Question DRIVER_POWER_STATE_FAILURE (BSOD) ntoskrnl.exe ?

[Sterrenstoof]

Hi!

I've been facing issues with my PC recently, and I can't seem to pinpoint the cause. My PC has been shutting off the past few days regularly at around the timestamps 23:30pm till 1:30am whilst gaming.
The issue has pursued me across two Windows installations now as I tried to reinstall my Windows to see if my issues would disappear, as I have been having frequent stuttering problems in games, and thus I think this bluescreen is some sort of indication as of why my system is having these instability issues... and maybe the minidump can lead somebody to a cause? Cause I can't get my head around it.

My setup:
OS: Windows 11 Pro
GPU: MSI RTX 4080 Suprim X
CPU: Ryzen 7800X3D
RAM: G.Skill Trident Z5 Neo RGB F5-6000J3040G32GX2-TZ5NR
MOBO: MSI x670-P Pro
PSU: Corsair RMX Shift 1200W

I've tried the following:
• Clean install of Windows 10 & 11
• Updated all drivers, and tried rolling back some drivers to pin-point the problem
• Run the latest BIOS revision by MSI
• Disabled EXPO to see if the instability miraculously disappears but it hasn't.
• And this: https://www.tomshardware.com/how-to/fix-driver-power-state-failure-error


I want to add that my PC's temperatures are fine, 68c GPU and 65-74c whilst gaming, though under heavier load situations the CPU can run up to 84c under a Be Quiet AIO.

If anybody with the skill of debugging my minidump file here can potentially see a cause of what hardware in my case is causing this I would greatly appreciate it, the instability has been giving me headaches and not knowing what could be the cause makes it even more frustrating as this system isn't old at all, the GPU is only a recent addition! :/

Here's a link to my minidump as this is the first time my PC created one for a while, though those blackscreen freezes that suspiciously look a lot like a bluescreen crash have been happening more often with no dumps created if anymore get created I can upload them: https://drive.google.com/file/d/1wJcD6_goY9H6boGHHUxLXwZyzOPDOobI/view?usp=sharing

 

[ubuysa]

Are you sure the ussues that caused the PSU replacement weren't actually motherboard issues?

 

[Sterrenstoof]

Yeah no, but I don't remember stuttering as bad as I do now. Though let me add that sometimes the situation is a lot more stable.

I can't really test this without purchasing a new mobo to test it on, which is a last resort after the SSD replacement and complete reinstallation, especially if I get BSODs again (blackscreens) which I haven't had in 3 days.

I'll run HWMonitor again, one time it shown 11.000 on the 12V (gotta say this was a while ago, and idk when it got this reading) but overall things look /okay/ to me.

https://imgur.com/a/buOboia

View: https://imgur.com/a/buOboia

 

[Sterrenstoof]

Tonight the wall sockets will get a checkup on them, to make sure this isn't coincidentally a electrical issue that causes it. I mean it sounds weird but it's still smart to check.

Anyhow, this is what HWMonitor sees after an hour:

https://imgur.com/a/IPWHsFx

View: https://imgur.com/a/IPWHsFx

 

[Sterrenstoof]

Alright well, even with AGESA 1.0.0.8 the crashing still occurs, whilst running the drivers from the motherboards download page.

Once my SSD arrives.. whenever it does cause it's shipment got postponed, I'll reinstall Windows and test again to see if that solves it..

If not I'll wait till Black Friday / Cyber Monday sales, and hope a motherboard drops in price, we're almost there anyway.

 

[Sterrenstoof]

It's been a while since I've posted here and got everybody up to date, and once again thanks to @ubuysa for helping me before.

The following happened during the last 3 months:

Late October
All electrical wiring was replaced in the house, outlets were replaced with ground connected ones etcetera, my setup has ground now and fresh electrical cables.

November
I've tested two different Power Supplies, my old one and my new one on both a RTX 3080 and RTX 4080 and still faced the same issue of my PC just turning off during extensive gameplay sessions, though when playing a game slightly heavier like CS the PC would often crash within a time span of 30-45 mins, making it way easier to reproduce.

