Davinci Resolve causing BSOD

Hey,
I just finished my to this time most complex project. And I'm having some problems with davinci resolve crashing or in most cases my whole PC dying giving me the BSOD with the error code: system thread exception not handled. I tried cheking my system logs (windows) to see what happened and I have a few Level: Warning, Error and Critical problems. But none of them seem to show the direct cause of my problems. I'm not very good at PC software so I could be incorrect.

So the problem i've had happens when I either render, edit or scrub through some fusion heavy clips. Just tried to open one of the fusion heavy clips and Davinci Resolve just froze. This is a screen shot of the fusion comp:

Screenshot 2025-01-20 200449.png (126.14 KiB) Viewed 3619 times

I have been trying to export my project but it started to freeze when I got to 21% completed (16.02 seconds in). So I tried to omptimize the media in the fusion comp at that part (16.02) and now my pc blue screen instead of the render freezing. I tired rendering the clips in the edit page and my PC blue screened again with the same error code.

I've had the same BSOD before when working with the fusion comp I showed above. The only things I suspect could be causing this is me not optimizing the medias I am using in the fusion comp. These media in nodes all have multiple different formats and probably different encodings as well.

I also think some of the media might be corrupted because a few months ago my RAM broke and caused a lot of problems. I reinstalled windows and all my programs on my C: drive then, but I have my media on my D: drive. And I wouldnt be able to find all the media I have again. So it isn't reinstalled. But when I play my fusion comp I haven't seen any time when the media in nodes isn't working.

One other thing that could be causing this is all the media I have in the media pool. I imported too many folders into the project as I didn't know this could cause problems. Now it takes a few seconds every time I open Davinci Resolve before I can do anything. I probably have at least 200GB of files in the media pool.

Windows is updated but I think I'm on Davinci Resolve 19.1.2. All my softwares should be up to date aswell and I'm using NVIDIAS Studio drivers.

I'm ready to clarify anything. Sorry if the details I'm giving are bad or confusing.

My PC specs are:
GPU: RTX 3080
CPU: i5 13600K
RAM: 16GB DDR4 (haven't received my replacement sticks from when the old ones broke)
Drives: I have Davinci Resolve on my C: rive (Kingston SKC3000S), and all my media on my D: drive (WD black m.2 nvme ssd), and all my cache on my E: drive (WD blue m.2 nvme ssd) all of my drives are m.2 NVME PCIE gen 4 1TB ssd:s.
PSU: Seasonic PRIME-TX 850W
My PC have extremely good cooling

Edvin, 16 GB of RAM is very low for any Fusion work.

But the BSOD "system thread exception not handled" points to a maybe old and unreliable hardware driver. Normally those could be one out of these: audio drivers, network drivers, and GPU drivers.

So it is very likely not related to a heat problem of your CPU.

Can you try the following:

* Open the device manager
** Open the network adapter node
-> Right click each of your available network devices (e.g. RealTek / Intel / etc. / ignore those WAN nodes) and select Properties and have a look at the driver tab

Post the exact device name here and its driver provider and version.

Do the same for your audio devices and GPU.

Also post the exact name/modell of each of your SSD drives. WD black/blue is not enough.

Then we can have a look if drivers or SSD firmware could pose a problem.

Next do this:

Step 1: Running a SFC Scan to check for corrupted system files

SFC automatically scans and finds issues with your system, after it will begin repairing them.

SFC has to be executed in an elevated Command Prompt and therefor can only be started by a system administrator or user with Administrator rights.

Open Elevated Command Prompt by following one of two methods:
Press the Windows + X at the same time, then choose Command Prompt (admin)
Press the Windows + W key at the same time, and then type “ cmd “ right click Command Prompt and choose Run as Administrator
Type the following command

Code: Select all

sfc /scannow

and hit enter

Wait for the results to come back.

-> If you get the message that "Windows Resource Protection did not find any integrity violations" then there were no system files damaged or corrupted, and Windows Resource Protection had nothing to repair, the problems you are experiencing with your system might have other causes.

-> If you get the message that "Windows Resource Protection found corrupted files and successfully repaired them" then Windows Resource Protection found and solved damage system files and / or corrupted system files.

-> If you get the message that "Windows Resource Protection found corrupted files but was unable to fix some of them" then Windows Resource Protection found the problem but was unable to fix them, like the message said. If that's the case we have to try another tool later.

