Job Submission

Running calculations on Hydra requires submitting a job to the job scheduler, which distributes resources in the cluster and organizes the execution of jobs in an orderly manner.

Hydra uses the Slurm Workload Manager as its job scheduler. VSCdocRunning Jobs has an extensive introduction to job scripting, submission and monitoring, as well as links to more advance topics.

The sections below cover specific use cases targeting the VUB clusters, from basic execution of jobs in parallel or on the GPUs to detailed instructions for specific scientific software applications available in the cluster.

• 1. Slurm Job Scheduler
  • 1.1. Command line tools
  • 1.2. Torque/Moab to Slurm migration
  • 1.3. CPU cores allocation
  • 1.4. Memory allocation
  • 1.5. Slurm partitions
  • 1.6. Job working directory
• 2. Job Queues
  • 2.1. Main job queue
  • 2.2. Test/Debug queue
• 3. Main Job Types
  • 3.1. Serial jobs
  • 3.2. Jobs for GPUs
  • 3.3. Parallel non-MPI jobs
  • 3.4. Parallel MPI jobs
  • 3.5. Hybrid MPI-threading jobs
  • 3.6. Task farming
  • 3.7. Job arrays
  • 3.8. Interactive jobs
  • 3.9. Data management
• 4. GPU Job Types
  • 4.1. GPU generation
  • 4.2. Memory settings of GPU jobs
  • 4.3. Single GPU jobs
  • 4.4. Multi GPU jobs
  • 4.5. Advanced: task distribution in GPUs
• 5. Job Checkpoints and Restarts
  • 5.1. Alternatives to checkpoints
  • 5.2. Checkpoints with DMTCP