Reset TDR occurred on GPUID + 4090 gigabyte OC
Posted by jimlaman8c@reddit | buildapc | View on Reddit | 143 comments
Anyone ran into this with a 4090. Black screen, recovers and logs to event viewer. I was just using chrome, nothing else.
In my case windows 11 + a 1000 watt evga SuperNova Platinum + 12900k cpu at 4.9ghz and 6400 ram.
I never had this occur with the previous 3080ti.
I ran DDU and the latest nvidia driver.
Any thoughts?
distracted_seagull@reddit
just started having this issue with my 2070 super recently. did set my performance to prefer maximum but it's happened once since then. have updated my drivers, see if that improves it. amazing it's happening to so many people without any kind of public statement by nvidia.
just to note for anyone searching, when the nvidia drivers crash event viewer will show 3 rows of errors:
\Device\000000cc Resetting TDR occurred on GPUID:100
\Device\000000cc Reset TDR occurred on GPUID:100
\Device\000000cc Restarting TDR occurred on GPUID:100
then a warning
Display driver nvlddmkm stopped responding and has successfully recovered.
spiralsnowagain@reddit
Still same problem. Anyone have a fix yet
No_Height4816@reddit
I have the same issue, did u find a fix? I have a 4070 super
Elegant_Carpet4359@reddit
I also have a 4070 super, and just started getting this issue. To me, it only happens once my monitor turns off after some time. I'd imagine it's something driver related due to the new update but that's just a guess.
selectexception@reddit
I also have this issue after monitors/display adapter has gone to sleep. It started after Win 11 24H2 update, it persisted over a fresh Windows installation.
assalariado@reddit
I have the same problem in the same situation, but with Windows 10. I disabled the monitor's automatic shutdown after 15 minutes; I hope that solves the problem. RTX3070 / 5800x / 32gb / W10 (22H2).
ZoteTheMitey@reddit
My 4090 keeps crashing nvidia drivers if I am watching a video on one display, then try dragging another window to a different monitor while the video is playing.
screens go black for a second as the nvlddmkm driver crashes and recovers.
Also once or twice in the last couple weeks, nvidia drivers BSOD the whole computer waking up from sleep.
This all just recently started happening. Card has been fine for over 2 years.
Did you find a fix?
I tried rolling back to 560.94 (august I think)...but the issue remains.
Comprehensive_Rise32@reddit
Wow, similar to my situation, I was also playing a video and moving windows here and there for some minutes before the screen turned black, with nothing happening, and several minutes later the computer restarted. It was just one single monitor in use, I was using MPC-HC, which I installed from Codec Guide, with MPC Video Renderer for RTX Video and RTX HDR, and SVP 4 for smoother videos. Also had a slight overclock using Nvidia's Automatic Tuning with the voltage and power sliders set to max.
Never had this happened before, I don't usually keep readjusting windows constantly when a video is playing, so I guess it's not something I have to worry.. about? Hmmm...
And of course I got here searching Event Viewer and plugging in those error codes in Brave Search.
Shinya150@reddit
I have a 5070 and im having this issue out of nowhere. It happened to me 2 times in the last month. I had this card for near a year now. Never had this problem before. Gonna update win 11 to latest and nvidia drivers too. Right now im using 591.74
ZoteTheMitey@reddit
this was an issue with a windows update. Once I got the next windows 11 update, the issue went away and I have not seen it since. about a year or more. I never had issues with the computer restarting. screen would just go black then recover a second later when moving videos between displays
Fan4eG@reddit
same. did u solve it after a year?
I've been struggling with this for over six months now. The only thing I haven't done in all that time is update and reinstall Windows itself (drivers etc don't count), and nothing has helped. Raising the priority using MSI Utility helped a little, but not much. Overall, I think Windows 11 is to blame.
ZoteTheMitey@reddit
Hello, yes mine is resolved I haven't had this issue since shortly after I posted this. I believe it was some kind of issue with Windows 11 and the nvidia drivers. I struggled with this for a long time until one day windows updated and poof never had the issue again.
I would say make sure Win 11 is up to date and make sure nvidia drivers are up to date
Fan4eG@reddit
So it is Windows after all, got it. Thank you!
With the current problems with Win 11, I really don't want to upgrade to newer versions and am sticking with 24H2 xd. Which one are you on?
selectexception@reddit
Try disabling the PCIE link state power management option in advanced power options. Seems to have fixed the wake up issue. Not 100% sure yet.
scatrinomee@reddit
Anything to report? that didnt fix my wake from sleep issue
ZoteTheMitey@reddit
it's already off. The wake issue is very rare for me it happened maybe twice in the last two or 3 weeks. Still too many of course, this card went over 2 years without a single issue until now
wtf is going on? It has to be some kind of windows update that caused this no? I've tried rolling back nvidia drivers and the issue remains. The only thing I can think of is something in windows 11 24H2 is causing it.
Meow81@reddit
The issue is also in Windows 10 x64 last version
selectexception@reddit
Yeah it started with the Windows update. But my main issue was the wake up thingy and it seems to have fixed itself now for me.
ZoteTheMitey@reddit
Yeah it's been a few days since mine crashed waking from sleep.
Do you have multiple monitors? Play a video on one either youtube or netflix or something, and keep trying to drag a different window around to your other monitors while the video is playing.
