IBM, GIT overclock chip to 500GHz

If you were wowed by those Japanese modders who overclocked their stock CPU to 7GHz, wait till you hear about IBM's latest foray into the world of ultra-fast computing: together with researchers from the digital camera-hating Georgia Institute of Technology, Big Blue has managed to overclock a chip to an unheard of 500GHz. Granted, the model they used already had a blistering native clock speed of 350GHz to begin with, but we're still floored that you could actually coax a small silicon wafer into operating at an incredible half-terahertz. As you probably suspected, there's no way to achieve speeds like this at room temperature, so the team froze their high-performance silicon-germanium chip to a super-chilly negative 451-degrees Fahrenheit, which is just eight degrees above absolute zero. Unfortunately, after learning about this breakthrough, electronics giant Sony apparently felt that consumers would no longer be impressed with their much-hyped Cell processor, so they've once again delayed the PS3 until IBM can make the liquid helium-cooled CPUs suitable for mass production . Expect the 500GHz PS3 to hit stores sometime in 2011.

[Thanks, bento-san]

[Guruboy @ Jun 20th 2006 11:00AM]

Come on Blass, do you really need to incorporate your bias against the PS3 into an article? It was really interesting until that completely ruined it.

[shirizaki @ Jun 20th 2006 11:01AM]

I wonder if this will birth the first computer freezer?

Chill your beer while using the internet.

[Corporate @ Jun 20th 2006 11:06AM]

"At 500 gigahertz, the technology is 250 times faster than chips in today's cellphones, which operate at 2 gigahertz"

O_o

2 Ghz cellphones?

[apt @ Jun 20th 2006 11:09AM]

It should be mentioned it was able to run at that 300 GHz speed *without the need for cryogenic freezing* - it ran that at room temperature.

It should also be mentioned that if the newer silicon-germanium tech is used, we could see 1 THz (yes, terahertz) processors running at ROOM TEMPERATURE

I'm curious as to the power consumption of these processors, and when we can get it in a cell phone :p

[Rick @ Jun 20th 2006 11:24AM]

G5 Powerbooks on Tuesday!

[David @ Jun 20th 2006 11:24AM]

As someone actually involved with this work, let me clarify a few misconceptions. This result is for a discrete transistor, not a processor or even a simple circuit. Further, the 500GHz refers to a figure of merit known as fT ("f sub T"), not to a clock speed. Such a high figure of merit might make the device suitable for use in high-speed logic gates operating at a couple of hundred GHz (although a complete processor would be slower). Such a processor wouldn't be that dense, however, since power dissipation with bipolar transistors is high compared with CMOS. More commonly, these devices would be used either in RF applications (which is the source of mentioning 2GHz cell phones ... the reference was to carrier frequency, not digital processor speed) or in high-speed serial links including transceivers for fiber-optic links.

[Eric @ Jun 20th 2006 11:30AM]

Georgia Tech is Never referred to as GIT, only GT.

[John Laur @ Jun 20th 2006 11:32AM]

I was just going to jump in and say this is not a CPU but looks like you have it straight from the horse's mouth.

It looks like the blogs are just going to regurgitate each other this time around. So much for reading!

[Chris @ Jun 20th 2006 11:33AM]

Thanks David. Very interesting indeed and definitely cleared up the rather ambiguous and vague article in the NYTimes.

I guess we can still expect such developments in silicon based components to increase the speed of modern day processors...eventually.

[Brian Baigorri @ Jun 20th 2006 11:53AM]

David, I'm curious as to your liquid helium setup. How long were you able to run this at those temps? And wheres the vacuum? My experience with nitrogen and helium always involves at least a few hours of vacuuming.

[James Baker @ Jun 20th 2006 11:54AM]

Oooooo everyone lets be all pedantic about the article. 'Georgia Tech is Never referred to as GIT, only GT.' Blaaa blaaa blaaa, no one cares. It's a bit of fun!!

