Slashdot Mirror
Super-fast Transistors On the Way
nbannerman writes "The BBC is reporting about a new kind of transistor, that recently set a world record of 110Ghz. From the article: 'To achieve the speed gain, researchers at the University of Southampton added fluorine to the silicon devices. The technique uses existing silicon manufacturing technology meaning it should be quick and easy to deploy.' The apparent applications for this process include mobile phones and digital cameras."
Maybe we should just get faster software.
Now remind me why exactly we need 110GhZ moblie phone processors?
int cents = 0;
cents += 2;
newbies will use this technology to grab the first post in every thread on ./?
added fluorine to the silicon devices
Not only will the transistors be faster, but whiter and shinier, they won't need to floss.
Rhymes that keep their secrets will unfold behind the clouds.There upon the rainbow is the answer to a neverending story
"Now remind me why exactly we need 110GhZ moblie phone processors?"
So the future will get here faster.
...might taint your precious bodily fluids.
...so this means that Flight Simulator X will run at 10fps instead of 5?
Paleotechnologist and connoisseur of pretty shiny things.
TFA didn't mention what effect this process has on CMOS devices, which is probably more relevant since this is what is used in most digital design these days (and a lot of analog as well). Bipolar devices take up more area and tend to consume a lot more power, among other things. But if this speeds up MOSFETs, then they're really on to something...
Back in the days of 486 I couldn't decode fullscreen MPEG1. MPEG2 couldn't be done at all because the processor wasn't fast enough. Are there any compresison codecs that we could use with a ridiculously fast phone CPU (or home computer, actually?)
Remember that when a CMOS gate is switching the current flowing through it increases. The faster the gate is able to switch, the less power will be used in the state change. Now the processor doesn't have to run at anywhere near that speed, but the fast transistor switch will minimize the power per cycle.
This transistor is for modulating the radiowawes, not for driving your java-games.
send + more == money?
The article says they did this with bipolar transistors. I recall from my intro electronics class that most integrated circuits are CMOS (built with field effect transistors) because in general they are faster and use less power than equivalent TTL circuits (built with bipolar transistors). If this is true, does this new process make TTL chips more attractive for (at least some) applications?
The vacancies are clearly left when the transistor dentists lose their jobs. The clusters are an obvious reference to the dentists all going by the name Beowulf.
It's a small world and it smells funny; I'd buy another if it wasn't for the money; Take back what I paid (SoM)
Tell me, what digital device would not benefit from shorter switching times?
:wq
This sounds like a plot to sap our vitality by adding Flourine to impurify our sacred bodily fluids ^H^H^H^H^H transistors.
Oh, for those who have never seen it, the silly reference is from Dr. Strangeglove.
Precisely! We agree completely.
Signed,
SACFTF
Slashdot Anonymous Cowards For The Future
-----EQUAL REPRESENTATION-----
Listen you "big A-C" Anonymous Cowards, we're against the future. Technology ruined our lives, remember when trolls used to live under bridges? Now we live in basements or apartments. We should go backwards, not forwards!
Signed,
SACATF
Slashdot Anonymous Cowards Against The Future.
Ripper: A foreign substance is introduced into our precious bodily fluids without the knowledge of the individual, and certainly without any choice. That's the way your hard core commie works. Mandrake: Jack... Jack, listen, tell me, ah... when did you first become, well, develop this theory. Ripper: Well, I ah, I-I first became aware of it, Mandrake, during the physical act of love. From Dr. Strangelove
-=- Many seek good nights and lose good days.
I believe this technique would speed up MOSFETs as well because they are saying that the added fluorine doesn't allow the boron to diffuse into the silcon as much. This means you'll have a cleaner line between the p-type and n-type dopped regions. In terms of MOSFETs you could inject the flourine under the gate so when you dope the silicon to create the source and drain you won't have overlap you normaly get under the gate. This means you could reduce the gate to drain and gate to source capacitances which kills the high frequencies.