Having bsods in DR too. Just stated learning the software, and I can assure you I haven't EVER had a single bsod in Premiere Pro or After Effects on my current rig for almost 3 years. DR has crashed several times already, but this is nothing, but I've had bsods 3 times during the last two weeks. This puzzles me alot, because DR is the only software that causes bsods. I've got a ton of plugins and different media editing software installed, but DR is "empty" - I installed it, imported my videos and started learning it. I didn't even fiddle with the preferences much. My dystem is upto date, Win11 auto updates, all the drivers are the latest.
14700k, 4080, 64 gb of ram. So again DR (also up to date) is the only software I'm getting bsods in. It normally happens when I switch to Fusion or back to the Edit tab od ehen there are some transitions played back.
Whetre shall I dig to?

Your BSOD points to a GPU driver issue:

https://mspoweruser.com/dxgmms2-sys-windows-11/

Your BSOD points to a GPU driver issue:

https://mspoweruser.com/dxgmms2-sys-windows-11/

Thank you, Robert, I'm trying these solutions right now, though I do believe these general solutions never help. Right now I deleted the cuda cache from DR folder and recached the neuro engines. The SFC scannow command showed no errors. Now is my memory being checked, then the checkdisc will start.
What I don't understand is - I have Topaz software which has ai engines in it, a lot of Blace plugins for Ae that are based on ai, I have a stable diffusion plugin (diffusae) that uses tens of different ai engines, even Premiere Pro uses ai which is shown by the GPU graph in the Win Task manager... Everything has worked for years. Not a single bsod. I do mean to say that the last time I saw a blue screen shot of death on my rig is some 4 or 5 years ago on Win 10, and that was on my old PC. Why do I see bsods in DR, what is so particular in its ai engines? If there's a problem, and as it turns out there are numerous complaints about this issue in DR 19 particularly, shouldn't Blackmagic raise this question with Microsoft (for DirectX) or NVidia (for RTX) to investigate and fix it?

I have the same problem of frequent BSOD with DR. I noticed that the problem occurs if I use DR immediately after using AVCLABS Video Enhancer AI. The problem does not occur if after using AVCLABS I reboot the PC before using DR.

Many people think that Resolve causes Bsdo be cause is during his work, but the real fact is ever another. The bado is not the problem, is a Symptom of problem.

Bsdo is an hardware fault that causes the lock of windows core, and happened for windows/hardware problems.

If you not reach it with other applications is only be cause most of app are worst written and not use really all resources of hardware.

In the same condition if you start a 3d render with gpu like cinebench 2024 for 10-20 minutes you reach the same Bsdo be cause use all resources.

Or if you test some of complex with demo of topaz video enhancer, which use heavily the gpu, the same.

Premiere and after not use all resources from release 2014, the latest multicore multi thread version of that software (ten years ago), at today if you see the activity monitor you could see low use of resources.

Check first heat of every cpu/gpu etc
Check driver error on log
Check memories (ones happens that the order which I put the exact same memories on 4 banks causes me strange problems, swap of them solved it)
Good luck


Inviato dal mio iPhone utilizzando Tapatalk

Many people think that Resolve causes Bsdo be cause is during his work, but the real fact is ever another. The bado is not the problem, is a Symptom of problem.

Bsdo is an hardware fault that causes the lock of windows core, and happened for windows/hardware problems.

If you not reach it with other applications is only be cause most of app are worst written and not use really all resources of hardware.

In the same condition if you start a 3d render with gpu like cinebench 2024 for 10-20 minutes you reach the same Bsdo be cause use all resources.

Or if you test some of complex with demo of topaz video enhancer, which use heavily the gpu, the same.

Premiere and after not use all resources from release 2014, the latest multicore multi thread version of that software (ten years ago), at today if you see the activity monitor you could see low use of resources.

Check first heat of every cpu/gpu etc
Check driver error on log
Check memories (ones happens that the order which I put the exact same memories on 4 banks causes me strange problems, swap of them solved it)
Good luck


Inviato dal mio iPhone utilizzando Tapatalk

Hi, carlomacchiavello, thanks for joining the conversation here.
I can asure you that I literally tested my PC with every possible tool once I bought it 2,5 years ago and I've never had any problems. So I ran Cinebench for an hour again today and it gave me no errors (and a total 34000+ points at 88 degrees which is alright for my 14700K). As well as OCCT. As well as the MemTest and FurMark. My CPU temps can go 80-85 degrees only in Premiere Pro or some encoding in Shutterencoder. These two are know for an extensive CPU usage. In DR I've never seen my CPU go higher than 65+ degrees. The GPU is the same - within 60-70 degrees, and most of the time it's 45-50. The memory tests I had started this morning finished with no errors as well, so as the checkdisc.
So I analyzed the system logs just now and found only some Resolve crashes I'd had:

Code: Select all

Faulting application name: Resolve.exe, version: 19.1.3.7, time stamp: 0x00000000
Faulting module name: ucrtbase.dll, version: 10.0.26100.3037, time stamp: 0x8481217d
Exception code: 0xc0000409
Fault offset: 0x00000000000a4c2e
Faulting process id: 0x2A3C
Faulting application start time: 0x1DB94E31E916C4E
Faulting application path: C:\Program Files\Blackmagic Design\DaVinci Resolve\Resolve.exe
Faulting module path: C:\WINDOWS\System32\ucrtbase.dll
Report Id: 2d4288f3-dd26-4f17-b11b-44910ac0d5b0
Faulting package full name:
Faulting package-relative application ID:

and something like this, not sure if it was the bsod of yesterday:

Code: Select all

The previous system shutdown at 18:28:40 on ‎14.‎03.‎2025 was unexpected.
- <Event xmlns="http://schemas.microsoft.com/win/2004/08/events/event">
- <System>
  <Provider Name="EventLog" />
  <EventID Qualifiers="32768">6008</EventID>
  <Version>0</Version>
  <Level>2</Level>
  <Task>0</Task>
  <Opcode>0</Opcode>
  <Keywords>0x80000000000000</Keywords>
  <TimeCreated SystemTime="2025-03-14T15:53:16.5947928Z" />
  <EventRecordID>24517</EventRecordID>
  <Correlation />
  <Execution ProcessID="3272" ThreadID="3356" />
  <Channel>System</Channel>
  <Computer>DESKTOP-4GLHG9P</Computer>
  <Security />
  </System>
- <EventData>
  <Data>18:28:40</Data>
  <Data>‎14.‎03.‎2025</Data>
  <Data />
  <Data />
  <Data>91056</Data>
  <Data />
  <Data />
  <Binary>E907030005000E0012001C002800F400E907030005000E000F001C002800F400600900003C000000010000006009000001000000B00400000100000000000000</Binary>
  </EventData>
  </Event>


I'll try to gather info if (when?) it bsods again, bot so far DR is the only software I've had for years which gives me pain. Though I like this VEP immensely...

I can't even remember when was the last time I got BSOD! Maybe 15 years ago! Who knows

From my 36 years of experience, I can say with confidence that it's not about the specs, it's about the brand which plays a big role in system stability and performance. Reputable brands do extensive testings to make sure all components, firmwares and drivers are %100 compatible which translates to a solid system.

Don't just run after the highest specs at the lowest prices, go with a well known and reputable brand to secure a dependable and reliable system.

Hi, carlomacchiavello, thanks for joining the conversation here.
I can asure you that I literally tested my PC with every possible tool once I bought it 2,5 years ago and I've never had any problems. So I ran Cinebench for an hour again today and it gave me no errors (and a total 34000+ points at 88 degrees which is alright for my 14700K). As well as OCCT. As well as the MemTest and FurMark. My CPU temps can go 80-85 degrees only in Premiere Pro or some encoding in Shutterencoder. These two are know for an extensive CPU usage. In DR I've never seen my CPU go higher than 65+ degrees. The GPU is the same - within 60-70 degrees, and most of the time it's 45-50. The memory tests I had started this morning finished with no errors as well, so as the checkdisc.
So I analyzed the system logs just now and found only some Resolve crashes I'd had:

Code: Select all

Faulting application name: Resolve.exe, version: 19.1.3.7, time stamp: 0x00000000
Faulting module name: ucrtbase.dll, version: 10.0.26100.3037, time stamp: 0x8481217d
Exception code: 0xc0000409
Fault offset: 0x00000000000a4c2e
Faulting process id: 0x2A3C
Faulting application start time: 0x1DB94E31E916C4E
Faulting application path: C:\Program Files\Blackmagic Design\DaVinci Resolve\Resolve.exe
Faulting module path: C:\WINDOWS\System32\ucrtbase.dll
Report Id: 2d4288f3-dd26-4f17-b11b-44910ac0d5b0
Faulting package full name:
Faulting package-relative application ID:

and something like this, not sure if it was the bsod of yesterday:

Code: Select all

The previous system shutdown at 18:28:40 on ‎14.‎03.‎2025 was unexpected.
- <Event xmlns="http://schemas.microsoft.com/win/2004/08/events/event">
- <System>
  <Provider Name="EventLog" />
  <EventID Qualifiers="32768">6008</EventID>
  <Version>0</Version>
  <Level>2</Level>
  <Task>0</Task>
  <Opcode>0</Opcode>
  <Keywords>0x80000000000000</Keywords>
  <TimeCreated SystemTime="2025-03-14T15:53:16.5947928Z" />
  <EventRecordID>24517</EventRecordID>
  <Correlation />
  <Execution ProcessID="3272" ThreadID="3356" />
  <Channel>System</Channel>
  <Computer>DESKTOP-4GLHG9P</Computer>
  <Security />
  </System>
- <EventData>
  <Data>18:28:40</Data>
  <Data>‎14.‎03.‎2025</Data>
  <Data />
  <Data />
  <Data>91056</Data>
  <Data />
  <Data />
  <Binary>E907030005000E0012001C002800F400E907030005000E000F001C002800F400600900003C000000010000006009000001000000B00400000100000000000000</Binary>
  </EventData>
  </Event>


I'll try to gather info if (when?) it bsods again, bot so far DR is the only software I've had for years which gives me pain. Though I like this VEP immensely...

uhm i'm not able to give you more suggest,
i checked and ucrtbase.dll is a basic dll from MSDNKIT for software dev, standard windows dll.

Edvin,

As a starting point you really need to get 32GB back on your machine. BMD themselves ask for 32GB as MINIMUM for Fusion. Your Fusion design looks complex and random crashes are normal when you have little RAM/VRAM. From what I see, Resolve is not great at dealing with lack of resources. It just quits when that happens!

BSOD might be another thing but I'd say get the PC back to 32GB as a starting point.

And don't listen to the "If you get an HP machine all your issues will disappear" please.

All of you with BSODs have Intel 13th and 14th Gen CPUs. Make sure these have not degraded. 13th gen and 14th gen have degradation issues.

I would check RAM first though.

ah yes, the Intel plague of 13th gen...

The degradation was the first thing I thought of. But there are BUTs. 1) I updated the Bios firmware every time it came out, from the first day of purchacing, and downvolted my CPU additionally. 2) I have to keep repeating that my rig passes EVERY possible test outhere. Should it be a CPU degradation, it would have happened elsewhere. Again, it's the first time I've seen a bsod for almost 3 years and it happened while using DR, that tells something. 3) The bsod shows it's a GPU (directX) related issue. The CPU bsod codes have different numbers and descriptions (like the infamous Watchdog error).

So again we're coming bach to DR... I've come across similar posts from DR users in Reddit. It's not a coincedence. I certainly understand it may be related to the drivers, because there are numerous complaints about the current NVidia drivers, as well as the OS in general for Windows 11 being a piece of glitchy crap. But still, DR is the only software using which I've had bsods, so... This matter should be investigated by the devs. Thanks god, DR saves nonstop, otherwise I'd have to go gack to Premiere Pro again...

Why not uploading a BSOD dump on elevenforums and ask for their opinion? I am not skilled when it comes to BSODs and there are many tools to diagnose that.
Elevenforums are usually quite helpful.

If no other torture stress can cause issues, I'd agree that the CPU might not be at fault. Also, what version of Resolve are you using? Free or Studio? Studio is GPU intensive, Free is using the CPU for many tasks.

Why not uploading a BSOD dump on elevenforums and ask for their opinion? I am not skilled when it comes to BSODs and there are many tools to diagnose that.
Elevenforums are usually quite helpful.

If no other torture stress can cause issues, I'd agree that the CPU might not be at fault. Also, what version of Resolve are you using? Free or Studio? Studio is GPU intensive, Free is using the CPU for many tasks.

Thanks for this Elevenforums tip, I'll try that.
It's Studio 19.1.3 build 7

Studio is definitely using GPU a lot.

basic stuff - and I don't mean to imply you don't know about your computer! - are you sure you're wired to the GPU port and not the iGPU one (I don't even know if your CPU has one)? Did you check in preferences that the correct GPU is being selected and did you try removing the automatic selection?

It's unlikely as this feels more a driver/windows issue but you never know.

Also, you might want to test the system to its "defaults" - that is, no XMP, no undervolt, no overclocks etc.
Don't get me wrong, undervolt is great, both my Ryzen and my GPU are undervolted but undervolting CAN cause instabilities. Yes, it should show up in your other torture tests for sure...

Studio is definitely using GPU a lot.

basic stuff - and I don't mean to imply you don't know about your computer! - are you sure you're wired to the GPU port and not the iGPU one (I don't even know if your CPU has one)? Did you check in preferences that the correct GPU is being selected and did you try removing the automatic selection?

It's unlikely as this feels more a driver/windows issue but you never know.

Also, you might want to test the system to its "defaults" - that is, no XMP, no undervolt, no overclocks etc.
Don't get me wrong, undervolt is great, both my Ryzen and my GPU are undervolted but undervolting CAN cause instabilities. Yes, it should show up in your other torture tests for sure...

