Optimizing Code by Selecting Compiler Flags using Parallel Genetic Algorithm on Multicore CPUs

The compiler optimization phase ordering not only possesses challenges to compiler developer but also for multithreaded programmer to enhance the performance of Multicore systems. Many compilers have numerous optimization techniques which are applied in predetermined ordering. These ordering of optimization techniques may not always give an optimal code further it is impossible to find a unanimous optimization phase ordering that will produce the best code. Finding an optimal sequence for even a simple code is not easy task since the search space for attempting optimization phase sequences is very huge. The focus of this paper is to implement an optimal compiler phase ordering using parallel genetic algorithm exploiting the advances of parallel programming on Multicore processors. Inherently Genetic algorithms are very well suited for modeling multiple individuals. Three methods were adopted in testing and comparing the performance of the Parallel Benchmark programs, firstly the Benchmarks were compiled without applying optimization flags, next the Benchmarks were compiled by randomly selecting the optimization flags and finally a parallel genetic algorithm was used to set the Compiler optimization flags for compilation. The parallel genetic algorithm outperformed the other two methods in this work. [email protected]