Slashdot Mirror
A Look at Photonic Clocking
zymano writes "In an article on the Electronic Engineering Times site James Siepmann shares a few thoughts about Photonic Clocking. Siepmann states: 'Copper interconnects are reaching their limit as data-transmission bandwidth and processor speed continue to rise. [..] Photonic clocking not only solves the limitations of electronic clocking, but also reduces jitter, skew, delay, crosstalk and power consumption while maintaining clock signal integrity for longer distances.'" Are Photonic Processors the next logical step, or will the almighty buck shuffle them aside because of cost?
Are Photonic Processors the next logical step, or will the almighty buck shuffle them aside because of cost?
Yeah, 'cause technology never gets cheaper. Hey, I've got an AT&T 8086 PC with a lovely green monitor that you can have for $5000, if you act now...
Hang on to it. In a few years, you can haul it down to Antiques Roadshow and have 'em tell you it's "worth between $2000 and $4000, but for insurance purposes..."
"Who are in control, they are not in control of anything - they don't even control themselves!" - Glen Beck
Aren't there like 5 or 6 orders of magnitude between optical frequencies and current clock rates? And last I checked, copper waveguide does
just fine for at least 3 or 4 of those. So how is it that we are "reaching the limitations" of copper?
Nice thing about a pulse of light is that it can be made to reach lots of places at the same time, or nearly so. Just a normal burst of light from a point source has a spherical wavefront, but this can be modified by optics in various ways. Having designed plenty of really fast stuff and having had to deal with skew problems, I can see the advantadge, if real use can be made of it. I think it might even be possible on silicon, which would be required for quick adoption -- after all, the LSI only has to receive, the clock light source can be made of anything. Making a hybrid of course drives costs way up, though. but at current profit margins for fast cpu's this may not be much of a real issue.
If photonic processors go into widespread usage, it will probably be because of the almighty buck and companies deciding that they can make more of it by producing photonic processors.
Profits and competition are the main reason for a lot of the recent advances in processor performance. Look at the processor introductions back when 486 and pentium processors were around and Intel didn't have any credible competition.
"When you sit with a nice girl for two hours, it seems like two minutes. When you sit on a hot stove for two minutes, it
This makes it sound like "the almighty buck" is the bad guy. I think this is one of those times when that's not the case. If fully photonic processors turn out to work best, then that's what we'll see. If they're not, and if the article claim that copper interconnects are reaching their limit is true, then we'll seem some hybrids. Rock on.
This whole article seems like an attempt to pad out a slow news day. Maybe we can turn this article into something useful, or at least more entertaining. We could start a flamewar! Yeah!
----------
<enganging fake troll mode>It's gotta be Photonic chips all the way man!!! Copper procs, yer all gonna burn in silicon hell!! yeah, burn baby burn! I unleash light-based clocking on all you 1nf1d3l5!!
(etc)
Furry cows moo and decompress.
Dude, you and your photonic chips can suck my voltage.
No, this isn't just a copper out, I'm serious. You photonic people seem to think that we're living in a dark age. If photonic processors were at all realistic, we all would have seen the light long ago.
Anyway, take your Star Wars chip and put it in a country where people have never seen George Lucas's masterpiece, Episode 1.
I actually read that as photonic cooking...but to each their own :)
Join the TWIT army now!
The article didn't say a whole lot, did it? It just said, "Gee, wouldn't photonic clocking be nice". It didn't say a whole lot about how, and whether it's feasible.
So, I'll quickly fill in what I know. To do clock distribution you need two types of components: waveguides and detectors. Let's assume you are going to work in silicon...
Waveguides function as the optical wiring, and includes things like bends and splitters. Although perhaps not trivial, it is relative straight-forward to make waveguides in or on silicon. Detectors, on the other hand, are not so easy, at least at the wavelength most people are interested in, 1550 nm. There's a number of people researching Ge growth for detectors on Si, and this does have promise, but it's not ready yet. Another option would be bonding InGaAs, but that might always be too expensive.
Now, if you want to do full up optical communication, on chip, you'll want modulators, too. These have been demonstrated by Intel and Cornell in silicon, but only at speeds around 1 Ghz. Optical amplifiers would be nice, too, and this has been demonstrated (using Raman amplification) by Intel and UCLA. (I'm not sure Raman amplification can give you the sorts of amplification and efficiency you really need, though).
