HPC Runner

hpc-runner provides a unified front end for submitting jobs to HPC schedulers (SGE, Slurm, PBS) and running them locally for testing. It abstracts scheduler-specific flags behind a consistent CLI and Python API, and uses Environment Modules to ensure reproducible tool environments across users.

Contents:

• Getting Started
  • Install
  • Quick start: submit a job
  • Pick a scheduler
  • First config
• Configuration
  • Where configuration is loaded from
  • Top-level sections
  • How jobs pick up config
  • SGE mapping notes
  • Environment variables
  • Tool option specialisation
  • Example: fully populated config (standalone file)
• CLI
  • Global options
  • hpc run
  • submit (config-driven daily driver)
  • hpc status
  • Other commands
• SGE Guide
  • Queues (default queue)
  • CPU / slots (parallel environment)
  • Memory and time resources
  • Custom resources
  • Native qsub passthrough
• Programmatic API
  • Installation
  • Quick start
  • Creating jobs
  • Submitting jobs
  • Monitoring jobs
  • Configuration-driven jobs
  • Job dependencies
  • Pipelines (multi-step workflows)
  • Array jobs