Skip to main content

December 2025 Maintenance Update

· 6 min read
Kristen Finch
Kristen Finch
Director of Research Computing Solutions

During this month’s maintenance window, we conducted routine security updates and performed hardware consolidation in the Data Center. The next maintenance is scheduled for Tuesday, January 13, 2025 (the second Tuesday of the month).

Notable Updates​

  • Node image updates – Routine updates and security patches.
  • We also installed new modules:
    • gcc/15.2.0
    • conda/Miniforge3-25.9.1-0 (read more below)

New Feature: Conda Module Now Available on Hyak Klone​

This month, we’re excited to announce that the Conda module is now available on Hyak Klone. Previously offered only on Tillicum, the module has now passed testing on Klone and is ready for full use. We highly recommend it as the primary way to manage Python environments on Hyak Klone due to its simplicity, reliability, and improved storage efficiency.

Using the Conda module means:

  • No miniconda installation required. Just load the module and start creating environments.
  • More efficient storage usage, leading to fewer issues with home directory quota limits.
  • Cleaner, isolated Python environments, easier debugging, reproducibility, and environment cleanup.

Below is a quick guide to help you get started.

Conda Environments​

Conda allows you to create isolated environments that include specific versions of Python, libraries, and tools.

Load Conda Module​

First, load the Conda module:

# remember on Klone, module commands are only available on compute nodes
module load conda

After loading the module, the conda command becomes available. You can now create and manage your own environments.

Create and Manage Conda Environments​

For example, create an environment named "myenv" with Python 3.12 and the NumPy package:

conda create --name myenv python=3.12 numpy

Activate the environment to use it:

conda activate myenv

List your available Conda environments:

conda env list

Remove an environment:

conda env remove --name myenv

To install additional packages in myenv environment, use conda install. Conda has several default channels that will be used first for package installation. If you want to use another channel beyond the defaults channel, you can, but we suggest that you select your channel carefully.

warning

By default, the system Conda stores environments in your home directory ($HOME/.conda/envs). We recommend installing Conda environments to your project directory under /gscratch/<group>/<myfolder> due to the limited storage space (10 GB) in your home directory. Please apply the steps by following the instructions below carefully.

Customize Environment and Package Locations​

There are two ways to specify where your Conda environments and packages are stored.

Option 1. Use --prefix for explicit paths​

Manually set the path to your Conda environment by --prefix and always activate your Conda environment with full path.

module load conda
conda create --prefix /gscratch/<myproject>/<myfolder>/myenv python=3.12
conda activate /gscratch/<myproject>/<myfolder>/myenv
conda install numpy scipy matplotlib
Option 2. Configure defaults in $HOME/.condarc​

To make this the default behavior, edit (or create) the file $HOME/.condarc:

envs_dirs:
  - /gscratch/<myproject>/<myfolder>/conda/envs
pkgs_dirs:
  - /gscratch/<myproject>/<myfolder>/conda/pkgs

This will place all of your environments and package caches in this directory by default, and you won't have to worry about specifying the full prefix to your environment when installing it or activating it.

Installing Packages with pip​

You can use pip inside a Conda environment to install Python packages. Our suggested use of pip is inside a conda environment. For example:

module load conda
conda activate myenv
pip install seaborn

This ensures that pip installs packages into the active Conda environment — not globally — making it easy to clean up completely when you are done.

See the best practices from Anaconda for using pip with Conda and pip documentation for more information.

New Tutorials Now Available​

This fall, our training sessions focused primarily on Tillicum, our new GPU-accelerated service. However, the core skills and workflows we covered apply equally well to Hyak Klone. With only minor adjustments to file paths or partitions, Klone users can benefit directly from these materials because the overall computing environments are very similar and the tools (modules, Slurm, containers, Python workflows, data management) work the same across both systems. We will offer more Hyak Klone trainings in 2026.

Stay informed by subscribing to our mailing list and the UWIT Research Computing Events Calendar.

Office Hours​

External Training Opportunities​

  • The NSF LEVEL UP AI (Launching an Educational Vision to Expand Leadership, Understanding, and Progress in Artificial Intelligence) project brings together educators, researchers, and professionals across the computing community to build a shared vision for expanding capacity and access to AI education. 01/07/26 - 12:00 AM to 01/08/26 - 11:59 PM EST Learn more and register.
  • 2 Postdoc positions in AI with deadlines arriving soon.
  • The eScience Institute offers the annual Winter School to students and lecturers interested in developing basic skills and knowledge of the tools used in data science. Gaining literacy in topics such as Python, R, Jupyter, and reproducible environments can be beneficial beyond STEM, including areas like global or public health, public policy, social sciences, social work, international relations, and business management. There are no prerequisites to take this course and there is no credit offered. Only UW faculty, staff, students (undergraduates and graduates) from all three campuses are welcome to apply. Deadline to apply: February 4th, 2026.

Having trouble? Get Research Computing support.

Happy Computing,

Hyak Team