Hacker News Reader

Comment by btown | Hacker News Reader

Photonic computing, to accelerate both logic gates and data transfer, is an incredibly broad and exciting field. While a lot of the promise is still in the lab, real advances are currently being commercialized.

https://spie.org/news/photonics-focus/marapr-2022/harnessing...

https://www.nextplatform.com/2022/03/17/luminous-shines-a-li...

View on HN

I always had the idea that before jumping to quantum, it would make sense to use photons for as many components as possible instead of the relatively slower, heavier, and much hotter electron.

I don't know enough about computing hardware to know how feasible each component is to be refactored this way, but it is indeed exciting. You could almost imagine such a "photon computer" as a computer which uses little to no energy (at least for the actual computing part), is extremly lightweight due to lightweight components, and never gets hot!

On a larger scale, the Meta Quest 2 uses a USB cable to plug into the computer so you can play VR games on your PC. The max length of the cable is something like 3 feet over copper. The link cable they sell switches from electric signals over copper to light over fiber, and then back to copper to get around the length limitations.

Not really the same thing but still cool!

Any decent quality standard copper USB-C cable works fine, typically 15ft

There's multiple data rates that it can run at, the highest has a reduced length over copper

It's interesting that they did that. One of the main reasons why you only see fiber optics in enterprise applications is that if you bend the cables too much, they're ruined. SFP+ hardware is cheaper and has lower latency than 10GBASE-T hardware, so in a vacuum it would be the obvious choice for everything.

I'm not sure if they just decided the conventional wisdom that consumers would ruin the cable was wrong, or if they figured out how to make them idiot-proof.

Sfp isn't copper vs fiber. It's just the standard for a modular plug on network hardware.

They have 25gbe twinax SFP28s that, as far as I'm aware, have essentially identical performance specs to using optical SFP28s with fiber. The fiber is thinner and can go further, but it also is more fragile and limited on bends. Latency is almost entirely driven by the switch and the NIC.

Photons in fiber do not move meaningfully faster than signals in copper, and speed of light in short datacenters runs is not the limiting factor In latency. Any meaningful processing will happen in electronics and converting from copper to fiber introduces more latency.

There's a good paper on it here: https://www.commscope.com/globalassets/digizuite/2799-latenc...

Tldr coax actually is faster (0.77c) than optical fiber (0.67c).

> Photons in fiber do not move meaningfully faster than signals in copper, and speed of light in short datacenters runs is not the limiting factor In latency. Any meaningful processing will happen in electronics and converting from copper to fiber introduces more latency.

My understanding is that it has nothing to do with the speed of light for short-run cables. 10GBase-T uses block encoding that adds a couple microseconds of latency (which dwarfs the impact of the speed of light at short distances) and fiber (including SFP+) doesn't need to.

Coax/twinax may not need the encoding schemes that twisted pair does, I'm not sure.

My current gig has a similar problem, with a piece of hardware that operates over USB. Cable lengths over 6m or so introduce sufficient voltage and current drop to cause intermittent issues with the sensor (which are really tough to troubleshoot if you don't already know about this).

The solution has been to use fiber optic hubs with the same behavior you describe - copper cables on either end to connect the hardware to the computing unit and a long run of fiber in the middle to make up the required length.

Pretty cool stuff!

It's probably not voltage drop that's making your device operate incorrectly. It's most likely the added capacitance and or inductance of the wires skewing the high frequency data signal out of spec.

Entirely possible it's voltage drop, I don't know the exact hardware, but at 6 meters, it's probably signal rather than power.

Since that cable is USB-C only, and I don't have USB-C ports on my gaming PC, I have a 5m (~16ft) USB-A to C cable. Pure copper all the way, just much heavier than a regular weight USB cable.

I have no issues with data rate or packet loss at all.

I have a 30m HDMI run that I bought several cables that were certified for HDMI 2.0b but all of them were pretty unreliable at that length. After returning a bunch I ended up buying a 50m (that's the shortest they had that would work) Optical HDMI cable and it's been smooth sailing since.

By "lightweight", do you mean the weight of photons?

you haven't noticed that copper wires sag under heavy current?

You jest, but

https://physics.stackexchange.com/questions/9615/why-do-powe...

It's a misconception that electrons are slower than photons. In a vacuum? Maybe. But you need a medium to use photons and in fiber photons go at 0.5-0.75c.

Electronic signals in copper propagate at somewhere from 0.66-0.8c.

The big benefit of photons is that they don't experience electrical interference, so you can often get a lot more bandwidth out of a arbitrarily sized photonic medium than an electronic one.

The actual latency of photons vs electrons is generally not relevant.

Just to unpack "electronic signals", what propagates is the disturbance ("signal"), not the electrons themselves. Like waves propagating at the beach, where water only moves back and forth a little, and slower.

And specifically what causes the electrons to move each other is the electromagnetic field. The boson of the electromagnetic field is the photon, so the signal is being carried from one electron to the next by photons. (Although the disturbance itself as a whole is not a photon; I think it's often described as a quasiparticle called a plasmon.)

Extremely relevant Veritasium video (skip the first two minutes):

https://www.youtube.com/watch?v=oI_X2cMHNe0

Any idea of instead of “traditional” fiber photons, if hollow core fiber is used if this changes the calculus?

> But you need a medium to use photons

Since the speed of light in a vacuum is the prohibitive speed limit of Relativity, I always felt we should develop a medium in which light moved faster than it did in a vacuum, and I swear I've read an article about such a material which was referred to in the article as "ruby," and the images of it reminded me of those glass things at the Fortress of Solitude in Superman (1978). If such a material exists, it would not only make photon computers very fast, but make it possible to violate causality without violating Relativity.

Reply on news.ycombinator.com