BSOD whea_uncorrectable_error and clock_watchdog_timeout

[Tintai]

Hello,

After Cmos reset and clean Windows 10 install I'm started getting some BSODs.
First time I've got whea_uncorrectable_error and some time later clock_watchdog_timeout.

Now I'm getting clock_watchdog_timeout all the time when I try to start AIDA64 stability test and in the middle game of CS:GO.

Here is dump from bluescreenview:
http://tintaiexcess.com/dmp.htm

And here is the dump files from "C:\Windows\Minidump":
http://www.mediafire.com/download/q2oy1mc5ku7girx/080315-7468-01.dmp
http://www.mediafire.com/download/ap25qa4ilr4gtc0/080315-7359-01.dmp

My CPU is running at 4.6G at 1.328V. RAM at 1866Mhz at 1.5V.
Stock looks stable.

Thanks
Tintai

 

[erocker]

So you don't get BSOD's when your CPU is stock? If so, time to re-do your overclock.

 

[Tintai]

Yeah it looks good at stock.
Can you tell my is that hardware issue or software? I don't know how to read from dump files.

Is that time for new CPU?

Also clock at stock looks weird. "3771Mhz" and idle "1605". I don't remember how it was before I started OC. It was a long time ago. At 4.6G it was "4.600" and "1600". Both at bus 100Mhz.

 

[Ferrum Master]

Lack of CPU voltage... up it up... or twiddle with vdroop settings, make them give more voltage at load.

 

[Toothless]

Might have to back off your OC a bit..

 

[Tintai]

One more thing.
Problems starts when I did a clean install Win10. Before I was running at Windows 8.1 and upgrade to 10. There was no issues.
After format and install W10 BSODs starts appears. So I don't believe that is hardware failure.

So I'm 90% sure that is OS issue but I can't back to 8.1 now.

Can someone check my dmps files? Also I will try at 4.2G.

 

[INSTG8R]

Well I run 1,35V for my 4.6 so don't be afraid to give it a little more.

 

[Steevo]

We saw the same issues when we all moved to X64 Vista, the additional functions on the CPU that are untested by almost all benchmarking and stability testing software are used and reveal minor hardware issues with otherwise "stable" overclocks.

 

[Tintai]

I'm not. It's not problem with hardware.
I've opened that dumps files and there is a line:

DEFAULT_BUCKET_ID:  WIN8_DRIVER_FAULT

But I don't have drivers for Win8. I'm running at Win10.
Someone can explain?

 

[Steevo]

I disagree, it is hardware, windows 10 is using more of the advanced instruction abilities to speed up the OS than Windows 8, just like when we took X86 stable overclocks into X64 and they failed, more data is being moved and used and the chances for failure are significantly higher, and the rounding that every processor does didn't add up in X64.

So as previously suggested, add more volts, or back down the overclock.


Windows 8, and 7 drivers work in Windows 10 as the kernel is very similar, and MS is updating hardware with whatever "stable" WHQL they have for the piece of hardware based off its Hardware ID.

I am fairly certain without even looking that the memory address referenced in your dump files will be very low, and may even be in the processor cache with a pointer to the RAM address.


Yep, your second BSOD referenced "3594f" which is in HEX, and correlates to "219,471" which is the number in Kb from 0 that was in use, so a "101" and "124"BSOD with a very low memory address to me smells exactly like more core voltage needed, or a lower overclock.

 

[Tintai]

Thanks for advice I will test at 1.35V.

@edit:
Still BSOD at 1.35V. I'm trying Auto now.

I have one question anyway. When first time I've got Win10 it was upgrade from Win8.1 and there was no problems with OC. Troubles started when I did a clean install. Can you explain this?

 

[brandonwh64]

This is the SAME thing that happened to me after windows 10 install and even one of the SAME BSOD's you got. I was rock stable with 7/8/8.1 but the night I do the windows install I started getting instabilities. I did a full CMOS default and it seems to be working good on stock settings.

 

[Tintai]

Ok so it's stable at 1.4V but it's to much. I'll try 1.370V.

 

[Ferrum Master]

I am feeding 1.45V on my 4.8GHz, and temps are fine... So chill...

These systems are living their last showtime...

But honestly... I think... Reading the anniversary Pentium problems, Windows is gone Linux style, it holds very fresh CPU microcodes and loads them independently, it is well known that 28th version of intel CPU microcode needed less voltage for stable clocks for Sandy Bridge, and the latter ones need more... I can bet the dog is burried there...

 

[Tintai]

For SB 1.4V it's little to much for me. I was running at 1.330V at 4.6G and now 1.3V can't even hold 4.6G at 1.360V for a second.
However I set 4.4G at 1.320V and seems stable. 200Mhz it's not a big deal for me.

 

[eidairaman1]

drop your overclock to stock and run it, W10 has some teething to get through

Hello,

After Cmos reset and clean Windows 10 install I'm started getting some BSODs.
First time I've got whea_uncorrectable_error and some time later clock_watchdog_timeout.

Now I'm getting clock_watchdog_timeout all the time when I try to start AIDA64 stability test and in the middle game of CS:GO.

Here is dump from bluescreenview:
http://tintaiexcess.com/dmp.htm

And here is the dump files from "C:\Windows\Minidump":
http://www.mediafire.com/download/q2oy1mc5ku7girx/080315-7468-01.dmp
http://www.mediafire.com/download/ap25qa4ilr4gtc0/080315-7359-01.dmp

My CPU is running at 4.6G at 1.328V. RAM at 1866Mhz at 1.5V.
Stock looks stable.

Thanks
Tintai