In a private message I have been asked for my opinion about http://semiaccurate.com/forums/showthread.php?t=8095&page=4.
Well, I must admit I am following silicon developments from an airliner's cruising altitude only. As long as POET is considerably faster, less power consuming and more integrative than silicon, I don't care too much about the nitty-gritty silicon details.
In my opinion three comparisons are fundamental:
-
How does top-notch silicon compare to top-notch POET? According to POET Technologies, 40 nm PET roughly levels 14 nm silicon performancewise and is ahead of it regarding less power consumption.
-
How do silicon and POET compare pricewise? How much would you have to pay to implement your application a) in silicon and b) in POET? Manufacturing of high-end chips with comparable properties will be cheaper with POET, because POET can do with a larger node size than silicon to achieve comparable characteristics. On the other hand, if you have a simple application where neither performance nor power consumption matter, you will be cheaper with, say, 130 nm silicon. There will always be some market for silicon solutions.
-
Which solutions can you (reasonably) implement only with silicon or only with POET? I believe that a) due to POET's capability to integrate electronics, optics, digital, and analog into the same die and b) due to POET devices that don't have silicon equivalents (e.g. the optical thyristor), we'll see a lot of novel solutions in the coming years that can only be implemented using the POET technology.
