beautypg.com

2 lsf-slurm external scheduler, Section 7.4.2 – HP XC System 2.x Software User Manual

Page 95

background image

additional capabilities at the job level and queue level by allowing the inclusion of several
SLURM options in the LSF command line. Refer to Section 7.4.2.

7.4.2 LSF-SLURM External Scheduler

An important option that can be included in submitting parallel jobs with LSF is the external
scheduler option: The external scheduler option provides application—specific external
scheduling options for jobs capabilities and enables inclusion of several SLURM options in
the LSF command line. For example, this option could be used to submit a job to run one task
per node when you have a resource-intensive job which needs to have sole access to the full
resources of a node. Or, if your job needed particular resources found only on a specific set
of nodes, this option could be used to submit a job to those specific nodes. There are several
options available for use with the external scheduler. Refer to the list in this section.

The format for the external scheduler is:

-ext "SLURM[slurm-arguments]"

slurm-arguments

can consist of one or more of the following

srun

options, separated

by semicolons:

SLURM Arguments

Function

nodes=min[-max]

Minimum and maximum number of nodes allocated to job.
The job allocation will at least contain the minimum number
of nodes.

mincpus=

Specify minimum number of CPUs per node. Default value is 1.

mem=

Specify a minimum amount of real memory of each node.

tmp=

Specify a minimum amount of temporary disk space of each
node.

constraint=

Specify a list of constraints. The list may include multiple
features separated by “

&

” or “

|

”. “

&

” represents AND-ed, “

|

represents OR-ed.

nodelist=

Request a specific list of nodes. The job will at least contain
these nodes. The list may be specified as a comma-separated
list of nodes, or a range of nodes.

exclude=

Requests that a specific list of hosts not be included in
resource allocated to this job. The list may be specified as a
comma-separated list of nodes, or a range of nodes.

contiguous=yes

Request a mandatory contiguous range of nodes.

When this option is added to an LSF command line, it looks like the following:

bsub -n num-procs -ext "SLURM[slurm-arguments]" [bsub-options]

[srun [srun-options]] jobname [job-options]

Refer to the LSF

bsub

command manpage for additional information about using the external

scheduler (

-ext

) option. See the

srun

manpage for more details about the above options

and their arguments.

Consider an HP XC system configuration where

lsfhost.localdomain

is the LSF

execution host and nodes

n[1-10]

are compute nodes in the

lsf

partition. All nodes contain

two processors, providing 20 processors for use by LSF jobs.

Example 7-2 shows one way to submit a parallel job to run on a specific node or nodes.

Using LSF

7-11