Why AI inference will remain largely on the CPU

It’s a complex argument, but there are good reasons why inference shouldn’t head into accelerators or GPUs

Sponsored Feature  Training an AI model takes an enormous amount of compute capacity coupled with high bandwidth memory. Because the model training can be parallelized, with data chopped up into relatively small pieces and chewed on by high numbers of fairly modest floating point math units, a GPU was arguably the natural device on which the AI revolution could start.…

ai model capacity compute cpu data feature floating point good gpu gpus head inference math memory natural numbers small sponsored training