BatchesΒΆ

Batches provide a way to run and manage multiple jobs together. This can be useful if you are running many related jobs for a design of experiments (DoE) or a parameter sweep, or if you use self-hosted licensing and you can only run one job at a time. Specifically, batches allow you to:

  • Submit multiple jobs at the same time.

  • Specify whether the jobs in a batch will run simultaneously (in parallel) or sequentially (in serial).

  • Specify whether all jobs in a batch should be stopped if one job fails.

  • Monitor the progress and cost of jobs in a batch.

  • Stop all jobs in a batch simultaneously.

You still need to create each job in a batch individually. When you add a job to a batch, the job will not start until the batch is started.

The jobs in a batch run on separate job clusters, so they can run on different hardware configurations and even in different data centers. However, all jobs in a batch must use a file system for storage. Spaces are not supported for batches.