Sign in with
Sign up | Sign in

Nvidia to Integrate ARM Processors in Tesla

By - Source: InfoWorld | B 25 comments

Nvidia has confirmed that it will be combining GPGPUs with ARM 64-bit processors in future Tesla products.

In an article posted by InfoWorld, the company's chief technology officer for the Tesla product line, Steve Scott, was quoted saying, "Tegra is going to become GPU computing capable in the not-so-distant future. Sometime this decade we are also going to start bringing integrated CPUs and GPUs together in the Tesla line".

'Sometime this decade' is not exactly clear, but we would take a guess that Nvidia is shooting for a release prior to 2015.

Scott was referring to ARMv8 processors, recently announced as Cortex-A53 and A-57 models and a 2014 released date. However, Nvidia was not mentioned by ARM in a row of current ARMv8 licensees, which include AMD, Broadcom, Calxeda, HiSilicon, Samsung and STMicroelectronics. However, the emerging microserver market is an opportunity for a natural evolution of Nvidia's business. Combining Tesla with ARMv8 cores will allow the company to compete in a segment that will be crowded with industry heavyweights such as Samsung, Qualcomm, Texas Instruments - and Intel on the x86 side. Nvidia will also compete with its arch rival AMD in a new market.

The InfoWorld article pointed to Nvidia's Project Denver, which will be based on using the ARMv8 architecture in Nvidia products. According to ARM, the Cortex-A57 processor will be able to provide up to three times the performance of its current flagship, the 32-bit Cortex-A15 design.

Contact Us for News Tips, Corrections and Feedback

Ask a Category Expert

Create a new thread in the News comments forum about this subject

Example: Notebook, Android, SSD hard drive

This thread is closed for comments
  • 4 Hide
    dragonsqrrl , November 1, 2012 9:36 PM
    Maxwell should see the first integration of ARM cores onto the GPU for Nvidia. It's good to see that's still on track. It's also good to see another hint/confirmation that the next Tegra SOC will have some sort of a unified compute capable GPU, finally moving beyond Nvidia's original mobile GPU architecture.
  • 0 Hide
    A Bad Day , November 1, 2012 9:56 PM
    Well, looks like Nividia has some catchup to do, though I wonder if they would fare better than Intel's history of not-the-greatest GPUs (before the HD4000s).
  • 1 Hide
    sonofliberty08 , November 1, 2012 10:18 PM
    i hope to see the day ARM replacing x86 in the very near future
  • Display all 25 comments.
  • 1 Hide
    Shin-san , November 1, 2012 10:25 PM
    That's pretty clever. This makes it more like a self-contained unit and let it do better traffic copping on the board instead on the main CPU
  • 9 Hide
    bloc97 , November 1, 2012 10:30 PM
    sonofliberty08i hope to see the day ARM replacing x86 in the very near future

    That would take quite a while... x86-64 is still the most popular platform and I think it will stay this way for a long time.
  • 5 Hide
    A Bad Day , November 1, 2012 10:42 PM
    sonofliberty08i hope to see the day ARM replacing x86 in the very near future

    Standards die hard. There's still a significant amount of software written for x86 compared to ARM, and business software tend to be the vast majority.
  • -1 Hide
    DjEaZy , November 1, 2012 10:44 PM
    ... AMD and nVidia implementing ARM now... intel is not... hmm.... AMD and nVidia haz GPU tech, intel haz not... where is this going and will the x86 be just relevant for legacy stuff?
  • 6 Hide
    eddieroolz , November 1, 2012 11:10 PM
    A computer contained within the GPU can give way to a huge card farm that can compute with more efficiency.
  • 3 Hide
    ekho , November 1, 2012 11:23 PM
    AMD cannot take a break.......
  • 5 Hide
    fb39ca4 , November 2, 2012 12:04 AM
    Tesla is for parallel computing, so will they have hundreds of ARM cores on each chip? That would make the system a lot more flexible. Currently GPUs are good for brute force number crunching, but are slow when it comes to things like if-then statements because a number of cores have to be working on the same instruction at the same time. With ARM cores, you should see much less of a performance hit when branching especially with stuff like branch prediction.
  • 1 Hide
    A Bad Day , November 2, 2012 1:33 AM
    The only downside is, many developers are still only going to use around four cores. Multicore software is usually harder to design than a singlecore software (deadlocks, etc).
  • 0 Hide
    silliconcrayon , November 2, 2012 4:29 AM
    "TESLA "
  • -1 Hide
    saturnus , November 2, 2012 5:47 AM
    bloc97That would take quite a while... x86-64 is still the most popular platform and I think it will stay this way for a long time.

    Not by a long shot. x86-64 is the most popular platform for laptops, desktops and servers. But by far the most popular computing platform is ARM.

    There's 25 billion ARM processors produced to date, and over 16 million is shipped every single day.
  • 2 Hide
    kronos_cornelius , November 2, 2012 6:15 AM
    That signals my super-computing Raspberry-Pi project is approaching :-)
  • 1 Hide
    ojas , November 2, 2012 7:08 AM
    Interestingly, Apple doesn't seem to be a ARMv8 licensee. Looks like the A6 architecture will be around for a long time.
    Wonder how it'll compare to Qualcomm and Samsung's ARMv8 based chips when they're released...

    sonofliberty08i hope to see the day ARM replacing x86 in the very near future

    Not me. I feel x86 will remain more powerful and eventually (as Intel tick-tock goes on) more efficient than ARM stuff.

    Maybe ARM may occupy the lower end one day, maybe they get this stuff into consoles. PCs will remain x86 for a long, long while. Probably as long as they can keep the die shrinks happening.
  • 3 Hide
    martel80 , November 2, 2012 7:24 AM
    ARM may soon find itself in a position similar to AMD - they will be able to put a huge number of cores on a chip but they will struggle with single-threaded performance. Vast majority of tasks people do on a PC today requires no more than 2 cores.
  • -1 Hide
    Anonymous , November 2, 2012 7:59 AM

    It's an interesting idea, basically being able to have the interest of the gpu to be able to do what any computer system does or say do without is probably an interest within to have for a gpu solution in terms of computing.

    Especially on interest to say within place of is that of what is probably an interest of computing rather with a gpu or cpu, why not have both in one?

    Be like saying otherwise being able to run any application from just the say gpu instance for display only has point or purpose for such when on without has no place more for.

    Arm processing essentially would almost cover the topic for given arm processing can consist of about anything to say of for.

    Why have a processor on the interest of general computing computing any specific interest for its place of processing when an arm processor for a place of such rather within additional device holding or not can be placed for a use of such, without reliance of support, expansion, and/or extension systems??

    Let alone with and/or on a system (seperate) that has a place of use to say for one.

    It's a good one, but one that might not find alot of place for use to as such a system as to say.
  • 1 Hide
    technoholic , November 2, 2012 8:13 AM
    AMD has APUs and now Nvidia. It seems to me that in a couple of years we will see a completely different type of market competition in a different field. So x86 (or x86-64 in this case) is no longer going to be the only market. So when we look at x86 realms, we see Intel as the giant but what about ARM world? Intel doesn't either have a competitive "APU" ( or should i say "SoC" to be more precise? )solution nor tools/licenses/wireframe to fight in the ARM world.
  • 0 Hide
    sonofliberty08 , November 2, 2012 9:12 AM

    there will be more competition on ARM world
  • 0 Hide
    captainblacko , November 2, 2012 9:49 AM
    but will it run minecraft?
Display more comments