Skip to main content

Source repo: sdsc-summer-institute-2021 | Branch: main | Last synced: 2026-04-24 10:27:17.425 UTC

Configuring Your Training Account

Each member of the class, and instructors have a training account with a default user environment customized with the right paths and software paths, including the conda/mininconda software.

Bash Environment​

*Note: Do NOT replace the .bashrc with your own. * Aliases and paths are pre-installed (see below) and if you replace the files you will not be able to run the example software.

Miniconda installs:​

Institute staff will provide instructions for installing miniconda for the workshop on the preparation day. Please do not re-install conda. If you have trouble, go to the Slack HelpDesk.

Default applications location:​

CHECK THIS You have been set up with a default user environment: you do not need have a .bashrc file. You will see the code in these folders by default

  • link to /cms
  • link to singularity for SDSC SI20

Node Reservations:​

GPU reservations​

*getcpu: "srun --pty --nodes=1 --ntasks-per-node=24 -p compute -t 3:00:00 --reservation=si2020resday1 --wait 0 /bin/bash" *getgpu: "srun --pty --nodes=1 --ntasks-per-node=6 --gres=gpu:1 --partition gpu-shared -t 3:00:00 --reservation=si2020resday1 --wait 0 /bin/bash" *getcpu1: "srun --pty --nodes=1 --ntasks-per-node=24 -p compute -t 1:00:00 --reservation=si2020resday1 --wait 0 /bin/bash"

CPU reservations​

For the class, we have created special SLURM queue reservations so our jobs will run faster. Please use the following aliases:

getcpu - one comput node for 3 hours
getcpu1  - one comput node for 1 hour

GPU reservations​

getgpu - one GPU for 3 hours (on days with GPU hands-on)
getgpu1 - one GPU for 1 hour (on days with GPU hands-on)

Using Jupyter Technologies​

Jupyter Notebooks​

** NEED TO UPDATE MPT 6/11/21** The command to run secure Jupyter notebook servers for this class is the start_python_hpc_notebook command (this will be in your path).

  • update to use Galyleo

JupyterLab servers​

*update to use Galyleo

Notebooks hosting Spark​

** do we need this section?**