PC Freezing

BriT

BriT

BriT

Soldato
Joined
27 Jun 2005
Posts
5,111
Location
Reading
Hi guys,

Had my PC in my sig for two years now with no problems but recently it has been locking up complety and have to power off by case button.

Thing is, is it doesn't do it all the time but more lately. Software issue? or BIOS update would help?

Could do a fresh windows install.
 
Firstly, open up Event Viewer and see if it says anything in the Admininstrative Events and then check your WHEA logger. If you don't have a WHEA logger, set one up to monitor and refer to it upon a reboot.
You don't mention your PSU but I would have thought a power issue would result in a shut down not a screen lockup but would probably help to know.
 
Oh, forgot to say:

Open up task manager or even better HWinfo64 and monitor what your temps are. It may be a question of reapplying some paste or adding some fans if your temps are causing issues. When was the last time you cleaned the PC out? Blocked air vents or clogged fans may be causing temp issues. You don't say what you are doing when it happens.
 
This is event viewer :



PSU is a 850w seasonic focus gold, a few years old now. I was playing MSFS 2020 for a few hours then froze near landing.

Ran cinebench a few times and max core was 80c so i think temps are fine.
 
Last edited:
Those events all look meaningless to me, would check for WHEA errors as suggested above.

What model is the Samsung Evo? Can you post the smart data? (crystaldiskinfo)

Do you have any wireless devices?

Anything changed with the PC recently?

Device manager is clear of errors?

BriT said:
PSU is a 850w seasonic focus gold, a few years old now. I was playing MSFS 2020 for a few hours then froze near landing.
Click to expand...
I believe that model was replaced by the GX in 2018, so could be 5+? I'd have expected it to age alright, but you could try power limiting/underclocking your 6900 XT.
 
Tetras said:
Those events all look meaningless to me, would check for WHEA errors as suggested above.

What model is the Samsung Evo? Can you post the smart data? (crystaldiskinfo)

Do you have any wireless devices?

Anything changed with the PC recently?

Device manager is clear of errors?


I believe that model was replaced by the GX in 2018, so could be 5+? I'd have expected it to age alright, but you could try power limiting/underclocking your 6900 XT.
Click to expand...

Hi

It's a Samsung 850 EVO 500GB, Saying 90% health in crystaldisk. It's very old as well so maybe its worth buying a cheap m2 drive and reinstalling windows? Has 22300 hours.



Although I have no other symptoms and doesn't freeze everyday?

I have wireless mouse and headset, no changes to t he PC.

Will check when I bought the PSU, but as you say should be ok.
 
Last edited:
BriT said:
It's a Samsung 850 EVO 500GB, Saying 90% health in crystaldisk. It's very old as well so maybe its worth buying a cheap m2 drive and reinstalling windows? Has 22300 hours.
Click to expand...
Reserved blocks used, bad blocks and ecc uncorrectable are all 0, so there doesn't appear to be any issue with the flash. There's nothing there that stands out as a problem. Have you done a SFC /scannow?

BriT said:
PSU was bought on 31st Dec 2018, so coming up to 5 years old.
Click to expand...
Would have been one of the last of that model sold, I'd imagine.
 
Tetras said:
Reserved blocks used, bad blocks and ecc uncorrectable are all 0, so there doesn't appear to be any issue with the flash. There's nothing there that stands out as a problem. Have you done a SFC /scannow?


Would have been one of the last of that model sold, I'd imagine.
Click to expand...

Ok cool. I will do the SFC/scannow.

Edit - ran that and it found a corrupt file and fixed it.
 
Last edited:
Setup a WHEA logger in Event Viewer by following the instructions here Setting up WHEA Logger

It starts about a minute in. Then you will be able to see after a lockup if it is one of your CPU cores (I think Event ID 18 or Event 19)

Open this up after a crash and see if there are any problems here which are unlikely to show up in the standard admin events
 
Noughtboy said:
Setup a WHEA logger in Event Viewer by following the instructions here Setting up WHEA Logger

It starts about a minute in. Then you will be able to see after a lockup if it is one of your CPU cores (I think Event ID 18 or Event 19)

Open this up after a crash and see if there are any problems here which are unlikely to show up in the standard admin events
Click to expand...

