Move Over Flash

[DragonMaci]

This is an interesting new development.

The newly discovered circuit element - called a memristor - could enable cell phones that can go weeks or longer without a charge, PCs that start up instantly, and laptops that retain your session information long after the battery dies.

Poorly thought out name, but it'd be nice to have a technology that allows such devices.

It also could challenge flash memory, which is now widely used in portable electronics because of its ability to retain information even when power is off.

Chips incorporating the HP discovery would be faster, suck up less power and take up far less space than today's flash.

This is quite impresive given that flash already uses little power than alternatives (e.g. platter based hard drives) and are faster than the alternatives.

The researchers discovered that the amount of resistance it exerts depends on how much electric charge had previously passed through it.

That characteristic gives the memristor an innate ability to remember the amount of charge that has flowed through it long after the power to it is turned off.

That means the circuit itself can be built with a memory function baked in.

Otherwise, data have to be stored in power-hungry transistors configured for storage.

That also takes up valuable real estate on microprocessors or requires separate memory chips.

This is also cool, as it would either result in smaller processors, which would need a smaller cooling system and less material used to make it, thus saving costs and resulting in a cheaper processor, or would leave more room to be used for building other features into processors, thus giving us more powerful processors.

All, in all a very cool discovery. I hope it succeeds.

EDIT 1: This article lists some more cool possibilities, such as replacing RAM and "make computers more intelligent by tracking data it has retained". It also states that it takes less time to turn on that DRAM and flash. And it can store more data than flash memory due to having denser cells. The memristor also has properties similar to a synapse, which would allow for much more powerful computing than what we have today.

Okay, it is about five years until it will be commercially available, but that isn't a long time to wait. I am looking forward to seeing what these memristors are capable of.

Edit 2: Does anyone have more information about this? I'd love to read more about them.

Edited May 1, 2008 by DragonMaci

[Benpercent]

Sounds great, but I'm skeptical as to whether or not we'll actually see in within five years due to current economic regulations (especially looking at monopoly laws). I remember a few years back I heard about a computer company coming out with a chip that sped up processes on a computer so much that it almost does everything instantly. Not only that, but it could also be applied to the internet, allowing data transfer at the same almost-instantaneous speed! Greater than a T3 line even.

This would mean great things for the company right? They could sell these chips, to computer and internet-providing companies alike, and make massive profits, right? Well, it's been a few years since I've heard anything about it and I notice no speak of quality improvements in the market, save for Alienware desktops (horribly expensive). With such a great invention under their belt, I'd imagine that the government saw what an advantage this company would get and warned them about their "anti-competitiveness".

Even more years back, when I was reading in a funeral home, I came across an article speaking of micro needles, which was a needle so small one literally can't feel it going in. This health care invention was supposed to help those who are fearful of needles and those who do not like the pain (such as the IV in the hand). I was in middle school then and this is about seven years later. Still no sign of them, and I'd imagine the production process would only be complex, not expensive (microscopic [used as an adjective] needles only need microscopic materials).

So this news, to me, is the carrot at the end of the treadmill.

[DragonMaci]

Sounds great, but I'm skeptical as to whether or not we'll actually see in within five years due to current economic regulations (especially looking at monopoly laws). I remember a few years back I heard about a computer company coming out with a chip that sped up processes on a computer so much that it almost does everything instantly. Not only that, but it could also be applied to the internet, allowing data transfer at the same almost-instantaneous speed! Greater than a T3 line even.

I don't think monopoly laws will be an issue. Hewlett Packard have been talking as if all computers will one day have this technology, which would mean that all computer companies would use it. That is far from a monopoly. They seem to be proposing an industry-wide use of it just like with flash.

This would mean great things for the company right? They could sell these chips, to computer and internet-providing companies alike, and make massive profits, right? Well, it's been a few years since I've heard anything about it and I notice no speak of quality improvements in the market, save for Alienware desktops (horribly expensive). With such a great invention under their belt, I'd imagine that the government saw what an advantage this company would get and warned them about their "anti-competitiveness".

That really is a guess with no base to it. There are many other factors, such as lack of funding or lack of proper business management, that could of caused that. Or failed marketing, or the company not yet being ready to release it or even produce it.

Even more years back, when I was reading in a funeral home, I came across an article speaking of micro needles, which was a needle so small one literally can't feel it going in. This health care invention was supposed to help those who are fearful of needles and those who do not like the pain (such as the IV in the hand). I was in middle school then and this is about seven years later. Still no sign of them, and I'd imagine the production process would only be complex, not expensive (microscopic [used as an adjective] needles only need microscopic materials).

