The serc partition is SDSS’s shared allocation on Stanford’s Sherlock cluster. It provides CPU, high-memory, and GPU nodes for Doerr School research workloads.
Use the shared resources considerately¶
SERC is a shared research resource. Request only the CPUs, GPUs, memory, and wall time your workload needs, and use batch jobs for substantial work. The Community Guidelines summarize current limits for interactive and batch sessions.
Find node details and scheduling guidance¶
The reference guide includes the current node classes, GPU configurations, memory characteristics, and practical advice for choosing resources:
If your workload requires a large or unusual allocation, contact sdss