Skip to main content

Information

How to use allocated resources

User Access

User access must be done using an ssh client, for that you need and ssh key created on your local machine, more on that in this link, if you need help accessing the machine or creating your ssh key, please contact us at ingrid.helpdesk@lip.pt

For users runing jobs on Cirrus-B cluster, the interface machine name is ui102.b.incd.pt.

Consumed resource and available quota

For users running jobs on Cirrus-B cluster you can use the reporting tools found on this link

General HPC information

Documentation on how to use slurm, creating your first jobs and all sorts of interactions with the resource manager, please take a look at these tutorials

If you are using slurm for the first time, we recommend starting with this tutorial

NOTE: The jobs running under the FCT grant must include 2 extra flags on the submission script "#SBATCH -p PARTITION" and "#SBATCH -q QOS", the "-p" is used to specify the partition in which you will run your job and the "-q" is used to specify the QOS with which you will run your job under. We will send you an email with the QOS attributed to your group. As for the partitions youdepends canon choose,the site you haveare twousing; possibilities:CIRRUS-A or CIRRUS-B.

  • "FCT"
    • SitePartitionMAX.TIMEComments
    CIRRUS-A (Lisbon)fct4 daysnormal priority
    CIRRUS-B (Minho)FCT4 daysnormal priority
    CIRRUS-B (Minho)FCT_Low_Priority1 daylow priority

    Example of submit script for jobsCIRRUS-A:

        $ cat myjob.sh
    #!/bin/bash
    #SBATCH --job-name=MyFirstSlurmJob
    #SBATCH --time=0:10:0
    #SBATCH --nodes=1
    #SBATCH --ntasks-per-node=16
    #SBATCH -p fct
    #SBATCH -q <group_qos>
    echo Hello

    Example submit script for CIRRUS-B:

        $ cat myjob.sh
    #!/bin/bash
    #SBATCH --job-name=MyFirstSlurmJob
    #SBATCH --time=0:10:0
    #SBATCH --nodes=1
    #SBATCH --ntasks-per-node=16
    #SBATCH -p FCT
    #SBATCH -q <group_qos>
    echo Hello

    For submission of job with normallower priority,priority:

    in
        which$ youcat canmy_low_prio_job.sh
    run#!/bin/bash
    jobs#SBATCH up--job-name=MyFirstSlurmJob
    to#SBATCH 4--time=0:10:0
    days#SBATCH long;--nodes=1
    
  • "FCT_Low_Priority"#SBATCH (currently--ntasks-per-node=16
    only#SBATCH vailable-p forFCT_Low_Priority
    cirrus.b.incd.pt)#SBATCH for-q jobs<group_qos>
    withecho lowRelaxed priority, in which you can run jobs up to 24 hours long;
    •
Hello
Back to top