RoR: Forum

Post 0

Wednesday, May 14 - 5:12pm

Link

Edit

"...may lead to instant-on PCs..." YAY!!!

Brede, this sounds very interesting.

Can you explain a little better for us idiots, or offer some links?

Post 1

Wednesday, May 14 - 7:10pm

Link

Edit

It's similar to the RAM used in USB memory sticks, but it goes further. The same memristor should hint at what it does: memory + resistance. What that means is that it 'remembers' whether you inputted a certain voltage or not and generally will not output anything unless you give it the same voltage (or close to it) in regards to how it works. Also, that means it can store particular states, which makes it better than the RAM since it's a single type of circuit that works best in this form (USB RAM memory sticks basically are NAND gates constructed in such a way as to 'latch' into a particular set of states, but they tend to wear out faster since often they have to be forced to output faster than what they can do under normal conditions to make it practical... Luckily, the wear with NAND memory is even so it would probably take several years of constant use for anyone using such memory to burn it out.).

The part about particular states is important, because circuits by their design are digital. They take a one or a zero and output a one or a zero, there's no middle ground, no maybes, and etc. Memristors have 'maybes' and middle grounds in regards to their state changes as I pointed out prior that they only respond under a similar voltage that activated them first. Thus, their memories can be more analog, smooth, but also granular in that there's a specific state that it can recall versus a computer system that must digitize (to encode) a particular state so it can store something about it. In essence, memristors would be like having an animal-like memory or maybe organic memory in this case, in which we could input a specific sort of information (perhaps raw images or sounds through some sensors) and it could store it as is without encoding it. That's the neat trick here. The immediately practical part of memristors is that mobile devices will be able to access such memories faster than with NAND RAM (I assume X multiple times otherwise there wouldn't be as much buzz about it...), so that means faster devices, possibly cheaper in the long run, and generally more stable (NAND RAM as I stated before burns out over time).

-- Brede

Post 2

Wednesday, May 14 - 8:00pm

Link

Edit

Oh, God, Brede. You lost me at USB. I majored in (Philosophy and) Biology. There are two subjects I simply cannot seem ever to understand. Electronics and Baseball Statistics. Next time you are in NYC I will buy you drinks until I think I know what you're talking about.

Thanks for trying.

Post 3

Thursday, May 15 - 10:02pm

Link

Edit

We need more facts over time. This is too potentially a hype. Google reveals hints and allegations going back 40 years.

We report on a tungsten-oxide-based, nonvolatile, electrically reprogrammable, variable resistance device as an analog synaptic memory connection for electronic neural networks. A voltage controlled, reversible injection of H+ ions in electrochromic thin films of WO3 is utilized to modulate its resistance. A hygroscopic thin film of Cr2 O3 is the source of H+ ions. The resistance of the device can be tailored and stabilized over a wide dynamic range (~four orders of magnitude), and the programming speed is modulated by the control voltage. The suitability of such a device in terms of its response speed, reversibility, stability, and cyclability for its use in electronic neural networks is discussed. Journal of Applied Physics is copyrighted by The American Institute of Physics.
Solid-state thin-film memistor for electronic neural networks
J. Appl. Phys. 67, 3132 (1990)DOI:10.1063/1.345390
Issue Date: March 15, 1990

Components for trainable systems
Hoff, M. Crafts, H. Angell, J. Low, P.
Stanford Univ., Stanford, CA, USA;
This paper appears in: Solid-State Circuits Conference. Digest of Technical Papers. 1962 IEEE International Publication Date: Feb 1962 Volume: V, On page(s): 92- 93 Posted online: 2003-01-06 17:31:36.0

US Patent 5315162 - Electrochemical synapses for artificial neural networks
US Patent Issued on May 24, 1994
"The Memistor is an electrochemical cell in which copper is either plated on or deplated from a carbon rod."
http://www.patentstorm.us/patents/5315162/description.html

Posted by Gian Mattia Porcelli | May 2, 2008 4:43 PM
Posted on May 2, 2008 16:43
James Whong:
"The reason that the memristor is radically different from the other fundamental circuit elements is that, unlike them, it carries a memory of its past"
This isn't right. Both the inductor and the capacitor have memory. The resistor is the only fundamental circuit element without memory.
http://blogs.spectrum.ieee.org/articles/2008/05/the_mysterious_memristor.html

The only Wikipedia entry is in Dutch.
http://nl.wikipedia.org/wiki/Memistor
Op 30 april 2008 maakt HP bekend een electronische component in de theoretische fysica ontdekt te hebben.
My translation: >> on april 30 2008 made HP known an electronic component in the theoritical physics [it's only the little words that are hard] discovery to have.
Deze component die op nanoschaal zou kunnen geproduceerd worden heeft geen energie nodig om data op te slaan, en kan data veel dichter opslaan dan een harde schijf en het kan het even snel oproepen als het ramgeheugen van een pc. Dit impliceert dat het beide in de toekomst zou kunnen vervangen.
{This component which can be produced at nano-scale gives an energy noting about data of its wooden shoes and windmills holding back the north sea smoking tobacco with Calvinists and enjoying the lush Verneer prints.}
http://nl.wikipedia.org/wiki/Memistor

(Edited by Michael E. Marotta on 5/15, 10:10pm)

Post 4

Friday, May 16 - 9:29pm

Link

Edit

Michael, one might think that but HP has it right there working for everyone in the laboratory. So, if you really want to believe that there can never be an analog computer go ahead. I'll be very happy for your proof to be presented in a scientific format and not in a format of questionable quality. Plus, I never go by any Wiki-article for any source material. Try the English Wikipedia if you want the most up to date articles which may give you links which may or may not reveal source material.

-- Brede
(Edited by Bridget Armozel on 5/16, 9:31pm)

Post 5

Saturday, May 17 - 6:01am

Link

Edit

My point was only that there were previous announcements of the "memristor:" Journal of Applied Physics and two IEEE publications.

I cited the Nederlands Wikipedia because that was the hit I got, the most recent citation to the news item you posted. The en.wikipedia.org on "Memristor" has since been edited and now contains the same basic information.

I do not know how you read into my words that I deny the possibility of an analog computer. No matter how I re-interpret my post, I fail to see that meaning in it. You misunderstood me completely. Lest we get too far off the track, let me state here that analog computers predated digital computers. That is a fact which "everyone" knows. If you can present a very early digital device, I am happy to learn to about it, but it is commonly accepted that analog came first.

The "memristor" has a forty-year history. The invention at HP is just another way to implement the same function.

(Edited by Michael E. Marotta on 5/17, 6:09am)

Post 6

Saturday, May 17 - 11:08am

Link

Edit

But HP is the first to make it economical. The first versions were basically built from other circuits (in fact the theorists of the memristor did this and found it wasn't practical for experiments let alone for production). That's the issue here: HP is going to make lots of money and I don't have any stock in them. :(

And thanks for the clarification.

-- Brede