WashU IT Research Infrastructure Services (RIS) new Compute2 high-performance computing cluster is already driving faster, more efficient research across disciplines. For one researcher, Yuanjian Zhang, who is working on global atmospheric simulations, the upgrades have made a measurable impact.
“Compute2 features the latest Intel Emerald Rapids CPUs, which are about 30% faster than those on Compute1,” he explained. “With faster internode communication and updated MPI libraries, performance improvements can reach 50–70% for large-scale simulations.”
His work relies on running massive MPI-based models that simulate global climate systems. Compute2’s improved architecture and expanded node capacity have proven critical for this kind of data-intensive research.
He also praised the system’s new Slurm scheduler and traditional HPC tools, such as modules and Spack. “I prefer the traditional workflow, but it’s great that Compute2 also supports Docker for those who need it. It offers more accessibility for everyone.”
Beyond performance, Compute2 also provides significant cost savings compared to commercial cloud services like AWS or Azure. “It’s faster, cheaper, and more accessible for WashU researchers,” he said.
With its latest-generation hardware, improved scheduling, and user-focused design, Compute2 is setting a new standard for computational research at WashU — enabling scientists to run complex models faster and more efficiently than ever before.
Ready to take your research to the next level? Those interested in taking their research to the next level are encourage to onboard now, and join the growing community of WashU researchers leveraging Compute2 for advanced computation and discovery.
For more information, check out the article: Compute2 – The Future of High-Performance Computing at WashU.