pearlmut (01/27/83)
The 100 million speedup estimate for Potato Chips is based on the idea of using protiens as the individual switching elements, rather than transistor gates. A molecule can twitch very quickly, and is very small, leading to extremely high signal propogation speeds, and incredibly high densities. Consider that modern VLSI has feature sizes on the order of 1 um, about the diameter of a yeast cell. Think of the processing power in a yeast cell, and think of getting all that power in a usable form, and you see why this would be so neat. The idea is not to use entire cells as swithing elements or gates. They are much too slow, and quite unreliable. By the way, shouldn't this discussion be on net.vlsi? Barak Pearlmutter,
wapd (01/27/83)
Maybe transitions at the molecule level could be fast, but does that necessarily translate into overall speed of a large processor ? The situation is probably analogous to that of VLSI today : things are fine if you can fit them all onto one chip, but any time you have to drive a pin it falls apart. Everyone dreams of putting so much stuff on a chip that you don't ever have to go across pin boundaries (called wafer-scale devices), but that technology is still several years off. The same will happen with bio-chips : for the first ten years they may be slower than VLSI, and then eventually they will pass it. Bill Dietrich houxj!wapd