Robert Colwell, the director of the microsystems group at the Defense Advanced Research Projects Agency, talked about Moore's Law in hi keynote speech at the Hot Chips conference. The engineer who once managed a Pentium-class processor design at Intel predicted that Moore's Law is running on its last legs, he said the ability of packing twice as many transistors on the same piece of silicon every two years will probably come to end as soon as 2020 at the 7nm node, or around 2022 at the 5nm node.
Colwell stressed that the end of Moore's Law will have big implications for the chip industry because it will mean the end of exponential growth in computing power. He added that engineers will make a bunch of incremental tweaks, but doubted that a new discovery would result in another exponential growth curve to replace Moore's Law.
Further details can be read at EE Times.
Moore's Law was a rare exponential growth factor that over 30 years brought speed boosts from 1 MHz to 5 GHz, a 3,500-fold increase. By contrast, the best advances in clever architectures delivered about 50x increases over the same period, he said.
Exponentials always come to an end by the very nature of their unsustainably heady growth. Unfortunately, such rides are rare, Colwell said.
"I don't expect to see another 3,500x increase in electronics -- maybe 50x in the next 30 years," he said. Unfortunately, "I don't think the world's going to give us a lot of extra money for 10 percent [annual] benefit increases," he told an audience of processor designers.