Examples from computational geometry that demonstrate the potential of using the thrust library to implement parallel processing on GPUs
Erik Wynters · Journal of computing sciences in colleges · 2013
This paper uses two classic problems in computational geometry to illustrate how GPUs can be used to achieve large speedups with minimal effort. The Thrust library, which supports rapid, high-level parallel coding on GPUs, is used to accelerate simple brute-force algorithms for solving these problems. Familiarity with parallel processing is increasingly necessary for software developers, and the ideas presented here can be used as in-class examples or as programming assignments that can introduce students to parallel processing using GPUs in appropriate college courses. The problems are easy to visualize and understand and demonstrate very well the speedups that can be achieved with parallel processing on GPUs.