(Sorry, I won't be able to respond to any replies; at least not until Monday. I'm off to bed and I'm not planning to be near a computer tomorrow).
Assuming that the clock circuity takes up 30% of the chip, wouldn't manufacturing a chip with both photonic and electrical circuitry be more expensive than just manufacturing a purely photonic chip?
Vintage computer adverts: http://www.vintageadbrowser.com/computers-and-software-ads
Here, let me explain. The processor is made from a neutronium-alloy, which siphons quantum energy from a Sea of Dirac to produce a photonic current. This is channelled via plasma conduits to the high-frequency phase conducer, where creates a precision clock signal with an error margin of less than 0.000000386 parsecs.
You look beautiful! Incidentally, my favorite artist is Picasso.
clockless CPU's, which of course--wait for it--don't need a clock. (I realize that other system parts may still need it.) Every once in a while, I hear a tiny thing about clockless chips, but it seems like the Big 2 don't want anything said about them.
Reminder: this comment is on topic.
The difference between spam and poop is that you don't have to dig through septic tanks looking for real food. -- Me
Distributing your clock with photons imples you have a photon wave guide. If you are going to build a photon wave guide then why not build an electrical wave guide. Electrical wave guides, like for example coax cable, have wave velocities that are faster than light in glass, so they would logically be even better. And you dont' need any special materials like you would for optical wave guides.
The problem might be that usually wave guides have to be the size of the wavelength to work right. ghz wavelength are larger than the chip. Thus you get forced towards the optical region by this considerarion.
But you can beat this two ways.
1) use negative index of refraction materials. Then the waveguide can be smaller than the wave length
2) use near field waveguides with amplification. When the wavelength is a lot larger than the waveguide then the wave becomes evanscent (decaying). So it can't propagate very far. But hey, that's okay because the chip is not very wide either, so we can tolerate some loss of signal. And we could toss in some amplification to offset it.
Some drink at the fountain of knowledge. Others just gargle.
Transistors don't need clocks, logic gates don't need clocks, but flip-flops do. The reason you need a clock is because the outputs of a bit of logic will be 'unstable' for a while the result is computed. The clock tells the next piece of the system when to read. In place of that, you'd need a 'done' signal, which would rase transistor counts quite a bit. Not to mention it would be very hard to find people who would know how to design these things. I think the future of the CPU involves different parts of the system operating on separate clocks, transferring data via a 'networking' type system. Computers connected via Ethernet don't need to have their clocks synched in order to work. Think of a simple instruction decoder. The decoder reads the instructions, and opens the right 'gates' in the CPU so that there is an electrical connection between the two registers and the ALU, and inside the ALU to the adder or subtractor, or whatever depending on what instruction you're trying to run. Then, the clock signals and tells the ALU that the registers are ready. Without the clock, the ALU might try to add the wrong things. (the ALU doesn't need a clock to work) In the future you could have some sort of system where the decoder just sends a message to the ALU telling it to setup the adder, and to the registry file to access these two registers. Then the register file will send the data to the ALU whenever it's ready.
autopr0n is like, down and stuff.
Consider the semiconductor. The way we work today is based on binary elemental logic-- on, off, unknown/neutral. Your basic light switch (SPST) is your basic computer, but it can't count too well.
The evolved state of computing uses Boolean logic to mosh states together into integer algebraic, then other kinds of math transformations.
Now, consider what light does, and how it flows. Light (actually this segment of the electromagnetic spectra) has different frequencies, at about the same data rate depending on media. No information there, except frequency differences and blendings of frequencies... lambda moshing.
You can modulate light, like any other electromagnetic phenomena. You can modulate information, therefore, onto light. It's done all the time. By adding information, you can blend things together, then demodulate them to see what happened as the change in information. This modulation mimics how ALUs/accumulators/CPUs work with logic states in some ways, but now we have to multiply the effect to get to significant digits and significant logic handling-- math by light modulation and the devices that can do that. But not densely, so far, in the calculative/logic-state change tracking sense.
What of these devices-- aye, thar's the rub. Is there an advantage to using light to do math? Not yet, really. It doesn't meet the state change efficiency model. One day, it might. Today, we lack the ability to make things dense enough. That's why photonic logic may fall short of expectations.
---- Teach Peace. It's Cheaper Than War.
Not a credible player.
