PDA

View Full Version : Stability Myth or Fact?


boorishid
11-23-09, 12:45 PM
The cooler the cpu the less voltage you need to get it stable? I just set up a two swiftech triple rads and a bonnie heater core in the same cpu loop only with two pumps, hoping to break 4ghz daily clock.

thideras
11-23-09, 12:49 PM
Yes, the lower the temperature, the more stable the CPU becomes and it may take less voltage to stay stable at those speeds. Why do you think people cool with dry ice and liquid nitrogen? ;)

It also depends on how much of a temperature drop you see. If it is 1-2c, you probably won't notice a difference. If it is 20c+ (or 100c+ in the case of extreme cooling), there is a difference. No way I could cool my QX9650 at 5.288GHz @ 1.6v on air or water.

oldbrave
11-23-09, 03:32 PM
No matter how many rads you put on a system or in a single loop, you CANNOT cool below the ambient air temp without some from of phase change cooling

hokiealumnus
11-23-09, 03:43 PM
Even if you get very, very close to ambient (see Spawn's crazy external triple-rad setup), it's doubtful you'd see much of a decrease in Vcore requirement, if you saw any at all. Stability only improves when you get cold, not just warm instead of hot.

doz
11-23-09, 03:57 PM
Id agree that any kind of cooling that uses ambient air to control it is NOT going to get you to run on lower volts, it will only help temp and give you a little bit more voltage to work with to achieve a higher clock.

Say with air on a CPU (lets say i7), you are hitting 95c @ 1.35v and can only hit 4.2ghz. Water will allow you to hit that but probably 15-20c less (say 75c). Well, now you have some more temperature to work with, so you could squeeze a bit more otu of your CPU by up'ing the voltage and getting to possibly 4.4ghz. I could almost guarantee that even if you dropped the temp 45c (from 95c to 50c), you WONT be running lower volts to achieve that 4.2ghz, youd only run alot cooler prolonging the life of the CPU and giving you mroe headroom.

Im a firm believer that temperature will kill a CPU faster than voltage. Id rather give an i7 1.45v w/ water @ 75c than give it 1.35v w/ air @ 90c.

BobbyBubblehead
11-23-09, 04:06 PM
myself im not suffering clock envy looking around at other peoples Q9650 speeds.
it just handles 4GHz on water.

you want more then I suspect you need a more extreme solution.

28fans, two cases, two rads push/pull still no more speed :)

Conumdrum
11-23-09, 04:27 PM
Your watercooling will remove any ambient issues with the chip. If it's a good chip, good mobo, capable memory I don't see why not. You might reach it you might not. All you can do is try. If it don't reach it due to bsods, increasing ram and NB voltage can help. You got plenty of rad, you might, might need to look at cooling the NB chip too. You got enuff rad etc.

It's like racing a dragster. Not fast enough? Move up a class, spend mo'money.

BobbyBubblehead
11-23-09, 04:31 PM
my chips rubbish already fried 4gig of ram trying to break past 4070 stable :cry:

I think its trying to tell me to leave off :rain:

Spawn-Inc
11-23-09, 08:05 PM
with my setup i wasn't able to reduce the vcore, but then again i didn't try. it works fine with where it's at so i will leave it at that.

BobbyBubblehead
11-23-09, 08:39 PM
with my setup i wasn't able to reduce the vcore, but then again i didn't try. it works fine with where it's at so i will leave it at that.

#Laughs#

is that canadian for I couldnt be botherd to try it out?

Spawn-Inc
11-24-09, 07:03 PM
#Laughs#

is that canadian for I couldnt be botherd to try it out?

lol, pretty much. i got my overclock so what do i care about vcore.

think the next cpu i get will be these complicated i9's...