The problem of Protein Structure Prediction (PSP) is known to be computationally expensive, which calls for the application of high performance techniques. In this project, parallel PSP algorithms found in the literature are being accelerated and ported to different parallel platforms, producing a set of algorithms that it is diverse in terms of the parallel architectures and parallel programming models used. The algorithms are intended to help other research projects and they have also been made publicly available so as to support the development of more elaborate prediction algorithms. We have thus far produced a set of 16 algorithms (mixing CUDA, OpenMP, MPI and/or complexity reduction optimizations); during its development, two algorithms that promote high performance were proposed, and they have been written in an article that was accepted in the International Conference on Computational Science (ICCS).
Workshop on Using Emerging Parallel Architectures for Computational Science
