Intel Israel heralds (optical) chip breakthrough

[Megadrive1988]

please bare with the seemingly off-topic post, i relate it to consoles later myself. 8)

http://www.haaretz.com/hasen/spages/416440.html

Intel Israel heralds chip breakthrough

By Oded Hermoni



A team of Israeli researchers at Intel has achieved a breakthrough in chip development that promises to change the world of computing and telecommunications within 5 to 10 years.




For the first time, the team succeeded in developing electro-optical chipsets based on silicon wafers capable of converting electronic signals to optic signals within the chip. They have the potential to be mass produced at the same cost as standard electronic chips. Currently, the manufacturing cost of an optical chip (which is not silicon based) runs into hundreds or even thousands of dollars.

According to Intel's assessment, the electro-optic chips developed during the past year and a half at the company's Jerusalem facility will replace the standard electronic chips used for communications between computer components, allowing this communication to be conducted at the speed of light - 10 times the current speed.

"Today, the fast processors operate at speeds of three gigahertz, but their surroundings still work at speeds of hundreds of megahertz and, therefore, don't succeed in exploiting their speeds," explained Amir Elstein, the co-CEO of Intel Israel and director of Intel's Jerusalem facility. "When the chips, the processor and the ports of the computer speak at the same speed, which will be about 10 gigahertz, the computer's capability will be totally different," he added.

The new development will also change the multi-leg appearance of today's chipsets. "There will still be several legs on each chip, but most of the information will be transfered via a single optic opening of one optic port," Elstein said.

An Intel press release explained how the new technology works: "Researchers split a beam of light into two separate beams as it passed through silicon, and then used a novel transistor-like device to hit one beam with an electric charge, inducing a `phase shift.' When the two beams of light are recombined, the phase shift induced between the two arms makes the light exiting the chip go on and off at over one gigahertz (one billion bits of data per second), 50 times faster than previously produced on silicon. This on and off pattern of light can be translated into the 1's and 0's needed to transmit data."

Patrick Gelsinger, senior vice president and chief technology officer at Intel, called this "a significant step toward building optical devices that move data around inside a computer at the speed of light. It is the kind of breakthrough that ripples across an industry over time, enabling other new devices and applications. It could help make the Internet run faster, build much faster high-performance computers and enable high bandwidth applications like ultra-high-definition displays or vision recognition systems."

Elstein said last week that the company has not yet completed planning the production of the new optical devices, but that Intel's Kiryat Gat plant may be involved. "This is the greatest R&D success. There is no need to build new factories - faster chips can be manufactured at lower cost, with the same production infrastructure used in existing facilities. We took a theoretical physical affect and, using existing infrastructure, moved it up to a level that was previously impossible to implement," Elstein added.

and

http://www.theinquirer.net/?article=15387

Looking through the glass, brightly


By INQUIRER staff: Sunday 18 April 2004, 13:49

ISRAELI DAILY Ha'Aretz said that an optical transmission scheme prototyped by CTO Pat Gelsinger at various of its forums in the past year is close to delivering huge frequency hits.
According to the newspaper, Intel Israel technicians have almost mastered the ability to meld optics and electronics meaning that faster computers are within sight of the techies.

The paper quotes Amir Elstein, joint CEO of Intel Israel, as saying that when everything is matched up, switches will be able to operate at 1GHz, leading to vastly faster computational capabilities.

According to an Intel press release, the Internet "will run faster".

You will recall that this was one of the features that Intel said the Pentium III would herald.

I assume IBM has similar projects going in the area of optical chips. of course, SCEI has a deep partnership with IBM on the Cell. I believe SCEI will stay with IBM into the generation after PS3.

do any of you technical minded gamers think that optical chips might be ready in time for PlayStation 4, Xbox 3, Nintendo 6, etc? which would be 6~8 years from now. if so, do you feel that that generation of consoles will be built with optical chips?

