View Full Version : when will new chip technology hit?
drguitarman
06-20-06, 04:20 AM
anyone here with any inside information on when new chip technology might go into production? eg diamond wafers, optical based chips, trigate transistors, reversible computing, simpe quantum computers, any other cool new tech? is anyone trying to make 3D chips? Anything to break past the possible barriers from leakage and so forth with current sillicon tech.
Redeemed
06-20-06, 06:02 AM
anyone here with any inside information on when new chip technology might go into production? eg diamond wafers, optical based chips, trigate transistors, reversible computing, simpe quantum computers, any other cool new tech? is anyone trying to make 3D chips? Anything to break past the possible barriers from leakage and so forth with current sillicon tech.
I'm no engineer, nor an expert on this subject- but I bet most he information you are after would be "Top Secret", and those who knew about it would be under NDA or something similar.
Your best bet might be to do a google search and just see what comes up.
Demon_82
06-20-06, 05:00 PM
IBM just ended testing some Si-Ge transistors wich are said to reach 300GHz on ambient temperature and 500GHz at 0ºC, so I think it can delay any other news by just expanding the reign of the silicon chips.
when will new chip technology hit?
december 18th 2009...no one really knows, not even the companies themselves. anyways they usually take bits and pieces of the new tech and find ways of incorperting it into there current product lines. no offese but this is realy a silly question considering the many techinical hurdles that need to be over come with new designs
IBM just ended testing some Si-Ge transistors wich are said to reach 300GHz on ambient temperature and 500GHz at 0ºC, so I think it can delay any other news by just expanding the reign of the silicon chips.
yea ive heard of some Si-Ge chips being designed. sounds incredible.
IBM just ended testing some Si-Ge transistors wich are said to reach 300GHz on ambient temperature and 500GHz at 0ºC, so I think it can delay any other news by just expanding the reign of the silicon chips.
not to criticize, but a small correction to your post: the chip ran at 500GHz at absolute zero which is something like -451 degrees celcius not 0 degrees celcius. though the 300GHz was at room temp, both are still amazing.
not to criticize, but a small correction to your post: the chip ran at 500GHz at absolute zero which is something like -451 degrees celcius not 0 degrees celcius. though the 300GHz was at room temp, both are still amazing.
ummm just to correct you. .i dont think anyone has reach absolute zero to this day. last i heard they have gotten very very close but never quite there.
ummm just to correct you. .i dont think anyone has reach absolute zero to this day. last i heard they have gotten very very close but never quite there.
sorry, i guess i should have said near absolute zero, i was just trying to point out the large difference between 0 degrees celcius and -451 degrees celcius or so. wasn't trying to ruffle any feathers
Edit: my bad, i remembered it wrong it's -459 degrees Farenheit, not celcius. -273 celcius as Red_shift pointed out. Thanks Red_shift. I knew it was negative four hundred fifty something on one of the temp scales, lol(damn memory!!!).
yea ive heard of some Si-Ge chips being designed. sounds incredible.
The real question is, how long before Jakup can get his hands on it?? :D
The real question is, how long before Jakup can get his hands on it?? :D
haha I can see JaKuP being a borg by 2010.:D
haha I can see JaKuP being a borg by 2010.:D
:rofl:
BWUAHAHAHAHA.
Greatest post about jakup. Ever.
Red_Shift
06-21-06, 11:53 PM
Originally Posted by retsam
ummm just to correct you. .i dont think anyone has reach absolute zero to this day. last i heard they have gotten very very close but never quite there.
sorry, i guess i should have said near absolute zero, i was just trying to point out the large difference between 0 degrees celcius and -451 degrees celcius or so. wasn't trying to ruffle any feathers
Actually 0 Kelvin is -273.15º Celsius. For a chip to run near 0 K it needs a massive cooler to keep the temperature down since ambient temperature is at ~300k and the chip might also produce significative heat if it's not a superconductor, if it is, it still has to be protected from ambient temperature.
300GHz @ ambient temperature seems impossible for now, so does 500GHz @ 0º Celsius. the chip running at such frequencies would generate alot of heat unless it is a superconductor but there are no superconductors at such temps.
anyone here with any inside information on when new chip technology might go into production? eg diamond wafers, optical based chips, trigate transistors, reversible computing, simpe quantum computers, any other cool new tech? is anyone trying to make 3D chips? Anything to break past the possible barriers from leakage and so forth with current sillicon tech.
Intel is currently developing trigate transistors along with some new tech, in 2009 intel will produce cpu's wth trigate; read this article : http://www.dailytech.com/article.aspx?newsid=2649
CaptNKILL
06-22-06, 01:24 AM
haha I can see JaKuP being a borg by 2010.:D
LMFAO!! :rofl
drguitarman
06-22-06, 01:39 AM
december 18th 2009...no one really knows, not even the companies themselves. anyways they usually take bits and pieces of the new tech and find ways of incorperting it into there current product lines. no offese but this is realy a silly question considering the many techinical hurdles that need to be over come with new designs
Actually I didnt want exact dates (as you said, this would be silly) I just want whatever estimates people have, if they have any. I've already got an interesting answer on those si-ge chips, looking them up now
myshkinbob
06-27-06, 01:32 PM
point of interest, the SiGe technology is only useful for analog signal processing, eg. radio frequencies, and not for digital signals that a computer works by. So SiGe means nothing to the cpu industry, and germanium is already used in strained silicon wafers by intel. :)
source:
http://arstechnica.com/news.ars/post/20060622-7117.html
TimeOut
06-27-06, 01:55 PM
In few years from now when new processors are released based on the Hewlett Packard's Crossbar Nano-technology patented last year.
http://www.hp.com/hpinfo/newsroom/press/2005/050201a.html
Soylent
06-30-06, 02:02 PM
point of interest, the SiGe technology is only useful for analog signal processing, eg. radio frequencies.
