• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

6700k overclocking, how did I do?

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

Arthur_Dayne

New Member
Joined
Aug 4, 2016
Hey,

(2133MHz RAM) 4.5GHz - stable @ 1.260VCore
Cinebench avg - 980
Avg temp - 47C, max 50C

(2133MHz RAM) 4.6GHz - stable @ 1.300VCore
Cinebench avg - 1002
Avg temp - 57C, max 63C

(2133MHz RAM) 4.7GHz - stable @ 1.300VCore
Cinebench avg - 1015
Avg temp - 57C, max 64C

(2133MHz RAM) 4.8GHz - stable @ 1.380VCore
Cinebench avg - 1036
Avg temp - 61C, max 65C

4.9GHz isn't stable no matter the voltage. I tried asking for help on the overclocking subreddit but didn't get any useful feedback. I was hoping you guys could help me out a bit - I'm still new to this.

Cooler - Thermalright Macho Rev.B + Arctic MX4 (got a bad application, I'll be re-doing it sometime)
Motherboard - Gigabyte Z170X UD3
 
The voltages look in line with where you are sitting with your overclock. I am assuming you are trying to find a 24/7 stable overclock. If this is the case cinebench is not the best application to test with.

I personally like to use Intel Burn Test to do my initial testing and then when I am happy with my results I do a final test on Prime95 to verify the results.

As I said earlier your voltage and temperatures look ok though. I would run Prime 95 blend at your 48x multiplier and stated voltage to see what you get. If that passes I would look at trying to up your ram speed a little as well.

I would not push your voltage past 1.4v trying to stabalize the cpu for daily usage so if you cannot get 48x stabalize you may have to back down to 47x for daily usage. You will honestly not really see a difference in a day to day usage unless you are running a heavy cpu like constantly encoding videos.
 
Yeah. Not sure how you confirmed stability as all you talk about is running benchmarks with Cinebench. That tool is far from an adequate stress tester. You can complete Cinebench and not be anywhere close to stable.

I still use both Prime95 and IBT as Lochekey does but more and more overclockers are moving away from those kind of stress testing tools because their iterations are stacked in such a way as to generate so much heat that without custom water cooling you may not be able to run them safely on the high end CPUs like the 6700k. More people are beginning to move to AIDA64 Extreme and Intel Extreme Tuning Utility for stress testing because they spread those intense iterations out out in a more natural way so as to keep the temps down and imitate real world computing tasks. After all, what we really want is to test stability rather than the capacity of the cooling system. So depending on how good your cooling is passing two hours of Prime95 blend or IBT on "very high" settings should confirm stability or altermatively, running and passing AIDA64 Extreme or XTU overnight should give you some peace of mind.

Another option is using an older version of Prime95 like v.27.7 that did not yet have the AVX 2 instruction set and does not generate so much heat: http://www.mersenneforum.org/showthread.php?t=16779.
 
Last edited:
Scores seem a bit off though, i got 1045 @4700mhz and 1061 @4800mhz. Does memory speed have any impact on Cinebench ?
 
Scores seem a bit off though, i got 1045 @4700mhz and 1061 @4800mhz. Does memory speed have any impact on Cinebench ?

Yes, ram has a modest impact on the score of cinebench as well as uncore speed. I'm not sure if the op has tried pushing the uncore or just the core. Depending what else he has going on in the background of his system his score may be affected as well.

I would not place to much stock in his scores versus others without knowing more system details like Windows version, type and amount of ram, etc. His scores seem to be scaling Ok as his overclock ramps so I do not think he is unstable from that perspective.

Speaking of system details. Arthur, can you post your system specs similar too what you see in Trents and Kenrous signatures, it well help with knowing what is going on in your system.
 
Yes, ram has a modest impact on the score of cinebench as well as uncore speed. I'm not sure if the op has tried pushing the uncore or just the core. Depending what else he has going on in the background of his system his score may be affected as well.

I would not place to much stock in his scores versus others without knowing more system details like Windows version, type and amount of ram, etc.

I was thinking more along the lines of unstable overclock, that's a big difference Cinebench-wise :p
 
I thought I read that it was version 26.6 for prime 95. So I downloaded that and ran it for 3 hours today. Then again for another hour and a half after 30 min of aida64. Ended on 1.290v in bios for 4.5Ghz. Actual operating vcore is 1.296. LLC level 7 keeps my vcore between 1.296-1.312 under max during p95 version 26.6 I posted a bunch of photos in another thread I posted. I hope version 26.6 is a good enough version to test with. I max temp was 71c but average was about 66c.
 
Yes, that is something else you can do if cooling is limited, i.e., stressing with an earlier version of Prime95. I have sometimes used Prime95 v.27.7 I believe. The deal there is that those earlier Prime versions don't contain iterations that hook some of the newer CPU core family instructions sets (AVX II I think) that cause those high temp spikes. Shaves off 10-15c on max temps.
 
26.6 uses SSE2, 27.* uses AVX1 and 28.5+ uses AVX2/FMA3 (but you can use latest version and use a command line in local.txt to disable AVX*/FMA* testing). Many users/techs/guides continue to claim that if you only game/browse 26.6 will be more then enough and from personal testing the voltage/temp drops are quite impressive.
 
Re memory performance, it affects score but you can count on +/-50 points in total score going from auto/2133 to something like 3600. Sometimes differences are visible after 2-3 runs. Considering that after overclocking you get over 1000 then it's not a big difference. About the same difference is on quad channel platforms.

On some new motherboards there are sometimes options which let to automatically change voltage when you run apps with AVX. On ASRock X99 Taichi it's available when you leave power saving options enabled and use Broadwell-E CPU.
 
On some new motherboards there are sometimes options which let to automatically change voltage when you run apps with AVX. On ASRock X99 Taichi it's available when you leave power saving options enabled and use Broadwell-E CPU.
Do they say AVX enable voltage change?
 
Do they say AVX enable voltage change?

Sorry my mistake, mixed it with something else. Option on my board says:
"AVX2 Negative Offset will reduce core frequency while core using AVX2 instructions"
so not voltage but frequency ... it's appearing only when I use Broadwell-E CPU
 
Back