Again this is a baseless guess. But more importantly, there are key differences here. The medical industry is one of the most heavily regulated industry while the computer industry is one of the least regulated industries. That is why it has such a innovation, a rate that, to the best of my knowledge, is unrivaled by any other industry. To my knowledge no other industry has a rate of innovation like the one Moore's Law refers to. In fact it looks like Moore's Law still has a lot of life left in it and will continue to do so for ages, or that if it does die it will only be because many technologies currently being worked on (eg optical processers and carbon nanofibre transistors) will see innovation happen at a faster rate than prescribed by Moore's Law. Computer technology innovation almost never (if ever) gets stopped by government. That is why the computer industry has so much innovation; they are free to do so.

Anyway, your post is beside the point. The point is that the technology is a cool and has a lot of potential.

EDIT: Besides, I don't think they could claim a monopoly on the technology. It is a type of resistor. Also, they didn't originate the idea of the memristor. A university professor did in 1971. Resistors are a concept no owned by any one company. At best they could only claim ownersip and thus monopoly on their implementation, which won't be blocked any more than Intel's or AMD's implementation of a CPU transistor will, or Samsung' or, Western Digital's implemtation of hard drive techologies will, or countless other implementations of computer technologies. You don't see any government agency moving to block Intel's monopoly on the use of their unique processor technology or AMD's, or VIA's, or Sun's, etc. Why would their implementation of a memristor be any different?

Also, bare in mind the implementations that Hewlett Packard have suggested are things they do not do. They have suggested their use in CPUs, a replacement for DRAM, as a flash memory replacement, in phones, and all sorts of other things HP do not do. They use tyhe CPU, RAM, etc made by other people. They don't make these things themselves, so clearly they intend to have others make memristors rather than do it themselves. In fact like with Flash I expect to see multiple companies using memristors. Intel and AMD in their CPUS, motherboards, etc; Intel, nVidia, and AMD in their graphics solutions; and many many more such competitive situations that are quite different to the monopoly you assume.

EDIT: The article I linked to below (post #5) confirms that a monopoly would not exist:

But he said understanding this new circuit element could be critical as companies attempt to build ever smaller devices.

Edited May 2, 2008 by DragonMaci

[Grames]

Thank you for this post! First I'd heard of it.

This is not just an invention, it is a breakthrough in theoretical treatment of electrical circuits! Physics and engineering textbooks must be rewritten to accomodate the new knowledge. Its a new integration of Hysterisis into passive circuit elements.

EETimes covers the story with a 3 page article. The memristor (a word formed from memory and resistor) has a new physics explanation on Wikipedia.

Electronic theorists have been using the wrong pair of variables all these years--voltage and charge. The missing part of electronic theory was that the fundamental pair of variables is flux and charge," said Chua. "The situation is analogous to what is called "Aristotle's Law of Motion", which was wrong, because he said that force must be proportional to velocity. That misled people for 2000 years until Newton came along and pointed out that Aristotle was using the wrong variables. Newton said that force is proportional to acceleration--the change in velocity. This is exactly the situation with electronic circuit theory today. All electronic textbooks have been teaching using the wrong variables--voltage and charge--explaining away inaccuracies as anomalies. What they should have been teaching is the relationship between changes in voltage, or flux, and charge."

"Hysteresis is a tell-tale manifestation of the fourth circuit element--the memristor," said Chua. "And Stan Williams is very smart to have realized that if you cannot explain something properly, then there must be a better explanation."

"The memristor is our salvation, because it works better and better as you make it smaller and smaller," said Chua. "The era of nanoscale electronics will be enabled by the memristor. This is not just an invention, it is a basic scientific discovery. It has always been there--we just had to face these nanoscale problems to realize its importance."

Chua postulated the existence of the memristor in 1971. This is the kind of progress for which Nobel prizes are awarded.

[DragonMaci]

This article says "memristor" is a short name and the full name is "memory resistor". That is a much more sensible name. Edited May 2, 2008 by DragonMaci

[DragonMaci]

ts a new integration of Hysterisis into passive circuit elements.

EETimes covers the story with a 3 page article. The memristor (a word formed from memory and resistor) has a new physics explanation on Wikipedia.

Ok, thanks, I will look at both articles and the hysterisis thing and then post here again.

[Grames]

If you are a member of the Institute of Electrical and Electronics Engineers you can download Chua's 1971 paper IEEE Transactions on Circuit Theory, Memristor-The missing circuit element, Chua L.

[DragonMaci]

If you are a member of the Institute of Electrical and Electronics Engineers you can download Chua's 1971 paper IEEE Transactions on Circuit Theory, Memristor-The missing circuit element, Chua L.

I am not a member. I don't have the knowledge and training and nor do I intend to ever do so. If there is any industry based organisation I will belong to in my future it will be equine based, but even that is unlikely.

[DragonMaci]

From the Wikipedia article:

It is sometimes also known as a flux capacitor

I assume the term "flux capacitor" refers to something very different to the one in Back to the Future at least in terms of purpose.

HP prototyped a crossbar latch memory using the devices that can fit 100 gigabit in a square centimeter. The highest-density Flash memories store 16 Gbit in the same area, for comparison.

