Slashdot Mirror
Frozen Chip from IBM hits 500 GHz
sideshow2004 writes "EETimes is reporting this morning that IBM and Georiga Tech have demonstrated a 500 GHz Silicon-germanium (SiGe) chip, operating at 4.5 Kelvins. The 'frozen chip' was fabricated by IBM on 200mm wafers, and, at room temperature, the circuits operated at approximately 350 GHz."
TFA wasn't clear... I assume this wasn't running a larger fully synchronized CPU with memory and multi-level cache at 500GHz, but is instead running a smaller number of transistors at that speed?
I can understand your concern. However, after IBM backs this up, it forces me to do more research (which, I haven't finsihed yet obviously).
Don't waste time... procrastinate now!
350 * 1024 * 1024 * 1024 (375 809 638 400) cycles per second divided by the distance light travels in a second (299 792 458 000 mm / s) is 1.2 mm. Just thought I'd throw that in.
Religion is what happens when nature strikes and groupthink goes wrong.
Since these temperatures only occurs naturally in space, why not build a super, big cluster of these things, hook them up to a satallite and launch it into orbit.
Wait a minute! These things can run at 350GHz (o_o) at ROOM TEMPERATURE! The fancy cooling is all nice and good, and 500GHz... well congratulations!
But heck, I'll take one even without the cooling.
You are right, sort of. This could be useful for some very specialized processors that are very simple but need to do these simple operations very fast.
A CPU like the one we use now in PCs can't go much higher than 10GHz simply because, at light speed, an electron wouldn't have enough time to make it through the long circuit paths before the next clock cycle.
Yes, space applications necessitate the need for radiation hardened processors. Due to the nature of the hardening process (special design rules, triple modular redundancy, etc.), these chips dissipate much more heat and operate at lower clock frequencies than the original processor they were based on. But these chips are still made at state-of-the-art facilities; they may not use the smallest processes out there, but technology like SOI (silicon-on-insulator) dramatically improves the radiation-tolerance of the process.
But, NASA is interested in using newer commercial-of-the-shelf processors (IBM PPC based) in order to increase the amount of processing power available on satellites. As an example, take a look at this project that I previously worked on. (Hopefully it will fly sometime in 2009)
Simple calculation: speed of light is 3*10^8 m/s, frequency is 5*10^11, so in one clock you can travel 3/5 millimeter at the speed of light. But electrical signals in copper and semiconductors travel at approx. two third of the speed of light, so in one clock an electrical pulse can travel roughly 0.4 mm. Your processor has to be way smaller than this, because all routes signals can take from anywhere to anywhere must be shorter than this distance. And let's forget entirely about phase-problems, synchronization, ... These things are now already causing difficulties in chip-design, at current speeds where signals can travel several centimeters.
I honestly do not expect that processor speeds will increase very much anymore. The past however has time after time proven everybody wrong that made that statement.
int main(void) {while(1) fork(); return 0;}
The terahertz band is the holy grail of telecommunications because at such high frequencies, we can begin to test communications based on gravitational waves instead of electromagnetic waves. This represents significant progress toward that end. Gravitational waves potentially are not limited to the speed of light, which might pave the way to real-time satellite communications with no lag or communications with operations on other planets with significantly shorter wait times.
Actually, if I recall correctly, they're translational energy states, not vibrational ones... Vibrational states are much higher energy. But perhaps it's been too long since p-chem.
Actually they can't even do 1GHz at light speed. But that's why we have pipelining, and current generation have between 10-20 pipeline steps..
Next gen with 20+ like the Pentium IV have however already flopped.
In theory you could have a 100GHz Pentium V with 100 pipeline stages. The problem is really that it most likely wouldn't be faster than a 2GHz Pentium M.
Pah! My flashlight runs at 750000 Ghz (7.5 x 10^14 Hz). Its portable, has a 12 hour battery life, lets me see in the dark AND sports a durable andonized aluminum casing.
Beat that IBM.
Nothing sucks like a Vax, nothing blows like a PowerMac G4