Recoil II 17.3 VIDEO_DXGKRNL_FATAL_ERROR and reinstalling nvidia drivers

[Namu]

Since I was getting systematic nvidia driver crash and bluescreens with code VIDEO_DXGKRNL_FATAL_ERROR every time I would open the map in Skyrim SE, I figured I'd replace the default drivers my laptop shipped with, with the latest 419.35 drivers from nVidia. Except now the RTX 2070 in my Recoil II 17.3 is missing from the Device Manager, will not appear nor give a prompt for installing any driver, and none of the 419.35 installers, DHC or standard, that I tried will pass the compatibility check. They all fail with:

This NVIDIA graphics driver is not compatible with this version of Windows.

This graphics driver could not find compatible graphics hardware.

What do I do now ? Did I just buy a 2000 Euro laptop with barely as much 3D processing power as my 2011 macbook ? Is there any version of Nvidia drivers that will work with the RTX 2070 shipped with this laptop ? What is the simplest option to have it recognized by the system again ?

Thank ye all

 

[Namu]

So I performed a cold restart and, lo and behold, this time the RTX magically showed up again in under Display Adaptors in the Device Manager. From there the DCH drivers finally agreed to install, I did a clean installation for added safety. Let's hope the card is not dying on me like I've seen reported around, it worked perfectly fine for the first couple weeks and I'd hate to have to ship this back for a replacement.

 

[Namu]

The screen-freeze to black and crash of nvidia driver (now 419.35) still happens as soon as I open the map in Skyrim or just walk outdoors randomly in specific areas (Solitude, and west of Falkreath), following which I can CTRL-ALT-DEL and sign out, but upon signing back in the RTX 2070 is gone from the Display Adaptors and I have only the integrated UHD630 left. Also, if I restart to bring it back on the laptop invariably bluescreens with the above-mentionned code VIDEO_DXGKRNL_FATAL_ERROR.

At least once it did not come back up in the Display Adaptors list after one reboot, in addition to the occurrence in my first post.

Oddly enough, testing with my other Steam games so far did not trigger this at all, I'll test this further with as many games I have on here. I ran Skyrim off the UHD630 and could not trigger anything either, but that's little evidence of anything as I had to turn off a great many graphic features (for reference, Skyrim gives lower framerate on the UHD630 than on the outdated 6750M through Cider-X on the 8-y-o Mac, lol). I'll have to test for repro with the RTX and low graphic settings + reduced modlist.

I'd be eager to blame it on Skyrim alone if not for the complete loss of the RTX that sometimes persist through a reboot, and the fact the exact same game + modlist ran with no issue for two weeks prior.

 

[ubuysa]

How old is the laptop?

Was it working ok when you got it or has it always been BSODing like this? That's an unusual BSOD, I've never seen that one before.

Did you buy this with Windows installed or did you install your own?

 

[Namu]

It was preloaded with Win10 Home, came in early February. Only thing I did was install my games and update the nvidia drivers to 419.17 with GeForceExperience (which I then deinstalled for being an obnoxious UI). It worked perfectly for 2-3 weeks, I completed the entire Dawnguard arc and then some, until it suddenly started freezing and losing the GPU from Devices, as described. Reducing things here and there only let it go a little further but eventually it would crash systematically when opening the map in-game. No other game has any issue.

At the moment I'm running Skyrim SE windowed and with vsync disabled, with all the same modded stuff as before (MO2 and SKSE files refactored, too, for good measure), and this seems to have averted the crashes entirely, so I'm willing to chalk it up to Bugthesda shennannigannery. More reading on the topic has suggested the game is acting retarded with triple-buffering vsync and certain nVidia GPUs of the 10xx and 20xx series. Unless the crashes return later I consider this was it.

 

[ubuysa]

At the moment I'm running Skyrim SE windowed and with vsync disabled, with all the same modded stuff as before (MO2 and SKSE files refactored, too, for good measure), and this seems to have averted the crashes entirely, so I'm willing to chalk it up to Bugthesda shennannigannery. More reading on the topic has suggested the game is acting retarded with triple-buffering vsync and certain nVidia GPUs of the 10xx and 20xx series. Unless the crashes return later I consider this was it.

Fingers (and toes) crossed then.

 

[Namu]

Sadly I've just had the same issue again in windowed mode (though it is proving oddly more resilient). I got a pair of events, ID 4101, in the System logs, that report both that "nvlddmkm stopped responding and has successfully recovered" - except it hasn't and the GPU is then unusable until I reboot.

 

[ubuysa]

Sadly I've just had the same issue again in windowed mode (though it is proving oddly more resilient). I got a pair of events, ID 4101, in the System logs, that report both that "nvlddmkm stopped responding and has successfully recovered" - except it hasn't and the GPU is then unusable until I reboot.

It's getting close to the point where you'll need to call PCS, but before you do there is one more troubleshooting 'test' that you really should do first; a clean reinstall of Windows. I realise you probably don't want to do that but it's something that PCS are almost certain to ask you to do, and if they don't it's likely to be the first thing they will try when they get it back.

BSODs generally happen, not in the driver code itself, but in the Windows kernel. If the kernel receives information or requests from a (bad) driver that are inconsistent or undoable then because of the elevated level at which the kernel runs the normal recovery routines are not available and so it BSODs. Usually these problems can be fixed by reinstalling the correct (or a working) driver, but it's not completely impossible that the problem is in the Windows kernel and only a clean reinstall will fix that.

I would suggest that you do a completely clean install, deleting all the UEFI partitions so that the system drive (or that part of it that you use for the system) shows as unallocated space. Select that unallocated space and click Next, the installer will create and populate all the required UEFI partitions and install Windows into the correct partition. Once Windows is installed run Windows Update repeatedly until no more updates are found. This will (should) install all necessary drivers as well as Windows updates. Windows 10 makes a very good job of finding the right drivers.

Then, and before you install other software or make other configuration changes, install Skyrim and see how it works then. If you get freezes or BSODs in that state it's most likely a hardware issue and you will need to call PCS - be sure to tell them you've done a completely clean install of Windows too.

 

[Namu]

I managed to get the same issue in a different game (Crossout). Right after starting to render, the game crashes to desktop, and I see the same pair of events in the system logs as when Skyrim freezes to black. Seems it's really unstable drivers at the root... And I should be complaining to nVidia instead.

Just to clarify: the BSoD is during the restart process. I suspect the kernel is trying to get the GPU to shut down, except it's in a degraded or faulty state.

 

[ubuysa]

Just to clarify: the BSoD is during the restart process. I suspect the kernel is trying to get the GPU to shut down, except it's in a degraded or faulty state.

Ah, that does change things somewhat. That said, PCS will really expect you to have tried a clean install before RMAing it. It does sound as though you have a hardware issue but it is important to do all the troubleshooting you can before it has to go back.

 

[Namu]

Thanks for your continued involvement - reading around I'm apparently not alone at all getting these 4101 "nvlddmkm stopped responding" events irregularly. None of the people fixed it with reinstalling Windows (nor by replacing the GPU), though, and in my case it's nothing to do with overheating (GPU never goes past 62C, which is a testament to the nice cooling on those larger 17" laptops). I'll use DDU, reinstall the March drivers as cleanly as possible, and block Windows Update from touching up those drivers, as this seems to hold the most promise for now.

I'm unsure how to boot into safe mode, though, can you advise ? Thanks again

 

[SpyderTracks]

Thanks for your continued involvement - reading around I'm apparently not alone at all getting these 4101 "nvlddmkm stopped responding" events irregularly. None of the people fixed it with reinstalling Windows, though, and in my case it's nothing to do with overheating (GPU never goes past 62C, which is a testament to the nice cooling on those larger 17" laptops). I'll use DDU, reinstall the March drivers as cleanly as possible, and block Windows Update from touching up those drivers, as this seems to hold the most promise for now.

I'm unsure how to boot into safe mode, though, can you advise ? Thanks again

In windows 10 it needs to be done through the settings menu:

https://support.microsoft.com/en-gb/help/12376/windows-10-start-your-pc-in-safe-mode

 

[Namu]

Last update: so I cleaned up display drivers with DDU, blocked+masked the 25.21.14.1749 display driver update in Windows Update using this:
https://support.microsoft.com/en-us...t-a-driver-update-from-reinstalling-in-window

Then I clean-installed the 419.35 drivers.

But I still get the crashes, in any game, at a specific point for each game (upon logging into Crossout, opening the map outdoors in Skyrim SE, applying graphic options in The Witness, etc.). I think I have narrowed down the circumstances, my testing has been confirming it only happens after the laptop was once put to sleep/hibernate. If the laptop was cold-powered up, the drivers are stable, and I can have hours and hours of gameplay with no trouble at all. Once it has been put to sleep and woken back, though, the issue WILL trigger sooner than later, usually at the very first opportunity.

I guess I'll just report these findings to NVidia.

 

[Oussebon]

I'd suggest logging the issue with PCS as well, in the event you need to have the system looked at or replaced. Better to preserve that option and have the problem / fault (if it is a fault) registered under the warranty period.

I don't have any expertise here, but one thing I'd be inclined to try if you've not already is to make sure the Windows power settings set to the High Performance plan with PCIe Link State Power Management disabled.
Also have the Nvidia GPU set to prefer performance and set as the default GPU in Nvidia control panel.

I might also be inclined to run Prime95 to check for CPU stability. Though be aware temps might get quite high in a laptop chassis. https://overclocking.guide/stability-testing-with-prime-95/ I'd be looking to see if there were any errors on any of the 'workers' (threads).

