• We should now be fully online following an overnight outage. Apologies for any inconvenience, we do not expect there to be any further issues.

How high the likelihood that 4670k or 3570k OC to ... [insert OC here] ?

flexy

Diamond Member
Sep 28, 2001
8,464
155
106
Someone on another forum is selling pre-tested and also optional de-lidded i5s/i7s. (It's not an US forum, just FIY).

His prices, despite that SOME chips are very great seem unreasonable to me.

For example, he just sold a very GOOD i5 i3570 which did 5ghz, but wants almost €300

I am not effing paying €300 for a last gen CPU, the difference is still only 10% to a "normal" OC, IMHO.

My question is simply how high is the chance that a, say

* i4670K OCs to.....4.2 or 4.3? ++
* i3570K OCs to.....4.4 or 4.5? ++

It's my understanding that those OCs should be rather common...which would mean I can save myself the hassles of buying a "pre-tested" CPU.
I would also say that a 4670k at €200-ish which does 4.3 would be overall better than an overpriced i3570, even if it may do 5ghz...taking into account that Hasswell on avg. is 10% faster anyway. Agree?
 
Last edited:

Kenmitch

Diamond Member
Oct 10, 1999
8,505
2,250
136
Roll the dice and pick up a 4670k. You should be able to get 4.2 or so like you think without too much hassle. Just make sure you get a decent cooler as they tend to run on the hot side....Dependent on another luck of the draw known as the thermal lotto.
 

aigomorla

CPU, Cases&Cooling Mod PC Gaming Mod Elite Member
Super Moderator
Sep 28, 2005
21,073
3,576
126
YMMV is ALWAYS YMMV!

Your millage may Vary....

There is no guarantee outside what the chip is made to do on stock settings.
There is even a small possibility the CHIP will be DOA, so i guess thats negative overclocking.
 

wilds

Platinum Member
Oct 26, 2012
2,059
674
136
Even a stock 4670k would make a drastic performance upgrade from your current (sig) platform.
 

JimmiG

Platinum Member
Feb 24, 2005
2,024
112
106
* i4670K OCs to.....4.2 or 4.3? ++
* i3570K OCs to.....4.4 or 4.5? ++

Those clock speeds should be no problem at all. My 4770K is rather crappy, but it did 4.3 GHz easily before de-lid (now 4.5 GHz). Some of the very crappiest ones only do 4.2 GHz, but that's still something like 98% of the performance you'd get at 4.3 GHz.

It's also impossible to know how this guy actually "pre-tests" his chips. Just 1 hour of Prime95? 48 hours of Aida64? Linpack 11 AVX2? Better to test and validate the OC yourself. It's in his interest to pass the CPUs so he doesn't lose money.
 
Last edited:

Idontcare

Elite Member
Oct 10, 1999
21,110
64
91
I would also say that a 4670k at €200-ish which does 4.3 would be overall better than an overpriced i3570, even if it may do 5ghz...taking into account that Hasswell on avg. is 10% faster anyway. Agree?

Yep. Get Haswell, consider delidding yourself (or not).
 

tolis626

Senior member
Aug 25, 2013
399
0
76
Yep. Get Haswell, consider delidding yourself (or not).

Seeing as this is a topic that concerns me and that you're one of THE guys to ask about delidding,how hard would you consider it to be on a scale of 1 to 10?
On topic,I'd definitely go for a new 4670K or 4770K.You can't know what the guy tests,how he tests it and if he can be trusted at all.Even if you somehow mess it all up,you can blame it on yourself and your bad luck,not some random guy on the internet.Believe me,it's better this way.Plus,you never know,you may hit a "golden" chip that can do 4.6+ on air.Delid it too and you'll be happy.Although I guess you could also be happy with a "crappy" one that "only" does 4.2.I would.
 

BrightCandle

Diamond Member
Mar 15, 2007
4,762
0
76
I don't think we ever really know what the likelihood of getting a particular overclock. We have a pretty good impression that the 4770k doesn't appear to do worse than 4.2Ghz for everyone but there is a lot of reporting bias in the numbers. If you can't overclock the chip well you don't publish the result, hence the poor results are typically under reported. Those that get great overclocks on the other hand are far more likely to publish how awesome it is and are thus over represented. So we can't really say that even 4.2 is guaranteed because we just don't know what the range of results really is.

Further to that the conditions you run your CPU in will be different to any test another party does. You may very well have a motherboard that isn't quite as good or higher thermals that make it unstable because you live in a warmer place or with more heating on. You might not be able to use the same cooling mechanism either (he might be using custom water loop with 12 fans).

But the level of testing and the type of stability achieved is really important if you are considering a binned CPU where the binning isn't done by Intel. A lot of review sites in the past have shown crazy overclocks with dangerous voltages and on chips specially selected by Intel/AMD for them. That doesn't mean what they did was 24/7 truly stable, because the quality of testing they will have done is very limited. So I would be quite wary and critical in asking about the precise testing procedure.