I5 2500k overclock unstable with p95??

[maverickdc]

Hi All,

ive run intel burn test for 4.5ghz on high and passed ran it on p95 and few minutes later BSOD, after loads of trial and errors i came to the conclusion i would rather keep it to 4.3ghz and keep temps safe but p95 just wants to keep BSOD me, intel burn again is fine, im on 4.3GHZ at offset of 0.135, tried 0.125 and P95 BSOD evan faster than 0.135 lol, whats going on? i thought upping the voltage would help but it seems to have made things worse???

quick edit: the max voltage it goes to is 1.248 on the 0.135 off set.

any ideas?

 

[neograndizer]

It's going to be dependent on your chip as well. I need an avg vcore of 1.3-1.35v to remain a stable OC at 4.3GHz. I hit the wall at 4.4GHz (I couldn't get it stable no matter what I did). I have mine set at offset + 0.015v. I tried offset - but my chip doesn't play nice and tends to crash when it goes idle.

You'll still need to play with the voltage to see what it needs to be stable.

 

[steddora]

IBT isn't a good way to determine small stability issues. IBT WILL however determine a big instability very quickly. IBT will also press your processor to higher temperatures then anything you could ever throw at it. So IBT is perfect for testing for initial stability and thermal issues. Prime95 is more thorough in the search for instability. It can determine very little instabilities. At that speed, you're going to want probably a fixed voltage of 1.3 or higher. My 2600k holds stable at 4.6Ghz with an offset of +0.020, however won't hold stable at 1.300v fixed. So play with the voltages and watch the temperatures.

Offsets are hard to understand as the processor demands a certain amount of power and you're controlling what it gets. When it's asking for 1.250v and you're at a +0.050v offset, you're telling the motherboard to give it 1.300v instead of the 1.250v it's asking for. So you have to account for those variables and the biggest thing; it's hard to determine where it will hold steady in it's need for voltage and as neograndizer said.... the idle side can be too low as well with offsets. This means it could run a prime95 session for 48 hours completely stable but once it backs off to idle it crashes because it doesn't have enough voltage to remain stable when idle.