Forums » Suggestions
Is it a bug ?
I don't know if this is a bug or just me clicking to fast to exit.
I can reproduce this.
If I click Quit too fast after Logoff, I get a read memory error poppup window.
Logout, wait a few seconds then Quit no errors.
Thanks
I can reproduce this.
If I click Quit too fast after Logoff, I get a read memory error poppup window.
Logout, wait a few seconds then Quit no errors.
Thanks
I have consistently experienced client crashes with this on Linux.
I don't know if this is a bug or just me clicking to fast to exit
Well, it's definitely not a Suggestion?
This would be more appropriate on Bugs. I'll move it there eventually, once a little time passes so people know where it went.
Well, it's definitely not a Suggestion?
This would be more appropriate on Bugs. I'll move it there eventually, once a little time passes so people know where it went.
We released a client patch with additional debugging info to track this down.
How often does this happen?
Please send reports via the BugReporter.
How often does this happen?
Please send reports via the BugReporter.
I can replicate this anytime I logoff then click too fast. Like this click, then click boom get this message.
However, if I loggoff and wait maybe 5 seconds to click quit I NEVER get error.
The memory message says something like DXGIWatchDogThreadWindow - Application Error
The instructions at 0xbunchofnumbers reference memory at 0xbunchofnumbers The memory could not be read.
Something like that. No big deal as long as I wait a few seconds first and it's not causing a problem for me but just thought you might want to know.
Windows 11, i9 7900x RTX 2080ti, ssd, 32gb
Sort of an old machine now but plenty fast still and windows 10 was upgraded automatically by windows to windows 11 etc. Just fyi.
However, if I loggoff and wait maybe 5 seconds to click quit I NEVER get error.
The memory message says something like DXGIWatchDogThreadWindow - Application Error
The instructions at 0xbunchofnumbers reference memory at 0xbunchofnumbers The memory could not be read.
Something like that. No big deal as long as I wait a few seconds first and it's not causing a problem for me but just thought you might want to know.
Windows 11, i9 7900x RTX 2080ti, ssd, 32gb
Sort of an old machine now but plenty fast still and windows 10 was upgraded automatically by windows to windows 11 etc. Just fyi.
If you can please continue triggering the issue, and then hopefully the Bug Reporter pops up, then submit the report and hopefully we can glean more information on the problem.
Although, as Google says, this is likely to be related to drivers, from their "AI Overview":
A "DXGI WatchDog Thread Window - Application Error" message usually indicates a problem with your graphics card or its drivers, specifically related to the DirectX Graphics Infrastructure (DXGI) component that manages communication between your application and graphics hardware.
What it means:
Watchdog timer triggered: The "Watchdog" part refers to a monitoring system within DXGI that checks for prolonged inactivity or errors in the graphics pipeline. When it detects an issue, like a driver crash or a stuck rendering process, it triggers this error message.
I'm not a huge fan of Google's AI overviews, as some of them are super wrong. But this one is reaching the same conclusion as I did after reading a bunch of deeper reports and articles. There's no question the subsystem involved as driver-related, and the watchdog failure is pretty suspect.
Once you submit some more bug reports for our records, I would try upgrading to the latest NVIDIA drivers?
Or, alternatively, you could switch to using the Vulkan renderer, which may stop using the DXGI driver codepath entirely, at least on x86-64 (but probably not on ARM64).
Although, as Google says, this is likely to be related to drivers, from their "AI Overview":
A "DXGI WatchDog Thread Window - Application Error" message usually indicates a problem with your graphics card or its drivers, specifically related to the DirectX Graphics Infrastructure (DXGI) component that manages communication between your application and graphics hardware.
What it means:
Watchdog timer triggered: The "Watchdog" part refers to a monitoring system within DXGI that checks for prolonged inactivity or errors in the graphics pipeline. When it detects an issue, like a driver crash or a stuck rendering process, it triggers this error message.
I'm not a huge fan of Google's AI overviews, as some of them are super wrong. But this one is reaching the same conclusion as I did after reading a bunch of deeper reports and articles. There's no question the subsystem involved as driver-related, and the watchdog failure is pretty suspect.
Once you submit some more bug reports for our records, I would try upgrading to the latest NVIDIA drivers?
Or, alternatively, you could switch to using the Vulkan renderer, which may stop using the DXGI driver codepath entirely, at least on x86-64 (but probably not on ARM64).