepler")
set(CUDA_NVCC_FLAGS "${CUDA_NVCC_FLAGS} --generate-code arch=compute_30,code=compute_30")
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