rsnt_translations
56,430
edits
Using node-local storage: Difference between revisions
no edit summary
No edit summary |
No edit summary |
||
| Line 68: | Line 68: | ||
or if you are chaining together several Slurm jobs to complete a long calculation. | or if you are chaining together several Slurm jobs to complete a long calculation. | ||
However, it will not preserve the contents of <code>$SLURM_TMPDIR</code> in the case of a node failure. | However, it will not preserve the contents of <code>$SLURM_TMPDIR</code> in the case of a node failure. | ||
See [https://services.criann.fr/en/services/hpc/cluster-myria/guide/signals-sent-by-slurm/ this page] | See [https://services.criann.fr/en/services/hpc/cluster-myria/guide/signals-sent-by-slurm/ this page] from | ||
CRIANN for an example script and detailed guidance. | |||
for an example script and detailed guidance. | |||
= Multinode jobs = <!--T:12--> | = Multinode jobs = <!--T:12--> | ||
| Line 98: | Line 97: | ||
<!--T:22--> | <!--T:22--> | ||
At | At <b>[[Niagara]]</b>, $SLURM_TMPDIR is implemented as <i>RAMdisk</i>, | ||
so the amount of space available is limited by the memory on the node, | so the amount of space available is limited by the memory on the node, | ||
less the amount of RAM used by your application. | less the amount of RAM used by your application. | ||
| Line 104: | Line 103: | ||
<!--T:23--> | <!--T:23--> | ||
At the general-purpose clusters | At the general-purpose clusters, | ||
the amount of space available depends on the cluster and the node to which your job is assigned. | the amount of space available depends on the cluster and the node to which your job is assigned. | ||
| Line 121: | Line 120: | ||
<!--T:25--> | <!--T:25--> | ||
The table above gives the amount of space in $SLURM_TMPDIR on the | The table above gives the amount of space in $SLURM_TMPDIR on the <i>smallest</i> node in each cluster. | ||
If your job reserves [[Advanced_MPI_scheduling#Whole_nodes|whole nodes]] | If your job reserves [[Advanced_MPI_scheduling#Whole_nodes|whole nodes]], | ||
then you can reasonably assume that this much space is available to you in $SLURM_TMPDIR on each node. | then you can reasonably assume that this much space is available to you in $SLURM_TMPDIR on each node. | ||
However, if the job requests less than a whole node, then other jobs may also write to the same filesystem | However, if the job requests less than a whole node, then other jobs may also write to the same filesystem | ||
(but | (but a different directory!), reducing the space available to your job. | ||
<!--T:26--> | <!--T:26--> | ||
Some nodes at each site have more local disk than shown above. | Some nodes at each site have more local disk than shown above. | ||
See | See <i>Node characteristics</i> at the appropriate cluster's page ([[Béluga/en|Béluga]], [[Cedar]], [[Graham]], [[Narval]]) for guidance. | ||
</translate> | </translate> | ||