Sure thing, the Monitor is connected to the GPU, the iGPU in my CPU (14700K) is used sometimes for decoding in DR - seen the graph spikes in the Performance tab - but generally everything is done by the GPU. To my surprise I saw the CPU go up today for the first time in 2 weeks of using DR - when I resolved the 3D camera, so like After Effects, DR must be a CPU process as well.
The preferences are all automatic - never changed anything there:

Screenshot 2025-03-17 173933.png (59.13 KiB) Viewed 2685 times

So since the GPU is used most of the time I think DR is automaticly set correctly. I asked in another discussion here how I should turn on the iGPU instead of the GPU to decode my footage, like I have it in Premiere Pro, but no one knows (the iGPU is greyed out, though it's used when the OpenCL renderer is selected on purpose).
As for the undervolt - I can see your point and that might be the reason but again the bsod screen says it's a dxgmms2.sys that causes the bsod which is a GPU-related process or file... IThe problem is that this issue does not manifest itself everyday - haven't had it for 4 days working in DR all this time. And should I remove undervolt I'll hit 100 degrees in render for sure, though in games it's barely 60 and even in DR it's up to 70. Render is the toughest process out there.
Another thing I've just found - a guy on my local forum posted this:

I confirm, the driver 551.86-desktop-win10-win11-64bit-international-dch-whql is installed, everything is fine. Recently there was 572.60-desktop-win10-win11-64bit-international-nsd-dch-whql (February 2025), everything was fine for a long time, and then a dxgmms2.sys error with a blue screen popped up. After removing this driver using the Display Driver Uninstaller utility and reinstalling the driver version 551.86, it was impossible to start the NVIDIA control panel, I had to reinstall Windows.

The guy works in Fusion 360, Blender and Cinema 4d - all the GPU extensive applications. So I think I'm gonna try the driver downgrade first once I come across a bsod again, because mine is a Studio 572.60.
And I think I'll better start a discussion on the NVidia forum too

I recommend removing the auto on the GPU - you never know.
The iGPU could be used for some 4:2:2 decoding I think, something nVidia does not offer. I'd imagine Resolve might be able to use your iGPU automatically when needed but if you're having instabilities it might be worth telling Resolve to use the 4080 only for now. As a test.

My 3070 stays much cooler in DR than when playing games - though I don't do complex things with Fusion or special effects.

THAT SAID, a 4080 should normally run pretty cool - I see tests where it normally stays at 60+ degrees with heavy gaming. What temperature do you get by running something reasonable, say Unigine superposition?

I know you're saying that the error is with a GPU related file but if your system is unstable because of undervolting, anything can happen. I think it's only reasonable to start from the default to rule out those possibilities. Your GPU should be designed to deal with Resolve/a game at its stock settings. If it does not, then it's time to have it re-pasted

Trying different drivers is also an option - I'd imagine you're using the Studio nVidia drivers?

I recommend removing the auto on the GPU - you never know.
The iGPU could be used for some 4:2:2 decoding I think, something nVidia does not offer. I'd imagine Resolve might be able to use your iGPU automatically when needed but if you're having instabilities it might be worth telling Resolve to use the 4080 only for now. As a test.

My 3070 stays much cooler in DR than when playing games - though I don't do complex things with Fusion or special effects.

THAT SAID, a 4080 should normally run pretty cool - I see tests where it normally stays at 60+ degrees with heavy gaming. What temperature do you get by running something reasonable, say Unigine superposition?

I know you're saying that the error is with a GPU related file but if your system is unstable because of undervolting, anything can happen. I think it's only reasonable to start from the default to rule out those possibilities. Your GPU should be designed to deal with Resolve/a game at its stock settings. If it does not, then it's time to have it re-pasted

Trying different drivers is also an option - I'd imagine you're using the Studio nVidia drivers?

I think much depends on the CPU itself and its ability to make the GPU use its full potential. With Superposition 1080p Extreme I get this result - it's only 66 degrees which if perfectly fine for a 2 year-old videocard:

Screenshot 2025-03-18 115247.jpg (113.84 KiB) Viewed 2476 times

I also ran the Furmark with the GPU-Z open:

111.jpg (813.76 KiB) Viewed 2476 times

I now am on the older studio video driver, so I'm gonna observe the behavior with it for now. I also emailed NVidia support and sent them all the needed data and system logs.

I think anyone who has other symptoms or just bsods for no reason in DR, should add their info here, in this discussion, too.

