A Survey of HPC Applications Appropriate for Execution on DSM and Grid Systems

Distributed memory Beowulf clusters and Grids are gaining more and more interest as low cost parallel architectures are meeting high performance computing demands. With the rise of cheap commodity cluster solutions, Distributed Memory approach has inspired many research projects and has lead to significant advances in this area. S oftware DS M (SDS M) systems enable the use of shared memory programming approaches in loosely coupled environments. S hared memory is appealing because it isolates the programmer from the underl ying communication scheme used to maintain consistency across the distributed system. On the other hand, modern organizations depend on network-attached personal desktop and mobile computer systems. A characteristic of these systems is that they are relatively resource rich (in terms of CPU power, memory, and disk capacity) but are utilized only for a fraction of the time during a day. Even during the time they are in use, their average utilization is much less than their peak capacity. Grid computing is basically coordinated resource sharing of these existing underutilized resources. These computing systems usually support embarrassingly parallel applications which consist of many instances of the same computation with their own data. The applications are usually involved with scientific problems, which need large amounts of processing capacity over long periods of time. In this paper we have surveyed various applications which can run on DS M and Grid environments individually. Based on the analysis of these applications we propose to identify applications which can be executed in an integrated model comprising of shared memory and grid-S MIG (Shared Memory Integrated with Grid). We are in the process of architecting and designing S MIG model and the applications we select will prove useful in testing this model.