is this overclock stable? - Overclock.net

consume · 07-10-09

I just recently bought an Asus p5ql pro and i just started overclocking my e6320. So far, i set the fsb to 358, and my ddr2 1066 ram to 717mhz(it's g.skill). It's now at 2.5ghz and while running prime95, im not getting any errors and the temp at 100% is 55 - 57c. Is this too hot? Should i not move my ram above the standard 533mhz? Am i frying my cpu/ ruining my ram at that temp/speed?

TwoCables · 07-10-09

Quote:

Originally Posted by consume

I just recently bought an Asus p5ql pro and i just started overclocking my e6320. So far, i set the fsb to 358, and my ddr2 1066 ram to 717mhz(it's g.skill). It's now at 2.5ghz and while running prime95, im not getting any errors and the temp at 100% is 55 - 57c. Is this too hot? Should i not move my ram above the standard 533mhz? Am i frying my cpu/ ruining my ram at that temp/speed?

It depends on the Prime95 test that you ran and how long you ran it for. The default test in Prime95 is the Blend test, but that's not good enough for determining maximum temperatures, nor is it good enough for determining the stability of a CPU overclock. However, these temps are still very cool (pun not intended), so don't worry.

As for the memory, try to keep it at 100% stock for now. If you tweak the memory while you tweak the CPU/FSB, then you'll have too many variables change, and then it could make finding the source of instability next to impossible. The stock settings are printed on the sticker that's usually found right on the memory sticks themselves.

Oh, and you'll know if the temps get too hot: the first result is frustrating instability (blue screens and lock-ups, possibly a random reboot). After that, it might not boot into Windows. After that, it might not even perform the POST, which is the Power On Self Test. But when the temps get WAY too high and reach the Tj. Max, which is about 80°C for this CPU, then the CPU shuts itself off.

But, stability testing is extremely important and vital, and I want to tell you that it can consume a few days to a week. So, please prepare for that.

consume · 07-10-09

Quote:

Originally Posted by TwoCables

It depends on the Prime95 test that you ran and how long you ran it for. The default test in Prime95 is the Blend test, but that's not good enough for determining maximum temperatures, nor is it good enough for determining the stability of a CPU overclock. However, these temps are still very cool (pun not intended), so don't worry.

As for the memory, try to keep it at 100% stock for now. If you tweak the memory while you tweak the CPU/FSB, then you'll have too many variables change, and then it could make finding the source of instability next to impossible. The stock settings are printed on the sticker that's usually found right on the memory sticks themselves.

Oh, and you'll know if the temps get too hot: the first result is frustrating instability (blue screens and lock-ups, possibly a random reboot). After that, it might not boot into Windows. After that, it might not even perform the POST, which is the Power On Self Test. But when the temps get WAY too high and reach the Tj. Max, which is about 80°C for this CPU, then the CPU shuts itself off.

But, stability testing is extremely important and vital, and I want to tell you that it can consume a few days to a week. So, please prepare for that.

Which test should i run in prime95?

MiniTuckin · 07-10-09

small ffts stresses just the cpu itself. I would do the blend test to make sure the ram/cpu/fsb are all stable

consume · 07-10-09

Quote:

Originally Posted by MiniTuckin

small ffts stresses just the cpu itself. I would do the blend test to make sure the ram/cpu/fsb are all stable

Ok, so i'm now running that. How long should i run it for and what's a safe distance away from the tjmax at load? And this test is making it hotter now, it just sits at 56-57c O.O

TwoCables · 07-10-09

Whoa, I'm slipping. I forgot to mention WHICH test to run instead of the Blend test. I'm sorry about that.

For rock-solid stability, make the overclock able to do the following:

Approximately 12 hours of the Small FFTs test (run it on a night just before bed when you have to basically leave to go somewhere as soon as you wake up).
Approximately 6-12 hours of the Blend test.
Approximately 5-10 passes in Memtest (Memtest86 or Memtest86+. Either one is fine. It only tests for errors, though. It's not a stress tester).

If your overclock can do all of this, then you know it's really stable and that you should never have to worry about being frustrated by instability. However, you can put the icing on the cake and make it absolutely certain that your overclock is petrified rock-solid stable by making it able to pass about 50 runs of the Linpack test. For that, please use LinX as it has the best GUI among all the Linpack testers available. Also, be sure to use "All" available memory as this gets the most out of the linpack test. Although, be aware that linpack testing makes the load temps 10-15°C warmer than the Small FFTs test, but this is safe.

TwoCables · 07-10-09

Quote:

Originally Posted by consume

Ok, so i'm now running that. How long should i run it for and what's a safe distance away from the tjmax at load? And this test is making it hotter now, it just sits at 56-57c O.O

See above reply. But regarding the temps: those are absolutely fine. But which temperature monitoring program are you using, and which Tj. Max is it using?

consume · 07-10-09

Quote:

Originally Posted by TwoCables

See above reply. But regarding the temps: those are absolutely fine. But which temperature monitoring program are you using, and which Tj. Max is it using?

coretemp and it's tj. max is 85C