Accelerating Scientific Application Performance via GPUs and the Xeon Phi

Temporary Supervisor

Associate Professor Peter Strazdins

Description

Scientific applications ranging from simple advection solvers to global atmosphere models may benefit greatly from acceleration by devices such as GPUs and the Xeon Phi.

Goals

This project will investigate the techniques and effectiveness of using accelerators to improve the performance of applications of interest. Candidates include a resilient solver for advection to the Met UM atmosphere model used for operational forecasts by the Bureau of Meteorology.

Requirements

Prior GPU experience is recommended for 12-unit project courses.

Updated:  1 June 2019/Responsible Officer:  Dean, CECS/Page Contact:  CECS Marketing