FEATURED Intel Haswell i7 4770K & i5 4670K Results Thread!

[MattNo5ss]

Yep, just enter your info into the forms in the original post and I'll update the spreadsheet.

 

[Rob94hawk]

So is the 5.1 ghz @ 1.3v legit?

 

[hokiealumnus]

They never responded, so I'm going with NO and it has been deleted. Thanks for the reminder.

 

[MattNo5ss]

Uploading this pic correctly...

View attachment 130096

Warning: If any submissions are made without images being hosted at overclockers.com, then those submissions will just be deleted.

 

[blazenarrow]

pretty good chip

I think I have a pretty good 4670k.

Using asrock z87 extreme6 with 2.1 bios I was able to boot up in windows @ 4.8 with 1.25 VCORE, and started super pi 32M but crashed.

4.7 was ultra stable.

The reason why I am not going to do this 24/7 is because the asrock sucks with adaptive voltages and so if I wanted to use 1.25 vcore it would not throttle voltage with cpu clock. So I am at 4.4 24/7 OC with max of 1.23 volts.

 

[Janus67]

I think I have a pretty good 4670k.

Using asrock z87 extreme6 with 2.1 bios I was able to boot up in windows @ 4.8 with 1.25 VCORE, and started super pi 32M but crashed.

4.7 was ultra stable.

The reason why I am not going to do this 24/7 is because the asrock sucks with adaptive voltages and so if I wanted to use 1.25 vcore it would not throttle voltage with cpu clock. So I am at 4.4 24/7 OC with max of 1.23 volts.

Ultra stable using what as a stability check?

SuperPI32M only stresses a single thread

 

[blazenarrow]

Ultra stable using what as a stability check?

SuperPI32M only stresses a single thread

I was just curious to get a quick feel to the chips OC ability while I was playing CS:GO, and so at the time I didn't run anything like Prime95. But it was stable for a while playing that until I tried 4.8, then decided to just go back to 4.4 with the ability to decrease the voltage.

 

[Janus67]

Fair enough, I would just hesitate to use the term 'ultra stable' when not using anything remotely close to a stability test

 

[blazenarrow]

Fair enough, I would just hesitate to use the term 'ultra stable' when not using anything remotely close to a stability test

I agree, I guess I was just so used to Pentium 4 & Core2duo days where it was either stable, or obviously not stable REAL quick.

Anyway, here is a little further stability testing. I mean I know it's only been several minutes but it's at least a stability test.

 

[harney]

I am still on with my under-volt project so far i have pulled it down to 1.128 @4.2 managed to pull down to 1.120 but still testing @4.2......i hit a barrier @ 4.3 wants over 1.15 to be stable but @4.2 with only 1.128 is crazy

 

[EarthDog]

Harney... what is your CPU's stock voltage? I havent seen (out of 3 samples, tiny I know) one of those CPU's above 1.05v stock speeds.

 

[harney]

Hi Earthdog

Flat 1.128v in bios

 

[EarthDog]

So all the values given are bios values? Ok. What are the values in windows? Specifically the LOAD voltage? That is what I would use as what is set in the bios, most of the time does not match what is in windows, note.

 

[Eldonko]

So if I'm interpreting your graph correctly the avg oc for a 4770K is 4.6? Is there a SS for the 5G result?

 

[MattNo5ss]

Yeah, 4611MHz is the mean. That first graph is kinda hard to understand If you go to the far right of the "Form Responses" tab, you can see the Mean, StdDev, and the Mean ± StdDev range.

 

[hokiealumnus]

I'll give a cookie to anyone that can graph standard deviation in google docs. That's the only reason I graphed it the way I did.

I suppose we should remove all results without screenshots since there's no proof.

Leading by example.

4.8 / 1.3V



4.9 / 1.33V



I didn't submit the second one yet. Frankly, I had forgotten I did it.

 

[Eldonko]

Should probably have some minimum stability standards to get more accurate data? Like 30 mins of aida or p95 or something like that. Otherwise you have people submitting 5ghz results with no ss that are more than likely 1m stable lol.

 

[Janus67]

Should probably have some minimum stability standards to get more accurate data? Like 30 mins of aida or p95 or something like that. Otherwise you have people submitting 5ghz results with no ss that are more than likely 1m stable lol.

Agreed on that. At least 30 minutes imo.

 

[Heater]

Heres something, it doesnt show AIDA but you should be able to tell by the temps. IF this doesnt work let me know and will get a 30 min stable SS with AIDA showing. Its not my 24/7 clock so i would have to redo it

 

[MattNo5ss]

It would be nice to see how long the test has been running in the screenshot as well. Thanks for responding so quickly

EDIT: I've PM'd the guys that made submissions before we started the screenshot requirement, so that we have the chance to update their submission.