That is quite the difference. I knew there was a lot more density in the memristor than flash, but not that much. And that is just a prototype. What will future generations of memristors be capable of?

Samsung has a pending U.S. patent application for a memristor similar to that described by Williams. Thus it is questionable whether Williams group is the originator of this structure.

There you go, Benpercent, we already appear to have two companies involved in memristors, rather than a monopoly.

[Benpercent]

I don't think monopoly laws will be an issue. Hewlett Packard have been talking as if all computers will one day have this technology, which would mean that all computer companies would use it. That is far from a monopoly. They seem to be proposing an industry-wide use of it just like with flash.

How does it get to that point?

That really is a guess with no base to it. There are many other factors, such as lack of funding or lack of proper business management, that could of caused that. Or failed marketing, or the company not yet being ready to release it or even produce it.

I do admit I should have Dogpiled before I typed. I am ignorant as to what terms I would use to look up the computer chip I described, but I found out a little bit about the micro needles. It seems it's still in the study stage, but according to this source, one of the problems they were facing in production was the fragility of the needle.

Again this is a baseless guess. But more importantly, there are key differences here. The medical industry is one of the most heavily regulated industry while the computer industry is one of the least regulated industries.

Didn't realize I made a package-deal. Sorry about that.

EDIT: Besides, I don't think they could claim a monopoly on the technology. It is a type of resistor. Also, they didn't originate the idea of the memristor. A university professor did in 1971. Resistors are a concept no owned by any one company. At best they could only claim ownersip and thus monopoly on their implementation, which won't be blocked any more than Intel's or AMD's implementation of a CPU transistor will, or Samsung' or, Western Digital's implemtation of hard drive techologies will, or countless other implementations of computer technologies. You don't see any government agency moving to block Intel's monopoly on the use of their unique processor technology or AMD's, or VIA's, or Sun's, etc. Why would their implementation of a memristor be any different?

So, a question for clarification, a monopoly wouldn't exist because the only protection the technology could get is in its fashion of implementation?

[DragonMaci]

How does it get to that point?

What are you asking?

Didn't realize I made a package-deal. Sorry about that.

Lol, nether had I.

So, a question for clarification, a monopoly wouldn't exist because the only protection the technology could get is in its fashion of implementation?

Basically, yeah. The concept of a memristor belongs to Leon Chua, a professor at UC Berkeley (though he is about to retire). HP at best only owns their implementation of the concept. In fast as already stated Samsung already owns a patent for a memristor similar to the HP Labs one, so HP may not even own that. They may be working with or on behalf of Samsung.

[Benpercent]

What are you asking?

Never mind. I figure it out after thinking about how they can only legally protect their implementation of the technology. The question I was asking was: " How does it get to the point that the technology gets used industry-wide?" Since only the implementation can be protected, I realized that pretty much it's only a matter of time for the technology to spread since other companies can buy the products, learn from studying the technology, and make their edits.

[DragonMaci]

Never mind. I figure it out after thinking about how they can only legally protect their implementation of the technology. The question I was asking was: " How does it get to the point that the technology gets used industry-wide?" Since only the implementation can be protected, I realized that pretty much it's only a matter of time for the technology to spread since other companies can buy the products, learn from studying the technology, and make their edits.

That is one way. Another way is that they can start from scratch and design their own way. After all the concept is an academic one available for free to any member of IEEE, meaning any company can develop an implementation of a memristor from scratch.

Edited May 2, 2008 by DragonMaci

[Prometheus98876]

Very very interesting, especially how the mathematical foundations have existed for 37 years already. I will have to look more into them later, but it doesnt look too complicated from a cursory glance

There are some people raising issues with the security of having things like "RAM" content "permentenlty" stored, but really I think that would be relatively easy to address, by techniques like setting the memristors contents to something useless so reading it wouldnt be of any use.

[Grames]

From nextbigfuture.com: HP plans to release first memristor memory alternative to flash and solid state drives in 18 months

HP intends to have an alternative technology to flash on the market in eighteen months, an alternative to DRAM in three to four years and, following DRAM, a replacement for SRAM, Stan Williams, Senior Fellow at HP, told the IEF2011 meeting in Seville this morning.

From the linked article at EETimes.com

One of the best things about the memristor memory is that it is a simple structure made using materials that are already common in the world's wafer fabs making CMOS-compatible devices relatively straight forward, he said.

This creates the prospect of adding dense non-volatile memory as an extra layer on top of logic circuitry. "We could offer 2-Gbytes of memory per core on the processor chip. Putting non-volatile memory on top of the logic chip will buy us twenty years of Moore's Law, said Williams.

Further out Williams said the memristor could be used for computation under a scheme called "implication logic" in a fraction of the area taken up in CMOS by Boolean logic. In addition a memristor device is a good analog of the synapse in brain function.

I guess the singularity is on track after all.