Slashdot Mirror
Fiber On Your Motherboard...Soon!
km790816 writes: "In this post I joked about wanting an optical bus on my PC. In the last week I've seen two articles from The Register and EETimes discussing the real possibility. Both mention high bandwidth and lower heat and power usage. Sounds good to me."
Is this really the next thing in technology we need? Seems to me that ability to attain high motherboard speeds isn't as much of an issues as getting one that is reasonably priced. Why do I have the feeling that fibre is not a cost-effective solution?
The bigger bottleneck in the system is still the time it take to seek and retrieve data from the hard drive, even more so with fibre busses.
--- RFC 1149 Compliant.
As nice as an optical bus on my m-board would be wouldn't there be a rather large slow down due to the encoding / decoding of the optical stream? If so wouldn't that eliminate any possible advantage it would have over my current wire based system? I mean wouldn't you have to have tranciever at every point on the optical bus and then have a bunch of sensors and electronics to decode the signal?
HOWEVER if it doesn't, does this mean that there will be random strips on my m-board that will glow from fiberoptic cables passing data back and forth.... I might have to build a clear case if something like that happens!
My feeling is that we are a long ways away from optical computers. Optical computers are envisioned to work in a fundamentally different way than the current manner that photonic systems such as telecommunication systems operate. The way telecommunications systems work right now is that they are electronic systems that are linked by devices that generate photons (a laser), that transmits photons (an optical fiber), and receives photons (a photo-detector). In these cases, the generation and detection of photons is an electron to photon to electron conversion process. When people speak about the prospects for optical computing, they are usually speaking about photons switching photons. This would require light itself to activate an optical switch. Thus, basic logic functions such as an AND gate would have optical inputs and outputs and would not involve an explicit photon-electron-photon conversion as discrete components. That is a lot harder to do. Electrons have charge and mass and they interact in a fundamentally different way than photons can, which have no charge and mass.
You're Just Jealous Because The Voices Are Talking To Me.
Of course it will have a 20 second "Slow Down, Cowboy!" timer. That's how long it will take to boot XP 2002. Linux will, of course, boot in .3 seconds on that motherboard.
t m
From the article: "But it may not take divine intervention to get more mileage out of copper interconnect. Intel claims it can reach speeds of 10 GHz and beyond in five to eight years using copper. "We're confident we can get to 10 GHz. And there's reason to believe we can double that," Pinfold said."
I'd put my money on copper; we're still using
gasoline, when hydrogen-powered cars have
been viable for years.
http://www.auto.com/industry/iwirn22_20010822.h
-wp
information is immaterial
Right now fiberoptics are a little scary for consumer grade appliances. They may look like ordinary wires, but they can shatter when you drop them, and it's impossible to tell. In addition, you have to clean the connectors with a special cleaning cloth (one-time use silk) every time you plug them into a new connector to prevent dust buildup.
So to me the real problem is a cheap fiberoptic motherboard connector that won't have shatter or dust buildup problems. I couldn't find any mention of this in the EETimes article - but then, it's not a real product yet, so how could it have technical challenges yet? (-;
Sure would be nice, though.
Honestly... :)
There is no way this is going to be useful in consumer grade pc's for a long, long time. The only possible use I can see is ultra-high-end servers and graphics boxes that cost >$200K and thats not for another 5 years. Right now, we have a glut of processing power in our pc's. Dual Athlon 1.5ghz? Are you nutz? I'm still amazed by how fast my 1ghz tbird is! We need processers and internal components that are more reliable and do more, not just do the same things faster.
Who the hell needs 10,000fps in quake, anyway...
Why is it so hot? Where am I going? What am I doing in this handbasket?
Also, Time magazine reported last year about this, and they pointed out that the kind of speed offered by fiber is the only real bottleneck to creating a truly self aware computer.
In that case, Time magazine is filled with idiots. Computers will never be self aware as long as they are the glorified calculators they are.
People who talk about self aware computers are usually ignorant of what computers do. They do not do incredible things, they do what they are programmed to do. One cannot program in self awareness. The closest we can get is a convincing emulation of self awareness. if you write a program to print "I think therefore I am" on the screen, the computer doesn't suddenly see any value or meaning in those words, simply a string of Ascii characters. Even with the hal project (remeber that article?), all they accomplished was a sophisticated simulation using years of statistical data of what a self aware organism may say, NOT self awareness itself. the greatest emulation is still an emulation. Science fiction paints a strange picture that powerful computers will eventually become sentient. This is mistaken. Build the worlds fastest and most powerful calculator, and you still have to press 1+1 in order to get it to answer the question -- and the computer will never ask the question...Unless we tell it to.
It's been a long time.
I am sooo tired of the bull-shit scientists on this site with the crackpot proposals that add a minimum of 3 new problems for every one old problem that their "idea" would fix.
:-P
This is the current proposal for the hardware setup, by a man in the know (not me):
"Levi has proposed an "encapsulated processor" concept whereby a CMOS device uses fiber-optic ports as the only connection to external chip sets and DRAM. The processor, which itself could contain two CPUs and cache memory in the core, would integrate a crossbar switch that connects the ports to the processors and cache memory.
The ports, each of which could sustain 40 Gbytes/s of data throughput in each direction, decode and multiplex signals for an optical subassembly containing vertical-cavity surface-emitting lasers (VCSELs), PIN receivers and the fiber interface. There would also be a short, low-power electrical link from the port to the processor, according to Levi's proposal."
