[slurm-users] Re: Scheduling issues with multiple different types of GPU in one partition

Kevin M. Hildebrand via slurm-users wrote:

bf_max_job_test is the default of 500. However sdiag says the number of times bf_max_job_test has been hit is zero, so that's probably not relevant.

Du you have weights defined? I've seen the scheduler insisting on waiting for nodes with lower weights when higher weighted ones were idling. An squeue --start will tell.

Gerhard