Nvidia AGX Xavier puts the brains in autonomous robots

A lot of talk around here revolves around powerful parts suitable for your latest personal computer, phone, workstation, or server. That list doesn't cover all the use cases for a juicy, tasty chunk of processing power. Enter the latest member in the company's Jetson family of "system-on-modules," the Nvidia Jetson AGX Xavier.

Nvidia says the AGX Xavier board has 10-W, 15-W and 30-W configurable operating modes, and is fit for robots, computer vision, medical instruments, and autonomous machines, among other applications. The company says Xavier offers the performance of a workstation in a board measuring 3.9" by 3.4" (10 x 8.7 cm). That claim is certainly bold, but then again, the system's specs are also generously sized.

According to Nvidia, Xavier's neural-network output is 32 TOPS. That impressive figure comes by way of Xavier's centerpiece, a Volta GPU with 512 shader ALUs and 64 tensor cores. The cerebellum in the chip is an eight-core Nvidia Carmel ARM v8.2 processor with 8 MB of L2 cache and 4MB of L3. The CPU is connected to 16 GB of LPDDR4x on a 256-bit bus.

The 9-billion-transistor Xavier includes two Deep Learning Accelerator engines and a vision processor. The module's video-handling capabilities can deal with up to four streams of HEVC 4K video at 60 FPS. Additionally, the SoC offers 16 lanes of PCIe 4.0 connectivity along with a bevy of miscellaneous I/O including Ethernet, multiple display outputs, USB, and an alphabet soup's worth of peripheral connectivity acronyms.

It's not difficult to think up more than a few use cases for this kind of tech, and Nvidia's PR notes an interesting one. Oxford Nanopore, a startup focusing on DNA sequencing, apparently built a hand-held computer based on the AGX Xavier. The startup named the device the MinIT, and claims that it can run DNA sequencing up to 10 times faster than a "standard laptop."

Interested developers can order Jetson AGX Xavier boards for the amount for $1099 a piece, so long as they're buying at least 1000. The price may be steep, but we figure the performance available on tap likely justifies it.

Comments closed
    • ronch
    • 11 months ago

    Just in time. I was just putting the finishing touches on my robot and was wondering what chip to plug in it. It’s called C3PO.

      • Krogoth
      • 11 months ago

      “We are doomed” chip

    • Bobs_Your_Uncle
    • 11 months ago

    Eh … honestly Xavier isn’t anything even [u<][i<]close[/i<][/u<] to being something "Special!" I'm certainly not dropping any cash on anything less than at [u<][i<]LEAST[/i<][/u<] a Spacely Sprocket or a Cogswell Cog! Even then I'm likely holding-out for an Elroy or a Judy. Heck, even an Astro barely merits a page click.

    • Anonymous Coward
    • 11 months ago

    So is nVidia making their own CPU cores for this, or is it based on a licensed core? It would be impressive if they were designing their own cores, given that they apparently don’t get much use in high-volume products. Maybe in cars… ?

      • renz496
      • 11 months ago

      if i’m not mistaken Carmel is pretty much Denver successor. if they were using stock ARM core then we should know which ARM core being use by now.

        • Anonymous Coward
        • 11 months ago

        I wonder if their core is too hungry for phone usage or something.

          • tipoo
          • 11 months ago

          Denver was just too…Weird. If they sured up the performance unpredictability I’d be interested in it being tested in a phone or tablet again, but Denver 1’s problem was the binary translation, it would do great in a straight line at the time but any unpredictable paths would be met poorly.

    • albundy
    • 11 months ago

    still too big to fit in Arnold’s head.

    • Neutronbeam
    • 11 months ago

    So…pocket Skynet? Our secretive, hidden, robot overlords welcome this development.

      • freebird
      • 11 months ago

      Did they teach it Isaac Asimov’s “Three Laws of Robotics”?

      Apparently, they didn’t read about those in China…
      [url<]https://www.auburn.edu/~vestmon/robotics.html[/url<] That information and the links must be getting blocked by the "Great Firewall of China". [url<]https://i.dailymail.co.uk/1s/2018/12/13/18/7393382-6493343-image-m-2_1544726766585.jpg[/url<]

        • Chrispy_
        • 11 months ago

        Sort of. Nvidia taught it [i<]their[/i<] three laws: [list<][*<]A Xavier may not ignore a customer or, through inaction, allow a customer to buy a competitor product. [/*<][*<]A Xavier must obey the Proprietary Nvidia code and avoid using open standards wherever possible. [/*<][*<]A Xavier must protect its own existence, either through sabotage, market segmentation, or developer bribery.[/*<][/list<]

    • djayjp
    • 11 months ago

    I think you mean cerebrum, not cerebellum 😉

      • Voldenuit
      • 11 months ago

      I think you mean Cerebro, bro.

      • Mr Bill
      • 11 months ago

      Cerebellum, because the ARM is just linking the Xaviar Cerebro with other tasks.

        • morphine
        • 11 months ago

        Ding ding, we have a winner!

    • Symmetry
    • 11 months ago

    I’m eagerly awaiting someone doing some in depth benchmarks on Carmel.

      • chuckula
      • 11 months ago

      Anybody trying to benchmark Carmel will get caught in an infinite traffic circle loop.
      Especially in autonomous vehicle tests.

      #KampmanGotThatOne

        • Wirko
        • 11 months ago

        “Infinite loop”, does it mean circling around Apple campus, within the event horizon of the RDF?

          • chuckula
          • 11 months ago

          As much as I love to bash Apple, this is an inside joke that Kampman will get due to his geographic location.

            • Wirko
            • 11 months ago

            I get it now. Poor Carmel, her daily commute to 1 Infinite Loop at an infinitesimal speed certainly is an awful experience, and her self-driving car isn’t programmed to break traffic rules when it should.

    • chuckula
    • 11 months ago

    Sorry Nvidia. But Xavier is no match for AMD’s project Jugganaut.

      • Growler
      • 11 months ago

      [url=https://i.imgur.com/MGAlrYm.jpg<]Charles, get out of my head![/url<]

Pin It on Pinterest

Share This