Wayne Mitchell works with Professor Tom Manteuffel on low-communication, parallel multigrid algorithms. While multigrid algorithms are the current method of choice for solving or preconditioning many large linear systems (especially those arising from the discretization of elliptic PDEs) in high-performance computing codes, the communication cost associated with these algorithms is one of the major challenges for their continued scalability as modern machines continue to add parallelism. His work focuses on developing and studying the convergence of novel algorithms which limit communication by increasing local computation on each processor.
Mitchell is originally from Dallas, Texas. He earned his undergraduate degree from Loyola University in New Orleans. There, he double majored in jazz studies (tenor saxophone) and mathematics. In his free time, he climbs, mountain bikes, skis, dances, and plays music.