Random Crashes (hard lock)

Man, all of this reminds of being a Windows Technician back in the 90's and how difficult it was to explain to some people that the problem wasn't with Windows it was with some aspect of their hardware or hardware drivers.

Regards,
SB
I did run sfc /scannow and it found and repaired errors. Everything seems fine so I hope whatever this was didn't bork my OS.
 
Well it crashed again. It went so long being perfectly stable that I'm not confident I'll be able to troubleshoot this. If I have to wait days to find out if whatever I changed fixed it, I could be at this for months. No tests that I run show any problems at all. Hours and hours of memtest and no errors, though it did freeze once during the test so I know it must be some kind of hardware problem (Windows wasn't even loaded). For now I pulled out one of the RAM sticks.

Does anyone know of any suite of hardware tests that might help? I've never had to deal with a demon quite like this.
 
Well it crashed again. It went so long being perfectly stable that I'm not confident I'll be able to troubleshoot this. If I have to wait days to find out if whatever I changed fixed it, I could be at this for months. No tests that I run show any problems at all. Hours and hours of memtest and no errors, though it did freeze once during the test so I know it must be some kind of hardware problem (Windows wasn't even loaded). For now I pulled out one of the RAM sticks.

Does anyone know of any suite of hardware tests that might help? I've never had to deal with a demon quite like this.

There are some things that are very hard to test for. For example, failing VRMs or capacitors (MB, video card, PSU, etc.). You've also mentioned that the Seasonic you are using is heavily used, so it's possible that while it's still generally providing clean power that it is occasionally allowing either insufficient voltage or "dirty" power through.

If you haven't done so yet (I can't remember) you can try setting everything to stock (so, no overclocks even "factory" overclocks). There's always a potentially for a slightly less than stable driver or TSR (old terminology) but something that is resident in memory that's fucking things up after a certain amount of time or after it's been doing stuff for a certain amount of time.

Try to keep an eye on whether it does it after doing a certain set of actions in Windows. For example, a program or series of programs or a certain time after a program has been run, etc. Intermittent locks after a long period of time (days/weeks) can be extremely difficult to track down.

Regards,
SB
 
There are some things that are very hard to test for. For example, failing VRMs or capacitors (MB, video card, PSU, etc.). You've also mentioned that the Seasonic you are using is heavily used, so it's possible that while it's still generally providing clean power that it is occasionally allowing either insufficient voltage or "dirty" power through.

If you haven't done so yet (I can't remember) you can try setting everything to stock (so, no overclocks even "factory" overclocks). There's always a potentially for a slightly less than stable driver or TSR (old terminology) but something that is resident in memory that's fucking things up after a certain amount of time or after it's been doing stuff for a certain amount of time.

Try to keep an eye on whether it does it after doing a certain set of actions in Windows. For example, a program or series of programs or a certain time after a program has been run, etc. Intermittent locks after a long period of time (days/weeks) can be extremely difficult to track down.

Regards,
SB
Unfortunately it seems to crash even when Windows isn't loaded (locked up during memory test) so I don't think it's a Windows/driver issue. I also think it's highly unlikely that 2 PSUs, one brand new and the other that is known to work fine are failing at the same time in the same way. I don't have much experience with bad RAM, but I feel like if that were the problem it would at least produce some error at some point during all the testing I've done. But maybe not.

So it could be memory, CPU, motherboard, or GPU (maybe?). Pretty confident it's not the SSD or the PSU.

If it crashes with this stick of RAM removed, I will swap it back in for the other stick. If it crashes again I'll put my GTX970 in there. As much as it sucks, RMAing the GPU is the least painful option. RMAing the CPU, RAM, or mobo will mean I'm out of a computer for a while. And if it's the CPU, RAM, or mobo I doubt I'll be able to tell which of these components is the problem unless I RMA them 1 at a time until it stops happening. So then I'm out of a computer almost indefinitely while each of these components is shipped back and forth. This truly sucks.
 
Last edited:
I also think it's highly unlikely that 2 PSUs, one brand new and the other that is known to work fine are failing at the same time in the same way.

Unlikely, but it happens. I know since it happened to me. My backup PSU about 15-17 years ago had some capacitors go bad just enough while it was sitting on a shelf for it to become periodically unstable. It's why I try to keep a lightly used PSU as a backup now, but one that has been used at least long enough to know it isn't one that will fail in the first month or two due to faulty components.

