i7 qpi and ram voltages.

setter

setter

setter

Caporegime
Joined
14 Dec 2005
Posts
28,067
Location
armoy, n. ireland
I was reading something earlier regarding qpi and ram voltages, and how they should be set. (quote below)

You must keep the difference between dram voltage and qpi voltage within 0.5v

Should this be the case at all times, im currently running a 920 DO on the following voltages at 3.6ghz, (HT and turbo enabled) bclk 173x21.

vcore, 1.18
pll, 1.8
qpi, 1.2125
dram, 1.64

Passed 10 runs of IBT on theese settings (maximum stress), do theese voltages look ok. Thanks for advice.
 
That's what I've read too. It's more applied in terms of "it's ok to go over 1.65V on the ram as long as qpi stays within 0.5V of it". QPI should be below 1.35V according to intel, PLL up to 1.88V. Some people find more stability with pll below 1.8V.

Your settings are fine though, with the exception of bsck which is too low.

I'm reasonably sure that enabling turbo makes my computer irrational, can't get 4.2ghz stable with it on but I can with it off. Strange and frustrating.
 
JonJ678 said:
Your settings are fine though, with the exception of bsck which is too low. .
Click to expand...
Yep i know mate,:o, gonna have a crack at pushing it a bit higher, ive seen a few folks starting on very high voltages/bclk then working downwards lowering voltages, im doing the opposite as i prefer to work from a stable base, takes a bit of time though.:)
 
Decided to push a bit higher, currently testing at 3.8ghz (183x21), failed IBT on the above settings after 2 rounds, upped vcore to 1.2500, qpi/vtt to 1.2500 and pll to 1.84, 10 out of 20 completed on IBT so far, highest temps so far, 66-61-63-61. If all goes well ill be trying for 4.0ghz
 
Thanks for the advice.:) Hopefully i can maintain stability on my current ibt test at 3.8ghz, should be a good foundation from which to push higher.
 
I doubt you needed the increase on pll somehow, but fair enough. I'm on around 1.25V on each of those for 200x20, averaging 60 under ibt on water. Your temperatures are very good for 3.8 on air.

I look forward to your results over 4ghz, as I'm stuck anywhere past 4.2.

edit: I recommend you disregard bluemax and increase the voltages only to what your system needs to remain stable.
 
Last edited:
Cheers jon, i may try lowering theese voltage settings just to see if the increases i used were maybe to much, temps are actually better than i expected, im using sharkoon silent eagle se pwm fans, (400-1200 rpm) currently running at full rpm (nice and quiet too) i could have replaced them with the silent eagle 2000's i have as front intakes but it would have meant breaking down the whole rig again to swap them round.
 
I'm more surprised that you saw any change in stability when increasing PLL, it doesn't seem to make any difference for me at all. If 1.25V is too much for 3.8, I'm sure you'll soon be running fast enough to need an increase anyway.

Its worth trying with turbo off as well to see what difference it makes, I think its unusual to find it derails everything but I'm not sure.
 
Ill definitely try knocking pll back to 1.800 for interest, i remember on my p5q deluxe and q6600 combo, when pushing it to 3.8ghz i needed a hefty jump on pll to 1.6, needed very little on my q9550, 1.54 iirc. Dont intend on staying at 3.8ghz for long so id say a few increases on qpi and vcore may be required.:)
 
Failed IBT after 16 passes, looks like im gonna have to do a bit more tweaking.

ibt38.jpg
 
I'm still having odd problems with my overclock, ive left it at 3.8ghz for now..

Off the top of my head I'm using 200 x 19 which lets my ram run at the natural frequency easier, currently my cpu voltage is at 1.25625, pll at 1.88 and qpi at 1.35.

Oddly enough though, 200x20 needs a bigger voltage increase for me than 191x21 which is so confusing, makes me think i may be doing something hideously wrong.

Edit: oh, and my ich and ioh are at 1.2V.
 
You must log in or register to reply here.
Back
Top Bottom