HwInfo crashing since a few days mostly everyday

M

maffle

Well-Known Member
Not sure whats causing this, if it is a new Nvidia driver or HwInfo update, but I am mostly seeing HwInfo crash events in Windows reliability report every day:

Beschreibung
Name der fehlerhaften Anwendung: HWiNFO64.EXE, Version: 7.64.5240.0, Zeitstempel: 0x65265eb7
Name des fehlerhaften Moduls: nvml.dll, Version: 8.17.15.4584, Zeitstempel: 0x65270167
Ausnahmecode: 0xc0000005
Fehleroffset: 0x00000000000a93e0
ID des fehlerhaften Prozesses: 0x0x37E8
Startzeit der fehlerhaften Anwendung: 0x0x1DA08969FC7F985
Pfad der fehlerhaften Anwendung: C:\Program Files\HWiNFO64\HWiNFO64.EXE
Pfad des fehlerhaften Moduls: C:\Windows\system32\DriverStore\FileRepository\nv_dispi.inf_amd64_4eef80c06561d01a\nvml.dll
Berichtskennung: 541e19b7-1e9b-4a16-bcd9-3d5027962925
Vollständiger Name des fehlerhaften Pakets:
Anwendungs-ID, die relativ zum fehlerhaften Paket ist:

Seems the crash has something to do with nvml.dll (Nvidia driver?). I am just seeing these crashes since a few days, didnt happen before.

HwInfo version 7.64-5240.

It started around 26.10.2023.

Looking at the time when it happens, it seems to happen when Windows is starting, it crashes in the background, and then starts itself again it seems and then it is working.

Sorry posted it accidentally under the DOS section, maybe you can move it in the other section.
 
Is this crashing on every start or maybe only during update of NVIDIA GPU drivers?
Please attach the HWiNFO Debug File of the crash so I can have a deeper look into that.
 
Hm. It might be, that it happened when the Nvidia driver was updating. 2 events seem to be identical when the Nvidia driver was updated, one is not.

1698856432541.png

1698856451154.png

1698856467233.png

1698856481783.png

It seems the event at 27.10.2023 06:32 and 31.10.2023 at 16:23 are when a Nvidia driver update happened. The crash on 26.10.2023 10:46 I dont have an event of Nvidia driver update though, the crash happened during Windows start/login though it seems.
 
After going through the Windows events, it seems the other event where HwInfo crashed, there was some weird action caused by "NVFLASH":

1698858228501.png

And I have a log file for it:

1698858260449.png

Seems Nvidia firmware updater did something at that time and HwInfo crashed because of that. (11:46 is 10:46 in the log because of time change)
 
Last edited:
hello,
i had to set GPUI2CNVAPI=0 (i2c without nvapi),hwinfo would not start without it(or crash opening sensors),might be different problem,using sensors with shared memory for external display...
 
korenchkin said:
hello,
i had to set GPUI2CNVAPI=0 (i2c without nvapi),hwinfo would not start without it(or crash opening sensors),might be different problem,using sensors with shared memory for external display...
Click to expand...

That's most probably a different issue, I'd need to see the HWiNFO Debug File of the problem.
 
So to conclude this, it seems all the events I had where HwInfo crashed were indeed caused by the Nvidia driver being updated or restarted, two updates of the driver, and one NvFlash event, which seemed to have disabled the driver for a short moment, reading through the log. Sorry for opening this thread. Is there no way for HwInfo to not crash if the GPU driver updates? Maybe that is not possible.
 
Well, I'm thinking about how to solve this but all ideas so far are quite complicated and would most likely require a restart of HWiNFO anyway.
Maybe if I could get a HWiNFO Debug File of the crash, I could find some less painful solution...
 
@Martin Cant you reproduce it on your System or do you know own a Nvidia GPU? Where can I find the crash log? Does it contain sensitive private data of my PC?
 
So there is no debug file until I enable debug logging? I would need to wait for the next Nvidia driver then.
 
I did some more analysis and I don't think this is a bug in HWiNFO but in GPU drivers, more precisely in the NVML layer.
This should be currently possible to work-around by disabling "NVML Support" in HWiNFO. Can you try that?
Anyway, I will continue to search for another workaround...
 
I was finally able to reproduce this problem and I think I have a solution. A workaround will be available in the next build of HWiNFO.
 
You must log in or register to reply here.
Back
Top