Optimizing Off-Chip Memory Access Costs in Low Power MPEG-4 Decoder

Haitham Habli, Johan Ersfolk, Johan Lilius, T Westerlund, J Nurmi

    Forskningsoutput: Kapitel i bok/konferenshandlingKonferensbidragVetenskapligPeer review

    Sammanfattning

    One of the main sources for energy consumption in modern video coding algorithms is caused by reading reference frames from main memory. Once a reference frame block is in the local memory, we need to predict in advance when the same reference frame block would be reread from the main memory. The prediction relies on the calculation of motion vectors according to their absolute addresses. The reduction of the number of read accesses can be attained by calculating which motion vectors point to a part of the reference frame that can be copied to a local memory. This increases data locality, and thus reduces energy consumption. The optimized memory access algorithm was integrated into low power MPEG-4 decoder architecture [1, 2] and modeled using SystemC. A cycle accurate model simulation with optimized algorithm showed an average of 10% increase on runtime performance compared to an unoptimized typical video decoding algorithm.
    OriginalspråkOdefinierat/okänt
    Titel på gästpublikation3rd International Conference on Information and Communication Systems
    FörlagACM Press
    Sidor
    Antal sidor5
    ISBN (tryckt)978-1-4503-1327-8
    DOI
    StatusPublicerad - 2012
    MoE-publikationstypA4 Artikel i en konferenspublikation
    EvenemangInternational Conference on Information and Communication Systems - International Conference on Information and Communication Systems
    Varaktighet: 3 maj 20125 maj 2012

    Konferens

    KonferensInternational Conference on Information and Communication Systems
    Period03/05/1205/05/12

    Nyckelord

    • Memory access
    • fine-grained
    • heterogeneous
    • low power
    • motion vector
    • reusable data
    • video decoder

    Citera det här