Frequently Asked Questions (FAQ)
Welcome to the Research Computing FAQ. If you cannot find the answer to your question here, please contact our support team or join our Slack Channel.
đŠī¸ Cloud Computing & Ursa Major
Q: Are Google Cloud GPUs free under the Ursa Major subsidy?
A: No. Under the strict Tier 1 subsidy rules (Project Polaris), Ursa Major covers baseline compute (e2-standard-2), secure enclaves, and strategic AI (Gemini). All Cloud GPU (A100, L4, T4), high-performance compute (c3, n2), and massive storage workloads must operate via the Tier 2 Direct Recharge model to your grant COA. If you need no-cost GPU access, please utilize the UCR HPCC.
Q: How do I request an Ursa Major allocation? A: Please review the Ursa Major Service Tiers and submit a request via our Intake Form.
Q: Can I run local open-source LLMs like Llama 2 or Ollama on Ursa Major? A: Yes, but running these models requires a GPU-enabled workstation (e.g., a T4 node). Because Cloud GPUs are excluded from the Tier 1 subsidy, this will incur a Direct Recharge to your lab. We strongly recommend running offline LLMs on the HPCC for subsidized access. See our Ollama Guide for details.
đž Data Storage
Q: Where should I store my active, hot research data? A: Active data being processed should reside on the HPCC GPFS Storage. For warm project data, we recommend the new Ceph RDS System (currently in Pilot) or SDSC Qumulo.
Q: I have terabytes of raw genomic sequencing data. Should I put it in Google Drive? A: No. Google Drive has a strict 500GB limit per account and is designed for document collaboration, not massive binary datasets. Please utilize Ursa Major Coldline Storage for long-term archiving or Ceph RDS for active storage.
Q: What is the cost of archiving data in Ursa Major? A: Coldline storage (for compliance archiving and backups) is fully subsidized under the Tier 1 pool at no cost to the researcher, provided usage remains within reasonable limits. See KB001: Data Storage Strategy.
đĨī¸ High-Performance Computing (HPCC)
Q: How much does it cost to use the UCR HPCC? A: The HPCC operates on a $1,000/year Flat Rate per Lab. This provides unlimited compute access for your entire research group.
Q: My compute needs exceed the capacity of the UCR HPCC. What are my options? A: We utilize a strict 4-Tier GPU & Compute Offload Hierarchy. If the local HPCC cannot meet your needs, we will assist you in migrating your workloads to the regional Nautilus Cluster, or help you apply for national allocations via the NAIRR Pilot or NSF ACCESS.
Q: How do I migrate my code from a Google Cloud VM to the HPCC? A: We use Apptainer (Singularity) on the HPCC instead of raw Docker containers. Please refer to KB006: Migrating Compute to HPCC for a step-by-step guide.
đ Security & Compliance
Q: I am working with HIPAA, NIST 800-171, or CUI data. Can I process this on the standard cluster? A: No. Highly regulated data (P4_CRITICAL) requires a formally approved Data Security Plan (DSP) and must be processed within a designated Secure Enclave. Please contact us immediately to initiate the enclave provisioning process.
Q: Can I use ChatGPT or Google Gemini with my unpublished research data? A: You may use the institutional, enterprise version of Google Gemini provided through UCR, as we have a Business Associate Agreement (BAA) ensuring your data is not used to train external models. Do not input sensitive or P3/P4 classified data into public, consumer-tier AI chatbots. Read our AI Safety Guide for more information.