The business pages of a few of the big dailies are covering a Nature paper about new developments in memristors—components whose resistance depends on a “memory” of previous currents.

Memristors are one of four fundamental circuit elements (the other three, for curious readers, are resistors, capacitors and inductors). Electrical engineers first theorized the existence of memristors in the 1970s, but it took nearly thirty years for Stanley Williams and his team at Hewlett-Packard Labs in Palo Alto, California to actually fabricate one. Williams used an ultra-thin film of semiconducting titanium dioxide sandwiched between two platinum leads to make his memristor (published in Nature in 2008).

Now the HP team has struck once again. In a paper in this week’s issue they describe using their memristor to perform several ‘stateful’ logic operations. In a nutshell, stateful logic means that the ‘state’ of the memristor acts as both the computer and the memory. That’s a pretty big change from current computers, which typically load data from memory, perform operations on it, and then send it back.

It should be no surprise then that coding a memristor is quite a bit different to coding a transistor. To perform their operations, the memristors had to be fed a series of voltages in sequence. Those voltages set and cleared the memristive circuits as appropriate.

It would take “substantial parallel operations” of many memristors to make the technology competitive with existing silicon, according to the paper. But that hasn’t kept Williams and HP from making some bold claims in the press: He told the New York Times he expected memristors could overtake flash memory within three years, and would beat out competing “”https://www.nature.com/news/2009/090925/full/news.2009.951.html">phase change memory", which is being produced by Samsung and others.

Credit: S. Williams/Nature