Dell workstations start carrying Nvidia Tesla processors

Nvidia has landed a big design win with its Tesla “GPU computing processors.” According to the company, Dell has begun offering Tesla C1060 cards in its Precision R5400, T5500, and T5700 workstations.

The R5400 is a rack system, while the T5500 and T7500 are both desktop-sized towers based on Intel’s new Xeon 5500-series processors.

As for the Tesla C1060, adding it to one of those Dell workstations will involve paying a hefty $1,699 premium. That card packs 4GB of memory and 240 stream processors, making it a sort of supercharged, compute-only sibling of Nvidia’s GeForce GTX 200-series GPUs.

If you’re struggling to think of a real-world application for those workstations, Nvidia provides an example in its press release:

“National Instruments is developing the control system for the European Extremely Large Telescope project, which upon completion will be the world’s largest. To tackle this computational challenge, we developed a CUDA interface with LabVIEW to simulate and control the M1 mirror consisting of 984 individual segments,” said Jeff Meisel product manager for LabVIEW at National Instruments. “A Dell workstation equipped with a single Tesla C1060, can achieve near real-time control of the mirror simulation and controller, which before wouldn’t be possible in a single machine without the computational density offered by GPUs.”

That’s not it, of course. Nvidia points out that people have used its GPUs for general-purpose computing in areas like oil-and-gas processing, medical imaging, financial computing, “GeoSciences,” and computational chemistry.

Comments closed
    • bdwilcox
    • 11 years ago

    Now Timmy can develop nuclear warheads at home.

    • ssidbroadcast
    • 11 years ago

    Apple, please take notes. Give the Mac Pro a workstation-class GPU, and a Tesla. And don’t raise the price!!!! >:o

      • Meadows
      • 11 years ago

      You can only have one of those, you know.

      Only the first one.
      At a 40% chance.

      • UberGerbil
      • 11 years ago

      Psst, a Tesla is really just a GPU without the parts that can actually produce images — you know, like the “OpenCL processor” that Adi keeps fantasizing about. If you have a GPU already, you get the computing power /[

        • ssidbroadcast
        • 11 years ago

        Sure. In that case, the Mac Pro should have an entry-level Quadro or FireGL. Not some crappy GeForce 150 :/

    • lycium
    • 11 years ago

    until we are offered real ieee754 and virtualised memory, its applications are still limited.

    oh well, at least intel are doing the Right Thing with larrabee 🙂

      • djgandy
      • 11 years ago

      Yep Larrabee has a massive advantage here. Only half the performance when doing dp, whereas Nvidia’s implementation is about 1/12th.

        • Anonymous Coward
        • 11 years ago

        Though one would suspect that nNidia is rather ahead in actual performance.

        • Silus
        • 11 years ago

        Oh yeah, such a massive advantage for a non-existent product!

    • CheetoPet
    • 11 years ago

    Meh, my 386 had a co-processor. It did the maths real fast like.

    • CampinCarl
    • 11 years ago

    One of my friends just scored an internship at Battelle, and one of the first things he will be working on (apparently) is converting a lot of their x86/x64 code to CUDA.

    • Buzzard44
    • 11 years ago

    But can it fold?

      • Jigar
      • 11 years ago

      Yes it can, but you better get something else … 295GTX will kick it any day.

      • Dr_b_
      • 11 years ago

      Will it blend? that is the question.

        • SomeOtherGeek
        • 11 years ago

        No, will it disappear, that is the question…

    • _Sigma
    • 11 years ago

    Why is geo sciences in quotes?

      • BobbinThreadbare
      • 11 years ago

      It’s not real scientific field, it’s an encompassing term for what I’m guessing is geology, meteorology, and oceanography.

    • UberGerbil
    • 11 years ago

    g[<*[http://arstechnica.com/business/news/2009/04/why-processors-need-high-finance.ars<]§

      • Anomymous Gerbil
      • 11 years ago

      Not really. We’re about to replace >100 blades with 2 (or maybe 4) NVIDIA Tesla devices for our modelling. Just a toe in the water, but I very much doubt we’re alone. You don’t need to be “high frequency” traders to need a shedload of compute power for pricing and risk models.

