期刊名称：International Journal of Computer Science and Information Technologies
电子版ISSN：0975-9646
出版年度：2011
卷号：2
期号：6
页码：2924-2931
出版社：TechScience Publications
摘要：Numerous parallel programming tools have been developed so far for supporting parallel programs. This paper presents performance analysis of wide range of parallel programming simulation tools. This paper also compares the features of different tools. PVM and MPI are most widely used standards for parallel and distributed computing. MPI has better performance in high performance massively parallel processing (MMPs) computer systems to provide highly optimized and efficient implementations than PVM. In MMP, all of the processing elements are connected together to be one very large computer. This is in contrast to the distributed computing where massive numbers of separate computers, connected through a network, are used to solve a single large problem. PVM is most suitable in heterogeneous networks to gain optimal performance. One may favor the other tools depending on the need. With the help of our performance comparison one can choose which one would be the better for a particular application.
