...
This indicates a mismatch between the resources specified to srun and those specified to SLURM, e.g. the number of tasks executed does not match the number of cores allocated. Make sure in your job script the resources match both in the SLURM request (i.e., SBATCH directives ) and srun command in your job script.in the flags to the srun command, e.g. if you need 24 cores:
| Code Block |
|---|
#!/bin/bash -l
[..]
#SBATCH --ntasks=24
[..]
srun --export-all -n 24 <..> |
Related articles
| Content by Label |
|---|
| showLabels | false |
|---|
| max | 5 |
|---|
| spaces | com.atlassian.confluence.content.render.xhtml.model.resource.identifiers.SpaceResourceIdentifier@a9e |
|---|
| showSpace | false |
|---|
| sort | modified |
|---|
| reverse | true |
|---|
| type | page |
|---|
| cql | label in ("slurm","job_step","more_processors_requested","srun") and type = "page" and space = "US" |
|---|
| labels | slurm srun more_processors_requested job_step |
|---|
|
...