SB Overclocking problems

MikeHiow

MikeHiow

MikeHiow

Permabanned
Joined
26 Oct 2004
Posts
7,540
Location
Isle of Wight
I seem to have hit a wall at 4.6GHz with my 2500K and P67P8 (Beta BIOS).

That is with;
Vcore: 1.34V (1.35 in CPU-Z)
Mem: 1.55V
PLL: 1.80625
SA: Auto (Although tried at 0.96875)
IO: Auto
LCC: Extreme (100%).

I tried 4.7GHz with as much as 1.4Vcore and 1.85 PLL but windows would not boot. It did not bluescreen, but simply froze on the little windows logo.

Any suggestions? Is 4.6GHz likely to just be my limit on this chip, or am I missing something/doing something wrong?

Temps seem OK with 4.6/1.34V with highs between 72 and 76 across the four cores.

Also, how do I get it to down the voltage when clocking down to 1.6GHz? Seems pointless for Speedstep to be enabled if it is staying at over 1.3v.
 
A bit of everything really, however I do not wish to clock it out of necessity, but rather just 'because I can'.
 
1.35V, over 70 degrees, and LLC set to "extreme" suggests we'll be seeing another "Sandy bridge died!" thread soon.

Still, I'm glad you're having fun. 4.6ghz stable is good (though I note you're not claiming stability anywhere). If more vcore doesn't help, trial and error with the other settings is the way to go. It's a new platform, so being one of the early experimenters carries a certain risk with it.
 
All my settings are within recommended limits (VCore was actually suggested to be OK upto 1.42V, but 1.34V for 24/7 use - I'm 0.01V over!). I'm sure I could knock the voltage down a bit yet, but I was planning on running Prime over night first.

As for LLC set to extreme, this just means I get the voltage I'm actually asking for, from what I can see - People with lower LLC just seem to be getting less than the VCore they set. :confused:

As for stability, I've done a 20 minute Prime session, played some Black Ops (now we all know how CPU hungry that is!) and done some heavy photoshop stuff - all without a glitch. But as I say, I shall run Prime overnight to double check.
 
I'm fairly sure 1.42V (at 70 degrees) would kill a 32nm chip. That isn't quite what llc is, but I suppose you're close enough to the mark. Has it been thoroughly tested at any lower frequencies? It's very difficult to determine what the minimum stable voltages are without testing repeatedly as the frequency creeps up.

Perhaps I'm just jealous though, as my old 920 isn't likely to make it to 4.6 however hard I try.
 
Perhaps you just have a chip that doesn't clock as much as others...4.6 might be all you can get

Would hate to have my chip running that hot though.I'm more used to 25c loads
 
Knocked LLC down to 75% and now voltage fluctuates between 1.336 and 1.344 - ran a quick prime and it seems ok thus far.
 
Justintime said:
Nothing wrong there imo, not sure on the voltage dampening bit though i've heard no reports of any rampant chip deaths of any generation due to this.
Click to expand...
I've had 1.41v at 80+ degrees at 4.8Ghz and survived, although the overclock wasn't stable. I don't think the SB chips are quite as fragile as some of the 'SB died' threads have suggested.
 
You must log in or register to reply here.
Back
Top Bottom