Input file generation for weak-scaling test of HemoCell

Weak-scaling analysis

The weakscaling.py command-line tool can generate a directory populated with submit scripts and simulation configuration files based on provided template files and cluster specifications. For instance

./weakscaling.py --cluster archer2 --slurm ../templates/archer.job -v

populates ./batch with a series of weak-scaling-* directories matching the provided node counts set in weakscaling.py. The script is not fully automatic, i.e. setting node ranges, simulation parameters, and other configurations are to be set manually. However, after that, all template replacements and POS-file generations are performed automatically.

Ultimately all batch/weak-scaling-*/slurm.job scripts can be submitted as jobs to the compute cluster.

Post-processing weak-scaling analysis

The postprocessing.py script extracts iterate times from timing files generated by HemoCell. It handles both the traditional space-based indented file format as well as JSON-based timing logs.