high performance computing on graphics processing units: hgpu.org

Using GPUs for general purpose programming is, nowadays, much easier than the previous years. In the very beginning were Brook-GPU or Close To Metal the approaches used for exploring the new possibilities of hardware accelerators. After that, CUDA and OpenCL were released. They had been adopted by many programmers due to theirs advantages but, however, both of them are quite complex for beginners. We need to find a way to leverage the programming effort otherwise, developers will spend most of their time focusing on device-specific code instead of implementing algorithmic enhancements. The recent advent of the OpenACC standard [1] for heterogeneous computing represents an effort in this direction. Users only need to annotate the parallel regions to be offloaded into a GPU. We have developed our own version of the standard, accULL. The main aim of this work has been to create a comparison between accULL and two different compilers that support the new standard: PGI an CAPS.