CUDA PARALLELIZATION
This project involves parallelizing C/C++ programs using the CUDA (Compute Unified Device Architecture) programming language. CUDA is used for developing programs for NVIDIA GPUs. In this project, I will learn about GPU programming and will apply the knowledge gained to develop and test efficient parallel programs. The programs that will be developed in this project will serve as test-cases for the Interactive Parallelization Tool (IPT) that is currently under development at TACC. All the hand-on work required for this project will be done on the Stampede supercomputer at TACC which is equipped with the K20 GPUs.