I'd agree your GPU seems perfectly healthy. I'm surprised that Resolve is more demanding than Furmark - I don't see that on my 3070 but I'd imagine you use some demanding effects. Can you name one for me to test here?

In any case, I hope you can sort that out and 11forums can help you!

Updated my videocard driver and the iGPU driver today (newer versions came out) and again got a bsod working in DR... And this is after almost a whole day of video editing in Premiere without any issues... I don't get it...
@Tony359
I'm using 2 glitch transitions from Red Giant Univers from Boris FX. Never had any issues with them, and the bsod just now happened when I was trying to retime a video clip with an optical flow on.

Code: Select all

Fault bucket , type 0
Event Name: BlueScreen
Response: Not available
Cab Id: 0

Problem signature:
P1: 7e
P2: ffffffffc0000005
P3: fffff80440e23f83
P4: ffffd98ae2f4ed58
P5: ffffd98ae2f4e540
P6: 10_0_26100
P7: 0_0
P8: 256_1
P9:
P10:

Attached files:
\\?\C:\WINDOWS\Minidump\031925-14375-01.dmp
\\?\C:\WINDOWS\SystemTemp\WER-29328-0.sysdata.xml
\\?\C:\WINDOWS\MEMORY.DMP
\\?\C:\ProgramData\Microsoft\Windows\WER\Temp\WER.da163f53-c302-4c3d-a76a-d4c45a4f5a47.tmp.WERInternalMetadata.xml
\\?\C:\ProgramData\Microsoft\Windows\WER\Temp\WER.414082ce-7bb8-497c-87f1-9b8be0aae221.tmp.csv
\\?\C:\ProgramData\Microsoft\Windows\WER\Temp\WER.bbee7b8b-9dea-4354-8e33-1e225a456f16.tmp.txt
\\?\C:\ProgramData\Microsoft\Windows\WER\Temp\WER.05f1c93c-8634-486a-8da4-bff3d462cd9b.tmp.xml

These files may be available here:
\\?\C:\ProgramData\Microsoft\Windows\WER\ReportQueue\Kernel_7e_e1435c8cfc659752d8d7fe22bc289ba4eee3551_00000000_dbca89d2-88de-4dbb-9b48-7522796ac2b8

Analysis symbol:
Rechecking for solution: 0
Report Id: 53a72957-036f-4c34-9d4f-51e563a60714
Report Status: 2049
Hashed bucket:


You meant Resolve, right?

Did you ask on elevenforums?
Did you try setting your gear to stock? Sorry for insisting on those points but any under/over_whetever can bring instability.

What about the older drivers you were using? I thought they were working for you?

Did you run a memtest overnight - or even more?

You meant Resolve, right?

Did you ask on elevenforums?
Did you try setting your gear to stock? Sorry for insisting on those points but any under/over_whetever can bring instability.

What about the older drivers you were using? I thought they were working for you?

Did you run a memtest overnight - or even more?

No, I actually meant Premiere. I've worked in it since this mornning. In the evening I launched DaVinci with my current project there. And after half an hour, fiddling with retiming, I got a bsod.
I tried the elevenforums, but no one responded there yet. I think once they do, they'll tell me to go to Resolve forums. It's a circle...
My PC is at stock now and still there's a bsod issue and only in one program. Nvidia support suggested these:

Code: Select all

Open NVIDIA control panel (Right click on desktop and click NVIDIA control panel) > Go to Manage 3D settings under 3D settings > Select Global settings > And now change these settings from Option 2.
- Power management mode - Prefer maximum performance
- CUDA GPUs - RTX 4080
- OpenGL rendering GPU - RTX 4080

3). Disable Windows Game Mode and Xbox Game Bar.
(Click on the Windows start button > Settings > Gaming > Xbox GameBar / Game Mode)

4) Disable Hardware Accelerated GPU scheduling
- Click on the Windows start button > Settings > System > Display
- Under the “Multiple Displays” section, select “Graphics settings”.
- Turn off the “Hardware-accelerated GPU scheduling” option

So, it didn't help. An older video driver is the last resort for me... Then I'll have to stop learning DR and go back to Premiere + Ae workflow...

Or go to Mac Mini M4.......

elevenforums and tenforums are usually quite good - PROVIDED you gave them all the data required in the rules. Give them some time. I doubt they'll send you back here, they'll tell you what's causing the issue.

I'm surprised, Resolve is pretty solid here. It tends to be unstable when you lack RAM or VRAM. But I don't think it's your case - though I don't use those effects. Bear in mind that Resolve is designed to edit Hollywood stuff: just wondering whether maybe you're using effect which would require even more powerful HW?

