VR is a lot of times crashing - - - it says the graphics card is the problem, but I have already replaced it?
VR is a lot of times crashing - - - it says the graphics card is the problem, but I have already replaced it?
Hey all, Firstly, I'll give some specs. In brackets are upgrades where I was also still having the issues: CPU: AMD 9800x3d (previously 7800x3d) GPU: Palit 5080 GamingPro OC (previously Merc 310 7900 XTX) GPU Riser: NZXT vertical GPU mounting kit Mobo: NZXT N7 B650E RAM: 2x16GB Corsair Vengeance 6000 CL30 PSU: Refurbed Corsair RM1000e FMod 80PLUS Gold VR: Meta Quest 3 Peripherals: Steering wheel, pedals, haptics..etc all on individual USB ports OS: Windows 11 Pro (26100.3915 24H2) GPU Driver: Geforce Driver 576.28 (though just seen that 2 days ago 576.40 was released, so going to upgrade) I primarily play iRacing, and don't play enough other games over PCVR to know if it's specific to the game, though have had enough crashes in other non-VR games to still raise the concern (Rust primarily). When I had my 7900xtx, occasionally whilst racing using PC Link (USB) my VR would crash the PC. From inside the headset, I'd get black borders and I could still look around (very laggy), but the image I was looking at was static. At the same time, my steering wheel would lock full left. Sometimes after around 20 seconds, my PC would recover and I'd get a GPU driver timeout error. My VR would still not work until a pc reboot either. If I went to EventViewer on these occasions, it showed me a LiveKernelEvent 141 or 142, and iRacings own logs would suggest that my GPU "Device Hung". A workaround to calm this down seemed to be to uninstall the full AMD Adrenaline package and go back to bare bone driver install, which still made it happen occasionally but seemingly less. Could have been pure coincidence. I did multiple DDU and reinstalls at the time. I kept up to date with the new drivers as they were released, tried the beta program, and the issue persevered throughout. I eventually got annoyed with it, and seeing that the nVidia range could get better performance on iRacing both through SPS and Foveated rendering, so bought a 5080 thinking it would solve all my issues (and picked up a 9800x3d at the same time), however I carried on getting the exact same crash symptoms and the LiveKernelEvent 141/142s, along with the some errors on EventViewer, which typically would be one or both of the below (I used to get EventViewer errors but since my wipe I can't see them of course): Somewhere around here I thought that maybe it's because, despite using DDU, I still had remnants of the old AMD drivers so did a full fresh install of windows and redownload of everything, yet it persevered. I've tried overclocking (both manually and from nVidias auto tool), upping power limit to 105%, underclocking (-200MHz on both clocks), lowering power limit to 95%, and nothing seems to help. I figured that being as I have a lot of USB peripherals, between steering wheels, pedals, haptics...etc. and that the VR headset was running on USB too, maybe it was a USB bandwidth/power issue. I bought a powered USB hub to run it through and made sure my VR was straight to a mobo port, however still continued. I've now started trying to run through Virtual Desktop over wi-fi. I have a great connection seemingly and for the most part I get good performance. However, I get the same crashes again. They seem to come out of nowhere; it's not like my frame rates start going up or anything.. it runs smoothly up until the exact moment everything just freezes. This time I'm still getting the nvlddmkm error, and also got this appear in the bottom right (though can't find a log where I can read the full message after it disappeared): At the time, it closed the game and crashed SteamVR, but everything else carried on running. I'm at a loss. So in summary, I'm still getting issues that look seemingly related to GPU despite having a new GPU from a different brand, installed on a fresh wipe of windows, and utilising a VR headset across both USB and/or wifi. I'm not fully convinced it's the games fault as I do get crashes occasionally across other games under seemingly higher load. These freezes typically tend to hang the whole PC and require a power off to recover. I've googled around and seen a few things. DDU and reinstall (done), change tdrdelay value (done), update drivers (done). Other's mention disabling XMP (no idea what this is), or PSU problems... I suppose my next steps are swapping PSU/Mobo, but I don't really want to carry on throwing money at it. What's the best way of determining once and for all what the issue is? I've ran stresstests/benchmarks, albeit not through VR, and get no issues at all. Is there any more detailed logging software I can run that will give me a definitive answer on what is causing the crash?
I ran an OCCT power supply test and it passed with no errors found. Yeah, that was a cheap way to save money or whatever it cost me back then, but it seemed good at the time. Are there any ways to check things while playing games and see if my power is dropping or spiking? Or are these changes so quick they aren't caught by loggers? I feel like I'm stuck in a guessing game where I have to figure out exactly what component is failing or which setting caused the crash, but I thought some software would just tell me right away.
Is this a new power supply? That's not really what matters here. Is there not enough electricity going to my graphics card?
Virtual Desktop is often full of bugs. Can we try using SteamVR instead? Maybe lowering how many cars show up will help if things get too graphically messy for me.
Before using Virtual Desktop (which used to run through SteamVR), I was using Oculus Link but still had problems. Max Cars is already at its lowest setting. I tried using AI and searched online, which led me down a path of analyzing the watchdog LiveKernelEvent dmp file. After pasting that into AI, here came the output:
While the errors above show it wasn't just the definitive thing I was hoping for, they did lead to some more discussion about settings. Based on that and the advice that followed, I have made these changes so far today: Disabled Hardware-accelerated GPU scheduling Installed AMD chipset drivers Added a Tdrdelay registry key with a value of 10 Disabled Nvidia audio in device manager Disabled SteamVR as the OpenXR codec, instead using VDXR turned on Performance mode in the nVidia Control Panel's power management.
I know what fixes it would be great, but I am more focused just on fixing it now than spending money right now. So don't worry about doing multiple things at once until a crash happens later. I might eat those words if they make things worse. I'm going to do some races tonight and hope I get a gauge to see if it's fixed or not, with the intention of doing nothing else until the next crash.
Other suggestions from here and AI/google are: New PSU Remove the vertical riser for the GPU Disable XMP
Thanks for posting about this. You're way more desperate than you look, especially since everyone else seems stuck in the exact same spot with the same upgrade path. I've tried everything several times—multiple reinstallations, even switching to a virtual desktop from link. It keeps crashing randomly and gives up on support advice. Has anyone actually fixed it or are we still stuck here? I'll keep this thread updated if something works for me but honestly, putting money and time into this setup is a huge waste because my whole iRacing thing just doesn't work anymore.