Proceedings of the 22nd International Conference on Parallel Architectures and Compilation Techniques (2004)
Antibes Juan-les-Pins, France
Sept. 29, 2004 to Oct. 3, 2004
James E. Smith , University of Wisconsin - Madison
Ashutosh S. Dhodapkar , University of Wisconsin - Madison
Kyle J. Nesbit , University of Wisconsin - Madison
AC/DC is an adaptive method for prefetching data from main memory. The basic prefetch method divides the memory address space into equal-sized concentration zones (CZones), and uses a global history buffer to track and detect patterns in miss address "deltas" (differences between consecutive addresses) within each CZone. When simulated with a realistic desktop memory system, CZone prefetching with Delta Correlations (C/DC) outperforms four other previously proposed prefetching methods. C/DC yields an average performance improvement of 23 percent when compared with no prefetching.<div></div> Adaptively is then added to the basic method. A tuning algorithm dynamically configures the CZone size and prefetch degree (i.e. the amount of data prefetched) on a per program-phase basis. Adaptive re-configuration provides additional performance improvements of 4% over C/DC. Overall, the Adaptive CZone / Delta Correlation (AC/DC) method outperforms other methods studied by 10%.
James E. Smith, Ashutosh S. Dhodapkar, Kyle J. Nesbit, "AC/DC: An Adaptive Data Cache Prefetcher", Proceedings of the 22nd International Conference on Parallel Architectures and Compilation Techniques, vol. 00, no. , pp. 135-145, 2004, doi:10.1109/PACT.2004.10016