So I have a very peculiar problem with my system, more specifically with the GeForce RTX 3090...
Specs: i7 9700K, 32GB RAM, RTX 3090, 970 EVO PLUS SSD, bunch of LL120s with a couple of Corsair's fan controllers, a SATA SSD, 3 HDDs, H110i SE 240mm Liquid Cooler.
I was one of the lucky ones to snag a 3090 FE at MSRP and have been gaming with it without problems until about two months ago (could be even before that, I'm not exactly sure). I've started getting BSODs (mostly whea_uncorrectible_error) but unfortunately none of these BSODs would register as a BSOD (the computer would show the BSOD for a split second or not at all before rebooting) or at least that's my theory, because the memory.dmp would never show these crashes when I checked them with the BlueScreenView app.
CPU temps only go up to 80-90 degrees during the toughest of the torture tests (and when the fan profile was set to quiet), GPU temps never exceed 70 degrees. (Memory junction is another story unfortunately, but as far as I can tell, it's around 80-90 when the crashes happen.)
Furmark, Prime95 would run fine for up to an hour (I've never had the patience to check longer because whenever something went wrong it happened within 20 minutes max.), but OCCT power test caused similar BSOD crashes and the PC wasn't able to pass any of the 3D Mark tests. So I started testing different components. Like I said CPU would complete benchmarks fine, swapped out RAM sticks and the problem was still there, but than I swapped my 3090 for a 2080 Ti and voila, no more crashes, everything was fine, could play all my games.
Before swapping the GPUs I checked the event viewer and all the crashes would happen after "File System Filter 'npsvctrig' (10.0, 2025-01-06T05:41:12.000000000Z) has successfully loaded and registered with Filter Manager." event happens. There's a lengthy thread about this somewhere but it unfortunately doesn't have any final solutions to the problem. Most suggestions I've read up to this point suggested a potential PSU problem, which made sense as the 2080 Ti uses less power than the 3090, so I suspected (also didn't want to think the problem lies with my 3090). Also my RM1000 was 7 years old at this point so I decided what the heck and ordered a 850 Watt power supply.
At first everything is fine and the problem looks like it's gone away. Previously failed 3DMark stress tests (FireStrike, Port Royal, etc.) are scoring 98-99%. One of the most problematic titles, Microsoft Flight Simulator is running smoothly. But after a few days, bam, the PC crashes again and after that things start happening more frequently as before. The only difference as opposed to the previous PSU, is that the PC can now pass the various 3DMark stress tests. But it crashes after 20-30 mins on Doom, after 10 seconds on Cyberpunk 2077 and only runs titles like Stardew Valley and Hades without any issues.
I turn towards the OCCT once more, power test, either crashes within 20 minutes or starts logging CPU errors, CPU and Memtests are fine, shader test is fine, VRAM test starts logging insane number of errors after a certain point, in the billions, but doesn't register any errors in the next run. I am about the go crazy at this point. So still not wanting to think that my 3090 is the problem, I borrow a 3090 from one of my friends, same exact FE model and having tested the system with the 2080 Ti the same day, with great results once again, the second 3090 also causes the exact same errors!!!!!
Since my buddy didn't have any problems with his 3090, I am now led to believe that even the 850 watt PSU is struggling with the 3090, which is insane and definitely shouldn't be any problem as even the wildest estimations calculate around 725 Watt power draw for my system. Oh, I haven't formatted the system yet, I hate formats, but this honestly feels like a hardware issue as all the BSODs I get are wheas.
So TL
R --> Started having crashes in graphically demanding games, the more demanding the game, the sooner the crash occurred. Dropped down to trusty old 2080 Ti, everything was fine. Switched PSUs, went back to 3090, problems started happening again, switched back to 2080 Ti, everything was fine. Switched to a DIFFERENT 3090 and crashes came back. No OC whatsoever, I even turned off the XMP profile. BIOS and all the other drivers are updated. DDUd multiple times. Tried all your TROUBLESHOOTING 101s, except formatting.
