Sign in / up

back to article Nvidia dollars lifted by Intel's Sandy Bridge ramp, payola

Nvidia has frenemy Intel to thank for its improving financial situation in its first quarter of fiscal 2012. Despite the fact that the graphics and CPU chip maker's revenues and profits for its first quarter ended May 1 were down – revenues fell 4 per cent, to $962m, and net income was down 1.7 per cent, to $135.2m – Wall …

COMMENTS

House rules Send corrections
This topic is closed for new posts.
  1. RJ

    Hehehe

    Love the quip about buying the hachet in AMD

    0 0
  2. Anonymous Coward
    Anonymous Coward

    GTX590?????

    GTX590?? What GTX590?? Nvidia manufactured about 2000 and then they sold out. Go to Pricewatch.com and see for yourself. The GTX590 is NOT available.

    The GTX590 is vapourware.

    0 0
    1. Adam T
      Reply Icon

      Pricewatch.com? :)

      You need to broaden your horizons, sir!

      There's plenty available from 3rd party sellers on Amazon but, admittedly, you'll pay a premium.

      0 0
      1. Bounty
        Reply Icon

        Really?

        "There's plenty available from 3rd party sellers on Amazon but, admittedly, you'll pay a premium."

        I just searched and I see like two 590 models available, of which, one is out of stock. The one that is in stock, is a limited edition card that's way overpriced.

        0 0
  3. Matthew Taylor
    FAIL

    CUDA is a tricky customer

    I've had a few run-ins with CUDA, and the problem I've found is that you need to know a lot about the architecture of the card, in order to get decent performance out of it, e.g. the memory layout of a "warp" (a small batch of items to be processed), and how the card pipelines texture fetches.

    I found I could easily raise or lower my performance by a factor of 4-5 by adjusting some parameters, and it wasn't clear why. I ended up just trying loads of combinations and picking the best - but then on another card, that wouldn't give good results. NVidia need to do a lot more in the compiler to abstract that stuff away.

    Having said that, it was over a year ago, so things may well have improved in that regard these days.

    1 0
    1. E 2
      Reply Icon

      Different card/GPU

      You can ask the GPU how many cores it has using a CUDA API function and then choose the correct number of threads accordingly. Which is not to say that knowing the correct number of threads does not involve a bit of black magic.

      0 0
  4. Paul Shirley

    good news, less nvidia=less system bugs

    Nice to see nvidia finally give up pretending they can make mboard chipsets. The pain these idiots have caused me down the years doesn't bear thinking about. Don't remember ever seeing one with all the advanced features working and too many with simple things like sata and basic nic buggy.

    Their gpu drivers haven't been a lot better. The total unwillingness to fix PAL tv issues was a long running problem. Afaik they still aren't taking it seriously.

    2 0
  5. Adam T

    greedy

    "Frankly, I wish our Professional Solutions business was growing faster, too," admitted Jen-Hsun Huang, president and CEO at Nvidia, in the call.

    Perhaps if they stopped trying to fleece professional workstation users with overpriced cards which are actually outclassed by the consumer cards, they'd find themselves with a booming professional sector.

    Pro users have money, but they're not stupid. Climbing into bed with Autodesk and getting them to "qualify" the cards is a trick nobody's falling for anymore.

    3 0
This topic is closed for new posts.

Other stories you might like