Sure, it might be the fastest silicon BJT, but as TFA alludes to, there are InGaAs HBTs that are functionally equivalent to BJTs and have cutoff frequencies of 710 GHz. Specifically, I'm talking about the one discussed in this paper by Milton Feng's group at the University of Illinois.
The article is very less on details.
Anybody can explain on the process here?
rajmohan_h@yahoo.com
Maybe OT but so true.. Just came home from work, someone asked me to write a test program for our product. So - 1h with Delphi ( works in Wine perfectly and compiles in Kylix for Linux people ) and I had a test performance program emulating 5000+ ( any number IP addresses.. ) terminals through raw sockets ( you build the headers.. ) creating tunneled test traffic and keeping statistics/timings/random test patterns/,, Now - the product doesn't do much except route the requests / replies, etc.. --- My 2.4GHz dual system is ~2% busy when the product running in a decent size IBM / AIX system tops 100%, 110tx / sec ( not bad BUT..) So - it is the sad truth, we do need faster programs !! It is not difficult, just KISS ! But unfortunately I have heard that faster HW solves the problem, what!!, 35+ years in this busines and never seen it - still waiting..
Now, of course as so many have said, this technology is ( at least ) not yet valid for faster computers but for digital processing ( big difference. ) Anyway - I think that the SW skills have somehow gone "a little" down lately - just IMHO.. And, of course, this is OT..
Why would the prime purpose of this be cameras and cell phones, rather then computers.
Klingon Software is not released, it escapes, inflicting terrible damage onto the enemy as it does
At 100 GHz the wavelength is 3 cm. A quarter wave line would be 0.75 cm. This thing is operating at a frequency well above that at which it is easy/feasible to use a printed circuit board. To operate at this frequency I would have to spend a whole pile of money so I could use hybrid IC techniques. Or I could figure out how to couple this device to waveguide. AARGH!
The magic word Slashdot asks me to type to prove that I'm not a robot is 'hospital'. How very appropriate 'cause that's where I would end up if I tried to use this sucker.
RTFA, this is talking about bipolar trannies, not CMOS.
We have transistors that can operate at 100GHz - we have for a while. They're called BJTs. Only problem is it would cost waaaaaay more to manufacture a 50M transistor chip than other semiconductor designs. You never know, but really I am highly doubtful of the cost of this.
http://en.wikipedia.org/wiki/W_band
Apparently this freq is known as W-band microwave radiation. Useful for millimetre wave radar apparently - you know the machine that can see through clothes at the airport? Wonder if this is why they are researching it? Anyone know better than me? Will my kids have real X-ray specs?
spoonerize "magic trackpad"
Hey, it is "standard" "silicon" process, but they compete with other GaAs/InP/SiGe bipolar transistors, not yoru garden variety CMOS FETs -- and for other technologies Ft of more than 100GHz is not unheard of. Neat trick, and you will see them in your cellphone front-end, maybe soon, but do not hold your breath for 20GHz processors (and if someone makes 'em, please *do not hold them with your bare hands*! -- they gonna be HOT!) ;-)
Paul B.
I'm curious to see what kind of heat these transistors would generate, as thats a growing problem already.
because if the CPU is fast enough then the software just won't have any time to fail, because it will be very occupied trying to not fall behind the processor. All of the software failures are due to high stress that software is experiencing and the faster the computers are the less time there is for the software to stress, thus it doesn't think about being bloated with all that dead weight and terrible algorythms that much and doesn't get depressed that easily.
You can't handle the truth.
I'd like to speculate about some uses for these new goodies....
1. Faster Cash Registers (Wal-mart, are you listening???)
2. Faster ATM's (though I admit it's usually the dufus in front of me & not the machine that's the problem)
3. Faster calculators (for all you toe counters out there)
4. Faster gas pumps (so we can fund the terrorists we're fighting more effectively)
5. Faster Coffee makers (I don't know if making the programmable stuff faster will actually make the brew faster, but one can only hope)
2 cents,
QueenB
HDGary secures my bank
This here's my BIPOLAR TRANSISTOR. I design analog circuits with it and it's got an Ft of 110 GHz.