 

[ubuysa]

What your seeing is a driver error. Most driver errors are not trapped until the Windows kernel gets control back and because the kernel runs at an elevated level the usual recovery routines are not available, that's why it BSODs - there is no other way out.

Driver errors like this have three possible causes; a fault in the driver, a fault in the Windows kernel, a fault in the device hardware.

Most often of course the problem is in the driver, but you seem to have done all you can to identify and correct that without much success. So ubuysa's rule #4 applies; if the problem is not where you're looking then it must be somewhere else.

I would suggest you do a fully clean reinstall of Windows and allow Windows Update to install all the necessary drivers. This will (should) eliminate the possibility of it being a fault in Windows, so if it works in that state then job done.

If it fails in that state it's more likely to be a hardware problem (it could of course still be an inherent Windows issue but you/we can't debug that). In either case (Windows error or hardware error) you're going to need to send it back to PCS I'm afraid.

 

[harryhatton105]

Hi Namu,

I'm glad I stumbled across this post because I have been experiencing the exact same problem. I bought the Recoil II with the RTX 2070 and in certain games I will freeze and upon restarting I get the same BSOD error message and like you described it also led to the disappearance of the RTX 2070 from device manager. I did a windows reinstall which brought the RTX 2070 back into the device manager but I seem to be getting the same stuttering and freezing and blue screen quite often. I emailed PCS to ask if this was a known issue and they thought it could be a driver/software issue but due to the reoccurence of this stuttering/freezing/BSOD I'm beginning to think something could be amiss with the GPU itself.

Did you get any further with your problem?

 

[G0975]

Hi all,

I too have had this system since new and have experienced the exact same thing. Funnily enough it only happens on one game in particular, that being WoW. In-fact it's just happened ten minutes ago. It's not happened on any other game in regular usage since I bought it early February. I've been running really demanding games and after under-clocking my CPU neither my CPU or GPU temps go above 70 when being played on my lap. Almost too good to be true for a laptop!

Back to the matter at hand, do not reinstall windows or use DDU to uninstall any drivers! It needs to be put out there to ALL Recoil II RTX series owners that we are currently only able to use modified drivers provided from PCs themselves. This is an e-mail I received this morning:

"The official NVIDIA drivers at this moment in time do not have your graphics card currently listed within the installation files, this is now bringing up errors with incompatibility for all our customers with the same chassis and it’s likely that NVIDIA’s new release will include this. "

To which I asked if we are ever going to be fully compatible with the response:

"Thank you for your email and to confirm we expect all our chassis to be fully supported by Nvidia. They’re usually on the ball with chassis manufactures but as TongFang are a little newer on to the scene I suspect this may be a reason why. CLEVO based computers usually have day 1 official nvidia support."

I expect the modified files to be slightly corrupt, hence the completely random crashes which disconnect the GPU and cause a blue screen upon reboot. I'm really hoping they rectify this soon as I'm keen to run the official drivers and put this issue right. The best thing we can all do is keep on at PCs about the issue and keep the pressure on to get it fixed.

 

[harryhatton105]

I did a full windows reinstallation last night (before I saw your post) as the card had disappeared from device manager but this did not bring it back into the device manager. But after leaving the laptop off last night and not plugged in, this morning the cards seems to have reappeared and I can reinstall all the drivers straight from Geforce experience which were causing the crashing and errors.

Its strange as we all seems to mention one game which is causing the crash as I've played plenty of different games on the laptop but this error only seems to occur with Sekiro.

I'm glad that this is mainly a software issue and as I honestly thought I had a faulty GPU. Hopefully either PCS or Nvidia sort something out with the drivers and this problem stops occuring.

 

[ubuysa]

I did a full windows reinstallation last night (before I saw your post) as the card had disappeared from device manager but this did not bring it back into the device manager. But after leaving the laptop off last night and not plugged in, this morning the cards seems to have reappeared and I can reinstall all the drivers straight from Geforce experience which were causing the crashing and errors.

Its strange as we all seems to mention one game which is causing the crash as I've played plenty of different games on the laptop but this error only seems to occur with Sekiro.

I'm glad that this is mainly a software issue and as I honestly thought I had a faulty GPU. Hopefully either PCS or Nvidia sort something out with the drivers and this problem stops occuring.

Since we have several people with the same problem it's vanishingly unlikely to be a hardware issue. I would suspect this will be either a driver issue or (perhaps more likely) a game issue - especially if it happens in a single game.

 

[harryhatton105]

Since we have several people with the same problem it's vanishingly unlikely to be a hardware issue. I would suspect this will be either a driver issue or (perhaps more likely) a game issue - especially if it happens in a single game.

What I meant is we all have one game which is causing the crash but that game is different for each of us, OP mentioned skyrim I think, G0975 mentioned WoW and I mentioned Sekiro. Shows its not an isolated issue to one game and this software issue is occuring across the board. Just got to hope for a speedy fix on one of the partys fronts (Nvidia/PCS).