SDSS owns a wide array of servers in Stanford’s Sherlock cluster, offering dedicated on-premises CPUs and GPUs to SDSS researchers. These resources are accessible and monitored through the serc partition.
Getting Access¶
You can get access to Sherlock and the serc partition by requesting an account from this page. If you are unable to use serc for any reason, please reach out to sdss
Using the SERC Partition¶
In general, serc can be accessed by setting the --partition flag to serc in many SLURM commands.
How busy is the SERC partition?¶
You can view how busy the serc partition is on this dashboard. The homepage of this dashboard shows an overall status for the latest data. There are specific pages for CPUs, Memory, and GPUs that show 2 week windows of data. If you see resources in queue higher than the total resources in serc, I would expect a longer than usual wait for your job to start.