Inderscience PublishersInderscience PublishersInderscience Publishers
  PUBLISHERS OF DISTINGUISHED ACADEMIC, SCIENTIFIC AND PROFESSIONAL JOURNALS

Article Abstract

Title: Off-loading application controlled data prefetching in numerical codes for multi-core processors
  Author: J. Weidendorfer, C. Trinitis   Email author(s)
  Address: Institut fur Informatik, Technische Universitat Munchen, D-85747 Garching bei Munchen, Germany. ' Institut fur Informatik, Technische Universitat Munchen, D-85747 Garching bei Munchen, Germany
  Journal: International Journal of Computational Science and Engineering 2008 - Vol. 4, No.1  pp. 22 - 28
  Abstract: An important issue when designing numerical code in High Performance Computing is cache optimisation in order to exploit the performance potential of a given target architecture. This includes techniques to improve memory access locality as well as prefetching. Inherent algorithm constrains often limit the first approach, which typically uses a blocking technique. While there exist automatic prefetching mechanisms in hardware and/or compilers, they can not complement blocking with additional prefetching. We provide an infrastructure for off-loading application controlled prefetching on a chip multiprocessor, allowing to further improve numerical code already optimised by standard cache optimisation. Clear benefits are shown for real workloads on existing hardware.
  Keywords: chip multiprocessing; multi-core processors; cache optimisation; data prefetching; numerical codes; high performance computing.
  DOI: 10.1504/IJCSE.2008.021109
  Access for editors and complimentary subscribers       Access for Subscribers   Purchase this Paper        We welcome your comments about this paper Comment on the Paper