Specs: i7 9700K, 32GB RAM, RTX 3090, 970 EVO PLUS SSD, bunch of LL120s with a couple of Corsair's fan controllers, a SATA SSD, 3 HDDs, H110i SE 240mm Liquid Cooler.
I was one of the lucky ones to snag a 3090 FE at MSRP and have been gaming with it without problems until about two months ago (could be even before that, I'm not exactly sure). I've started getting BSODs (mostly whea_uncorrectible_error) but unfortunately none of these BSODs would register as a BSOD (the computer would show the BSOD for a split second or not at all before rebooting) or at least that's my theory, because the memory.dmp would never show these crashes when I checked them with the BlueScreenView app.
CPU temps only go up to 80-90 degrees during the toughest of the torture tests (and when the fan profile was set to quiet), GPU temps never exceed 70 degrees. (Memory junction is another story unfortunately, but as far as I can tell, it's around 80-90 when the crashes happen.)
Furmark, Prime95 would run fine for up to an hour (I've never had the patience to check longer because whenever something went wrong it happened within 20 minutes max.), but OCCT power test caused similar BSOD crashes and the PC wasn't able to pass any of the 3D Mark tests. So I started testing different components. Like I said CPU would complete benchmarks fine, swapped out RAM sticks and the problem was still there, but than I swapped my 3090 for a 2080 Ti and voila, no more crashes, everything was fine, could play all my games.
Before swapping the GPUs I checked the event viewer and all the crashes would happen after "File System Filter 'npsvctrig' (10.0, 2025-01-06T05:41:12.000000000Z) has successfully loaded and registered with Filter Manager." event happens. There's a lengthy thread about this somewhere but it unfortunately doesn't have any final solutions to the problem. Most suggestions I've read up to this point suggested a potential PSU problem, which made sense as the 2080 Ti uses less power than the 3090, so I suspected (also didn't want to think the problem lies with my 3090). Also my RM1000 was 7 years old at this point so I decided what the heck and ordered a 850 Watt power supply.
At first everything is fine and the problem looks like it's gone away. Previously failed 3DMark stress tests (FireStrike, Port Royal, etc.) are scoring 98-99%. One of the most problematic titles, Microsoft Flight Simulator is running smoothly. But after a few days, bam, the PC crashes again and after that things start happening more frequently as before. The only difference as opposed to the previous PSU, is that the PC can now pass the various 3DMark stress tests. But it crashes after 20-30 mins on Doom, after 10 seconds on Cyberpunk 2077 and only runs titles like Stardew Valley and Hades without any issues.
I turn towards the OCCT once more, power test, either crashes within 20 minutes or starts logging CPU errors, CPU and Memtests are fine, shader test is fine, VRAM test starts logging insane number of errors after a certain point, in the billions, but doesn't register any errors in the next run. I am about the go crazy at this point. So still not wanting to think that my 3090 is the problem, I borrow a 3090 from one of my friends, same exact FE model and having tested the system with the 2080 Ti the same day, with great results once again, the second 3090 also causes the exact same errors!!!!!
Since my buddy didn't have any problems with his 3090, I am now led to believe that even the 850 watt PSU is struggling with the 3090, which is insane and definitely shouldn't be any problem as even the wildest estimations calculate around 725 Watt power draw for my system. Oh, I haven't formatted the system yet, I hate formats, but this honestly feels like a hardware issue as all the BSODs I get are wheas.
So TL
R --> Started having crashes in graphically demanding games, the more demanding the game, the sooner the crash occurred. Dropped down to trusty old 2080 Ti, everything was fine. Switched PSUs, went back to 3090, problems started happening again, switched back to 2080 Ti, everything was fine. Switched to a DIFFERENT 3090 and crashes came back. No OC whatsoever, I even turned off the XMP profile. BIOS and all the other drivers are updated. DDUd multiple times. Tried all your TROUBLESHOOTING 101s, except formatting.
Facebook
Twitter
Instagram