Intel Haswell 4670k + 4770K Overclocking guide

[acceldrag]

Thermalright HR-22

[ProKoN]

Thermalright HR-22

 i would make sure its mounted proper

 

i prefer to spread my thermal paste on haswell cpus

 

is your fan cranking up to high speed under load?

 

can you add a second fan to do push- pull?

[ProKoN]

that cooler seems decent.

 

are you stress testing with xtu?

 

[acceldrag]

The paste is ideally spread through many tests and googling...

Proper mounting is one of my concerns...

I am dissapointed because i bought this specific cooler because it's kinda beasty and i really wanted to oc with no major probs,at least from the cooling system...

yeah it cranks up under load

yeah xtu

[ProKoN]

The paste is ideally spread through many tests and googling...

 

Proper mounting is one of my concers...

I am dissapointed because i bought this specific cooler because it's kinda beasty and i really wanted to oc with no major probs,at least from the cooling system...

 

yeah xtu

 

 

compare your coolers.

 

test each at 1.2V and see how much more heat is dissipated by the aftermarket cooler to stock.

[acceldrag]

I thought i would avoid that...to get everything out of the case but seems i have to...ok thanx guys

i will get back to you when i have more info on that.

[acceldrag]

One quick notice in case i won't be able to remove the cooler soon,

i stressed it with xtu and stock speeds-voltage and it showed 49-50 degrees...i guess we can compare that to other stock voltages and speeds with aftermarket coolers right?

The difference will be 2-8 degrees max among the coolers...

[DSD27]

 

i stressed it with xtu and stock speeds-voltage and it showed 49-50 degrees...

That is normal..

So you only had 93ºC when overclocking?

(sorry, I didnt pay attention to your temps yesterday, I only saw your multiplier and voltage :\)

 

Mine is 4.2GHz at 1.16V and goes to 65º under load. My cooler is True spirit 140

[acceldrag]

Yeah 93 with 4,4Ghz and 1,23 Vcore

While playing games i am around 65 max

[DSD27]

Yeah 93 with 4,4Ghz and 1,23 Vcore

While playing games i am around 65 max

thats really weird  :huh:

playing agmes with the cpu overclocked or at stock?

What kind of games?

[acceldrag]

I mean 65 while overclocked...

all kind of games, no matter what the max temp is 65, never over that...

[DSD27]

I mean 65 while overclocked...

all kind of games, no matter what the max temp is 65, never over that...

thats fine.

I dont know why you have 93 when stress testing... Are you using adaptive voltage?

 

test with Aida64

[acceldrag]

No i am using manual voltage

same thing happens with aida64.

[terrytek]

This should be pinned, for people to see how to OC it.

[JantsoP]

Might come in handy in the future.... Plz pinn already. Anyway bookmark thank you... 

[Gilbert_Sarip]

Thanks for the guide,
will try it  on mine,
will post results soon

[Ensaumneak]

Hey ProKon, is it okay if i go below 1.2 volts for 4.4 Ghz? This is my first time overclocking so im a bit in the dark. ATM im at 4.4 Ghz/1.2 V and stable.

 

[ProKoN]

 

Hey ProKon, is it okay if i go below 1.2 volts for 4.4 Ghz? This is my first time overclocking so im a bit in the dark. ATM im at 4.4 Ghz/1.2 V and stable.

 

 

 

yes it is ok to go below 1.2V

[Faceman]

@ProKoN I am wondering if the Extreme auto-tuning provided by Asus's AI Suite 3 on the Asus Z87-A is a good, safe, and stable overclock?

 

 

I have a i5-4670k and Hyper 212 EVO with just 1 fan, running at 1500rpm

 

Using the Extreme auto-tuning provided by Asus, it set my system to 4.7Ghz at 1.275v.

CPU Load-line Calibration = Level 8

CPU Current Capability = 140%

CPU Power Phase Control = Extreme

 

 

I stress tested by running Prime95 in blend for 20minutes, stable. Then Unigine Heaven @ max settings for another 40min just for good measure, total 1 hour.

 

The only slightly concerning thing to me was the temperatures.  I was getting as high as 84C, and might have continued to rise if I went longer than an hour, but typically it ran in the mid to high 70Cs.

[adriel604]

you are too vague about describing what you have tried . seems like your a bit all over the place. you really need to follow the guide step by step. if you are unwilling to post a screenshot identical to this one, i cant help you sorry.

 

 

Big noob here sorry in advance but how did you get 47ghz at an insanely low voltage of 0.968?

[ProKoN]

Big noob here sorry in advance but how did you get 47ghz at an insanely low voltage of 0.968?

cpuz voltage reading was wrong when i posted this screenshot.

 

the cpu was set for 1.41V. look at the xtu values on the right hand side of the screen.

[ProKoN]

@ProKoN I am wondering if the Extreme auto-tuning provided by Asus's AI Suite 3 on the Asus Z87-A is a good, safe, and stable overclock?

 

 

I have a i5-4670k and Hyper 212 EVO with just 1 fan, running at 1500rpm

 

Using the Extreme auto-tuning provided by Asus, it set my system to 4.7Ghz at 1.275v.

CPU Load-line Calibration = Level 8

CPU Current Capability = 140%

CPU Power Phase Control = Extreme

 

 

I stress tested by running Prime95 in blend for 20minutes, stable. Then Unigine Heaven @ max settings for another 40min just for good measure, total 1 hour.

 

The only slightly concerning thing to me was the temperatures.  I was getting as high as 84C, and might have continued to rise if I went longer than an hour, but typically it ran in the mid to high 70Cs.