I think this coming generation: PS3, X2, N5 will probably push silicon close to its limits. okay well, even if not to the limit of silicon, there might not be room for a massive enough improvement over the PS3, X2, N5 systems using silicon AGAIN with the PS4 generation. see what I am trying to say?

the only way I see getting another huge leap beyond PS3, X2, N5, using silicon chips, is to use many dies, each with many many cores. but that would probably be too expensive for consoles which typically have 2 or 3 major chips because sucessful consoles can't cost more than $300~$400 at launch.

I know we are MANY years away from such radical theoretical things like quantum computing (which i have no understanding of whatsoever)
but things like optical processors seem within reach, within LESS than 10 years.

the way I see it, optical processors are probably the key to achiving things like complete complex real-time raytracing and/or true global illumination or radiosity, and other currently impossible realtime graphics challenges now facing everyone.

 

[pc999]

wow...
I would like to see this in next gen , xb2 ,n5 ...

 

[I.S.T.]

I keep hearing notihng but good coming out of Intel's Isreal unit. I'm falling in love with it.

 

[Megadrive1988]

disclaimer: this is half-serious, half-for-fun

PlayStation / PSone - silicon

PlayStation 2 / PSX / PStwo - silicon

PlayStation Portable - silicon

PlayStation 3 - silicon

PlayStation 4 - silicon or optical

PSP sucessor - silicon or optical

Playstation 5 - optical

Playstation 6 - optical or bio-tech

PlayStation 7 - bio-tech

PlayStation 8 - quantum!

*Deadmeat rolling in his grave*

j/k Deadmeat, I like some of your posts

 

[jvd]

just because they had break throughs doesn't mean it will be ready for prime time anytime soon (i'm talking about the next 10 years ) Then after that it will still be a few years till all the bugs of mass production would be worked out. I would say most likely 2015 or so we can see optical. Remember silicon has a good life spand yet and tons of money has been invested in it. SO you can expect them to milk it for all its worth

 

[Megadrive1988]

that's why I think PlayStation 5 would go optical almost for certain, which could come sometime after 2015 (if PS4 does not). but now that moves it to the more distant future.

 

[cthellis42]

If it stands a good chance of being produced at high volume with low cost, then we'll see it take off. If it doesn't produce much better results than where silicon goes between now and then, it will see limited spreading. If another method closer to conventional is an option and is ALSO high volume with low-enough cost...? We may go there instead.

This is a huge "we'll see."

 

[Farid]

Thanks for the news megadrive,

the last time i heard news about optical processor they were only able to perform as DSP.
So (except if this device/architecture is more complex), it's still almost useless in a gaming environement. No?

According to an Intel press release, the Internet "will run faster".

The Intel PR BS "force", is strong , very strong.

 

[Deadmeat]

...

Personally, I don't believe in optic processors, unless you can put a million optic switches on a chip. This thing will be useful for optic routers, however.

I would place my bet on the recently demoed superconducting CPU from Japanese university. The gate density is acceptable, and you have virtually no heat coming out of it so you can hit 100 Ghz or more..

 

[Farid]

Re: ...

I would place my bet on the recently demoed superconducting CPU from Japanese university. The gate density is acceptable, and you have virtually no heat coming out of it so you can hit 100 Ghz or more..

Do you have a link presenting that technologie deadmeat, it looks indeed interresting, thank you in advance.

 

[Megadrive1988]

^

would also like to see any info on this tech ^__^

 

[Deadmeat]

...

I posted it here. Sonic locked it immediately. He doesn't tolerate anything anti-SCEI and anti-CELL.

 

[Farid]

Re: ...

He doesn't tolerate anything anti-SCEI and anti-CELL.

Nobody does Deadmeat, anti-anything topics are useless. Express your ideas is a fact, and highly respectable, but doing a whole topic and interpreting news so they can enter a "strange anti-SCEI logic", is indeed inane.

