Heterogeneous parallel computing platforms and tools for compute-intensive algorithms: a case study

In this chapter we analyze the development of parallel programs for heterogeneous architectures, that is, workstation composed by multiple CPUs and accelerators, considering different programming models and tools, both commercial and freely available. In particular, we discuss the provided support and achievable performance for each of them with respect to some widely used, computationally intensive algorithms such as the convolution and the N-body algorithms. The aim is to provide a clear measure of the different efficiency figures with respect to the programming paradigm considered, the adopted tool, and the achievable peak performance.