Cilk and MPI : the impact of the number of sub-tasks in parallel programming
SaeAm Chung, KyuHo Lee, Jongsu Yi, JunSeong Kim · ITC-CSCC :International Technical Conference on Circuits Systems, Computers and Communications · 2007
Inexpensive high performance computer systems combined with high speed networks and machineindependent communication libraries have made cluster computing a viable option for parallel application programs. In this paper, we examine the use of two different parallel programming models on a cluster system: MPI and Cilk. We consider both performance and programming complexity of the two systems across several common application programs. The experimental results show Cilk’s advantages in both performance and complexity when its dynamic work scheduling nature is applicable. MPI, however, can always receive a reward of competitive performance for some extra programming effort