Back in the day, "real" computer manufacturers scoffed at Intel. IBM would only let them produce the chips for the PC after Intel found another manufacturer willing to produce the part in case Intel tanked. The PC was nothing to boast about compared to the mainframes of the day.
Slowly but surely, Intel grew to become the monster they are today. The turning point was somewhere near the Pentium II, when Intel machines were beginning to be used as engineering workstations. Profits truly are the source of competition and progress. Back then, the PC market was small, and improvements came quickly only because things were relatively simple. Now, everyone wants a piece of a growing pie, and companies are innovating as fast as possible.
LaForge: Gee, I don't know... These RocketIO bus are mighty archiac and PCI even more so.
Data: RocketIO is rated at 9.8 Gbps and PCI-X6 is rated at even slower rate of 8 Gbps aggregated.
LaForge: Yeah, right. Nothing compare to our photonic bus of 980 THz over each of the 2^1024 channels.
Scotty: Ayie! Why don't they get with the program, laddies? I kin nev'r understind them, bloody buses.
The "almighty buck" won't shuffle it aside...people will. If the processor is not cost-effective then it won't catch on. However, the processor may become the SUV of chips. It may not be the most cost-effective solution but a purchaser may feel it helps him compensate for his undersized penis.
You know, when I started writing this post that's honestly not where I was going. I was going to make some point about how marketing may overcome the possible lack of value of the chip (a la VHS vs Beta) but then the post just headed south...
It might not be compensation, instead it might be the benefit of the power savings if it's significant enough. If this would double the uptime of a battery powered notebook it might be worth it even if it costs more.
The man who trades freedom for security does not deserve nor will he ever receive either. - Benjamin Franklin
"I for one welcome our Almighty Buck overlords... personally. Just keep sending them my way."
Don't think of it as a flame---it's more like an argument that does 3d6 fire damage
A processor, greatly simplified, is a collection of logic gates. These logic gates, greatly simplified, are nothing more than modulators. In hardware design, the modulation of the electrical signal indicates the result of the logical function of the circuit. Electrical impulses are measured in cycles/second.
:)
Photons can achieve frequencies in vast excess of current processor speeds. The function of a photonic logic gate would be measured by simple amplitude modulation. A photon has a frequency and an amplitude. Using a photon with the energy of a gamma ray would be _FAST_, have negligable heat loss due to the friction which plagues electronic processors, and the amplitude of the photon could be easily modulated by passing through different materials. Different materials of different refractive indeces and transparencies (see fiberoptics) would be the photonic equivalent of electronic resistors and capacitors.
I can only wait for the development of photonic processors.
fast as fast can be. you'll never catch me.
I thought the article said "platonic" clocking... I was thinking... I would hope they loved their clock..
Note to self: Don't read Slashdot too late at night..
No, really. As soon as it gets anywhere near a point where there's a large-enough market for it, it'll be sold. Witness the present collection of chips - do you *really think that the majority of people using computers need a P4 at >3 GHz? No, they don't. The minute the niche market of gamers (and, yes, it'll be gamers) who can afford it is large enough, it'll hit the market.
Ethernet does rely on synchronised clocks. You might be misstating that Ethernet doesn't have a clock line, meaning there is no dedicated wire with a clock signal on it.
There is a high-precision clock on every Ethernet card. An Ethernet frame has a 64-bit preamble with Manchester encoding. That preamble adjusts the skew of the receiver clock so that it's synchronised with the transmitter clock. If the synchronisation didn't occur, you wouldn't know when to latch the data on the line and you couldn't receive a frame. The synchronisation occurs on every Ethernet frame and the precision of the clock must be high enough that the synchronisation lasts for the length of a frame.
Async architectures will likely use a similar technique. The subsystems won't be driven by a system-wide clock line, as in the existing synchronous architectures, but the various clocks in subsystems will certainly be synchronised.
I find it amusing that you are picking one part of a piece of technobabble, and ignoring the rest. If you were going to correct the technobabble:
* Neutronium may not even be physically possible, and certainly would be instantly highly explosive in Earth-conditions. It certainly wouldn't "alloy", and has nothing to do with the theoretical Dirac Sea.
* Zero-point energy is not related to processors
* You can "channel" photons, but not a "photonic current", which is at best a term to describe the amount of current that can be produced by receiving a stream of electromagnetic radiation.
* What would "plasma conduits" have to do with either light or current?
