A Source to Source Automatic Parallelizing Compiler for a Cluster of Workstations

The large number of sequential programs that takes lot of computational time with single processor needs to be ported onto the cluster of workstations to harness its computational power and to decrease the running time of the programs. One way to achieve this is to rewrite the sequential programs that fit in message- passing paradigm, but this is a time consuming and error prone process. This paper describes a parallelizing compiler that automatically converts the sequential programs in to parallel programs for a COW and distributing the data in such a way that the communication between the processors is minimized. Since nested for loops are the core of scientific and engineering applications, which access large arrays of data, this paper deals with parallelizing the perfectly nested for loops present in the source code.