I do not recommend using the ai suite automated OC tool. I tried it and had very poor results. However if it works for you by all means use it. if you start bluescreening all the time i would suggest doing things manually.

 

84c is ok as a max temp. remember p95 will overvolt your cpu if the vcore mode is set to "adaptive". make sure it is set for "manual" if stressing with prime95

[Faceman]

I do not recommend using the ai suite automated OC tool. I tried it and had very poor results. However if it works for you by all means use it. if you start bluescreening all the time i would suggest doing things manually.

 

84c is ok as a max temp. remember p95 will overvolt your cpu if the vcore mode is set to "adaptive". make sure it is set for "manual" if stressing with prime95

Yes, I did that.  I allowed the Asus suite to set the values itself, then I changed it from adaptive @ 1.275 to manual @1.275.  It was stable for me, I didn't run Prime for longer than an hour, but it worked, also worked just fine in these games: Skyrim, Assassin's Creed IV, Metro 2033, and Titanfall.

I am somewhat skeptical about the automated overclock though because the results I am getting seem to be well above what a manual OC is able to achieve.  4.7Ghz on just 1.275v??  Seems too good to be true.

 

I am not sure if it is a true 4.7Ghz either, the way it is shown is as:

1 core = 4.7

2 cores = 4.6

3 cores = 4.6

4 cores = 4.5

 

The CPU cache remains at 38

 

Another thing that surprised me, is the score I was getting in XTU when I ran the benchmark.  I was only getting 759 "marks", and when I compared my score with others universally, I couldn't make sense of it.  Some people were scoring in the 1000s, but only had 4.0Ghz OCs, so I am not sure if that benchmark score directly correlates to performance.  Reminds me of the old adage, you either play the benchmarks or you play the games.

[ProKoN]

Yes, I did that.  I allowed the Asus suite to set the values itself, then I changed it from adaptive @ 1.275 to manual @1.275.  It was stable for me, I didn't run Prime for longer than an hour, but it worked, also worked just fine in these games: Skyrim, Assassin's Creed IV, Metro 2033, and Titanfall.

I am somewhat skeptical about the automated overclock though because the results I am getting seem to be well above what a manual OC is able to achieve.  4.7Ghz on just 1.275v??  Seems too good to be true.

 

I am not sure if it is a true 4.7Ghz either, the way it is shown is as:

1 core = 4.7

2 cores = 4.6

3 cores = 4.6

4 cores = 4.5

 

The CPU cache remains at 38

 

Another thing that surprised me, is the score I was getting in XTU when I ran the benchmark.  I was only getting 759 "marks", and when I compared my score with others universally, I couldn't make sense of it.  Some people were scoring in the 1000s, but only had 4.0Ghz OCs, so I am not sure if that benchmark score directly correlates to performance.  Reminds me of the old adage, you either play the benchmarks or you play the games.

so you dont have a true 4.7GHz OC.

 

Asus per core overclock performs well but i prefer all my cores to run at the same speed

 

my 4670k runs at 4.7GHz 1.3V.

 

good luck at finding any decent ring bus stability with the asus board

 

i score low on the xtu benchmark I cant explain why.

 

here is the best place to compare OC performance. Well maintained and very popular thanks @Jumper118

 

https://docs.google.com/spreadsheet/ccc?key=0AlC81MjwelBgdEZNV3l6aHl1eUNwSUR4Rml0MXMzN1E&usp=sharing#gid=0

 

http://linustechtips.com/main/topic/62476-post-your-cinebench-r15-scores/

[Faceman]

so you dont have a true 4.7GHz OC.

 

Asus per core overclock performs well but i prefer all my cores to run at the same speed

 

my 4670k runs at 4.7GHz 1.3V.

 

good luck at finding any decent ring bus stability with the asus board

 

i score low on the xtu benchmark I cant explain why.

 

here is the best place to compare OC performance. Well maintained and very popular thanks @Jumper118

 

https://docs.google.com/spreadsheet/ccc?key=0AlC81MjwelBgdEZNV3l6aHl1eUNwSUR4Rml0MXMzN1E&usp=sharing#gid=0

 

http://linustechtips.com/main/topic/62476-post-your-cinebench-r15-scores/

My mistake, I quoted you the wrong speeds. This is how it really is:

 

1 Core = 4.7

2 Cores = 4.7

3 Cores = 4.6

4 Cores = 4.6

 

Same 1.275v

 

When I am not running a stress test, should I leave the voltage as adaptive? or leave at a manual 1.275?

 

     Thank you again for starting this discussion, giving us a very thorough, detailed, and organized guide for us novices to conduct our "experiments"  I am new to overclocking, and just starting to get a grasp of how everything fits in place together, your guide is by far the most comprehensive and explained resource I have come across.  I did attempt your method of going through each individual variable one by one, as all scientific experiments should be conducted, but my patience wore thin, and I was getting too many blue screens.

     I know most people advise against the automated OC, but as far as I can tell, my results are pretty good when compared to others using the guide you provided.  I have yet to blue screen with the automated OC, and am noticing increased responsiveness in my games.  For others out there, if you lack the patience to do it properly, because the auto-OC is not the proper way, merely an an option.  It is often labeled as a "Do not even attempt" but it works for me.  As always, every chip, and every configuration is different, but for me, the auto-OC is working and I am pleased with my results.  I wanted to share because while reading this entire thread, I gathered that you @ProKoN enjoy hearing all opinions, circumstances, situations, and stories so that your wealth of knowledge on the subject can grow.  Here is one situation where the auto-OC has worked, and worked very well.  If at any point I encounter some problems, I will retract my statements about the success I have had with the automated Asus Suite III Extreme Tuning Overclock.