Windows 1909 sluggish although all resource usage low ???

JoK

Weaksauce
Joined
Feb 24, 2017
Messages
97
Hi all,
I've run out of ideas here.
I've got a Dell XPS with Win 10 1909 and the last week or so it becomes sluggish in the sense that when I move the mouse or drag a window the system freeses momentarily (but for enough time to notice) before it performs the action. Even now that I am typing the letters do not appear in natural rate.

I am keep in an eye on resource usage in task manager and monitor and all is quite low (10-45%). The only value I see spiking is the GPU usage. Desktop Window Manager uses gpu even up to 80-90% for a moment and then down again. I can't say that the freese happens when this spike occurs...I would expect DWM to behave like this as it is responsible for rendering images and screens.

I have used malware scanners, antivirus and CCleaner alike apps but nothing fixes it.

I would appreciat eany ideas

Thanks
 
I had something similar to your issue. I used system restore and selected a restore point about a week earlier. Seems to have fixed what ever was messed up.
 
I have used malware scanners, antivirus and CCleaner alike apps but nothing fixes it.
And those are probably the reason why your computer is now swimming in tar. Try uninstalling your antivirus and all the 'protection' and see if it fixes the situation. Note that full uninstall and reboot is necessary, simply turning them 'off' will not remove the system hooks.
 
And those are probably the reason why your computer is now swimming in tar. Try uninstalling your antivirus and all the 'protection' and see if it fixes the situation. Note that full uninstall and reboot is necessary, simply turning them 'off' will not remove the system hooks.
Yes I thought of it as well but I found the problem.

The laptop has two graphic cards, an integrated (GPU-0) and a "high-performance" (GPU-1) one. I had changes the global choice to the high-performance instead of the integrated which is the default setting. This created all the sluggish behaviour.

Which takes me to another point of why would this happens? I know this is off-topic to this thread. When I check the performance tab in task manager, GPU-1 is hardly used. I even tried to change Netflix app to use it but for some reason it falls back to GPU-0.
 
And forgot to mention that I disabled the hardware acceleration from the browsers. I think this plays a role too
 
I have pretty much the same - I have a Dell E6540 with the Intel 4600 and AMD 8790M along with the AMD Control Center to select high performance or power saving. The only setting I've played around with was the graphics application setting. I have a couple of games that I manually selected high performance. The AMD piece does not run by default. It only turns on when it encounters a graphic type application.
 
Im guessing LatencyMon would show some interesting results? would answer a few questions anyway
 
That shouldn't be happening on your other GPU. What exactly is it? Then do a full uninstall and reinstall of the latest driver.
 
primetime Thanks for the suggestion. I ran both LatencyMon and WhySoSlow.

The setup is Intel(R) Core(TM) i7-7700HQ CPU @ 2.80GHz, The integrated GPU: Intel HD Graphics 630 and the other one is NVIDIA GeForce GTX 1050

Those apps primetime suggested put some light to the situation although they both report that the system is running quite well. The primary issue they both identify is that there is CPU throttling. Then, I turned to Intel Extreme Utility to check the actual CPU freq and it does confirm that there is power throttling set up. This makes the cpu to work at 25-30% of it's nominal frequency.

So I will now see how to lift this.

But the other issue with the two cards remains a mystery to me. I did some reading and most sources seem to suggest that the high-performance card kicks in when 3D capabilities are needed. I am not very convinced that this should be the case though. I mean I can see in task manager that the integrated GPU operates at 80% and the other one 0% for most of the time. Wouldn't be better is the load is split---or not?

Another conclusion I draw was that the change to graphic settings as to which app uses which card does not seem to be respected most of the times. In short, apps or rather Windows do what they want. As an example, I found impossible to direct Netflix app to use NVIDIA....if a video streaming app is not going to be benefited by a "high-performance" card, then which app will? Notepad?

Thanks people for the help
 
The Nvidia activates when there's 3D graphics, Netflix is 2D and apparently doesn't make use of hardware acceleration. Which is only natural especially if you used a browser with acceleration turned off. Why did you do that?
 
So, the external cards are used only when hardware acceleration and 3d is required? So, given I do not play any games, my system must not use it a lot.

About disabling hardware accelerator in the browsers, I did some reading and the consensus was that this is a feature of the past and modern processors do the required tasks anyway. Maybe I should turn it back on. But as said, I didn't notice any difference so far.
 
Mmmmm...
Let us know what else you find out. I suspect as in the past MS close something that was causing security issues and in doing so slowed things down. Kind of like what happens on the interstate. Close one exit and it slows everything down.

Also, though it might not be exactly your issue. I have noticed that the recent browser updates have made the memory handling a lot more sloppy. Seem like there are a lot of resources being used all the time. have you checked to see if there might be something with a driver that was changed or updated by MS that is part of the problem?
 
Quick Update:

Very soon I noticed that the processor was working at very low speed (~ 0.8GHz when the base frequenxy is 2.8GHz). I used ThrottleStop to observe the reasons for limiting down the frequency while the other resource usage was low to moderate. It seems that power and thermal throttling was causing it.

I then decided to add thermal pads to the motherboard and replace the thermal paste on the cpu and gpu. To my surprise, when I opened the case, Dell has thermal pads on two chipsets and on the SSD. I replaced them, added some more and replaced the thermal paste. I should also add pads on the memory chips but I didn't this time.

Now the laptop is much cooler and the cpu runs at high frew almost all the time (~3.xxGHz) while the temp is kept at ~70C. I undervolted the cpu as well and now even with heavy tasks, and some heavy testing which do 99% cpu usage for a few minutes, the temp doesn't go above 88C.

I am quite happy with the situation although not sure if having the cpu working at turbo freq all the time is good thing. What controls my concern is the low temps.

LatencyMon for some time wasn't identifying any problems but the last 3-4 days it says that the machine can not handle real time vid/audio. I tested expensively with amazon prime, two virtual machines running and some other tasks and the vid/audio works fine with the temps kept at good level. It also shows that the problematic driver is wdf01000.sys with lot of ISR and dxgkrnl.sys with high DPC couns. The first is a vital windows components and till now I haven't found a way to solve this issue. Some posts say it relates to the Wifi but I use an Ethernet cable for internet so not really sure whether the LatencyMon is right. I get the feeling though that all works fine.

dxgkrnl.sys is the kernel for DirectX and again wasn't able to find a solution to drop the dpc down.
 
Quick Update:

Very soon I noticed that the processor was working at very low speed (~ 0.8GHz when the base frequenxy is 2.8GHz). I used ThrottleStop to observe the reasons for limiting down the frequency while the other resource usage was low to moderate. It seems that power and thermal throttling was causing it.

I then decided to add thermal pads to the motherboard and replace the thermal paste on the cpu and gpu. To my surprise, when I opened the case, Dell has thermal pads on two chipsets and on the SSD. I replaced them, added some more and replaced the thermal paste. I should also add pads on the memory chips but I didn't this time.

Now the laptop is much cooler and the cpu runs at high frew almost all the time (~3.xxGHz) while the temp is kept at ~70C. I undervolted the cpu as well and now even with heavy tasks, and some heavy testing which do 99% cpu usage for a few minutes, the temp doesn't go above 88C.

I am quite happy with the situation although not sure if having the cpu working at turbo freq all the time is good thing. What controls my concern is the low temps.

LatencyMon for some time wasn't identifying any problems but the last 3-4 days it says that the machine can not handle real time vid/audio. I tested expensively with amazon prime, two virtual machines running and some other tasks and the vid/audio works fine with the temps kept at good level. It also shows that the problematic driver is wdf01000.sys with lot of ISR and dxgkrnl.sys with high DPC couns. The first is a vital windows components and till now I haven't found a way to solve this issue. Some posts say it relates to the Wifi but I use an Ethernet cable for internet so not really sure whether the LatencyMon is right. I get the feeling though that all works fine.

dxgkrnl.sys is the kernel for DirectX and again wasn't able to find a solution to drop the dpc down.
If the laptop had warranty you should have just returned the thing. If you fix it yourself Dell will continue to produce crap.
 
Here is mine when I took it apart to improve performance and cooling. Not a very good job that someone did applying thermal paste. Notice the thermal pad on the left side? It touches the small components that sit around the GPU. I've actually replaced the thermal pad on the heat pipe with brass shims that I bought at Amazon. I also bought some new thermal pads from Amazon to replace those on the heat pipe that touch the components around the GPU. When I received the shims, I made sure there were no sharp edges and they were smooth. I then applied some arctic sliver. I ran some games and the maximum temperature is now 80C.

BPMyf53.jpg
 
Last edited:
Here is mine when I took it apart to improve performance and cooling. Not a very good job that someone did applying thermal paste. Notice the thermal pad on the left side? It touches the small components that sit around the GPU. I've actually replaced the thermal pad on the heat pipe with brass shims that I bought at Amazon. I also bought some new thermal pads from Amazon to replace those on the heat pipe that touch the components around the GPU. When I received the shims, I made sure there were no sharp edges and they were smooth. I then applied some arctic sliver. I ran some games and the maximum temperature is now 80C.
That's what the thermal paste looked like in mine as well. But I have it 2-3 years already so I assume it naturally wears off
 
That's what the thermal paste looked like in mine as well. But I have it 2-3 years already so I assume it naturally wears off
If they use thermal paste that evaporates away, it's a clear case of failure by design...
 
  • Like
Reactions: JoK
like this
Thanks for the update. That is good to know and one of those little tidbits to put into the mental file for later use!

Ditto B00nie, designed obsolescence. That way they can get you to buy new hardware sooner than you need to. Good thing for the most part prices have come way down.
 
  • Like
Reactions: JoK
like this
Back
Top