Re:TSMC and Graphcore Prepare for AI Acceleration on 3nm
by robin_reala on Thursday August 27, 2020 @01:09PM
(#24295852)
It’s worth pointing out that these node names are only mildly related to feature size at this point.
[ Parent
| on HN ]
Re:TSMC and Graphcore Prepare for AI Acceleration on 3nm
by api on Thursday August 27, 2020 @01:12PM
(#24295883)
They are more like minimum possible feature size, not average feature size. The real metrics you want to look at are things like transistor density, power density, power use, etc.
Re:TSMC and Graphcore Prepare for AI Acceleration on 3nm
by microcolonel on Thursday August 27, 2020 @01:18PM
(#24295990)
And even those things need to be assessed in the context of a specific design.