VideoHelp Forum




+ Reply to Thread
Results 1 to 9 of 9
  1. The original article from New Scientist... http://www.newscientist.com/news/news.jsp?id=ns99994274

    I wouldn't mind a few of these...

    ------------------------------------------

    New chip gives PCs supercomputing muscle

    18:04 14 October 03

    NewScientist.com news service


    A computer chip that will enable personal computers to perform some calculations as fast as some supercomputers was unveiled on Tuesday.

    Developed by ClearSpeed Technologies, based in California, the CS301 chip is capable of 25 gigaflops - 25 billion "floating point" calculations per second. These arithmetical calculations are also a common measure of computing power.

    A desktop Pentium processor operates at a few hundred million flops, while some of the most powerful computers in the world operate at few hundred gigaflops. Putting around 20 ClearSpeed chips into a few personal computers could potentially provide the sort of power normally only found in a supercomputer built from hundreds of parallel processors or specialised hardware.

    The CS301 works as a supplementary component to a regular processor. A chipset carrying one or two of the chips can be plugged into a normal PC like a graphics card and perform intensive calculations on behalf of the machine's normal processor. The chip is also very power-efficient, consuming only three watts and ClearSpeed is working on a version for laptop computers.

    "The goal here is to enhance supercomputers at one level," says Tom Beese, CEO of ClearSpeed. "But also to deliver a power-efficiency that means you can put a few of chips inside a laptop, running along side a Pentium, and have a gigaflop laptop."


    Protein modelling

    The CS301 would be especially suited to arithmetically intensive scientific applications such as protein modelling or geological data analysis. Beese says the chip is fast and efficient because it has been designed almost entirely to focus on performing mathematical calculations with around 70 per cent of its surface dedicated to number crunching.

    ClearSpeed plans to start selling a PC-compatible version of the microprocessor to research companies and universities within the next few months. A price has yet to be finalised but Beese says a single chip will initially cost around $16,500.

    Many supercomputers are built from large arrays of off-the-shelf processors, although there is also a growing return to the use of specialised hardware. The world's fastest supercomputer, NEC's Earth Simulator, is made from specialised components. It is theoretically capable of 35 thousand gigaflops or 35 trillion floating point operations per second.

    Details of the CS301 chip will be announced at the Microprocessor Forum 2003, which takes place in California this week.


    Will Knight


    -----------------------------------------

    Regards.
    Michael Tam
    w: Morsels of Evidence
    Quote Quote  
  2. Член BJ_M's Avatar
    Join Date
    Jul 2002
    Location
    Canada
    Search Comp PM
    i don't get it -- if it plugged into the pci bus (even pci2) - its going to be only passing data at the speed of the bus --- which isnt that fast in comparison with say the memory buss.

    and $16,500 isnt chicken feed ...
    "Each problem that I solved became a rule which served afterwards to solve other problems." - Rene Descartes (1596-1650)
    Quote Quote  
  3. Member The village idiot's Avatar
    Join Date
    Apr 2002
    Location
    Adrift among the STUPID
    Search Comp PM
    All that in a PCI slot? Is there enough bandwidth to support that many calculations? Maybe I'll look for info later.
    Hope is the trap the world sets for you every night when you go to sleep and the only reason you have to get up in the morning is the hope that this day, things will get better... But they never do, do they?
    Quote Quote  
  4. Yes, that is so. The real advantage will be putting about 30 of those on a card, say one for each frame in a GOP and letting it process an entire GOP.

    Of course a low end P4 chip on the same card to coordinate them wouldn't hurt.

    I would like to see a TMPEnc that supported this. Of course, I'm still waiting on a TMPEnc that does AC3 encoding also
    Just what is this reality thing anyway?
    Quote Quote  
  5. Member Faustus's Avatar
    Join Date
    Apr 2002
    Location
    Dallas, TX
    Search Comp PM
    Ok so chip=fast bus=@ss. I don't get the point.
    Quote Quote  
  6. Член BJ_M's Avatar
    Join Date
    Jul 2002
    Location
    Canada
    Search Comp PM
    Originally Posted by i_am_dave
    Yes, that is so. The real advantage will be putting about 30 of those on a card, say one for each frame in a GOP and letting it process an entire GOP.

    Of course a low end P4 chip on the same card to coordinate them wouldn't hurt.

    I would like to see a TMPEnc that supported this. Of course, I'm still waiting on a TMPEnc that does AC3 encoding also
    hmmm - $495,500 plus cost of system to support it (lets say another $150,000) = $645,000

    to RUN TMPGenc ? :P

    for $645,000 i can buy right now a 128 proccessor sgi (comes with 192GB of memory and of course 64bit cpu's) with 8 real time rendering pipes (real time HD in fact) .. and that would be rendering from raw data (as from putting textures on wireframes) or from stored or piped in data ..
    "Each problem that I solved became a rule which served afterwards to solve other problems." - Rene Descartes (1596-1650)
    Quote Quote  
  7. Member The village idiot's Avatar
    Join Date
    Apr 2002
    Location
    Adrift among the STUPID
    Search Comp PM
    Originally Posted by BJ_M
    Originally Posted by i_am_dave
    Yes, that is so. The real advantage will be putting about 30 of those on a card, say one for each frame in a GOP and letting it process an entire GOP.

    Of course a low end P4 chip on the same card to coordinate them wouldn't hurt.

    I would like to see a TMPEnc that supported this. Of course, I'm still waiting on a TMPEnc that does AC3 encoding also
    hmmm - $495,500 plus cost of system to support it (lets say another $150,000) = $645,000

    to RUN TMPGenc ? :P

    for $645,000 i can buy right now a 128 proccessor sgi (comes with 192GB of memory and of course 64bit cpu's) with 8 real time rendering pipes (real time HD in fact) .. and that would be rendering from raw data (as from putting textures on wireframes) or from stored or piped in data ..
    But think how fast it would render if you were only using 2 pipelines
    Hope is the trap the world sets for you every night when you go to sleep and the only reason you have to get up in the morning is the hope that this day, things will get better... But they never do, do they?
    Quote Quote  
  8. Член BJ_M's Avatar
    Join Date
    Jul 2002
    Location
    Canada
    Search Comp PM
    only 1/4 the speed // sgi's work with a unified memory graphics system .. the more piplines - the greater the speed ...
    "Each problem that I solved became a rule which served afterwards to solve other problems." - Rene Descartes (1596-1650)
    Quote Quote  
  9. Member
    Join Date
    Aug 2003
    Location
    NE Ohio
    Search Comp PM
    Originally Posted by BJ_M
    i don't get it -- if it plugged into the pci bus (even pci2) - its going to be only passing data at the speed of the bus --- which isnt that fast in comparison with say the memory buss.

    and $16,500 isnt chicken feed ...
    I am NO expert on this.....

    I believe that these special applications like "protein modelling or geological data analysis" use a pretty ordinary computer on the front end and the supercomputer to do lots of calculations. So the PC only sends a short query (that requires lots of calcs) to the supercomputer... the supercomputer does the crunching and then sends back a small answer.

    Therefor the bus speed is not inportant.

    check out this to see what I mean:
    http://www.stanford.edu/group/pandegroup/folding/

    JSB
    Quote Quote  



Similar Threads

Visit our sponsor! Try DVDFab and backup Blu-rays!