The C4 cluster is high-performance compute resource used by a UCSF cancer researchers. It is designed to handle a large number of analyses run in parallel by a large number of groups and users. Despite this, there will times when the demand is higher than the supply, resulting in longer queueing times and possibly also slower processing times when the jobs are started.
In order to maximize the efficiency of the cluster and decrease your (and others) queuing and processing times, please try make yourself familiar with the following guidelines:
Identify and specify the amount of resources your jobs need (⇒ better utilization and load balancing)
Make your script agile to Slurm resources allocated to your job (⇒ scalable and lower risk of using resources not requested)
Work toward local scratch on the compute nodes (⇒ faster processing and less load on shared disk I/O)