Doerr School of Sustainability, Center for Computation (SDSS-CC)
Sherlock 4.0 SERC update
SERC will be replacing older SH02 machines and adding capacity in Sherlock 4.0 (SH04). Stay tuned for details!
Introduction
The Doerr School of Sustainability, Center for Computation (SDSS-CC) provides a variety of high performance computing (HPC) resources to support the SDSS research community in performing world renowned research.
Resources available to SDSS researchrtd include cluster-computing, interactive compute nodes, GPU computation, memory-intensive nodes, data storage, and more. See examples below:
Sherlock HPC shared SERC partition - Sherlock 3.0 nodes: 200 Standard (32 cores, 256GB), 8 Performance (128 cores, 1024GB), 10x8 A100 GPU
- Sherlock 2.0 nodes: 24+ nodes, (24 cores, 192/384GB) 2 x 4 V100GPU
- 1.35 PB + Oak storage
- Sherlock public partitions (owners, normal, dev, gpu, bigmem)
- Cloud Compute Resources (e.g. Google Earth Engine & Cloud Platform)
Mission:
To advance research and scholarship by providing access to high-end computing, training, and advanced technical support in an inclusive community at SDSS.
Scope of SDSS-CC:
High performance computing (HPC) enables scientific and engineering advances for many research groups in the Doerr Stanford School of Sustainabilty (SDSS). HPC encompasses a constantly evolving set of resources including multi-core CPU and GPU processors and other advanced hardware, multi-node cluster computing, as well as single-core jobs that require computing resources beyond individuals’ laptops and desktops. SDSS-CC is the SDSS organization tasked with assessing the HPC needs of research groups and providing or facilitating access to HPC resources at Stanford or elsewhere (e.g., supercomputer clusters run by NSF or other agencies, Amazon Web Services, Google Cloud Platform, Microsoft Azure). SDSS-CC invests in HPC resources using funding provided by SDSS faculty, affiliates programs, and general SDSS funds. SDSS-CC also funds system administration and staff who provide user consulting, support, and training. When appropriate, SDSS-CC partners with the University SRCC team to provide these services. User support and training services provided by SDSS-CC include consultations with faculty and/or their research groups to provide specific recommendations for the most appropriate hardware or other HPC resources to meet group needs, up-to-date instructions for accessing and utilizing SDSS-CC resources, consultations with individual or research groups to develop HPC workflows to optimize utilization of SDSS-CC resources, and assistance with getting access to resources beyond those that SDSS-CC can provide (e.g., allocation proposals for national supercomputer centers). SDSS-CC is distinct from and complementary to SDSS IT, in that SDSS-CC is focused on supporting computational research.
Eligibility:
SDSS-CC resources are available to most members of the SDSS community. Please contact the SDSS-CC team to determine which resources would best meet your needs.