/laue-parallel

Scripts and tools to parallelize coded aperture post-processing via spatial decomposition on ALCF's Theta and Polaris machines

Primary LanguagePython

laue-parallel

Scripts and surrounding infrastructure to run cold in parallel via spatial decomposition on the Polaris supercomputer.

Repository Organization

  • laue_parallel.py performs the parallel computation, and recon_parallel.py reconstructs the individual process outputs from the first script
  • These scripts are controlled by command-line arguments described in the scripts and by the cold config files located in configs
  • analysis contains various scripts that are useful for debugging output and logging files
  • runscripts contains shell scripts with HPC settings embedded to queue debug and individual production runs
  • prod_scripts contains python scripts to generate and run batches of jobs needed for full production runs

Environment Setup

Pip installable requirements can be found in requirements.txt. These are additional requirements added on top of the base anaconda environments provided by ALCF. Cloning the environment is recommended before attempting to add or modify packages. Additional external requirements and github links are described requirements.txt commented lines.