Please create an account to participate in the Slashdot moderation system

 



Forgot your password?
typodupeerror
×
Patents AMD

New AMD Patent Proposes Teleportation To Make Quantum Computing More Efficient (tomshardware.com) 48

A team of researchers with AMD have filed a patent application that looks toward a more efficient and reliable quantum computing architecture, thanks to a conventional multi-SIMD (Single Instruction Multiple Data) approach. Tom's Hardware reports: According to the application, AMD is researching a system that aims to use quantum teleportation to increase a quantum system's reliability, while simultaneously reducing the number of qubits necessary for a given calculation. The aim is to both alleviate scaling problems and calculation errors stemming from system instability. The AMD patent, titled "Look Ahead Teleportation for Reliable Computation in Multi-SIMD Quantum Processor," aims to improve quantum stability, scalability, and performance in novel, more efficient ways. It describes a quantum architecture based on quantum processing regions: areas of the chip that hold or can hold qubits, lying in wait for their turn on the processing pipeline. AMD's approach aims to improve on existing quantum architectures by actually reducing the number of qubits needed to perform complex calculations -- via the science fiction-esque concept of quantum teleportation.

AMD's design aims to teleport qubits across regions, enabling workloads that would theoretically require in-order execution to become capable of being processed in an out-of-order philosophy. As a quick refresher, in-order execution features dependencies between one instruction and the next, meaning that a workload has to be processed sequentially, with later steps dependent on the previous step being fully processed and its result being known before the chip can move ahead with the computation. As you may imagine, there are chip resources (in this case, qubits) that sit idle until it's their time to perform the next calculation step. On the other hand, Out-of-order execution analyzes a given workload, figures out which parts of it are dependent on previous results and which are not, and executes every step of the instruction that doesn't require a previous result, thus improving performance via increased parallelism.

AMD's patent also includes a look-ahead processor embedded into the architecture, tasked to analyze the input workload, predict what steps can be tackled in parallel (and those that can't), and appropriately distribute the workload across qubits, using a quantum teleporting technique to deliver them to the required quantum processing, SIMD-based region. How this quantum teleportation occurs isn't described in the patent -- it looks like AMD is playing its cards close to its chest on this one.

This discussion has been archived. No new comments can be posted.

New AMD Patent Proposes Teleportation To Make Quantum Computing More Efficient

Comments Filter:
  • Filing a patent doesn't mean they're actually "working in it", it just means they filed a patent.

    • Re:Working on it? (Score:4, Interesting)

      by drinkypoo ( 153816 ) <drink@hyperlogos.org> on Thursday September 02, 2021 @08:45AM (#61756365) Homepage Journal

      Which, frankly, is a travesty of the whole system, and actually discourages progress. Because if you can get a patent for just imagining something, but don't have to work out any of the details, it decreases the value of actually doing it because whoever actually does figure it out now has to either work around your patent, or share the credit and thus the profits.

      Nobody should be allowed to patent anything without proving that they can actually do it, period.

      • by Anonymous Coward

        Amen. If you can't deliver it in box, now, today, it shouldn't be patentable.

      • by PPH ( 736903 )

        and actually discourages progress

        They are called blocking patents. They keep people from pursuing competing technology. Like oil companies generating a lot of solar/wind/battery patents.

      • I think there is a sweet spot for patients in terms of Complexity vs Ingenuity.
        We tend to like patents for the invention that is simple, however someone ingenious and original. The "Why didn't I think of that" type of invention. That once you see it, it just seems so practical. For that case, we would want to reward the inventor on such an idea.
        However they are so many others are solution to a novel problem that needs to be solved, where such an invention can be created by any smuck who has came up with

  • by Mascot ( 120795 ) on Thursday September 02, 2021 @06:06AM (#61755867)

    How this quantum teleportation occurs isn't described in the patent

    Then it shouldn't be granted a patent, now should it. A patent is supposed to contain enough information for someone else to be able to implement the invention. I know, it's been decades of this crap and it's not going to change, but it still irks me that this has become the norm.

    • by tragedy ( 27079 )

      What's really frustrating is that if they develop the technology, the patent will be treated as valid to protect it. If someone else develops it, they can use the patent to claim ownership over the invention, but if the patent expires and then they develop it, they will just get to file a brand new, more specific, patent and it will be treated like a new filing.

