Use this table to find out how to convert basic Auger tags to slurm options. There is no convertion from INPUT/OUTPUT related Auger tags to slurm option. For production jobs we recommend using swif2, which provides the job flow management, job grouping and many more advanced functions.
Auger Tag | Slurm option example | Note |
PROJECT | --account=halld | run job using halld account (click here to find all user's accounts) |
TRACK | --partition=production | use production partition (click here to find available partations) |
OS | --constraint=general | use any OS nodes |
NODE_TAG | --constraint=farm18 | use farm18 nodes only (click here to find all avalible node features) |
JOBNAME | --job-name=test_slurm | |
TIME | --time=2:00:00 | 2 hours walltime |
MEMORY | --mem-per-cpu=250 | 4GB memory (250MBx16) |
CPU | - -ntasks=16 | 16 cores |
DISK_SPACE | --gres=disk:1000 | 1000MB (1GB) disk allocation |
--mail-user=<username>@jlab.org | ||
--output=/farm_out/%u/%x-%j-%N.out | output file: /farm_out/<userName> /<jobName>-<jobId>-<hostname>.out | |
--error=/farm_out/%u/%x-%j-%N.err | error file: /farm_out/<userName> /<jobName>-<jobId>-<hostname>.err | |
INPUT_FILES | ||
OUTPUT_FILES | ||
SINGLE_JOBS | ||
MUITI_JOBS | ||
INPUT_DATA | ||
OUTPUT_DATA | ||
OUTPUT_TEMPLATE |