CPU Adaptive Mode core voltage

[Saptex]

So i have overclocked my i7-6700k skylake processor on an Asus Rog Maximus Ranger VIII 170 to a stable overclock at 4,635 MHz with 1,360 V ( I know it could be allot for that overclock, but I am currently in a "wiggle zone"). So after watching Linus' video about overclocking the skylake processor, he said that when the overclock was stable and good the CPU Core/Cache Voltage should be turned on Adaptive Mode. And so i did, with putting in the values of Additional Turbo Mode CPU Core Voltage to 1.360 V and the Offset Voltage to auto (not sure why I need to set a certain offset voltage since it would just take whatever voltage it needs to operate on idle). But then when I booted up Windows and opened up CPU-Z, the core voltage spiked up to 1.440 V. I did not like that... Anyone that has run into this thing and has a fix for it?

And one other thing that I'm wondering about is what the Offset Mode Sign is.

Thanks

[Masada02]

I can't exactly help with the issue at hand, but I can say that for Skylake 1.45v is perfectly fine for a 24/7 overclock. I want to say someone at Intel mentioned this at one of the technology shows or something when the CPUs were just being released.

[Enderman]

that is why you should not use adaptive mode

set the voltage manually

[Saptex]

Just now, Masada02 said:

I can't exactly help with the issue at hand, but I can say that for Skylake 1.45v is perfectly fine for a 24/7 overclock. I want to say someone at Intel mentioned this at one of the technology shows or something when the CPUs were just being released.

hmm... so okay, it's just that in the video Linus said that there's kind of a max voltage that the processor can handle (1.4 in the video) before the lifespan i effectively. Now i know Linus can make mistakes, but just to be sure.

[Saptex]

3 minutes ago, Enderman said:

that is why you should not use adaptive mode

set the voltage manually

But then the processor would get an unnecessary amount of voltage, wich would lead to more heat being produced.

[Enderman]

Just now, Saptex said:

But then the processor would get an unnecessary amount of voltage, wich would lead to more heat being produced.

no, you set the voltage manually at a setting that is not unnecessary

set it to where its just enough to keep the CPU stable

do you know how to OC properly?

[Patrick Bateman]

 try override mode..

i can use adaptive on my z97 without spikes like that. its auto that is mostly the problem

[Masada02]

1 minute ago, Saptex said:

hmm... so okay, it's just that in the video Linus said that there's kind of a max voltage that the processor can handle (1.4 in the video) before the lifespan i effectively. Now i know Linus can make mistakes, but just to be sure.

He's generally pretty conservative with voltage recommendations. He also said that 1.3 was the max safe for Haswell but that is definitely not the case. It all depends on your cooling really. If you have good cooling, you can use more voltage.

Here's some reference material:

http://www.hardocp.com/article/2015/08/05/intel_skylake_core_i76700k_ipc_overclocking_review/7#.VrEWmrIrLmE

http://club.myce.com/f184/intel-skylake-overclocking-guide-339700/

http://www.overclock.net/t/1570313/skylake-overclocking-guide-with-statistics

http://www.tweaktown.com/guides/7481/tweaktowns-ultimate-intel-skylake-overclocking-guide/index5.html

The Tweaktown page has some really good info btw.

[Saptex]

1 minute ago, Enderman said:

no, you set the voltage manually at a setting that is not unnecessary

set it to where its just enough to keep the CPU stable

do you know how to OC properly?

Well no as this is my first time actually trying to OC manually, but I just thought that because of what happens when things gets too much voltage than it uses.

[Enderman]

2 minutes ago, Saptex said:

Well no as this is my first time actually trying to OC manually, but I just thought that because of what happens when things gets too much voltage than it uses.

well adaptive voltage always overestimates the needed voltage by a ton, so manually setting the voltage is far better

[Saptex]

5 minutes ago, Masada02 said:

He's generally pretty conservative with voltage recommendations. He also said that 1.3 was the max safe for Haswell but that is definitely not the case. It all depends on your cooling really. If you have good cooling, you can use more voltage.

Here's some reference material:

http://www.hardocp.com/article/2015/08/05/intel_skylake_core_i76700k_ipc_overclocking_review/7#.VrEWmrIrLmE

http://club.myce.com/f184/intel-skylake-overclocking-guide-339700/

http://www.overclock.net/t/1570313/skylake-overclocking-guide-with-statistics

http://www.tweaktown.com/guides/7481/tweaktowns-ultimate-intel-skylake-overclocking-guide/index5.html

The Tweaktown page has some really good info btw.

Thanks, those were actually really helpfull. Thanks

[Saptex]

2 minutes ago, Enderman said:

well adaptive voltage always overestimates the needed voltage by a ton, so manually setting the voltage is far better

okay, I'll try running som tests to check the temperatures

[Saptex]

20 minutes ago, Enderman said:

well adaptive voltage always overestimates the needed voltage by a ton, so manually setting the voltage is far better

So as you were saying it actually ran alittle bit cooler when the computer ran all the start-up programs, since it runs at a much lower voltage. So thank you!

[ThomasD]

Generally speaking the idea behind switching from fixed to adaptive after stress testing is that whatever voltage was stable with fixed voltage should probably not be demanded under more normal operation while set at adaptive.

But this is not always the case.  the chip's own algorithms generally err on the side of guaranteeing stability, and so will ramp up voltages faster than might be necessary.

If normal operation causes the voltage to spike above that of what was proven stress-stable then you can either go with whatever fixed value you had (because it will still be lower than the adaptive) or you re-examine your overclock to see if it really was all-that-stable to begin with. 

[Kurtis444]

I had the same issue with my Asus z97m-plus while over clocking a 4790k. I found a stable voltage using manual voltage, then switched to adaptive and kept lowering the voltage and tested it in Windows until it matched. With manual I was stable at 1.265v, and with adaptive (if I remember correctly) I have it set at 1.00v with a negative offset of 0.005v to achieve the same max voltage.

Its a pain that you need to do a trial and error, but Linus had the same issue with his z97 oc guide. It's the one complaint that I have with the Asus boards.