IBM's 'frozen chip' claims speed record

[D_o_S]

IBM Corp. and the Georgia Institute of Technology Tuesday claimed they have broken the silicon speed record, thanks in part to a "frozen chip."
IBM and Georgia Tech claimed that they have demonstrated the first silicon-based chip capable of operating at frequencies above 500 GHz by cryogenically "freezing" the circuit to minus 451 degrees Fahrenheit (4.5 Kelvin).

The experiments, conducted jointly by IBM and Georgia Tech, are part of a project to explore the ultimate speed limits of silicon germanium (SiGe) devices, which are said to operate faster at cold temperatures.

The chips used in the research are from a prototype fourth-generation SiGe technology fabricated by IBM on 200-mm wafers. At room temperature, the circuits operated at approximately 350 GHz.

Source: EE Times (http://www.eetimes.com/news/semi/showArticle.jhtml?articleID=189500692)

[RickyG512]

so why dont we have cpus at 350 ghz

[cmberry20]

Because AMD dont want to produce CPUs faster than 3Ghz. :laugh:

[Jimmy 2004]

I would love to see how much I could overclock my CPU when it was cryogenically frozen. Still - 350GHz. Would you love to have on of those dual core!

[K2Storm]

What the hell can you do with a 350 GHz Processor ......... :rolleyes:

Anyway ............ You still will need to wait for all the other hardware to catch up ....... Ram, Video Cards .... :slap:

[Polaris573]

so why dont we have cpus at 350 ghz

Because this technology is a very recent breakthrough. It takes years for technology to go from reasearch and development into commercial use.

[Alec§taar]

Ithe first silicon-based chip capable of operating at frequencies above 500 GHz by cryogenically "freezing" the circuit to minus 451 degrees Fahrenheit (4.5 Kelvin).

"Fahrenheit -451"

(For some reason, that 'rings a bell' here...)

APK

[newtekie1]

Yes, lets go burn some books!

[Alec§taar]

Yes, lets go burn some books!

Ah, well... since it is in the NEGATIVE direction? Let's go write some!

:)

(The length of most of my posts ought to make a good "war & peace" sized one... lol!)

APK

[BigD6997]

hummm ibm needs to come back to pcs and make cpus, cuz i want my cpu runnning at 350ghz

[POGE]

Its probably very simple with like 100 transisters...

[Azn Tr14dZ]

Does anyone know what's after GHz? First MHz, then GHz, now what?

[Thermopylae_480]

Ray Bradbury's latest New York Times #1 Bestsellar, Fahrenheit -451: The Temperature at which Books Freeze. Buy it today in paperback, only from Del Rey.

Oh, by the way. Something is messed up with EE Times temperature calculations. -451~F is equal to -268~C. Since absolute zero is 0 Kelvin or -273~C that would be impressive. Not only would that mean that IBM had broken the CPU record, but also been the first to reach absolute zero, defied physics by preforming work at absolute zero, and then actually dipped below absolute zero, to a temperature of -5 K on a temperature scale that begins at zero. IBM should be rather proud today.

4.5 K is pretty low, but if they actually decreased the chip to those temperatures, the correct numbers would be more around, -277~C and -466~F.

I e-mailed the author, maybe he'll fix it?

Terahertz comes after Gigahertz.

For everyone who ever wanted to know the progression it is:

kilo
mega
giga
tera
peta
exa
zetta
yotta
xona
weka
vunda

If you're concerned about farther than that well...

[Homeless]

I think I would be scared to death if they ran a benchmark on that

[newtekie1]

Oh, by the way. Something is messed up with EE Times temperature calculations. -451~F is equal to -268~C. Since absolute zero is 0 Kelvin or -273~C that would be impressive. Not only would that mean that IBM had broken the CPU record, but also been the first to reach absolute zero, defied physics by preforming work at absolute zero, and then actually dipped below absolute zero, to a temperature of -5 K on a temperature scale that begins at zero. IBM should be rather proud today.

4.5 K is pretty low, but if they actually decreased the chip to those temperatures, the correct numbers would be more around, -277~C and -466~F.

I e-mailed the author, maybe he'll fix it?


