Partition organization

Since July 25, 2022, a new partition organization was introduced in EXPLOR. The new layout is shown below:

Attention

The general allocation in the std partition now includes nodes from the former freeXXX partition. If you do not select any special nodes, your JOB may be terminated due to the priority of JOBS on MyXXX nodes. To avoid this, consider selecting appropriate nodes while excluding private nodes. For instructions, please check items (4.3 and 5).

Association table

Partition	Hostname	# nodes	# CPU /node	CPU memory (Mb)	Node memory (Gb)	FEATURE
debug	cna[01]	1	64	2000	118	BROADWELL,OPA,INTEL
std	cna[02...64]	53	32	3750	118	BROADWELL,OPA,INTEL
std	cnb[01...62]	30	32	3750	118	BROADWELL,OPA,INTEL
std	cnc[01...64]	51	32	5625	182	SKYLAKE,OPA,INTEL
std	cnd[01...12]	11	16	3750	54	IVY,IB,INTEL
std	cne[01...16]	16	8	15000	118	BROADWELL,INTEL,HF
std	cnf[01...08]	5	8	12000	86	SKYLAKE,OPA,INTEL,HF
std	cng[01...12]	8	40	4500	182	CASCADELAKE,OPA,INTEL
std	cnh[01...02]	2	8	94000	758	CASCADELAKE,OPA,INTEL,HF
std	cni[01...16]	16	40	9000	354	CASCADELAKE,IB,INTEL
std	cni[16...32]	16	40	4500	182	CASCADELAKE,IB,INTEL
std	cnj[01...64]	64	48	5200	246	EPYC3,IB,AMD
std	cnk[01...08]	8	8	22500	182	CASCADELAKE,IB,INTEL,HF
std	cnl[01...04]	4	24	241	241	EPYC4,IB,AMD
std	cnl[05...18]	14	32	10000	500	EPYC4,IB,AMD
gpu	gpb[01...06]	6	32	3750	118	BROADWELL,OPA,P1000,INTEL
gpu	gpc[01...04]	4	32	2000	54	BROADWELL,OPA,GTX1080TI,INTEL
gpu	gpd[01...03]	3	24	3600	86	CASCADELAKE,OPA,T4,INTEL
gpu	gpe[01...02]	2	40	4500	182	CASCADELAKE,IB,RTX6000,INTEL
gpu	gpf[01]	1	32	3750	118	CASCADELAKE,L40,INTEL
myXXX	---	--	--	--	--	--

myXXX associations table

Partition	Hostname	# nodes	#CPU	Memory (Gb)	Preemption
mysky	cnc[53,55-57,59-63]	9	32	182	Oui
myhf	cnf[02-06]	5	8	86	Oui
mycas	cng[01-02,04-05,07,09,11-12]	8	40	182	Oui
mylhf	cnh[01-02]	2	8	758	Oui
mylemta	cni[01-16]	16	40	354	Oui
mystdcas	cni[16-24,29-32]	8	40	182	Oui
mystdepyc	cnj[01-64]	64	48	246	Non
myhfcas	cnk[01-08]	8	8	182	Oui
mylpct	cnl[02-04]	3	24	241	Oui
mygeo	cnl[05-18]	14	32	500	Non
myt4	gpd[01-03]	3	24	86	Oui
myrtx6000	gpe[01-02]	2	40	182	Oui
mylpctgpu	gpf[01]	1	32	118	Oui

Note

The hostnames cnXX correspond to CPU nodes and gpXX to GPU nodes.

Submission instructions

(1) All submissions must always include

#SBATCH --account=MY_GROUP

or

#SBATCH -A MY_GROUP

(1.1) Special information – MyXXX submission with a different project association

To use a different project association when you have multiple projects, remove the option #SBATCH -A/--account from your script and add it externally on the command line:

sbatch --account MY_GROUP my_subm_scrit.slurm

or

sbatch -A MY_GROUP my_subm_scrit.slurm

MY_GROUP should be your project identifier; you can verify it in your terminal prompt:

[<user>@vm-<MY_GROUP> ~]

(2) General cases where you do not need a special machine

Attention

The general allocation in the std partition now includes nodes from the former freeXXX partition. If you do not select any special nodes, your JOB may be terminated due to the priority of JOBS on MyXXX nodes. To avoid this, consider selecting appropriate nodes while excluding private nodes. For instructions, please check item (4).

