Random BSOD's on Win10

[Pablo]

Hi experts, this is a really weird case... this computer can run perfectly for a whole month and the out of the blue starts crashing on every boot for 2 days... to just stop crashing for another month... That behaviour makes it extremely difficult to test for solutions.

AMD A10-5800K
RAM 8Gb DDR3 (2 x 4Gb)
Mobo Gigabyte F2A55M-HD2 (latest BIOS installed)
GPU GeForce GTX 750Ti
SSD Kingston A400 240Gb Sata 2,5 inches
PSU Corsair CS750M 3 months old (bought it thinking it would solve it.. wrong)
Win10 x64

Do the minidumps point in any direction?
Thanks in advance

 

[ubuysa]

My initial thoughts on looking at these dumps is that RAM is the most likely cause. That also would explain why the BSODs are so rare, it's not until you use some particular RAM location that you get a problem. The main thing that makes me think RAM are the failing addresses in three of your dumps, all with 0xC0000005 exception codes (invalid memory reference)...

122723-7125-01.dmp     ffefb301`364c0a20=????????????????
121923-11593-01.dmp    ff7fb703`86c86b70=??
121323-6390-01.dmp     ff7fc38e`5d095698=????????????????

These are all non-canonical addresses, that means they are not within the range of acceptable addresses. Kernel addresses all start with 0xFFFF, but you can see in these that the third nybble (from the start) is not 0xF as it should be.

I'd suggest that as a first step you test your RAM...

1. Download Memtest86 (free), use the imageUSB.exe tool extracted from the download to make a bootable USB drive containing Memtest86 (1GB is plenty big enough). Do this on a different PC if you can, because you can't fully trust yours at the moment.
2. Then boot that USB drive on your PC, Memtest86 will start running as soon as it boots.
3. If no errors have been found after the four iterations of the 13 different tests that the free version does, then restart Memtest86 and do another four iterations. Even a single bit error is a failure.

 

[Pablo]

Forgot to clarify that I already ran Memtest for a whole day (about 10 or 11 passes) and not a single error was found. Thinking it could be a sneaky error (as it can work flawlessly for a month) I removed the RAM and installed a stick I have laying around. Proceeded to let the computer run for about a week nonstop with

shutdown -r -t 300

command at startup, to make it continously reboot in a safe manner.
Not a single error either.
But that same behaviour could have been replicated with the oringal RAM... that's what makes it so annoying, trial and error can take months!

[EDIT] I will install my RAM stick again and redo the continous rebooting test. I'll post back results.

 

[ubuysa]

Memtest86, to which I linked is IMO a better memory test, bit let's assume your RAM is probably ok for now, in which case we're probably looking for a rogue driver. Please enable Driver Verifier as detailed here: Driver Verifier Instructions - BSOD - Windows 11, 10, 8(.1), 7 and Vista. If you've never enabled Driver Verifier before you may want to read through those instruction before you make any changes.

We expect BSODs with Driver Verifier so please be sure to keep all dumps. Not also that because Driver Verifier performs additional checks on selected drivers there may be a small performance impact.