Mine crashes pretty regularly doing that ever since whatever update windows did broke something.
selectexception@reddit
Yeah, 4 monitors. This has not been an issue for me. Do you have the NVidia control panel setting for maximum performance?
ZoteTheMitey@reddit
nope...not firefox. moving edge window around as the native netflix app played a video crashed the nvidia drivers.
ZoteTheMitey@reddit
yes I've tried that also. I may have narrowed it down to Firefox actually. Switched to edge and it hasn't crashed yet. will need to test further.
Intelligent-Wolf-825@reddit
THIS IS MY EXACT ISSUE NOW. 3080 here Can you DM we can maybe crosstab and see whats common between our setups. Dragging Chrome across monitor (every once in a while) causes the TDR reset for me but then causes a reboot. It also occurred on different browsers so its not just chrome.
Delete132@reddit
did u solve it?
Intelligent-Wolf-825@reddit
Yes. DDU, uninstalled some windows updates, restarted pc then reinstalled an older gpu driver. All fixed
Delete132@reddit
Thanks, I was getting desperate but DDU helped me too!
No_Height4816@reddit
I just used the ddu. Installed every display driver. Restarted and did it again for good measure. Didn’t have the problem since
Intelligent-Wolf-825@reddit
Same
christopherw@reddit
I have a 4090 FE and it's recently begun doing this. Thought at first it was due to stress tests I ran a few times following a PBO adjustment, but even on this stable setup it still occasionally does it - did it while I was simply web surfing just now!
I get the same errors in event viewer. I'm still on Windows 11 23H2 (I'm preventing it from upgrading to 24H2 because I use a WMR VR headset), I also DDUed the drivers in Safe Mode last night and reinstalled fresh to see if that would prevent it. I do not use the Nvidia app.
I'm now considering DDUing and rolling back to even older drivers, something like 566.36 (Feb 2025) or even 551.86 (March 2024), because I can't recall this happening on those older builds.
Meow81@reddit
Did you solve ? fix it?
christopherw@reddit
I ended up DDUing and rolling back to 566.36 (566.36-desktop-win10-win11-64bit-international-dch-whql). I also did some system file integrity checks and checked for any outstanding driver and BIOS updates. I also DDU removed and reinstalled AMD chipset drivers for good measure as I noticed one component was outdated. No black screens since then, so far...
Meow81@reddit
I have the same problem with it constantly crashing Windows Explorer
christopherw@reddit
Since I did the DDU and install of 566 my card's been OK, no more drops to black. If Explorer is crashing but your display output is staying on, it might be something else. Does anything show in Event Viewer for driver crashes?
scatrinomee@reddit
Just checking, how have things been on 566
christopherw@reddit
Hi, I stayed on 566.36 for quite a long time, until we were well into 572 - I upgraded again around 572.60, then I've gone through 572.70, 577.00 and I'm currently running 580.97. All have been absolutely fine.
Meow81@reddit
event viewer shows 3 rows of errors:
\Device\000000cc Resetting TDR occurred on GPUID:2060
\Device\000000cc Reset TDR occurred on GPUID:2060
\Device\000000cc Restarting TDR occurred on GPUID:2060
then a warning
Display driver nvlddmkm stopped responding and has successfully recovered.
ZoteTheMitey@reddit
My 4090 keeps crashing nvidia drivers if I am watching a video on one display, then try dragging another window to a different monitor while the video is playing.
screens go black for a second as the nvlddmkm driver crashes and recovers.
Also once or twice in the last couple weeks, nvidia drivers BSOD the whole computer waking up from sleep.
This all just recently started happening. Card has been fine for over 2 years.
Did you find a fix?
I tried rolling back to 560.94 (august I think)...but the issue remains.
christopherw@reddit
Not found a fix yet. DDUed my drivers, installed older versions, problem persisted, put newer ones on, problem persisted. I'm also running the latest publicly available VBIOS according to the GPU-Z database (I actually submitted mine to the DB as it was a couple of versions newer than the ones on there already) but it's still the unchanged, presumably original shipping version, from \~2023.
I noticed that if I resume the machine from sleep or reboot, and open Discord, it occasionally does it just once
I've started to just accept that the graphics subsystem crashes but restarts and reinitialises the card, after 30-60 seconds. Annoying, but so far it's always come back.
Some other discussions around it:
I've also seen other comments about motherboard XMP stability at certain configurations potentially affecting the GPU, it may even be a motherboard BIOS issue because I updated the motherboard BIOS - from memory, around the same time the crashes started.
Bit puzzled about the whole thing tbh!
ZoteTheMitey@reddit
same issues here...same troubleshooting steps to no avail
Not a motherboard BIOS issue. I think it's a win 11 24h2 issue.
I would not mess with your bios lol. I thought I would downgrade to gigabyte F3 bios from latest F4d bios last night just as part of troubleshooting. I've been on gigabytes F3 bios before with no issues. It's even the latest stable version. F4d is a beta version. But anyway, I flashed F3 last night.
This morning I went to disable memory context restore setting before I went to work, just to test if that made any difference with the nvlddmkm crashes. and all of a sudden my bios freezes when I try to save settings. I have to keep restarting with the reset switch on my case. The settings do save, and it does boot into windows. But I noticed the dreaded "00" motherboard debug code and red cpu LED when I was in bios.
it absolutely kills me that I'm stuck at work and didn't have time to try to flash back to F4d to fix it. Hopefully it's just a fluke and it lets me flash back and there is no issue with the CPU now....
so yeah....don't flash your bios unless you absolutely have to lol.
