Re:TSMC and Graphcore Prepare for AI Acceleration on 3nm
by lotyrin on Thursday August 27, 2020 @02:42PM
(#24297082)
I guess if we hit the edge of feature density, we might start building things out of 3d FPGA type designs of homogeneous modules but because of heat, the capacity of such a design is ultimately going to be limited by its surface area, not its volume, no?
[ Parent
| on HN ]
Re:TSMC and Graphcore Prepare for AI Acceleration on 3nm
by craftinator on Thursday August 27, 2020 @10:00PM
(#24300679)
My CPU is a sphere!