backtop


Print 80 comment(s) - last by nugundam93.. on May 8 at 9:03 AM


The researchers formed memristors by crossing platinum nanowires with another wire, with a thin dab of titanium dioxide at the junction. Each wire is a seperate memristor, yielding 17 memristors, each 50 nm (150 atoms appr.) wide. The memristor remembers even after powered off.  (Source: J. J. Yang, HP Labs.)
New discrete component from HP joins the ranks of the resistor, capacitor and inductor as the fourth major circuit element

Yesterday Hewlett-Packard (HP), best known as a leading personal computer manufacturer, announced what may be one of the most significant electronics breakthroughs of the decade.  Researchers at HP Labs, the central research center for the company, confirmed the existence of the previously theorized fourth fundamental circuit element of electrical engineering.

The new component is called the “memristor” -- a word blend of "memory" and "resistor".  The physical working model and the mathematical model of the component were presented side by side in a paper in the journal Nature, yesterday.  Four researchers at the lab, led by R. Stanley Williams, presented the device which retains the history of information passed to it.

The device could make for computers that need no boot-up, never forget, use less power, and associate memories much like the human mind.  Such possibilities were long considered the realm of science fiction.  The realization of the device was 37 years in the making, and many had come to think it would never be created.

The component was initially theorized and named by Leon Chua, a distinguished faculty member in the Electrical Engineering and Computer Sciences Department of the University of California at Berkeley, in 1971.  Chua's paper argued that the new component was a fundamental fourth element of electronics with unique properties which the other elements did not have.

By leveraging experience in nanoelectronics, Williams was finally able to realize Chua's creation, over three decades later.  Williams was ecstatic about the success.  He stated, "To find something new and yet so fundamental in the mature field of electrical engineering is a big surprise, and one that has significant implications for the future of computer science.  By providing a mathematical model for the physics of a memristor, HP Labs has made it possible for engineers to develop integrated circuit designs that could dramatically improve the performance and energy efficiency of PCs and data centers."

The device could eventually make dynamic random access memory (DRAM) obsolete.  In current systems, active computers store data in DRAM, but must shuffle the information to and from a magnetic hard disk or a flash drive, nonvolatile forms of memory.  Furthermore, when the computer is turned on, the DRAM must be initially loaded from the magnetic memory.  These processes consume both time and energy, slowing computing and raising the energy and heat envelopes of systems.

A memristor would need no boot up as its data would be exactly how it was previously left.  Data could theoretically be read and wrote directly to and from memristors, eliminating the need for hard drives, except possibly for backup storage.

With the advent of “cloud computing” -- the transition of data storage to the online world this device becomes even more valuable and timely.  The IT infrastructure that's growing to support cloud computing uses thousands of systems, multiplying the energy costs of ram usage exponentially.  The new component could dramatically reduce the power, and thus the expense of such systems, as well as helping to protect user data and reducing load times.

One key problem to data centers has always been the possibility of a power loss.  The memristor essentially would take away the problem, as barring complete circuit destruction; the data would survive a power outage.  The type of memory also offers the possibility of continuously learning and adapting systems, similar to the human brain.  Such systems could be used in facial recognition technology, as well as in enabling advanced biometric security and privacy features.

Williams is no stranger to innovation -- he founded HP Labs’ Information and Quantum Systems Lab and has been its director ever since.  The lab strives to develop advances in the realms of mathematics and physical science useful to computing.  The lab has logged many advances in nanoelectronics and nanophotonics, but the memristor may well go down in history as its most significant contribution.



Comments     Threshold


This article is over a month old, voting and posting comments is disabled

RE: So....
By jordanclock on 5/1/2008 4:01:13 PM , Rating: 3
The whole point of having RAM and hard drive is because hard drives are too slow to actually run a program from, but RAM doesn't have the density to store everything you need (and keep it).

So this would, indeed, remove the need for two separate devices. You could use a memristor device that has enough space to provide storage like a hard drive, but it would also be used as the RAM too, most likely as a large swap space.


RE: So....
By Master Kenobi (blog) on 5/1/2008 4:08:51 PM , Rating: 2
Alright I will use your method then. Under your method you still need two devices. They can both be comprised of these new memristor devices, however you still need two. One that can be cleared on demand (Swap space/Ram) and another that stores data. If you were unable to dump swap space/ram in favor of a fresh set of the last working data, your completely screwed in the event of a software glitch that is sitting in RAM.


RE: So....
By darkpaw on 5/1/2008 4:40:13 PM , Rating: 3
Two devices yes, but likely both memory based. The sooner they can eliminate the mechanical parts (HDD) the better.


RE: So....
By Master Kenobi (blog) on 5/1/2008 5:21:02 PM , Rating: 2
I'm in favor of Holographic based disk drives tbh. Memory can be whatever your fancy is.


RE: So....
By casket on 5/1/2008 8:41:06 PM , Rating: 3
Not necessarily two devices. You can partition this memory... Reserve the first 2 gigs for ram, and the rest for storage.


RE: So....
By Natfly on 5/1/2008 4:48:31 PM , Rating: 2
My point was that all I knew was that this can be used to store data.

The possible density, speed, read/write cycles, data retention life I have no clue about. So depending on these factors it could be used as a replacement for any type of storage whether it be cpu cache, dram, magnetic hard disks, or flash memory.

The only thing I've gathered is that it is much simpler than the above storage methods.


RE: So....
By Zoomer on 5/3/2008 9:41:41 AM , Rating: 3
2This is huge.

Just about the simplest memory that can be created now would be the latch. This is how it looks like:
http://en.wikipedia.org/wiki/Image:SR-NOR-latch.pn...

These aren't really commonly used because of the restricted combination. More commonly used D-flip flops are even more complex.
http://thalia.spec.gmu.edu/~pparis/classes/notes_1...

This would allow the elimination of both these gates, as well as the intermediate wires. Elimination of gate delay, delay from intermediate wires, etc will lower power consumption and increase the response time (read: higher max clock) of the component.


RE: So....
By s12033722 on 5/3/2008 7:31:42 PM , Rating: 1
Actually, the simplest memory that can be created now is a capacitor and sense transistor, otherwise known as DRAM. You are describing SRAM which is all but obsolete.

Also, I seriously doubt that this element will be used in anything major anytime soon. This just describes how to make an element. No read/write circuitry exists, it appears to be extremely hard to manufacture, and no details exist about speed, data retention, etc. This is interesting, but it is only a curiosity for the moment.


RE: So....
By Jedi2155 on 5/5/2008 5:38:27 AM , Rating: 2
There was a mention in an article somewhere of the speed being about 1/10th what current memory technologies are at but it doesn't mention which tech in particular.


RE: So....
By 9thZergKiller on 5/6/2008 1:02:52 PM , Rating: 3
quote:
You are describing SRAM which is all but obsolete.

SRAM's a lot faster than DRAM. The requirement for "sense transistor" in the DRAM makes it slow. Thus, the SRAM is not "all but obsolete."

quote:
Also, I seriously doubt that this element will be used in anything major anytime soon. This just describes how to make an element. No read/write circuitry exists, it appears to be extremely hard to manufacture, and no details exist about speed, data retention, etc. This is interesting, but it is only a curiosity for the moment.

Agreed there at least.


"I'm an Internet expert too. It's all right to wire the industrial zone only, but there are many problems if other regions of the North are wired." -- North Korean Supreme Commander Kim Jong-il














botimage
Copyright 2014 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki