SLURM-new - Revision history

GiulioFontana: /* Overall resources available to a QOS */

2026-04-27T14:34:57Z

Overall resources available to a QOS

← Older revision		Revision as of 14:34, 27 April 2026
Line 93:		Line 93:

	<pre style="color: lightgrey; background: black;">		<pre style="color: lightgrey; background: black;">
	sacctmgr list qos format=name%-11,grpTRES%-34		sacctmgr list qos format=name%-11,grpTRES%-34 \| grep -Ev "normal\|r-limits\|s-limits"
	</pre>		</pre>

Line 101:		Line 101:
	Name GrpTRES		Name GrpTRES
	----------- ----------------------------------		----------- ----------------------------------
	~~normal~~
	nogpu cpu=48,mem=384G		nogpu cpu=48,mem=384G
	gpuheavy-20 cpu=56,gres/gpu:4g.20gb=4,mem=896G		gpuheavy-20 cpu=56,gres/gpu:4g.20gb=4,mem=896G

GiulioFontana: /* SLURM Quality of Service (QOS) */

2026-04-27T14:33:44Z

SLURM Quality of Service (QOS)

← Older revision		Revision as of 14:33, 27 April 2026
Line 87:		Line 87:

	See [[Singularity#Building Singularity images\|Building Singularity images]] for directions about building Singularity container images.		See [[Singularity#Building Singularity images\|Building Singularity images]] for directions about building Singularity container images.

	~~== Restricted QOSes ==~~

	Some of the QOSes are not available to M.Sc. students. See [[#Research users and students users\|Research users and students users]] to understand the differences between user categories in Mufasa and to find out what access you have to system resources.

	== Overall resources available to a QOS ==		== Overall resources available to a QOS ==

GiulioFontana: /* SLURM partitions */

2026-04-27T14:32:36Z

SLURM partitions

← Older revision		Revision as of 14:32, 27 April 2026
Line 155:		Line 155:
	:: maximum duration of a job run on the partition, in format ''[days-]hours:minutes:seconds''		:: maximum duration of a job run on the partition, in format ''[days-]hours:minutes:seconds''

			See [[#Default values\|Default values]] to get the full list of the features of a partition.

			Partitions are used to limit the set of QOSes that jobs run on that partition can use.

			What partition is accessible to user <code><user></code> and what QOSes are associated to such partition can be determined with command

			<pre style="color: lightgrey; background: black;">
			sacctmgr show assoc where user=<user> format=partition,qos -P
			</pre>

	~~See [[#Default values~~\|~~Default values]] to get the full list of the features of a partition.~~		which provides an output similar to this:

			<pre style="color: lightgrey; background: black;">
			Partition\|QOS
			r-jobs\|build,gpu,gpuheavy-20,gpuheavy-40,gpulight,gpuwide,nogpu
			</pre>

	== QOSes used to set limits to job numbers ==		== QOSes used to set limits to job numbers ==

GiulioFontana: /* QOSes used to set limits to job numbers */

2026-04-27T14:28:50Z

QOSes used to set limits to job numbers

← Older revision		Revision as of 14:28, 27 April 2026
Line 186:		Line 186:
	:; MaxJobsPU		:; MaxJobsPU
	:: maximum number of jobs from a single user that can be either running or queued		:: maximum number of jobs from a single user that can be either running or queued

	~~The list of QOSes associated to the partition(s) accessible to user <code><user></code> can be inspected with command~~

	~~<pre style="color: lightgrey; background: black;">~~
	~~sacctmgr show assoc where user=<user> format=partition,qos -P~~
	~~</pre>~~

	~~which provides an output similar to this:~~

	~~<pre style="color: lightgrey; background: black;">~~
	~~Partition\|QOS~~
	~~r-jobs\|build,gpu,gpuheavy-20,gpuheavy-40,gpulight,gpuwide,nogpu~~
	~~</pre>~~

	== Partition availability ==		== Partition availability ==

GiulioFontana: /* QOSes associated to partitions */

2026-04-27T14:26:48Z

QOSes associated to partitions

← Older revision		Revision as of 14:26, 27 April 2026
Line 158:		Line 158:
	See [[#Default values\|Default values]] to get the full list of the features of a partition.		See [[#Default values\|Default values]] to get the full list of the features of a partition.

	== QOSes ~~associated~~ to ~~partitions~~ ==		== QOSes used to set limits to job numbers ==

	In Mufasa 2.0, two special QOSes are used to set ~~the~~ limits to:		In Mufasa 2.0, two special QOSes called <code>r-limits</code> and <code>s-limits</code> are used to set limits to:
	* the number of jobs by the same user that can be running in the system at any given time		* the number of jobs by the same user that can be running in the system at any given time
	* the number of jobs by the same user that can be either running or queued (both types of jobs count) at any given time		* the number of jobs by the same user that can be either running or queued (both types of jobs count) at any given time

	~~These~~ special QOSes are different ~~for researchers~~ and students, and ~~they are liste~~		The special QOSes are two because they set different limits to [[#Research users and students users\|Research users and students users]]. Their features can be inspected with command

	~~associate limits on the number of jobs that a single user can to partitions~~ <~~code~~>r-~~jobs</code> and <code>~~s-~~jobs~~</~~code~~> ~~limits~~		<pre style="color: lightgrey; background: black;">
			sacctmgr list qos format=name%-11,MaxJobsPerUser,MaxSubmit \| grep -E "r-limits\|s-limits"
			</pre>

			which provides an output similar to the following

			<pre style="color: lightgrey; background: black;">
			Name Partition MaxJobsPU MaxSubmit
			----------- ----------- --------- ---------
			r-limits r-jobs 2 4
			s-limits s-jobs 1 2
			</pre>

			where

			:; MaxJobsPU
			:: maximum number of jobs from a single user that can be running

	~~A job run on~~ a ~~given partition~~ can ~~only use the QOSes associated to the partition.~~		:; MaxJobsPU
			:: maximum number of jobs from a single user that can be either running or queued

	The list of QOSes associated to the partition accessible to user <code><user></code> can be inspected with command		The list of QOSes associated to the partition(s) accessible to user <code><user></code> can be inspected with command

	<pre style="color: lightgrey; background: black;">		<pre style="color: lightgrey; background: black;">
Line 179:		Line 193:
	</pre>		</pre>

	~~The~~ output ~~of this command is~~ similar to this:		which provides an output similar to this:

	<pre style="color: lightgrey; background: black;">		<pre style="color: lightgrey; background: black;">

GiulioFontana: /* QOSes associated to a partition */

2026-04-27T14:16:58Z

QOSes associated to a partition

← Older revision		Revision as of 14:16, 27 April 2026
Line 158:		Line 158:
	See [[#Default values\|Default values]] to get the full list of the features of a partition.		See [[#Default values\|Default values]] to get the full list of the features of a partition.

	== QOSes associated to ~~a partition~~ ==		== QOSes associated to partitions ==

			In Mufasa 2.0, two special QOSes are used to set the limits to:
			* the number of jobs by the same user that can be running in the system at any given time
			* the number of jobs by the same user that can be either running or queued (both types of jobs count) at any given time

			These special QOSes are different for researchers and students, and they are liste

			associate limits on the number of jobs that a single user can to partitions <code>r-jobs</code> and <code>s-jobs</code> limits




	A job run on a given partition can only use the QOSes associated to the partition.		A job run on a given partition can only use the QOSes associated to the partition.

GiulioFontana: /* Resources available to a QOS */

2026-04-27T14:04:10Z

Resources available to a QOS

← Older revision		Revision as of 14:04, 27 April 2026
Line 92:		Line 92:
	Some of the QOSes are not available to M.Sc. students. See [[#Research users and students users\|Research users and students users]] to understand the differences between user categories in Mufasa and to find out what access you have to system resources.		Some of the QOSes are not available to M.Sc. students. See [[#Research users and students users\|Research users and students users]] to understand the differences between user categories in Mufasa and to find out what access you have to system resources.

	== ~~Resources~~ available to a QOS ==		== Overall resources available to a QOS ==

	The maximum amount of resources that a QOS has access to (available to the running jobs using the QOS, collectively) can be inspected with command		The maximum amount of resources that a QOS has access to (available to the running jobs using the QOS, collectively) can be inspected with command

GiulioFontana: /* SLURM Quality of Service (QOS) */

2026-04-27T14:03:39Z

SLURM Quality of Service (QOS)

← Older revision		Revision as of 14:03, 27 April 2026
Line 34:		Line 34:

