2600k overclock bsod in bios but stable if done in windows

[egh]

Hi all,

i7 2600k
Asus p8p67 deluxe board
kingston hyperx 8gb 1.65v 9-9-9-27 ram.

I just got my Sandy Bridge rig and was playing around with the overclocks but i encountered a pretty weird situation.

I was playing around with the TurboV EVO tool in windows and was able to get 4.7ghz stable at 100 bclk, 47 multi, 1.25 vcore.

my bios always boots at around 4.4ghz (eztuned setup) so i was thinking i would go put my OC directly in the BIOS so i wouldnt always have to get set it back to 4.7ghz after i boot.

went in bios, set bclk back to 100, 47 multi and 1.25vcore, 1.65v ram... BSOD after windows screen,

if i boot in windows at 4.4ghz, open turboV EVO and set the same values for 4.7ghz.... super stable and fast.

i dont touch anything else.. are there other values in the bios i should be playing with?

Thanks

 

[tweakboy]

So stick to the one that works.

 

[stahlhart]

Why are you running 1.65V RAM in a P67 board?

 

[egh]

thats what was automatically set by the motherboard when doing the easytune OC.

edit: dont scare me like that, i checked everywhere and it seems that 1.65v ram works just fine on P67 motherboards. even the kingston site put the ram i have as their p67 compatible ram.

 

[kevykay]

1.65V is fine. I'm running my ram at that too, so I don't think that's causing the problem.

1.25 V vcore is extremely low for 4.7 GHz, so that's probably your problem. I'm not sure why you aren't seeing that problem when you set that voltage in windows, but I'd expect 1.35 V for 4.7 ghz.

 

[KIAman]

Logic would say that Windows bootup calls something that is more sensitive to CPU OC instability than Windows idle.

I was playing around with the TurboV EVO tool in windows and was able to get 4.7ghz stable at 100 bclk, 47 multi, 1.25 vcore.

^ how did you come to that conclusion? Did you run a bunch of stability tests?

 

[Kenmitch]

I'd set the vcore to more than 1.25v in bios....Most likely when OC thru windows your seeing the vcore after droop or drop! Maybe set it to 1.275v or so and give it a try!

 

[d4a2n0k]

Theres a lot of confusion with what happens with the turbo ratios on these CPU's. Your 2600k is 3.4 Ghz at stock speed but will turbo up to 3.8 Ghz. BUT, only 1 core will hit 3.8. The 2nd core will hit 3.7, 3rd core 3.6 and 4th core 3.5. At stock speed the 4 cores will not hit 3.8 Ghz at the same time. They will also not stay at the turbo speed forever, even under load. There are time limits to this, all of this is way over my head so I dont know how long etc. Read here: http://www.intel.com/support/processors/corei7/sb/CS-032279.htm

Now, when you are overclocking things change a bit. If you notice in the "AI Tweaker" section of your BIOS there is the "Turbo Ratio" setting. There is "by all cores" (can adjust in OS) and "by per core" (cannot adjust in OS). If you set it to "by per core" you have to plug in the turbo ratio 4 times in your case for a 2600k. Using this option the CPU will turbo all 4 cores to say for instance 44 and all 4 cores will hit 4.4 Ghz at the same time. When you overclock in Windows, it cannot set all 4 cores turbo ratio. This can only be done in the BIOS.

So, when you overclock in Windows, you are only overclocking 1 core with the other 3 at reduced speeds. Thats why you can hit higher speeds at lower voltages because its a lot less stress on the CPU compared to overclocking all 4 cores in the BIOS.

 

[taltamir]

if i boot in windows at 4.4ghz, open turboV EVO and set the same values for 4.7ghz.... super stable and fast.

What makes you think it is super stable? what kind of burn in tests have you run?

My guess is that it is unstable OC that causes errors in one of the more demanding steps during the windows startup process, causing a BSOD. Sooner or later you would come across something else that would cause it to error.
It is better to just dial back your OC a bit.

 

[egh]

What makes you think it is super stable? what kind of burn in tests have you run?

My guess is that it is unstable OC that causes errors in one of the more demanding steps during the windows startup process, causing a BSOD. Sooner or later you would come across something else that would cause it to error.
It is better to just dial back your OC a bit.

I ran Prime95 for a good 30 minutes without errors, which is good enough for me.

what d4a2n0k makes a lot of sense, i went in bios and just jacked the vcore to 1.35 instead and now it can boot up and is also stable although a bit hotter hehe but still under 70c loads (prime95)

 

[Kenmitch]

I ran Prime95 for a good 30 minutes without errors, which is good enough for me.

what d4a2n0k makes a lot of sense, i went in bios and just jacked the vcore to 1.35 instead and now it can boot up and is also stable although a bit hotter hehe but still under 70c loads (prime95)

Just drop the vcore down a little at a time and test for stability. You may be able to shave off a couple of degrees at load.