Automatic program parallelization for multicore processors
Institute of Applied Informatics, Wroclaw University of Technology, 50-370 Wroclaw, Wybrzeze Wyspianskiego 27, Poland
In Proceedings of the 8th international conference on Parallel processing and applied mathematics: Part I (2010), Volume 6067/2010, pp. 236-245
@article{kwiatkowski2010automatic,
title={Automatic Program Parallelization for Multicore Processors},
author={Kwiatkowski, J. and Iwaszyn, R.},
journal={Parallel Processing and Applied Mathematics},
pages={236–245},
year={2010},
publisher={Springer}
}
With the advent of multi-core processors the problem of designing application that efficiently can utilize it performance become more and more important. Moreover developing programs for these processors requires from the programmers some additional, specific knowledge about the processor architecture. In multi-core systems efficient program execution is the main issue. It can even happen that switching from sequential to parallel computation can lead to decreasing of performance. The paper deals with the short description of SliCer, the hardware independent tool that parallelizes serial programs in automatic way depending on the number of available processing units by creating the proper number of threads that can be later execute in parallel.
February 8, 2011 by hgpu
