Intel Solidifies Deep Learning Portfolio with Varied Architectures and Packages

Discussion in '[H]ard|OCP Front Page News' started by cageymaru, Sep 12, 2018.

  1. cageymaru

    cageymaru [H]ard|News

    Apr 10, 2003
    Gadi Singer is the vice president and general manager of the Artificial Intelligence Products Group at Intel. In an interview with Ed Sperling of Semiconductor Engineering, he discusses how Intel is evolving to meet the ever changing requirements for deep learning. He believes that Xeon processors are well suited for deep learning but there are other solutions needed that range from sub-1 watt to 400 watts. Intel's ability to leverage the data center, edge and system integration will be key to their future in creating a solid portfolio of products for the deep learning field.

    There are three elements. One is that we need a portfolio, because our customers are asking for it. You need solutions that go from the end device, whether that's a security camera or a drone or a car, to a gateway, which is the aggregation point, and up to the cloud or on-premise servers. You need a set of solutions that are very efficient at each of those points. One element of our hardware strategy is to provide a portfolio with complementary architectures and solutions. Another element is to further make Xeon a strong foundation for AI.
  2. BSmith

    BSmith [H]ard|Gawd

    Nov 9, 2017
    They are trying to construct a new leader for the company using the AI they are developing just to try and avoid more mis-steps. Yeah Intel, aint foolin no one!
  3. arnemetis

    arnemetis 2[H]4U

    Aug 2, 2004
    Xeon processors are well suited for landfills until they fix/remove speculative execution and come up with a new architecture.
  4. pcgeekesq

    pcgeekesq [H]ard|Gawd

    Apr 23, 2012
    "AI is hot, so buy our Xeon processors for it. They weren't designed for AI, but they make us a lot of money, so of course they are the perfect solution for your AI needs."

    Once upon a time, Intel had great engineers leading its product development.
    That died, replaced by optics-driven* product development, when Intel decided to go with the Pentium 4 "NetBurst" architecture.
    It's been downhill ever since.

    *optics-driven: "how buzz-worthy the product looks,"
    as opposed to marketing-driven: "how well it satisfies the customer's needs"
  5. Twisted Kidney

    Twisted Kidney 2[H]4U

    Mar 18, 2013
    All I ever see when I read about this shit is "Derp Learning".
    J3RK likes this.
  6. J3RK

    J3RK [H]ardForum Junkie

    Jun 25, 2004
    Let's use our spare cycles to crystalize our roadmap, and leverage that deep learning! Ping me, and we'll set up some meetings to plan a time to discuss the primary action items, and solidify the synergy between deep learning and our silicon semiconductor engineering.
  7. S-F

    S-F Gawd

    Aug 5, 2010

    Conroe? Intel trounced AMD there.

    Just trying to be clear. I'm not trying to be in Intel's corner. I want them put to the thumb screws and torches as much as the next guy. Well, unless the next guy is blindly dedicated to Intel for some reason.
  8. joobjoob

    joobjoob Limp Gawd

    Jun 29, 2004
    Well Intel welcome to AMD shoes where your foundries are holding your engineers back. All while a wealthy competitor blows right past you.