Do you really need this?
Every Linux operating system at the department comes with its own software modules. What you need may already be offered, albeit probably in a different (usually newer) version. If there is software missing, let us know, and we can probably install it, but for older versions you may need to fall back to the software stack offered by RHEL8 or RHEL9.
Note: For Fedora we have migrated to only offering software distributed by the European Environment for Scientific Software Installations, so on Fedora, when you do module avail you will frist have to pick an EESSI version before you pick which the software you want. You will typically see EESSI/2023.06, EESSI/2025.06, and EESSI/2026.06 as options. Loading an EESSI version and then doing module avail again will show the software collection available for that EESSI version. If you don't see EESSI available in your module list, drop us an email and remember to state the machine this happens on.
EESSI should also be available on RHEL8 and RHEL9 machines at the department.
For more information about EESSI, please see their excellent documentation.
How do I use it?
Type rhel8 or rhel9 at a command prompt. You should now be seeing something like this:
terjekv@ninhursaga:~$ rhel8
[RHEL8] - terjekv@ninhursaga ~$ From here on, you have access to RHEL8 modules via the module command, as if you were on RHEL8. As a comparison:
terjekv@ninhursaga:~$ rhel8 [RHEL8] - terjekv@ninhursaga ~$ module avail PyTorch -------------------- /opt/uio/modules/rhel8/easybuild/modules/bio -------------------- Taiyaki/5.1.0-20200617-fosscuda-2019a-Python-3.7.2-PyTorch-1.2.0 ------------------- /opt/uio/modules/rhel8/easybuild/modules/devel ------------------- PyTorch/1.1.0-foss-2019a-Python-3.7.2 PyTorch/1.2.0-foss-2019a-Python-3.7.2 PyTorch/1.2.0-fosscuda-2019a-Python-3.7.2 PyTorch/1.4.0-foss-2019b-Python-3.7.4 PyTorch/1.4.0-fosscuda-2019b-Python-3.7.4 PyTorch/1.6.0-fosscuda-2019b-Python-3.7.4-imkl PyTorch/1.6.0-fosscuda-2019b-Python-3.7.4-MKL PyTorch/1.6.0-fosscuda-2019b-Python-3.7.4 PyTorch/1.7.0-fosscuda-2019b-Python-3.7.4-imkl PyTorch/1.7.0-fosscuda-2019b-Python-3.7.4 PyTorch/1.7.1-foss-2020b PyTorch/1.7.1-fosscuda-2020b PyTorch/1.9.0-fosscuda-2020b-imkl PyTorch/1.9.0-fosscuda-2020b PyTorch/1.11.0-foss-2021a-CUDA-11.3.1 (D) -------------------- /opt/uio/modules/rhel8/easybuild/modules/vis -------------------- torchvision/0.7.0-fosscuda-2019b-Python-3.7.4-PyTorch-1.6.0-imkl torchvision/0.7.0-fosscuda-2019b-Python-3.7.4-PyTorch-1.6.0 torchvision/0.8.1-fosscuda-2019b-Python-3.7.4-PyTorch-1.7.0-imkl torchvision/0.8.1-fosscuda-2019b-Python-3.7.4-PyTorch-1.7.0 torchvision/0.8.2-fosscuda-2020b-PyTorch-1.7.1 Where: D: Default Module [...] [RHEL8] - terjekv@ninhursaga ~$ exit terjekv@ninhursaga:~$ module avail PyTorch -------------------- /opt/uio/modules/rhel9/easybuild/modules/all -------------------- PyTorch-Lightning/1.7.7-foss-2022a-CUDA-11.7.0 PyTorch-Lightning/1.7.7-foss-2022a (D) PyTorch/1.12.0-foss-2022a-CUDA-11.7.0 PyTorch/1.12.0-foss-2022a (D) Where: D: Default Module [...]
Notice that the module list includes the OS version, rhel8 vs rhel9 in the paths.
I can't access my files!
By default the environment only provides access to your home directory and your current working directory. So, if your files are on a research volume on kadingir, start the environment from where your files are, like so:
terjekv@ninhursaga:~$ cd /mn/kadingir/my_research_volume_000000/ terjekv@ninhursaga:/mn/kadingir/my_research_volume_000000$ rhel8 [RHEL8] - terjekv@ninhursaga /mn/kadingir/my_research_volume_000000$
Now you can access all files and directories, as well as subdirectories, under /mn/kadingir/my_research_volume_000000/