MetaFork: A Metalanguage for Concurrency Platforms Targeting Multicores

Abstract. We present MetaFork, a metalanguage for multithreaded algorithms based on the fork-join concurrency model and targeting mul-ticore architectures. MetaFork is implemented as a source-to-source compilation framework allowing automatic translation of programs from one concurrency platform to another. The current version of this frame-work supports CilkPlus and OpenMP. We evaluate the benefits of the MetaFork framework through a series of experiments, such as nar-rowing performance bottlenecks in multithreaded programs. Our exper-iments show also that, if a native program, written either in CilkPlus or OpenMP, has little parallelism overhead, then the same holds for its OpenMP or CilkPlus counterpart translated by MetaFork.