Stabalising I7 4770k

[SamyBoy13]

Hi there can cpu input voltage stabilize the overclock as I need 1.13v core voltage for 4.2ghz with temps 80 Celsius degree under prime to get it stable and on 1.1v I get blue screen after a while?

[g8ina]

Moved to a more appropriate area.

[SamyBoy13]

Thank you

[Gronich26]

Here is a reasonable overclocking guide which helped me.

http://www.overclockersclub.com/guid...l_4770k_guide/

As 1.13v sound like the default voltage for your chip, you should try increasing the voltage to around 1.2v. I would only start worrying about your Prime temperatures if you are consistantly hitting 90o+ You will generally be hitting a lot lower if you are gaming.

[Spreadie]

I know Haswell runs hot but 80 degrees for 4.2GHz seems too bloody hot to me. What cooler are you using?

[Kovoet]

That seems a bit to hot. What cooler and what case. Mines running at between 60 and 67.

[Gronich26]

Now that I see, from another thread, you are using a Hydro H100i your 80oC+ temperatures are way to high for 1.13v@4.2Ghz.

[SamyBoy13]

I just re run OCCT in Linpack with AVX instructions at 3.5ghz with lowest voltage that was stable and at 4.2 with lowest voltage without blue screen I only put voltage to override (v core at 1.15 at 4.2ghz with all core synchronized I didnt change anything else) Here are screen shots of temps for both



Sorry for late reply and thanks for previous replies guys.

Also my fans at h100i are running at about 900rpm

[SamyBoy13]

Anybody help???

[Domestic_Ginger]

Raise the fan speed (900 is pretty slow)or reseat your cooler.
Run with the side panel off you case.
If you found the lowest voltage your CPU runs at with oc it would just be about improving cooling.

[excalibur2]

Well 80 degrees seems hot to me as well and will only get worse on a hot summer's day, although if not into heavy gaming (or whatever that stresses a CPU) then the cpu will run cooler and you would have no problem......did you use e.g. arctic silver thinly between the cooler and chip ?

[SamyBoy13]

I uses mx-4 thermal paste. I will try to run it without side panel. Also I dont think this is issue with thermal paste application as I have re sited my cpu quiet few times and temps don't seem to be dropping. Usually -+2 degree which I would say it's measurement error. Also I have re run the stress test while running my fans on h100i at 1056 rpm and temperatures seems the same although it's hotter now by 7 Celsius degree than when I was testing previously

[excalibur2]

Well I know you are trying to get everything stable but according to this site (as posted) http://www.overclockersclub.com/guid...0k_guide/3.htm for 4.2ghz your CPU temp should be appx 68 degrees and they are using CPU-Z

[SamyBoy13]

I tried their settings and I tried to tweak them but no luck above 1.16v I get my cpu throttling down and at 1.14v i get to 87 low 90's Celsius degree under prime. I think I just have a bad chip. I will re-seat the cpu with different thermal paste tomorrow as I will ned o go to my friends house who have silver 5 or something like that. Thank you for your support but I have no idea what to do.

[excalibur2]

Well your mX4 is good stuff and from sites comparing, there is not much difference between them all...I've been using Arctic silver II for about 10 years and all I do is smear a thin layer and with no problems in the CPU getting too hot when overclocking...but maybe I've been lucky with CPUs that work on lower volts.
For interest for myself I dropped my 2500k to 4.2ghz and was getting max 71 degrees on test with a cheap cooler i.e. Freezer 7 PRO Rev.2 h'mm which has got me thinking in that I should check the temp when my grandson is playing games at 4.4ghz.
Anyway let us know if you have solved the problem.

I wasn't sure how much a game stresses a CPU so looked up this site and found it interesting, all about overclocking a CPU and playing games:- http://blog.szynalski.com/2012/11/25...verclocked-pc/

[Bonebreaker777]

Instead of using Prime95 for testing (which test you were using? Blend,Small FFTs or Large FFTs), did you consider using game based benchmarks for stability testing?