This contains:
- water.inp: water cluster input for the single-GPU/single-tile case
To adjust for more devices,
ngroup
should be changed to equal the number of GPUs or tiles to run on. - run_aurora.sh: runscript for Aurora for the single-GPU/single-tile case For more devices, the number of MPI ranks launched should be adjusted
- run_polaris.sh: runscript for Polaris for the single-GPU/single-tile case For more devices, the number of MPI ranks launched should be adjusted