BSOD occurs often. (resolved)
BSOD occurs often. (resolved)
Hi everyone,
I'm experiencing several BSODs lately, often showing errors such as irql_not_less_or_equal or hypervisor issues. I ran memtest on my RAM and all tests passed. I also stress-tested the CPU and GPU, and there were no problems even at low temperatures. The only part I couldn't verify was the motherboard (B350m Tomahawk, in use for three years).
What I did:
- Updated drivers for the GPU using ddu
- Reinstalled chipset and CPU drivers
- Tested each RAM individually
- Switched RAM positions to see if crashes changed (initial: slots 2 & 4, then 1 & 3, then 2, then 1)
- Used BlueScreenView to analyze the dump files; BSODs were linked to ntoskrnl.exe, but I’m unsure how to resolve them
- Ran sfc /scannow without issues
- Did not overclock any devices
- Reset CMOS
At this stage, I’m unsure what further testing is needed.
I have the dump file available for analysis if anyone wants to help.
https://www.mediafire.com/file/2yaxqhz6m...p.rar/file
verify your power cord
also check your power supply and graphics card
ensure the chipset driver and BIOS are both current
Power cord connected fully. I'm running Rosewill capstone 650w and Zotac rtx 3070 Twin Edge with Ryzen 5 3600.
I've updated the chipset and bios too.
Update: After replacing a new pair of RAM, it still crashed, but the RAM seems fine.
Almost no details regarding this bug check have been shared. As the title implies, it functions as a Hyper V exception.
I also noticed that so I attempted to turn off virtualization in BIOS. It still crashed, but the unusual aspect is that when virtualization is enabled, the BSOD error changes to a Hypervisor Error. When virtualization is disabled, the error differs, usually showing rql_not_less_or_equal. Can I try testing the CPU somehow? I ran stress tests and it worked without any issues.
Check this out!
Rich (BB code):
IRQL_NOT_LESS_OR_EQUAL (a)
A try was made to reach a pageable or invalid address during an interrupt request level (IRQL) that exceeds the allowed limit. This typically happens when drivers use incorrect addresses.
If a kernel debugger is accessible, retrieve the stack trace.
Arguments:
Arg1: 0000000000000080, memory accessed
Arg2: 0000000000000002, IRQL
Arg3: 0000000000000001, bitfield:
bit 0 : value 0 = read operation, 1 = write operation
bit 3 : value 0 = not an execute operation, 1 = execute operation (only on chips supporting this level)
Arg4: fffff80709c2ef21, address that was referenced
The system experienced two crashes on this file! The second crash triggered a Hyper v exception once more. Likely connected to CPU interrupts – addresses being accessed.
Rich (BB code):
8: kd> KnL
# Child-SP RetAddr Call Site
00 ffffc88c`4a3a9a78 fffff807`09c2de29 nt!KeBugCheckEx
01 ffffc88c`4a3a9a80 fffff807`09c29289 nt!KiBugCheckDispatch+0x69
02 ffffc88c`4a3a9bc0 fffff807`09c2ef21
nt!KiPageFault+0x489
03 ffffc88c`4a3a9d58 fffff807`09bcb587 nt!memset+0x61
04 ffffc88c`4a3a9d60 fffff807`09d42560
nt!HvcallInitiateHypercall+0x57
05 ffffc88c`4a3a9de0 fffff807`09d6589a nt!HvlNotifyRootCrashdump+0x68
06 ffffc88c`4a3a9e50 fffff807`09c18087 nt!KeBugCheck2+0xca
07 ffffc88c`4a3aa5c0 fffff807`09c97d4e nt!KeBugCheckEx+0x107
08 ffffc88c`4a3aa600 fffff807`09c921bc nt!MiCheckSystemNxFault+0x1f638a
09 ffffc88c`4a3aa640 fffff807`09a30213 nt!MiRaisedIrqlFault+0x200e38
0a ffffc88c`4a3aa690 fffff807`09c2917e nt!MmAccessFault+0x363
0b ffffc88c`4a3aa7b0 fffff807`09af7420
nt!KiPageFault+0x37e
0c ffffc88c`4a3aa940 00000000`00000000
nt!KxReleaseQueuedSpinLock
Prime95. In fact, based on my experience, it’s not always the case that every CPU fails under stress testing. It might be wiser to test with a different processor.
I just tested prime95 and occt, and some cores aren't working properly. Is this the issue? OCCT reported errors right away when I started using it.
https://imgur.com/a/3hW30VV
There are numerous elements influencing Prime95's performance. Let's review the tests conducted. Examples include Small, Smallest, Large...
Blend is a demanding process that impacts various components. It can be challenging to grasp. For the CPU, it's best to opt for the simplest and most minimal tests.