Say that we have two jobs/tools, which will be grouped by the system:
job1|job2
When I try to set specs for a job2, it looks like they are ignored, and only the specs of job1 are used.
In occasions this can be OK, but usually it would lead to errors. For example, I feel like the memory requested should be the addition of the memories of all the jobs in the group, and the number of CPUs requested could be the maximum along the jobs in the group. Avery value has to be processed in a different way, but the point is that right now there is no place where this processing takes place.
Say that we have two jobs/tools, which will be grouped by the system:
When I try to set specs for a job2, it looks like they are ignored, and only the specs of job1 are used. In occasions this can be OK, but usually it would lead to errors. For example, I feel like the memory requested should be the addition of the memories of all the jobs in the group, and the number of CPUs requested could be the maximum along the jobs in the group. Avery value has to be processed in a different way, but the point is that right now there is no place where this processing takes place.