Please create an account to participate in the Slashdot moderation system

 


Forgot your password?
Close
typodupeerror
×
Science

Massively Parallel Computer Built From Single Layer of Molecules 46

Posted by Soulskill from the thin-client-solutions dept.
djeps sends in this excerpt from the Physics arXiv Blog: "Japanese scientists have built a cellular automaton from individual molecules that carries out huge numbers of calculations in parallel. ... At the heart of their experiment is a ring-like molecule called 2,3-dichloro-5,6-dicyano-p-benzoquinone, or DDQ. This has an unusual property: it can exist in four different conducting states, depending on the location of trapped electrons around the ring. What's more, it's possible to switch the molecule from one to state to another by zapping it with voltages of various different strengths using the tip of a scanning tunnelling microscope. It's even possible to bias the possible states that can form by placing the molecule in an electric field. Place two DDQ molecules next to each other and it's possible to make them connect. ... When one molecule changes its state, the change in configuration ripples from one molecule to the next, forming and reforming circuits as it travels."
This discussion has been archived. No new comments can be posted.

Massively Parallel Computer Built From Single Layer of Molecules More

Massively Parallel Computer Built From Single Layer of Molecules

Comments Filter:

  • Last year's news (Score:4, Informative)

    by gzipped_tar ( 1151931 ) on Friday October 28, 2011 @11:41AM (#37870328) Journal

    This is impressive discovery, but it's no longer news. The paper was published in April 2010: http://dx.doi.org/10.1038/nphys1636 [doi.org] Admittedly the authors only recently uploaded a copy to arXiv on October 17, but can we not pretend this is some breaking news for nerds?

    • This is impressive discovery, but it's no longer news. The paper was published in April 2010: http://dx.doi.org/10.1038/nphys1636 [doi.org] Admittedly the authors only recently uploaded a copy to arXiv on October 17, but can we not pretend this is some breaking news for nerds?

      Where did you get the idea that /. was about breaking news? The stuff that shows up on here is usually two of: interesting, breaking, accurate. Thankfully, the editors often choose accurate over breaking.

      • Re:Last year's news (Score:4, Funny)

        by Oswald McWeany ( 2428506 ) on Friday October 28, 2011 @12:07PM (#37870726)

        Indeed- many things I read on slashdot I read online elsewhere the week before.

        (I still come over to the story to compulsively comment even if I have nothing useful to say)

        I don't mind the delay- gives me time to gather my thoughts on the issue first.

        I've got a great article on Microsoft's next OS, Windows 7, I'm planning on submitting tonight- supposedly it's going to fix all the problems in Vista...

        • I've got a great article on Microsoft's next OS, Windows 7, I'm planning on submitting tonight- supposedly it's going to fix all the problems in Vista...

          Surely you mean Windows ME?

      • Re: (Score:1)

        by Anonymous Coward

        Where did you get the idea that /. was about breaking news? The stuff that shows up on here is usually two of: interesting, breaking, accurate.

        You forgot advertisement.

    • At least the pdf is no longer paywalled.

    • but can we not pretend this is some breaking news for nerds?

      Yes we can, if we realize that part of the news here is that the paywalled scientific publications are not, well, publications as far as the greater public is concerned.

  • Cricket anyone?

  • Nice summary in MIT Tech Review. Rightly points out that parallel computation is the key to energetically efficient processing, but doesn't mention the first thing that came to my mind, namely that many ion channels expressed by neurons in the brain also exhibit multiple conductance states. I wonder if the computerized intelligence that eventually destroys us all will use arrays of these things in its robo-neurons.
  • Very cool stuff, but... How is this useful information for those of us who aren't chemical engineers?

    I'm still waiting for quantum processors [sciencedaily.com] and biological hard drives [technotips.org] to hit the market.

    • Re:Practical Applications? (Score:4, Interesting)

      by arnoldo.j.nunez ( 1300907 ) on Friday October 28, 2011 @12:04PM (#37870666)
      I research in the related field of memristors. While I agree with skepticism, someone has to first demonstrate that the technology can work. There are typically grants handed out by the government (e.g. SBIR) that spurs interest in first showing that it works over a half a year period, in second developing that idea into more production-worthy product over two or so years, and in finally taking the training wheels off to let people find their own funding to start a small business around a relatively new technology.
    • It's interesting, because certain cellular automata are Turing-complete. Thus, in theory, you could write, say, a C++ compiler for this (or similar) thing.

