(What Would you do) Computer Constantly Crashes ... GPU ? CPU ?

[neko77025]

Ok .. So I am 100% confused at what is causing my issue. .. It down to Motherboard , CPU, GPU or Maybe PSU? ( I know , i know that almost everything)



So about 3-4 months ago My PC started to Randomly reset or Blue screen ( whea_uncorrectable_error ). Thing is , it was so rare it was hard to reproduce. I would do something and the issue would not happen for 3-4 days .. I would assume what I did Fixed it . And this has gone on for 4months now. (When I say its rare .. understand ... I am Disabled so I am on my computer everyday for 10 hours .. and it will happen once every 4-5 days.) (now it just resets , not seen A bluescreen / whea_uncorrectable_error sense windows reinstall)

I can play games for 4 -5 hours Str8 with out issues ... No artifacts nothing ... for 3,4,5 days ... Nothing goes wrong. (warthunder, Cyperpunk, rimworld, Rocketleauge, Pantfinder)

First off, Did not think it was GPU, temps never got hot .. never had Artifacts. And games would run fine for 3 hours. (every time I have had gpu issues , this is nothing like it ). So dont seem like GPU issue.
Assumed it was drivers or ram .. maybe windows.

again .. would to something .. it would stop .. assumed it was fixxed ...

Ran all the ram test ... Move them to arranged them in ever configuration with 2 or 4 sticks .

Updated every driver (and all software) ..

Reinstalled windows ...

Updated Bios ...

Unplugged ... everything not needed (USB, Sata)

Rebuilt the PC, made sure every cable is nice and tight, New thermal paste...

CPU / GPU are not overheating...

I Left GPU benchmark running for 4 hours no crash (3dmark)

AIDA 64 .... seems to be fine when I run it.

Assumed it was my New SN750 ... even RMAed it ... (its not .. not even got it back yet) .. (SORRY WD .. I WAS WRONG)

also have new Mouse keyboard too... still got issues.


... What I do know ...

its never Crashed when playing Rimworld.

it has crashed on War Thunder, Rocket Leauge and Cyperpunk. (that make me think GPU)
It has also crashed when loading A VM.
Its also Crashed A few time with Launching War Thunder.
But again .. right after the crash i can play for hours with out issue. ....


Its Just so strange .. that i can go so long without issue but with heavy use...
After I reinstalled windows ... i went light on installing software slowy adding stuff ... got about 2weeks and it started doing it. Uninstalled last software added .. still happens.


At this point ... I need to start replacing parts ... Any ideas ... also what you replace first ?
PSU is the easiest .. so thinking that.
I dont think CPUs really go bad ... so saving that for last.
GPU is going to be pain to RMA.... and will be out of a computer for a bit .. ugh
Motherboard ... have to get a new one. (not looked at x570 in a bit ... whats good and cheap for a 5900x)

 

[shovenose]

Test RAM for several hours would be my first thought.

 

[neko77025]

Test RAM for several hours would be my first thought.

I have

 

[kilo]

Can you go to event viewer, then filter by source, "bugcheck" and get the hex code for us?

It's likely ram failing. You can use memtest (burn and image and boot to it). It's been a while since I used memtest but I'm sure there's a youtube tutorial on it. If you're shy about that, windows has a built in memory tester called "Windows Memory Diagnostic". Restart, and make sure to press F1 to use the extended tests, this will take a very long time, but should let you know if it's memory failure.

If everything passes, you may be able to use prime95 with 1 thread and all your ram to replicate memory being utilized within the OS.

Sometimes, rarely, VRAM BSOD look like RAM BSOD even without artifacts.

You might have implied it, but can you confirm that is does or doesn't BSOD in and out of a game?

 

[neko77025]

Went to Event viewer, Create Custom View: Logged( Any time) Checked all Event levels, By Source (picked Bugcheck from drop down), All Users / All computers .. Ok.
(Finds Nothing)

If I look at the event for times of crash.

Its A Kernel-Power , Event ID 41, Trask Category (63), Keyword (70368744177664),(2)
Details Tab

Hope that helps.

I did run Windows Memory Diagnostic the extened stuff .. Took forever with 64gb of ram. It passed. I have also had it crash with just 2 sticks in ... and swapped it out for the other 2 sticks and it still crashes .. (that why i dont think its ram)

If everything passes, you may be able to use prime95 with 1 thread and all your ram to replicate memory being utilized within the OS.

Have not done this .. will try that tonight.

