You can also use the newer squeue -O option to look at TRES for the job. Here's an example of a job that requested one node and two CPUs, with --mem-per-cpu=1750M:
# squeue --noheader -O tres -j 109296_1 cpu=2,mem=3500,node= On Wed, Apr 13, 2016 at 6:34 PM, Lyn Gerner <[email protected]> wrote: > Hi Felix, > > Memory is one of the job parameters that is not available automatically in > an environment variable. You could have your user run the following command > inside the job script (which takes advantage of the extreme configurability > of squeue output formatting), but they will need to understand whether > they've requested memory per node or per CPU. In this example, I submitted > a single-core job w/memory per node of 2G (--mem=2G): > > # HeapMem=$(squeue --noheader -j $SLURM_JOB_ID --format=%m) > > # echo $HeapMem > > 2G > > > Regards, > > Lyn > > On Mon, Apr 11, 2016 at 5:13 AM, Felip Moll <[email protected]> wrote: > >> I have a user that needs to gather some submission parameters of his job >> after it's submitted, for example using an environment variable like it is >> done in other cases like: SLURM_TASKS_PER_NODE, SLURM_JOB_NODELIST, etc. >> >> Specifically, I need to get the total memory of the job. >> >> What do you suggest me to do to set the environment with this info? >> >> He need this information to configure on run-time the heap, etc. >> >> Regards, >> Felip M >> >> >> *--Felip Moll Marquès* >> Computer Science Engineer >> E-Mail - [email protected] >> WebPage - http://lipix.ciutadella.es >> > >