[Grant @ Jun 20th 2006 12:22PM]

I've become convinced that Blass is actually Devin Faraci from CHUD.com.

Both of them take otherwise well-written articles, and proceed to completely cock them up with lame, jokey biases.

And it's tough for me to say that -- I'm no Sony-backer. Seriously, you'd do well to keep the unrelated rhetoric to a minimum.

[David @ Jun 20th 2006 1:03PM]

"David, I'm curious as to your liquid helium setup..."

Can't answer that directly. My involvement was with the development of the prior version of this transistor (the 200+ GHz version ... at room temperature ... that we first reported around 2002). While I've tracked how my colleagues have advanced the transistor to the present generation, I'm no longer directly involved except in measuring the RF noise properties. Also, only the device development along with measurements in the room temperature range (say, in the 0-125C range) were performed at IBM. The cryo testing was GT's part of the collaboration.

[Josh! @ Jun 20th 2006 1:03PM]

I agree with Grant. You've become completely too biased on this website. While the occasional joke is occasionally funny, I think that you might just be overdoing it. I could be wrong, though.

And I really hope that any of you don't tell me, "If you don't like it, don't read it!" That may be the lamest comeback that you could possibly use. My reasoning is that the site wasn't always this biased, and it's much more enjoyable when it's neutral as opposed to bashing the hell out of anything and everything.

The jokes are seldomly funny, so stop it!

[nub @ Jun 20th 2006 1:05PM]

i disagree... rhetoric in all forms in entertaining.

[Atanas Boev @ Jun 20th 2006 1:12PM]

In my neighbourhood if one shouts freeze everybody runs like hell not just the CPUs :P

[Tom @ Jun 20th 2006 1:13PM]

"I'm curious as to the power consumption of these processors, and when we can get it in a cell phone"~apt

Put that processor in the Sidekick 3, and I guarantee Paris Hilton will say "That's Hawt"... shortly before bursting into flames.

[Michael @ Jun 20th 2006 1:24PM]

What can you do with this type of power? I wonder how fast windows would start up with that kind of speed. Probably too many bottlenecks from other hardware. Well at least when other things catch up we know the processor side is taken care of.

[Jitty @ Jun 20th 2006 1:27PM]

I think the 500ghz ps3 your talking about will actually be the PS4. ;)

[Geir E @ Jun 20th 2006 1:39PM]

You wonder how fast you could boot windows on this? My good man, this is the minimum requirement to boot Vista at all.

[TC @ Jun 20th 2006 2:14PM]

'David, I'm curious as to your liquid helium setup. How long were you able to run this at those temps? And wheres the vacuum? My experience with nitrogen and helium always involves at least a few hours of vacuuming.'

Yeah, I too like to get the chores done before I mess around with cold gases too....

[treycranson @ Jun 20th 2006 2:33PM]

So 500GHz and still no cure for cancer?

[Octavus @ Jun 20th 2006 3:28PM]

I might be a little partial because I am going to UIUC, but here they were able to get a transistor to run at 604ghz over a year ago.
http://www.newscientist.com/article.ns?id=dn7253

[David @ Jun 20th 2006 3:48PM]

"I might be a little partial because I am going to UIUC, but here they were able to get a transistor to run at 604ghz over a year ago.
http://www.newscientist.com/article.ns?id=dn7253"

You missed that the report claims a record for a silicon transistor (actually, a SiGe transistor, but the SiGe film is grown on Si and is fully compatible with Si processing). The report you quote is for an InP/InGaAs device. Such so-called "III-V" devices have always been faster than Si-based devices at any given point in time, but are much, much more costly to produce and can't be integrated nearly as densely due to much poorer yield.

You can produce a chip with mostly standard CMOS and several tens of thousands of SiGe bipolar transistors through into some critical performance paths for extra speed. You can't integrate an InP/InGaAs device into such a chip at all.

[GT fan @ Jun 20th 2006 4:43PM]

Its GT not GIT.

