IBM says they’ve developed speedier transistor -- 100x faster

[Acert93]

Device more than 100 times faster than best PC chips sold today

The transistor achieved a speed of 500 gigahertz, which is more than 100 times speedier than the fastest PC chips sold today, and about 250 times faster than the typical mobile telephone chip, Meyerson said.

That speed was hit only when IBM researchers, working with counterparts from the Georgia Institute of Technology, cooled the transistor to near absolute zero, but Meyerson said the device still ran at 300 gigahertz at room temperature.

Clay Ryder, president of Sageza Group, a technology market research firm, said the breakthrough should lead to faster processors, but ones that will run far below the top speed demonstrated by IBM.

"We can build a (race car) that can go 240 miles per hour, but is that what you're going to drive to work? No, but you learn things that you can put in mass-produced cars," Ryder said.

As neat as it sounds, but the comments below make me think this is limited, at least right now:

Most improvements in chip speeds over the years have come from shrinking the size of transistors, but IBM's approach is to tweak the silicon on the atomic level, meaning that transistors can be designed from the ground up with very specific applications in mind.

"That means you can have Babe Ruth-style scenarios where you step up and point the bat to left field and nail a shot there," Meyerson said.

But hey, I would be pretty content to see a "mere" 10x increase in frequency over a 5 year period on chips with similar architectures to what we have today.

 

[pc999]

Cool, at least it make us have more hope in silicon.

 

[standing ovation]

This sounds like bug spray for the multi-core colony. :neutral:

 

[Zaphod]

The transistor achieved a speed of 500 gigahertz, which is more than 100 times speedier than the fastest PC chips sold today, and about 250 times faster than the typical mobile telephone chip, Meyerson said.

Cellphones run at 2GHz now?! Nokia N91 is 312MHz, no? Methinks someone is missing a zero.

 

[Frank]

Sounds good, but they would need a better way to get that precision than is available with current and near-future wafer steppers.

 

[Basic]

Cellphones run at 2GHz now?!

Not the CPU, but part of the RF does. Maybe that's what he meant.

 

[Zaphod]

Not the CPU, but part of the RF does. Maybe that's what he meant.

Right forgot about that. GSM frequencies are up to 1990 MHz, I think, so that's close enough. In the context it was mentioned it still sounds weird, though.

 

[soylent]

The fastest silicon-based transistor before this was IBM's 210GHz SiGe heterojunction bipolar transistor from 2001. That's been used almost exclusively for high-speed optical and wireless communication.

Why? I'd guess power demands would be absolutely heinous and routing of clock signals a complete pain in the ass if you tried to do a big 'ol monolithic slice of the stuff like you'd need in a CPU.

But hey, I would be pretty content to see a "mere" 10x increase in frequency over a 5 year period on chips with similar architectures to what we have today.

I don't think this will have any effect on frequency scaling in regular CPUs since it's a specialty transistor that appears completely aimed towards simple tasks such as sampling and de-multiplexing an optical signal down to frequencies 'normal' asics can handle.