        • Game_boy
        • 11 years ago

        What is the difference between Tesla and the much cheaper Geforce cards? If it’s drivers alone then surely building a custom application for the Geforce would circumvent that?

          • wibeasley
          • 11 years ago

          The main advantage is 4GB of RAM (vs 1 GB).

        • UberGerbil
        • 11 years ago

        Yes, but the point wasn’t that the financial guys were the only one to _[

      • Meadows
      • 11 years ago

      Computational chemistry is probably a funny official name for folding.

        • pedro
        • 11 years ago

        Computational chem is the brute force method of drug discovery.

          • wibeasley
          • 11 years ago

          Computational chemistry is more than than just pharmaceuticals. Here’s a popular software package that my wife uses: §[<http://www.gaussian.com/index.htm.<]§ But I'd be surprised if they create a CUDA/OpenCL version any time soon, considering I don't think they've embraced Open MPI -but I'm forgeting the details.

        • BobTheBacterium
        • 11 years ago

        Folding is a small subset of computational chemistry

      • MadManOriginal
      • 11 years ago

      The answer is *right there* in the picture you link. It says ‘two high performance graphics cards’ not ‘two high performance Tesla/GPGPU/CUDA/whatever cards.’

        • wibeasley
        • 11 years ago

        You’re right. Dell’s ‘Tech Specs’ tab and the customizable options mention only Quadro cards, not Tesla cards. From that announcement, I was thinking the focus was on the Telsa C1060 card.

          • MadManOriginal
          • 11 years ago

          They’re just kind of generic product feature pictures anyway. I’m pretty sure anyone looking to get one of those would realize that the Tesla option would go in the place of those graphcis cards 😉

    • Trymor
    • 11 years ago

    Arrg.

    • Trymor
    • 11 years ago

    Good for them. It’s interesting to see the shift in computing from the Mhz race, to efficiency to multi-core to GPU assisted computing to the GPU integration race.

    • tesla120
    • 11 years ago

    I love my processors!

      • Jigar
      • 11 years ago

      Funny, you choose the same ID …

        • tesla120
        • 11 years ago

        I have had it years before any one thought about the processor

          • Meadows
          • 11 years ago

          That might require proof, because you appeared in 2008 around here and Tesla devices appeared in 2007.

            • tesla120
            • 11 years ago

            I dont post in the forums, only on news articles. as a result my account was deleted due to in activity and had to be recreated. I have been here since I was introduced to it by Moose17145 in my freshmen year of college at mankato, and don’t think i created the account until fall semester of my sophomore year. (2006) I chose the name because at the time I was interested in electrical engineering, taking the name tesla after an inventor that I admire and 120 because I needed a number and it seemed appropriate considering his work on AC.

            I also was building my desktop at the time which is why I started coming here more often. once it was done i started folding on TR’s team under the name of the desktop i built Balthasar. I folded quite a bit and then stopped. my last WU was 2007-11-06 00:19:34 which was the semester I moved to UND. meaning I had to have been here before that to fold that much for TR when i was running F@H.

            I have also used the same screen name on eggxpert and registered it there on 4/10/2007 long after I came here and used the name.

            now, is there any reason I need to keep proving myself or does it seem more plausible that I created a screen name after who is named rather than a processor?

            • moose17145
            • 11 years ago

            troof.

            • Chrispy_
            • 11 years ago

            :O

            I hope that doesn’t happen to me. I rarely post outside of these comments sections and my account going inactive would reset my registration date.

            What will I do if someone called “Fields” disputes the originality of my name when the Intel ChrispyTech line appears next year? I’d better hit up the patent office: They love important issues such as this, and maybe in a decade I can sue Intel and retire \o/

Pin It on Pinterest

Share This