run_many

This routine runs a set of models.

Each model is run in single processor mode, so the parallelism here is job parallel.

Command line usage (if any):

usage:

run_many.py -jobs x -np y filename

-njobs

is the number of jobs to run simultaneously

-np

is the number of processors per job

filename

is a list of the .pf files to run

Description:

Primary routines:

doit - oversees multiprocesing of the jobs run_one - runs each job steer - parses the command line

Notes

At present, there is no checking that the various .pf files do not drop into interactive mode, or that the routine stalls. This really should be added. Also, we don’t currently gather information on whether the runs succeeded.

History:

190319 ksl Coding begun

Functions

doit([filename, nj, np, outputfile])	Do something magnificent
get_no_jobs(jobs)	Check how many jobs are running
read_file(filename[, char])	Read a file and split it into words, eliminating comments
run_one(filename[, command])	Run and verify a single model
steer(argv)	Process the command line