This shows you the differences between two versions of the page.
| Both sides previous revision Previous revision | Last revisionBoth sides next revision |
| software:gromacs:farber [2019-08-29 17:22] – [Batch job] anita | software:gromacs:farber [2019-08-29 17:23] – [Batch job] anita |
|---|
| ===== Batch job ===== | ===== Batch job ===== |
| |
| On Farber, the template (''/opt/templates/gridengine/gromacs.qs'' specifies a number of resource options to provide the best performance. The number of slots specified for ''NPROC'' should be a multiple of 20 in conjunction with specifying [[farber:runjobs:queues#farber-exclusive-access|exclusive access]] to allow your job to run on the node(s) required based on the number of slots requested without any other jobs running on those node(s) too. Please remember exclusive access on Farber rounds up NPROC to a multiple of 20. In this case, set the WANT_NPROC variable to the actual core count you want. The script will "load balance" that core count across the N nodes the job receives. By default the template uses double-precision (''md_rund''), however uncommenting ''USE_SINGLE_PRECISION=YES'' will specify the single-precision version (''md_run'') instead. | On Farber, the template (''/opt/templates/gridengine/gromacs.qs'' specifies a number of resource options to provide the best performance. The number of slots specified for ''NPROC'' should be a multiple of 20 in conjunction with specifying [[abstract:farber:runjobs:queues#farber-exclusive-access|exclusive access]] to allow your job to run on the node(s) required based on the number of slots requested without any other jobs running on those node(s) too. Please remember exclusive access on Farber rounds up NPROC to a multiple of 20. In this case, set the WANT_NPROC variable to the actual core count you want. The script will "load balance" that core count across the N nodes the job receives. By default the template uses double-precision (''md_rund''), however uncommenting ''USE_SINGLE_PRECISION=YES'' will specify the single-precision version (''md_run'') instead. |