1. Home
  2. /
  3. Documentation
  4. /
  5. Job Submission
  6. /
  7. HTCondor Batch System
  8. /
  9. Introduction

Introduction

HTCondor is a modern batch system which replaces LSF, PBS or Torque systems on many computing sites. We installed our computing capacities to this system during the first months of 2018.

There are many suitable tutorials and Howto documents about HTCondor, see Useful Links to HTCondor Documentation. The most useful information on HTCondor job submission can be found in HTCondor Batch at farm.particle.cz presented by Petr Vokac at Golias Farm Users Meeting in February 2018.

Job submission can be done from our standard user Interfaces (ui.farm.particle.cz).

Default limits for job resources within HTCondor:

  • walltime = 1200 s
  • CPU cores = 1
  • memory = 2 GB (per one core)
  • scratch = 20 GB (per one core)
  • universe = vanilla (an execution environment for a normal user job)

Default max limits for job resources within HTCondor:

  • max walltime = 2 weeks
  • max CPU cores = 128
  • max memory = 512 GB (per job)
  • max scratch = 1 TB (per job)

Be aware that
– user HOME directory is not set for jobs by default in HTCondor
– no swap is allowed for jobs
– job disk space is fully isolated, jobs do not see other job disk spaces

Přejít nahoru