Anyway, thanks for the reply Deadmeat.

edit: here's the thread, but there's no link (the picture comes from news.yahoo.co.jp).

http://www.beyond3d.com/forum/viewtopic.php?t=10345

The world's first superconducting CPU. Burns 1/1000th the power of traditional CMOS devices and runs 10x as fast, so it solves the power consumption problem altogether. The downside is that the CPU has to be cooled to -269 C to run. This chip shown has 5000 transistors. The first commerical version is expected in 2010

The original title of your topic was really inappropriate, btw, Deadmeat.

 

[cthellis42]

Actually, the thread was locked by jvd, because--as usual--you take completely disconnected concepts and try to connect them via your well-known paranoia to CELL.

In fact, both the concept and the body text didn't have any connection and didn't even mention CELL... You'd stuck it in your title " Japanese University builds a CELL killer...." which only served to attract flames until the thread was past useless.

Not that it was anything but a look at "cute tech" similar to what we have in this thread. It's small chip (but with enormous comparative landscape, as seen in the picture) that runs fast with little power consumption... because it was cooled to 4-degrees above absolute zero.

Mmm... Laboratory conditions...

Just what the heck such a thing would resemble and be used for by 2010 is anybody's guess. And also as usual, you didn't post any other links or references to let people know what the source of the information was, and if there was any information presented other than the few sentences you quoted.

 

[cthellis42]

...and now that we've finished that quick detour, let's make sure the same crapola doesn't happen in this thread, either. ^_^

 

[JasonLD]

I don't think this thread has anything to do with consoles...

 

[Paul]

Silicon will be with us for a long time, the fact that Intel has heavy investments(and still plans to have) in it still echo's this point.

It's not a matter of us finding a new medium because we cannot hit a certain clock, it's about using the most out of the medium we have. In the near future clock speed will becoming largely irrelevant, but you'll hear more of a "How many cores" instead of a how many Mhz type of tune.

You can also expect to see more efficient designs instead of raw brute force ones, extremely parallel ones, distributed, and so on.

Now seems like the time of era in where companies are going to try different designs to see what works, and everyone will follow.

There are many many eyes on STI with Cell and I am willing to bet there are tons of eyes on this Israel group aswell, along with everyone else(your DNA, quantum) - Though these two are a decade or more off from having any practical consumer applications.

Everyone is looking at the prospects of each and seeing what turns out pretty good and what doesn't, and if something doesn't turn out good they'll think "why" and will see if it's just a matter of money issues, time, design or that the idea (for example extreme parallel) is just a waste of time and they will move to another idea to base things off of.

 

[pc999]

I will bet in whatever gave more profits to them

 

[jvd]

Actually, the thread was locked by jvd, because--as usual--you take completely disconnected concepts and try to connect them via your well-known paranoia to CELL.

In fact, both the concept and the body text didn't have any connection and didn't even mention CELL... You'd stuck it in your title " Japanese University builds a CELL killer...." which only served to attract flames until the thread was past useless.

Not that it was anything but a look at "cute tech" similar to what we have in this thread. It's small chip (but with enormous comparative landscape, as seen in the picture) that runs fast with little power consumption... because it was cooled to 4-degrees above absolute zero.

Mmm... Laboratory conditions...

Just what the heck such a thing would resemble and be used for by 2010 is anybody's guess. And also as usual, you didn't post any other links or references to let people know what the source of the information was, and if there was any information presented other than the few sentences you quoted.

actually the thread was locked because anything deadmeat says is flamed to hell. The topic was bad but it was changed. There was nothing negative about it . Even so its no diffrent then the tons of posts about nintnedo or ms being doomed or sony being king. The reason why it was locked is because the others flaming in it instead of talking about the tech. Notice in this thread everyone can talk about it with no problems. But deadmeat posts it and its flamed.

That is a problem

 

[London Geezer]

Notice in this thread everyone can talk about it with no problems. But deadmeat posts it and its flamed.

That is a problem

Ok, then the problem is not DM, it is EVERYONE ELSE... Wow... :? :|
Enough off-topic talk for today.