Anyone with a SUNet ID should have access to Stanford's Linux computing resources, which can be accessed by ssh using your SUNet credentials and are documented here. These include nodes for
- low-intensity computing such as email (cardinal.stanford.edu)
- general interactive computing, including most analysis with runtimes <1 day (rice.stanford.edu)
- non-interactive high-performance computing (wheat.stanford.edu, oat.stanford.edu)
(updated 2020-06-12)
Free computing, easy to access, good for jobs needing ~10-100(ish) cores. Anyone with an SUNet ID should be able to access it, though you may need to email your faculty mentor to request an account. Jobs are submitted using the slurm workload manager, and you can access the system using ssh with your SUNet credentials at [SUNetID]@login.sherlock.stanford.edu. When you log in, you'll have access to a home directory (you can get the path via the bash variable $HOME) and a scratch directory where you should put large output files (the relevant path will be accessible through the bash variable $SCRATCH.) For KIPAC, the partition line in your slurm job files can be something like #SBATCH --partition=kipac,iric,hns.
Useful links:
- Some other wiki information is here
- [Official Sherlock page] (http://sherlock.stanford.edu/mediawiki/index.php/Main_Page), including introductory documentation
- Status page