I can't try to fix it for another 5 hours =`[
lactose_intoleroni@reddit
Nope. I'm on 23H2 and been having the same problem.
ZoteTheMitey@reddit
I found a workaround by disabling hardware acceleration in any browsers installed. For me it was Firefox and Edge (which the native Netflix app in windows uses)
It's been weeks, no crashes since disabling hardware accel in my browsers. Not sure what the cause is. Some issue with widevine or video codecs maybe? not sure. But yeah it should fix it for you too. Go into your browser settings and uncheck hardware accel.
lactose_intoleroni@reddit
I tried this and it unfortunately didn't work for me. My issue is that when I lock my screens at night and let my PC idle, the monitors just refuse to wake up at all whenever I come back to it the next day.
christopherw@reddit
I'm on W11 23H2 (cannot upgrade to 24H2 until I can figure out a decent replacement for my WMR VR headset), so one more data point to contemplate. I have the Asus ProArt X670-E, recently flashed to latest stable BIOS and some stable and gentle XMP and AMD PBO settings, which I've run for many months with zero issues.
kevinq@reddit
FWIW DDUing fixed this for me
scatrinomee@reddit
I have the issue after sleep as well
westie1010@reddit
Exactly the same, freshly reinstalled windows last night, new drivers on everything possible, haven't even launched a game on it and it crashed last night, Had to PSU cycle.....
No_Height4816@reddit
I downloaded the ddu driver uninstalled and uninstalled mine then restarted and repeated. Then reinstalled and it worked
lactose_intoleroni@reddit
What? lol
Perception5@reddit
they used DDU uninstaller* to uninstall their drivers. they restarted their computer and reinstalled their drivers.
stavik96@reddit
Have the 3090 and updated to 572.47 on the 20th of February, had hardware error 117 on the 22nd, 26th, 27th, 28th(117 is apparently related to GPU) and today while it didn't list it as a hardware error on reliability monitor event viewer does list the issue OP has with nvlddmkm. Did actually use DDU to uninstall driver yesterday and update to 572.60 as the nvidia app failed installing the latest driver, but seems issue is still there.
grainofsands@reddit
Did you manage to resolve the issue?
selectexception@reddit
If you ever find a solution to this please let me know, I have had this for over a month now and reinstalling Windows did not help.
Meow81@reddit
Did you have even this text in event viewer ??
The description for Event ID 153 from source nvlddmkm cannot be found. Either the component that raises this event is not installed on your local computer or the installation is corrupted. You can install or repair the component on the local computer.
HBizzle26@reddit
I am getting this with it.
Meow81@reddit
Did you find a fix ?
distracted_seagull@reddit
sorry if i did i've no idea what i did but i'm not having the problem any more.
Skindrix@reddit
i have a 4060 ti , I have nothing older to compare my issue to , i got this pc around the same time windows rolled out that problematic update but it seems to only do it while im browsing , it has not happened while playing a game
WhoDrankTheRedRum@reddit
3060 Ti started randomly doing this a few months before I replaced it with a 5070, Did fresh install of Win 11 after getting the 5070.
Same thing randomly occurs with 5070, I now believe its 25H2 related based on timing of when it started on 3060 Ti but still not 100%, ran memtest86 and no errors, PSU is only 2 years old (Corsair RM750x Gold with 10 year warranty).
Replying to old post to share my experience with it so far, luckily doesn't happen very often and usually fully recovers but it can cause several programs to crash when it happens.
AstroflashReddit@reddit
PSA:
I was encountering this TDR error on my system recently and it would only occur on closing an application / game. I had zero stability issues launching or during gameplay.
The culprit ended up being the fact that I had enabled "Memory Context Restore" in my BIOS/UEFI. My mobo performs memory training on every boot by default, leading to slow post times. I enabled the above option to reduce POST times, but this obviously led to memory instability issues in my case.
(Specs: 7600x // MSI B650M-P// 6000Mhz 32GB CL30 // 3080 FE // 650W EVGA Gold)
If this isn't your issue, I can summarize the other reddit tabs you're about to open for you.
It's important to note that while the nvlddmkm TDR error would suggest a Nvidia driver/hardware issue (and it could be), the real cause could always be something else causing the nvlddmkm driver to crash (RAM instability in my case)
Unfortunately, this error doesn't seem to have one common root cause - AND, it could be either hardware or software...
Here are some of the various "fixes" I've seen that may be worth trying:
Some people have managed to fix this issue by replacing their 12VHPWR Nvidia adapter, while others fixed it with a new GPU altogether. YMMV.
absolutelynotaname@reddit
I have the same specs and this happened to me during gameplay, disabling Memory Context Restore has fixed it for me so thank you very much for posting.
AstroflashReddit@reddit
Glad I helped someone!
absolutelynotaname@reddit
Wait why did you delete the comment then? It would've been helpful if someone else found it like me
AstroflashReddit@reddit
I didn't?
alimbu57@reddit
what's going on here?
AstroflashReddit@reddit
Describe it as you see it. I'm confused
RitualST@reddit
I know that this is an old topic, but I was plagued by this problem on my 4090 GPU where a cold boot would lead to black screen and I had to force a restart of my PC and only then it would be ok.
Yesterday I grew sick of this issue and decided to spend some time to try to resolve it. It seems I was successful. I have performed several actions 1 of which is optional imho and I will get back to the old settings.
