elgato_gpu.pbs

#!/bin/bash

### This will be the researcher who supported your use of the HPC.
#PBS -W group_list=[your researcher here]

### Researchers can buy in for priority queueing. However, the amount of time 
### they can use for this is limited.  Everyone has access to unlimited
### windfall.  However, any priority jobs will go first.  The queues on El Gato
### are windfall, standard, high_pri
#PBS -q [your queue here]

### GPU nodes have 16 cores of standard processors.  When requesting a GPU node,
### it is important to request the whole node.  The request below shows a
### request for a single GPU.  On El Gato, there are more single GPU nodes than
### double gpu nodes.  Requesting single GPU nodes will shorten your queue wait
### time.  With GPU requests, it's important to also configure your software to
### recognize the GPU resources.  This will also inform how many you request.
### If you need help with configuration or deciding what resources to request,
### please reach out to the HPC consult team.
#PBS -l select=1:ncpus=16:mem=250gb:pcmem=16gb:ngpus=1

### This is the amount of time you think your job will take to run.
### Not sure how long it will take?  Request the max of 10 days.
### (240:00:00)                  #PBS -l walltime=240:00:00
### This request shows 5 minutes #PBS -l walltime=00:05:00
#PBS -l walltime=[your time here]

### This will show which node your job is running on.
echo 'This script is running on:'
hostname

### GPU test program
nvidia-smi

### If you have any trouble or questions, please contact:
### hpc-consult@list.arizona.edu