Ivy Bridge Overclocking Guide (Extreme LN2 Section Guide Included)

[sin0822]

looks very gooD!!

[Ace.]

yea ist ok for the first Test

[sin0822]

nice man keep it coming!

[Ace.]

4Ghz it is possible to run under 1V!!!
My Board is the min. 1V so i can't change Voltage more down

[sin0822]

yes yes that looks pretty col. How much do you need for 5GHz? 1.4v?

Try to set DVID offset to the negative if possible. What board do you have?

[herderien]

what is the influence of CPUPLL, i mean what happens when you add/remove juice besides reducing temp or influencing cold bug ?

[sin0822]

that is about it, it feeds internal clock generators, i mean not every voltage helps OC. I found the PLL doesn't do much, it can help though i guess.

[dinos22]

wow nice!

4Ghz it is possible to run under 1V!!!
My Board is the min. 1V so i can't change Voltage more down

[owikh84]

4Ghz it is possible to run under 1V!!!
My Board is the min. 1V so i can't change Voltage more down

win7 SP1 + prime95 v2.74?

[cma6]

Sin, thanks for the superb O/C guild to Ivy Bridge. However, if one is on air or water, it seems that one should stick with the i7 3930K as per the Anand Tech article. What do you think of that perspective: the high tmps on the Ivy Bridge make it unsuitable for air/water O/C. http://www.anandtech.com/show/5763/u...-on-ivy-bridge

[Ace.]

@ sin0822

i try it with Offset and the the minus Offset is 200 so I do not come under 1 volt
Mainboard is the Biostar TZ77XE4
I play Batlefield 3 and all the stuff ist game stable with 1V 4Ghz thats good

@ owikh84
Prime 26.6 Build 3 is best in the Moment all other Prime Version to many bugs

[owikh84]

@ sin0822

i try it with Offset and the the minus Offset is 200 so I do not come under 1 volt
Mainboard is the Biostar TZ77XE4
I play Batlefield 3 and all the stuff ist game stable with 1V 4Ghz thats good

@ owikh84
Prime 26.6 Build 3 is best in the Moment all other Prime Version to many bugs

Ok, but 26.6 is not coded to utilize AVX. So expect 27.4 or latest IBT/LinX to load more vcore.
Let's try IBT/LinX in win7 SP1.

[ToyTen]

Thanx for this very useful topic sin0822

[Allubz]

4Ghz it is possible to run under 1V!!!
My Board is the min. 1V so i can't change Voltage more down

What cooling are you running? Those load temps are quite nice, as 4GHz is still fine for daily use and this would save a lot in small / badly cooled builds.

[stasio]

Prime95 v27.6 Build 3 is latest.

[Ace.]

Prime95 v27.6 Build 3 is latest.

This is the latest version, but 26.6 brings faster error as 27.6. why? i dont know

[Ace.]

4.7Ghz is already hot

[sin0822]

Sin, thanks for the superb O/C guild to Ivy Bridge. However, if one is on air or water, it seems that one should stick with the i7 3930K as per the Anand Tech article. What do you think of that perspective: the high tmps on the Ivy Bridge make it unsuitable for air/water O/C. http://www.anandtech.com/show/5763/u...-on-ivy-bridge

Ivy can take the high temps just fine, the transistor is newer, it is moving more current per unit area, it will run hotter, it should run hotter. its leakage is also increased thus it will waste a little more power than previous gens, and it will run hotter. However what is surprising is that ivy uses less power than Sandy.

So what scares people about heat? Is it the temperature or the amount of actual heat released in the air? The bottom line is if you run a Sandy bridge CPU or SBe CPU at 4.8ghz all day and leave it in your room, your room will be VERY hot at the end of the day, however if you leave a Ivy bridge CPU running at 4.8ghz it's temps will be higher but your room wont be as hot. This is becuase Ivy bridge will consume less power, but show a higher temp.

IMO this wont be fixed as we move down in size of the processor, my bet it that with haswell this will get worse. So there is only one thing we can do, realize that the temperature isnt' always very important as long as you stay below a certain point. That point for Sandy Bridge was 80C with a TJ-Max of 95C, with ivy bridge the TJ-Max is 105C and the CPu wont throttle until ~102C, this means that the CPu can take about 10C higher than Sandy on average. So I will now start recommending keeping below 90C because of this. Also I messed with ivy for a long time on high volts, like 1.5V+ and I did the same with Sandy. Sandy degraded VERY fast, 1.55V would be 1.57v the next week to do the same 5.2ghz OC, it kept increasing up to 1.62v until i said just forget it. On Ivy bridge this number has not increased at all, it seems to be a much more resilient CPU and seems like it can handle a lot. now I am not sure if one day it will just let out and ide, but you can always buy the Intel Tuning protection program and have some fun.

Remember guy you can get a warranty from intel to mess the CPu up to hell with OCing and they will replace it for you! Even though i doubt you will kill the CPU.

[MannerRev]

Is there any word on what Vcore would be 24/7 safe for a 3770k @ 4.6-4.8? Right now I have 4.6 @ 1.272V but I'm scared to push higher Vcore for 24/7

[sin0822]

id say you will hit limits with temperatures first. keep under 1.3 imo

[BababooeyHTJ]

This is a great guide. The tip on bumping up the vccsa when clocking the memory saved my ass.

[BAR*B*Q]

Thanks for this guide. My first oc attempt for a long time. Could you please provide information about voltage reading points on this motherboard? If possible also for the Z77MX-D3H (TH).

[sin0822]

show me a picture of the back and I will show you

[BAR*B*Q]

Hi,

cool, here a picture with with a mounting plate of the mx d3h th. I hope that is not a problem.
thank you so much.



cheers

[BlindFreddie]

Great guide/thread. Thanks, sin.
** Why not stickied/moved or copied to Guides section yet? **

Also see threads on de-lidding: 3770k IHS Removals - CPU temp dropped from 79C to 71C
Warning on de-liffing: Apart from probably voiding your warranty, there seems to be a problem with the new high-performance TIM that you put under the IHS needing to be replaced after a number of heat/cool cycles. Liquid metal TIMs like Coollaboratory Liquid Pro & Liquid Ultra seem to be OK.
No-one knows for sure why this happens but the leading theory is that different amounts of expansion/contraction of the metal IHS vs the silicon die shear and break up the TIM layer, and the effect is a bit like re-using TIM.
Dunno why it doesn't seem to have been reported in the days of running naked CPUs (no IHS) or with GPUs (no IHS) - ??