      Of course, since it's still in the lab, there's no immediate practical application and probably there won't be for another decade, similarly to the examples you quoted.
  • Or at least that's the best explanation I could think of. I wonder how stable this molecules are: If they degrade or are sensitive to light.
  • Past a certain point, aren't most uber-computers just going to end up being used to institute and maintain tyranny (aka death of privacy) - if not by Fatherland Security, then by corporations (assuming these will remain two separate entities)? Careful what you wish for, qubit-wise.
    • They had no problem running the "Fatherland" or practicing oppressive tyranny starting back in the late 1930's and they certainly didn't need a super computer to do it. And do try to understand the difference between privacy and anonymity. If you want to participate in today's civilization you cannot avoid sacrificing a little of each.

    • And with all those people to track, a machine that can perform many parallel operations, is just what they need!

  • With regards to the "one molecule affects the next", there's this kid theory that still considers the universe a type of aether -- the one that was debunked in the Michaelson-Morley experiment last century. You could consider existence to be a big blob of jello and matter as we know it is just perturbations in the jello. Perturb the jello *here* and it will affect the jello *there* to a certain extent, depending on distance, other perturbations, etc. Jello could be swapped with spacetime, of course, but

  • FTFPDF
    Writing, erasing and retrieving information: In Fig. 3a we demonstrate the
    sequential writing of a state 1 matrix on a state 0 surface. The states are stored as static
    information until spontaneous pattern evolution is triggered externally. By scanning the
    surface at -1.68 V one can reset all molecules to state 0, thus erasing the information. To
    retrieve information the surface is scanned at ~0.2 V (Figure 1d).

    Logic gate: The effects resulting from Rule 3 appear similar to the interactions in
    the Billiard

  • Moleculatronic Computer - 125% beam attack! [wikipedia.org]

  • Cool, but not a CA and not parallel (Score:3)

    by Hentes ( 2461350 ) on Friday October 28, 2011 @12:56PM (#37871332)

    This is an awsome project, but the researchers make some claims that are not true. First, this is not a CA, as molecules affect other molecules in a big radius not just their neighbours. Second, a computer is not massively parallel just because it's realized on a CA [quinapalus.com]. That's like saying that silicon-based chips are massively parallel because each of the great number of electrons "computes" its path on its own.

    • This is an awsome project, but the researchers make some claims that are not true. First, this is not a CA, as molecules affect other molecules in a big radius not just their neighbours.

      So isn't that just a highly connected CA? What about a CA where each cell is connected to all the rest - it might behave very differently to a more grid-like CA, but it still counts as one.

      Second, a computer is not massively parallel just because it's realized on a CA [quinapalus.com].

      The image in that link looks like a non-parallel computer in a CA. So, yes, you can throw away the advantages of parallelism if you like; what's your point? They are claiming that their setup could be parallel, not that it must be.

      • Re: (Score:2)

        by Hentes ( 2461350 )

        So isn't that just a highly connected CA? What about a CA where each cell is connected to all the rest - it might behave very differently to a more grid-like CA, but it still counts as one.

        Yes, you can define CAs in a very general term so that everything counts as one but that makes it impractical to model anything with them. And in this case the physical arrangement of the molecules also counts, and if I understood correctly, external electric fields are also applied.

        They are claiming that their setup could be parallel, not that it must be.

        Of course, you can make any computer parallel, for example by using two of them. I was reflecting to the claims throughout the paper like this:

        As an alternative to serial logic operation, von Neumann demonstrated parallel comput

        • Well you're absolutely right that the important thing is what it can do, not what it is called :) I'm not sure how the connectivity of a CA-like computer affects its function. The brain, for example is connected both locally (to nearby neurons) and globally (long-distance axons). I'm no neuroscientist, however, so I don't know how dense the network is.

          Apologies, I thought that you were claiming the opposite - that you can make parallel computers serial. I suppose that replicators in CAs are serial, but I as

    • Re: (Score:2)

      by naasking ( 94116 )

      First, this is not a CA, as molecules affect other molecules in a big radius not just their neighbours.

      You're operating under the assumption that "neighbour" is a spatial definition. X neighbours Y if X can affect Y in one time step. This is often correlated with spatial proximity, but need not be.

  • So, how do I overclock this?

Slashdot Top Deals

No skis take rocks like rental skis!

Close