Interactive Julia

Opening Julia Form

  • The first step is to select the Interactive apps and then click Julia in the Open OnDemand dashboard.

  • Then you will see a form which needs to be filled to launch the Julia server on compute nodes.

ood-julia-form.png

Partition

  • Partitions are work queues that has computational nodes in it to run the jobs. The partitions available in Colibri are general-cpu, general-gpu, interactive-cpu, interactive-gpu, backfill, inbre, and prism.

  • However, you will see only the list of authorized partitions available for your Colibri account and hence, you won’t see the partitions which you don’t have authorization in the select dropdown.

  • Select one of the partitions where you want to submit your job. Refer to the page → Partitions in Colibri to select the ideal partition for your job.

  • Jobs on the backfill partition are subject to being stopped and requeued.

Walltime(Hours)

  • Sets the Walltime for your job. The job gets killed if the Walltime you set gets elapsed. The maximum Walltime for each partition is listed in the table below.

    Partition Max Hours

    backfill

    338

    general-cpu

    169

    general-gpu

    169

    inbre

    169

    interactive-cpu

    25

    interactive-gpu

    25

    prism

    169

CPUs(Threads)

  • Sets the Number of CPUs(Threads) for the job. The default and minimum value is 1 CPU thread.

Memory (GB/Gigabytes)

  • Sets the memory for the job in terms of GB. The memory you select is available to all assigned threads for the job. The default and minimum value is 1 GB.

GPUs

  • GPUs are available only in the partitions that contain GPU nodes: general-gpu, interactive-gpu, backfill, inbre, and prism. If you select one of these partitions, you can see the GPUs dropdown enabled.

Node Selection

This is an optional node selection feature for submissions to Slurm. It allows users to specify a particular node or any node with specific feature flags.

Selection Options

There are four available options for node selection:

  1. Disabled: This is the default setting. When "Disabled" is selected, the application will not send any specific node preferences to Slurm during job submission.

  2. Node List: With this option, users can specify a particular node (one single node) where the job should run.

  3. Node Features (--constraint): Users can select this option if they want to run their job on any node that meets a specific feature.

  4. Node Features (--prefer): This option is similar to the "--constraint" option, but it only indicates a preference. If it’s not available, it might be scheduled elsewhere.

Email Address

  • If you want to get notified via email when the job starts/fails/ends, enter the email address associated with the Colibri account.

  • You can leave the text field blank if you don’t want job notifications.

Julia Working Directory

Julia Working Directory

If left empty, the working directory points to your /home directory by default. To switch to a different directory, click the Select Path button.

Switch Directory

Select the desired directory you want to work with. After you make the selection, the Working Directory for the interactive job gets updated.

User Interface

julia-user-interface.png
  • Pluto.jl, VSCode, Jupyter Lab and Jupyter Notebook interfaces are available. Select one for the interactive job.

Submission Environment

You can select your desired submission environment from the drop down list. For example, to use an Apptainer image, select the Container option.

Submission Environment
  • Basic Modules

  • Container

This option allows you to launch the interactive application with the default software modules.

Instead of you using Envirumental Modules, you can use an Apptainer/Singularity Container. You just need to select the appropriate SIF image.

  • When using Pluto.jl as user interface, make sure that Pluto.jl is added in your Apptainer/Singularity Container.

Container

Launching the Interactive Session

  • Once all the form fields are set, click the Launch button. The interactive session will be launched and the below output is shown.

Julia Output
  • The above output shows that the interactive Julia session has been launched successfully. You will see the Connect button only if the resources you have requested becomes available. Else, you might need to wait for some more minutes for the resources allocation and then the Connect button appears.

  • Click the Connect button after it appears.

The interactive session won’t be launched if you try to submit your job to the partition that you don’t have access to. Also, the page gets reloaded and you need to select the value for the form fields again since the form gets reset.

Opening Julia in Jupyter Lab/NoteBook

  • On clicking the Connect to Jupyter button, the Jupyter Notebook opens in a new tab like below. Select the appropriate julia kernel. image::ondemand/ood-julia-jupyter-open.png[Julia - Jupyter]

Killing the Interactive Session

Go to My Interactive sessions tab and select the Delete button of the current interactive session.

Julia Delete

A confirmation pop-up will appear and press Confirm to delete the session.

Whenever you are done with the work, it’s always recommended to kill/delete the interactive session running on the compute nodes and free the resources.