MY PC Is Asus TUF Gaming X670-E with Ryzen 7 7800x3d and MSI Suprim 4090
0) I have updated drivers for AM5 chipset for my Mobo, itself it didn't do anything but I just wanted to mention this.
1) Disable Memory training - this is unnecessary step but I just wanted to make sure that my rig is as stable as possible:
In Bios: Extreme Tweaker > DRAM Timing Control > Memory Context Restore -> Enabled
Honestly I doubt it would do much by itself, but as mentioned I just wanted to make sure my system is as stable as possible.
2) Power Options:
Disable PCI Express Link State Power Management
3) (and this is what I think did the trick) 0 Fan mode
In my GPU Settings it's possible to enable 0 Fan mode. I think that for some reason GPU doesn't start correctly with this option turned on. I have disable it and the issue never came back.
Keeping my fingers crossed for others I'll update this post if it comes back at any time.
alimbu57@reddit
others report of similar fixes
plus
a) possible to replug all power cords into other outlets,
b) reseat RAM and GPU fully,
c) replug or CHANGE Powerconnector to GPU
and most prevalent either temporary mitigation or permanent fix
d) change global nvidia performance Mode (or only for crashing game or browser) to prefer maximum PERFORMANCE
as the most likely cure, however since I only have 5 TDR per month so far in first month which only take 30-60 seconds with no crashes or permanent freezes I don't want to increase power consumption globally yet
SYD_APEX@reddit
News?
Acceptable-Cry-2605@reddit
Problem sollved after putting in the new card.
SYD_APEX@reddit
Guys, I think i found the problem.
I had the same issues like you all have.
Picture of the Error: https://postimg.cc/gallery/wJc3JvY
At the end I controlled the power cable of the 4090, pluged it out and in again. But, I don´t think that was the problem!
After the hardware check I directly installed FurMark. I thought I could provoke the error by running the GPU burner Test.
At the first start, FurMark told me that is has to update the DirectX11. It automatically did that and what should I say?
The problem is solved!
No issues during gaming any more!
Nikita1409@reddit
Hey, so all you did was update DirectX11?
SYD_APEX@reddit
Yes, but since a cuple of days the problem is there again. I will check what I installed the last days and see if there is any corelation between.
Some one else an idea?
alimbu57@reddit
ur card set to maximum performance nvidia power setting?
Mine only TDR 5 times in first month of new PC Win 11, which takes 30-60 seconds with 0 Crashes or permanent freeze, so I don't want higher electricity cost as a fix if it stays at 5 TDR per month
Feistiestdisc0@reddit
Also seeing this with a 4090, but a zotac OC. Problems first started in January.
alimbu57@reddit
Fixed? Problem on a Zotac 4070 Ti A.E.A.
only 5 TDR 0 Crashes in first month of new PC / Win 11,
same GPU 2 years old PC win 10 - 0 TDR
Empty_Debate@reddit
My understanding is that this is a known issue by NVIDIA:
https://www.nvidia.com/en-us/geforce/forums/geforce-graphics-cards/5/502996/rtx-4090-driver-crashing-constantly-without-any-lo/
https://www.reddit.com/r/nvidia/comments/11k1dyq/nvidia_confirms_new_driver_is_causing_cpu_spikes/
jimlaman8c@reddit (OP)
the issue seems to have been gone now for a while, maybe the last 1 or two driver updates cured it.
InternationalBand666@reddit
im on the latest with new pc .. still happened one time
alimbu57@reddit
still happening 2026
lifelite@reddit
Coming back to old threads to post my resolution to this issue.
My issue was caused by the fact I have a Samsung Odyssey G93sc monitor (very high resolution, refresh rate, HDR, etc), which due to Display Port 1.4 not having the bandwidth, required heavier use of Display Stream Compression (DSC). The time it took for that to kick in was around the same time it took for my monitor to go into standby, thus causing a sort of a recursive loop which would trigger a crash on the NVidia drivers, which I'm sure was exasperated by the fact I have 3 high resolution monitors.
To resolve this, I replaced the DP 1.4 with HDMI 2.1 (using a confirmed high speed HDMI cable). I've yet to encounter this issue since. Yep, that's it.
In retrospect, I should have known to use the best input connection that was out, but due to my legacy knowledge, I just assumed DP was better than HDMI (because historically....it always had been).
Hopefully this helps someone out!
Pox4eveR@reddit
Having this issue and it was HELL to debug it. I then went to even viewer to check the first occurrence and it was the day i bought my Alienware AW3225QF (4K@240hz+HDR+Dolby Vision) monitor...
It seems this is the most correct answer. I have a 4k@240hz+HDR+Dolby Vision monitor + 3 2K@144 monitors and that was exploding the GPUs Display Engine. Bought a cheap 1060 for 30 euros and tossed all 3 auxiliar monitors there, no issues yet.
Probably just shitty vendors GPU? I have a MSI Ventus 3X 5070 Ti OC.
I started the RMA process on MSI but damn, they are picky, I have to do tons of shit before sending it.
Miserable_Anteater62@reddit
I've used a 2.1 HDMI since I got the card over a year ago but I'm now running into this issue. 4090 Strix with an LG C2 OLED.
uacnix@reddit
GOD DAMN IT, A MONTH LATER AND AFTER RMA OF MY LAPTOP, IT FINALLY STOPS CRASHING!
