©ASUSTEK COMPUTER INC. ALL RIGHTS RESERVED.

Xeonnemesis · ‎12-21-2015

System: (Current Settings)
Core I7 6700k
Asus Hero 8 Bios 1102 (currently)
16gb Gskill DDR4 F4-3200C16D-16GVK 2133mhz
Asus Strix 980ti OC 359.06
Samsung 850 Pro SSD
Seasonic 860W PSU
Windows 10 Home 64bit

I’m using HWIN64 v5.10-2700 for my temps/voltage readings

Backstory:
Sytem was built 9/29/15. Everything appeared to work as intended using the latest drivers & bios (0802) at the time . After several days of running on auto settings to confirm stability, I started playing with an OC to see what the setup could do. I wa s able to get 4.6ghz @ 1.285v using the XMP profile/manual voltage mode and the core multiplier set to 46. Temps were never a problem, 68-70c under stress testing. The system ran in this configuration for 5 weeks with no issues. Suddenly and for some unknown to me reason, I now had issues.

Realbench 2.41 would crash or fail to run various tests, failing randomly on any one of the four segments. Witcher 3 would sometimes crash in a loop where I could log in for approximately 1 minute and then crash to the desktop. Windows programs would randomly hang but no blue screens. At this point I thought that maybe my OS was corrupted so I reinstalled Windows 7, problems continued during stability testing. I took this oppurtunity to take the free upgrade to Win 10 to see if that makes any different, however the issues I’m having do not appear to be specific to either OS. I was now getting occasional BSOD’s in Win10 along with some random freezing. The more I attempted to tweek things, the more freezing I would experience. At one point the system would hang anytime I opened a browser window regardless of flavor (Chrome/FF/Edge). I ended up reloading the OS and setting all bios settings to their defaults.

Note that while I had removed my OC to troubleshoot this issue, I was still using manual voltage to prevent high spikes in vcore. From day one the mobo appeared to provide too much voltage even on stock defaults for my liking and causing unnecessary heat. My goal with this system has been cool & quiet.

Current Situation:
At this point, the only way I can keep the system stabile is for everything in the BIOS to be set to stock defaults. Any changes I make end up crashing RealBench and in some cases causing a BSOD, most commonly for a “page fault in nonpaged area wof.sys” error. If I make any changes to the voltage/multiplier/XMP settings, the system will freeze or blue screen. I contacted Asus to see about RMA’ing my mobo and they recommended rolling back my bios to 1102 as I was running the latest 1202. The bios change does not appear to affect my results, I still can’t change any of the default voltages or use xmp without freezes/crashes. Windows Event Viewer does not appear to have any useful information other than I am seeing some ACPI errors (ACPI device 5) around the time of the BSOD.

As it stands today, with all settings on default and a fresh 10 OS, here are the tests I was able to run in the default settings without failing/crashing:
• HCL – 14.5gb pass for one hour
• Memtest86+ one hour
• Intel Burn In – one hour pass
• Realbench 2.41 – Full stress test / 16bg one hour pass

This would indicate that there is nothing glaringly wrong with the hardware itself though the auto voltages have me concerned since the VID will jump to around to a max of 1.43-1.44v while the vcore will max around 1.44v. This is at the stock 4.2 turbo clock and balanced windows power settings.The VID voltages seem to drop under load during handbrake staying under 1.3v and at times I’m seeing a delta of .05v between the vcore/VID readings. I’m dumbfounded that the chip is getting so much more votlage than when I had my OC in place at 4.6 running under 1.3v.

I have spent much of the last 6 weeks trying to figure this out to no avail. I’m at a loss with nothing to really point at directly since everything appears to work if left on defaults. Kind of defeats the purpose of both a K series cpu and a ROG mobo if I can’t tweek anything.

