I don't think so.
Ivy Bridge will not degrade at all, unlike Sandy Bridge & SB-E that degrade so fast.
I have been using Very High/Turbo LLC since one year ago.
Tortured up to 100C, my chip is still performing as perfect as brand new.
Ivy Bridge is as fast as lightning, as strong as a steel.

Is that a stable OC? What's your max temp and what cooler are you using?

Perfect! Nice chip you got there

Stock cooler can't go higher. Put it under water is your best choice. But nothing can beat a delidded chip, which will offer 15-20C improvement.

A happy Ivy Bridge owner spotted.
Now it's time to play with mem OC.

Got this PSC kit today... G.Skill Flare 2000C7



stock 2000 7-9-7


quick test on 2400 7-11-7

Good job bro.
What vdimm, VCCSA, VCCIO?

I suspect there is a bad contact between your IHS or die with your waterblock/ heatsink base.

No idea with D14.


Samsung should be able to take any voltage. I would bump it up to 1.65v etc it won't die without heatspreader.

Yes it could be board issue or RAM chip itself has hit max capability.
I must agree with that statement, Asus & ASRock have been awesome at mem OCing whilst MSI is not doing good due to rubbish BIOS (sorry )

No matter how good is your cooler, even a quad TEC, you wont be able to tame the Ivy Bridge because there is still a culprit underneath the IHS, until the IHS is delidded, black adhesive glue is removed and TIM is reapplied. Then only you will see the improvement. So guys, dont waste your money for expensive cooling gear, if you dont plan to delid your chip.

No idea bro what's happening but H100 is supposed to perform as good as the D14. You should check your radiator mounting so that your pull fans are intaking cool air.
For the cost of H100 + 4x GT, I think better get a RASA RX240 kit or even the RS360, beat those out of craps lol. More interestingly, RASA kit is upgradable.

This post has been edited by owikh84: Apr 13 2013, 03:27 PM

It's been quite a while I don't run stress test. So recently I had some time to play with these TridentX kits.



G.Skill TridentX 2666 10-12-12-25-1T @ 1.65v (Samsung IC)


I have never seen anybody LinX @ DRAM 2800MHz successfully yet, so here is it now

G.Skill TridentX 2800 11-14-14-25-1T @ 1.66v (Hynix CFR IC)


Running 11-13-13-25-1T now... Don't know whether my stuffs can make it or not. Stay tuned...

This post has been edited by owikh84: Apr 13 2013, 05:25 PM

Pls show us some screenshots. How did you run the prime95 and for how long? Using custom blend with at least 90% RAM stressed?
What version of prime95? Older versions than 27.7 don't have AVX instruction to max stress.
Supposedly prime95 is slightly lower temps than LinX after a few hours whilst LinX can shoot up drastically within seconds.

my 3570K (undelidded) @ 4.6GHz with 1.376v cooled by two 360 rads in none airconded room with ambient as hot as 35C.



So i think to hit 105C with your 240 rad is quite normal. That's why I said earlier.

It's ok. There is nothing wrong with the wb mounting.

I got one extra delidded 3770K 5.0 1.42v LinX stable IMC 2776 max to let go for RM2k. Those who got money please PM me.

Actually 3770K should yield 110+ GFlops, 3570K 125+

Not bad at all. X60 performs really good like any reviews recommended. List updated.

Yupe, sometimes GFlops could be fluctuating depending on voltages ie vCore, vdimm, VCCIO, VCCSA, RAM timings, as well as RAM amount selected.


Welcome to the club. Your temps look fine. Try start OC to 4.0 with stock voltages then slowly go up further.

Nothing new for me, it's been discussed earlier that from my own experience prime95 requires more juice than LinX. You might wanna have a look at post #4 of this thread.



Another thing I saw your GFlops is not 110+, which is not really showing a correct figure for 3770K. Most probably a slight bump on voltage is needed to get higher GFlops. Good luck

This post has been edited by owikh84: Apr 15 2013, 02:49 PM

I have been using 5.0GHz @ 1.48v for 24/7, haven't seen any degradation since the first day.
Ivy Bridge is a lot more durable than Sandy Bridge in which the latter is a hell of degrading craps.
In conclusion, IB will not degrade at all, until you put an extremely high voltage on it say 1.9-2.0v.

Batch is meaningless, total crap shoot. It's all about draw of luck.

Voltages could be vCore, vdimm, VCCSA, VCCIO, vPLL.

Seems like AIDA64 Stress Test uses its own engine so just noticed that they have updated their version to support AVX instruction.
I will try it later when I have free time. A quick note from me is:

The higher the core temps and heat generated, the more stressful the test.
The lower the vCore under load (vdroop), the more stressful the test.

This post has been edited by owikh84: Apr 15 2013, 09:07 PM