    • Re: (Score:3, Informative)

      by SirSlud ( 67381 )

      It depends what they're patenting. You don't need to describe everything in a patent - a patent for something that keeps ice cream cold doesn't need to describe how to make ice cream, because you're not patenting the mechanism to make ice cream.

      So this patent may be looking to lay claim on the architecture that involves having teleported qbits but not the invention of how you teleport them. (Maybe that's in a separate filing, or outside the scope of the claim)

      • by gtall ( 79522 )

        Hey, neato. So if I come up with an architecture that uses the magical properties of pink unicorns, then I can get it patented. How you get them for your device is your problem, but if you do, I get a cut. I suppose you could use white unicorns, but they don't have the magical properties.

      • by PPH ( 736903 )

        So this is akin to a patent that claims to perform some prior art "using the Internet". Before the Internet was invented.

        The way I read the "implement by a person skilled in the field" requirement is that if required the parts aren't in existence, you can't implement it. So you can't patent it.

      • by Mascot ( 120795 )

        I agree, you don't need to describe every underlying technology. But that technology should exist. Being able to patent ideas as opposed to actual inventions, is why we're having today's issue with patents and patent trolls. The fact that software can be patented at all is a problem in itself, but the larger issue is the ability to patent broad notions as opposed to specific technical designs. To run with the ice cream example, you should not be able to patent a box to keep ice cream cool, before refrigerat

    • > but it still irks me that this has become the norm.

      The old system may have been nicer, but at this point realize that USPTO is part of the corporatocracy and exists mostly to ensure that huge corporations don't face much competition from new market entrants.

      The Patent system is now primarily the enemy of people who wish to advance the progress of science and the useful arts.

      cf. No state shall make anything but gold and silver good for the payment of taxes. The last remains of the old republic have bee

    • These patents should be both approved and rejected at the same time. How it is enforced will be influenced by those who view the patent. This will create a state where there can be identical duplicate patents that its legal status will always be in question, and perhaps can be in different phases of the legal process at any random point.

    • How this quantum teleportation occurs isn't described in the patent

      Then it shouldn't be granted a patent, now should it. A patent is supposed to contain enough information for someone else to be able to implement the invention. I know, it's been decades of this crap and it's not going to change, but it still irks me that this has become the norm.

      The patent can be a theory based on some given facts more or less, I believe. The term "teleportation" is what strikes me as open to interpretation, just guessing.

  • Relevant xkcd (Score:5, Interesting)

    by JoshuaZ ( 1134087 ) on Thursday September 02, 2021 @06:09AM (#61755873) Homepage
    Relevant xkcd https://xkcd.com/465/ [xkcd.com] thinking of quantum teleportation as "science fiction-esque concept" is not helpful. Quantum teleportation involves taking a particle, entangling it with another set of particles, so the original state no longer exists, and then applying a new unitary transformation at the end to the entangled particles to recover the original particle's state. The most interesting thing is that during this process one doesn't find out directly the state of the original particle. But thinking of it as "teleportation" is only useful in a very, very limited sense.
    • AFAIK and if I understand quantum teleportation correctly - which I probably don't - it has the potential to become the holy grail of communications: hacker-proof, interference-proof, basically unlimited bandwidth on unlimited distances. i.e. real-time internet between Earth and Mars, for example. Or unmanned probes being controlled remotely in real time.

      This would change everything. The only wires we'd ever need would be to power things efficiently. No more cables for TV, phone, internet. No more satellite

      • It seems like you have some misconceptions here. First of all note that there is an idea for eavesdropping-proof communication using quantum bits, but it doesn't involve quantum teleportation. See quantum key exchange https://en.wikipedia.org/wiki/Quantum_key_distribution [wikipedia.org]. Second, all processes here are limited. You cannot use any quantum process to communicate faster than the speed of light unless we're wrong on very basic things. This is a consequence of the No Communication Theorem which essentially says
    • good point. i wish i'd read ahead before i posted. entanglement theory.... =)
  • AMD only needs to work on resizing them to fit in the chip packages but I presume they already filed a patent for that one.

  • What's their definition of "quantum" here? It seems to be the "infinitely fast computer" based on discussions here in the past.. but an infinitely fast chip would need infinite power... so we'd move the limit of how fast a computer can go to the power supply instead of the processor. I'm not sure if this patent is just a buzzword document.

    • The "infinitely fast" part is only to transmit information between one part of the system to another, it doesn't change the speed of the parts processing the information.

      As an example, just because my ping is 5ms doesn't mean the person on the other end will receive my email 5ms later.