Georgia Institute of Technology is abbreviated GT. Ive never even seen it as GIT before.

[Kibiyama @ Jun 20th 2006 5:56PM]

Okay, we get it. You don't like Sony. That's fine, keep it to yourself next time and don't ruin a perfectly interesting article.

There are plenty of people that aren't happy with the PS3 as it is right now, myself included, but I don't toss it into everything I say.

By the way, check your site traffic after posting one of these articles. I generally stop coming to engadget for a week or so after I see one of these unnecessary comments, and I'm sure a few other readers do, too.

[Exbzurg @ Jun 20th 2006 10:07PM]

@kibiyama
dude relax, you'll luve longer...
seriously its just a joke man. a it was a smalll one. At least its was vaguely related to the article, IBM is helping with the cell processor. but still it was just a joke.
blogs joke all the time so no big deal.
you dig?

[Exbzurg @ Jun 20th 2006 10:10PM]

arg meant to say live longer. stupid keys, they're like all next to each other.

[mike @ Jun 21st 2006 1:11AM]

It's not just a joke to a reporter. The comment was almost completely unrelated to the article being reported. It seemed quite random. Showing personal bias against any person, company, idea, or anything ruins your professional credibility, regardless of whether readers like myself love or hate Sony. It's fine to make jokes in your articles, in fact, that's one reason I love this site, but it was a bad joke that did more to show you have a bias against Sony than actually elicit humor. Readers can't trust reporters who so blatantly show their personal bias in a professional article. Some probably did find the comment funny, but it did more to damage your credibility than anything else. I'll have to keep and eye on the name of the author of whatever Engadget article I read now- not all of Engadget's writers have the same high level of credibility, it would seem.

[MattUK @ Jun 21st 2006 3:28AM]

Cool....literaly.

500GHz processor....should run Windows Vista OK, and they propably be realised at the same time!!!

[Shermin @ Jun 21st 2006 4:39AM]

what the? everybody, chill out!
It was only a joke - and a good one at that.
i'm still laughing. especially at all the uptight loosers who are getting their knickers in a knot.
heh heh heh

[Pirateinmymind @ Jun 21st 2006 11:49AM]

Like many are saying that PS3 thing was a joke. It's somewhat related to the article, and actually in my opinion shows no biase against Sony. They are taking forever on the PS3, to joke that they are going to to delay again is just that a joke. Not Sony-hating writing in hopes of starting wars.

[jamminman @ Jun 21st 2006 4:51PM]

holy crap guys... it's just a joke, why do you care. this is engadget... a BLOG... not CNN.

[Daniel @ Jun 22nd 2006 12:38AM]

Hey at least this blog's still less biased than CBS news.

[FormlessCloud @ Oct 4th 2006 1:05PM]

Wow... What was the age old saying, sticks and stones will break my bones but words will never hurt ( I said this to someone who was standing next to quite a few stones, seriously... I've never again spoken this after that experiance).

At the end of the day I fail to see how those words could have affected anyone, come on! are your shares in Sony falling at the speed of a raging meteroite because of what was written?

> Haha... didn't like that did you? > What? Of course im have a laugh.

This reminds of my early days when I was loyal to SEGA, could never speak bad of them infront of me, I can remember writing them a e-mail saying how I would become one of the richest men in the world just so I could help them with their dying console the Dreamcast (may it rest in peace). I eventually sold the console because I couldn't bear to keep it any longer especially when I found out that the same processors were going to be used in washing machines.

Well... I certainly hope they have a moderation facility and block this comment, I dont even know what I was trying to acheive, looks like it failed whatever it was. Should stop typing now but its sort of relaxing getting all these words of my chest, kind of makes up for the depressing day I had at work.

They wanted to cancel, They were in a contract... I told them but they didn't listen, They wanted to speak to a Manager, but there was none.... WHAT WAS I SUPPOSED TO DO?!?

Seriously though If you've just read everything I have just typed you and me need some long sessions of counseling, If you have any suggestions, feel free to share them.