UIUC Creates World's Fastest Transistor Again

An anonymous reader writes "The University of Illinois has developed (again) the world's fastest transistor operating at over 500 GHz. They used an indium phosphide based wafer, and super-scaled dimensions. The device kind of looks like a spaceship." Milton Feng, the professor in charge of the team behind the transistor, admits that their ultimate goal is a terahertz transistor, which given their previous achievements, doesn't sound too lofty.

Re:DARPA

[localghost]

DARPA funds a lot of scientific research. This is a good thing. It doesn't neccessarily affect them directly, but advancements such as this will likely benefit everyone, so it's worth it for them to put money into.

Also, it isn't a chip, it's a single transistor.

500GHz?!! I'll change my job!

[WARM3CH]

When I started designing hardware circuits, the world was much more beautiful. You could understand everything that your small micro-processor based system did, downto the function of the BJTs in the TTL devices down there... Then Intel started the 1GHz race and I had to learn a great deal of RF techniques to just design my next PCB. And now 500GHz?!!! At this rate, a few years later I'll have to learn more about RF and then eventually optics than next hot FSM synthesis algorithm! I guess I'd better change my job, start something more calm and steady, like paiting or ...

Don't get your hopes up...

[Escaflowne]

Too bad current Computer Technology doesn't use indium phosphide and indium gallium arsenide. It would take years for fabs just to adjust to a new material and yield decently.

Also as someone stated, it's just one transistor not the hundreds of millions that are in current technology (all acting in "harmony").

Then again, this is a great discovery and a step in the right direction. I'm very proud of my Alma Mater. Too bad I didn't have a class with Professor Feng.

Are you ready for lots of latency?

[G4from128k]

At 1 THz, it will take more than 40 clock cycles for a signal to move across a 1/2 inch die of the CPU. And it will take 320 clock cycles for a round-trip to a memory location just 2 inches away. (And that is assuming the signals travel at the speed of light in a vacuum, not the slower speed found in metal traces or optical fibers.) Should make it interesting for chip designers.

don't worry about it too much

[lingqi]

it's a 509GHz *TRANSISTOR*, not a chip. even for the transistors on a P4, they also operate at a "speed" much faster than the actual chip operations - after all, to squeeze 3+ GHz out of a chip, which has tons of gates connected one after another, isn't exactly a "everybody switch at once" deal.

besides, for real high speed stuff people are moving toward serial on PCB anyway, parallel just doesn't work anymore past a certain point due to the increased capacitance that's caused by traces getting tighter with eachother (need more traces for more pins)...

Almost all (i'd wager to say "all" but there might be some tiny companies i don't know about) FPGA manufactures include serdes (serializer / deserializer) ports on their chips, usually more than one - those go at 6+GHz (faster ones due out are 10GHz), but PCB still handles that because it's only a few pins compared to, a DDR bus.

But it's still planar

[freidog]

which means (even if they produce a FET version) it's still going to have the terrible electrical characteristics we see in today's transistors. Lots of bleeding and heat in the off state. I'd much rather see people focusing on something like Intel's trigate [electronicstalk.com] transistor. While current transistors can handle and 8 or 10 ghz CPU, nothing will dissipate the KWatt or so the chip would dissipate.....

Re:Improvement rate

[igny]

If you plot those 3 points on a plane you will see that the dependence is not linear. I tried to fit a curve through those points and got that

y=3000/x^0.4

where x is size (nm), y is speed (GHz). 1000GHz will be reached at ~15nm.

Re:don't worry about it too much

[lingqi]

hmm... if you use differential pair and bury it between layers of power / gnd planes (and surround it by the same), it's not AS bad...

yes yes I know it's still a pain, but I don't think it's the end of world as people seem to make it sound like; tis all.

Re:Read the link you linked to. Mod parent down.

[almadenmike]

You can check out Gordon Moore's original paper via this Intel site -- http://www.intel.com/research/silicon/mooreslaw.ht m -- which says Moore's Law refers to "an exponential growth in the number of transistors per integrated circuit..." The notable chart in the paper itself has on the vertical axis: Log (base 2) "of the number of components per integrated function."

Re:Faster and faster

[Anonymous Coward]

Faster = necessairly smaller. Smaller = more can fit in a tight space.

Do the math.

Still anticipating the optronics revolution

[BlueCoder]

This is good for academic study just to see what can be achieved but the industry should be focused on nano engineering and laying the foundations for optronic design. Processors are just too hot and power hungry; it's a dead end. Time to move into the 21st century with optical circircuitry.

I have little doubt that an equivelent optical pentium processor, or any other processor of choice, could be created now for a big chunk of change that would be 10 to 100 times more powerful at least, using at quarter of power though probably requiring the space of cabinet. The equivelent of old solid state computers. (I gaurantee you that at least the NSA and multitary have this already but their development rarely contributes to the commercial sector since they like to keep technology to themselves.) The commercial commutity should have already done this and have started refining the technology to reduce it to the size of a standard cpu case and be ready to release a product within a year. With such a new technology breakthroughs would happen daily yet anything produced would be more powerful while requiring less power.

Industry is behind where they should be because they are wasting time further developting lithography and smaller transistors. Optronics is a slam dunk and far more deserving to have the money thrown at it that is currently being spent pushing the limits of electronics.

It's too bad ...

[Anonymous Coward]

... in some ways, because most of the really high-speed transistors are BJTs. Since BJTs leak power constantly (not just when switching, as does CMOS), their application to entire chips is limited.

They are still useful in very small, critical, high-speed portions of chips, so that's great. But unless we can reach these speeds with CMOS (or some other kind of technology), then we're going nowhere anytime soon.

Re:Improvement rate

[Anonymous Coward]

It's only three points. You can draw an awful lot of curves through three points.