From kde-core-devel Thu Mar 28 20:20:32 2002 From: Christian Esken Date: Thu, 28 Mar 2002 20:20:32 +0000 To: kde-core-devel Subject: Re: Comparing Startup speed of KDE2 to KDE3 X-MARC-Message: https://marc.info/?l=kde-core-devel&m=101734741428757 On Thursday 28 March 2002 18:34, David Faure wrote: > On Thursday 28 March 2002 13:52, Christian Esken wrote: > > The following table shows the results: "max" ist the highest start > > time measured. "avg" is the average of the 4th and 5th run. > > > > =09KDE2=09KDE3 > > max=0931s=09=0916s > > avg=0916s=09=0910s > > Yes!!! This is very cool. All those people complaining about KDE being = too > slow will be happy with KDE3. Yes!!! But actually I was very frustrated for some time as well, as logging in c= ould=20 need up to 3 Minutes, until I found the trick. See the end of the mail fo= r=20 more info. > > The test was probably not 100% fair, as under KDE2 I have more > > applications installed: > > > > chris@donald:~> ls -1 /opt/kde2/bin/ | wc > > 281 281 3038 > > chris@donald:~> ls -1 /opt/kdehead/bin/ | wc > > 183 183 1993 > > > > So I recalculated the KDE3 measurements, using (281/183 =3D1.53552) a= s > > linear correcting factor (please be aware: it is just a "intelligent > > guess" to correct the measurement like this): > > I don't understand. Why does having a different number of applications > matter? Startup time is only about the applications being started, not > about those installed on the harddisk and unused.... There *could* be numerous reasons. I am not into the design of all=20 applications, but here are some (hopefully) reasonable arguments: Probably kicker startup time (due to more *.desktop files =3D> bigger men= u=20 structure), basic startup times because more Mimetypes are there, and=20 whatsoever. But as I said: I just tried an "intelligent guess" to be a bit more "mean= " to=20 KDE3 - but it did not help: KDE3 is still faster. BTW: Did I mention until 3 weeks ago I had *HORRIBLE* KDE2 startup times. The record was something like 3 Minutes from clicking "Login" until I cou= ld=20 use the KDE Desktop. Usual startup time was 40-60 seconds. It was the same version (~ 2.2.1) as I did the tests with, but with my re= gular=20 user. The trick was to do a *MAJOR* cleanup up ~/.kde/config : I threw away ne= arly=20 everything below ~/.kde . Exceptions were those obviously "good" files li= ke ~/.kde/share/apps/konqueror/bookmarks.xml ,=20 ~/.kde/share/apps/kmail/addressbook , some config files (e.g. kmail) and = some=20 more. After this cleanup startup time was reduced to 20-30 seconds. Christian -- Is Unix ready for the desktop? See http://www.kde.org The Christian Esken |/ Desktop KDE Developer |\ Environment esken@kde.org KDE - The net transparent free Unix Desktop for everyone