Stability issues on Windows 10 - frequent BSOD

[Violet_Shift]

I'm putting this in hardware because I'm almost certain the problem is something hardware related.

Over the past few months my computer has had random stability issues - often blue screens of death when I'm doing nothing particularly odd.

For the longest time I ignored it - it wasn't a particular concern to me.

However, for my work I now have to do long (24 hour plus) program runs (I do computational biology, and I'm using my home PC because I'm still unable to return to the office on account of corona), and I've found it impossible to actually have one of these run to completion. I'm beginning to suspect something is wrong with my computer.

Often, my code will crash out with memory allocation errors, but more often the entire system will go down with it and I'll get a WHEA_UNCORRECTABLE_ERROR or SYSTEM_THREAD_EXCEPTION_NOT_HANDLED BSOD, which is exactly the same kind I was getting when I was getting random BSOD's when gaming or browsing or whatever.

I've ran windows memory tester, and I will run TPU memtest64 the next time I'm not actually sitting at the PC, but I'm worried it could be something else - replacing my RAM would probably be the least painful hardware problem to deal with.

At the moment I'd just appreciate advice for other things to test if the culprit isn't the memory. I'd blame my own code if this wasn't happening when gaming or running Firefox or any other amount of mundane PC tasks.

My specs are as follows -

i9-9900k stock
16GB G.Skill Trident Z RAM - DDR4-3600 CL16
MSI Tomahawk Z390 motherboard
RTX 2080 Super GPU (although this shouldn't matter)

 

[Shrek]

Latest BIOS?

https://www.msi.com/Motherboard/support/MAG-Z390-TOMAHAWK

How to Fix the "SYSTEM THREAD EXCEPTION NOT HANDLED" BSOD Stop Code in Windows 10 (makeuseof.com)
"The "SYSTEM THREAD EXCEPTION NOT HANDLED" is a common Windows 10 error that occurs mainly due to an outdated or incompatible driver."

 

[Bill_Bright]

Before spending money on RAM or anything else, you need to verify power. Since everything inside the case depends on good, clean stable power, I recommend you swap in a known good PSU of ample horsepower to see if the problem persists.

RTX 2080 Super GPU (although this shouldn't matter)

Well, remember, graphics cards typically are the most power hungry devices in our computers, or at least a close second behind the CPU. So yeah, it certainly can matter.

Over the past few months...

But was it rock stable before then? Any other changes around the time this problem surfaced?

What are your temps?

 

[Violet_Shift]

Thanks for the help. I suppose I jumped to blaming memory first because of the WHEA_UNCORRECTABLE_ERROR crashes.

I'll flash the motherboard BIOS soon - but before that, the other issues...

Before spending money on RAM or anything else, you need to verify power. Since everything inside the case depends on good, clean stable power, I recommend you swap in a known good PSU of ample horsepower to see if the problem persists.

I have a Corsair TX650 that I bought at the same time as this system and unfortunately I don't have access to another PSU right now (most of my computer crap is on the other side of the planet back home). I'll try removing the graphics card and seeing if that helps my stability at all since as you said, they're power hungry.

But was it rock stable before then? Any other changes around the time this problem surfaced?

What are your temps?

I wouldn't say rock stable, but the stability issues have been steadily getting worse, and they've been getting worse much more quickly in the past week.

At the moment, CPU temperature is 47 degrees and GPU temp is 52 degrees.

Under serious load, it will go up to about 60-70 degrees CPU temp, which rapidly returns to 50 degrees after I ease the load. I've cleaned all of my filters on my fans just in case that's a problem.

 

[DeathtoGnomes]

what does the Windows Event Viewer say about the error?

 

[Violet_Shift]

what does the Windows Event Viewer say about the error?

I can only find the "critical - kernel power" errors. No more details - I can't find any search results for WHEA_UNCORRECTABLE or SYSTEM_THREAD_EXCEPTION when that's exactly what popped up on the BSOD.

 

[Tetras]

I'd put the RAM back to stock (2666 @ 1.2v or lower), if you haven't tried that already.

Do you have any old devices/drivers?

 

[Violet_Shift]

I'd put the RAM back to stock (2666 @ 1.2v or lower), if you haven't tried that already.

Do you have any old devices/drivers?

I've not changed ram voltages at all, I wasn't aware this was a stock overclock. I might try that if nothing else works.

This entire PC is new (well, 2019 new) and as far as I'm aware up to date on drivers.

 

[GreiverBlade]

I can only find the "critical - kernel power" errors. No more details - I can't find any search results for WHEA_UNCORRECTABLE or SYSTEM_THREAD_EXCEPTION when that's exactly what popped up on the BSOD.

"If you receive a blue screen error (also known as a stop code), it means that your PC has shut down suddenly to protect itself from data loss. If you see the text “WHEA_UNCORRECTABLE_ERROR,” it means that a hardware error has occurred."

