Desktop Window Manager (Win10) vs KDE/GNOME

JazzSinatra

Member

Thanks: 1
Messages: 29

#1
Has anyone done performance comparison between Windows' window manager (Desktop Window Manager) and KDE and/or GNOME? With performance, I mean resource usage (RAM, CPU etc.). I would like to know how well Microsoft can compete with the open source community on this area.
 

SirDice

Administrator
Staff member
Administrator
Moderator

Thanks: 6,225
Messages: 27,200

#3
Windows 10 feels faster than KDE
It's probably highly depended on the graphics card driver. If you run KDE through software-rendering (no 2D/3D acceleration) everything will appear to be faster.
 

ShelLuser

Son of Beastie

Thanks: 1,377
Messages: 2,935

#4
I don't think you can really compare these two. Not in a fair manner anyway because there are so many other things going on 'under the hood'. A regular FreeBSD box for example has several processes running in the background, DBus and Hald being the most obvious, but the same can be said about Windows. It wouldn't even surprise me if Windows had more and more intense processes running than FreeBSD.

For what's it worth I'm under the impression that KDE4 and Windows 7 are pretty much on-par where responsiveness is concerned. But the fun fact is that I'm also comparing apples and oranges. My Windows 7 is a 32bit machine whereas my FreeBSD box is a 64bit server which also happens to run KDE from time to time (I'm usually not logged in).
 

SirDice

Administrator
Staff member
Administrator
Moderator

Thanks: 6,225
Messages: 27,200

#5
I'm really struggling to figure out a way to measure this in an objective way[*]. Open lots of windows and time that? How could one measure the performance of a window manager without, inadvertently, measuring the rest of the OS with it?

[*] It "feels" faster isn't objective.
 

kpedersen

Daemon

Thanks: 336
Messages: 1,151

#6
Perhaps a bandwidth comparison via VNC? Though I guess that is based on the assumption that the more pixels change on the screen, the slower the DE is.

Running Gnome in Cygwin is a little unfair because of some overhead but perhaps running this and comparing CPU usage over time moving some windows about vs Microsoft's DE could give some measurable data?
 
Top