      • The "system bus" can only run as fast as the central processor (the CPU) or otherwise there'd be a need for a large amount of cache memory before the processor... the 6 Gbps speed of Sata connections can't be realized in system that don't have 6GHz processors, and Moore's Law isn't there yet.

        • The "system bus" can only run as fast as the central processor (the CPU) or otherwise there'd be a need for a large amount of cache memory before the processor... the 6 Gbps speed of Sata connections can't be realized in system that don't have 6GHz processors, and Moore's Law isn't there yet.

          Congratulations, you have achieved a state of confusion so dense you're not even wrong. Might as well complain that 6 gbps SATA can't work until you can use a live trout to unclog a drain. It's just as irrelevant as your complaint.

          • Nah, he's just wrong.

            He forgot to account for bus width, which is stupid and ignorant and lots of other stuff, but it's not an otherworldly level of stupid. It's typical.

            That, or it was a clever enough troll to spawn a whole discussion thread

    • No, quantum computers are not infinitely fast. In particular, we have clear cut limits on what they can do. We know that essentially anything a quantum computer can do, a classical computer can do also if given a lot more time, roughly an amount of time which is exponential in the time it takes the quantum computer. More formally, BQP https://en.wikipedia.org/wiki/BQP [wikipedia.org] the set of things a quantum computer can do in polynomial time is contained in EXP https://en.wikipedia.org/wiki/EXPTIME [wikipedia.org] the set of things a
      • Well, part of the problem with factoring positive integers is Intel's intentional decision to put basic math functions such as addition, subtraction, multiplication and division at or near the center of the chip, which means there's problems removing the heat generated by any known cooling system (fan or liquid). You can't compare this theoretical computer one based on the "GigaChip" (a 1 GHz processor from Intel or AMD... because at the time AMD had the Intel designs available to it, and was collecting cas

        • That really doesn't have anything to do with this at all. The difference is architecture independent. The best algorithms we have on a classical computer is the number field sieve https://en.wikipedia.org/wiki/General_number_field_sieve [wikipedia.org]. Given a sufficiently large number, a quantum computer will do things in polynomial time while the classical one will not.
          • Again, no numbers in your post. Wikipedia that you link to considers the advantage of that theory is only felt at a gogol (10 to the 100th) or higher... which is quite out of range of today's 64-bit chips.

            • Again, no numbers in your post. Wikipedia that you link to considers the advantage of that theory is only felt at a gogol (10 to the 100th) or higher... which is quite out of range of today's 64-bit chips.

              Hmm? If your question is when these become practical that's a difficult one to answer. That's to some extent a function of the number of qubits you can have. And right now the technology is in its infancy. For example, the largest number factored successfully with Shor's algorithm right now is 21 https://ui.adsabs.harvard.edu/abs/2012NaPho...6..773M [harvard.edu]. But as things improve, quantum computers are getting better. Where the cross-over point will occur isn't by itself clear. But if our current understanding of

  • Such a lost opportunity for using another buzzword filled with false promises.
  • There is no actually working quantum computing system in existence. Hence you cannot make them "more" reliable or "more" anything.

  • This technology will do wonders for my commute, now that going to the office is back on the table.
  • This is something nobody is likely to infringe upon. But I bet it looks good in the portfolio.

    Maybe they will get the competition to waste time trying to implement the idea.
  • by thegreatemu ( 1457577 ) on Thursday September 02, 2021 @10:57AM (#61756907)

    Why is it, on a site devoted to science and technology, every story about quantum computing is treated like it's completely made-up snake oil? Quantum computing is very real, and there is a ton of research being poured into it. Quantum teleportation is also a very real, demonstrated phenomenon. It's not even that complicated if you remember at least some basic linear algebra.

    • Quantum teleportation is also a very real

      O really? Quantum Teleportation is based on Quantum Entanglement which is just the equivalent of two matched marbles each in a box shipped long distance then opened, looked at, then smashed at an opportune time.

      -at least that's what all the discussion in these types of articles claim.

  • --Jeff Goldblum / Brundelfly enters the chat...--
  • As if having to deal with the Heisenberg's Uncertainty Principle wasn't enough when I go looking for my quantum computer, now it will be able to teleport too? Luckily, I put an AirTag on it -- and a bell.

Every nonzero finite dimensional inner product space has an orthonormal basis. It makes sense, when you don't think about it.

Working...