I have been lurking here for a while sifting though the topics searching for a solution to my BSOD problem. As I did not find one, I now post for myself. Here is my story: In January 2016 I upgraded my self-built Desktop-PC with new hardware and progressed from Vista x64 to Win 10 x64 trough clean installation. My system now sports:
CPU: Intel i5 Skylake 6600K (not overclocked)
RAM: G.Skill RipJaws V DDR4-3200 CL16 Dual Kit 2 x 4 GB
MB: MSI Z170A Gaming Pro
Graphics: Asus HD 7950 DirectCU-II PCIe 3.0 x 16 3 GB
SSD: Samsung 850 EVO 250 GB
HDD1: Samsung HD103UJ 1 TB
HDD2: Samsung HD103SJ 1 TB
Sound: X-Fi Xtreme Gamer Soundcard
Power: beQuiet 500W PSU
Here are some screencaps of system info from CPU-Z:
A screencap of bluescreen viewer with all recent minidumps and bugcheck strings:
During the summer I encountered few bluescreens which only occurred after waking the system up from standby/sleep/hibernation mode. I suspected that this was caused by a driver which could not handle the power state and did not care too much.
Since the beginning of December, the frequency and severity of the crashes increased greatly. I sometimes got small bootloops, where the system would crash even after normal startup/cold boot after some time - a thing my PC never did before - and re-crash multiple times during the following boot-process. The system started working again after a few loops, but as I said I never witnessed this behaviour before. I did not install any new hard- or (significant) software in December. Right now, the system crashes randomly during day-to-day usage such as gaming, browsing the web and watching streams, I could not identify any specific trigger. Any help would be appreciated as I don't know what to do next to fix this nagging issue.
Here is what I already did:
- Enable driver verfier: I used my PC with driver verfier for over 48 hours following the tutorial of jcgriff2. I don't think it detected any faulty driver/crash the system or throw a bluescreen.
- Run Memtest86 V7.1 for over 9 hours (19 passes) with no errors reported
- Updated all hardware drivers (including UEFI, graphics, audio, chipset)
- Load-tested the system with Prime95/Furmark and surveilled the temperatures of graphics/CPU -> they were in acceptable range
- Tried to analyze the Minidumps with WinDbgX64 to find the error. I couldn't identify anything as I am not knowledgeable enough. One common "string" in all of them was memory_corruption
- Run sfc /scannow from an elevated command prompt (no errors found or corrected)
- Run chkdsk /r /f at boot time for my system SSD (no errors found or corrrected)
- Run a full system scan with Malwarebytes Anti-Malware (no threats found | I don't use another antivirus apart from Windows-internal tools)
I attached the relevant files according to the stickied BSOD posting tutorial.
Apart from completely reinstalling the system - my next and last resort measure - I am running out of options. I would greatly appreciate any help which could aid me on my journey towards a stable system.
With kind regards,
----------------- Major Edit 11. Dec:
During some load-testing with Prime95 blend test, I encountered another instance of these "small" bootloops I described above. After about ~ 1 minute of Prime95 the system halted. So now I can give detailed information about this error:
- The system initially failed with an IRQL_NOT_LESS_OR_EQUAL BSOD during Prime95 and no other programs running
- During following boot-up we got a KERNEL_SECURITY_CHECK_FAILURE bluescreen
- While the following boot, the system threw a BAD_SYSTEM_CONFIG bluescreen stop and by (automatically) entering the Windows system recovery utility, we were able to start the system normally
Note that only the first BSOD produced a memory dump, which I attached at this post, too. The file for this specific crash bears the name "PrimeCrash Minidump.zip". Again, the common denominator of all this minidumps in WInDbg "!analyze -v" is the generic string "memory corruption", but please see for yourself. Again I thank you in advance for helping me getting a stable system again.
----- Small After-Edit:
I reran the Prime95 blend-test after I could reboot correctly. Right now it has been running perfectly for about sixty minutes, the temps are in an acceptable range. Screencap of the process delivered here:
It becomes clear that I still fail to find an "algorithm" to crash the system consistently!
After approx. 2 hours of consistent Prime95 blend-test (no other usage at point of stop error) the system crashed again with a new and unique bug check string: DEVICE_QUEUE_NOT_BUSY (reported by BlueScreenViewer app). I uploaded the corresponding minidump zipped with the name "PrimeCrash_2HoursIn". Thanks for looking at it!
Edited by Coridan75, 11 December 2016 - 08:57 AM.