GROMACS version:2020.4
GROMACS modification: no
I did the command: cmake … -DGMX_BUILD_OWN_FFTW=ON -DREGRESSIONTEST_DOWNLOAD=ON
and I get back:
Check for working NVCC/C compiler combination - broken
– /usr/bin/nvcc standard output: ‘’
– /usr/bin/nvcc standard error: 'In file included from /usr/include/cuda_runtime.h:83,
from :
/usr/include/crt/host_config.h:138:2: error: #error – unsupported GNU version! gcc versions later than 8 are not supported!
138 | #error – unsupported GNU version! gcc versions later than 8 are not supported!
| ^~~~~
’
CMake Error at cmake/gmxManageNvccConfig.cmake:182 (message):
NVCC/C++ compiler combination does not seem to be supported. CUDA
frequently does not support the latest versions of the host compiler, so
you might want to try an earlier C++ compiler version and make sure your
CUDA compiler and driver are as recent as possible.
Call Stack (most recent call first):
cmake/gmxManageGPU.cmake:208 (include)
CMakeLists.txt:578 (gmx_gpu_setup)
Please I need help
Thank’s in advance
Ricardo
The version of GCC used to build gromacs needs to be compatible with the version of CUDA you are using. What are the versions of GCC and CUDA you are using?
Very similar issue here: what was the solution to your problem
?
gromacs-2021.1 with
cmake … -DCMAKE_C_COMPILER=gcc-10 -DCMAKE_CXX_COMPILER=g+±10 -DGMX_BUILD_OWN_FFTW=ON -DGMX_GPU=CUDA -DCUDA_TOOLKIT_ROOT_DIR=/usr/lib/cuda -DREGRESSIONTEST_DOWNLOAD=ON
message:
– Found CUDA: /usr/lib/cuda (found suitable version “10.1”, minimum required is “9.0”)
– Adding work-around for issue compiling CUDA code with glibc 2.23 string.h
– Check for working NVCC/C++ compiler combination with nvcc ‘/usr/bin/nvcc’
– Check for working NVCC/C compiler combination - broken
– /usr/bin/nvcc standard output: ‘’
– /usr/bin/nvcc standard error: 'In file included from /usr/include/cuda_runtime.h:83,
from :
/usr/include/crt/host_config.h:138:2: error: #error – unsupported GNU version! gcc versions later than 8 are not supported!
138 | #error – unsupported GNU version! gcc versions later than 8 are not supported!
| ^~~~~
’
CMake Error at cmake/gmxManageNvccConfig.cmake:229 (message):
NVCC/C++ compiler combination does not seem to be supported. CUDA
frequently does not support the latest versions of the host compiler, so
you might want to try an earlier C++ compiler version and make sure your
CUDA compiler and driver are as recent as possible.
Nvidia’s Cuda compiler (nvcc) only works with certain versions of GCC. The solution is to use a compatible version: according to the message, the latest supported is GCC version 8, so installing and using that should fix it.
Regards,
Petter
Thanks so much - it works with gcc/g++ reverted to v 8
Also note that each cuda generation tends to improve compatibility. For example, CUDA 11.3 supports gcc 9.x, you can find the compatibility matrix on the Nvidia install guide. They’ll likely always be behind the most cutting edge release, but they’ve gotten fairly good at staying within a generation or two of gcc trunk.