* There is no such thing as a "phase conducer"
The thing that makes technobabble amusing is that it's *wrong*. It's pseudoscientific. It's fake. If something was theoretically possible, and terms were used correctly, it wouldn't be technobabble.
You look beautiful! Incidentally, my favorite artist is Picasso.
AMD will also sue Intel now, in fear that the last part comes true.
Slash-for-Thought
And move on to Photonic CANNONing. After all, the Borg are gonna hit is one of these days, and not just through Bill
I have a feeling that quantum computing will happen before photonic computing. That's just me though...
-illumina+us "I put on my robe and wizard hat..."
erm ... a waveguide is a waveguide, no matter what kind of terminators you use. The pertinent condition is to support propagation modes.
Judging by what's coming out of Hollywood lately, the other thing must be shit.
--S
-- sigs cause cancer.
Few years ago, I read the book "Asynchronous Circuit Design" by Chris J. Myers, in which it is explained the basics for designing digital asych logic by using special techniques and gaining advantages in power consumption and systems speed and, to stay on topic, no clock to be propagated through the silicin at all.
The real questions are:
- why these tecniques never gained importance (the first studies are from S.Unger in 1969)?
- what about the EDA industry?
-- See you, UncleScrooge
>>Sorry to parent, but people seem to be taking this seriously so I gotta point out that this is BS so hopefully noone takes this seriously...
Oh? Read on.
>"Consider the semiconductor."
>>Ok, here is the parent posts first fundamental misconception. Digital doesn't necessarily mean semiconductor. Say, for example CDs which encode digital data using light.
No, you misconstrue it. Transistor logic is what's used to do state changes that amount to the various relationships that form what a CPU does. Go back to your basics. We need the equivalent of optical accumulators in dense forms to make photonic processors feasible.
>"By adding information, you can blend things together, then demodulate them to see what happened as the change in information."
>>that isn't how light works, the waves superpose ontop of each other, just like every other kind of wave...
>"This modulation mimics how ALUs/accumulators/CPUs work"
>>No, no it doesn't.
Yes, it can. When you modulate, endowing information, and sum the modulations, you do the same thing as changing states in the lowly semiconductor when merged modulations are seen by an optical accumulator or reflective accumulator like your own eye. Yes, you've changed the information to create new states that accumulate information. Then it needs to be stored or moved on to be changed again to suit the calculative ends of the program.
>>There is a difference between storing and transmitting information and actual computation. You need some kind of devices which emit light depending on light inputs implementing AND OR and NOT logic.
Yes, we agree that to satisfy Boole's needs, this must be done.
>>Not helping is that this guy sounds like an idiot:
>>WTF does "mosh" mean anyway?
Sum, integrate, push together to form a meta value. I can tell you've never had a good punk rock experience.
>>"integer algebraic"?
How much about processor theory do you understand? Is integer algebra via binary summation foreign to you? Perhaps you're not familiar with integer algebra as the root method to obtain the basics of microprocessors.
>>"the electromagnetic spectra" (I guess theres more than one of the electromagnetic spectrum)
I can see that you're on a tangent, here. Yes, there are all sorts of subsections of the spectrum. Consider these spectra as finite sections of frequency ranges. There is one total spectrum. But Ethernet, a baseband technology, starts at 0hz and goes to varying heights of frequency; it's not modulated onto another carrier.
>>"your basic light switch is your basic computer" see above, a SWITCH is one bit of memory -- computation means implementing AND OR or NOT, at a minimum.
An SPST switch is either zero, or one. It's the most elemental binary calculator there is. Run your program and the switch closes, or it doesn't-- it's a computer and you control the logic state. But because it has either state, it isn't a static value. You can add logic onto multiple switches in numerous ways. You can then build truth tables, and so on. This is how the first computational devices were built via binary logic accumulators.
So, mod me down if you'd like. Moderation isn't the point-- photonic CPUs are the point. Building dense arrays of photonic sensors that can have state changes as a result of merging light sources (think loosely of colors, like red and yellow merging to orange, with orange as the new piece of information) can have future application. We can modulate light, change its frequency, make it do tricks by interrupting it in various modulations, then pushing them onto an observeration point to discern changes.
The modulations and summations can be programmed. The result of these inputs are an output. That's what we do in computing--> have inputs, process them, and do something with the outputs, now on a grand scale with the evolution of microprocessor integration capabilities and the surrounding chipsets that make use of I/O.
---- Teach Peace. It's Cheaper Than War.