So it could be memory, CPU, motherboard, or GPU (maybe?). Pretty confident it's not the SSD or the PSU.

If it crashes with this stick of RAM removed, I will swap it back in for the other stick. If it crashes again I'll put my GTX970 in there. As much as it sucks, RMAing the GPU is the least painful option. RMAing the CPU, RAM, or mobo will mean I'm out of a computer for a while. And if it's the CPU, RAM, or mobo I doubt I'll be able to tell which of these components is the problem unless I RMA them 1 at a time until it stops happening. So then I'm out of a computer almost indefinitely while each of these components is shipped back and forth. This truly sucks.

At least you have a plan of attack, it's all about attempting to narrow down the possibilities. Not everyone can keep stuff around to swap in, I try, but don't always have backups (like if I have to invest in a new MB platform). Thankfully AM4 lasted long enough that I have both a 1600x and a 3700x as backup as well as an x370 MB.

Regards,
SB
 
Unlikely, but it happens. I know since it happened to me. My backup PSU about 15-17 years ago had some capacitors go bad just enough while it was sitting on a shelf for it to become periodically unstable. It's why I try to keep a lightly used PSU as a backup now, but one that has been used at least long enough to know it isn't one that will fail in the first month or two due to faulty components.



At least you have a plan of attack, it's all about attempting to narrow down the possibilities. Not everyone can keep stuff around to swap in, I try, but don't always have backups (like if I have to invest in a new MB platform). Thankfully AM4 lasted long enough that I have both a 1600x and a 3700x as backup as well as an x370 MB.

Regards,
SB
Now I kind of wish I'd gone with DDR4. I have tons of that laying around. I'd never even seen a DDR5 stick until I built this computer.

Right now my solution is I'm setting up a timer that goes off every 5 minutes reminding me to save my game in Valheim. Losing 20+ minutes of progress from a crash is devastating.
 
Last edited:
I’ve swapped in the GTX970 since nothing I do with the RAM makes any difference.
 
Not sure you have tried these suggestions,

look in your Device Manager and determine if some components/devices are not attached or have errors. Look at each device tab(s) for issues/errors.
are you using all "default" motherboard settings with no manual changes?
try using only your OS ssd and detach others to see is crash occurs?
 
Not sure you have tried these suggestions,

look in your Device Manager and determine if some components/devices are not attached or have errors. Look at each device tab(s) for issues/errors.
are you using all "default" motherboard settings with no manual changes?
try using only your OS ssd and detach others to see is crash occurs?
Yea all the drivers are good, no unknown devices. It was starting to crash pretty often, but not since I swapped the GPU. Again I'm just gonna wait a few days before I make any claims about having fixed it. I was suspect of the GPU early on. I would get random AMD driver crashes. I didn't think much of it, but this has turned into these lock ups and has been happening more and more frequently.
 
If you suspect is the GPU you might want to try Display Driver Uninstaller (DDU) to make sure all driver components are removed from the system, and then try the AMD card again and reinstall the GPU driver.
 
How’s your house power? Any brownouts? Do yoh have an especially sensitive light bulb that’ll dim or flicker with lower house voltage? i have some cheap LEDs that’ll dim when the circuit is loaded. Easier than using a multimeter.

It’s more likely a component is at fault but who knows.
 
Last edited:
If you suspect is the GPU you might want to try Display Driver Uninstaller (DDU) to make sure all driver components are removed from the system, and then try the AMD card again and reinstall the GPU driver.
I tried all that before. No difference. Since putting in the GTX970 and installing the latest NVIDIA driver I've had no crashes.
How’s your house power? Any brownouts? Do yoh have an especially sensitive light bulb that’ll dim or flicker with lower house voltage? i have some cheap LEDs that’ll dim when the circuit is loaded. Easier than using a multimeter.

It’s more likely a component is at fault but who knows.
The power is clean and reliable. System is on a UPS just in case.

I will RMA the 6700XT tomorrow assuming I don't get any more crashes. I've been gaming like crazy these last 2 days and it's been all good.

The GTX970 actually runs Valheim just fine. Clearly I was CPU limited on my old 3770K.
 
