Does anyone at HN know where the memristor is going? From my perspective this seems like exactly what is needed to maintain d/dt[computing power/$] during the next decade, but given how bleeding edge this is, it is difficult to find good information about just what is going on.
Are there any good articles or papers about how to make gates from memristors, for instance? It would be interesting to see what all the fuss is about.
It is really interesting that they're applying this to FPGA's. Reconfigurable computing would be great, if it was a) easy to program for and b) cheaper. This could solve the cheaper part, and make it even more powerful at the same time. However, I don't know how hard it is to program reconfigurable computers.
FYI: Reconfigurable computing is like parallel computing, but instead of running all of the threads on commodity hardware, they are each dynamically implemented on a FPGA. This makes a super computer with thousands of simultaneous processes much cheaper, and take far less space, if you can program the system.
Wikipedia says that reconfigurable computing systems commonly increase speed 4 orders of magnitude and decrease power consumption by one order of magnitude. (that's right, order of magnitude) However, this is in spite of the fact that FPGAs are behind commodity hardware by almost the same amount. If memristors make a decent dent in the hardware disparities, then reconfigurable computing could be huge.
Are there any good articles or papers about how to make gates from memristors, for instance? It would be interesting to see what all the fuss is about.