Our research group modifies existing programs to use the GPU to obtain significantly better performance. On a typical project, we work with researchers in other disciplines to understand the problem they are trying to solve and develop an efficient program for solving their problem. We then modify the program to run significantly faster by using the multiple cores in a GPU. This is done using the CUDA programming API or the OpenCL programming API. The URL given above will soon have a description of one of our projects.
Ability to program in C or C++. Knowledge of multi-threaded programming helpful but not required.
There is no specific starting or ending date. We have on-going projects and new projects starting. Undergraduate students can start in any semester and continue for multiple semesters according to their interest.
Primary duties are writing and debugging C or C++ code. Some programming of multiple cores and participation in group meetings to understand the goals of a particular implementation effort.
November 11, 2013