BUG: dwm.exe uses memory leakage with Intel HD Graphics 630

[Cody]

Hi,

dwm.exe (Desktop-Window Manager) uses high memory with Intel Intel HD Graphics 630. This problem has been reported before:

https://answers.microsoft.com/en-us/windows/forum/all/desktop-windows-manager-dwm-memory-leakage-in/c2968fb0-ee26-4483-9f08-dfa19149cd24

https://www.reddit.com/r/windows/comments/iqttw5/windows_10_dwmexe_high_ram_usage/g4umzyn/?utm_source=share&utm_medium=web2x&context=3

Here's a screenshot of my dwm ram usage after 2 hours (1.6 GB):

Please, this bug is already known for 3-4 months, and I do not want to reinstall a 2019 driver...

[HKlum]

@Almighty1 " I just discovered that my v30.0.101.1002 somehow is back at v30.0.100.9976 " 🙂

[TurboTacho]

Hi Almighty1,

I understand everyone is frustrated since this issue has been left floating for a year or more. Just that I feel that sometimes we have to think from a different perspective (i.e. from the mods, devs, and others). I'm not saying that this is being very personal but I do hope that we are all at least not that heavily pessimistic and working together to help Intel devs develop a proper resolution, that would benefit the greater demographics around the world.

I know some of our feedback context may sway off the rails, which is why having HKlum to keep a lookout, as well as everyone else for each other, is pretty much appreciated. At least we know we're on the right track at most! 😀

Anyway, off the sidetrack and into the main topic, regards to your Windows rolling back the drivers on its own, or even overriding your installation of your graphics driver, perhaps you might want to try disabling Windows Update (there's an app called WUB if I can recall) to see if it does suppress and give you greater flexbility testing out the experimental drivers.

[Almighty1]

@TurboTacho  - What you said is correct that we should all work together and hope for a proper resolution from Intel in the near future.  Ofcourse I am sure COVID-19 has something to do with it.  I have a Swiftpoint ProPoint Mouse which under S0 Modern Standby, the system idle timer would work correctly but with S3, the mouse software is preventing the system idle timer from counting so the display never goes to sleep.  This was under Windows 11 and I provided some diagnostic logs that they requested and after about 3 weeks, they provided a new version of the software which fixed the problem for me but not the rest of the world because the latest release is still older than the test version that fixed the problem but just saying that Intel's developers can also have some tools available for the people who are willing to help to provide the data they need to fix the problem.   The memory problem with dwm.exe is almost as bad as the explorer.exe problem with Windows 11 which they fixed in 22000.120 partially after just 1 week of complaints but the real fix is in the 24xxx.x versions of Windows 11 in the development channel and never made it to 22000.194 or 22000.282.  Atleast it's good to know that Gabriella has been able to reproduce the issue and it's moving in the right direction.  I will need to spend some time to actually read through the entire thread to have a better understanding of all the details about the memory leak.  As for the rollback, I will have to look into that further on my own when I have a chance since I hate rebooting if I can avoid it on weekdays.

[HKlum]

@TurboTacho That post was meant to the post above of @Darch , he edited his post and deleted the important part, that he already tested 30.0.101.1002  and that it didnt fix anything, which I answered to, that it cant fix anything, because it is from before Intel even started to work on the fix. Thats all.

[TurboTacho]

I've got your message there. And yes, based on the timestamp I did believe that indeed there's non-continuity between when it was highlighted by Gabriela and the dev team.

Whereas for Darch's case, I guess it might be a few flag switches that may have made the difference. Unfortunately that is beyond my knowledge or understanding. Hence I wasn't able to test out the MPO flags on my end, which may have hindered clues by the Intel dev teams.

But I take this partially due to my lesser knowledge of experimenting even further with in-depth settings, for fear of not being able to reverse the procedures once there's a resolution found.

[Almighty1]

@HKlum - That was why that post was confusing the first time I read it because @Darch's original post was gone so it was hard to see what you replied to so the only thing I got out of it was that v30.0.101.1002 was out.  I actually misread it thinking that was the first version that had the fix instead of the work for the fix was after that driver was released which is just a clear misunderstanding.

[HKlum]

@Almighty1 Even this is off-topic I tried to help you a few times with the black screen issue, which, I say again, has nothing to do with this thread, yet I wanted to give you some hints, because I know this issue too, I experience over time, with my 9570. I told you to disable windows driver updates, which will cause issues, it will randomly roll back drivers, like you seem to have already noticed now yourself. So if you use Windows pro, use gpedit to disable it. First step. Afterwards update to latest driver and keep eye on if it stays on that version with no roll back happening. Keep MPOs disabled. Revert back to modern standby, Intel drivers became more and more buggy if you use S3 because they dont support S3 anymore and wont test them for S3 compatibility, can cause black screens. Then I dont advice to use S3 because it causes a firmware bug on the 9570 which causes a permanent 2W drain, even if you say you use power chord always, it keeps the system unstable over time, and also has a permanent runaway process of 10% on one core toggling around. Then, try undervolting plus a power limit with ThrottleStop, black screens can also mean thermal issues, with newer drivers using more or different resources triggering it more and react worse to thermal issues or voltage spikes, I could give you my settings if you like, and see if that stabilize your issues. I also advice to remove all XBOX and gamebar apps via powershell script, because that also caused black screens randomly for me, I could link you an older post about this. This is some bug it seems with DWM too MS never fixed so far.

[Almighty1]

@HKlum - Thanks for the pointers.  I'm actually on Home and not Pro.  Yes, I noticed whenever there is a new version of a driver that Windows has never seen before, it will always roll back after I install and I will have to install the driver I want over it and it will not roll back again.  I know about the drain because I believe there is a thread on Notebookreview and other places that talks about this in detail.  Undervoting was what I was trying to do except I've never found a good guide on how to do it and actually test for stability.  I didn't use S3 for the power management part because the other thing I noticed is that with the stock power plans, it seems like the i8750H 2.2Ghz processor will only hit around 2.5Ghz max while with S3, it hits 3.1Ghz.  That's why I was saying it's hard to really know where the black screens comes from as the screen is off unless there was a way to log what's happening in the background.  For now, I am actually trying to test things so I can report back to the other thread I started before finding this one before trying other things which takes dedicated time.  I remember the main issue with turning on S3 was that the machine will be so hot and not go to sleep which has the issues you mentioned but wasn't the resolution mentioned in the post here:
http://forum.notebookreview.com/threads/xps-15-9570-owners-thread.817008/page-294#post-10999744

Are you maffle by any chance because the chances of running into someone else who has a 9570 and actually knows about all the issues in that much detail is rare.  

[Darch]

Sorry for that, i didn't foresee this conversation and edited my first post(i wrote "will edit if problem returns" but anyways)

Tl:Dr; We have to wait new drivers.

[boneadventure]

Hello! Please tell me where can I download 30.0.101.1002 driver?

[HKlum]

@boneadventure Irrelevant, it doesnt fix the issue of the memory leak. No official way to download it so far from Intel just 3rd party. I wouldnt recommend it.

[boneadventure]

Thanks. It remains only to wait for the official solution...

[Ansuel]

released on 2 nov let me install it and tell you the real build time.

Nope it's older... it's just a driver for the new 12gen cpu

[Anonymous]

Hello!

When watching a video in the chrome browser for a long time, there is a memory leak in the DWM.

Laptop MSI GS65

Intel(R) UHD Graphics 630: 30.0.100.9955

NVIDIA GeForce GTX 1060: 30.0.14.9649

[Reck]

After several days without reporting any change on my part (The same problem of memory leak when starting the computer and without doing anything increases said consumption), today I experience an error with DWM.exe and the igd10iumd64.dll module, according to the monitor Reliability the module had an error and caused the application to crash, then after the error the screen was black for a few minutes since the system was still working because I was watching a video on YouTube, so just wait and windows indicated this failure As a hardware error, I know that it may not have anything to do with the main problem, but I also report it in case it is a problem derived from the main one.

Windows 11 Build 22000.282

Intel UHD Graphics: 30.0.100.9955

NVIDIA GeForce GTX 1650:  30.0.14.9649

[Stefan3D]

@Reck 

In your case find an application crash dump. It should be stored in one of the following folders:
c:\Windows\System32\config\systemprofile\AppData\Local\CrashDumps\
c:\Users\%username%\AppData\Local\CrashDumps\
c:\Users\%username%\AppData\Local\Temp\

Actually an application crash dump is easier to debug (at least for me) because it contains a program counter.

[Reck]

hi @Stefan3D ,

that dump file is not found in any of the paths you mention, it only appears in the reliability monitor with their respective error codes, versions, exception code and additional information codes, but the dump file is not in the paths what do you tell me.

[Stefan3D]

Did you go to the 1st path as admin?

On my rig i can see there a dump from another system app.

[Reck]

Yes, in fact I get dump files from other crashes, but dwm.exe does not appear anywhere

[TurboTacho]

Hello @Gabriela_Intel , @Bryce_Intel ,

I hope you guys are still going good there. Just a friendly knock on the door to check things and progress on the recent new discoveries that rolled off a few weeks ago here.

Hope to hear from you all soon

Cheers!

[Gabriela_Intel]

Here's the latest update:

We have another proposed fix in the works. Microsoft has tested the driver with the fix for a week and the good news is that they didn't observe any spikes.

We're doing another round of testing at the moment to check from our side as well.

By the way, we have a driver release planned for this week that will not contain the fix mentioned above. The fix will most likely be included in the following release and if it does, I'll post here to confirm so you can all test it out.

Thanks everyone for your patience! And thanks @TurboTacho for the friendly reminder!