Aug. 8, 2024 — The visionary team behind the 2004 paper, “GPU Cluster for High Performance Computing,” has been awarded this year’s SC Test of Time Award. Led by Professor Arie Kaufman of Stony Brook University, the team’s work laid the foundation for the GPU-powered supercomputing landscape that dominates every major national lab and every single one of the top supercomputers in the world.

The award-winning team included Kaufman, Dr. Suzanne Yoakum-Stover, Dr. Zhe Fan, and Dr. Feng Qui. Their research is perhaps the first documented application of GPUs for high-performance computing (HPC), a concept that has since become standard practice in HPC and artificial intelligence. More than the common standard, GPU computing today represents the largest paradigm shift to data center computing in decades.

Reflecting on their work, Fan recalled: “I remember someone visited our VisLab, listened to us about this work, and commented, ‘Researchers like to do crazy things.’ At that time, it did feel a little bit crazy. But looking back at it now, the idea of GPU clusters was obvious and inevitable.”

The team’s innovative approach was inspired by the improved FLOPs per dollar that GPUs provided compared to traditional CPU hardware. To leverage this, they created a cluster of 30 GPU nodes to simulate airborne contaminant dispersion in New York City’s Times Square. This simulation achieved speeds 4.6 times faster than CPU cluster implementations of the time, showcasing the immense potential of GPU computing.

Yoakum-Stover fondly remembers the collaborative spirit that drove their success.

“Under Professor Kaufman’s leadership, I remember the VisLab as a hotbed of innovation with a diversity of brilliant, hard-working students,” she noted. “I’m sure I learned more from them than they did from me, and I am grateful and proud to have been part of VisLab history.”

The legacy of innovation continues among the team members. At Mission Focus, Yoakum-Stover serves as chief scientist, where she oversees the development of data analytics products deployed on infrastructure that has its roots in the work conducted at VisLab. Meanwhile, Fan has forged a successful career at Google, currently applying his expertise to the cutting-edge field of autonomous vehicles with the Waymo division.

In the medical sector, Qui contributes his skills at Siemens Medical Solutions, focusing on advancing medical imaging and visualization software. Finally, Kaufman continues his research career, where he is one of the most cited researchers in visualization of all time.

Congratulations to Professor Kaufman and his team on receiving this prestigious award!

Paper Abstract

Inspired by the attractive Flops/dollar ratio and the incredible growth in the speed of modern graphics processing units (GPUs), we propose to use a cluster of GPUs for high performance scientific computing. As an example application, we have developed a parallel flow simulation using the lattice Boltzmann model (LBM) on a GPU cluster and have simulated the dispersion of airborne contaminants in the Times Square area of New York City. Using 30 GPU nodes, our simulation can compute a 480×400×80 LBM in 0.31 second/step, a speed which is 4.6 times faster than that of our CPU cluster implementation. Besides the LBM, we also discuss other potential applications of the GPU cluster, such as cellular automata, PDE solvers, and FEM.

Z. Fan, F. Qiu, A. Kaufman, and S. Yoakum-Stover, “GPU Cluster for High Performance Computing,” SC04: Proceedings of the 2004 ACM/IEEE Conference on Supercomputing, Pittsburgh, PA, USA, 2004, pp. 47-47, doi: 10.1109/SC.2004.26.

---

Source: Waleed Atallah, SC24