Thanks, have set that up.
 
BriT said:
Thanks, have set that up.
Click to expand...
Great. Now after your next freeze, check to see that it isn't one of your cores falling over.

If it isn't, we can turn to software after your SFC scan showed corruption (although corrected).

But a Windows reinstall is the nuclear option because a clean install means you have to note all your added programs which would need to be downloaded and reinstalled.

Let's rule out hardware or driver issues first. You may want to update/reinstall Radeon software first. You should run DDU first to wipe it clean, then download latest and install. There are YT videos but essentially download DDU, install and run. I forget if you have to do this in safe mode or not.

Then reboot. Then install Radeon software and drivers.

Out of interest, has the PC been overclocked or tweaked in any way other than setting RAM to XMP?
 
Noughtboy said:
Great. Now after your next freeze, check to see that it isn't one of your cores falling over.

If it isn't, we can turn to software after your SFC scan showed corruption (although corrected).

But a Windows reinstall is the nuclear option because a clean install means you have to note all your added programs which would need to be downloaded and reinstalled.

Let's rule out hardware or driver issues first. You may want to update/reinstall Radeon software first. You should run DDU first to wipe it clean, then download latest and install. There are YT videos but essentially download DDU, install and run. I forget if you have to do this in safe mode or not.

Then reboot. Then install Radeon software and drivers.

Out of interest, has the PC been overclocked or tweaked in any way other than setting RAM to XMP?
Click to expand...

Sounds good. I have the latest GPU drivers installed but not used DDU.

My RAM is set to 3000mhz as XMP would cause instability.

I've not overclocked the CPU but have set a set speed at 5ghz for P cores and 3.8ghz for the E cores with a set voltage, turned off the save states.
 
Cannot see your Sig so just assumed it is all AMD build.

Haven't run Intel since my 2500k days.

Might I suggest running a RAM test to start on your next boot and then reboot. It will usually take quite a while. Bank on at least an hour.

If this passes, then I would suggest taking it back to stock for a few days. If you want to start slow, reduce both set speeds by 10% - 20% along with some lower voltage - I think each stage is something like 0.0125v.

Try one stage . If stable two stages. If stable up speeds a little. Sometimes, CPUs become less stable running under or over volts (I know as my 5700x has had to have a couple of bumps up to the undervolt over the last 5 months).

If that doesn't work, take it to stock and then slowly work your speeds and volts upwards. Or by undervolting, just your voltage downwards.
 
BriT said:
Sounds good. I have the latest GPU drivers installed but not used DDU.

My RAM is set to 3000mhz as XMP would cause instability.

I've not overclocked the CPU but have set a set speed at 5ghz for P cores and 3.8ghz for the E cores with a set voltage, turned off the save states.
Click to expand...
You should run memtest just in case it's the memory.
 
BriT said:
Edit - ran that and it found a corrupt file and fixed it.
Click to expand...
One corrupt file isn't something I'd worry about.

BriT said:
My RAM is set to 3000mhz as XMP would cause instability.

I've not overclocked the CPU but have set a set speed at 5ghz for P cores and 3.8ghz for the E cores with a set voltage, turned off the save states.
Click to expand...
You're technically not running at stock for either the memory or the CPU, then? That makes troubleshooting difficult, because to rule them out as a cause, you'd have to check that both the memory and CPU are 100% stable in all conditions (e.g. idle, light load, heavy load) operating in the way you have configured them. Is the memory voltage @ the manufacturer recommended, or at 1.2v? Does XMP change other voltages that you haven't set manually?

With the CPU: having the CPU run at a fixed clock is pretty complicated nowadays, have you done much stress testing?

If it was me, I'd disable all of your manual tuning and put everything back to stock, including running the memory at 2133 or whatever it boots at.
 
Thank's fro the advice guys. I mean it's been fine for 3 days now with PC on 12 hours a day but will continue to monitor it. If it happens again I will reset to default settings as suggested and go from there.
 
You must log in or register to reply here.
Back
Top Bottom