Late November
I've put back my other GPU into the case (4080), as there was no difference between the two crashing just the exact same making the suspicion that something else in the PC was wrong, directly aimed for the CPU, though it being heavily unconfirmed and during stress tests Prime95, Cinebench you name it the CPU wouldn't crash, and the RAM neither shown any sort of errors during memory tests like memtest, test this with two different kits.

December
I spent my time with family, whilst frustration about my setup still not functioning the way it should when I had time to game on it.. but not much I could do as just buying parts whilst still not knowingly exactly what could be the culprit but there's just no time for it.

January
I finally shot myself in the foot, and listened to some feedback around myself to try a entirely different system so I ordered a Intel based system with a i9-14900K, and a Z790 board from ASUS, as well as a new kit of RAM to try my luck, I did keep my original cooler and fans and GPU, so the MSI RTX 4080 Suprim X is still part of the setup and so are my Lian Li fans, and my Be Quiet Cooler as well as the Corsair RMx Shift 1200W.

Well while it felt better the first 1 hour of gaming, I mean it feels better overall I managed to ONCE again crash the PC during a gameplay session, this time it created no dump files, just like it did before so I shrugged it off and just went to try again, just to face the same crash, so I put the game to windowed borderless and guess what at this point only the game crashes ??? This is basically the same behavior I experienced on the other setup.

Yesterday a few people wanted to game, and I hopped on too and I had to endure at least 2 BSODs with a dump file created, and 10 game crashes the moment something graphical intensive happens the PC <Mod Edit> the bed basically, I've uploaded the dump files for anybody to read out like @ubuysa did, but my suspicion is that it's the DPC taking way too long once again on the GPU and I might be one of the unluckiest people to walk around to be owning not 1, but 2 faulty GPU's lol.

https://drive.google.com/file/d/1E6ywy34kVeXM7i0K34PtrvB8io-A0Zlb/view?usp=sharing

It be nice if somebody can confirm it's the GPU through the dumpfiles available, though it being the exact same behavior as of previous it must be, and thus will I probably RMA this product, bit upset that both my graphic cards have the same though. Especially with the 4080 barely being 6 months old now.. :/

EDIT:
I disabled the ASUS boost, and that removed one of the BSODs, seems to be some default mobo settings ain't fantastic, still sitting with a load of stutters. Gonna test tmrw a GPU that isn't mine to see if it results in the same deal, because that would mean something else is causing it.. and that would mean I need to look at the peripherals which I've done already once, but could've sworn stuff was alright..

 

[Sterrenstoof]

Update, 3rd GPU tested same result a DPC_Watchdog_VIOLATION within a timespan of 3 hours, at this point I replaced my entire PC and it might be in the peripherals where the issue underlies, though my keyboard and mouse are barely a year old, and my monitor too... I might as well just quit trying entirely cause this is pointless.

 

[ubuysa]

The two dumps you uploaded were CLOCK_WATCHDOG_TIMEOUT bugchecks (0x101) and we really need a kernel dump to debug those. A minidump only contains status for the processor that took the BSOD and for 0x101 that's not the processor that failed, that's why we need the kernel dump so that we can look at all processors. Sadly there is only ever one kernel dump stored, for the most recent BSOD, and your latest 0x133 BSOD will have overwritten the 0x101 kernel dump.

The 0x101 BSOD happens because a processor failed to respond to the clock synchronisation interrupt. It is remotely possible that a flaky driver could have blocked that processor and masked it for interrupts causing the 0x101, but that's about as rare as hen's teeth. Most usually a 0x101 BSOD indicates a flaky processor and thus a faulty CPU.

Try downloading the Intel processor Diagnostic Tool and see what that has to say about your CPU. You might also run Prime95 to stress your CPU, but beware that this WILL make it run hot, so monitor temps. Ask if you've not run Prime95 before.

If you see another 0x101 BSOD please copy the file C:\Windows\Memory.dmp to a temp folder somewhere to preserve it in case you get a second BSOD. Then upload that dump file (that's the full kernel dump). That will allow us to see (hopefully) whether the failing processor was blocked by a flaky driver.

