Exploiting Parallelism Available in Loops Using Abstract Syntax Tree

Abstract

Performance of a program depends on two factors: better hardware of the executing machine and exploiting parallelism for concurrent execution. Loops with multiple iterations provide efficient parallelism in an application, and are used to reduce overall execution time and also to increase performance. Abstract Syntax Tree (AST) can be used as an effective tool for exploiting parallelism at the compiler level. This definitely saves time and automates the decomposition of a parallel job that is to be executed in parallel framework. AST can be used to detect loops in the source code, therefore this approach can be used to design a new parallel computing framework where simple codes written for normal machines can be parallelized by the framework itself.