i7 6700K random freezes in idle

[idata]

I have same issue with random freezes in idle as many others, here is my config:

Intel Core i7-6700K

Asus Maximus VIII Gene bios 1701

SSD Intel 750 with Asus Hyperkit

GeForce GTX 760

64GB RAM Kingston

Windows 10

The PC passes all validation test, memory is ok (2 days in memtest with no error), it is perfectly stable in all stress tests.

The random freezes occur daily when PC is idle, during browsing or video playback. I have 3 monitors, when freeze happens they display random color (like white on 1st, light blue on second, pink on third) and only hard reset is possible.

I checked voltages in bios and there are so many that I got completely confused. I tried to increase CPU_PLL voltage from auto to some value slightly higher than standard (which is what BIOS displays in hint).

I found in bios these items: CPU_PLL, Termination CPU_PLL, Internal CPU_PLL and Eventual CPU_PLL.

Increasing them changed behaviour - approximately every 5 minutes nvidia display driver crashed, monitors went to same state as during freeze - random colors, but then windows recovered the driver and the PC did not freeze.

I then increased the values a bit more and my windows 10 did no boot, so I loaded bios defaults and activated XMP memory profile.

Now I see much higher voltages in HWINFO than before. CPU_PLL voltage was raised above 1.2V (previously 0.978), another increased voltages are System Agent Voltage, VCCMP, CPU Standby voltage and may be some other.

After 1 hour the freeze happen again.

BTW HWINFO shows CPU PLLs OC as 0.584V, this one did not change and I think it is wrong value.

I built two other PCs with same memory and CPU, motherboard is Asus Z170M Plus, one with AMD graphics and Samsung SSD 950 Pro, other uses CPU graphics and Samsung NVMe SM951, both are perfectly stable.

Any suggestions?

[CCris2]

Hi @Marwyn,

Please do a bios reset first

Then set memory to XMP, and CPU Core/Cache Voltage to Adaptive mode.

Test and if you still have problems give a reply here. If you have blue screens please note them also.

[idata]

Hi

I already tried this, bios reset, XMP and also adaptive mode on core/cache voltages, with no effect.

However, I removed GeForce from the PC, I use now integrated graphics with 2 monitors only and it has been running already 36 hours with no issues.

I will wait several more days and then I plan to try it with AMD graphics card.

[CCris2]

Try disable integrated graphics, and use only the gtx 760 and see how it works.

If you use the integrated graphics + discrete graphics you will have to increase VCCSA, VCCIO and maybe PCH +1.0V voltage slightly.

If that 64 GB memory kit is 4x16 then you will have a big problem to find the right VCCSA and VCCIO voltage. VCCSA can go up to 1.3V, VCCIO up to 1.25V. Don't go straight to those voltages. Do small increments, till you find the stable voltages.

I would like to know what is the frequency of the memory kit. What is the model.

My suggestion is to increase first VCCSA maybe you will be lucky and get system stable.

[idata]

GTX 760 (DP + 2xDVI) and IGP disabled was my previous unstable state.

[idata]

Here is how it looks in frozen state:

[CCris2]

That frozen state looks like Image off, it seems that is a problem with gtx 760.

Please update the video driver to the last one from nvidia website:

http://www.nvidia.com/download/driverResults.aspx/103913/en-us NVIDIA DRIVERS GeForce Game Ready Driver WHQL

and if you have same problem try to swap video cards and check then.

[idata]

I am not sure what Image off means, it is crashed display driver, the very latest one from NVIDIA, but several previous versions exposed same behaviour.

As I wrote in first post, if I tweak some PLL voltages, it becomes very frequent, looks exactly like this, but instead of freeze windows are able to recover the display driver.

[CCris2]

That CPU PLL doesn't do much difference, Try increase VCCSA to 1.25V, and VCCIO to 1.15V

[idata]

Ok, I will try it and report back.

[idata]

Hi all,

Marwyn, I'm just following on this thread and, I was wondering if you were able to resolve this if there is any update. Regards,Hellen.

[idata]

Hi Hellen,

no, I have been quite busy since my last post and I need to finish my work first.

My PC has been perfectly stable since I removed GeForce from it.

I plan to try it out again this weekend.

Regards

Marwyn

[idata]

Hi Marwyn,

If there is something else I can do for you, please let me know. Regards,Hellen.

[idata]

Hi all,

Yesterday afternoon I put back the GeForce and increased VCCIO and VCCSA by the smallest possible steps.

Comparing current HWINFO values with previous screenshot I see these changes (old value : new value)

VTT 1.160V : 0.984V

VCCSA 1.128V : 1.152V

IMC 1.136V : 1.168V

VCCST_A 1.163V : 0.988V (same as VTT)

CPU PLL 1.209V : 1.228V

VCCMP 1.138V : 1.169V (same as IMC)

I am not sure why some values are smaller, but so far the PC appears to be stable.

Fingers crossed.

[CCris2]

Hi Marwyn,

At least now you know where is the problem

[idata]

Hi Ronin,

Yes, it seems solved. Thank you for your help.

Best Regards

Marwyn

[idata]

Thanks, Ronin for the help on this case, great to know that it was helpful to you.

Regards,Hellen.

[idata]

Hi All,

I am back with some bad news. 5 days after I changed voltages my PC froze again. I then raised voltages several times but the only result was it started freezing more often.

Then I said enough and bought ASRock Z170 Extreme7+ mainboard. The issue is gone, I have uptime almost 2 weeks with default bios settings.

Regards

Marwyn

[CCris2]

Hi Marwyn ,

Now you have my mainboard, hope is all good with it .

[idata]

Hi Ronin

yes, this board is rock solid and because it has also 3 M.2 slots I bought 2 pieces of Samsung NVMe drives for raid 0.

3.2 GB/s read/write experience :-)

Regards

Marwyn

[CCris2]

Ehehe, i don't have what to say