Strategy for Improving Memory Locality Reuse and Exploiting Hidden Parallelism

In recent years, methods for analyzing and parallelizing sequential code using data analysis and loop transformations have been developed. These techniques have proved remarkably successful, and have been used to move from sequential to parallel codes, or to improve efficiency of existing parallel codes. Our research focuses on Fortran code optimisation for parallelisation in Shared Memory architectures by using data analysis and loop source-to-source transformations. Our optimisation strategy, although designed for OpenMP directives, is sufficiently general to be used for pure Fortran code. Our algorithm has been implemented as a tool called Automatic Guidance Module (AGM), and have received high evaluation scores from our industrial partner