Stateside chip boffins say they have developed a radical new method of building memory, which will smash through the "brick wall" that Moore's Law is about to run into. The underlying technique was thought to be dependent on the use of graphite, but in fact this has now been shown to be untrue. A plucky grad student studying at …
Whats the catch?
What's the catch? This is a possible memory unit that ticks pretty much every conceivable box. It seems as though it can be made relatively easily using currently existing techniques, from a low cost and abundant material, has the capacity to made in 3D for high density storage, and the ability to make gates considerably smaller than they are now? This is more or less the holy grail of memory. Surely this is way too easy whats the catch?
Paris, 'cos she knows the catch...
Switching signals involves converting electrical energy into heat. Compact switch layouts have less area to dissipate the same amounts of heat. 3D structures have even more problems getting the heat away from the switches. Two ways around this are to use less energy per switch transition (which raises other problems in itself such as thermal and electrical noise swamping the signal) and the other is to use materials and switch technologies that don't degrade when hot (which causes knockon problems with packaging and assembly technologies -- a ceramic-encapsulated chip that can run OK at 400 deg C will desolder itself from a conventionally-built planar).
But surely ...
... a smaller switch requires proportionately less energy.
Nobody believed him.
Good job that guy. Pint for you, mate.