HPC@UC

Providing UC Researchers Access to SDSC’s High-Performance Computing Resources and Expertise

The San Diego Supercomputer Center (SDSC), an Organized Research Unit of the University of California, San Diego, is making major high-performance computing resources available to the UC community. This program, HPC@UC, is being offered in partnership with the UC Vice Chancellors of Research as well as campus CIOs. Awards under this program are intended to help UC researchers who are new to high-performance computing expand their overall research program. Specifically, this program is designed to:

  • Broaden the base of UC researchers who use advanced computing
  • Provide an on-ramp to UC researchers who are just getting started with high-performance computing by giving them access to cyberinfrastructure that complements what is available at their campus
  • Seed promising computational research
  • Facilitate collaborations between SDSC and UC researchers
  • Help UC researchers be successful when pursuing larger allocation requests through NSF’s Advanced Cyberinfrastructure Coordination Ecosystem: Services & Support (ACCESS) program, and other national computing programs.

Advanced cyberinfrastructure (CI), including high-performance computing (HPC) systems, is critical to advancing science and discovery across a wide range of research domains. SDSC operates some of the most advanced CI in the nation, including a petascale computing system for conducting complex numerical simulations, and high-performance storage systems for moving, analyzing, and storing massive amounts of data from simulation and experiment. SDSC’s HPC applications specialists, data scientists, and systems administrators provide the support and expertise required to make maximum use of these resources. Access to the resources is ubiquitous over high-speed networks between the campuses, with specialized hardware that enables data movement at speeds in the tens of Gigabits per second.

Resources Available

Resources offered through this program include the following:

  • Expanse supercomputer: A ~5 PFlop/s system featuring 728 nodes, each with two 64-core AMD EPYC 7742 processors, 256 GB memory; and 4 large memory nodes, each with two 64 core AMD EPYC 7742 processors and 2 TB of memory;
  • Expanse GPU: 52 GPU nodes:  36 NVIDIA V100 GPU Nodes; each has 4 GPUs/node;
  • Data Resources: Over 12 PB of high-speed storage made available via Lustre parallel file systems, as either short term Performance Storage used for temporary files, or long term, non-purged Project Storage that persists for the life of the project. A Durable Storage resource provides a second copy of all data in Project Storage file system;
  • Applications: A large installed base of applications for HPC and big data analytics;
  • Expertise: SDSC staff have broad expertise in the application of advanced computing and stand ready to assist you in making the best use of these resources for your research.

Learn more about SDSC's HPC resources.

Trial Accounts on Expanse: Before submitting an allocation request to HPC@UC, you may also want to consider requesting a Trial account on SDSC’s Expanse system. These are allocations for 1000 core-hours and are intended for very quick assessment of Expanse to see if it meets your needs. These requests are fulfilled within 1 working day. You are eligible for an HPC@UC allocation even if you already have a Trial account.

Eligibility and Review

  • Applications may request up to 500K core-hours of computing, associated data storage, and access to SDSC expertise to assist their research team. Awards are active for one year.
  • Applicants must not have an active award in the NSF’s Advanced Cyberinfrastructure Coordination Ecosystem: Services & Support (ACCESS) program.
  • The expectation is that these awards will lead to larger, formal allocation requests on national HPC systems that are available through the ACCESS program. SDSC staff will assist in developing these allocation applications.
  • Applications are reviewed on an ongoing basis. Applicants will be notified within 10 business days of the review decision.
  • Applicants must not have a previous HPC@UC award on Expanse.