Implementation of parallel algorithm for heat equation using SkelGIS library, CUDA and SISAL

In this paper we present three different algorithms based on finite difference method (FDM) for solving heat equation. The common point of these algorithms is that all of them have been developed in parallel. While, each one of them has been differently implemented using SkelGIS library, Streams and Iterations in Single Assignment Language (SISAL) and Compute Unified Device Architecture (CUDA). Halstead metrics have been used for evaluating the programs.

[1]  David C. Cann,et al.  A Report on the Sisal Language Project , 1990, J. Parallel Distributed Comput..

[2]  Yao Zhang,et al.  Parallel Computing Experiences with CUDA , 2008, IEEE Micro.

[3]  David C. Cann,et al.  SISAL versus Fortran: a comparison using the Livermore Loops , 1990, Proceedings SUPERCOMPUTING '90.

[4]  Vikas Kumar,et al.  Designing a parallel algorithm for Heat conduction using MPI, OpenMP and CUDA , 2015, 2015 National Conference on Parallel Computing Technologies (PARCOMPTECH).

[5]  A. P. Wim Böhm,et al.  Reducing communication by honoring multiple alignments , 1995, ICS '95.

[6]  Olivier Pironneau,et al.  Introduction to Scientific Computing , 1998 .