Notes on AWS cloud platform with GPU

ls /contrib/Niki.Zadeh/spack/var/spack/environments/myproject/.spack-env/view/Linux_x86_64/22.9/comm_libs/mpi/bin/
env.sh  mpiCC    mpif77   mpirun      ompi_info    ompi-submit   ortecc      orte-dvm   orterun      orte-top  oshmem_info  shmemfort
mpic++  mpicxx   mpif90   ompi-clean  ompi-ps      ompi-top      orte-clean  orte-info  orte-server  oshcc     oshrun       shmemrun
mpicc   mpiexec  mpifort  ompi-dvm    ompi-server  opal_wrapper  orted       orte-ps    orte-submit  oshfort   shmemcc

First log into the compute node that has GPU attached

[Niki.Zadeh@cadiscretionarygpu-11 ~]$
salloc -N 1
ssh nikizadeh-cadiscretionarygpu-00011-1-0001

Then on compute node [Niki.Zadeh@compute-0001 ~]$

. /contrib/$USER/spack/share/spack/setup-env.sh
export SPACK_USER_CONFIG_PATH=/contrib/$USER/.spack 
spack env activate myproject
which nvfortran
\rm ./gpu_offload_test2d; nvfortran -O2 -mp=gpu -stdpar gpu_offload_test2d.f90 -o gpu_offload_test2d;./gpu_offload_test2d