Alan Humphrey - Research Computer Scientist
Most of my past research has been focused on formal verification of concurrent systems, specifically the Message Passing Interface (MPI) and verification tools like In-situ Partial Order - ISP and its integration within the Eclipse Parallel Tools Platform (PTP), to which I am a committer. I have also studied the application of SMT solvers to symbolic verification tools for CUDA C kernels. Most of this research was done under the advisement of Dr. Ganesh Gopalakrishnan while completing a Bachelor of Science in Computer Science from the University of Utah.
I am a software developer at the SCI Institute, and also working on a PhD in Computer Science at the University of Utah. I work with Dr. Martin Berzins to improve the scalability and performance of the the Uintah Computational Framework (UCF). Uintah now demonstrates excellent strong and weak scalability up to 256,000 cores on the DOE Titan system and up to 768,000 cores on DOE Mira system for challenging fluid-structure interaction problems using adaptive mesh refinement (AMR). I've been responsible for migrating Uintah to hybrid CPU/GPU architectures such as NSF's Keeneland System and DOE Titan with the development of a heterogeneous task scheduler and runtime system that allows Uintah to dynamically dispatch computational tasks to both CPU cores and available GPUs on-node. My recent focus has been on the development of a scalable GPU-based radiation transport model to run on the heterogeneous, DOE Titan system in preparation for machines like DOE Summit. This work is driven by the target problem for the Utah Carbon-Capture Multidisciplinary Simulation Center, an NNSA, PSAAP II Center funded from March 2014. This project aims to use simulation science at petascale and eventually exascale to accelerate the design of the next generation of clean coal boilers that will improve clean coal technologies.
As part of a Multi-Scale Multidisciplinary Modeling of Electronic Materials Collaborative Research Alliance funded by the U.S. Army Research Laboratory, I am also working to incorporate molecular dynamics capabilities into Uintah. The goal of this work is to develop quantitative understanding of materials from the smallest to the largest relevant scales to advance the state of the art in electronic, optoelectronic and electrochemical materials and devices.