MODS PIN THIS! ULTRA-WIDE DISPLAYS HAVE COMPLETELY DIFFERENT AND BIGGER REQUIREMENTS and apparently GPUs can't handle it very well.
I have a 165Hz, 3440x1440p HDR main monitor, along with 2 FHD on both sides and before that, it was connected like this (to Legion Pro 7i with 4090) - UW1440p via USB-C to DP 1.4 Adapter, marketed as "TB4 240Hz 8K" and each of the FHD via DP to USB-C to 2xDP,1xHDMI, 3xUSB HUB
And it was crashing like hell, but ONLY on desktop. In games it was rock solid.
So then I've tested it with only that main monitor -of course via the same cable, cause who would've noticed, especially when every bugcheck ended with nvlddmkm.sys or dxgkrn.sys. Then I took that laptop to my office, where I have another monitor, except its got very same specs (also UW1440p, HDR, 165Hz, except it was Philips and the one in home is iiyama). Laptop crashed like before.
I wrote a post on Lenovo Legion regarding this, and of course some "UserGPT" told me that my GPU is broken and I should RMA.
Except I just got it from RMA, where they replaced entire mobo and heatsinks, so I played with it for a week and opened another RMA, waited 3 weeks and got it returned with apparently "driver issues".
Now go ahead and ask yourself people, does the average service desk tech guy have access to high-end monitors on hand? Of course not!
Right now I got myself a HDMI 2.1 cable, but apparently iiyama GB3467WQSU has only HDMI 2.0 and that's barely a 100Hz at UW1440p, but I can live with it not crashing while in desktop.
sk3tchcom@reddit
Having this issue with a PNY RTX 5080 OC. Same rig worked fine with other cards in the past several months (9070 XT Reaper, 9070 XT Red Devil).
Ssyl@reddit
Same card, same issue. I have it with a 9800X3D, Gigabyte B650M, latest BIOS, chipset drivers, and gpu drivers as of 2025-11-16.
sk3tchcom@reddit
I did an RMA and then sold the card I received and moved on. All good since.
Ssyl@reddit
What card did you get? I sent it in for RMA and they sent it back saying there was nothing wrong with it, unfortunately.
sk3tchcom@reddit
Oh dang sorry to hear that.
I go through a lot of cards but I’m settled for now on a 5090 TUF that I got for $1999 in that rig.
Ssyl@reddit
$2k? Damn, I'd get one if I could get it at MSRP.
HBizzle26@reddit
I am having the issue with a Gigabyte OC 5080. Bought an open box from Newegg and can return it. Thinking this is memory related and going to send it back.
sk3tchcom@reddit
Check your BIOS settings for any PCIe sleep-related settings (power saving) - disable them and see if it occurs.
HBizzle26@reddit
Dram Module errors started popping where the board wouldn't post. Appears to be a bad mobo. Replacing it as it is still within my 30 day window.
sk3tchcom@reddit
Damn dude sorry. Thats an annoying part to have to replace.
HBizzle26@reddit
Second board with this build. Originally had an Asrock Taichi x870e, but they are killing 9800x3D chips left and right. Pulled everything off it and returned it and grabbed an open box Aorus Master x870e off Newegg. Now I get why it was an open box. Went to micro enter today and got a brand new MSI x870e tomahawk. Realized I don’t need the feature set of either of those boards, just solid VRM and ram support.
sk3tchcom@reddit
Nice. Exactly. The fancy boards are cool and all but rarely does it matter to the average gamer. I’ve got two ASRock boards (Taichi with 9950X3D and Steel Legend with 9800X3D - both X670E) no issues. Seems it may be looking like more of an AMD issue versus just ASRock. I’d swap out too at that stage, though.
HBizzle26@reddit
This ended up being the mobo. Swapped the PSU, Ram, and it still happened. Swapped mobo with same components and not a single problem of far.
sk3tchcom@reddit
Nice dude! Glad it’s sorted. This stuff is so annoying and stressful.
jayhxmo@reddit
This issue took out 5 years out of my lifespan.
I had a 3090 that would crash every now and then like your 4090, randomly. Ran benchmarks and it was fine. The randomest actions would crash, nothing that was reliably reproducible.
It turned out there was something wrong with the power port on the wall that I had my machine plugged into. Something about it being slightly unstable, other appliances were fine, but it was the culprit of the occasional crash. The A/C repair man discovered that for me.
Solution was to plug it into a different wall. Works totally fine now. I hope this might help another poor soul...
Darrylboio@reddit
Maybe this is also the reason why my pc is being weird.. i think my pc is plugged into a splitter of some sort, one that gives more plugs. Should try this.
Gitmonay07@reddit
That makes sense. I came upon this error bc static electricity(i think) when moving my chair over and the wheels bump the desk, sometimes my right monitor will go black. Its the only one plugged into the power strip mounted to the underside of my desk. both monitors are wall mounted and the pc tower is on the floor. Anyways, lately when it happens both end up going black and staying black/unresponsive until i hard reset. I decided to look in event viewer to see if it showed any info and seen this error. So maybe ill do some checking with the outlet/surge protector.