IBM's fastest process before this switched at something like 200 GHz, and that's been available since 2001.
Diamond based processors were also aimed at networking and signal processing AFAIK. And their power density at high frequency was absolutely attrocious and completely unacceptable to anything but tiny, tiny ASICs.
Researchers at Nippon Telegraph and Telephone Corp. (NTT) have set their sights on a transistor that runs at 200GHz while dissipating 30W/mm², and they believe impurities are the only thing holding them back.
30 W per mm²? That's 3 KW for a 100 mm². Which is more like 4-5 KW at the power outlet. It never was meant for normal, monolithic, processors.
And quantum computers don't make much sense for normal problems. Their charm comes from the possibility of transforming non-polynomial time problems for classical computers into polynomial time for a quantum computer.
Such as finding the factors of a product of two huge prime numbers, simulating QM systems efficiently, search algorithms and making truly secure communication possible(as well as breaking most/all of current cryptographic hashes).
ntel Corporation researchers today disclosed they have developed new technology designed to enable next era in energy-efficient performance. Intel's research and development involving new types of transistors has resulted in further development of a tri-gate (3-D) transistor for high-volume manufacturing. Since these transistors greatly improve performance and energy efficiency Intel expects tri-gate technology could become the basic building block for future microprocessors sometime beyond the 45nm process technology node.
Sounds to me like trigate transistors is slated for some time early next decade.
As for carbon nano-tubes, no-one has figured out a way to massproduce chips with them efficiently. Sure you can move them around one by one in tiny experimental systems but that's far, far removed from useful implementation.
On chip optical interconnects built using current silicon tech are surely useful, but it's mainly for mitigating propagation delays in the very thin copper wires of modern chips and allow further miniaturization and lower latency communication. They're intended use is to keep current silicon tech working for a while longer rather than making a radical break with it to some entirely new tech.
In few years from now when new processors are released based on the Hewlett Packard's Crossbar Nano-technology patented last year.
http://www.hp.com/hpinfo/newsroom/press/2005/050201a.html
The step from building a simple component to figuring out a way to mass-produce them cheaply, reliably and building efficient tools for developing chips using the process is as far removed from building a bottle rocket to landing a man on the moon.
30 W per mm²? That's 3 KW for a 100 mm². Which is more like 4-5 KW at the power outlet. It never was meant for normal, monolithic, processors.
The section you quoted stated it dissipated heat at 30W per mm², not draws that amount of power.
Soylent
06-30-06, 05:47 PM
The section you quoted stated it dissipated heat at 30W per mm², not draws that amount of power.
One implies the other even if you take it to mean that it is _capable_ of dissipating 30W per mm². It's no use having something capable of dissipating that much heat if it doesn't draw remotely near that much power.
If you take it literally, like I did, it means their target is to have it dissipate 30W per mm², which it can't do unless it consumes 30W per mm² of power. No matter what cooling you use, dissipating more heat than you generate means cooling the device. And you can't keep that going for any extended period of time. In real systems your cooling will be less efficient the smaller the difference in temperature between the device and cooling medium and you'll asymptotically reach an equilibrium point where you're removing as much heat as is generated. Which you're taking to be much lower than 30W per mm²; why then would you let it get extremely hot before you start cooling it, just because you can?.
Si-Ge technology is nothing new. It's widely used in mobile communication equipment. Germanium it is employed for producing heterojunction bipolar transistors or as a strain-inducing layer for MOS transistors.
Over the years, several alternative semiconductors have been proposed. Fact is that Silicon is still the best overall compromise.
Other semiconductors like Gallium Arsenide (I think a GaAs transistor currently holds the crown for the device with the highest cutoff-frequency) or Sapphire have their fields of use but they won't replace Silicon. At best they will be integrated into current silicon technology like happened to Germanium (histhorically, the first semiconductor devices were based on Germanium cristals).
mike686
07-01-06, 05:45 PM
haha I can see JaKuP being a borg by 2010.:D
It would only be a matter of time before he would overclock himself. :D
Capt. Picard
07-02-06, 07:24 AM
Nalugrl would soon follow. Then this forum. Then us. Because ...
Resistence is futile.
It would only be a matter of time before he would overclock himself. :D
Did someone say Quad-jAkUp?
lol. Now that's a scary thought :p
|MaguS|
07-02-06, 11:45 AM
Did someone say Quad-jAkUp?
lol. Now that's a scary thought :p
Imagine all the spam on the forum...
Capt. Picard
07-02-06, 03:06 PM
Imagine what Nalugrl would say if Jakup "overclocked" himself for better "performance".
:naughty:
vBulletin® v3.7.1, Copyright ©2000-2012, Jelsoft Enterprises Ltd.