CMake failing for gnugpu on Perlmutter-GPU

[xylar]

When I run:

module load cmake

mkdir -p build_omega/build_pm-gpu_gnugpu
cd build_omega/build_pm-gpu_gnugpu

export METIS_ROOT=/pscratch/sd/x/xylar/spack_pm-gpu_test//dev_polaris_0_3_0_gnugpu_mpich/var/spack/environments/dev_polaris_0_3_0_gnugpu_mpich/.spack-env/view
export PARMETIS_ROOT=/pscratch/sd/x/xylar/spack_pm-gpu_test//dev_polaris_0_3_0_gnugpu_mpich/var/spack/environments/dev_polaris_0_3_0_gnugpu_mpich/.spack-env/view

cmake \
   -DOMEGA_BUILD_TYPE=Release \
   -DOMEGA_CIME_COMPILER=gnugpu \
   -DOMEGA_CIME_MACHINE=pm-gpu \
   -DOMEGA_METIS_ROOT=${METIS_ROOT}\
   -DOMEGA_PARMETIS_ROOT=${PARMETIS_ROOT}\
   -DOMEGA_BUILD_TEST=ON \
   -S /global/u2/x/xylar/e3sm_work/polaris/add-omega-ctest-util/e3sm_submodules/Omega/components/omega/ \
   -B . 

I'm seeing:

...
-- Configuring done
CMake Error: Error required internal CMake variable not set, cmake may not be built correctly.
Missing variable is:
_CMAKE_CUDA_WHOLE_FLAG
CMake Error: Error required internal CMake variable not set, cmake may not be built correctly.
Missing variable is:
CMAKE_CUDA_COMPILE_OBJECT
-- Generating done
CMake Generate step failed.  Build files cannot be regenerated correctly.

[xylar]

I'm now able to build CTests but I'm seeing:

no CUDA-capable device is detected

over and over, specifically:

0: terminate called after throwing an instance of 'std::runtime_error'
0:   what():  cudaGetDeviceCount(&m_cudaDevCount) error( cudaErrorNoDevice): no CUDA-capable device is detected /global/u2/x/xylar/e3sm_work/polaris/main/e3sm_submodules/omega/develop/externals/ekat/extern/kokkos/core/src/Cuda/Kokkos_Cuda_Instance.cpp:275

[xylar]

I'm re-testing this now (with --gpus=4).

[xylar]

Yep, with --gpus=4, all the tests pass for me now!

[xylar]

I think this is fixed, so closing.