Linux on ISG Blog · D-MATH

Torch

Mon, 22 Dec 2025 14:33:30 +0100

https://github.com/pytorch/pytorch/blob/main/RELEASE.md#release-compatibility-matrix

It is necessary that the required CUDA library is installed in the system, normally this is under ls -lad /usr/local/cuda*.

…in a python project

https://docs.astral.sh/uv/guides/integration/pytorch/

First create a new uv project:

export UV_CACHE_DIR=/scratch/userdata/${USER}/CACHE
export UV_NO_CACHE=1

cd /scratch/userdata/${USER}

# uv will create a virtual env for each project
uv init new-uv-project
cd new-uv-project

Customise the file pyproject.toml:

[project]
# the following supports CUDA 12.8 w/ CUDNN 9.10.2.21
requires-python = ">=3.10,<=3.15"
dependencies = [
 "torch>=2.9,<=2.10",
 "torchvision",
 "torchaudio",
]

[tool.uv.sources]
torch = { index = "pytorch" }

[[tool.uv.index]]
name = "pytorch"
# experimental with 2.9 + 2.10
#url = "https://download.pytorch.org/whl/cu130"
# stable support
url = "https://download.pytorch.org/whl/cu128"
explicit = true

Now add packages defined in the pyproject.toml to the new project, the packages will be installed inside a virtual environment:

Tensorflow

Mon, 13 Oct 2025 13:52:32 +0200

Resources:

To be able to use tensorflow you need to install it for your project. We recommend to use uv for creating your tensorflow python project.

Setup your project

uv will create and use a python virtual environment transparently so there is no need to create one separately. Use the storage on /scratch/userdata/$USER for your project, this is on the local disk of the ada-Xs as not all have the same python versions so the virtual environment must be recreated for every ada-X.

Access a remote linux desktop with ThinLinc

Fri, 16 May 2025 16:21:50 +0200

We are using a solution based on ThinLinc from Cendio to easily connect to a remote Linux desktop inside the D-MATH network from anywhere.

Requirements

You need

either a Wi-Fi connection at ETH, as described here (with the “math-guest” profile),
or a working VPN connection to the D-MATH network, as described here (with the “math-guest” profile)

and

a ThinLinc client for your operating system (external link) Note: on our managed computers, the ThinLinc client is already installed

If you do not want or cannot install the ThinLinc client on your (private) device, you may try to use it via browser at this URL:https://tango.math.ethz.ch (the correct Wi-Fi or the VPN connection is still needed!). This solution may also work on a tablet.

Linux Environment Modules at the D-MATH

Fri, 09 May 2025 08:12:22 +0200

With Environment Modules, or simply called Modules, it is possible to dynamically expand the active user environment and to remove it again without closing the shell.

At the D-MATH modules are used for:

providing applications in different versions,
customising the personal library environment,
setting environment presets for institutes and groups.

Modules are grouped by their functionality in different folders. These module collections are referenced by their MODULEPATH.

Default collections

Every bash-shell starts with two collections.

Sagemath

Tue, 16 Jul 2024 15:09:50 +0200

https://hub.docker.com/r/sagemath/sagemath/

Since Fedora 39 sagemath is not included as a package anymore. There are several possibilities you can use sagemath on Fedora, let’s highlight two:

Using the official docker image
Installing it with conda

Podman

sagemath commandline

As our installation provides a fully configured rootless podman configuration, you can for example use the following commands to run the sagemath image:

# connect to a central client
ssh ada-X
# start tmux, alternativly you can use screen
tmux
# create a .sage folder
mkdir $HOME/.sage

# run sage with your home mounted to /home/sage/mydata and a mounted .sage
podman run --user 0 -ti -v $HOME/.sage:/home/sage/.sage -v $HOME:/home/sage/mydata sagemath/sagemath:latest

# to detach from the tmux session use the following keystrokes
#CTRL+B then D

To add an additional folder you can put another -v parameter, for example:

