i7 6700k OC Advice

Toggle sidebar Toggle sidebar
N

nromonoski

Commendable
Jul 28, 2016
6
0
1,510
Could someone help me get an idea these are normal overclocking results?

I'm using an i7 6700k, Gigabyte Z170x-Gaming 6 motherboard, and Kraken X61 cooler. While setting the CPU core ratio to 45 (4.5GHz) and vcore to 1.34v the system seems to be stable and reaching max temps of around 80'C while stress testing with x264, Intel XTU, Cinebench, and Handbrake encodes. Idle temps are around 38'C with 'Performance Mode' set on the cooler.

I feel like I've seen a lot of cases where people are achieving greater frequencies, at lower voltages, and lower temperatures. Did I just get a bad chip, or is my cooler not working very well? My GTX 1070 is even idling around 48'C. Something just seems a little off with my results.

Here's some additional info that may be worth noting...
- vcore LLC = 'Auto'
- XMP is enabled to give a memory frequency of 3000MHz
- I live in Phoenix, and its the middle of summer, so my ambient temperatures are kind of high (26'C)
 
Solution
MeanMachine41
Sometimes running two monitors together can cause erroneous readings with some MBs.
You can run the other tests in AIDA64 for DIMMs and GPU and extend the test periods for say 30mins. readings only with AIDA64 without HWMonitor.
Also try Realbench v2.43 as an alternative tester, If you pass that test and benchmark, your system is stable.
Sort by date Sort by votes
MeanMachine41

MeanMachine41

Judicious
May 8, 2014
5,195
336
33,390
Hi nromonoski :)

Can you not achieve stability with Vcore set to 1.25V. ? Every small increase in voltage has an exponential increase in temperature. Not all Intel CPUs perform the same however try to lower the Core voltage as 1.34V is a little high.
Stress testing can reveal deficiency's in voltage, (Vdroop)
Load Line Calibration LLC helps when on high however watch your temperatures.
I use HWMonitor and AIDA64 trial version for testing and monitoring.
Your Kracken X61 is a good unit and should cope or your TIM may need replacing.

Temps at idle should be 10-15C above ambient and 60-65C under load is ideal.

Also list your PSU as a stable good quality and efficient power supply is important.
 
Upvote 0 Downvote
N

nromonoski

Commendable
Jul 28, 2016
6
0
1,510


Thanks for the reply MeanMachine! To answer your PSU question, I have a EVGA 850W G2.

Update on the OC... Dropped the score to 1.25v and turned LLC on high like you suggested, which booted the comp, but crashed in stress tests. So I bumped the vcore to 1.26v (LLC still on high) and it seems stable. I was able to run all of the stress tests I listed and the temps are slightly cooler. Now I'm looking at around 36'C idle and 77'C max under load, although I do notice that there are random spikes in CPU usage and temperatures with just the monitors open (using NZXT CAM and CPUID HWMonitor).

Does this sound better? I'm glad I was able to drop the vcore, but is it ok to keep LLC on high? HWMonitor shows a near constant vcore of 1.248v-1.260v, but the VID was anywhere in the range of 1.232v to 1.432v. If 1.34v was high the 1.432v spikes make me nervous. Also, I would have thought the temperatures would be lower by dropping the voltage so much.
 
Upvote 0 Downvote
MeanMachine41

MeanMachine41

Judicious
May 8, 2014
5,195
336
33,390
That's a decent PSU you have and should not be a problem.

I would say you have reached the limit of your OC nromonoski with that chip. Further testing will tell.

Drop your frequency a little to 4.4GHz at 1.25V, LLC stay on high and run the test again. I want to see if its more stable and there is little or no difference whilst gaming at the slightly lower frequency.

For testing use HWMonitor and AIDA64 trial version and put them side by side on your desktop.

Run the test for CPU and FPU for ten mins and observe HWMonitor with all system voltages showing.
Take a screen shot at 10min mark and upload the pics to IMIGUR, then obtain the BB url and link it here.
Stop the test if temps climb to 80C.

I await the result.
 
Upvote 0 Downvote
N

nromonoski

Commendable
Jul 28, 2016
6
0
1,510
I ran CPU and FPU at the same time. FPU gave similar temps on its own so I figured it was ok to run both simultaneously.

Here are screen shots of 4.4GHz/1.25vcore/LLC High around 10 mins into the test and one at idle.
(I also ran the original tests I was using and the results were similar. Max 79'C.)

Fi0Euoo.png

4gGBpj5.png


I dont understand how the temps from all the vcore voltages I've used are pretty much the same. The 4.5GHz/1.34v gave me about 80'C under load as well.
 
Upvote 0 Downvote
MeanMachine41

MeanMachine41

Judicious
May 8, 2014
5,195
336
33,390
Your temperatures are OK under load at 4.4GHz and Vcore @ 1.248V.

Look at Voltages on all rails during the test. The readings indicate: 3V reading 2.040v, 5V reading 3.367V and 12V reading 0.048V.
The test indicates a serious problem with your PSU as all rails are out of acceptable limits.
If the readings are correct, I'm surprised the system did not fail the test as this would cause serious instability.

Get the unit tested under load or Swap the unit out with a known working unit of the same Wattage to determine this.
Normally EVGA gold certified units are a good PSU and come with a ten year Warranty so your covered.
 
Upvote 0 Downvote
N

nromonoski

Commendable
Jul 28, 2016
6
0
1,510
Oh wow! That's not good. Thanks for catching that! Using a voltmeter is probably the best way to test it, rather than using software right? Also, do you think you could post a screen shot of yours for reference? I don't have another PSU to use and compare to so I'm not sure how much variation there's supposed to be while its in use.
 
Upvote 0 Downvote
MeanMachine41

MeanMachine41

Judicious
May 8, 2014
5,195
336
33,390


Using a Multimeter is not the best idea without the system being under load.
Here are my HWMonitor readings:
NJuFIaf.jpg


Best to Swap the unit out.
 
Upvote 0 Downvote
N

nromonoski

Commendable
Jul 28, 2016
6
0
1,510
MeanMachine, I looked at the voltages listed in the BIOS and re-ran AIDA64 to check the voltages and got the following. These voltages seem to be more in line with what you'd expect. I guess HWMonitor must be reading the voltages wrong?
dtheoQT.jpg

H6y9aAM.png


 
Upvote 0 Downvote
MeanMachine41

MeanMachine41

Judicious
May 8, 2014
5,195
336
33,390
Sometimes running two monitors together can cause erroneous readings with some MBs.
You can run the other tests in AIDA64 for DIMMs and GPU and extend the test periods for say 30mins. readings only with AIDA64 without HWMonitor.
Also try Realbench v2.43 as an alternative tester, If you pass that test and benchmark, your system is stable.
 
Upvote 0 Downvote
Solution
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom