• DocumentCode
    2448072
  • Title

    A framework for FPGA functional units in high performance computing

  • Author

    Koltes, Andreas ; O´Donnell, John T.

  • Author_Institution
    Dept. of Inf. & Math., Univ. of Passau, Passau, Germany
  • fYear
    2010
  • fDate
    19-23 April 2010
  • Firstpage
    1
  • Lastpage
    8
  • Abstract
    FPGAs make it practical to speed up a program by defining hardware functional units that perform calculations faster than can be achieved in software. Specialised digital circuits avoid the overhead of executing sequences of instructions, and they make available the massive parallelism of the components. The FPGA operates as a coprocessor controlled by a conventional computer. An application that combines software with hardware in this way needs an interface between a communications port to the processor and the signals connected to the functional units. We present a framework that supports the design of such systems. The framework consists of a generic controller circuit defined in VHDL that can be configured by the user according to the needs of the functional units and the I/O channel. The controller contains a register file and a pipelined programmable register transfer machine, and it supports the design of both stateless and stateful functional units. Two examples are described: the implementation of a set of basic stateless arithmetic functional units, and the implementation of a stateful algorithm that exploits circuit parallelism.
  • Keywords
    coprocessors; field programmable gate arrays; FPGA functional units; I/O channel; VHDL; coprocessor; generic controller circuit; high performance computing; pipelined programmable register transfer machine; register file; specialised digital circuits; stateful arithmetic functional units; stateless arithmetic functional units; Application software; Communication system control; Coprocessors; Digital circuits; Field programmable gate arrays; Hardware; High performance computing; Parallel processing; Registers; Software performance; FPGA; interface;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Parallel & Distributed Processing, Workshops and Phd Forum (IPDPSW), 2010 IEEE International Symposium on
  • Conference_Location
    Atlanta, GA
  • Print_ISBN
    978-1-4244-6533-0
  • Type

    conf

  • DOI
    10.1109/IPDPSW.2010.5470769
  • Filename
    5470769