• DocumentCode
    2994191
  • Title

    Performance and power consumption trade-offs for a VLIW DSP

  • Author

    Ibrahim, Mostafa E A ; Rupp, Markus ; Habib, S. E -D

  • Author_Institution
    Inst. of Commun. & RF Eng., Vienna Univ. of Technol., Vienna, Austria
  • fYear
    2009
  • fDate
    9-10 July 2009
  • Firstpage
    1
  • Lastpage
    4
  • Abstract
    Since performance and power consumption optimizations are crucial issues in embedded systems, it is necessary to find a trade-off between these optimization goals. This paper explores the performance and power trade-offs of VLIW processor, specifically the Texas Instruments TMS320C6416T DSP. We evaluate the effect of the global performance optimizations as well as a specific architecture feature on the power consumption of the targeted processor while running typical digital signal and image processing algorithms. We assess the specific C64x+ architecture feature, software pipelined loop (SPLOOP), effect on the power consumption and the performance as well. The binaries used in this study were generated using the Texas Instrument C/C++ Compiler, which allows control over the whole set of optimizations.
  • Keywords
    digital signal processing chips; embedded systems; multiprocessing systems; pipeline processing; power aware computing; program compilers; C64x+ architecture; Texas Instrument C/C++ Compiler; Texas Instruments TMS320C6416T DSP; VLIW DSP; digital signal processing algorithm; embedded systems; image processing algorithm; power consumption optimizations; software pipelined loop; Computer architecture; Digital signal processing; Embedded system; Energy consumption; Image processing; Instruments; Optimization; Signal processing; Software performance; VLIW;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Signals, Circuits and Systems, 2009. ISSCS 2009. International Symposium on
  • Conference_Location
    Iasi
  • Print_ISBN
    978-1-4244-3785-6
  • Electronic_ISBN
    978-1-4244-3786-3
  • Type

    conf

  • DOI
    10.1109/ISSCS.2009.5206135
  • Filename
    5206135