This BSOD seems diffrent then the first ones (or before windows reinstall) .. First ones were whea_uncorrectable_error and system would hang on the BSOD ... had to hard reset comp.

Now after windows reinstall ... the BSOD shows .. but it is so fast .. not even sure what it says ... comp just resarts... but it dont start wtih whea_ . (but they were both (Kernel-Power , Event ID 41)

Right now (after reinstall) it BOSD after about 2 hour of WT, and 2x when I launched games WT and Pathfinder. Before windows reinstall .. it BSOD when launching A VM and one time opening chrome I think.

 

[RJARRRPCGP]

You didn't select a valid bugcheck entry in the event log.

 

[joemama]

Any commonality about the crash timings? Does it only crash when playing games?
I would eliminate the CPU, GPU and ram from possible sources since AIDA 64 and ram tests show no issue.
There's also other possibilties of unstable power from your house.

 

[ir_cow]

Random restart is often a drop in power or bad overclock. blue screen is hardware error related. blue screen could be from software crash as well.

First step is to turn off all overclocks. If it happens again, you can eliminate that being the issue.

 

[kilo]

Its A Kernel-Power , Event ID 41, Trask Category (63), Keyword (70368744177664),(2)
Details Tab
View attachment 260896

Hope that helps.

Kernel power failures, like the one above don't typically occur at a BSOD, since the system will have power to dump the memory to your system.

If you don't see any bugcheck events, but you do see the hard resets, it'll be PSU that needs replacing (you likely already know this).

However, if perhaps something was filtered incorrectly in event viewer and you do find the error, it should indicate which component is failing, which is usually RAM (just raw chance).

Just to rule out RAM once and for all, perhaps if you get no BSOD on prime95, you could also try turning off XMP.

If it is the PSU, then I suggest picking up a cheap (sub $20) PSU tester if you don't have a multimeter.

Edit: I forgot to mention that you can change the system dump settings in control panel > system > advanced system settings > advanced > startup and recovery. Turn off automatically restart for starters. If you get change "automatic memory dump" to "complete memory dump", that'll be all 64GB of your ram written, in case that's needed.

 

[Bill_Bright]

PSU is the easiest .. so thinking that.

Ensuring quality power is also the smartest, most logical place to start. Since EVERYTHING inside the computer case depends on good, clean, stable power, you need to ensure you are providing it. So starting by swapping in a known good PSU is simply a good, and standard electronics troubleshooting practice.

Also, every home and every computer user should have access to a AC Outlet Tester to ensure the wall outlet is properly wired and grounded to Earth ground. I recommend one with a GFCI (ground fault circuit interrupt) indicator as it can be used to test bathroom and kitchen outlets (outlets near water) too. These testers can be found for your type and voltage outlet, foreign or domestic, (like this one for the UK, or this one for German outlets) at most home improvement stores, or even the electrical department at Wal-Mart. Use it to test all the outlets in the home and if a fault is shown, have it fixed by a qualified electrician.

That said, I always recommend all computers be powered through a "good" UPS with AVR. Note it is the AVR (automatic voltage regulation) feature that makes a good UPS with AVR so important and valuable. It is also important to note that destructive and disruptive power anomalies don't have to originate from the "grid". A faulty high wattage device in the home can create them too. And battery backup power during a power outage is just a minor, bonus feature. If someone comes back and says you don't need an UPS if you rarely lose power, they didn't read past the first sentence in this paragraph before pouncing. I only mention that because it happens all the time.

 

[Count von Schwalbe]

Possibly swapping your GPU for another if you have one could isolate if that is the issue. Based on the issues you are having, it is unlikely but possible.

I also recommend testing other circuits in your house.

 

[TheoneandonlyMrK]

You tested it yes but still I think it memory related, probably actually the IMC getting hot.
No Ryzen support's 4x 3600 dimms even 2x 3600 isn't guaranteed to work.

Set at jedec(jefeck does seem better though) speeds off 2400 or whatever the low speed is and it wouldn't do this I would wager.

And I say this as someone who cured random crashes via these means.

 

[neko77025]

You tested it yes but still I think it memory related, probably actually the IMC getting hot.
No Ryzen support's 4x 3600 dimms even 2x 3600 isn't guaranteed to work.

Set at jedec(jefeck does seem better though) speeds off 2400 or whatever the low speed is and it wouldn't do this I would wager.

It happens with just 2 sticks .. (tested both set ea by them self) .. with or with out XMP .. (tested as it will happen no matter what)

Possibly swapping your GPU for another if you have one could isolate if that is the issue. Based on the issues you are having, it is unlikely but possible.

