Using GPUs with Slurm: Difference between revisions

tweak header levels and TOC
(change advice from --gres=gpu to --gpus-per-node)
(tweak header levels and TOC)
Line 2: Line 2:
<translate>
<translate>


== Introduction ==
= Introduction =


To request one or more GPUs for a Slurm job, use this form:
To request one or more GPUs for a Slurm job, use this form:
Line 22: Line 22:
For general advice on job scheduling, see [[Running jobs]].
For general advice on job scheduling, see [[Running jobs]].


== Available hardware == <!--T:1-->
= Available hardware = <!--T:1-->
These are the GPUs currently available:
These are the GPUs currently available:


Line 70: Line 70:
please see the specific instructions on the [https://docs.scinet.utoronto.ca/index.php/Mist#Submitting_jobs SciNet web site].
please see the specific instructions on the [https://docs.scinet.utoronto.ca/index.php/Mist#Submitting_jobs SciNet web site].


== Selecting the type of GPU to use == <!--T:16-->
= Selecting the type of GPU to use = <!--T:16-->


<!--T:37-->
<!--T:37-->
Line 80: Line 80:
For example, molecular dynamics code requires high double-precision performance, for which T4 GPUs are not appropriate.
For example, molecular dynamics code requires high double-precision performance, for which T4 GPUs are not appropriate.
In such a case, make sure you include a type specifier.
In such a case, make sure you include a type specifier.
= Examples =


== Single-core job == <!--T:3-->
== Single-core job == <!--T:3-->
Bureaucrats, cc_docs_admin, cc_staff
2,879

edits