is it possible to increase the priority of a previously scheduled job, so that it’d be run immediately after a GPU is free?
I know that you can set a per-user priority, but per-job – haven’t seen so far.
The use-case is following: I am running a moderate “grid screen” of parameters for one job, and at the same time e.g. cleaning particles with roudns of 2D classification. I’d like to keep cleaning, without having to re-set the screen again. Is it possible?