Core i7/X58 Overclocking Thread

[dejanh]

Someone please correct me if I'm wrong, but as far as I know the DRAM Bus voltage can go over 1.65 as long as it remains within 0.5 volts of the QPI/DRAM core voltage.

This is pure speculation. Until there is an official confirmation from Intel on this you are pretty much safe to assume you are cooking your CPU by going over. Either that or you take the plunge in the deep, cross your fingers, and hope for the best. I just know that if my R2E puts a big red warning that says "do not go over this" I probably should be a bit worried.

Just my $0.02.

[Classic Satch]

This is pure speculation. Until there is an official confirmation from Intel on this you are pretty much safe to assume you are cooking your CPU by going over. Either that or you take the plunge in the deep, cross your fingers, and hope for the best. I just know that if my R2E puts a big red warning that says "do not go over this" I probably should be a bit worried.

Just my $0.02.

Lol, I agree. And thank you for your response to my question on my RAM voltage.

[003]

This is pure speculation. Until there is an official confirmation from Intel on this you are pretty much safe to assume you are cooking your CPU by going over. Either that or you take the plunge in the deep, cross your fingers, and hope for the best. I just know that if my R2E puts a big red warning that says "do not go over this" I probably should be a bit worried.

Just my $0.02.

It was confirmed by the intel rep on XS... forgot his name, but I'm sure somebody can find where he says that.

[dejanh]

It was confirmed by the intel rep on XS... forgot his name, but I'm sure somebody can find where he says that.

I've only seen this coming from other forumers, not from official Intel reps. It makes no sense anyway because if that was trully done so, then why not just make it so that your QPI voltage limits your DRAM voltage. That way if you up QPI voltage you get more room to up DRAM voltage.

[D3mon_Hunt3r]

Actually a few posts up you can read that table that states 1.85 maximum safe for Vdimm voltage as long as you have the max safe Vtt of 1.35 (the 0.5v margin.)

[freshy]

14 hours prime still going @4.1ghz, HT on, turbo on , might stop it

[Lukee]

Well, the best I can do with either 20x multiplier or 20x + turbo is 195 x 21

Freshy - Looks like we are similar. i ran Intel Burn test for a while and that was stable, and now I'm running Prime 95. Only 45 minutes into it, but normally my system freezes in 5 max so this is a great sign.

I'm putting her under water this weekend so we'll see what temps are then!!

[dejanh]

Actually a few posts up you can read that table that states 1.85 maximum safe for Vdimm voltage as long as you have the max safe Vtt of 1.35 (the 0.5v margin.)

Indeed you are right. So why the heck are they saying to limit DRAM voltage to 1.65V? Maybe 1.75 is reasonable considering that absolute max is 1.875V.

By the way, I spent some more time looking into the 20x multiplier issue and I think I am on to something...I posted this in another thread but this one is a lot more active so I'll bring it up here just in case...

Basically, after some more tests, I am starting to think that this has something to do with heat. If I lower the ambient temperature I can run stress testing longer at 20x200 then when I leave it at normal. Also, if I crash at 20x200 and then restart to try the run again, I cannot do it. But if I turn the system off for a few minutes, then power it on, I can run 20x200 again. Why in the heck would this be the case though when even higher volts and temps have been stable under say 19x216? Could something be making the CPU too sensitive to heat at 20x200? Some protection circuit tripping when it is not supposed to? Any ideas why this would be the case?

[BiOvOrE]

Indeed you are right...

Pfft, speculation...

Anyways, my latest stable(ish) overclock on air for those interested.




Stable for an hour or so at this stage. vCore is 1.385 and QPI/DRAM is 1.35. The QPI/DRAM core voltage seems to have more effect on stability at the ~4Ghz range for me. Gonna Prime overnight and see how it goes.

[Bootsy]

Can't tell you where I got the block...

I have reseated this block on the cpu three times now...I think this is just a really hot CPU! It idles nicely at 47C 1.45V....but as soon as you load down all four cores and the four virtual cores, this thing heats up like a furnace! Maybe i am missing something...maybe it just can't take that much voltage at 4.2 GHZ??

-M

Here is my chip @ 4.2 all 8 threads 1.45 with the same block 5hrs+ into a prime run not even close to 100c. My chip idles at 33c are you still getting high temps?

[freshy]

Well, the best I can do with either 20x multiplier or 20x + turbo is 195 x 21

Freshy - Looks like we are similar. i ran Intel Burn test for a while and that was stable, and now I'm running Prime 95. Only 45 minutes into it, but normally my system freezes in 5 max so this is a great sign.

I'm putting her under water this weekend so we'll see what temps are then!!

yep , but you got HT off ? and also lower voltages

that last prime test for 14 hours i done @ 4.1ghz i had the vcore 1.45v and the qpi/vtt 1.435, thought id try drop voltages a bit got 4.1ghz stable @ 1.425v vcore stopped it after 5 hours prime, im just trying to get qpi/vtt down a bit now , so priming now at 1.425vcore and 1.375v qpi/vtt, see what happens, i really think we gonna hit 4.2ghz sooner or later, might try give it a bit of qpipll v next run @4.2ghz, but first i wanna try drop voltages as much as possible for 4.1ghz

and under water sounds good

edit...dejanh i really dont think its a temp issue as my temps are the same even with the 20 x multi , but it just isnt stable, i really hope a newer bios fixes it, also i was just reading back and roller said that his 20 x 200 works fine his is a CO revision, mine is a co/c1 revistion,

[Lukee]

Freshy, I agree. It is only a matter of time we hit 4.2.

Yes, HT is off. I figured that I need to find a stable OC with HT off first, then I can turn HT on, but to be honest, I play mostly games and do some encoding. What concerned me was the temps without HT on. I wanted to wait until she was under water to try it with HT.

[modboxx]

Are these overclocks 24/7 stable or we just playing around? I just sold of my 780 q9550 combo and i'm looking to pick up a rampage 920 combo was wondering if i should expect 4ghz.

I will be wc once they release the hold down plate for my fusion.

[Splave]

bootsy what is you WC setup, nice work

[Lukee]

Are these overclocks 24/7 stable or we just playing around? I just sold of my 780 q9550 combo and i'm looking to pick up a rampage 920 combo was wondering if i should expect 4ghz.

I will be wc once they release the hold down plate for my fusion.

I would need to test mine a little bit longer, but Intel Burn Test was stable for about 10 runs and Prime95 small FFT was stable for a bit over 2 hours so a little more testing is needed, but to be honest, I'm not on the PC more than 3 -4 hours anyways, so for ME it's stable.

[Classic Satch]

Well, I saw the issues some were having with the 20x multiplier so I upped the Bclk to 211, dropped the multiplier to 19 and lowered the voltage to 1.39. 12 hours later of Prime95 and still stable with HT on.



Time to drop the voltage further and retest.

[freshy]

just crashed after 1 hour of prime @ 4.2ghz HT on, turbo on


turbo off tho ha classic ?

[Lukee]

Well, I saw the issues some were having with the 20x multiplier so I upped the Bclk to 211, dropped the multiplier to 19 and lowered the voltage to 1.39. 12 hours later of Prime95 and still stable with HT on.



Time to drop the voltage further and retest.

Nice. I was able to get 211 x 19 to work, but only with HT off.

[Classic Satch]

just crashed after 1 hour of prime @ 4.2ghz HT on, turbo on


turbo off tho ha classic ?

I'll try it with Turbo on once I find the lowest voltage that I can run 4.0 stable with. I don't think Turbo is that important tbh.

Nice. I was able to get 211 x 19 to work, but only with HT off.

I had tried 200x20 with 1.41v and it kept crashing after about 30 minutes. Lowered the multiplier and voltage and upped the Bclk and whaddya know? It's stable. Go figure.

[freshy]

199 x 19 + turbo 4.179ghz hope its still going when i wake up

[bmilos]

@Bootsy - what are your other settings? I am running at the same 200x21 @ 1.43750 on a GTZ too, and my load temp is about 88c. Is the CPU the only block in your loop? I also have my NB in my loop. Maybe thats the root?

[rge]

With all new F4 bioses TDP unlimited turbo... need to find settings for those. F3 sweetspot on mine was 1.35 bios 1.31 load for 4/4.2 turbo.

On bios F4j, 1.38 bios 1.34 load for 4.1ghz 8hrs so far...settings on pic where I can remember them myself

Need 1.39-1.4 bios for 4.14, and ? 1.42 bios for 4.2...have to decide between those after figuring out exact settings.

[Classic Satch]

Wow, I'm comparing temps of you guys with similar setups and it looks like my Exos-2 LX is holding up pretty well. I was too lazy to build a custom water cooling system and went with the LX and a Apogee GTZ water block for my new build. Comparable performance from what I can see so far.

[Hot'n'Chilly]

Question: What were your temps at stock speed before you began to OC? I'd like to compare my stock-clock temp to yours before I OC and then compare my OC temp to yours and see how they compare.

[Classic Satch]

Question: What were your temps at stock speed before you began to OC? I'd like to compare my stock-clock temp to yours before I OC and then compare my OC temp to yours and see how they compare.

I believe they were around the 30-33 range, depending on the core. Of course, my new 920 was only at stock speed for as long it took to load Vista 64.