I also recommend testing other circuits in your house.

Yea, dont have access to another GPU , with out buying one. (would, if new ones were not about to drop)

I order A new PSU (evga G5 850w) Old GPU is A G5 750w (it is about 1 1/2 years old)

Will do some of the other stuff later today.

 

[Bomby569]

not all hardware errors are hardware errors, i learned that the hard way. Hardware fails but can be because of software requests. I would stay uninstall all unnecessary crap like rgb stuff for example.

 

[A Computer Guy]

... whea_uncorrectable_error ...

Seems like a memory thing.

One time because of an aggressive ram overclock I had a ram problem when only AceCombat 7 would cause the PC to instantly reboot, after 30 minutes or so but memtest was always fine.

You might just be running at the edge of stability for some reason.
You can try to manually dropping your RAM speed down to DDR4-3200 and see if your problem goes away.
(like others have said remove any overclocks CPU, RAM, GPU, etc...)

Can you post your RAM specs along with the voltages and timings you are using including SOC, IF speed. It might be helpful.

 

[Deleted member 185158]

TridentZ neo 4x 16gb 3600Mhz C16 (F4-3600C16D-32GTZNC)

Try pulling 2 sticks from A1 and B1 and do some more testing.

 

[A Computer Guy]

TridentZ neo 4x 16gb 3600Mhz C16 (F4-3600C16D-32GTZNC)

Try pulling 2 sticks from A1 and B1 and do some more testing.

4 x 16gb so you have 64GB ram correct?
You can probably set XMP or AUTO and then manually downclock the frequency to DDR4-3200.
Post your voltages. DRAM, SOC, etc...

 

[MarsM4N]

I wouldn't suspect a hardware failure from the start without checking for simple software/driver errors.
I had the issue a while back & I could track it down to a incompatible audio driver. Fixed it by deinstalling the driver in the device manager. But it could be any driver. I got BSOD's super random, once a day, sometimes twice, then for a week nothing & then it came again. So make sure you got the newest & correct drivers for your hardware.

I would first check the "Event Manager". When did it first occur? What did you do on that day? Did you install anything?

You can also try these steps: How to Fix a Windows 10 Kernel Power Error in 5 Easy Steps
And this: WinDBG - The Basics for Debugging Crash Dumps in Windows 10

With WinDBG you can analyze the Windows crash dumps, you can go through the file & search for hints about what is causing the crash.

 

[neko77025]

TridentZ neo 4x 16gb 3600Mhz C16 (F4-3600C16D-32GTZNC)

Try pulling 2 sticks from A1 and B1 and do some more testing.

I have ran Every Combination of my ram and the slots.
Set 1 only A1/B1
Set 1 only A2/B2
Set 2 only A1/B1
Set 2 only A2/B2
XMP on or OFF

not all hardware errors are hardware errors, i learned that the hard way. Hardware fails but can be because of software requests. I would stay uninstall all unnecessary crap like rgb stuff for example.

Their is no RGB software installed ...

The last thing I had installed before it started again was Glasswire .. and I have uninstalled it. But it still happens after uninstall.

Only software on system right now . (everything is Up todate)

Ryzen Master (just for looking at CPU temps)
Discord
Google Drive
Nvida's Geforece Experiance
Track Ir (hat controler for War thunder)
Gigabyte Realtek sound card software
AIDA64 Extreme
Plex (Client)
HDHomerun (Client)
KeePassXC
VeraCrypt
Steam
War Thunder
Rim world
Pathfinder WoTR
3DMark
ShareX

You can also try these steps: How to Fix a Windows 10 Kernel Power Error in 5 Easy Steps

I have done all the steps but the last one .. New PSU will be here Tomarrow.

 

[Zach_01]

Went to Event viewer, Create Custom View: Logged( Any time) Checked all Event levels, By Source (picked Bugcheck from drop down), All Users / All computers .. Ok.
(Finds Nothing)

If I look at the event for times of crash.

Its A Kernel-Power , Event ID 41, Trask Category (63), Keyword (70368744177664),(2)
Details Tab
View attachment 260896

Hope that helps.

I did run Windows Memory Diagnostic the extened stuff .. Took forever with 64gb of ram. It passed. I have also had it crash with just 2 sticks in ... and swapped it out for the other 2 sticks and it still crashes .. (that why i dont think its ram)


Have not done this .. will try that tonight.