	<pre style="color: lightgrey; background: black;">		<pre style="color: lightgrey; background: black;">
	Name Priority MaxWall MaxJobsPU ~~MaxSubmit~~ MaxTRES		Name Priority MaxWall MaxJobsPU MaxTRES
	~~---------~~-- ---------- ----------- --------- --------- --------------------------------------------------------------------------------		----------- ---------- ----------- --------- --------------------------------------------------------------------------------
	nogpu 4 3-00:00:00 1 cpu=16,gres/gpu:3g.20gb=0,gres/gpu:40gb=0,gres/gpu:4g.20gb=0,gres/gpu=0,mem=128G		nogpu 4 3-00:00:00 1 cpu=16,gres/gpu:3g.20gb=0,gres/gpu:40gb=0,gres/gpu:4g.20gb=0,gres/gpu=0,mem=128G
	gpuheavy-20 1 1 cpu=8,gres/gpu:3g.20gb=0,gres/gpu:40gb=0,gres/gpu:4g.20gb=2,mem=128G		gpuheavy-20 1 1 cpu=8,gres/gpu:3g.20gb=0,gres/gpu:40gb=0,gres/gpu:4g.20gb=2,mem=128G
	gpuheavy-40 1 1 cpu=8,gres/gpu:3g.20gb=0,gres/gpu:40gb=1,gres/gpu:4g.20gb=0,mem=128G		gpuheavy-40 1 1 cpu=8,gres/gpu:3g.20gb=0,gres/gpu:40gb=1,gres/gpu:4g.20gb=0,mem=128G
	gpulight 8 12:00:00 1 cpu=2,gres/gpu:3g.20gb=1,gres/gpu:40gb=0,gres/gpu:4g.20gb=0,mem=64G		gpulight 8 12:00:00 1 cpu=2,gres/gpu:3g.20gb=1,gres/gpu:40gb=0,gres/gpu:4g.20gb=0,mem=64G
	gpu 2 1-00:00:00 1 cpu=8,gres/gpu:3g.20gb=1,gres/gpu:40gb=0,gres/gpu:4g.20gb=0,mem=64G		gpu 2 1-00:00:00 1 cpu=8,gres/gpu:3g.20gb=1,gres/gpu:40gb=0,gres/gpu:4g.20gb=0,mem=64G
	gpuwide 2 1-00:00:00 2 cpu=8,gres/gpu:3g.20gb=0,gres/gpu:40gb=0,gres/gpu:4g.20gb=1,mem=64G		gpuwide 2 1-00:00:00 2 cpu=8,gres/gpu:3g.20gb=0,gres/gpu:40gb=0,gres/gpu:4g.20gb=1,mem=64G
	build 32 02:00:00 1 cpu=2,gres/gpu:3g.20gb=0,gres/gpu:40gb=0,gres/gpu:4g.20gb=0,gres/gpu=0,mem=16G		build 32 02:00:00 1 cpu=2,gres/gpu:3g.20gb=0,gres/gpu:40gb=0,gres/gpu:4g.20gb=0,gres/gpu=0,mem=16G
	</pre>		</pre>

GiulioFontana: /* SLURM Quality of Service (QOS) */

2026-04-27T14:03:08Z

SLURM Quality of Service (QOS)

← Older revision		Revision as of 14:03, 27 April 2026
Line 28:		Line 28:

	<pre style="color: lightgrey; background: black;">		<pre style="color: lightgrey; background: black;">
	sacctmgr list qos format=name%-11,priority,maxwall,MaxJobsPerUser~~,MaxSubmit~~,maxtres%-80 \| grep -Ev "normal\|r-limits\|s-limits"		sacctmgr list qos format=name%-11,priority,maxwall,MaxJobsPerUser,maxtres%-80 \| grep -Ev "normal\|r-limits\|s-limits"
	</pre>		</pre>

GiulioFontana: /* SLURM Quality of Service (QOS) */

2026-04-27T14:02:53Z

SLURM Quality of Service (QOS)

← Older revision		Revision as of 14:02, 27 April 2026
Line 56:		Line 56:
	:: maximum number of jobs from a single user can be running with this QOS		:: maximum number of jobs from a single user can be running with this QOS
	:: (note that there are also [[#Research users and students users\|other limitations]] on the number of running jobs by the same user)		:: (note that there are also [[#Research users and students users\|other limitations]] on the number of running jobs by the same user)

	~~:; MaxSubmit~~
	~~:: maximum number of jobs from a single user that can be running or queued with this QOS~~

	:; MaxWall		:; MaxWall