Slashdot Mirror
Cooling Down Hot Processors
DonnaMai writes "Face it: the only scorching hot thing you want with a chip is salsa. Any other overheating is potentially counterproductive, and can be downright damaging to the microprocessor -- or other components. This article uncovers potential ways to chill the chips."
When I was a kid we had to cool our chips by using our little brothers as a heatsink.
What we really need is a spare, low-power, mimimal processor without all the fancy extensions that you can switch to when you're just, say, reading a webpage or email, or such.. you could integrate this right into the motherboard and completely shut down your processor when you're not using it for real stuff. IMHO... maybe an engineer will give me a reason this is unreasonable.
Best way to cool your processor is to move to Canada. Hands down.
Ice cubes work well.
They don't last very long, though.
Perhaps we should be working on a better ice cube!
"Rocky Rococo, at your cervix!"
Dr. Trevor Mudge (U. Michigan) came to give a lecture at my University last year. He had an interesting proposal which I suspect is probably going to end up being used in nearly every architecture. The energy usage of a procesor is proportion to the square of the voltage - so dropping it as much as possible is desirable. The only problem is that once you get too close, you start getting bit level errors. He proposes to use a shadow register to keep track of values as they pass through and detect bit errors automatically, and route around them. If run at the optimal voltage (1.4 volts) a razored process will see a dramatic drop in energy consumption with a virtually-nonexistant hit to processing power.
To make laws that man cannot, and will not obey, serves to bring all law into contempt.
--E.C. Stanton
I don't know what the problem is, I like hot chi... oh wait, you're talking about chips, nevermind...
hack a day
"Fieththa".
Nope, didn't work. Roof is still tender.
Karnal
How do I cool processors? Simple: I underclock them. Even a 10-20% less MHzs is usually enough to get rid of a noisy fan, i.e. the most stupid idea in the history of personal computers. Most of today's computers are I/O-bound anyway (Moore's law) so there is no performance loss whatsoever. Seems like an obvious solution.
Sincerely,
Pan Tarhei Hosé, PhD.
"Homo sum et cogito ergo odi profanum vulgus et libido."
There was nothing new or innovative in the article, and it had the depth of Paris Hilton as far as actual real world cooling suggestions.
/. I forgot for a moment...
There are a ton of different solutions out there both onchip and off including aircooling via different heatsink designs, watercooling, peltier cooling, and self contained refrigeration units.
This article barely scraped the surface of anything useful or interesting related to cooling.
Oh wait, this is
When I first got my Prescott chip, it ran *way* too hot. Realized that the stock thermal pad was just acting as insulation, so I scraped it off and replaced it with Ceramique. It still ran warm, so I superglued a piece of 3" PVC pipe to my case fan. Now air blows right onto the processor area, and the CPU temps are great. I highly recommend the ducting. Cheap, easy, and oh-so-geeky.
Here it is in the "mostly finished" stage:
Picture 1
Picture 2
DAMN YOU OCTODOG! DAMN YOU TO HELL!
Color this mechanical engineer disappointed.
Software piracy is victimless theft.
Someone needs to figure out an efficient way of makeing use of the huge surface area on the lid of a laptop for cooling. When in operation, it's facing away from you, so you wouldn't feel all the heat from it. The problem is tranferring the heat to a part that has to hinge away from the area that's making the heat. Plus there might be problems if it transfers too much heat to the LCD screen rather than to the air on the surface away from the user. It just seems a shame not to be able to take advantage of all that surface area.
It would --- but there would be other problems.
The first one is the most simple: silicon's expensive. Really expensive. The more units you can slice off that wafer the cheaper the units are. Making the die bigger simply for thermal reasons isn't going to wash with the chip manufacturers. They already glue the die to a metal backing plate, which gives you much the same effect anyway.
The second one, however, is the most crucial one. Electricity is slow. Electrical impulses travel at about 2/3 c through copper and a touch less through silicon (IIRC, I can't find the figures to check). This means that the bigger your die is, the longer it takes the impulses to travel from one side of it to the other.
A 1GHz clock fires every 10^-9 seconds; since the speed of light is 3x10^8 m/s, this means that the impulses are going to travel about twenty centimetres between clock pulses. For a 4GHz clock, it'll be about 5cm. There's a lot more wiring than that folded up inside the die; and it gets worse --- particular things happen at particular times throughout the clock cycle, and where you are in the clock cycle now depends on how long the wire is that connects you to the clock. Making sure everything happens in sync is a nightmare.
There are solutions to all of this; asynchronous designs which don't use clocks, offloading functionality to special-purpose processors like GPUs so you don't need as fast a main processor, radically different approaches like Cell, optical transports so you can route signals through each other, etc, but basically there are loads of good reasons why you need the die to be as small as humanly possible.