Software stacks

The software stacks are shared for the whole cluster and are generated using the Spack package manager

Types of available software stacks

Currently there are two types of stacks available.

current

stable stack throughout the semester
new stack every semester (build from stratch)
“old” stacks still valid with current_<date>
3 ways to access (further details see Usage):
- directly with spack commands
- via modules
- via environment

Paths:

stack: /home/data/software/current
environment: /home/data/software/current/ipsy_env

experimental

stack that will be updated regularly and on short notice user requests
updating instead of building from scratch can lead to multiple versions of software (e.g. python ~gcc@11.0 vs python ~gcc@11.1)
same access methods as in current are available, but due to multiple versions not unambiguous
-> recommended method: use environment to load stack

Paths:

stack: /home/data/software/experimental
environment: /home/data/software/experimental/ipsy_env

Usage

There are multiple ways to use the software stacks, depending on how much control you want over what packages are loaded (and in which way):

ipsy env
directly via spack
via modules

load ipsy environment

Instead of loading every package individually an evironment can be activated containing only the newest version of each package.
This is the recommended way for most users.

$ . /home/data/software/current/ipsy-env/activate

all packages inside of the environment are now loaded and can be used directly.
e.g.: to start python:

$ python

To see what software is available run:

module avail

via modules

For every installed package spack generates a module file in addition meaning that classic module commands can be used as well.

Again shell support has to be enabled

$ . /home/data/software/current/env.sh

to see what packages are installed run

$ module avail

to load a package

$ module load python

or if a specific version is needed, load the entry shown in module avail