Intel Researchers Create Tri-Gate Transistors

[MacRumors]

According to Macworld, Intel's researchers have created tri-gate transistors that better insulate circuits, which could enable either a 45% increase in speed or a 35% reduction in power used when compared to today's processors. This will help Intel to extend Moore's law, which states that that number of transistors on a chip would double about every two years, or in layman's terms, chip processing power will approximately double every two years.

Chip manufacturers have had difficulty extending Moore's law as chip geometry shrinks below 90 nanometers and frequencys escalate to beyond 2 Ghz. In both cases, chips begin to leak more electricity and run less efficiently. One solution is to build multi-core chips, which the industry as a whole has already adopted.

However, different chip manufacturers are playing with other methods of further increasing chip efficiency below 90 nm. IBM is placing research money into carbon nanotubes, and has made some impressive breakthroughs.

For its part, Intel is placing its bets on tri-gate transistors.

"Compared to carbon nanotubes, it is far easier to build," [Mike Mayberry, director of components research and VP of Intel's Technology and Manufacturing Group] said. "The problem with carbon nanotubes is that no one knows how to put them in a particular spot except by moving them one at a time. Even our smallest chips have millions of transistors, so that is an insurmountable challenge. [...] [Tri-gate transistors] will be an option for chips somewhere beyond 45 nm -- in the 32 or 22 nm mode -- so that gives us confidence we can continue scaling Moore's Law into the next decade," said Mayberry.

Macworld states that Intel could begin using the technology by 2010.

 

[PlaceofDis]

i can't profess that i have much knowledge on what all of this means, but hey i guess it means there won't be any stagnation. which is good.

 

[BornAgainMac]

Too much info for me. Seems like a great Woz level topic.

 

[zv470]

Does Tri-Gate mean we will finally move from Binary Computing to Ternary Computing?

Switch on... switch is between... switch off... etc...?
Yes... maybe... no... etc...?
White... gray... black... etc...?


1012010010202102010002222202012202020200120021

 

[ImNoSuperMan]

Oh no.... We`ll be gettings chips which are 45% faster at 35%less power in 5 years???? So I guess everone on MR will start advocatng WAIT TILL YOU GET TRI GATE CHIPS. THEY`RE GONNA SMOKE THE CURRENT LINEUP. WAIT WAIT WAIT. Dont buy a Mac before we have this Chip in it. It`s just a wait of 5 yers. Cant you hold off for 5 small years???


LOL
This a tribute to all those posters on MR who`ll always tell you to "WAIT TILL THE NEXT REV. IT`LL BE SO MUCH BETTER without even thinking bout the frustration building up with each day of waiting. But the real prolem is that they are correct most of the times. .

 

[bigandy]

Why does it always seem to me like intel are taking the easy way out?

I think Carbon Nanotubes will be the way to go... But with IBM running that research, will it ever go mainstream?

 

[DavidCar]

Here's some photos of trigates:

http://news.com.com/2300-1006_3-6082265-1.html

 

[weitzner]

Why does it always seem to me like intel are taking the easy way out?

I think Carbon Nanotubes will be the way to go... But with IBM running that research, will it ever go mainstream?

IBM will figure out how to do it, then intel will steal it, perfect it and put it everywhere. it doesn't really matter who comes up with this stuff, if it's clearly the best solution, it will be adopted by everyone.

 

[~Shard~]

Too much info for me. Seems like a great Woz level topic.

I'm an Electronics Systems Engineer, so I love gobbling this kind of stuff up!

This is a very cool advancement from Intel, however as the article states, we probably won't be seeing this technology in the marketplace for another 4-5 years. Plus, it's kind of a no-brainer to say, "Technology is going to be much better in the future!" Duh.

Personally, I believe carbon nanotube technology is the way to go and offers far more potential, but I'll have to read up more on these tri-gate transistors to be fair.

 

[Abstract]

Carbon nanotubes are the way to go. Everyone knows that. Everyone.



(I have no idea what I'm talking about. Never listen to me, even if you think I'm making perfect sense)

 

[kainjow]

Apple already made nanotubes

 

[thejadedmonkey]

Good. This means the current lineup (once it goes merom) won't be outdated for a while.

