Open ramadugu opened 8 months ago
It should work but I have to update my compiler. Beware that gipaw is not GPU-aware yet.
Hello, following up on this question, can there still be speed ups achieved, if GIPAW uses CUDA accelerated subroutines like fftw or, wont this work?
No idea, I'll try
Awesome, thank you very much!
I am trying to compile QE 7.2 with CMAKE and adding the -DQE_ENABLE_PLUGINS="gipaw" -DQE_ENABLE_FOX=ON but that fails. My question is, does gipaw plugin compile with NVHPC version 23.5? I can provide more details if that helps in determining what might be causing the issue for me.
CMAKE: cmake -DCMAKE_C_COMPILER=mpicc -DCMAKE_Fortran_COMPILER=mpif90 -DQE_ENABLE_CUDA=ON -DQE_ENABLE_MPI_GPU_AWARE=ON -DNVFORTRAN_CUDA_CC=80 -DOpenMP_C_FLAGS=-fopenmp=lomp -DOpenMP_C_LIB_NAMES="libomp" -DOpenMP_libomp_LIBRARY="/hpc/software/spack_v20d1/spack/opt/spack/linux-rhel7-x86_64/gcc-10.4.0/nvhpc-23.3-exveznevn5vxtasxdjnusl262dfdamvf/Linux_x86_64/23.3/compilers/lib/libomp.so" -DQE_ENABLE_OPENMP=ON -DCMAKE_BUILD_TYPE:STRING=RELWITHDEBINFO -DQE_ENABLE_OPENACC=ON -DQE_ENABLE_FOX=ON -DQE_ENABLE_PLUGINS="gipaw" -DCMAKE_Fortran_FLAGS="-fast -O3" ../
Thanks, Sai