I have been following these threads as well but there are no solutions as of this posting.
http://www.tomshardware.com/forum/id-2830772/skylake-build-randomly-freezing-crashing.html

https://rog.asus.com/forum/showthread.php?77522-Maximus-VIII-Hero-Freezeing-(ACPI-error-)

Does anyone have any idea as to what I should do next? I don’t have any big box stores close to me to try and swap parts so if I do any parts swapping on my own, I’ll have to pay for restocking fees if it does not resolve the issue.

Any info/comments are appreciated and I apologize if I left out any pertinent information.

Thanks,

J

Nate152 · ‎12-21-2015

Hello Xeonnemesis

It's likely your cpu needs a little more vcore for 4.6GHz or your ram is unstable. With the vcore set to auto your cpu will get over volted some so it's good to set it manually like you are.

Reset the bios to defaults and try a little more vcore, once you're sure it's 100% stable then start working with the ram. It could be one or the other this is why I suggest to do the cpu first.

ROG Strix 1070 OC Guide - ROG Spatha X Rapid Fire - ROG Falchion Touch Panel - z270 BCLK Overclocking - Xsplit Gamecaster Guide - ROG RX 460 OC Guide

Xeonnemesis · ‎12-21-2015

Thank you for your reply!

So what I've seen so far when adjusting the cpu voltage in the bios is that I either need nearly .100 + in offset mode or 1.45 in manual mode just to keep the system stable at it's stock clocks. This seems insanely high for stock settings, but maybe I'm wrong and it's ok for the chip. I have read not to take this chip over 1.35v on air, but my temps are never an issue and always remain below 70c (25c at idle w/ cstep @ 800mhz). The issue is that 1.35v just doesn't seem to be enough voltage, assuming that the readings I'm getting are correct.

To be honest, this is my first intel chipset since the Prescott days and maybe I don't currently have enough understanding as to how the Z170 platform should be working to properly troubleshoot this.

My typical testing process involves making a change in the bios and then booting into windows and running RealBench for the benchmark. If it passes then I consider it good enough to keep testing. In most cases, RealBench will hang or crash so then it's back to the bios to try something else or reset the defaults. In theory it could be possible that RB is not the best application for testing in this situation and I'm open to any suggestions folks may have to better confirm system stability.

Thanks again for your time.

J

Chino · ‎12-21-2015

The voltage spiking issue was solved from BIOS 11xx and onwards. So you shouldn't have any problems on the 1202 BIOS. As for using Realbench as a stability testing tool, I find it requires more voltage to achieve stability. Also there is a new Realbench version updated for the Skylake platform. Make sure you're using the latest.

Xeonnemesis · ‎12-21-2015

Chino wrote:
The voltage spiking issue was solved from BIOS 11xx and onwards. So you shouldn't have any problems on the 1202 BIOS. As for using Realbench as a stability testing tool, I find it requires more voltage to achieve stability. Also there is a new Realbench version updated for the Skylake platform. Make sure you're using the latest.

Thanks Chino. I was using RB 2.41 and just managed to download the RB 2.42 today. Which one were you referencing for better Skylake compatibility?

J

NemesisChild · ‎12-21-2015

Xeonnemesis wrote:
Thanks Chino. I was using RB 2.41 and just managed to download the RB 2.42 today. Which one were you referencing for better Skylake compatibility?

J

RB was recently updated to 2.42 to accommodate Z170/Skylake.

Intel i9 10850K@ 5.3GHz
ASUS ROG Strix Z490-E
Corsair H115i Pro XT
G.Skill TridentZ@ 3600MHz CL14 2x16GB
EVGA RTX 3090 Ti FWT3 Ultra
OS: WD Black SN850 1TB NVMe M.2
Storage: WD Blue SN550 2TB NVMe M.2
EVGA SuperNova 1200 P2
ASUS ROG Strix Helios GX601

Xeonnemesis · ‎12-21-2015

I just installed and attempted to run RB 2.42 in benchtest mode.

