The way I implemented separable compilation it doesn't use nvcc to link.
It relies on the host compiler to do the linking. I'm not sure at first
blush why you are having this problem. One thing that should probably be
done is to change this line of code in FindCUDA.cmake:
# For now we are i
On 12/16/2013 04:25 PM, MARTIN, ROBERT S CTR USAF AFMC AFRL/RQRS wrote:
> Nvlink fatal : Unsupported file type '/usr/lib/openmpi/lib/libmpi_cxx.so'
>
> If I replace the '/usr/lib/openmpi/lib/libmpi_cxx.so' with
> '-L/usr/lib/openmpi/lib/ -lmpi_cxx', it works. When I look at the ldd
> output, it
I had similar issues using cmake to build for android, in that it
wanted to use full path to libraries, which were not the same as the
target device.
I used a macro to replace 'target_link_libraries' with this macro
'my_target_link_libraries'; would have to replace the AND __ANDROID__
with a speci
CMake Group-
I'm trying to build a CUDA project with NVCC. The project uses 'separate
compilation'. Unfortunately, 'separate compilation' breaks in a variety of
ways when I try to use the version in FindCUDA.cmake. First, it doesn't
include CUDA_NVCC_FLAGS in the -dlink phase so it immediately
