View Item 
        •   Utrecht University Student Theses Repository Home
        • UU Theses Repository
        • Theses
        • View Item
        •   Utrecht University Student Theses Repository Home
        • UU Theses Repository
        • Theses
        • View Item
        JavaScript is disabled for your browser. Some features of this site may not work without it.

        Browse

        All of UU Student Theses RepositoryBy Issue DateAuthorsTitlesSubjectsThis CollectionBy Issue DateAuthorsTitlesSubjects

        Coarsening functional parallelism using intelligent search algorithms

        Thumbnail
        View/Open
        thesis Oscar Leijendekker.pdf (362.1Kb)
        Publication date
        2016
        Author
        Leijendekker, O.C.
        Metadata
        Show full item record
        Summary
        The increase in parallelism in modern-day computer architectures requires programs capable of exploiting that parallelism. With the goal of automating the process of creating parallel implementations, we want to extract implicit parallelism from a program. This can be done easily when dealing with functionally pure languages but doing so may generate a parallel implementation that spends more time on communication than it saves on computation. To this end, we present a method capable of iteratively reducing the amount of parallelism until a proper balance between computation and communication is reached. Our method may be able to find more useful implicit parallelism than simply selecting tasks based on their size, as has been done in previous work.
        URI
        https://studenttheses.uu.nl/handle/20.500.12932/22816
        Collections
        • Theses
        Utrecht university logo