First thing I noticed as that it parsed my system info correctly for the first time:
CPU:Intel Core i7 6700K - Speed:803.5 MHz - Multi:8.0 - Bus:100.4 MHz
RAM:1071.4 MHz (1:16) - 16GB - 15-15-15-35-2T
MB:MAXIMUS VIII HERO - FW:1102 - Desktop
OS:Windows 10 Home
GPU1:NVIDIA GeForce GTX 980 Ti

The image editing failed out with a gimp.exe error, but I just hit enter to close the dialog boxes and keep moving through the steps. The Handrake section cleared successfully, VID's were around 1.270v. It then went into the OpenCL and I noticed my VID's jumped to 1.375v for approximately 1 minute before the system BSOD with a " irql not less or equal" error.

Is it expected that the core VID would increase significantly more during OpenCL vs Handbrake?

I'll run another RB session here shortly and see if I get different results.

Thanks for help everyone!
J

NemesisChild · ‎12-21-2015

You should not attempt CPU overclocking with the CPU voltage set to Auto.
1.45v should not be needed for stock clocks, so something else in your system is likely unstable.
RB is actually one of the best stress tests available since it's uses open source programs.

Which memory slots do you have you modules installed in?
Are you using XMP profile or manual memory timings in the bios?
Have you run memtest on each stick individualy?

Intel i9 10850K@ 5.3GHz
ASUS ROG Strix Z490-E
Corsair H115i Pro XT
G.Skill TridentZ@ 3600MHz CL14 2x16GB
EVGA RTX 3090 Ti FWT3 Ultra
OS: WD Black SN850 1TB NVMe M.2
Storage: WD Blue SN550 2TB NVMe M.2
EVGA SuperNova 1200 P2
ASUS ROG Strix Helios GX601

Xeonnemesis · ‎12-21-2015

NemesisChild wrote:
You should not attempt CPU overclocking with the CPU voltage set to Auto.
1.45v should not be needed for stock clocks, so something else in your system is likely unstable.
RB is actually one of the best stress tests available since it's uses open source programs.

Which memory slots do you have you modules installed in?
Are you using XMP profile or manual memory timings in the bios?
Have you run memtest on each stick individualy?

Thanks NemsisChild.

Currently I am running all settings in default/auto, including both RAM at 2133mhz and core ratio at Auto (4.2ghz). No overclocks at all as I'm still trying for stability. I should not need anything above 1.25v for stock clocks since I had a very solid OC at 1.285v when I first built the system. 1.45v really freaks me out but with the default bios settings, that's where it ends up as max voltage, though not usually sustained voltage, just spikes. Typical sustained voltage under testing is anywhere from 1.295 to 1.375v depending on the test and how the system is feeling at the moment. Still even 1.375v on stock clocks seems way too high.

As for the RAM, currently I am not running an XMP profile at all so it's just using the auto settings/timings (1.2v/2133mhz). As stated in the manual, I am using DIMM_A2 and DIMM_B2. When testing the modules, I only used DIMM_A2 and ram memtest for approximately 30 minutes each. When using HCL, I ram both sticks together (A2&B2) and opened 8 instances of HCL at 1812mb each in order to use up as much of the system memory as possible while not affecting the remaining 1.5gb that the system was using. All tests passed when I ran it for one hour, no errors.

J

Chino · ‎12-21-2015

Xeonnemesis wrote:

Currently I am running all settings in default/auto, including both RAM at 2133mhz and core ratio at Auto (4.2ghz). No overclocks at all as I'm still trying for stability. I should not need anything above 1.25v for stock clocks since I had a very solid OC at 1.285v when I first built the system. 1.45v really freaks me out but with the default bios settings, that's where it ends up as max voltage, though not usually sustained voltage, just spikes. Typical sustained voltage under testing is anywhere from 1.295 to 1.375v depending on the test and how the system is feeling at the moment. Still even 1.375v on stock clocks seems way too high.
J

If you find that your voltage is still spiking despite using the newest BIOS, input the stock voltage manually instead of leaving it on Auto. That should solve the spiking problem.

Asus Maximus Hero 8 / 6700k – Random Freezes/BSOD /RealBench Crashes