Cloud TPU Pods: AI Supercomputing for Large Machine Learning Problems (Google I/O'19)

Поделиться
HTML-код
  • Опубликовано: 25 ноя 2024

Комментарии • 13

  • @bobsalita3417
    @bobsalita3417 5 лет назад +3

    Would be nice to see TPU2/TPU3/GPU comparison chart supplemented with transistor counts, wattage.

    • @martingorner6233
      @martingorner6233 5 лет назад +1

      Those numbers are not public but the only reason we are doing TPUs is that they offer a huge datacenter density and power advantage (in one word: cost) over GPUs, for ML workloads.

  • @bobsalita3417
    @bobsalita3417 5 лет назад +1

    Thanks. Great information. Tensorflow 2.0 isn't mentioned. Would like to know 2.0 implementation challenges, timelines, anticipated performance differences.

    • @martingorner6233
      @martingorner6233 5 лет назад

      Keras/TPU support is planned in TF 2.1 and the TF 2.1 RC0 has just been published!

  • @ShaneTheMathGuy
    @ShaneTheMathGuy 3 года назад

    Not sure why TPU does not support training multiple models on the same instance. This is not quite efficient for small models...

  • @bobsalita3417
    @bobsalita3417 5 лет назад

    Compute cost is an important cloud metric but what about BYO system (GPU) costs? A slide on BYO (GPU) costs vs GPU/TPU cloud would aid decision making.

    • @MartinGorner
      @MartinGorner 5 лет назад

      The model I trained on stage on a 128 TPUv3 cores cost a little under $50. A TPUv3-128 is roughly the equivalent of 150 powerful GPUs on this workload. You can look up the market price of GPUs + interconnect. I hope this helps you make the buy on site / rent from cloud decision.

  • @visionscaper
    @visionscaper 4 года назад

    Usually 5 V100s will be in one node, I don’t believe that one TPUv2 is as fast as 5 V100s, not by a long shot.

  • @РодионЛукашов-п5у
    @РодионЛукашов-п5у 2 года назад +1

    8:20 watch

  • @bhuvaneshs.k638
    @bhuvaneshs.k638 5 лет назад +1

    Nvidia Tensor Core vs Google's TPU

  • @xavierpan515
    @xavierpan515 5 лет назад

    We need TPUv4!