headerlogo
scyourway

SC Conference - Activity Details



Jacket: The GPU Engine for MATLAB

Authors:
Gallagher Pryor  (AccelerEyes)
Tauseef Rehman  (AccelerEyes)
John Melonakos  (AccelerEyes)
James Malcolm  (AccelerEyes)
Posters Session
Tuesday,  05:15PM - 07:00PM
Room Rotunda Lobby
Abstract:
In this work, we present a novel approach to GPU computing with our software, Jacket: The GPU Engine for MATLABĀ®. Jacket is not another GPU API nor is it another collection of GPU functions. Rather, it is simply an extension of the MATLAB language to new GPU data types, 'gsingle' and 'gdouble'. Jacket provides transparent overloading of MATLAB's CPU-based functions with CUDA-based functions. Jacket includes automated and optimized memory transfers and kernel configurations. Furthermore, Jacket uses a compile on-the-fly system that allows GPU functions to run in MATLAB's interpretive style. Jacket brings the speed and visual computing capability of the GPU to MATLAB programs. Further, Jacket lowers the barrier to entry to GPU computing for programmers. Within minutes of download, programmers, most of whom have never attempted GPU computing, are able to start running their code on the GPU. In our poster, we will provide a visual description of Jacket's novel components. Further, we will bring laptops to allow users to experience Jacket firsthand.
   IEEE Computer Society  /  ACM     2 0   Y E A R S   -   U N L E A S H I N G   T H E   P O W E R   O F   H P C