 

[chibianh]

Apple already made nanotubes

but they're not carbon

 

[Max on Macs]

I'm all for carbon nanotubes to be honest. It'd be great to bring the PC into the carbon cycle

 

[Yamson]

but they're not carbon

Well, technically they are carbon-based.

 

[Timepass]

IBM will figure out how to do it, then intel will steal it, perfect it and put it everywhere. it doesn't really matter who comes up with this stuff, if it's clearly the best solution, it will be adopted by everyone.

not so much as steal. Depending on the agreement the makers have ibm may just give it to them. There is quite a bit of reverse engineering in the CPU market.

Intel first 64 bit chips where very clearly an revevise engineering AMD64 chip. Plus there is a lot of sharing in the industity. AMD could of put Hyper threading in there CPU and there been no law suit for it because of the agreement AMD and intel have where they share a lot of the stuff and are required to share quite a bit because everything need to be compatible with the software. Though intel may have more heavy restrition put on them by the goverment because they are the power house that much I dont know. But I do know there is a lot of infomation sharing in the industry

The computer industry as a whole is a great example of engineering at it best.
Things that the average person doesnt understand or know is the question an engineering ask when solving a problem.
1. what am I trying to make/do.
2. Has anyone solve a problem like this before.
3. if so what was there solution.
4. How can I take that solution and adapt it to my problem or make it better

They go though that before coming up with a new solution because it is faster and cheaper to adapt a working solution already.

An engineering solution is normally not always the best possible solution to the problem but it is a working pratical solution to the problem. To get the best solution will take to much time and money to get to it. An engineering solution takes a lot less time and money to figure out. It is a very good answer and solution but rarelly is it the best possible solution and it always a work in progress on how to make it fast cheaper and better.

 

[CallmeKenneth]

So is tri-gate a bit like triple core?

 

[Collin973]

Interesting stuff. New technology is always very interesting to read about. I wonder how soon this kind of equipment will make it consumers.

 

[gauriemma]

Here's some photos of trigates:

http://news.com.com/2300-1006_3-6082265-1.html

Yeah, "trigates" are cool. But I'm holding out for one of these:
http://www.vargr.com/pages/renders/images/StarGate.jpg

 

[Macmaniac]

but they're not carbon

But they contain carbon since they are made out of OIL!

 

[thejadedmonkey]

I have this feeling that the tri-gate stuff will be out first, and really give us a nice speed boost. A few years after that, IBM will have their carbon nanotubes ready and intel will adopt a nanotube approach, possibly combining the two, but giving us an even greater speed boost.

And by "us", I mean the world.

 

[bmarker]

From what I understand ...

This does not mean a move from binary. I'm guessing the wink means it wasn't really a serious question but it's one I hear often enough.

It's tri-gate because there are 3 paths from source to sink. There are 3 gates in the same amount of space a single gate usually occupies. This is why it can pass more power without leaking or take less voltage to activate. There are also dual gate designs but Intel is working on 3. It's still all on-off.

Of course, I could always be wrong.

 

[thogs_cave]

Personally, I believe carbon nanotube technology is the way to go and offers far more potential, but I'll have to read up more on these tri-gate transistors to be fair.

Given the amount of research into carbon nanotubes in all sorts of applications (like construction[0]), I suspect that long-term it will be the more viable technology. Intel's comments say to me: "Yeah, carbon nanotubes are cool, but it'll be hard to use them." Now it's hard, but in five years, who knows?

[0] For example, the Shimizu Mega-City Pyramid. See:
http://en.wikipedia.org/wiki/Shimizu_Mega-City_Pyramid

 

[mozmac]

Apple, Intel, and Nintendo should all merge. Then they could incorporate the new "Tri-force" technology into their chips. Macs would play Zelda music as they start up. Then, whenever you discovered something new about your computer, it would play the "Secret Discovered" noise that we've all come to love. You know what I'm talking about. "Do do do do do do do do do do!"

 

[SiliconAddict]

Tri-gate will prob win. From what I've read its the easiest and cheapest to implement. Nanotubes will eventually get here but until an easier way to use the tech is found its going to be a back burner technology. that being said it could simply be a matter of finding an innovative way to manipulate nanotubes. Something that could become apparent overnight.