Hi,
Once on a beatiful day Saturday 18 May 2002 02:45, Rolf Magnus wrote:
> I compared compilation times between gcc 2.95 and some gcc3 HEAD versio=
ns
> several times, and now I used the 3.1 release version to find out if my
> results are valid for it, too. Each time, compilation times with 3.1/HE=
AD
> were about twice as long as with 2.95, and I don't see any numbers at a=
ll
> from you. Still you call my numbers "total bogus", and I have to say I
> don't believe you. I did only try qt-copy right now, but I don't see wh=
y
> this should be different for kdebase. But if your compilation times are
> really equal for 2.95 and 3.1, then I'd like to find out why they aren'=
t
> for me. Did you do anything special?
I did not do anything special, because I have no idea how to. I am using =
a cvs=20
snapshot of gcc 3.1 from 20020423, but it is not possible, that such a bi=
g=20
change has been ade between this version and gcc 3.1 is it? Well, as for=20
compilation time I have not noticed any differences in time between my ol=
d=20
debian with 2.95 and my pld with 3.1. kdebase compiles in 3-4 hrs on both=
=20
systems and it does not differ that much, 1h at best, but I have found th=
at=20
apps compiled with gcc3 start faster. I apologize for using the "total bo=
gus"=20
(it sounds offensive and I surely did not want to offend you) phrase, wha=
t I=20
ment was not the numbers (numbers do not lie), I used it in reply to:
-----
Once on a beatiful day Friday 17 May 2002 17:40, Rolf Magnus wrote:
> Conlcusion: gcc 3.1 is slow, even much slower than 2.95.3 :(, and app
> startup times will go up again.=20
-----
What I wanted to say, is that it is not true, that the app startup times =
will=20
rise again. Because they do not., but maybe it is because I am using=20
objprelink.
--=20
!: Piotr Szymanski | LinuxPL Dev. Team Member | KDE i18n-pl coordinator
@: djurban@linuxpl.org | The website is coming=20
#: GG: 2300264 | ICQ: 12622400=20
%: "bo jesli ktos sie modli Pan Bog w nim oddycha" - ks. Jan Twardowski.