Link to GitHub Repository
UCR Research Computing

High-Performance Computing Center (HPCC)

The High-Performance Computing Center (HPCC) at the University of California, Riverside (UCR) provides a state-of-the-art computing infrastructure essential for computationally intensive research and instruction. It serves a broad community of over 150 research groups and more than 800 active users across diverse disciplines, including life sciences, engineering, computational studies, and statistics. The HPCC is committed to offering these resources at a low cost to all UCR researchers and affiliates. A significant NSF MRI grant recently facilitated the acquisition of a Big Data HPC Cluster, further expanding its capabilities.

Key Features & Resources

The HPCC provides a robust environment for demanding computational tasks:

Computational Resources

  • CPUs: Over 16,000 CPU cores (a mix of AMD and Intel processors).
  • GPUs: Approximately 230,000 CUDA cores, featuring Nvidia A100, P100, and K80 GPUs.
  • Memory: Nodes equipped with 512GB to 1TB of RAM.
  • Network: High-speed InfiniBand (IB) interconnect for efficient data transfer between nodes.
  • For more details, visit the HPCC Hardware Overview and Hardware Details pages.

Storage System

  • Capacity: Approximately 5 Petabytes (PB) of parallel GPFS-based disk space.
  • Architecture: A central, high-availability storage system directly mounted on all CPU and GPU nodes, allowing immediate processing of data without transfers.

Software Ecosystem

  • Availability: Over 1,000 software packages and community databases are readily available.
  • User Requests: Support for user-requested software installations (see “Software Installation Requests” below).
  • Explore the full range on the HPCC Software Page.

Accessing HPCC

Eligibility

Access is available to UCR researchers, staff, students (with PI sponsorship), and external collaborators affiliated with UCR research projects.

Account Requests

  • Procedure: Email account requests (new user or lab) to support@hpcc.ucr.edu.
  • Required Information:
    • Full name, NetID, and email address of the user(s).
    • Full name, NetID, and email address of the Principal Investigator (PI).
    • Requests should ideally come from the PI or have the PI CC’ed.
  • New Lab Registration: If the PI’s lab is not yet registered, provide a COA (Chart of Accounts) for the annual subscription fee and any optional additional data storage.
  • Once access is granted, follow the Login Instructions available in the HPCC Manuals.

PI-Based Annual Subscription

  • Fee: $1,000 per year for each PI’s lab.
  • Includes:
    • Access for all registered lab members to HPCC’s computational infrastructure.
    • Standard data storage quota (see below).
    • Access to the extensive software library.
    • Free attendance at HPCC workshops.
    • Up to 1 hour of free consultation services per month.
  • CPU Quotas: Standard quotas are 384 CPU cores per user and 768 CPU cores per lab. Jobs exceeding these quotas will be queued until resources become available.
  • For complete details, refer to the HPCC Recharging Rates PDF.

Storage Options & Policies

  • Standard User Quota: 20 GB per user account.
  • Rented Big Data Storage:
    • Cost: $1,000 per 10TB of usable and backed-up storage space per year. Smaller units (e.g., 100GB) are also available.
    • Backup: Data is backed up to a secondary server room with snapshotting for additional security (10TB usable equates to nearly 30TB raw disk space).
    • This storage can be shared among all users in a registered lab.
  • Owned Big Data Storage:
    • Model: A lab/PI purchases storage hardware (e.g., hard drives) per HPCC specifications.
    • Annual Support Fee: $260 per 10TB of usable and backed-up storage space.
    • This option is often cost-effective for storage needs exceeding 40TB.
    • Owned storage is accessible only to the PI’s group or those granted access by the PI.

Compute Node Ownership

  • Model: A lab/PI purchases compatible computer nodes (e.g., with supported network cards).
  • Benefits:
    • Priority access to owned nodes via a dedicated queue.
    • The lab’s overall CPU quota is increased by the number of owned CPU cores.
  • This is suitable for labs requiring continuous access to hundreds of CPU cores with minimal queue waiting times. Examples of hardware configurations can be found on the HPCC Access Page (links to Thinkmate, Gigabyte).

Special Access Programs

Departmental Cluster Membership

  • This model allows departments to purchase and integrate compute nodes for broad access, such as for large undergraduate classes, where individual PI-based membership is not practical. CPU quotas per user are typically lower under this model.

Using HPCC for Classes

  • Instructors wishing to use HPCC resources for UCR classes must coordinate with HPCC system administrators via support@hpcc.ucr.edu at least 4 weeks before the class starts.
  • Discussions should cover the number of user accounts, software requirements, class-specific Slurm partitions, data storage, and other needs.

External User Accounts

  • Eligibility: Granted to external collaborators with a strong research affiliation with UCR.
  • Requirements: Both the UCR PI and the external collaborator must maintain an HPCC subscription.
  • Approval: Subject to annual review and approval. Requires completion of the External Usage Justification Form.

Software Installation Requests

  • Procedure: Registered users can email software installation requests to support@hpcc.ucr.edu.
  • Timeline: Requests are addressed in the order received. Simple installations may take 1 to a few days; complex ones may take longer.

Support & Documentation

  • Primary Support Channel: For all inquiries, software requests, and issues, email support@hpcc.ucr.edu.
  • Consultation: Up to 1 hour per month is included with the PI-based annual subscription.
  • HPCC Manuals: Comprehensive user guides are available at hpcc.ucr.edu/manuals/.
  • Workshops: HPCC staff offer regular workshops. Schedules and topics can be found on the HPCC Events Page.
  • Main HPCC Website: hpcc.ucr.edu

Citing HPCC & Facility Information

  • Acknowledgement in Publications: If HPCC resources contributed to published research, please acknowledge the center. Guidelines are available on the HPCC Acknowledgement Page.
  • Facility Description for Grants: A detailed description of HPCC facilities for use in grant applications can be found here.

Contact Information

Security Knowledge Base Cost Models Guidelines About Us