Summary: | Wallpaper slideshow - changing wallpaper increases cpu usage | ||
---|---|---|---|
Product: | [Unmaintained] plasma4 | Reporter: | Mohd Asif Ali Rizwaan <maarizwan> |
Component: | containment-desktop | Assignee: | Plasma Bugs List <plasma-bugs> |
Status: | RESOLVED DUPLICATE | ||
Severity: | normal | CC: | jtamate, mmtsales, pranavg.india |
Priority: | NOR | ||
Version: | unspecified | ||
Target Milestone: | --- | ||
Platform: | Arch Linux | ||
OS: | Linux | ||
Latest Commit: | Version Fixed In: | ||
Sentry Crash Report: |
Description
Mohd Asif Ali Rizwaan
2011-06-20 12:22:04 UTC
I've run callgrind to plasma-desktop when this wallpaper presentation is enabled. Most of the time is spent in Qt resizing the image to the current screen size. (qSmoothScaleImage). The second CPU hungry method is to read the jpeg and transform it to a Qt image. (QImage::ConverToFormat). Therefore, two solutions to this problem: Create a bug report to Qt about those methods being slow. Change the images to be shown to the size of your screen with an image manipulation program such as digikam, gimp,.... therefore, no need to scale the image. Another solution from KDE, to launch the thread that does the slide show in a thread started with QThread::LowestPriority priority (even if Qt documentation says it will not make any difference in Linux, someday it will). *** Bug 273184 has been marked as a duplicate of this bug. *** I've experienced this for a long time... The desktop even freezes for a brief while when the wallpaper is being changed. It seems to me that the high CPU usage and the little freezes are due to the fading in/out effect when switching images. Maybe if the new image would just replace the previous one, without the transition effect, this would solve the problem. Of course it would be nicer to keep the soft transition between images, it is beautiful this way, but if it is difficult to solve this issue while keeping the fading in/out, I think it's better to remove it. Maybe this is a duplicate of bug #209152? (In reply to comment #5) > Maybe this is a duplicate of bug #209152? Well spotted, thank you. *** This bug has been marked as a duplicate of bug 209152 *** |