UCSD Aims to Make HPC an ‘Everyday Toolbox’ for Undergrads, Not a ‘Rare Privilege’

Key Takeaways

• Undergrads gain hands‑on access to Expanse supercomputer
• Course covers molecular mechanics, dynamics, Monte Carlo simulations
• NSF funds allocation transforms teaching platform
• Expanse hosts 93k CPU cores, 208 V100 GPUs
• Program prepares students for AI‑centric careers

Summary

The University of California, San Diego has launched a program that gives undergraduate students direct access to Expanse, the university’s 5‑petaflop supercomputer. Partnering the School of Computing, Information and Data Sciences with the San Diego Supercomputer Center, the initiative integrates high‑performance computing into a nano‑engineering course. Funded by an NSF allocation, students run molecular dynamics, Monte Carlo and other simulations on over 93,000 CPU cores and 208 Nvidia V100 GPUs. The effort aims to make HPC a routine tool for future engineers, not a niche privilege.

Pulse Analysis

High‑performance computing has traditionally been confined to graduate labs and national facilities, leaving a skills vacuum at the undergraduate level. As data‑intensive research and AI accelerate, employers increasingly demand engineers who can scale simulations and process massive datasets. UC San Diego’s initiative directly addresses this gap by embedding HPC resources into a core nano‑engineering class, allowing students to transition from textbook theory to real‑world computational experiments.

The centerpiece of the program is Expanse, a 5‑petaflop system featuring more than 93,000 CPU cores, 208 Nvidia V100 GPUs, 220 TB of DRAM and 810 TB of NVMe storage, all linked by a 100 GB/s HDR InfiniBand fabric. Leveraging NSF‑provided ACCESS allocations, the course enables undergraduates to perform molecular mechanics, energy minimization, statistical mechanics, molecular dynamics and Monte Carlo simulations. Faculty and graduate mentors guide students through model building, result interpretation, and iterative design, mirroring workflows used in national labs and industry R&D.

Beyond technical training, the program signals a broader shift toward democratizing advanced computing in higher education. By exposing a larger, more diverse student body to HPC and AI tools, UC San Diego cultivates a pipeline of talent ready for data‑driven roles in academia, government, and the private sector. The model also encourages other institutions within the California university system to adopt similar curricula, potentially reshaping the national workforce’s computational competency. As more universities follow suit, HPC may become as commonplace in undergraduate labs as standard desktop software today.