A Gamma-calculus GPU-based parallel programming framework

The General Purpose GPU computational model changes the way parallel processing can be achieved. It is becoming more attractive to carry out parallel tasks on GPU devices. The sequential part of the application runs on the CPU whereas the computationally-intensive part is accelerated by the GPU. GPUs provide a multithreaded high level of parallelism with hundreds of cores. For high performance computing developers, the GPU cores offer a higher magnitude order of raw computation power than CPU. In this paper we propose an efficient parallel programming framework based on the GPU devices. This framework adopts the Gamma formalism as an abstract model for making parallelism less difficult. The software developer has only to specify the action to be curried-out on any atomic portion of data. The framework will then run the given action simultaneously on the GPU cores.