AMD64 running temperature question

I have a San Diego 3700 which is overclocked at 2.8Ghz. Its load temperatures average at 55 degrees. Is this too hot for long term usage, im not sure. vCore is now 1.55v btw

PS: temps are sometimes as high as 59 degrees

Load shouldnt be anywhere near that high. If you're running that hot, you need better cooling, better thermal interface material or some case fans. My Opteron is at 1.6v and only hits 45*c MAX under load!

I wouldn't like to go over 55C at anytime and would prefer to keep it under 50C. My hunch would be that aiming to keep the CPU between 50-55C, if you can't cool it anymore, it would keep going till your next upgrade. But that is a guess from me.

55 is still well within AMD's specifications, but it's not ideal. Keeping it under 50 is a realistic goal. Doub't you'll see much below 40 unless you're on water or serious high end air.

so my unstable OC at 2.8 is likely to be temperature related?

1.6v and only 45 underload? thats alotta juice, id bet its probably closer to 55 under load...

LeAd_Poisoning said:

so my unstable OC at 2.8 is likely to be temperature related?

Click to expand...

No, infact very unlikely that it's unstable do to temps. Those core temps are well within spec. Though most people, myself included would say that 59 is too high for longevity and continual use.. 59 should not cause an unstable OC. I had my AM2 3800+ running over 60, as high as 65 and it was stable. I knew those temps were not good for longevity though so I went with a Zalman CPU cooler to get my temps back in the high 40's/low 50's.

What are using to pull your temps? I have found a large discrepancy from 2 or 3 different programs and from what my BIOS says.

Consider a good heatpipe cooler, the more pipes the better. I'm running ninjas on all my Big horses and none of them reach more than the lower 40s in this summer heat.
Keeping it below 60 improves longevity and the lower the better. It's always possible that temps above 50 could cause crashes due to hot spots that can exceed 70C but never detected in the thermal diode(s). Consider heat patterns to be similar to what you see in brain activity scans on an MRI. One part can heat up above the detected temp due to the fact that part is not close to the diode(s). The temps are detected for the total effect of the chip and is located near the hotest point determined by the engineers. This is the technical reason why we run all these superpi's, prime95, cpuburn, 3D Mark etc just to use as much of the proc as possible. Running the chip cooler allows faster heat dissapation and limits these hotspots from reaching higher than spec.

almstsobur said:

What are using to pull your temps? I have found a large discrepancy from 2 or 3 different programs and from what my BIOS says.

Click to expand...

ASUS PC Probe

Whats the best skt 939 air cooler i can get then?

I think the Akasa Amber Series case fans i bought might be total garbage, silent but not enough air flow

freeagent said:

1.6v and only 45 underload? thats alotta juice, id bet its probably closer to 55 under load...

Click to expand...

Nope - CoreTemp is reporting it at that, as well as Central Brain Identifier - one using the thermal sensor on the chip - the other using the sensor on the board thats connected to the chip

I have an Arctic Freezer 64 Pro on this, biiiiig fan on it

LeAd_Poisoning said:

Whats the best skt 939 air cooler i can get then?

I think the Akasa Amber Series case fans i bought might be total garbage, silent but not enough air flow

Click to expand...

The Arctic Cooling Freezer 64 Pro that I have is pretty good - cost me £17 at YoYo Tech in Central London, and mounts pretty easily....