At the Flash Memory Summit in August 2009, updates on several new technologies involving NAND flash were presented to conference attendees. One of them was given by Stan Williams, Hewlett-Packard senior fellow and director of Quantum Science Research, and it involved something called the "memristor," a term condensed from "memory resistor."
On that day, Williams described the memristor this way: "This is sort of the missing element of the processor puzzle. It takes its place alongside the resistor, capacitor and inductor [as the fourth basic circuit element in chip engineering]. And it could change the way we do IT."
In summary, let's just say adding a memristor to a solid-state NAND flash drive can be like putting it on steroids.
Since flash media already owns the fastest I/O speeds known to IT science, increasing that speed tenfold or by a higher magnitude-HP's conservative estimate at this time-is certainly an intriguing proposition for processor engineers and IT systems makers.
On April 8, HP Labs published an update on advances in memristor research. These findings are also detailed in a paper published the same week in the journal "Nature" and written by Williams and five other researchers who work at HP's Information and Quantum Systems Laboratory, headquartered in Palo Alto, Calif.
HP Labs has six other locations around the world, in Bangalore, India; Beijing; Haifa, Israel; Bristol, England; St. Petersburg, Russia; and Fusionopolis, Singapore.
Following two years of research, Williams and his team discovered that the memristor has more capabilities than was previously thought. The team said in its report that "in addition to being useful in storage devices, the memristor can perform logic, enabling computation to one day be performed in chips where data is stored, rather than on a specialized central processing unit."
Bringing the logic closer to the data is key
The idea of distributing logic directly into a dedicated processor, instead of exclusively in a CPU somewhere away from the data, is a revolutionary move. Where data and processors are physically close is always where the best performance is found. Google's home-grown systems have proven this for more than a decade.
"Memristive devices could change the standard paradigm of computing by enabling calculations to be performed in the chips where data is stored," Williams said. "Thus, we anticipate the ability to make more compact and power-efficient computing systems well into the future, even after it is no longer possible to make transistors smaller via the traditional Moore's Law approach."