An examples directory is available under the elegant directory in the source code. User can run a simulation with given lattice and input files according to the version of the MPI implementation.
For example, we can choose the Pelegant_ringTracking1 example to run the parallel elegant on 11 processors (10 working processors) with the following commands: 1
1) For MPI-1, we can use the following syntax:
mpirun -np 11 Pelegant manyParticles_p.ele
2) For MPI-2, mpiexec is strongly encouraged to start MPI programs, e.g.:
mpiexec -np 11 Pelegant manyParticles_p.ele
One can run another simulation with the serial version of elegant for comparison at the same time:
elegant manyParticles_s.ele
The contents of the input files are same, while ``_p'' and ``_s'' are corresponding to the input file of Pelegant and elegant respectively. User should check the elegant manual to prepare the input file.
In principle, user can run simulations on any number of processors. We have tested the program with more than 30,000 cores on the Intrepid (Bluegene/p) supercomputer at Argonne National Lab. While the number of processors can not be more than the number of particles to track for most of simulations, as it will not use the resource efficiently. For certain simulations, such as parallel optimizations, the number of CPUs is limited to the number of scenarios to be simulated for a parameter set.