Moodle Development Environment

Thu, 20 Jun 2024 14:25:07 +0200

Develop moodle

Resources:

The following setup can be used to set up a moodle development environment, with rootless podman:

# D-MATH link podman-compose
ln -s /usr/bin/podman-compose ~/bin/docker-compose

# D-MATH clone moodle-docker repo
git clone https://github.com/moodlehq/moodle-docker.git
cd moodle-docker

# D-MATH set variables
export MOODLE_DOCKER_DB_PORT=15432
export MOODLE_DOCKER_APP_VERSION=18080
export MOODLE_DOCKER_SELENIUM_VNC_PORT=15900
export MOODLE_DOCKER_WEB_PORT=18000

# Change ./moodle to your /path/to/moodle if you already have it checked out
export MOODLE_DOCKER_WWWROOT=./moodle

# D-MATH add custom configuration
cat <<EOF > local.yml
services:
 webserver:
 volumes:
 - "${MOODLE_DOCKER_WWWROOT}/../moodledata:/var/www/moodledata"
 db:
 environment:
 PGDATA: /var/lib/postgresql/data/pgdata
 volumes:
 - "${MOODLE_DOCKER_WWWROOT}/../database:/var/lib/postgresql/data"
EOF

# Choose a db server (Currently supported: pgsql, mariadb, mysql, mssql, oracle)
export MOODLE_DOCKER_DB=pgsql

# D-MATH creating files with group access for moodle code
umask 007

# Get Moodle code, you could select another version branch (skip this if you already got the code)
git clone -b MOODLE_404_STABLE git://git.moodle.org/moodle.git $MOODLE_DOCKER_WWWROOT

# D-MATH add folder for moodle data
mkdir moodledata
mkdir database

# D-MATH setgid for automatic setting group permission
chmod 2770 moodle
chmod 2770 moodledata
chmod 2770 database

# Ensure customized config.php for the Docker containers is in place
cp config.docker-template.php $MOODLE_DOCKER_WWWROOT/config.php

# D-MATH set the permission to 33 inside the container for moodle code
podman unshare chown 33 -R moodle
podman unshare chown 33 -R moodledata
podman unshare chown 999 -R database

# Start up containers, choose docker.io to download the images
bin/moodle-docker-compose up -d

# Shut down and destroy containers
#bin/moodle-docker-compose down

Above we set the webserver port to 18000 so this means you now need to connect to http://localhost:18000 in your browser to see the application.

minikube (Kubernetes Development)

Tue, 31 Oct 2023 12:07:58 +0100

If you want to develop something for kubernetes, you may need to run minikube. This can be done on our linux installations. The following commands should start a working “kubernetes cluster” on your installation:

mkdir ~/bin/

wget https://storage.googleapis.com/minikube/releases/latest/minikube-linux-amd64 -O ~/bin/minikube
chmod 755 ~/bin/minikube

cd ~/bin/
./minikube config set rootless true
./minikube config set driver podman
./minikube config set container-runtime containerd

./minikube start --cni calico

After minikube has started, you can check all running pods with the following:

Linux (Evolution) local mail archive

Wed, 17 May 2023 11:01:41 +0200

This describes how to create a local folder in which you can move emails, which will still be available in the evolution mail client. So the mails will be removed from the mailserver and saved into a folder, which you have to backup.

Requirements:

You already use evolution as your mail client
You make backups of your local files and folders

First create a folder in the filesystem, which will hold your mails:

Howto use docker at D-MATH

Wed, 17 May 2023 10:31:22 +0200

To use docker at D-MATH is to use podman -> https://podman.io/

With podman you can use all docker commands just write podman instead of docker in the command, all the parameters stay the same.

With podman we are able to provide a possibility that a user without administrative rights can download and run docker images.

Example 1

The following example will run the latest docker image of alpine linux (https://hub.docker.com/_/alpine).

podman run -ti -v $(pwd):/home/test alpine:latest