Paul is an SRE and HPC automation engineer by day and wears his physicist hat by night to solve physics and engineering problems on computing systems using FOSS software. Motivated by a passion to bridge gaps between science, engineering, and computing, Paul stands on the proverbial shoulders of giants to demonstrate that difficult, time-consuming calculations can be codified to scale out and return desired results in a timely manner, and enjoys sharing that learning process with others.

Presentations

23x

Accelerating Physics: Multi-GPU Scaling, for Science!

Building on the fundamentals of multi-GPU architectures, this talk dives deep into solving the 2D Burger's equation—a cornerstone of fluid dynamics. I will explore the critical relationship between grid resolution and time-to-solution, demonstrating how to effectively leverage multi-GPU compute to optimize performance for complex engineering and physics problems.

See Presentation
22x

Solving Engineering Problems Using Multi GPU Computing

Wait, so you're saying adding more GPUs to a problem won't automatically speed things up? In this talk, I will present how enterprise datacenters, HPC shops, and home lab workflows can benefit from using multi GPU computing. Examples from computational fluid dynamics and AI/ML will highlight how leveraging multiple GPUs can reduce time to solution, and improve scalability for large, multidimensional domains. Attendees will gain insights into science's interest in GPU computing, problem decomposition, scaling to multiple GPUs, and code optimization with multiple GPUs in mind.

See Presentation