Darrylboio@reddit
i think i found my people, this is happening to me now at rare times but had it happen just now after getting on steamvr..
aCh0zenBeing@reddit
I am now getting this issue with a 4080 .. randomly go into standby mode.. PC freezes and comes back .. Event viewer shows "Reset TDR occurred on GPUID:100" .. I noticed in Afterburner clock speeds drop to 210/405mhz core/mem and I have to restart PC to get my normal clockspeeds back. I also noticed issues with stuttering in games if I don't restart as it has issues with clockspeeds bouncing from sub 500mhz to 2.5ghz (core) but like I said, if I restart system all is normal.. until it happens again.. Only recent changes in the last 30 days was new gfx drivers and windows 11 updates. Never had this issue before.
Ecstatic-Fish-3296@reddit
Hello ! My last week was plagued by this exact same error after I bought an use RTX 3070. I have spent most of my free time looking up fixes and from what I have learned, this issue has multiple causes and many more solutions. Everyone was telling on forums what worked for them and so will I. I have upgraded from windows 10 to windows 11, clean install without keeping anything. After installing and setting everything up and with the automatic nvidia driver install from windows I have installed DDU, turned off device installation automatic drivers ( advanced system settings > hardware > device installation settings > no ), turned off fast boot ( power settings > choose what the buttons do > fast boot disabled ( change settings that are currently unavailable if it is greyed out ), turned off group policy “ Do not include drivers with Windows Update ( only for windows 11 pro, look up on the web how to turn it off, I forgot ). I have then booted into safe mode without networking, unplugged my ethernet cable, uninstalled the driver with DDU, booted into windows normally and installed only the driver, most recent one 576.88 without the nvidia app AND only the graphics driver from the custom install option, no phyisX no audio no nothing else + check the clean install box.
aCh0zenBeing@reddit
Love how nvidia rushes out the 5000 series mean while the 3000 and 4000 are still plagued with issues on newer titles. Intel and Nvidia need to get their shit together. Everyone is just chasing benchmarks but stability seemed to take a back seat. I can wait another 6-12 months to upgrade, just give me something stable out of the box. People shouldn't need to undervolt CPUs and GPUs out the box in order to be stable. At these price points, it should come stable and tuned with optimal settings out the box.
SushiOnix@reddit
If this helps anyone, I recently had this problem.
I was able to reproduce it by quickly unplugging and replugging the DP cable (1 second). No problem with HDMI.
After 5/10 attempts, the screen would go black and the PC would restart after a few seconds.
I tried everything, but the problem was actually with the BIOS… Since I updated the BIOS, I can no longer reproduce the problem.
If you updated your BIOS, try rolling back or updating it to the latest version.
You can also try changing some settings related to the PCI-E port.
My configuration: Win 11 24H2 + ROG Strix 4090 OC + ROG STRIX X670E-E
Dagos@reddit
I have this issue when I take a screenshot. Absolutely kills my games/streams. I have a RTX 3060 and a i9.
austinguru713@reddit
I did some more research as I was dealing with BSOD and PC resets as well, and apparently its tied to power management and TDR.
Someone on a forum suggested changing power settings in Nvidia Control Panel
Manage 3d settings > Power Management Mode > change to Prefer Maximum Performance
and they also suggested downloading a TDR manipulator, set to 10
So far I seem to of mitigated the issue with no restarts after doing those 2 actions.
Now If I can just figure out what is causing the overall issue.
Meow81@reddit
which value in" tdr manipulator" should be set to 10 ??
Meow81@reddit
Tdr delay to 10. in Tdr Manipulator program
Rapzid@reddit
Been happening with my 3070 since I bought it off ebay a couple years ago. Sometimes it recovers others it reboots. Never in a game though oddly enough; seems to occur most often while using Chrome.
Meow81@reddit
to me too. rtx 2060 6gb
Micio922@reddit
Same..... it only happens when using a web browser (tested with both edge and chrome) to stream crunchroll, netflix, or youtube. I can play MH Wilds, Helldivers 2, Civ 7, or any game for hours on end and no problem..... Youtube takes anywhere from 2 to 10 minutes and it crashes. Running Win 11 with a 3060 12gb OC
Sauceman9000@reddit
Getting the same issue on a 3090 FE and a 1070. F***. Happens randomly does anyone in here leave their computer 24/7 on?
Aggressive-Bug-1769@reddit
I do, but I only get this on boot. GTX 1080.
lebanonjon27@reddit
I am able to reproduce this crash now reliably. I suspect it has something to do with NVIDIA non-standard power saving of clocking down to PCIe 1.0 at 2.5GT/s when the link is idle. PCIe is supposed to go into an L1 state for saving power, not what NVIDIA is doing. I have all PCIe power settings disabled in the BIOS
System Configuration:
Microsoft Windows 11 Professional (x64) Build 22621.1702 (22H2)
CPU: 13th Gen Intel(R) Core(TM) i9-13900K
Motherboard Model: ASUS ROG STRIX Z790-E GAMING WIFI
BIOS Manufacturer: American Megatrends Inc.
BIOS Date: 03/29/2023
BIOS Version: 0904
Motherboard Chipset: Intel Z790 (Raptor Lake-S PCH)
GPU: Zotac RTX 4090, (ZOTAC Gaming GeForce RTX 4090 Trinity OC 24GB GDDR6X 384-bit 21 Gbps PCIE 4.0 Gaming Graphics Card)
Original Device Name: NVIDIA GeForce RTX 4090 (AD102-300/301)
Issue:
The description for Event ID 0 from source nvlddmkm cannot be found. Either the component that raises this event is not installed on your local computer or the installation is corrupted. You can install or repair the component on the local computer.