Inetellectual response to this idea is what was wanted, not bullshit ideas involving reflecting light off the inside of the case
--chris
But there's more to it than that. Other than CPU upgrades, the problem with a common bus in the past has been that the bus itself is a limiting factor. Think of commonly used buses and other interconnects, whether PCI, SCSI, IDE, the CPU/RAM FSB, etc. Every one of these has gone through multiple iterations of getting faster. Similiary, every time there was an improvement in backplane performance, you'd need to upgrade your backplane. Typically, during such an upgrade, you also want to upgrade other components, like CPU & RAM - so the most efficient way to do this is with a single motherboard that contains it all.
If it were possible to set up a backplane that had humongous speeds that far outstripped anything the components were capable of, the backplane approach might make more sense. Still, something like that sounds expensive, and actually adds complexity to systems from the point of view of manufacturers and even end users.
I just wanted to address two types of comments I've seen posted here:
* Encoding / decoding speeds are done at the speed of the medium. Encoding and decoding optical signals doesn't have any more overhead than PCI or IDE. The spec. writers and endec designers are well aware of these issues. That's why technologies like 10Gb Fibre Channel or Eithernet aren't ready yet -- not because we can't transmit at that speed, but that we can't build an entire NIC to sustain those speeds. (Give us some time: we'll be there soon enough.)
* Serial interfaces like Fibre Channel and Infiniband (and even Gigabit Eithernet) aren't replacing SCSI. They are replacing what you think of as SCSI: the 50 or 68-pin cable in your case. But SCSI is the protocol being used to talk to all those FC & Gig-E storage devices. SCSI over FC is called FCP (see T11's specs for more on FC). For Gig-E, most companies are looking into iSCSI, iFCP or FCIP (SCSI over IP or SCSI over FC over IP) for SAN-to-SAN communications. I forget the name of the spec for SCSI over Infiniband, but it pretty much rips it's ideas from the above specs. (sorry, no links for Gig-E and Infiniband at the moment: start at T10 or The SCSI Trade Association)
BTW, I refer to "serial interfaces" above instead of "optical interfaces" because a lot of this is actually copper. Most likely, Infiniband on the motherboard will be copper and off the motherboard it will be optical. Most of the Fibre Channel equipment I have isn't "fibre" but copper.
However, optical fiber would eliminate interferance, which seems to be the main barrier on clock speed. Heat would likely be reduced also, and cranking up the clock-speed would likely eliminate the latency issues. Not to mention the cool-factor inherent in optical.
What would be really cool would be to replace firewire and USB with fiber. There are hybrid fiber coax systems that could provide whatever power you're mouse/keyboard/etc would need, up to a certain point anyway. It probably wouldn't be enough to power an external drive.
Under capitalism man exploits man. Under communism it's the other way around.
I don't know about this specific instance, but in general, optical is better than electrical because it's not vunerable to electromagnetic interference causing noise on the signal. I also know that modern CPUs and other high speed components generate a lot of noise - ever tried to use an AM radio next to a computer, it's not easy. This noise both consumes power, and makes it difficult to route connections, if they are too close then they will interfere with each other and not work. If the route is too long, then it won't work either. Using optical connections could mean that the layout of boards could be simplified, and therefore use less power.
Somebody much more intelligent than I am (I forget who it was) made the following observation:
When man first tried to fly, we imitated the birds. We made feathery wings, flapped them, and promptly fell. It wasn't until someone (Bernoulli?) figured out the concepts behind flight that we realized that it wasn't the feathered wings that did the job, but the lift they created. Developing the Principles of Flight led to Flying Machines.
In a similar manner, contemporary AI simply imitates the human brain by making loads of calculations. Onve we get to the root principles behind thought itself, then we can make a self-aware artificial doohicky. (Can we even really call it a computer at that point?) Without the Principles of thought, AI's will be intelligent expert systems, but not self-aware.
Geez... Perhaps I should have posted this in the AI story! Anyway, let the (-1 Offtopic)s begin! My karma can take it.
What you need to check is the cache miss rate, I think this can be done in windows under the performance module or whatever. Believe it or not, most people only need 256k of cache and most can get by with 128k, this is why people could get away with celerons even though people laughed at them for their puny cache. Some multimedia apps need more cache (rare), hence the xeon chip. Increasing the cache beyond this point usually is futile since most data which falls outside the range is seemingly placed in a random part of memory, and sram is very expensive. I may be wrong on a couple points but I'm pretty sure the average hit rate looks similiar the graph of log(x)+ e.
Do you have any references for your assertion that the human brain in fact works by computing?
Change is inevitable.
Progress is not.
No - you're completely missing the design.
Imagine you're inside one of these next-gen computers. The bus inside the computer supplies power and low-frequency signalling. Arrayed across the mother board and daughter cards are these next-gen optical IO chips.
Instead of an opaque case these chips have a window transparent to whatever frequency is being used. Wherever on a traditionial chip the circuitry would head off to a lead in this case there's a tiny solid-state laser & adjacent reciever (with some support circuitry.)
Whenever a signal needs to be sent the laser serving as an optical IO point fires. They may differ in frequency, they may use coded pulses of light, however it works they'd be addressable. These picosecond flashes of light illuminate the interior of the PC bathing the other components in varying degrees of brightness.
Whatever other component is being address recieves the signal with it's own optical IO point and acts on it, replying back with it's own coded flash of light.
No line-of-sight is required as long as the primary reflective surfaces in the case have a high enough albedo and sufficient light scattering ability. If you need an anology imagine a bunch of kids flashing signals to each other with flashlights in the woods. Oftentimes one won't see another hidden behind a tree but the light reflecting off nearby bushes reflect the signal.
Some of the proposed benefits:
I don't read ACs: If a post isn't worth so much as a nom de plume to its author then I wont bother either.