I don't design digital circuits with bipolar devices. I design digital circuits with CMOS devices. Bipolar sucks power but it runs fast. CMOS sips power but it run's slower.
And if I'm going to design anything usefull with it, that thing is going to operate at about 1/10th of the cut-off frequency (Ft).
This ain't about 110 GHz CPUs.
This is about Op Amps and Phase Lock Loops.
I know that Circuits 101 was a long time ago for some of you folks, but really.
Flourine is about as nasty a substance as you'll find anywhere. It's used in oil refineries. Even a slight exposure to the gas will cause all the calcium to leech out of your bones and you'll die screaming in agony as you very literally turn to a jellyfish. Not to worry though, the process only takes 4-6 hours.
You REALLY don't want to live anywhere near an oil refinery. Seriously.
...who reads something like this, and hearkens back to the days when people thought stuff like this would lead to some sort of golden age, and then to hear that it will simply make "better phones and cameras" is kind of disappointing. I know the prior attitude was a lot of idealistic pie-in-the-sky, flying-car, jet-pack, white building, monorail nonsense. Still though. I miss it. Oh... bring back Donald Duck with his doors to the future, and the nuclear powered airplane. Please? For just a moment?
For all intensive purposes, "whom" is no longer a word. That begs the question, "who cares"?
"Or I could figure out how to couple this device to waveguide. AARGH!"
I work in satellite communications and couple waveguides all the time.*
*Alright now! Get your mind out of the gutter.
--
"The magic word Slashdot asks me to type to prove that I'm not a robot is 'hospital'."
Mine's "unproven".
"Same thing with IBM's big SiGe push - great for RF but doesn't seem to have made any inroads to logic, probably due to cost issues and technical problems that make SiGe potentially unsuitable for logic but don't really affect their RF performance."
IBM SiGe is a BiCMOS process. This means it integrates SiGe bipolar transistors with standard CMOS. The bipolar component of the technology is not intended for logic, except perhaps for low-density, very-high-speed logic in limited patches. Rather, the bipolar is intended as an analog/RF device, with the integrated CMOS handling the logic. Ideal applications include chips that combine an RF or high-speed analog front end with a digital baseband, which is why the technology has made great inroads into cellular front ends, WiFi, GPS, optical transceivers and test instrumentation circuits (i.e. front ends for high-speed oscilloscopes). The reason you don't see SiGe making huge logic inroads is two-fold: (a) the bipolar doesn't do enough in such applications to justify the cost of the extra processing steps needed to add the bipolar and (b) since the technology leverages existing CMOS baseline processes, the CMOS is typically a generation behind the most state-of-the-art CMOS only process. In summary, SiGe is best suited to chips that require integrating very high performance RF/analog with logic but where the logic doesn't need to be the very fastest.
That said, SiGe bipolar transistors hit the 110GHz (referring to an RF metric known as f-sub-T) point about 5 years ago and it's not clear that these devices are really that much cheaper (much of the cost of SiGe comes from other add-ons such as resistors, capacitors and the like, required to make complete RF circuits, rather than from the growth of the SiGe layer).
Q: What does the "B." in Benoit B. Mandelbrot stand for? A: Benoit B. Mandelbrot
Just a few weeks ago there was an article about IBM in conjunction with Georgia Tech, supercooled reaching 500ghz, room temp was at about 300ghz.
Hos is this new one a world record at 110ghz?
As part of my CS education I had a course in DSDM or the Dynamic Systems Development Method. The triangle consisted of money, time, features. Two corners were fixed (usually money and time) and with this triangle we could make the customer see that the feature part can't also be fixed. In comes the MoSCoW treatment, etc, etc.
Who reads BBC news for scientific discovery?
