• Notice regarding image hosts and attachments. See here.

  • Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

***Official Ivybridge Overclock Thread***

Those events were from prior tests not from this one there just still on the log list I had tried out an unstable overclock and reverted back to this one which was stable previously and tested it again for even longer and it has been stable as before whcih was only for a few hours and this test for nearly 8 over all 24 hours stable with no errors with this oc.

Just the temps P_P
 
You're using extreme/max LLC, which will give a lower load voltage. If you were to run with LLC disabled you'd probably need 1.35v in the bios with a vdrop+vdroop of 1.25v under load (estimates).

Personally I'd stick to medium LLC or below, but opinion is split on the evils of LLC.
 
Not necessarily because Vdroop (voltage drop under load) would mean that you aren't actually running at 1.35V when the cpu is loaded. I haven't played with Ivy Bridge but 11 degrees does seem like a lot of variation in temperature between cores, I normally only get about 5 degrees variation under stress testing.

I'd give it a shot without LLC if not long term, at least to see what you can achieve.
If you are happy with your best non-LLC result then stick with that IMO.
 
So what exactly do you suggest lol, I would rather not have to use fixed volts all the time but i really want to get the temps down O_O so i could maybe increase.

I dont really understand Vdroom and such things much tbh.


PS- Ive reseated repasted the difference in core temps is always like that really frustrating really.
 
Last edited:
WHEA logger on your event viewer means an instability. Upp the vcore or expect a blue screen.

I had 4.5 stable at your volts and after 3 days of non-stop folding I got 2 blue screens in a day of normal use.

Can take weeks to fully suss out your cpu's true stability that's why people post their 1.2s and 1.3v OC as they are gradually raising it over the weeks when blue screens happen. A day of stress testing is nothing really.

This.
 
So what exactly do you suggest lol, I would rather not have to use fixed volts all the time but i really want to get the temps down O_O so i could maybe increase.

I dont really understand Vdroom and such things much tbh.


PS- Ive reseated repasted the difference in core temps is always like that really frustrating really.

de lid your beast...... its a good chip, i see 5GHz easy for you ;)
 
Vcore is something we're all familiar with, in the case of Allnamestaken's chip it is showing as 1.192V in the CPU-Z screenshot.

Vdrop is when you set the Vcore in the BIOS but the voltage displayed when the processor is idle is lower than that value. For example if you set Vcore to 1.25V but when you boot CPU-Z shows 1.23V that is a Vdrop of 0.02V.

Vdroop is when the Vcore falls even further once the processor is loaded. Continuing the example above, the processor now at 100% load is displaying 1.20V in CPU-Z, thus it has a Vdroop of 0.03V (quite a small Vdroop, I've had a 1.4V Lynnfield chip droop to 1.28V).

LLC (Load Line Calibration) is a motherboard feature to eliminate Vdrop and Vdroop. It means that if you set the Vcore to 1.25 it will stay very close to 1.25 (although it's not going to stay perfectly on 1.25 the voltage won't droop or drop as much). The advantage of this is that you aren't pushing a dangerous amount of voltage through your chip when it's not loaded, it should reduce idle temps and it's easier to predict what your load voltage will be so it can speed up your overclocking. The downside is that it causes voltage spikes when your processor changes loads (for example when it starts or finishes a task). This has been documented but whether it causes damage long term is controversial; some think that although it is a high voltage the total current through the chip in that time (since the spikes last milliseconds) isn't sufficient to cause damage. I think it's best to use as little LLC as possible, that doesn't mean none. Do I enable it on my chips? Yes but I try to avoid it if I can. For a benchmark I'll turn it all the way up, for 24/7 clocks I'll probably still use it but on a lower setting (some motherboards have a range of values, I think the Asus RoG boards have 10 settings).

Allnamestaken, I suggest turning off LLC or setting it to its lowest value, upping the voltage a bit (probably to about 1.26V) then trying for a maximum overclock again. Once you figure out how high you can go (you will be limited by temperatures) then you can make the judgement call on whether you want to use LLC or not. It's a controversial topic and I can only give my opinion, ultimately it's your call. Although I doubt LLC will be enough to fry a chip, I'd rather play it safe with my 24/7 clocks so use less than maximum.

Related links
http://www.anandtech.com/show/2404/6
http://www.hardwarecanucks.com/foru...3-asus-p5q-pro-p45-motherboard-review-14.html
 
All of my motherboards have used manual voltage (where you set the value you want).
If yours uses offset then just set the Vcore to default and see if the Vcore is reported anywhere in the BIOS. That will give you the base so you can just add the offset to that. If the exact Vcore isn't reported in the BIOS anywhere then you will have to boot and check with CPU-Z what the default value is (although it won't be as precise because of Vdrop).
 
So say i set it to offset and its 1.2 and i have a vdrop to like 2 or 3 then it will only go up or down by a certain amount from that voltage? I just need to get a offset that goes to the voltage I want at load aka 1.2 volts for 4.5 i think.

Im just not sure the temps are gona be so great for an extra bump in volts considering what my temps are now.

I had an WHEA error while playing lol after nearly 2 days stable nothing is causing me problems but i checked the event viewer and it said one had happend so back to messing around with it i guess?
 
I think you need to find where your cpu is stable at before trying to tinker with vdrops and such. Sounds like you're a fair bit off finding the sweet spot for your OC. Don't be surprised to eventually be raising the volts to 1.25 or even 1.3v for a stable 4.5Ghz. Stulid I remember early in this thread got his stable at 1.4v @4.5Ghz.

It would be extremely rare to get a 100% stable chip at those speeds in under 1.2v so don't be too worried in going over. If temps are a problem lower the frequency until your cooling solution is solved or improved.
 
That's it, you want to achieve the ideal load voltage as that's when the most current will flow through the CPU and thus the most "wear and tear" will occur. Also most important for stability. The higher voltage when it's not loaded isn't important so long as it isn't ridiculously high. Of course temperatures are still important and will still be the limiting factor.

If the error is not causing problems it might be ok. It can cause corrupt data although if you were playing a video game it's not of critical importance so probably nothing to worry about. If you were wanting to keep your current settings with LLC enabled (which would probably be fine) then just bump the voltage by one increment (often 0.0025V) and check in a few days to see if another error has occurred.
 
I need to figure out how to keep my offset voltage so that im not always at 1.192 or what ever voltage i get to.

Will take a few days I guess.

Im so tempted to delid because of the seeminly crap core temps seems like some utter gunk under there.
 
If I can get it at max load around 70C instead of this ridiculous 80+ Then hell yeah but then I guess ill try some better Paste first and wait till i get a full cooling loop in, instead of this crappy 620.
 
One other thing My Pll voltage is disabled completely in bios lol.


(SEEM to have it stable at 4.6Ghz but the 100% load temps hit about 91C considering tjmax is 105C I think I may see how many mhz I can push out of it)

Realisticly I know that at load in games I wont hit over 70C anyway, and once I buy some new paste and not this gunk i had in the house temps will drop also.

Im going to see how far I can go Ill try not to go past 95C if I do im going to back down even if stable lol.


Im at 1.248v at the moment using offset voltage so it reverts to about 1.08 something at idle.
 
Last edited:
Back
Top Bottom