If the event originated on another computer, the display information had to be saved with the event.
The following information was included with the event:
\Device\Video3
Restarting TDR occurred on GPUID:100
The message resource is present but the message was not found in the message table
Symptom: The window goes blank for a few seconds, then recovers. Sometimes this causes a system freeze
Issue: System freezes, need to hard power cycle
Steps to reproduce reliability within 2 minutes
Set 'Power management mode' to 'prefer maximum performance'
Set BIOS slot PCIe link speed forced to PCIe 4.0 x16 at 16GT/s
Boot into windows, freezes within 2 minutes
Happens much more frequently on the latest BIOS, it was only once ever few weeks before, now it is 2-3 times per day.
This definitely has something to do with PCIe, because when I put the card in slot 2 on the PCH, it negotiates at PCIe 4.0 x4, but does NOT crash. Huge bummer for Intel and NVIDIA validation teams.
cevikmert42@reddit
Hey, did you solve this problem? I have same setup with you. I also confirm that reproduciability of the error as you described.
lebanonjon27@reddit
There are 2-3 issues, which makes this very tricky to debug.
BIOS - 0904: 4090 didn't work at PCIe 3.0 or 4.0. 3080Ti did not work at PCIe 4.0, but semi stable at 3.0 (one of these events per day). The difference is, the 3080Ti recovers and doesn't crash - I just had the app close and the screen go blank for a few seconds. Confirmed the same behavior on my 3070. So you still see the same error in windows event viewer.
So...with the ROG STRIX Z790-E GAMING WIFI BIOS 1202: PCIe 4.0 on the 3080Ti and NVIDIA driver 535, about one event per day
installed the 4090 on the AMD rog-strix-x570-e-gaming. Did not work at PCIe 3.0 OR 4.0.
ended up RMAing the card to Zotac. They sent new card, and put it in the Z790 at PCIe 4.0 x16. playing now for a day with no crashes. So this was a combo of PCIe link issue, driver issue, and inability for NVIDIA 4 series to recover from these link level resets. In my case, also appears a bad card. (which is odd because of the failures of the other two 3 series cards above).
you may want to check if the card is perfectly in the PCIe slot. The first 4090 I had was just off with the bracket, this is not the smoking gun, but may be important.
NVIDIA really has to get their shit together on this...
Ronicraft@reddit
Ive been having the same issue recently, my desktop icons disappear along with my task bar when this happens and I have to restart to get them back. Would this narrow the specific cause down to a solution? Im just tired of games randomly crashing on a card I spent 1500$ on
selectexception@reddit
That is explorer.exe crashing.
spiralsnowagain@reddit
Do you happen to know how I can limit crashes until I am able to get a new card? I'm having the exact same problem constantly but I can't return it since i bought the card off Facebook Marketplace some time ago.
lebanonjon27@reddit
If I recall setting it to PCIe 3.0 in bios reduced the number of link retrains that triggered the crash
spiralsnowagain@reddit
Thanks so much
SmoothLubedLinear@reddit
Gents, thank god I have found this thread, I purchased my high end system last November and have been having issues for the past 7 months, I have a ASUS GeForce RTX 4090 ROG STRIX OC 24GB Video Card w/ PS - ASUS ROG Strix 1200W Gold Fully Modular Power Supply - Aura Edition, G.Skill Trident Z5 RGB 64GB (2x 32GB) DDR5 6400MHz CL32, Intel Core i9 14900K 24 Core & ASUS ROG STRIX Z790-F GAMING WIFI II, I started to observe issues where by my displays would lose input and the only way to return any function to the system was to hard restart the system, I was racking my brain for months and believe me I tried everything, every analysis command you could run on cmd, memory tests, swapping memory modules/ changing the slots, re-cabling the whole system and reseated all components, tried clean booting, fresh installed windows, upgraded and downgraded Nvidia drivers using the DDU tool, over/ underclocked the system, undervolted the CPU and GPU, tried different cables both HDMI + DP, changed power settings to performance on both NCP and on Windows, changed the TdrDelay in the registry, changed PL1/PL2 wattage limits to the Intel recommended levels being PL1 125W and PL2 253W, took it to a professional PC retailer for debugging, they replaced the already brand new motherboard and I was still having issues, I then got the CPU replaced with another brand new one, the issues worsened, I was sick of this so I sent the whole system back to the original merchant I purchased the components from and they then replaced the GPU after deeming it to be faulty, I got it back last week and it has been smooth sailing until today.. Where the system black screened with error nvlddmkm ID 153. At this point I'm lost on what to do because I just want to love my system however it is simply unstable, I want to be on the bleeding edge but if this is to cost to do so it's not worth the mental degradation.
theseawoof@reddit
Did you end up having any more issues after receiving a new card? I have the same problem and tried every possible fix in the book. About to return the card unless it is an issue/setting I am missing.
lebanonjon27@reddit
Rma has been completely stable, I chased a ghost for months when it was bad card. Sigh…
theseawoof@reddit
Basically where I'm at.. salty and afraid to experience it again lol
Additional-Pipe-7357@reddit
im having the same issue except it started with an RX5500XT and is still happening with my new 1650 super.
some days it never happens some days it happens when i launch a game or am watching youtube. ive tried all fixes apart from just switching out my mobo psu and ram but i have no way to be sure which one wud be the culprit. i switched out my 5500xt and its worked perfectly for the new owner and this 1650 super started having the same issue after about a week of use. i ran all psu ram gpu and cpu benchmarks not a single problem but the moment i play a game thats not even demanding it just goes to black and gpu fans ramp up.
local--yokel@reddit
I've had this issue on and off. I think it's power related in all cases. Possibly a bad driver (I'm using 561 as of 2/21/25 which seems to be one of the most stable) controlling power management in a way that isn't interacting with a system well.
