O problema do fractal de Mandelbrot como comparativo de arquiteturas de memória compartilhada GPU vs OpenMP

In many areas of science and engineering large computational problems are common. Parallel processing techniques are use to solve this problems. This work show the characteristics of parallel processing using graphics cards from Nvidia, with the help of CUDA framework. The CUDA/GPU techniques will be applied to the generation of Mandelbrot’s fractal problem. In addition we generate results using OpenMP techniques. We will show serial and parallel implementations, as well as validation of codes and comparative performance between approaches.

Bruno Pereira dos Santos , Dany Sanchez Dominguez, Esbel V. Orellana.

Contatcs: bruno.ps@live.com, dsdominguez@gmail.com, evalero@uesc.br

Founding agency: PIBIC-CNPq.