These example scripts can be run with the USER-CUDA package, assuming you built LAMMPS with the package and the precision you want.

Note that these benchmark problems are identical to those in the examples/cuda directory which use the USER-CUDA package.

You can run any of the scripts as follows. You can also reset the x,y,z variables in the command line to change the size of the problem.

With the USER-CUDA package on 1 GPU:

lmp_machine -c on -sf cuda < in.cuda.melt.2.5 lmp_machine -c on -sf cuda -v x 6 -v y 6 -v z 6 < in.cuda.phosphate

With the USER-CUDA package on 2 GPUs:

mpirun -np 2 lmp_machine -c on -sf cuda -pk cuda 2 < in.cuda.melt.2.5 mpirun -np 2 lmp_machine -c on -sf cuda -pk cuda 2 < in.cuda.rhodo

CPU-only:

lmp_machine in.cuda.melt.2.5 mpirun -np 4 lmp_machine < in.cuda.melt.5.0 mpirun -np 8 lmp_machine -v x 1 -v y 1 -v z 2 < in.cuda.rhodo

Note that with the USER-CUDA package you must insure the number of MPI tasks equals the number of GPUs (both per node).