There are two problems here:
1. You're not allowed to submit a job with minimum cores larger than the cluster available cores. This limit can be removed with a setting from a recent
Update 3 QFE:
Set-HPCClusterProperty -DisableResourceValidation $newValue
2. You can
oversubscribe the system with more cores as you needed so that you have more jobs in running state, for example, if you want 100 jobs each asking 1 core, you can oversubscribe your node to 25 cores from 4 cores.
Qiufang Shi