diff --git a/libensemble/tests/scaling_tests/forces/readme.md b/libensemble/tests/scaling_tests/forces/readme.md index e9e5499b1..27ad70fcb 100644 --- a/libensemble/tests/scaling_tests/forces/readme.md +++ b/libensemble/tests/scaling_tests/forces/readme.md @@ -26,21 +26,42 @@ To run forces as a standalone executable on N procs: A random sample of seeds is taken and used as imput to the sim func (forces miniapp). -Modify build_forces.sh for target platform and run to build forces.x +Modify build_forces.sh for target platform and run to build forces.x: ./build_forces.sh -To run with one manager and N-1 workers: +Then to run with default local comms (multiprocessing) with one manager and `N` workers: + + python run_libe_forces.py N + +To run with MPI, set `USE_MPI = True` in `run_libe_forces.py`. +Then to run with one manager and `N-1` workers: mpirun -np N python run_libe_forces.py - -Application parameters can be adjusted in the file run_libe_forces.py. + +Application parameters can be adjusted in the file `run_libe_forces.py`. To remove output before the next run: ./cleanup.sh +### Running with Balsam + +To run with balsam, set `USE_MPI = True` and `USE_BALSAM = True` in `run_libe_forces.py`. +You need to have followed the instructions to install `balsam` and set-up/activate a database. +(See https://github.com/balsam-alcf/balsam). + +Then to test locally, run the `balsam_local.sh` script. The default runs with 2 workers. + + ./balsam_local.sh + +The running jobs can be seen inside the balsam database dir `