11 ms·
New Optical Switch Up to 1000x Faster Than Transistors
- fiddlerwoaroof 5y agoDoes this work at normal temperatures? Or does it need impractical cooling setups?
- sleepy_keita 5y ago> the new optical switch works at room temperature
- neltnerb 5y agoThese sound a lot like room temperature optical qubits... I'm certain I am missing something but this sounds pretty neat.
- brokensegue 5y agothey aren't quantum
- westurner 5y agoAre they polarized photons though?
- rocqua 5y agoThey probably aren't in a superposition, let alone entangled with each other.
- neltnerb 5y agoGotcha, if I am reading this right the mechanism for the switch is quantum but the property is not. So no different in that regard than in semiconductors where the underlying behavior is due to quantum mechanics but the actual computing elements are functioning like analog non-linear elements. Thanks!
- fiddlerwoaroof 5y agoThis is what I get for skimming the article.
- kfprt 5y agoIf this is the transistor moment what technology would an IC equivalent require?
- max_ 5y agoFirst we need a new paradigm of computing that is consistent with how light behaves.
- singularity2001 5y agoImprecise neural networks?
- Lichtso 5y agoSome questions from someone who is simply interested in the topic and seeks to expand their understanding. 1. For electrical circuits we are interested in properties like resistance, capacity, inductance etc. for passive elements and in the characteristic curves for active ones. What are their equivalents in photonic circuits? Which properties matter here? 2. In digital computing I think we abstracted the electrical analog properties and the differential equation level of the problem away pretty much. Can't we achieve that with photonic ones as well? Or do you mean a paradigm shift in higher level things like memory bandwidth and latency vs caching etc. ? 3. When I look at the state of photonic circuits I think miniaturization and fabrication are the most obvious problems. Let's assume for a moment that these can be solved, what else makes photonic circuits worse than electric ones?
- max_ 5y agoUnfortunately I am no expert. Most of these questions go beyond my knowledge. But you should definitely watch this video [0]. [0]: https://youtu.be/EwueqdgIvq4 https://youtu.be/EwueqdgIvq4
- dvh 5y ago3-4 football fields
- dogber1 5y agoIn reality, all-optical computing is mostly a terrible idea: fundamentally, it cannot reach the integration density of electronics. It boils down to the elementary differences between Fermions (electrons, neutrons, etc.) and Bosons (photons, etc.). Their intrinsic behavior determines the interaction with matter, i.e. conductive/absorptive properties. As a result, optical wires (waveguides) have to be sized roughly at a wavelength (hundreds of nm), whereas electrical wires can be much smaller (<30nm and below). Suppose you want to build an amplifier: all the claimed speed benefits of this optical device would vanish in the path delay of the feedback loop. But just like graphene, carbon nanotubes, and other fads, you can publish fancy papers with it.
- jcims 5y ago(I swear I’m not in Fridman’s payroll.) As a layperson I found this episode with Jeffrey Shainline an interesting discussion tangential to the topic of optoelectronic computing. The basic gist was that photons are good for communication, electrons are good for compute. https://youtu.be/EwueqdgIvq4 https://youtu.be/EwueqdgIvq4
- rewq4321 5y agoThe specific timestamp for that part of the conversation: https://youtu.be/EwueqdgIvq4?t=2793 https://youtu.be/EwueqdgIvq4?t=2793
- hrgiger 5y agoNot in the domain but as far as I understand from video for the counter argument: "Adding a matter of component and creating this kind of liquid of the light you add interactions in the system so practically light interacts more efficient with the light" https://youtu.be/Kv25Dw-IuCM?t=675 https://youtu.be/Kv25Dw-IuCM?t=675
- WithinReason 5y agoThis makes me wonder if transmitting data optically would help with the gradual lowering of the data/compute ratio over time: https://sites.utexas.edu/jdm4372/files/2016/11/Slide16.png https://sites.utexas.edu/jdm4372/files/2016/11/Slide16.png
- perihelions 5y agoI *think* this is a recent-ish preprint of the same paper they're talking about. (The current Nature version is neither open-access nor accessible on Sci-Hub). https://arxiv.org/abs/2005.05811 https://arxiv.org/abs/2005.05811 (E.g., both papers contain this [0] text string pointing to the same URI. But, the arXiv preprint doesn't mention the Nature submission). [0] "All data supporting this study are openly available from the University of Southampton repository at https://doi.org/10.5258/SOTON/D1374 https://doi.org/10.5258/SOTON/D1374."
- hrgiger 5y agoAlso IBM announcement here: https://research.ibm.com/blog/single-photon-switching https://research.ibm.com/blog/single-photon-switching
- kragen 5y agoIt's very regrettable the author of this article omitted the citation to the paper. Why did they do that?
- perihelions 5y agoIt's linked in the last paragraph.
- kragen 5y agoOh thank you, I'm an idiot. Somehow I overlooked the underline: https://www.nature.com/articles/s41586-021-03866-9 https://www.nature.com/articles/s41586-021-03866-9
- perihelions 5y agoIf you lack access to it, there's a draft preprint on the arXiv, https://arxiv.org/abs/2005.05811 https://arxiv.org/abs/2005.05811
- kragen 5y agoThanks!
- rpmisms 5y agoFrom the discussion here, it sounds like this is going to be a specialized tool for now. Perhaps we'll end up seeing optical/ic hybrid chips before too long
- vekerdyb 5y agoI always knew all those RGB lights will come in handy one day
- Lichtso 5y agoWould this make sense in domains which are not about computing, but simply routing light (given it is miniaturized)? E.g. using this instead of mirror arrays in movie projectors?
- friseurtermin 5y agoCould something like this be useful in space? With less heat generated (?), less power requirements (?), and immune to SEU(?)
- Veedrac 5y ago> A new optical switch is, at 1 trillion operations per second, between 100 and 1,000 times faster than today's leading commercial electronic transistors No, a CPU's clock speed is not the speed at which its transistors can transition. Clocks are multiple gates long (roughly, 10-20), and are not the only delays in a system. So a CPU running at a clock speed of 5 GHz could reasonably be switching each transistor at a rate approaching 100 GHz. (Most of the time each individual transistor is not switching, since only their last switch within a clock cycle is typically meaningful.)
- pickledcods 5y agoThis is absolutely amazing and is the perfect gift for the research I'm working on: fractal logic and calculating with structures. The concept is to break down (algebraic) operators into a single fundamental operator and a single reference value (zero/null). And because there is only one operator and one value, they can be removed from the system. What remains is the "essence of information" in the form of connections, opening a new realm of problem solving, especially in the realm of SAT solving. It's not about what the operator does, it's about how they are connected together. The "fractal operator" is a switch of which one input is inverted. The expressions/structures are stored as vectors and evaluated by activating them sequentially. For some years I visualised it being implemented with photons instead of electrons. And assuming that only a single photon is needed to set the switch state. This is a wish come true! [0] https://github.com/RockingShip/untangle