HPC/Applications/lammps/Package GPU: Difference between revisions

From CNM Wiki
< HPC‎ | Applications‎ | lammps
Jump to navigation Jump to search
Line 23: Line 23:
  …
  …
  pair_style      lj/charmm/coul/long'''/gpu''' 8.0 10.0
  pair_style      lj/charmm/coul/long'''/gpu''' 8.0 10.0
Alternatively, use the following command line options:
  lmp_mpi-gpu '''''-sf gpu -pk gpu 1''''' -in ''file.in'' > ''file.out''


=== Job file example ===
=== Job file example ===

Revision as of 18:11, March 23, 2018

Package GPU

  • Provides multi-threaded versions of most pair styles, all dihedral styles and a few fixes in LAMMPS; for the full list:
    1. In your browser, open http://lammps.sandia.gov/doc/Section_commands.html#comm
    2. Search for the string /cuda.
  • Supports one physical GPU per LAMMPS MPI process (CPU core).
  • Multiple MPI processes (CPU cores) can share a single GPU, and in many cases it will be more efficient to run this way.

Usage

  1. Use the command package gpu mode first last split near the beginning of your LAMMPS control script. Since all Carbon GPU nodes have just one GPU per node, the arguments first and last must always be zero; split is not restricted.
  2. Do one of the following:
  3. In the job file or qsub command line, request a GPU #PBS -l nodes=...:gpus=1 (referring to the number of GPUs per node).
  4. Call the lmp_openmpi-gpu binary.

Input file examples

package gpu force 0 0 1.0
package gpu force 0 0 0.75
package gpu force/neigh 0 0 1.0
package gpu force/neigh 0 1 -1.0
…
pair_style      lj/charmm/coul/long/gpu 8.0 10.0

Alternatively, use the following command line options:

 lmp_mpi-gpu -sf gpu -pk gpu 1 -in file.in > file.out

Job file example

#PBS -l nodes=...:gpus=1
…
mpirun … lmp_openmpi-gpu -in infile