THIS IS WORK IN PROGRESS - PULL REQUESTS & SUGGESTIONS HIGHLY WELCOME
This repository collects configurations of simulation codes on public compute resources for quick and easy setup in AiiDA.
The registry computers and codes database are released as JSON files on: https://aiidateam.github.io/aiida-code-registry/
In the following we'll take the example of Piz Daint, a HPC system at the Swiss National Supercomputing Centre.
- Navigate to the
daint.cscs.ch
folder in the GitHub web interface - Select the partition you would like to run on, for example
hybrid
- Click on the
computer-setup.yaml
file and click on the "Raw" button to get a direct link to the file
Now use this link to set up the computer directly via the verdi
command line:
verdi computer setup --config https://raw.githubusercontent.com/aiidateam/aiida-code-registry/master/daint.cscs.ch/hybrid/computer-setup.yaml
You can overwrite any of the parameters provided in the yaml file by appending the corresponding option to the command, e.g. --label my-computer-label
to overwrite the default computer label daint-hybrid
.
Note: Alternatively, you can first create a local clone of the aiida-code-registry
and pass the the local file path of the yaml file to the --config
option.
Some computers require specific configuration options (e.g. to jump over a login node) and provide a dedicated computer-configure.yaml
file.
You'll find it in the same folder:
verdi computer configure ssh daint-hybrid --config https://raw.githubusercontent.com/aiidateam/aiida-code-registry/master/daint.cscs.ch/hybrid/computer-configure.yaml
At this point, you should be able to successfully run:
verdi computer test daint-hybrid
The daint.cscs.ch
folder contains a codes
subfolder with configuration files for individual codes.
Just pick the ones you need and set them up:
verdi code setup --config https://raw.githubusercontent.com/aiidateam/aiida-code-registry/master/daint.cscs.ch/codes/cp2k-8.1-hybrid.yaml
Note: You will be prompted for the computer label (which you can avoid by appending --computer daint-hybrid
to the command).
We highly appreciate help in keeping the configurations up to date and adding new simulation codes & computers.
- Fork this repository
- Add your computer / code
- Create a Pull Request