Question Intermittent and Random Lockups (No BSOD)

[Brainonska511]

My 4-year-old Skylake rig in my sig will randomly lockup (screen goes blank or just turns completely green; nothing in the Windows event log beyond unexpected shutdown), maybe once a month, and requires a hard reset to fix. There seems to be no reason behind why - it's not overheating (I can't recall a time where it locked up while gaming; only while browsing the web), and I don't think there are any other problems. I also cannot reproduce this problem at all on my own. Am I right to assume that the likely culprit is the PSU?

 

[VirtualLarry]

Well, Skylake is a bit more than 3 years old at this point, more like 5-6.

Is this using the iGPU, or a discrete GPU? I have, or had, before my all-core OC, the same problem with my R5 3600 rig w/ RX 5700 reference.

Then again, I'm mining.

 

[Brainonska511]

Well, Skylake is a bit more than 3 years old at this point, more like 5-6.

Is this using the iGPU, or a discrete GPU? I have, or had, before my all-core OC, the same problem with my R5 3600 rig w/ RX 5700 reference.

Then again, I'm mining.

Haha, my bad. It is almost 4 years old now that I look back (built April 2016). It's using a discrete GPU (R9 380X). It's honestly had this problem since the day I built it - it's just so random and relatively rare that I never bothered to look into it till now. All the parts were new, never used it for mining.

 

[Steltek]

What model of PSU do you have, and how old is it? Also, what make and model of system drive are you running?

What does the Windows Reliability History look like (In Action Center, under Maintenance)? Does it show anything failing there?

Also, have you run SFC /scannow from an elevated command prompt to check for damaged system files? A damaged file system can also cause lockups, so running CHKDSK /f on your system drive couldn't hurt as well.

If nothing else works, try re-installing your motherboard chipset and USB drivers. I have an old Sabertooth Z77 system that I have to do this on every 9 months or so, otherwise it gets very flaky and starts having all sorts of problems.

 

[Brainonska511]

What model of PSU do you have, and how old is it? Also, what make and model of system drive are you running?

What does the Windows Reliability History look like (In Action Center, under Maintenance)? Does it show anything failing there?

Also, have you run SFC /scannow from an elevated command prompt to check for damaged system files? A damaged file system can also cause lockups, so running CHKDSK /f on your system drive couldn't hurt as well.

If nothing else works, try re-installing your motherboard chipset and USB drivers. I have an old Sabertooth Z77 system that I have to do this on every 9 months or so, otherwise it gets very flaky and starts having all sorts of problems.

Everything is from April 2016. PSU is an XFX TS 550W. The system drive is a Samsung 850 EVO, and the Samsung Magician software doesn't show any issues.

There's been nothing in the Windows reliability center. Scan disc has never found any errors.

As for the drivers, I could try that, but I've definitely installed newer ones over the years with no change in results.

 

[VirtualLarry]

There's been nothing in the Windows reliability center.

nothing in the Windows event log beyond unexpected shutdown)

This seems slightly contradictory. If there are unsafe Windows Shutdowns, they should be showing up in Reliability Monitor.

Does this happen only at idle? Only at load? Or randomly, either case?

I might consider, using a vcore positive offset voltage, if your board supports that. Either +0.050V (50mV) or +0.100V (100mV). Check your max / load temps after doing so, as you might cause your CPU to run hotter. (Well, almost certainly.) Maybe that's all it needs.

If that doesn't "fix" the issue, then we can delve further.

Or maybe try swapping the PSU, maybe yours was just flaky from the start.

You have an older and somewhat power-hungry AMD GPU. Have you kept your drivers continuously updated over the last few years? Any change in behavior noticed with updated drivers?

As for the drivers, I could try that, but I've definitely installed newer ones over the years with no change in results.

Maybe try removing the GPU, and running off of the iGPU, or possible dropping in a less power-hungry GPU, maybe a GTX 1650 or something not from AMD, and see if the problem still happens. If it goes away, the problem might have been your GPU. (Or PSU marginality.)

 

[Brainonska511]

This seems slightly contradictory. If there are unsafe Windows Shutdowns, they should be showing up in Reliability Monitor.

I guess I should say, they show up, but they're totally useless information:

"The previous system shutdown at 8:55:24 AM on ‎2/‎18/‎2020 was unexpected."

When I've looked in the event manager as well for other critical errors, they're always about as useless as that. No memory dumps available, no error codes to check...

And there are no "problem reports" in the reliability and problem history window.

Does this happen only at idle? Only at load? Or randomly, either case?

If I step away from my computer for a few minutes, I've never come back to it being frozen. The only time I can recall it freezing is under light use: web browsing, maybe some music playing, but my wife tells me that it's also frozen once or twice when she would use it to play video games (before I built her a machine at Christmas). But it is also very random and overall, very rare - I can go for weeks without this issue popping up.

