4 ms·

Also TPU v1,v2 and v3 were ASICs, but since v4 they have added some new features so they have a lower performance/watt which is quite near Nvidia's power draw.

by ilove196884 2y ago

Also TPU v1,v2 and v3 were ASICs, but since v4 they have added some new features so they have a lower performance/watt which is quite near Nvidia's power draw. I think Hopper is at 700W and TPU are around 600W.

YetAnotherNick 2y ago
Power draw doesn't matter in cloud. All that matters is performance/price for the task in hand.
- jeffbee 2y ago
  In the cloud, opex == energy.
  - talldayo 2y ago
    Which is why you don't send 4 TPUs to do 1 GPU's job.
    - jeffbee 2y ago
      If the TPUs cost less, that implies that they draw less power. If they cost more then nobody will use them.