Too bad BMD staff only seems to answer to posts about Resolve not starting

If you have Studio you can email then for support. They might be able to help you - maybe they can see something from DR logs.

Sorry I can't help you more...

Mac Mini M4: I have one but you wouldn't compare the Mac Mini M4 to a 4080, right?? Let's be serious. It's an amazing machine but it's not replacing a £2000 PC!

So there are two of us - a guy on Reddit has the same exact bsod code in DR like me. I'm relieved - I'm no crazy

https://www.reddit.com/r/davinciresolve ... _of_death/

So there are two of us - a guy on Reddit has the same exact bsod code in DR like me...

Apparently, you've never uploaded any logs or Windows .dmp crash files, but the other user on Reddit did, so I analyzed his Windows .dmp kernel dump file using Windbg. Observations from his crash:

- The failure occurred in dxgmms2.sys ; DirectX Graphics Memory Management System driver, Windows kernel component that manages GPU memory.

- Specific crash point: dxgmms2!GetVidMmGlobalAllocFromOwner+0x16 - The failure happened during GPU memory allocation operations.

- Exception type: Access violation (0xC0000005) - The driver attempted to read from an invalid memory address (0xffffdc8100000001).

- Process context: "System" - This confirms the crash happened in a kernel-mode thread, not within the DaVinci Resolve process itself.

Kernel-mode stack backtrace:

dxgmms2!GetVidMmGlobalAllocFromOwner
dxgmms2!VIDMM_PAGE_DIRECTORY::ExpandLargePagePte
dxgmms2!VIDMM_PAGE_DIRECTORY::CommitVirtualAddressRange
dxgmms2!CVirtualAddressAllocator::CommitVirtualAddressRange
dxgmms2!VIDMM_GLOBAL::PageInOneAllocation
dxgmms2!VIDMM_GLOBAL::ProcessDeferredCommand
dxgmms2!VIDMM_WORKER_THREAD::Run
dxgmms2!VidMmWorkerThreadProc
nt!PspSystemThreadStartup
nt!KiStartSystemThread

This is clear case of a kernel-mode driver defect. The driver is mishandling memory operations related to GPU resources.

If you will upload one or more of your Windows .dmp files to a file sharing site I can access, I will analyze them. They are usually in C:\Windows\Minidump, with filename something like 030425-9500-01.dmp.

A BSOD is caused by a hardware or driver bug. There is no software code in Resolve that can be changed to fix a BSOD. I'm not an expert at Windows, but if you upload your .dmp file, I'll analyze it.

If you will upload one or more of your Windows .dmp files to a file sharing site I can access, I will analyze them. They are usually in C:\Windows\Minidump, with filename something like 030425-9500-01.dmp.

A BSOD is caused by a hardware or driver bug. There is no software code in Resolve that can be changed to fix a BSOD. I'm not an expert at Windows, but if you upload your .dmp file, I'll analyze it.

Didn't realize they were so easy to find, thank you. Here they are: https://drive.google.com/drive/folders/ ... sp=sharing
These are all I have in that folder

Your three kernel-mode dumps indicate the same problem, and are very similar to the Reddit user's crash. Observations:

- The failure occurred in dxgmms2.sys ; DirectX Graphics Memory Management System driver, Windows kernel component that manages GPU memory.

- Specific crash point: dxgmms2!VIDMM_PAGE_DIRECTORY::ExpandLargePagePte+0xd3 - The failure happened during GPU memory allocation operations.

- The Windows VIDMM (Video Memory Manager) maintains page tables similar to the CPU's memory management system, translating GPU virtual addresses to physical locations in either VRAM or system memory.

- In your three crashes, a kernel-mode instruction attempts to read from virtual memory address 0x0000000000000048: This very close to address zero. In the Reddit crash, it was trying to read virtual memory from the highest part of the address map. Very low and very high address blocks are flagged as "out of bounds" because malfunctioning code tends to reference those. This preventatively crashes the system instead of letting it read invalid data, which would be worse.

- The analyzed dump file states PROCESS_NAME: System. The crash occurred in the kernel-mode System process, a special process that hosts many kernel-mode threads, including those for device drivers and memory management. It tells us that the failure happened during autonomous system operations, not directly tied to any user application's thread of execution. This further reinforces that the crash is happening in system-managed memory operations that no application could reasonably control or work around.