Intel "step by step"
To fix it, try the following:

• Get the latest updates with Windows Update. Go to Settings --> Update & security --> Windows Update, and then select Check for updates. Reboot system.
• If that does not work, try the followings:
  • Set the BIOS to its default values. Reboot system.
  • Start Windows in Safe Mode, and check if there are any driver errors in the device manager.
  • Check if there are any hardware issues.
    • Try to boot with minimal hardware settings (for example, one DRAM, less hardware components on your system).
    • Try the Windows 10 app called Windows Memory Diagnostic to check for memory issues.
    • Test the CPU by using the Intel® Processor Diagnostic Tool which can be downloaded here. For more information about the tool, see The Intel® Processor Diagnostic Tool.

"System Thread Exception Not Handled is a Windows error which comes with the Blue Screen of Death (BSoD). The Blue Screen of Death normally means that the system was not able to handle anything critical and Windows crashed to keep the hardware and software safe."

well there is countless result about these 2 codes ...

for my side i had a lot of BSOD lately too most were Kernel Power 41 and a few hardware related errors ... (Intel system too ... 6600K Seasonic M12 II 750w 4x4 HyperX Predator 2400) since i was fed up with Intel (messing the OC via microcode update, which luckily was solved for some but sadly not for me) i decided to change Mobo/CPU/PSU (MSI Tomahawk B550/R5 3600/Proton 1000w) initially after a few "memory management" BSOD decided to change the RAM too (although the memory issue was related to the CPU i got to replace the 6600K ) but since then i got 0 BSOD


my take would be PSU, tbf for me just the PSU swap would probably have been sufficient ...but since i was not sure, and genuinely hoping it would not be my aging 1070 and luckily it was not the case...
TX series PSU aren't bad at all, so was my M12 II but it was a 750W and my 1070 is nowhere near that 2080, i took a PSU in the wattage range a PSU calculator gave me with all the variable (including powered on at least 8hrs a day ) the 1000w is overkill (specially with a 65w 3600) nonetheless (well it was the cheapest in stock, even cheaper than a RM850M) ofc my overkill splurge on hardware do not make me sure, that, if i had only changed the PSU... everything would've been fine after all ID 41 can also mean duplicate driver (unlikely) third party AV soft (very unlikely ... WinDefender is enough) power settings (unlikely, never changed them ) Win fast boot (unlikel... yaddah yaddah) recent update in HW (errrr ... aside one SSD added ... i can't see anything else) or faulty HW (aye, sure was .... )

 

[Tetras]

I've not changed ram voltages at all, I wasn't aware this was a stock overclock. I might try that if nothing else works.

This entire PC is new (well, 2019 new) and as far as I'm aware up to date on drivers.

Then I would assume enabling XMP changed the RAM voltage, because there's no chance it would run 3600 stable at stock voltage (1.2v). Usually the system would fail to boot entirely.

The max stock mem frequency for a 9900k is 2666.

Devices: I was referring more to things like sound cards, USB peripherals, wireless adapters, etc.

 

[Violet_Shift]

It turns out that XMP had overvolted to just over 1.3V - still, not enough to cause alarm. I disabled both XMP and Game Boost, but we'll see if this has any impact.

 

[Tetras]

It turns out that XMP had overvolted to just over 1.3V - still, not enough to cause alarm. I disabled both XMP and Game Boost, but we'll see if this has any impact.

If XMP is enabled (or you manually set any frequency higher than 2666) then it should almost certainly be at least 1.35v. I'd check the specs of your modules, but I'd be very surprised if they were less than 1.35v.

If it was really running 3600 Mhz at only 1.3v then it could definitely lead to instability.

 

[Violet_Shift]

Since disabling XMP and Game Boost I've had no instability. I'll keep an eye on it but I think that was the culprit. Thanks for helping me find it, everyone.

 

[eidairaman1]

I'm putting this in hardware because I'm almost certain the problem is something hardware related.

Over the past few months my computer has had random stability issues - often blue screens of death when I'm doing nothing particularly odd.

For the longest time I ignored it - it wasn't a particular concern to me.

However, for my work I now have to do long (24 hour plus) program runs (I do computational biology, and I'm using my home PC because I'm still unable to return to the office on account of corona), and I've found it impossible to actually have one of these run to completion. I'm beginning to suspect something is wrong with my computer.

Often, my code will crash out with memory allocation errors, but more often the entire system will go down with it and I'll get a WHEA_UNCORRECTABLE_ERROR or SYSTEM_THREAD_EXCEPTION_NOT_HANDLED BSOD, which is exactly the same kind I was getting when I was getting random BSOD's when gaming or browsing or whatever.

I've ran windows memory tester, and I will run TPU memtest64 the next time I'm not actually sitting at the PC, but I'm worried it could be something else - replacing my RAM would probably be the least painful hardware problem to deal with.

At the moment I'd just appreciate advice for other things to test if the culprit isn't the memory. I'd blame my own code if this wasn't happening when gaming or running Firefox or any other amount of mundane PC tasks.

My specs are as follows -

i9-9900k stock
16GB G.Skill Trident Z RAM - DDR4-3600 CL16
MSI Tomahawk Z390 motherboard
RTX 2080 Super GPU (although this shouldn't matter)

Doesnt specify power supply of course