Difference between revisions of "Job management with LSF"

From ScientificComputing
Jump to: navigation, search
Line 38: Line 38:
 
Users can access the computing resources on the cluster solely through the batch system. On the ETH HPC clusters, we use the LSF batch system to manage computing jobs.
 
Users can access the computing resources on the cluster solely through the batch system. On the ETH HPC clusters, we use the LSF batch system to manage computing jobs.
  
== What are the steps to use the compute nodes on the cluster? ==
+
== What are the steps? ==
 
* Log in to a login node on the cluster
 
* Log in to a login node on the cluster
 
* Transfer your data to the cluster
 
* Transfer your data to the cluster

Revision as of 10:34, 1 February 2021

250px

     1. Submit a job

     3. Monitor a job

     4. Job output

Why should I use the LSF batch system?

Users can access the computing resources on the cluster solely through the batch system. On the ETH HPC clusters, we use the LSF batch system to manage computing jobs.

What are the steps?

  • Log in to a login node on the cluster
  • Transfer your data to the cluster
  • Load necessary modules
  • Prepare a BSUB command with LSF options which request the computing resources that you need
  • Submit a job with a BSUB command line or a job script
  • Wait for your job to run
  • Your job is run on compute nodes
  • Get your job results and output