Speakers
Description
The Triton Shared Computing Cluster (TSCC) at the San Diego Supercomputer Center is evolving beyond a traditional high-performance computing system into an AI-for-science platform that enables researchers to integrate artificial intelligence into scientific discovery. TSCC provides campus researchers with access to computational resources through both Condo (system purchase) and Hotel (pay-as-you-go) programs, offering NVIDIA A100, A40, RTX6000 Ada, RTX 3090 Ti, and V100 GPUs in a Rocky Linux and Slurm-managed environment. In addition to a broad suite of scientific software and libraries, TSCC supports modern AI workflows for research across domains such as computational biology, genomics, molecular simulation, and data-intensive science.
This poster introduces the AI resources available to TSCC users, including access to open-weight large language models (LLMs) through the National Research Platform (NRP) and the TritonAI Developer API at UC San Diego. NRP provides a robust and continuously expanding collection of openly accessible foundation models hosted on national-scale cyberinfrastructure, enabling researchers to experiment with state-of-the-art LLMs without managing their own deployments. For researchers requiring greater flexibility, TSCC supports secure deployment of user-hosted models through vLLM and reverse-proxy services, enabling private and isolated AI environments on campus GPU resources for custom models, sensitive research workflows, and advanced experimentation.
The poster also highlights the use of agentic coding tools, including Claude Code, OpenCode, and Crush, running directly within the TSCC environment. These tools can operate inside researchers’ project directories and software environments, allowing them to assist with software development, workflow creation, debugging, code optimization, Slurm job submission, and scientific computing tasks while working directly with users’ applications, scripts, and datasets. Example use cases demonstrate how AI-assisted development can accelerate scientific workflows and improve researcher productivity.
Finally, we present TSCC’s consulting, training, and research collaboration opportunities. Beyond providing computing resources, TSCC offers user support for HPC and AI technologies, including assistance with software installation and configuration, lowering the barriers for domain scientists to adopt AI methods in their research. Researchers can also collaborate with SDSC cyberinfrastructure experts on scientific computing, large-scale data analysis, AI-enabled workflows, and emerging computing technologies, helping bridge advances in artificial intelligence with real-world scientific research.