After a Job has finished running, it can be relaunched on different compute type. This allows you to write code on a lower level machine and then test it with different GPU tiers. Data in Persistent Storage will be available on clones as long as you stay within the same storage region.
Storage regions are as follows:
- East Coast (NY2) - GPU+, P5000, C2, C7
- GCP West - K80, P100, TPU
- In the console, in the Job section, click on the Job you want to rerun:
- Enter your New Job Name and click ReRun and Select the VM type from the dropdown:
- Job is Running on new Machine Type:
- You can clone to a different region type but anything you have in Persistent Storage (/storage) will not carry over.