Important

The private MyXXX partitions that are non-preemptible are: mystdepyc (cnj[01-64]) and mygeo (cnl[05-18]).

(2.1) any type of machine in std

#SBATCH --account=MY_GROUP
#SBATCH --partition=std
#SBATCH --job-name=Test
#SBATCH --nodes=1
#SBATCH --ntasks=4

or

#SBATCH -A MY_GROUP
#SBATCH -p std
#SBATCH -J Test
#SBATCH -N 1
#SBATCH -n 4

(2.2) any type of machine in gpu

#SBATCH --account=MY_GROUP
#SBATCH --partition=gpu
#SBATCH --job-name=Test
#SBATCH --nodes=1
#SBATCH --ntasks=1
#SBATCH --gres=gpu:2

or

#SBATCH -A MY_GROUP
#SBATCH -p gpu
#SBATCH -J Test
#SBATCH -N 1
#SBATCH -n 1
#SBATCH --gres=gpu:2

(3) Hosted hardware – continues as before

The mystdcasXXX partitions will now be accessible through a single mystdcas partition

(mystdcaslemta, mystdcasijl, mystdcascrm2) ==> mystdcas

#SBATCH --account=MY_GROUP
#SBATCH --partition=mycas
#SBATCH --job-name=Test
#SBATCH --nodes=1
#SBATCH --ntasks=4

or

#SBATCH -A MY_GROUP
#SBATCH -p mycas
#SBATCH -J Test
#SBATCH -N 1
#SBATCH -n 4

(4) Precise node selection

Selecting specific nodes is done via the FEATURES shown in the association table above. Examples:

#SBATCH --constraint=SOMETHING_FROM_FEATURES

(4.1) Select nodes from the former sky partition

#SBATCH --account=MY_GROUP
#SBATCH --partition=std
#SBATCH --constraint=SKYLAKE,OPA,INTEL
#SBATCH --job-name=Test
#SBATCH --nodes=1
#SBATCH --ntasks=4

or

#SBATCH -A MY_GROUP
#SBATCH -p std
#SBATCH -C SKYLAKE,OPA,INTEL
#SBATCH -J Test
#SBATCH -N 1
#SBATCH -n 4

(4.2) Select nodes from the former p100 partition

#SBATCH --account=MY_GROUP
#SBATCH --partition=gpu
#SBATCH --constraint=BROADWELL,OPA,P100,INTEL
#SBATCH --job-name=Test
#SBATCH --nodes=1
#SBATCH --ntasks=1
#SBATCH --gres=gpu:2

or

#SBATCH -A MY_GROUP
#SBATCH -p gpu
#SBATCH -C BROADWELL,OPA,P100,INTEL
#SBATCH -J Test
#SBATCH -N 1
#SBATCH -n 1
#SBATCH --gres=gpu:2

(4.3) Exclude all nodes from the former freeXXX/MyXXX machines and select all other legacy nodes (std, sky, ivy, hf)

#SBATCH --account=MY_GROUP
#SBATCH --partition=std
#SBATCH --constraint=NOPREEMPT
#SBATCH --job-name=Test
#SBATCH --nodes=1
#SBATCH --ntasks=4

or

#SBATCH -A MY_GROUP
#SBATCH -p std
#SBATCH -C NOPREEMPT
#SBATCH -J Test
#SBATCH -N 1
#SBATCH -n 4

(5) Restart/Requeue preempted JOBS

This is not a feature for JOBS that end with an error. It is a feature for JOBS that may have been removed from execution by the preemption rule. If you want to submit to all machines in the STD partition—even if on some machines your JOB may be preempted to give priority to a higher-priority job—you can use the job requeue feature (--requeue):

#SBATCH --account=MY_GROUP
#SBATCH --partition=std
#SBATCH --requeue
#SBATCH --job-name=Test
#SBATCH --nodes=1
#SBATCH --ntasks=4