_ Record_fT.pdf
Summary:
http://eprints.ecs.soton.ac.uk/12112/
pdf:
http://eprints.ecs.soton.ac.uk/12112/01/2006_Kham
"Fix it"
That's what we need in our landfills. Flourine. It goes will with the other hazardous chemicals found in modern electronic devices. Maybe in the future, you will have to pay to throw away an old pda like we do today with desktop machines.
Get a sense of proportion. There's very little fluorine in these new transistors, and it's chemically bound to the silicon. There's more fluorine in toothpaste, and a lot more fluorine in common refrigerants.
Contribute to civilization: ari.aynrand.org/donate
Maybe the BBC could get a technical editor, with the minimum amount of knowledge necessary to match the caption of the picture with the photograph, because I can't see how a bare printed circuit board has millions of transistors.
Programmer: There is no speed issue that cannot be solved through the use of better hardware.
Engineer: There is no hardware issue that cannot be solved through the use of well-written software.
Discuss.
It's not plugged in?
Oh, say does that Star-Spangled Banner entwine / The myrtle of Venus with Bacchus's vine?
- Eliminate all products that are completely unusable
- Favor the buggiest and most poorly implemented but still usable product
I am not just talking about Microsoft. If I studied the history more closely, there might be an underlying principle like "first to market" that would explain this apparent stupidity. On the money making side, it seems to be a huge advantage to have a *usable* product out first. If you misjudge your corner cutting, and it isn't actually usable, you lose. But presenting a polished product later, even at a lower price, seems to be doomed.The problem for Microsoft is that while they succeeding in emancipating themselves from the chains of DOS and Win16, Win32 is still mired in bad design decisions of Win95 - which live on in the name of backward compatibility. The "shattering Windows" attack exemplifies the problem. If I were them, I would create a new Win64 API that at least doesn't repeat the same mistakes as Win32, and run Win32 applications in a virtual machine in their new OS. The compatibility VM can have efficient multimedia by providing pass through drivers a la Xen and Win4Lin.
The transistors fabricated in the article are bipolar, not CMOS. Bipolar transistors don't have gates, they have a bases.
Anyway, back to CMOS. Assuming equal voltage (which isn't true as voltages are decreasing with smaller processes), if the gate can be modeled as a capacitor, the power consumed would be larger for a higher frequency since voltage would be the same, but current would increase due to the lower impedance of a capacitor at a higher frequency. You can think of it as having to pull of the same amount of charge, but at a faster rate, thereby requiring more current. I believe the decreased power typically comes from the fact that the faster transistors are usually created with a smaller process and therefore use a lower voltage.
A pity we can make 110 Ghz transistors and we can't solve the hard drive bottleneck.
I AM MAN!!!
Hmmm... so then they are 'Commie' transistors...?! I jest I jest :)
Everyone knows that the only reason you add fluorine to anything is for mind control! Soon, my tin-foil hat will need a anti-fluorine upgrade
Why bother with 110Mhz when you can get 1-3Thz:P
http://www.rochester.edu/news/show.php?id=2585
Mil spec Galliuim Arsenide, but still.
Fluorine doped silicon should be allot less hassle to manufacture and a butt load more eco friendly to boot assuming it all works as reported.
I have a sense of proportion. The isn't that much arsenic either, but millions of chips together might be a problem, INABCL.
But why is the hardware CPU chip itself STABLE, and FAST, and released on SCHEDULE?
Maybe software engineers should do some basic hardware engineering to learn the processes of making good stuff tm
So use stable/schedule software guys to write 90% of your software, and use a few smarty pants guys to write
the core speed deamon code for the specifics that require it.
Liberty freedom are no1, not dicks in suits.
But still funny. Making fun of the misfortunes of others is still great material.
... they feed that toostpaste only to Anonymous Cowards ;)
--- I am known for the ones who want to find me on the net. Is that a privacy risk or a privilege? One might wonder..