Kernel-mode stack backtraces (one shown, all three identical):

dxgmms2!VIDMM_PAGE_DIRECTORY::ExpandLargePagePte
dxgmms2!VIDMM_PAGE_DIRECTORY::CommitVirtualAddressRange
dxgmms2!CVirtualAddressAllocator::CommitVirtualAddressRange
dxgmms2!VIDMM_GLOBAL::PageInOneAllocation
dxgmms2!VIDMM_GLOBAL::ProcessDeferredCommand
dxgmms2!VIDMM_WORKER_THREAD::Run
dxgmms2!VidMmWorkerThreadProc

This is clear case of a kernel-mode issue. The Windows dxgmms2 driver is mishandling memory operations related to GPU resources. More precisely, the Windows DirectX Graphics Memory Management driver (dxgmms2.sys) is experiencing a null pointer dereference during GPU virtual memory page table operations.

Is it Microsoft's fault or a third-party GPU driver's fault? While the crash is manifesting in Microsoft's dxgmms2.sys, it's entirely possible—and even common—that the root cause originates in a vendor's kernel-mode driver. But we have no visibility of that solely from this dump. Making that determination would require kernel-mode debugging using two connected PCs with full debug symbols for all involved drivers.

It should be obvious from the above that this (and other similar BSOD cases) cannot be pursued as a Resolve problem. Attempting to do so wastes time that could be better spent with support personnel familiar with low-level system areas.

Brief tutorial on how to use WinDbg to analyze a BSOD:

Wow, @joema4, thank you for such an indepth analysis... The only question I have now - is what shall I do with that info...

Wow, @joema4, thank you for such an indepth analysis... The only question I have now - is what shall I do with that info...

I don't even have a Windows computer, but I ran WinDbg on my Mac via Parallels to try and help you out. Although I did low-level Windows work years ago, I've used only Macs since 2013, so I'm doing my best with old knowledge. Here are some forums that might be able to give suggestions for a Windows BSOD problem:

Tom's Hardware Forums: https://forums.tomshardware.com
Bleeping Computer Forums: https://www.bleepingcomputer.com/forums/
ElevenForums: https://www.elevenforum.com
Reddit Tech Support: https://www.reddit.com/r/techsupport/
Reddit Windows Help: https://www.reddit.com/r/WindowsHelp/
Reddit Build A PC Forum: https://www.reddit.com/r/buildapc/
Microsoft Community: https://answers.microsoft.com/en-us
Hardware Manufacturer Support (whoever made your motherboard).
Overclocking Forum: https://www.overclock.net
Reddit Overclocking Forum: https://www.reddit.com/r/overclocking/

Thanks @joema4 that is just amazing. I will study the tutorial you linked, I'd like to learn how to understand a BSOD!

Andrew,
BMD are probably aware and will try to push a resolution - either a workaround on their end of a fix from either MS or nVidia. Could using a different "SuperScale" be a temporary solution?

Andrew,
BMD are probably aware and will try to push a resolution - either a workaround on their end of a fix from either MS or nVidia. Could using a different "SuperScale" be a temporary solution?

Well, I've tried to cause a bsod today the whole day, using all the Superscale options, moving the timeline with effects like crazy, jumping back and forth between Edit and Fusion tabs and using the effects that made my gpu sweat as hell, and I don't wanna be too optimistic but I didn't get a bsod. Though tomorrow is the heavy work and if it runs without issues I'll start excluding possible causes. Right now I'm on a 566 something NV driver with no xmp and no downvolting. Is the Superscale is going to cause it again, I'll be sure it's not my system, but Superscale or something else in DR plus something else in my system - rather a driver or OS

Yet another opinion: https://www.reddit.com/r/24hoursupport/ ... are_button

Hey everyone, I’m jumping in a bit late here, but I’m experiencing the same issue. I was working on a Fusion composition that included a 3D model and a hologram effect, and now every time I try to open the project, I get a BSOD. I analyzed the dump file, and it points to the same issue with the dxgmms2.sys module. I’m not sure how to resolve this—has anyone managed to fix it? Here are my specs:

GPU: NVIDIA GeForce RTX 5070 (no overclocking applied).

Drivers: NVIDIA Studio 572.83 (dated March 18, 2025). I’m planning to switch to the Game Ready driver to see if it helps.

DirectX: Version 12.

CPU: Intel(R) Core(TM) i7-14700K.

RAM: 64 GB with XMP enabled (I’ll test disabling XMP to check for stability).

Since I can consistently recreate the problem by simply opening the project, I ran HWInfo to monitor GPU activity and logged the sensor data leading up to the crash. Surprisingly, I noticed that the allocated and dynamic GPU memory values were at zero across several rows, even though I was working on a resource-intensive 3D project. This feels like a potential memory management issue tied to the drivers.

If anyone has encountered something similar or has any suggestions, I’d love to hear your input! Thanks in advance.