1. This forum section is a read-only archive which contains old newsgroup posts. If you wish to post a query, please do so in one of our main forum sections (here). This way you will get a faster, better response from the members on Motherboard Point.

Advantages of Dsp Processor over Risc

Discussion in 'Embedded' started by swami, Jul 19, 2007.

  1. swami

    swami Guest

    list some of the advantages of using dsp processor over risc processor
    for applications like mpe4 encoder.

    what i feel is dsp processor executes more instructions in a single
    but this consumes more processing power.

    can some body put much more strong reason for it.
    swami, Jul 19, 2007
    1. Advertisements

  2. swami

    Tim Wescott Guest

    DSP processors are optimized for implementing algorithms that feature a
    mix of "regular algorithmic stuff" plus a large number of highly
    repetitive tasks, like vector dot products (a FIR filter implementation
    is essentially a vector dot product). All of the DSP chips in current
    production that I know of will do a vector dot product in one clock
    cycle per element in the vector; they scream at this.

    As a consequence of this focus on vector dot products and other highly
    repetitive tasks, DSP processors are less able to do other things really
    well. Specifically, the ones that I have worked with have required
    large context switches to maintain the vector dot product state machine,
    or they will not execute an interrupt while they're performing a vector
    dot product, or they store their vector dot product state machine in
    registers that are inaccessible to to use code, which makes any kind of
    a normal 'RTOS-like' context switch impossible.

    RISC processors are optimized for implementing just about any algorithm
    that comes down the pike, without emphasis on any one thing.
    Consequently, they don't scream through vector dot products the way a
    DSP chip will, but they are at least as good, or better, at performing
    other operations, and they don't have the nasty context switching
    difficulties that DSP chips do, so an RTOS is more efficient to
    implement (if you feel you need one).

    I'm not sure how this applies to the efficiency of an MPEG4 encoder,
    particularly since there are a lot of different ways you can encode
    MPEG. It really boils down to whether you are doing the kinds of things
    that a DSP chip is good for -- if you are, it will be to your advantage
    to use one. If you're doing a lot of function calls, memory searches,
    indexing, and other related stuff, and not a lot of vector dot products,
    then a RISC chip will probably be just as good as any DSP, and cost you
    less in other ways.


    Tim Wescott
    Wescott Design Services

    Do you need to implement control loops in software?
    "Applied Control Theory for Embedded Systems" gives you just what it says.
    See details at http://www.wescottdesign.com/actfes/actfes.html
    Tim Wescott, Jul 19, 2007
    1. Advertisements

  3. swami

    Pete Fenelon Guest

    List homework topic. Kthxbye pls.

    Pete Fenelon, Jul 20, 2007
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.