Regarding your most recent 0x133 BSOD, a flaky processor can cause all manner of random bugchecks, and now that you've been having 0x101 it's important to be sure your CPU is good before looking elsewhere.

BTW. Make sure there are no overclocks or turbo modes enabled, and not just on the CPU, also disable any XMP profiles you may have on your RAM and run that at it's base frequency.

 

[Sterrenstoof]

Figured that much after I had it a third time, a quick google led me to the Asus Enhanced BOOST option, which I've disabled. Haven't experienced this one for a bit, but though the DPC_WATCHDOG_VIOLATION, which blames the GPU's is back, 2x RTX 3080 and 1x RTX 4080, making me believe it has nothing to do graphically, or with the entirely newly built PC at this point.


I've rebuild the entire PC, to literally face the same issue making me believe it's a peripheral at this point, though funnily enough I'll be rebuilding my old PC which had this exact same issue this weekend, plug it in for testing and see what that does, the only main differences between the two will be one is AMD, the other is Intel but their cooler, and their fans are different brands, plus the case. I'll be wildly amazed if that somehow fixes it.

I'd the unluckiest person walking this planet when it comes to hardware if somehow this CPU is already flaky.. Prime95 ran fine last time I tested it, it's the first thing I did in attempt to crash it, it makes the PC lagg for sure but with the all core load that seems sensibile.

I'll see this weekend or next weekend, my schedule is unpredictable and I just can't tell when I get around to things testing, it's a huge disappointment and I am investing more of my free time diagnosing and trying to fix this godforsaken setup, than that I am enjoying it. I'm at a point of selling it all and not looking back.

EDIT: The CPU passed the diagnostic test from Intel.

 

[ubuysa]

Upload the 0x133 dumps if you like and I'll take a look.

Perhaps you might try building the bare minimum, just a CPU, cooler, one RAM stick, one drive and nothing else. Do a clean Windows install with all updates and drivers but with nothing else installed. See whether that's stable. If it is, then add components one at a time, testing carefully after each one. If it's stable with all components added then install third-party apps and devices one at a time, testing thoroughly after each one.

 

[Sterrenstoof]

It's only happened once, last night cause I barely have time to recreate it and it always happens during fullscreen gaming, and not borderless windowed which instead crashes the game to desktop. It always happens after 3-4 hours of gaming, same as on the previous setup you've analyzed the dump files for.. at this point having replaced it all, tested multiple kits of ram, 3 different GPU's, 2 motherboards, and both a AMD and Intel CPU, and two seperate PSU's, I do not see myself undo the entire PC again just to go testing each individual item bare, as I don't think this issue is within my computer, and something else is conflicting it to cause this behaviour from outside the case, inside the peripherals or even over the LAN. Unfortunately the LAN I have not much say in, so if it's in there it's books closed.

It's either electric (power grid) related, which I doubt due to all the cables that been replaced but heck.. I dunno, at this point it can be anything, but I just can't twist my head around the fact that it really likes to blame the GPUs, no matter which one just is stupidly weird, at this point I am not even gonna exclude the monitor even though that sounds ridiculous.

https://drive.google.com/file/d/1uCovykMbUh_ogK9jkM2GMfOHvGp5FWaR/view?usp=sharing

I feel like this isn't fixable, and having tried to troubleshoot it for 6 months its came to a point that it causes stress, rather than joy in this hobby.

I'll be reconstructing my old PC soon, when I got time on my hands and probably drop a update here if it's behaving any different than the current one, which well if it does will give me hope that it's fixable, if it's just the same bs, then idk. 😀

PS: Thanks for helping btw c:

 

[ubuysa]

Fair enough on the rebuild, I'd feel the same way TBH.

Just a thought. Is the video cable (HDMI/DP/etc.) leaving the PC routed anywhere close to WiFi or Bluetooth antennas? I have seen a problem in the past where the display HDMI cable ran between the two WiFi antennas and that cause graphics issues. It's worth ensuring that the graphics cable is nowhere near any radio sources.

