BSOD's mostly when idle

G-Man

G-Man

G-Man

Soldato
Joined
17 Oct 2002
Posts
3,737
Location
Scotland
So I put together an upgrade and its been giving BSOD's almost once a day minimum but almost always only when its idling overnight, only twice has it happened under load.
The BSOD error is pretty much always different, the last one was KMODE_EXCEPTION_NOT_HANDLED in ntoskrnl.exe.

Specs are:
Ryzen 7 7800X3D
be quiet! Pure Rock Slim 2
Asus TUF Gaming B650-Plus WIFI
TeamGroup Vulcan EXPO 32GB PC5-48000C38 6000MHz

Reused the following parts from old build:
Radeon RX5700XT (Powercolor if it matters)
Superflower 750W PSU
2 Samsung 970EVO SSD's
Corsair 230T Case with 2 fans at front, 2 at top and 1 back.

Running completely stock without even enabling EXPO on the RAM.
CPU temps are floating around 45-50 when idle and spiked to about 80ish under load.
GPU temps are about 60C when idle and it has reached 103C under load.

Was initially using Windows 11 but have reinstalled 10 to see if it made any difference, so far it hasn't.

Using latest BIOS from Asus and latest chipset/GPU drivers from AMD. Have tried the chipset drivers from Asus which seem to be newer than the ones AMD have but no help.

Ran memtest86 for about 9 hours with no issues and no errors.

This is what bluescreenview showed this morning (unfortunately due to going back to W10 I don't have a record of the previous BSOD's but they were things like SYSTEM_SERVICE_EXCEPTION and IRQL_NOT_LESS_OR_EQUAL

ji6aELx.png


I'm beyond stumped at this point and very much regretting the upgrade.
Any ideas welcomed.
 
Usually nothing, if it wasn't for the events about the previous shutdown being unexpected event viewer wouldn't even notice the crash

This morning (around the time of the crashes with no bug check string in the above screenshot) there is one listed under HAL as "The system watchdog timer was triggered." and one under Kernel-Power as "The system has rebooted without cleanly shutting down first. This error is caused because the system stopped responding and the hardware watchdog triggered a system reset."
Those have never been logged previously.
 
Any suggestions on benchmark? I've ran the little bench in CPU-Z with no issues but it's quite short.

While I'm at work I've left it running off the iGPU with the 5700XT removed as a test.
 
Turned every power saving option I could find off although the actual profile is still set to balanced.

No BSOD's or reboots while I was out at work all day with the 5700XT removed.

Ran cinebench and it survived the standard benchmarks on both the iGPU and with the 5700XT reinstalled.

5700XT:
gJd2YKC.png



iGPU:
0GZ5Zro.png


Only just noticed the advanced options with the 30min stablity test so I'll try that after raid in WoW. Did give the graphics card a blast with the compressed air to clean out the dust, maybe helped but not confident.


Tee Hee Johnson said:
I think people suffer from this when using the optimisation curve to undervolt the cpu, in idle conditions the power was too low and could cause a crash.

Not massively relevant to yourself since you haven’t messed with anything, but it could hint at perhaps a power delivery issue?
Click to expand...

As part of the tests earlier I did try setting the VSOC to 1.25v manually as per a random suggestion on reddit about Asus boards and the auto setting. Didn't help and as you can imagine the temps were deffo higher. Right now the BIOS settings are literally the optimised defaults option with a couple things turned off (virtualisation, fTPM, Wifi).
 
Had another BSOD while alt tabbing out of WoW, this time MEMORY_MANAGEMENT which of course makes it sound like the RAM but the RAM passed 9 hours of memtest86
 
It managed to survive a full night and day without crashing only to BSOD again this morning.

Now running with 1 stick of RAM removed but god knows how long it'll take to do anything.
 
Knowing my luck the stick of RAM I've left in will be the working one.


On an amusing note, the RAM I got isn't actually on the QVL list for the motherboard.


Would much prefer faulty RAM to faulty CPU, I can afford to buy replacement RAM so I'm not stuck without a PC waiting on it being returned. Can't afford that for the CPU....
 
Last edited:
keef247 said:
It'll be the ram not behaving running at the XMP speed. Turn XMP off and set the ram to the standard OEM recommended compatibility that AMD specify - I don't know what this is for AM5 as I'm on AM4. I had this problem and no matter what XMP profile I used, it wouldn't behave.
Click to expand...

RAM has been running stock speed pretty much since the first BSOD.
 
Stock as in its running at 4800 instead of using the EXPO profile to run at 6000.


Cyber-Mav said:
Wrong way to go around this. If you suspect its the ram then isolate it by running memtest86 if it passes then run testmem5 using anta absolute config.
Take things from there
I wont be surprised if the issue is the cpu. Ryzen cpus are notoriously unreliable
Click to expand...

9 hours of memtest86 found nothing with both sticks in so I'd be surprised if it finds anything with 1 stick.

Remember most of the BSOD's are happening when the PC is idle.
 
Running since Saturday on one stick of RAM. No crashes at all.

Swapped to the other stick today and already 2 BSOD's so I'm very much leaning towards faulty RAM.
 
You must log in or register to reply here.
Back
Top Bottom