Since when is -268C colder then -273C? They did not hit absolute zero, they reached 5 Kelvin, or 5 degrees over absolute zero. Since Kelvin is just C-273, the units are actually the same so: -273C+5C=-268C

[Sasqui]

Oh, by the way. Something is messed up with EE Times temperature calculations. -451~F is equal to -268~C. Since absolute zero is 0 Kelvin or -273~C that would be impressive. Not only would that mean that IBM had broken the CPU record, but also been the first to reach absolute zero, defied physics by preforming work at absolute zero, and then actually dipped below absolute zero, to a temperature of -5 K on a temperature scale that begins at zero. IBM should be rather proud today.

Hey good info, but confused about what you're saying... -268 > -273 ,so they didn't go below absolute zero. Am I missing something? Cheers.

[Thermopylae_480]

Ah, damn minus sign bunked my calculation. I feel like an idiot. What the hell was I thinking. Oh well. The original numbers are right. :nutkick: Sorry guys.

[OneCool]

Does anyone know what's after GHz? First MHz, then GHz, now what?

THz = Terahertz (I think)


Im how much power a 350ghz CPU would pull!! :respect:

[Sandman]

Its probably very simple with like 100 transisters...
Exactly, even at 500GHz, this thing prolly couldn't boot a lightweight linux distro. If this technology were practical enough to be used in a desktop environment, you'd already be seeing ads and headlines.

[XooM]

its a very very simple communication I/O chip and it's clock frequency is nothing spectacular. Intel had a 1 thz transistor in 2001, shortly followed by a 3.33 thz transistor from AMD. however, SiGe chips liking cold for clocking bodes well for overclocking AMD's next line of chips which will be using SiGe technology.
This chip isn't a CPU in any way, shape, or form. no shit it couldn't boot a linux distro, it doesn't have registers or ADD operators or... anything. its an IO chip ffs!

[FLY3R]

What the hell can you do with a 350 GHz Processor ......... :rolleyes:

Anyway ............ You still will need to wait for all the other hardware to catch up ....... Ram, Video Cards .... :slap:


Lots and Lots of Folding!!!

[BigD6997]

"i just bought the new 1vundahert cpu"- said by someone 150 years from now, vundahert sounds realy realy weird

[i_am_mustang_man]

Lots and Lots of Folding!!!
palabra!

[KennyT772]

yeah it was a very simple micro processor. but just imagine what cpus could do if they ran at say 10ghz at room temp....

[GeneralDodo]

its a paperweight!

[magibeg]

I honestly think i'ld settle for my home computer to be running at around say.... 200ghz at room temp with peak temps hitting 25C. And i want as much performance per mhz as a conroe or better. It would calculte pi to the 32 million before i even had a chance to hit the button!

[KennyT772]

by the time that happens we will be in the age of quantum computing. there will be no monitors, no keyboards, no video cards...there will just be a small box (run by its own power station of course) with a cord we plug into our visual cortex

[Jimmy 2004]

Any Kenny wakes up from his dream of the future...

[saumesh]

If such a high speed processor is developed then there must also be motherboards which support this type of chip.
I don't think that the present motharboards will be comptable with this processor, because such a high speed will rise the problem of cooling.

[saumesh]

The data transfer rate for optical fiber is much greater than that of the copper wires.
All the data Bus in the computer system are copper wired ,so this limits the data transfer rate maximum upto 10Gbps.
:toast: If optical fiber is also used for data transfer betwwen various components of system then, ofcourse, this can increase data tranfer rate anywhere between 30 to 60Gbps.
:nutkick:

[gerrynicol]

I thought this was only 1 transistor circuit, and not a full CPU??

[Dippyskoodlez]

so why dont we have cpus at 350 ghz


Because you people dont quite understand the difference between "microprocessor" and "CPU".....;)

I thought this was only 1 transistor circuit, and not a full CPU??


This would be incredibly important news if it was a full CPU.

In no way shape or form is it even close to a modern CPU..


Perhaps you could look at it like this.. Its a section of the cpu ;) Intel had a 10Ghz "microprocessor" about 5 years ago... and now we are at 500ghz.. good improvement I'd say :)