 

[Sterrenstoof]

Yes, there's was a wifi point right next to my setup as my room is located in a crucial location, it's also the only cable coming up to this room which is connected to the point, through that point to the PC.. also the AC unit is right above the PC basically, I mean effectively it cools the PC during the summer but I have no idea if it emits any sort of RF.

Though, today the point was replaced and maybe this one causes less interference? Fingers crossed xD

Not much I got to say in the matter of the wifi point location, but I might be able to one day test my PC in a different space, but that's probably in a few months :/

 

[Sterrenstoof]

Well 1 positive side effect of the replacement of the network parts, my VR headset finally works again lol. I mean the slight stutter is still present no matter what I do, but I can finally play VR so I might be able to spend some time smacking cubes in Beat Saber if I can't find a solution to the DPC stuff, unless it somehow has solved itself.

@ubuysa is it possible that a monitor could cause this issue with the 0x133? I dunno, I am just searching around a haybale but I been wondering if it could cause any weird issues with graphics cards, seeming it's so universal across 3 cards, and 2 PC's basically. Been hearing this loud ticking noise during usage at times, more audible than when you click a pen, at the rear left bottom where also the power plug enters the monitor. I been thinking that it could be the panel warming up causing it to expand and making this noise. But it sounds very loud.

 

[ubuysa]

@ubuysa is it possible that a monitor could cause this issue with the 0x133? I dunno, I am just searching around a haybale but I been wondering if it could cause any weird issues with graphics cards, seeming it's so universal across 3 cards, and 2 PC's basically. Been hearing this loud ticking noise during usage at times, more audible than when you click a pen, at the rear left bottom where also the power plug enters the monitor. I been thinking that it could be the panel warming up causing it to expand and making this noise. But it sounds very loud.

Quite possibly. A DPC is used to run the back-end of device interrupt processing, the DPC code is part of the device driver package. I don't know how the graphics DPC code works of course, but if it waits for the monitor to tell the graphics card that some operation has completed, and that delays the DPC completion then yes, the monitor could be involved.

 

[Sterrenstoof]

Alright, I might test that sometime soon when I can get my hands on a different monitor that has the same refresh rate as mine, or when I get this one replaced as I was thinking of going OLED later down the line.

After reading out my last dump, did you find anything related to the ethernet driver? How's that one doing DPC wise? We've figured out not too long ago that our current modem is having some issue with it's DHCP, after resetting the device it's still been doing that, we claim it to be a faulty modem it's a Fritzbox, so it's being replaced but what if that's been causing my setup to go haywire since August/September?

Would it be possible that something pings/packets my PC from the network, causing the driver to freeze up in which my game bricks up, and causes the DPC_Watchdog_Violation or in other cases a GPUID:100 error? The games I experience it in are online, the moment it's fully offline or I'm running a graphical test that's offline it just simply doesn't do it.

 

[ubuysa]

You mean the Jan 24th DPC_WATCHDOG_VIOLATION dump? If that is your most recent BSOD then upload the full kernel dump - it's the file C:\Windows\Memory.dmp and it will be large. I'll take a look for you to see whether the network adapter is indicated anywhere.

DHCP isn't going to cause BSODs, only a failure to obtain a network address.

It is possible that the network adapter driver is blocking a processor (causing the 0x101 bugchecks) and it's possible that it's causing the 0x133 bugchecks too, but that doesn't mean that it's something coming in from outside. If the network adapter is hanging then that will be because the driver is bad.

We can enable Driver Verifier if you like, that is designed to test every third-party driver as it's loaded. There are a few downsides to enabling Driver Verifier, and it needs to be enabled in the right way. If you want to try that let me know and I'll give you some detailed instructions.

 

[Sterrenstoof]

I no longer have that memory.dmp, if I get a new one I'll drop it here but since I am using the system less idk when that'll be.

I will try a third PC soon, borrowed one and if that one displays the exact same problem in my room? Then it's probably electric related still, or it must be ethernet cause I just tested each USB device individually by removing them, turning on a game and letting it idle for 3.5 hours, and if it crashed it ain't that device for sure.

If that other PC somehow doesn't show any of these symptons and runs entirely fine? This one's gonna find itself with the scrap cause 6-7 months with this bs is just borderline tiring.