New
#1
Display freezes need to reboot, GPU issue/relating to nvlddmkm??
Hi, been sent here from the Win11 sister forum where this thread is essentially about the same issue I'm having - just for Win11
I've been having an issue on my 4 year old Win 10 rig that's been going on intermittently for several months now (sometimes it happens multiple times in a day, sometimes once a week, sometimes goes away for longer but it always returns!). See below a list of symptoms, troubleshooting steps done as well as some screenshots and my V2 collector log.
Any help would be greatly appreciated.
V2 collector log on Google Drive
SYMPTOMS
- Screen freezes regardless of system load (happened when gaming, when completely idle, when only having base apps like browser + Spotify open, or also while not doing anything load intensive)
- Fans, drives, motherboard etc all still running - other evidence of activity: Sometimes sound will still play for a while, or sound will loop the last few seconds for a while, or sometimes when pressing buttons it will still make Windows sounds showing some OS functionality still running just graphics being dead(?)
- Cold reboot needed to recover
- Sometimes display won’t come back after cold reboot - dual monitor setup only showing “no signal” while fans and other hardware are running/LEDs on/showing activity - in that case I usually have to unplug one or both of my monitors from the GPU (1x HDMI, 1x DPort) and try another cold reboot. Can plug both monitors back in afterwards
- Sometimes there will be a few seconds of stuttering frames and audio before the crash
- Sometimes there will be a Windows pop-up that app.exe (e.g. Overwatch, Discord, Firefox) has been blocked from accessing hardware. This may not always lead to a full crash but the app affected will crash
- nvlddmkm driver errors shown in Event Viewer in some cases right before the crash or before the Windows pop-up
- Other times there is no log evidence in the files, last log entries are usually a few minutes before crash
- I noticed that while temperatures/power usage doesn't seem to be abnormal, logs for GPU show some weird data values moments before the crash happens (e.g. temp logged as 0 so assume the machine doesn't receive any data from sensors) and logging resumes normally after a successful reboot. I use OpenHardware Monitor. See screenshot of data in question from a crash that happened today (21/01/23).
Logging data:
Eventviewer logs for the same time stamps:
Troubleshooting done so far:
- Completely reinstalled drivers multiple times, including the use of DDU from safe mode, multiple versions of Nvidia drivers over the last 6+ months
- Updated drivers for pretty much anything else available
- Opened up my computer and checked that all parts are seated correctly, dusted off parts (my setup is designed quite well, barely had any dust!)
- Tried a few combinations of power plugs and power sockets in case of dodgy power setups
- Changed power options to high performance in case it’s due to power issues
- Changed timeout + recovery settings in registry as per How to Fix ‘Display Driver nvlddmkm Stopped Responding’ on Windows 10/11
- Turned off hardware acceleration in various apps incl Windows overall, browser, discord and games that use it
- Reseated GPU
- Checked that GPU isn’t overclocked
- Underclocking via MSI Afterburner
- Temperature checks/logging didnt seem to be a temp issue although GPU is running on the warmer side when gaming (Overwatch/Cyberpunk up to 83°C), after lowering graphics settings in Overwatch runs at around 50°C
- Reinstalled Overwatch as it’s the game I play the most
- Disabled a few gimmicks from Geforce experience (Shadowplay, ingame overlay etc)
- Memtests, multiple - came back ok
- Removed lots of crap/superfluous apps from PC, incl most apps that could interfere with GPU settings and that I no longer need
- Calculated PSU capacity required for the setup I have: I have a 600W BeQuiet Bronze, Estimates vary from 300W upwards, only reaches 600W if I massively overcalculate everything just in case? However this does not explain crashes when the PC is almost idle?
- Scannow - sometimes came back with errors but resolved itself in between, still did a few windows repairs via DISM just in case
- Looked at sensor logs for temperature anomalies/power surges etc - only abnormalities are around GPU sensors, see screenshot - temperatures not higher than usual/or definitely not high enough) - but noticed that temperature logging had stopped / a temperature of 0 was logged a few mins before crash (e.g. logged half an hour of temps around 45°C for GPU, then 5 mins before crash only 0°C, then logging resumed after a few minutes when the reboot was complete) - see screenshot above
Haven’t tried yet:
- Pretty much anything BIOS related - not feel comfortable
- Flash GPU - dont want to risk
- Complete clean reinstall of Windows
- Replacing parts like GPU, PSU or monitors (might test a new monitor, want a new one anyway..)
Last edited by Nuku2; 21 Jan 2023 at 14:03.