NVIDIA Unveils Turing GPU for the RTX Quadro Lineup

Discussion in '[H]ard|OCP Front Page News' started by Kyle_Bennett, Aug 14, 2018.

  1. Kyle_Bennett

    Kyle_Bennett El Chingón Staff Member

    Messages:
    53,596
    Joined:
    May 18, 1997
    Jensen kicked off the RTX Quadro cards by showing that now-famous Stormtrooper ray-tracing demo running in real time on a single Turing GPU with its Tensor Cores. The Tensor Cores provide up to 500 trillion operations per second that fill in where the real time ray-tracing comes up short. Deep learning on your ray-tracing to put it simply. Turing is 6X faster at rasterization than Pascal. The new RTX Quadro cards price out at $2,300 for the Quadro RTX 5000, $6,300 for the Quadro RTX 6000, and $10,000 Quadro for the RTX 8000.

    Stormtrooper video.

    Ray-traced graphics offer incredible realism. Interactive graphics driven by GPUs offer speed and responsiveness. The two now come together in the greatest leap since the invention of the CUDA GPU in 2006, NVIDIA CEO Jensen Huang announced Monday.

    Speaking at the SIGGRAPH professional graphics conference in Vancouver, Huang unveiled Turing, NVIDIA’s eighth-generation GPU architecture, bringing ray tracing to real-time graphics. He also introduced the first Turing-based GPUs — the NVIDIA Quadro RTX 8000, Quadro RTX 6000 and Quadro RTX 5000. And he detailed the Quadro RTX Server, a reference architecture for the $250 billion visual effects industry.
     
  2. illli

    illli [H]ard|Gawd

    Messages:
    1,099
    Joined:
    Oct 26, 2005
  3. Kyle_Bennett

    Kyle_Bennett El Chingón Staff Member

    Messages:
    53,596
    Joined:
    May 18, 1997
    Quadro RTX Professional GPU highlights:

    • New RT Cores to enable real-time ray tracing of objects and environments with physically accurate shadows, reflections, refractions and global illumination.
    • Turing Tensor Cores to accelerate deep neural network training and inference, which are critical to powering AI-enhanced rendering, products and services.
    • New Turing Streaming Multiprocessor architecture, featuring up to 4,608 CUDA cores, delivers up to 16 trillion floating point operations in parallel with 16 trillion integer operations per second to accelerate complex simulation of real-world physics.
    • Advanced programmable shading technologies to improve the performance of complex visual effects and graphics-intensive experiences.
    • First implementation of ultra-fast Samsung 16Gb GDDR6 memory to support more complex designs, massive architectural datasets, 8K movie content and more.
    • Nvidia NVLink to combine two GPUs with a high-speed link to scale memory capacity up to 96GB and drive higher performance with up to 100GB/s of data transfer.
    • Hardware support for USB Type-C and VirtualLink, a new open industry standard being developed to meet the power, display and bandwidth demands of next-generation VR headsets through a single USB-C connector.
    • New and enhanced technologies to improve performance of VR applications, including Variable Rate Shading, Multi-View Rendering and VRWorks Audio.


    GPU Memory Memory with NVLink Ray Tracing CUDA Cores Tensor Cores


    Quadro RTX 8000 48GB 96GB 10 GigaRays/sec 4,608 576

    Quadro RTX 6000 24GB 48GB 10 GigaRays/sec 4,608 576

    Quadro RTX 5000 16GB 32GB 6 GigaRays/sec 3,072 384
     
  4. DukenukemX

    DukenukemX 2[H]4U

    Messages:
    3,846
    Joined:
    Jan 30, 2005
    Just wanna emphasize that what Nvidia is doing with Ray-Tracing is not impressive. You can download a Hybrid Ray-Tracing Demo right now, that was made in 2012 that uses DX11. So why does Nvidia requires a RTX graphics card with most likely their API?

    http://rigidgems.sakura.ne.jp/



    They updated the demo recently.

     
    fs123, Darth Kyrie, Revdarian and 3 others like this.
  5. PaulP

    PaulP Gawd

    Messages:
    624
    Joined:
    Oct 31, 2016
    Starting at only $2300, I'm sure they will be selling like hot cakes!
     
  6. gulguran

    gulguran Limp Gawd

    Messages:
    255
    Joined:
    Jul 30, 2015
    Was it the 8000 that did this demo? or a lower end card? Thanks :)
     
  7. B00nie

    B00nie [H]ardness Supreme

    Messages:
    7,444
    Joined:
    Nov 1, 2012
    Turin or turding?
     
  8. spintroniX

    spintroniX Gawd

    Messages:
    901
    Joined:
    Apr 7, 2009
    I wonder how the $2300 quadro compares to the $3000 Titan V for deep learning?
     
  9. cyclone3d

    cyclone3d [H]ardForum Junkie

    Messages:
    12,415
    Joined:
    Aug 16, 2004
    Raytracing has never looked very realistic to me. Everything is too perfect.

    It has it's place, but real-time raytraced games are pretty much NOT going to be realistic looking.
     
  10. J3RK

    J3RK [H]ardForum Junkie

    Messages:
    8,562
    Joined:
    Jun 25, 2004
    That's why you then have to apply a pass of "Dirt-Tracing (R)" That dirt layer is what makes things realistic.
     
    ol1bit likes this.
  11. BloodyIron

    BloodyIron 2[H]4U

    Messages:
    2,726
    Joined:
    Jul 11, 2005
    One vendor lock in demon for another. How about Vulkan instead?

     
    Revdarian and ol1bit like this.
  12. Zarathustra[H]

    Zarathustra[H] Official Forum Curmudgeon

    Messages:
    25,290
    Joined:
    Oct 29, 2000
    And let me guess, it uses a proprietary Nvidia standard that's going to fragment the market further and lock in consumers to their brand.

    I know, this is Quadro, so not consumer....

    ...yet
     
    Darth Kyrie likes this.
  13. J3RK

    J3RK [H]ardForum Junkie

    Messages:
    8,562
    Joined:
    Jun 25, 2004
    Doesn't matter if they do. It's not like the industry isn't fragmented as all hell anyway. At least there aren't 50 different brands of GPU like there were back in the 90s and early 2000s. Obviously developers and artists see value in NV's way of doing things or they wouldn't be able to get away with doing things the way they do. Also, until they get some real competition from someone, they really are the only game in town. I'm not saying that's good for the industry, but they'll play ball if someone gives them a reason to. So far... Nobody is... It's not their fault that they act like a big business because they're able. Someone has to rise up and push them to compete and be more open. Literally nobody is doing it, or is even up to the task. AMD is the closest, but they're too busy with CPUs right now.

    Also, like you say, this is Quadro, so it's content development. Once out of that stage, things can change. Sure they'll make it much easier to integrate with their own consumer parts, APIs, etc. Why wouldn't they? But, they still want people to buy their stuff, and buy the software that runs on it, so they have to at least play "kinda" nicely. (largely fake or not)
     
  14. Trimlock

    Trimlock [H]ardForum Junkie

    Messages:
    14,964
    Joined:
    Sep 23, 2005
    So, one Turing out did 4 Volta’s...

    That’s incredible.
     
  15. DNMock

    DNMock Limp Gawd

    Messages:
    190
    Joined:
    Apr 16, 2015
    Why are these multi-thousand dollar commercial GPU's using GDDR6 instead of HBM 2/3?