For complete SLURM documentation, see https://slurm.schedmd.com/. Here we only show simple examples with system-specific instructions.
HAL Slurm Wrapper Suite (Recommended)
Introduction
The HAL Slurm Wrapper Suite was designed to help users use the HAL system easily and efficiently. The current version is "swsuite-v0.1", which includes
srun → swrun : request resources to run interactive jobs.
sbatch → swbatch : request resource to submit a batch script to Slurm.
Usage
- swrun -q <queue_name> -c <cpu_per_gpu> -t <walltime>
- <queue_name> (required) : cpu, gpux1, gpux2, gpux3, gpux4, gpux8, gpux12, gpux16.
- <cpu_per_gpu> (optional) : 12 cpus (default), range from 12 cpus to 36 cpus.
- <walltime> (optional) : 24 hours (default), range from 1 hour to 72 hours.
- example: swrun -q gpux4 -c 36 -t 72 (request a full node: 1x node, x4 node, 144x cpus, 72x hours)
- swbatch <run_script>
- <run_script> (required) : same as original slurm batch.
- <job_name> (required) : job name.
- <output_file> (required) : output file name.
- <error_file> (required) : error file name.
- <queue_name> (required) : cpu, gpux1, gpux2, gpux3, gpux4, gpux8, gpux12, gpux16.
- <cpu_per_gpu> (optional) : 12 cpus (default), range from 12 cpus to 36 cpus.
- <walltime> (optional) : 24 hours (default), range from 1 hour to 72 hours.
- example: swbatch demo.sb
New Job Queues
| Partition Name | Priority | Max Walltime | Nodes Allowed | Min-Max CPUs Per Node Allowed | Min-Max Mem Per Node Allowed | GPU Allowed | Local Scratch | Description |
|---|---|---|---|---|---|---|---|---|
| gpu-debug | high | 4 hrs | 1 | 12-144 | 18-144 GB | 4 | none | |
| gpux1 | normal | 72 hrs | 1 | 12-36 | 18-54 GB | 1 | none | |
| gpux2 | normal | 72 hrs | 1 | 24-72 | 36-108 GB | 2 | none | |
| gpux3 | normal | 72 hrs | 1 | 36-108 | 54-162 GB | 3 | none | |
| gpux4 | normal | 72 hrs | 1 | 48-144 | 72-216 GB | 4 | none | |
| cpu | normal | 72 hrs | 1 | 96-96 | 144-144 GB | 0 | none | |
| gpux8 | low | 72 hrs | 2 | 48-144 | 72-216 GB | 8 | none | |
| gpux12 | low | 72 hrs | 3 | 48-144 | 72-216 GB | 12 | none | |
| gpux16 | low | 72 hrs | 4 | 48-144 | 72-216 GB | 16 | none |
Traditional Job Queues
| Partition Name | Priority | Max Walltime | Min-Max Nodes Allowed | Max CPUs | Max Memory | Local Scratch (GB) | Description |
|---|---|---|---|---|---|---|---|
| debug | high | 4 hrs | 1-1 | 144 | 1.5 | None | designed to access 1 node to run debug job |
| solo | normal | 72 hrs | 1-1 | 144 | 1.5 | None | designed to access 1 node to run sequential and/or parallel job |
| ssd | normal | 72 hrs | 1-1 | 144 | 1.5 | 220 | similar to solo partition with extra local scratch, limited to hal[01-04] |
| batch | low | 72 hrs | 2-16 | 144 | 1.5 | None | designed to access 2-16 nodes (up to 64 GPUs) to run parallel job |
HAL Example Job Scripts
New users should check the example job scripts at "/opt/apps/samples-runscript" and request adequate resources.
| Script Name | Job Type | Partition | Max | Nodes | CPU | GPU | Memory (GB) | Description |
|---|---|---|---|---|---|---|---|---|
| run_debug_00gpu_96cpu_216GB.sh | interactive | debug | 4:00:00 | 1 | 96 | 0 | 144 | submit interactive job, 1 full node for 4 hours CPU only task in "debug" partition |
| run_debug_01gpu_12cpu_18GB.sh | interactive | debug | 4:00:00 | 1 | 12 | 1 | 18 | submit interactive job, 25% of 1 full node for 4 hours GPU task in "debug" partition |
| run_debug_02gpu_24cpu_36GB.sh | interactive | debug | 4:00:00 | 1 | 24 | 2 | 36 | submit interactive job, 50% of 1 full node for 4 hours GPU task in "debug" partition |
| run_debug_04gpu_48cpu_72GB.sh | interactive | debug | 4:00:00 | 1 | 48 | 4 | 72 | submit interactive job, 1 full node for 4 hours GPU task in "debug" partition |
| sub_solo_01node_01gpu_12cpu_18GB.sb | sbatch | solo | 72:00:00 | 1 | 12 | 1 | 18 | submit batch job, 25% of 1 full node for 72 hours GPU task in "solo" partition |
| sub_solo_01node_02gpu_24cpu_36GB.sb | sbatch | solo | 72:00:00 | 1 | 24 | 2 | 36 | submit batch job, 50% of 1 full node for 72 hours GPU task in "solo" partition |
| sub_solo_01node_04gpu_48cpu_72GB.sb | sbatch | solo | 72:00:00 | 1 | 48 | 4 | 72 | submit batch job, 1 full node for 72 hours GPU task in "solo" partition |
| sub_ssd_01node_01gpu_12cpu_18GB.sb | sbatch | ssd | 72:00:00 | 1 | 12 | 1 | 18 | submit batch job, 25% of 1 full node for 72 hours GPU task in "ssd" partition |
| sub_batch_02node_08gpu_96cpu_144GB.sb | sbatch | batch | 72:00:00 | 2 | 96 | 8 | 144 | submit batch job, 2 full nodes for 72 hours GPU task in "batch" partition |
| sub_batch_16node_64gpu_768cpu_1152GB.sb | sbatch | batch | 72:00:00 | 16 | 768 | 64 | 1152 | submit batch job, 16 full nodes for 72 hours GPU task in "batch" partition |
Native SLURM style
Submit Interactive Job with "srun"
srun --partition=debug --pty --nodes=1 \ --ntasks-per-node=12 --cores-per-socket=12 --mem-per-cpu=1500 --gres=gpu:v100:1 \ -t 01:30:00 --wait=0 \ --export=ALL /bin/bash
Submit Batch Job
sbatch [job_script]
Check Job Status
squeue # check all jobs from all users squeue -u [user_name] # check all jobs belong to user_name
Cancel Running Job
scancel [job_id] # cancel job with [job_id]
PBS style
Some PBS commands are supported by SLURM.
Check Node Status
pbsnodes
Check Job Status
qstat -f [job_number]
Check Queue Status
qstat
Delete Job
qdel [job_number]
Submit Batch Job
$ cat test.pbs #!/usr/bin/sh #PBS -N test #PBS -l nodes=1 #PBS -l walltime=10:00 hostname $ qsub test.pbs 107 $ cat test.pbs.o107 hal01.hal.ncsa.illinois.edu