Driver power state failure error occurs with ntoskrnl.exe.
Driver power state failure error occurs with ntoskrnl.exe.
The system has remained stable since the GPU was removed and the SSD was disabled. Although the issue persists, there seems to be another approach to test hardware health beyond safe mode, which I prefer for personal use. I’m sharing this LatencyMon report, as the PC screen went blank during testing and then came back when I tried to restart it. This could indicate a spike in latency, possibly linked to the 4080 chip or gaming activity. The readings reached as high as 2500-3000, which is unusual under load. Since I’m less familiar with latency issues, it’s possible this is causing all my system problems.
It's pointing out that the graphics driver is the oldest running DPC, which might explain the ongoing graphics problems. It appears there could be an issue in the graphics section still present. Could you attempt using the graphics card in a different PCIe slot? I just reviewed your initial message and noticed you've already installed both Windows 10 and Windows 11 from a bootable media. Was that done via a custom install? If so, Safe Mode might not be useful since continuing with a clean install still shows the problem, indicating a hardware concern.
If you installed Windows another way, try reinstalling from a bootable media, selecting a custom installation, and then deleting all UEFI partitions. Look for the unallocated space that appears and click Next. The installer will set up the right partition structure and install Windows.
I usually rely on bootable media (USBs) for installations and formatting drives where I install Windows. Anything essential I need is backed up elsewhere, but those are mostly personal files and not running software.
I’ll check what GPU-Z reports, since this could be a PCI issue I’ve seen online, even though I’m skeptical.
I’m confused about it, and honestly I’m starting to think the problem might be with the motherboard—I’ll need to confirm if my GPU fits in the larger slots.
I am currently transferring a BSOD file to my Google Drive, and it occurred again with the 3080 this time. I am also uploading a memory dump, though I don’t understand what it displays. I was playing GTA5 when it happened. Since the problem isn’t related to the graphics card but is blamed on it, I’m wondering if replacing all my SSDs this weekend might help. If that doesn’t resolve it, I’ll need another Windows installation, which would mean a fresh install to test, and hopefully that fixes the issue. Otherwise, I’ll focus on the motherboard. DumpandMemoryDUMP.rar
drive.google.com
Another update has been posted.
The image clearly shows the problem, and it’s still swapping my SSDs while bringing the 4080 back into the case. It looks like the main issue was a BIOS problem. Hopefully, fixing the micro stutters with a chipset driver update will help, and we can revert to Windows 10 once the new parts arrive.
The problem in that latest dump - the 0x133 with an argument 1 value of 0x1, indicating that a collection of DPCs ran for too long - needs the kernel dump to debug properly, and you provided that so thanks!
Without going into boring details we use the kernel dump to dump the event trace buffers and then writing out the DPC/ISR event log entry as an etl file that the Windows Performance Analyser can read. In there we look at the DPC/ISR times and sort the collection of DPCs on total run time, the one at the top is the longest running. Here's the WPA output for your kernel dump...
https://imgur.com/a/b2RFTDJ
The DPC at the top is nvlddmkm.sys with a run time twice as long as the next longest running DPC (tcpip.sys). The problem here is clearly graphics related - and we already know it's a hardware issue.
You've tried multiple drivers using DDU each time. You've tried a different graphics card (using DDU again). And you still have the problem, so we need to look at other hardware. That leaves the motherboard or the PSU as the most likely - and the only way to test those is by swapping them I'm afraid.
I'll monitor for updates, as the newest BIOS with AGESA 1.0.0.8 has reduced these spikes in LatencyMon significantly. I've observed one spike reaching 4000 in just three hours of heavy use within the user latency section, compared to spikes up to 5000 within ten minutes of LatencyMon activity. DPC spikes have stayed near 1000 from Nvidia's driver under gaming load.
If this problem persists, the motherboard may need replacement, and the power supply unit is unlikely to be the issue—unless they shipped a defective PSU. :/
Thanks for your assistance!
I woke up in the night thinking about your problem. Don't worry, there's no additional charge for out of hours consultations!
🤣
Whilst nlddmkm.sys is the longest running DPC there (by far) the DPCs for tcpip.sys and Wdf01000.sys do run longer than Microsoft recommend - 100 microseconds. Wdf0100.sys runs for almost 600 microseconds and tcpip.sys for over 600 microseconds. That got me wondering whether the games in which you have problems access the Internet in real time? If so, it's possible that a networking problem is slowing the graphics operation down, making nvlddmkm.sys run for too long and the collectively the cause a 0x133 BSOD?
The tcpip.sys driver is a Windows driver, but lower down in the driver stack will be your third-party network adapter driver. Check that the latest driver for your network adapter is installed. Better still, try a different adapter. If you connect via WiFi try connecting via cable (even if it means moving the PC temporarily). If you connect via cable then try connecting via WiFi. If it's a networking issue, swapping adapters should make the effects go away.
The Wdf01000.sys driver is the high-level Windows Driver Foundation driver, a whole host of third-party drivers are written based on the WDF libraries and any one of these lower down in the driver stack could be causing problems that might end up slowing nvlddmkm.sys down, especially if they are in the network stack too (and thus delaying tcpip.sys). Try temporarily disabling (unchecking) all third-party drivers in the network adapter properties box, especially things like security products etc.
I might be barking up the wrong tree here. I might even be barking in the wrong forest! But if the problem is not where you're looking then it must be somewhere else...
🤣 it seems it’s not ideal that it wakes you up from sleep!
Mostly the games I play are always online, I only play singleplayer a few times like Spiderman, Uncharted and RDR2. Among those three, RDR2 has the worst micro stutter while still being the most optimized.
I’m planning to update the drivers as well. I’ve turned off the wifi feature on my motherboard because I use a cable connection (glassfibre) and prefer the full 1 gigabit. But I might try again once the new SSD is installed with Windows 10 back on.
I’m downloading drivers from Realtek:
https://www.realtek.com/ja/componen...
as their vendor hasn’t released updated versions yet. The MSI site also has the same problem at least.
UPDATE: I’ve got the latest version