What ever happened to laser cooling? Get a pure substance to heat up and use out-of phase laser light to slow the atomic thermal-vibration and cool it down.Kinda like how noise cancellation works.
This is a well known problem of quantum tunneling associated with process scaling. One solution is to change the structure of the transistor from the classic MOSFET to a FinFET transistor where the gate wraps around the channel. This creates a gate that can block the leaking current across the channel. Of course, FinFETs are still in their experimental stages, but are seen as most likely to replace our current MOSFET technology.
100* C? Holy s**t. You've got some kind of problem - that's not normal for any processor. Thermal paste misapplied maybe?And, the new Core processors are great on heat and power. Don't diss intel, they've done right by completely abandoning new P4 stuff, and rightly so.
This tech might filter down to desktops too, but they're trying to solve the problems of cooling datacenters and main frames, somehow I doubt they were too concerned about making new ways to cool G5's
Is it "smart" to just throw the heat way? Heat is energy after all, why not recycle it? There are many technologies available that have the potential of making this possible, from piezoelectic materials to micro and/or acoustic sterling engines.
100c sounds like a misreading-- That said, my Prescott 3.0 ran at 75c load during a Connecticut winter, with a lapped thermal take heat sink and perfectly applied arctic silver, I sent it in and they said it was "faulty" and sent me a 3.2 Northwood, runs at 45c load on the same heat sink. IBM is an intelligent company, there is a lot of money in finding the right cooling solution and throwing a patent on it.
jasonpriniMay 29, 2006
What ever happened to laser cooling? Get a pure substance to heat up and use out-of phase laser light to slow the atomic thermal-vibration and cool it down.Kinda like how noise cancellation works.
metadffMay 29, 2006
This is a well known problem of quantum tunneling associated with process scaling. One solution is to change the structure of the transistor from the classic MOSFET to a FinFET transistor where the gate wraps around the channel. This creates a gate that can block the leaking current across the channel. Of course, FinFETs are still in their experimental stages, but are seen as most likely to replace our current MOSFET technology.
tovelingMay 29, 2006
100* C? Holy s**t. You've got some kind of problem - that's not normal for any processor. Thermal paste misapplied maybe?And, the new Core processors are great on heat and power. Don't diss intel, they've done right by completely abandoning new P4 stuff, and rightly so.
ninjathisMay 29, 2006
TL:DR
alienzMay 29, 2006
This tech might filter down to desktops too, but they're trying to solve the problems of cooling datacenters and main frames, somehow I doubt they were too concerned about making new ways to cool G5's
transfireMay 29, 2006
Is it "smart" to just throw the heat way? Heat is energy after all, why not recycle it? There are many technologies available that have the potential of making this possible, from piezoelectic materials to micro and/or acoustic sterling engines.
linkin4May 29, 2006
100c sounds like a misreading-- That said, my Prescott 3.0 ran at 75c load during a Connecticut winter, with a lapped thermal take heat sink and perfectly applied arctic silver, I sent it in and they said it was "faulty" and sent me a 3.2 Northwood, runs at 45c load on the same heat sink. IBM is an intelligent company, there is a lot of money in finding the right cooling solution and throwing a patent on it.