Running jobs on M3

Launching jobs on M3 is controlled by SLURM, the Slurm Workload Manager, which allocates the compute nodes, resources, and time requested by the user through command line options and batch scripts. Submitting and running script jobs on the cluster is a straight forward procedure with 3 basic steps:

• Setup and launch

• Monitor progress

• Retrieve and analyse results

For more details please see:

• Partitions on M3
  • Compute partitions
  • GPU partitions
  • Restricted partitions
• Checking the status of M3
  • The STATUS field explained
• Slurm Accounts
  • Default accounts
  • Setting the account for a job
  • Questions about slurm accounts
• Getting started with job submission scripts
• Running Simple Batch Jobs
  • An example Slurm job script
  • Cancelling jobs
• MPI on M3
• On CentOS 7
  • Software
  • Load the MPI compiler
  • Compile the software
  • Running the software
• On Rocky 9
• UCX_NET_DEVICES
• Running Multi-threading Jobs
  • An example Slurm Multi-threading job script
• Running Interactive Jobs
  • Submitting an Interactive Job
  • How long do interactive jobs last for?
  • Reconnecting to/Disconnecting from an Active Interactive Job
• Running GPU Jobs
  • Running GPU Batch Jobs
  • Compiling your own CUDA or OpenCL codes for use on M3
• Running Array Jobs
  • An example of Slurm Array job script
• QoS (Quality of Service)
  • How to run jobs with QoS
  • Explanation
• Features & Constraints
  • An example of using constraint in the job script
  • Features Available
• Checking job status
  • Method 1: show_job
  • Method 2: Slurm commands
• Project Allocation
  • Project Space
  • Questions about allocation
• Diagnosing problems with jobs
  • QOSMaxGRESPerUser
  • Pending jobs
  • CPU, Memory, and Desktop Job Limits
  • Job Submission rejected