I've RMAed the GPU. No idea how long it will take. They said they would repair it and send it back, but I really hope they send me a different card. Not confident they'll find anything wrong with my card since it appears to work fine for upwards of a few hours at a time.
 
Keep in mind that the 6700 XT is a more power intensive GPU so it could still be the PSU.

I'll be crossing my fingers that it was the Video card, however. :)

Regards,
SB
 
Keep in mind that the 6700 XT is a more power intensive GPU so it could still be the PSU.

I'll be crossing my fingers that it was the Video card, however. :)

Regards,
SB
Yea I'm just worried they'll test it for an hour and decide there's nothing wrong with it. Every time I ship it back to them it's on my dime. Lol.
 
I finally got the card back after 2 weeks, and they seem to have done exactly what I feared they would do. In the service log they checked it out for a couple hours and then shipped it back, marked as "reparied". I can't imagine they actually fixed anything. Gonna find out tonight.

This really blows because the card was essentially DOA, but I stupidly decided to troubleshoot on my own until it was past the Newegg return window. Now if I have to send it back to MSI again I will be back on the GTX970 for another 2 weeks (they ship fedex ground). And I'll also have to pay for shipping to them again.

I was thinking I should've gone NVIDIA, but I don't know of any NVIDIA options that come with a reasonable amount of VRAM for under $600 except for the 3060 which is not very good.
 
Last edited:
It started crashing again last night and it's now worse than it was before. I went ahead and ordered a 4070 that will be here Sunday.

I spoke with MSI and the customer support guy told me they replaced some component on the board, so it is legitimately defective. They're sending me a shipping label so at least I won't have to pay for that again.

My AMD experiment failed so I'm back to NVIDIA again but once the 6700XT gets back from MSI in a couple weeks I'll give it to my dad. He recently got into DCS world. I just hope they actually replace the card this time. It was DOA but since I spent some weeks troubleshooting, once I'd determined that the GPU was the problem I could no longer return it to Newegg. Feels like I got punished for trying. I explained all this to MSI in my original RMA and asked if they could just replace it but they didn't give a fuck and sent me back the same broken card anyway. Don't see myself buying anything from them again.
 
I want to be clear that I don't blame AMD for this. Shit happens and I got unlucky. I'd even be okay with MSI if they'd actually fixed the (brand new) card. Ultimately I can't be bothered to go back and forth in the RMA process for months like this. Really I don't know who can, especially if you don't have a spare GPU or money to buy another one.

Edit to avoid my first quintuple post: The 4070 is tremendously faster than the 6700XT. I wasn't dissappointed with the 6700XT, and the brief look I took at 4070 benchmarks had me tempering my expectations, but for the games I'm playing the 4070 completely dominates. In Witcher 3 with raytracing it's like 3-4x faster and doesn't stutter at all. Went from being unplayable to silky smooth. In X4: Foundations I'm able to run 4xMSAA with no visible performance degredation which was impossible on the 6700XT. I'd viewed this as more of a side-grade going into it with minor perforance improvement but I am pleasantly surprised by how much faster it is. Very nice for what was essentially an impulsive rage-purchase.
 
Last edited:
I want to be clear that I don't blame AMD for this. Shit happens and I got unlucky. I'd even be okay with MSI if they'd actually fixed the (brand new) card. Ultimately I can't be bothered to go back and forth in the RMA process for months like this. Really I don't know who can, especially if you don't have a spare GPU or money to buy another one.

Edit to avoid my first quintuple post: The 4070 is tremendously faster than the 6700XT. I wasn't dissappointed with the 6700XT, and the brief look I took at 4070 benchmarks had me tempering my expectations, but for the games I'm playing the 4070 completely dominates. In Witcher 3 with raytracing it's like 3-4x faster and doesn't stutter at all. Went from being unplayable to silky smooth. In X4: Foundations I'm able to run 4xMSAA with no visible performance degredation which was impossible on the 6700XT. I'd viewed this as more of a side-grade going into it with minor perforance improvement but I am pleasantly surprised by how much faster it is. Very nice for what was essentially an impulsive rage-purchase.

You to try hose path traced games that were unplayable on AMD đź‘€ ;)
 
Back
Top