I might consider, using a vcore positive offset voltage, if your board supports that. Either +0.050V (50mV) or +0.100V (100mV). Check your max / load temps after doing so, as you might cause your CPU to run hotter. (Well, almost certainly.) Maybe that's all it needs.

I think my motherboard had some power saving setting that was off by default, but when turned on, it would cause freezes within minutes, consistently, so I had kept that at the default setting. But I'll have to look into the vcore positive offset voltage: everything right now is at the defaults for my Gigabyte H170N-WiFi board. I don't think I'm running the latest BIOS (F22e), but I have updated that in the past, with no difference in this intermittent issue. edit: turns out I am running the latest bios.

Or maybe try swapping the PSU, maybe yours was just flaky from the start.

Yeah. I'm keeping an eye out for a decent PSU to drop in price. I don't have a spare lying around at the moment to swap out. I do have a PSU tester, but I've yet to go into the case and actually check, since it would be a time sink to undue a lot of the cable management. My long-term goal is to pick up a decent PSU and swap it, let it run for a few weeks to see if there are any changes, and simultaneously use the PSU tester on the XFX supply. If it comes back bad, I believe it's still under its 5-year warranty.

You have an older and somewhat power-hungry AMD GPU. Have you kept your drivers continuously updated over the last few years? Any change in behavior noticed with updated drivers?

Maybe try removing the GPU, and running off of the iGPU, or possible dropping in a less power-hungry GPU, maybe a GTX 1650 or something not from AMD, and see if the problem still happens. If it goes away, the problem might have been your GPU. (Or PSU marginality.)

I have kept the drivers up-to-date. No changes in the system's behavior. I've also never seen any visual indications that the graphics card might have problems (eg, in-game artifacts). Another long-term goal was to swap out the GPU, as part of a "renewal" to squeeze more life out of this desktop.

The hardest part is because it's so intermittent, I'd need to run it for weeks before I'd know if 1 thing actually solved the problem.

Mainly, I've just been fishing for ideas as to what to check. Thanks!

 

[VirtualLarry]

my Gigabyte H170N-WiFi board.

That rings a bell. Did you check the interwebs for reviews, see if anyone else is having intermittent lockups / hangs with that board?

 

[Torn Mind]

Hard Drive can screw things up with freezing. Run the system on the IGP to determine if the GPU is causing it.

 

[Brainonska511]

That rings a bell. Did you check the interwebs for reviews, see if anyone else is having intermittent lockups / hangs with that board?

I had checked around, but nothing ever came up beyond maybe something in some power settings of the BIOS, like ErP.

ErP was off, but I did just disable the dummy load feature in the power (which was set to auto).

 

[Meghan54]

Well, I don't believe that XFX ps was one of the better ps's on the market anyway.

But.....are you OC'ing the cpu? If so, I'd first return the cpu to stock settings and see if that helps fix it.

Then memtest the thing.

Unstable OC and a bad strip of memory are the two most common causes of random lockups from what I've seen over the years. Of course, it could be the power supply, but since the lockups are happening under light/low load, I'd wonder if the ps is actually the culprit....most times, power supply issues rear their ugly heads when under heavy loads when the ps cannot keep up with the output the computer is calling for under load.

 

[Brainonska511]

Well, I don't believe that XFX ps was one of the better ps's on the market anyway.

I remember checking at the time, it was a higher at the time (Tier 2 based on the giant list posted on Tom's Hardware back in 2016)

But.....are you OC'ing the cpu? If so, I'd first return the cpu to stock settings and see if that helps fix it.

Then memtest the thing.

Everything is stock. Previous runs of memtest have always shown no issues.

Unstable OC and a bad strip of memory are the two most common causes of random lockups from what I've seen over the years. Of course, it could be the power supply, but since the lockups are happening under light/low load, I'd wonder if the ps is actually the culprit....most times, power supply issues rear their ugly heads when under heavy loads when the ps cannot keep up with the output the computer is calling for under load.

Yeah, I thought it may be related to light load as it may trip some ground fault protection circuitry.

The only thing I've tried since the last lock-up is turning off "power loading" in the Gigabyte BIOS power options (Auto --> Off). Now, I just have to run it this way for a few weeks to see if it will still lock up...

 

[Brainonska511]

I thought I'd give an update to this saga:

I just replaced the XFX TS 550W PSU I was using with a new Corsair RM550x I picked up during BF. Prior to installation, my PSU tester confirmed the Corsair supply is perfectly in spec (ie, 12V rails were exactly 12.0V). Now to play the waiting game to see if I get any more lockups.

On a related note, after I got the XFX supply out, I tried that on the PSU tester - everything is technically within spec: 12.4V (-12V), 12.3V (+12V2), 12.3V (+12V1), 5.0V (+5V), 3.3V (+3.3V)