Prime95 users, a question.

Status
Not open for further replies.

Squirtis

In Runtime
Messages
172
Do you guys always use Small FFTs when using the program to test your CPU temps and stability? I find the temp difference between Large and Small FFTs to be quite large, so I was wondering what the majority of you guys use. Thanks.
 
small ffts to test cpu stability.
large ffts to test temperature.
blend coupled with 3-4 passes of memtest to test overall stability and ram.

IMO, all of these need to run at least 8-10 hours. some say less, some say more.
 
that depends on your cooling solution.
small ffts heat up more because it is using the cache much more.
if you have an air cooling set up, small ffts would test that solution more than large ffts.
if you have water, small ffts are not a stressful test at all. i guess i should have asked before i made my post.
and yes, i made a mistake in my post, large fft will stress the cpu much more than small (i always get them backwards like left and right lol).
so, small ffts for heat and cooling testing.
large for cpu stability and power consumption.
blend for overall stability.
BTW, i have found that if an overclock passes small fft and fails large fft, it is usually a NB/memory dimm voltage issue, not a cpu voltage issue.
sorry for the confusion.
 
thanks for your answers and knowledge, now my next question.

If my CPU temps reach above 60C in Small FFTs but stay around 50C in Large FFTs, should I be scaling back my OC because of temp issues? or should I base my CPU temp/stability judgment upon the Large FFTs.

As of right now Large brings my cores to around 50-53C and Small brings me to 60C after an hour testing. I OCed my Q6600 to 3.2 GHz.
 
no, 60* is not a problem AT ALL for that cpu.
let it run for several hours and see how hot it gets. it should be at the upper limit by about 1-2 hours or so. use coretemp, it records minumum and maximum temps in its control panel.
what are your idle temps? what cooler are you using? what is your cpu core voltage?
 
well i'm not at home right now, but off the top of my head. My idle temps are around 30-35C, my heatsink is a Thermaltake Big Typhoon with AS5...1 week into priming the paste, in an Antec 900.

My voltage is all wonky. The default was pretty high at 1.3250, but CPUZ only read 1.300. I OCed to 2.8 and put the voltage to 1.350. Last night i pushed my OC further to 3.2 GHz and moved my voltage up in the BIOS, to 1.40, but CPUZ still only reads 1.325, about. It's weird cuz CPUZ is reading around the same voltage, but when i OCed to 3.2 GHz and didn't increase the voltage in the BIOS it was unstable and i got a BSOD. What's up with the BIOS voltage and the actual reading? Is there a way to minimize this disparity?

thanks for the help, this message board has proved invaluable to me.
 
yea i've noticed some you have that feature when you post your BIOS pics. Mine, the eVGA 680i, does not. Is there any other way to remedy the problem? Is it even a problem? Should I only worry about the actual voltage output that CPUZ reads?
 
Status
Not open for further replies.
Back
Top Bottom