Disabling PCI Express Native Power Management and PCI Express Clock Gating seems to help or eliminate it. I also think it could be the power supply itself, or the wall outlet, possibly both. I have questionable wiring in my house, and it seems to have started once I moved here. I don't like the NV driver change to prefer maximum performance as that skyrockets power usage, but it may help some people or be required.
CardiologistLeast640@reddit
never had the issues in 2 years, after latest nvidia driver, when playing Hogwarts Legacy, i'm getting random pc crash with ->
WingedBunny1@reddit
Well either your gpu has an issue or the driver has an issue. Id try an older driver and see if it happens again, if so there might be a defect in the gpu. Another unlikely cause could be that a failsafe failed and a 1 changed to a 0 or vice versa. This can happen in any electronic at any time but usually doesnt occure because of failsafes, how often has it happened? If only once and nothing else is happening now, maybe the best thing to do is to leave it as is and you dont actually have a problem.
jimlaman8c@reddit (OP)
Unfortunately this is about the 4th time and I've had the last 3 drivers on too maybe its the gpu
WingedBunny1@reddit
Im sorry to hear it. You still within warranty I assume? Would be the only thing I seem sensible.
jimlaman8c@reddit (OP)
I just found this info, apparently a known thing
https://www.overclock.net/threads/massive-rtx-4090-problems-driver-or-hardware.1801381/
https://www.nvidia.com/en-us/geforce/forums/geforce-graphics-cards/5/502996/rtx-4090-driver-crashing-constantly-without-any-lo/
Fix : In nvidia control panel set the card to 'Prefer maximum performance' as a work around until a real fix is found.
Ronicraft@reddit
Does this still work?
-skyhook-@reddit
For what it is worth, I seem to be experiencing this when I use the Windows 10 "night light" feature with multiple monitors.
Big-Echo-7705@reddit
Same here for a "couple" of times now
gtx 1660 oc
screen goes black. cpu cooler turns up to 50..60% (by listening), pc non-reactive until restart
checked my eventlog..normally would only log sth from kernel (unexpected shutdown), recovery from bugcheck, but no indication of what might be the issue
changed ram, changed ssd, changed OS (now running win 11)
today, i noticed a new error (nvlddmkm, seems to come from nvidia drivers)
>>The description for Event ID 0 from source nvlddmkm cannot be found. Either the component that raises this event is not installed on your local computer or the installation is corrupted. You can install or repair the component on the local computer.If the event originated on another computer, the display information had to be saved with the event.The following information was included with the event: \Device\Video3BusReset TDR occurred on GPUID:900The message resource is present but the message was not found in the message table
Flowyflo99@reddit
Did you find a fix? My next step would also be to test other hardware. It happened for me in Win 11 but on a different Win 10 install it also doesn't work.
Familiar-Birthday191@reddit
Changing the Nvidia CP 3d to prefer Performance worked for me on RTX 4090
No-Safety2565@reddit
what this caused is that on my 4090 idle power draw went from 15 to 56W
mcbainas@reddit
I Just bought a PNY 4090 Verto triple fan, after using it for about 3-4 hours this problem started. Tried everything. I think Ill exchange the card Monday.
theseawoof@reddit
How'd this conclude? My PNY 4080 super is having this issue, tried everything. Ready to return
MrRightclick@reddit
I can see this thread is still somewhat active, so posting my 2 cents.
This is an issue with all RTX GPUs as far as I've seen and scoured the web for answers. It's been happening for a long time (years), and so far there has been no official fix. People who have contacted Nvidia for help have been getting responses where they state it's not their issue, it's an issue with either an ATI component you might have, or then it's something else. ATI says it's not their issue, and states it's Nvidia's fault. People with only Nvidia and Intel components are also getting this issue.
The "fixes", which I've found and done over the two years, which work like a temporary cure:
I've had this issue for over 2 years now with my 2060s. It started happening randomly, went from bad to worse, then a bit better with the 'fixes', then to horrible again.
The description for Event ID 0 from source nvlddmkm cannot be found.
It haunts my dreams, and has happened 3-4 times today. I have just reformatted my PC (Win10), and reinstalled Win11 instead. It's a blank PC. It's still happening.
I'm ready to say don't buy Nvidia. You might get this issue and spend hours upon hours trying to fix it, or you might not. RMA your card, and buy something else if possible.
theseawoof@reddit
I'm jumping ship. Bought a new 4080 super and wasted two weeks trying to troubleshoot with every fix known to the internet. Spent $1100 and this is unacceptable. People want to talk all day about amd driver issues but want to whistle past the graveyard about these Nvidia issues 😭 I use Davinci Resolve and Adobe Premiere for video editing as we as game, I read the software is optimized for Nvidia but id be willing to run a second system with my 3070 if needed in case the 7900xtx can't hang. I was super excited for the 4080 super but this bummed me out, I don't want to return and buy a new one only to have the same issue