Page 1 of 2 12 LastLast
Results 1 to 10 of 16

Thread: NVIDIA Announces "Pascal" Next-Gen GPU Family

  1. #1
    Join Date
    Jan 2007
    Posts
    15,130

    Default NVIDIA Announces "Pascal" Next-Gen GPU Family

    Phoronix: NVIDIA Announces "Pascal" Next-Gen GPU Family

    While NVIDIA's recently-announced Maxwell graphics architecture is doing terrific under Linux, Pascal was announced today as their next-generation GPU come 2016...

    http://www.phoronix.com/vr.php?view=MTY0Mjk

  2. #2
    Join Date
    Jun 2011
    Posts
    1,036

    Default

    Unless the situation has really changed I don't really see the point of NVLink, because when PCIe3.0 came out Tom's hardware did some test and found that there was minimal benefit to it over PCIe2.0, at that point those lanes weren't really the bottleneck.

  3. #3
    Join Date
    Sep 2008
    Location
    Vilnius, Lithuania
    Posts
    2,601

    Default

    Will the next GPU line be called "Delphi"? Or perhaps "COBOL"?

  4. #4
    Join Date
    Apr 2007
    Location
    Arctic circle, Finland
    Posts
    298

    Default

    Quote Originally Posted by Luke_Wolf View Post
    Unless the situation has really changed I don't really see the point of NVLink, because when PCIe3.0 came out Tom's hardware did some test and found that there was minimal benefit to it over PCIe2.0, at that point those lanes weren't really the bottleneck.
    It's for supercomputing not related for gaming, oh and think about hsa or something similar...:
    NVIDIA today announced that it plans to integrate a high-speed interconnect, called NVIDIAŽ NVLink™, into its future GPUs, enabling GPUs and CPUs to share data five to 12 times faster than they can today. This will eliminate a longstanding bottleneck and help pave the way for a new generation of exascale supercomputers that are 50-100 times faster than today's most powerful systems.
    So Volta was renamed to Pascal, hrh will be quite hilarious to conjugate it in finnish.

  5. #5
    Join Date
    Oct 2011
    Posts
    40

    Default

    I've watched press conference and didn't get a thing about price for Titan Z. It seems Nvidia has its own math.
    Titan x 1 = $1000
    Titan x 2 = $2000
    But Nvidia says that slapping those to chips on one board would cost $3000.

    Also realtime car demonstration using renderfarm was something out of this world. He said its cheap and then says 17 machines each $50 000 is ok to run one car almost in realtime. And the car didn't even look that good to begin with. Even though i liked that they modeled it completely, detail by detail, crazy modelers.

    I also don't really understand a thing about PCIe being the bottleneck when it was never the case, neither for OpenGL nor for OpenCL/Cuda performance. Infact AMD said goodbye to crossfire bridges because of the same reason of PCIe being enough for all needs. And i though Nvidia will do the same but somehow they don't.

    I liked the part when he said that all GPUs will give linear performance once you connect them with NVlim(renamed SLI i presume). I really hope he was talking about graphics because for compute tasks it always was the case.

  6. #6
    Join Date
    Apr 2007
    Location
    Arctic circle, Finland
    Posts
    298

    Default

    Hmm tegra k1 development board, called jetson TK1 @192$(not usually devboard price like 500$ or something, a bit pricey still):
    http://nvidianews.nvidia.com/Release...stems-ad8.aspx

  7. #7
    Join Date
    Mar 2012
    Posts
    121

    Default

    Quote Originally Posted by tuke81 View Post
    Hmm tegra k1 development board, called jetson TK1 @192$(not usually devboard price like 500$ or something, a bit pricey still):
    http://nvidianews.nvidia.com/Release...stems-ad8.aspx
    nvlink shoudl die, it's a dangerous trend to have.

  8. #8
    Join Date
    Sep 2010
    Posts
    474

    Default

    Stacked DRAM be good.
    Like that it's faster. It's an important improvement.

    Don't like the NVlink.
    Last edited by plonoma; 03-25-2014 at 04:19 PM.

  9. #9
    Join Date
    Sep 2012
    Posts
    287

    Default

    Oh, bloody hell.
    For a second there I thought Nvidia were creating a new shading language
    based on Pascal.

  10. #10
    Join Date
    Apr 2007
    Location
    Arctic circle, Finland
    Posts
    298

    Default

    Quote Originally Posted by plonoma View Post
    Stacked DRAM be good.
    Like that it's faster.

    Don't like the NVlink.
    I don't believe we will ever see nvlink in consumer space. It's for hpc, and especially power8 processors for better bandwidth for unified memory between gpus and cpu(heck it's developed with ibm for ibm supercomputers).

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •