Bug 469532 - KDE laggs horrible when io is active
Summary: KDE laggs horrible when io is active
Status: RESOLVED DUPLICATE of bug 469753
Alias: None
Product: plasmashell
Classification: Plasma
Component: generic-performance (other bugs)
Version First Reported In: 5.27.4
Platform: Other Linux
: NOR normal
Target Milestone: 1.0
Assignee: Plasma Bugs List
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2023-05-09 14:49 UTC by Moritz
Modified: 2025-08-13 22:44 UTC (History)
5 users (show)

See Also:
Latest Commit:
Version Fixed In:
Sentry Crash Report:


Attachments
System monitor (222.41 KB, image/png)
2023-05-09 14:49 UTC, Moritz
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Moritz 2023-05-09 14:49:02 UTC
Created attachment 158800 [details]
System monitor

SUMMARY
KDE is lagging horrible on all sorts of io events. It also affects other applications.

STEPS TO REPRODUCE
1. Copy files from one drive to another (or extract an archive, ...)

OBSERVED RESULT
KDE is lagging as hell

EXPECTED RESULT
Working fine, no lag spikes

SOFTWARE/OS VERSIONS
Operating System: KDE neon Testing Edition
KDE Plasma Version: 5.27.4
KDE Frameworks Version: 5.106.0
Qt Version: 5.15.9
Kernel Version: 6.0.19-060019-generic (64-bit)
Graphics Platform: Wayland
Processors: 12 × AMD Ryzen 5 5500U with Radeon Graphics
Memory: 17.4 GiB of RAM
Graphics Processor: AMD Radeon Graphics
Manufacturer: LENOVO
Product Name: 82KC
System Version: Lenovo V14 G2 ALC

ADDITIONAL INFORMATION
I do have an NVME SSD installed, I can also see this on a lot of other devies that run KDE software. I am observing this for years, but now (I am just copying 50 GB over the network) it is really REALLY annoying.


See the system monitor attachment.
Comment 1 Nate Graham 2023-08-04 20:22:18 UTC

*** This bug has been marked as a duplicate of bug 469753 ***
Comment 2 Moritz 2025-04-10 21:05:10 UTC
This is actually not a duplicate of the mentioned bug, the other bug describes high dolphin cpu load.


So reproducing this is quite easy with cheap consumer ssds. You copy files and the ssd cache fills up and then io operations on the ssd get really slow and pretty much stop. That is when everything (including the cursor) starts lagging and there are way less screen updates, applications (like dolphin) start to freeze and eventuelly (when the os flushed all file system changes and the ssd can write again) become response again.

Probably something (maybe in kwin?) is waiting for something from the ssd. Maybe it is also some io bug in the kernel? Did not track this one down.
Comment 3 Nate Graham 2025-04-11 18:38:15 UTC
It sounds like a hardware issue honestly.
Comment 4 Moritz 2025-04-18 14:29:09 UTC
Kind of, but more by design (or by money).

"Only" enterprise ssds are better in this case, but consumer ssds are just shit. I am observing this on 3 of my computers. 

But KDE has its issues there too, because I am kind of thinking there is some issue of doing or waiting for disk io in the rendering thread. In the application and probably also in the window manager, because applications that don't interact with io are often hanging/freezing.
Comment 5 Méven 2025-05-30 09:09:42 UTC
Using hotspot https://github.com/KDAB/hotspot on plasmashell process during the lag would be very helpful.

I suspect we do some synchronized IO on the busy device, when we can probably avoid it.
But spotting which part of plasmashell does it is not easy, using hotspot can identify it.

kwin might be concerned too, but you can't use hotspot on it, or you would break your input...
Comment 6 John Kizer 2025-06-13 16:59:17 UTC
Hi - I noticed in the System Monitor screenshot that there was a lot of network activity shown - does the interface lag you're describing occur when copying between two standard user folders on the same local, physical device, or does a remote/network device need to be involved somehow?

Thanks!
Comment 7 Moritz 2025-06-15 16:52:12 UTC
I don't think so (the attachment is from 2023).
Comment 8 Moritz 2025-06-15 16:52:59 UTC
*** Bug 410760 has been marked as a duplicate of this bug. ***
Comment 9 Nate Graham 2025-08-13 22:44:52 UTC

*** This bug has been marked as a duplicate of bug 469753 ***