This BSOD seems diffrent then the first ones (or before windows reinstall) .. First ones were whea_uncorrectable_error and system would hang on the BSOD ... had to hard reset comp.

Now after windows reinstall ... the BSOD shows .. but it is so fast .. not even sure what it says ... comp just resarts... but it dont start wtih whea_ . (but they were both (Kernel-Power , Event ID 41)

Right now (after reinstall) it BOSD after about 2 hour of WT, and 2x when I launched games WT and Pathfinder. Before windows reinstall .. it BSOD when launching A VM and one time opening chrome I think.

Kernel Power Event ID 41 usually logs when an unexpected shutdown happens. The cause would be different on that case.
Look for WHEA errors

 

[Deleted member 185158]

I have ran Every Combination of my ram and the slots.
Set 1 only A1/B1
Set 1 only A2/B2
Set 2 only A1/B1
Set 2 only A2/B2
XMP on or OFF

And even a single stick in slot A2 also XMP on and off?
What are the results of the above listed, still shutting down Kernel 41 sudden power loss?

Not a resolve, but you can try and disable "power down mode" in the memory timings menu. Should be right next to "gear down mode" Do some testing with power down mode disabled and see if that lowers the frequency of the restarts.

lmk

 

[Athlonite]

Also could be suffering from nVidia's power spikes putting the PSU into OCP/OPP mode

 

[neko77025]

Heads up ... seems like it was the PSU.

Did replace it and seems to have stopped ...... (I hope).

(noticed when I did go to replace it , the old PSU was very hot to the touch) (side with out the fan on it ).


Anyhow thanks all

 

[neko77025]

So .... It started happing Again .. out of nowhere. On the 1st .. Was So Confused... angry and sad. Have had no issues for over 2 months.

HOWEVER!!! .. I THINK I KNOW WHATS IT IS !!!

Fore the last 3 weeks I have done NOTHING buy play War Thunder ALOT. No new software .. Zero hardware changes. In fact the only things I did were on 1st. Because My TV died .. I could not watch Netflix / Disney+

Xbox Game Bar ... when that is on it causes BSOD. Assumed , it software and someone else had to be having this issue (even tho I cant find them) .. Started watching all video posted in the last 2 months that had to do with WHEA_UNCORRECTABLE_ERROR or BSOD... Found this one

I remember I wanted to install the Netfix app vs webpage .. because I was only getn 1080p .. Have to use the Windows Store .. when I did .. it said something About Xbox game bar.

their is also A short cut like Win+G ... think i might of did that in the past .. cause i remember it popping Up.


But I have NO clue why Xbox game bar BSOD mine and his system.

Other oddty ... even tho i have it set to off ... when I Press WIN+G it still opens but the Xbox Game bar .. (but setting still show it off)

 

[Damir4361]

So .... It started happing Again .. out of nowhere. On the 1st .. Was So Confused... angry and sad. Have had no issues for over 2 months.

HOWEVER!!! .. I THINK I KNOW WHATS IT IS !!!

Fore the last 3 weeks I have done NOTHING buy play War Thunder ALOT. No new software .. Zero hardware changes. In fact the only things I did were on 1st. Because My TV died .. I could not watch Netflix / Disney+

Xbox Game Bar ... when that is on it causes BSOD. Assumed , it software and someone else had to be having this issue (even tho I cant find them) .. Started watching all video posted in the last 2 months that had to do with WHEA_UNCORRECTABLE_ERROR or BSOD... Found this one

I remember I wanted to install the Netfix app vs webpage .. because I was only getn 1080p .. Have to use the Windows Store .. when I did .. it said something About Xbox game bar.

their is also A short cut like Win+G ... think i might of did that in the past .. cause i remember it popping Up.


But I have NO clue why Xbox game bar BSOD mine and his system.

Other oddty ... even tho i have it set to off ... when I Press WIN+G it still opens but the Xbox Game bar .. (but setting still show it off)

Had a simliar issue with one of my computers ,

tried updating m/b bios & swapping out with another motherboard, Swapped video card and updated videocard firmware, SSD drives firmware updated, Swapped ram brands and ram speeds went from 4400-

3600,changed p/s 1000watt Antec to 1300Watt Antec, re installed windows a few time to no avail, to be honest the only thing which pointed me in the right direction was this program whocrashed

which a technician from MSI australia suggested in trying prior to sending the motherboard back for rma.

The Link is https://www.resplendence.com/whocrashed

in the end it ended up being a thermal issue which without whocrashed would of never found.

Good Luck.