Re: GtkPerf data for GTK+ 2.6 vs GTK+ 2.10 vs Maemo-GTK+ 2.6 (on a Nokia 770)

["Kalle Vahlman" <kalle vahlman gmail com>]

2006/7/21, Tommi Komulainen <tommi komulainen nokia com>:
> On Fri, 2006-07-21 at 00:26 +0300, ext Kalle Vahlman wrote:
> > 3. Maemo-GTK+ 2.6
> > atk-1.10.3-0osso1
> > glib-2.8.6-1osso5
> > gtk+-2.6.10-1.osso2
> > pango-1.8.1-1osso10
>
> The gtk+ version pretty old, we're at 2.6.10-1.osso23 already. (Or was
> that a typo?)

It was a cut&paste from the dpkg -l output so yes, probably the '3'
got cut off..

> > I ran GtkPerf 0.4 first a couple of times with low iteration count for
> > each setup to get the libs to memory[1] and then ran it for 1000
> > iterations (should be enough for resolution, at least it took ages to
> > finish :). All tests used the default builtin theme of GTK+.
>
> I wonder how much the theme engine is affecting the performance here,
> and how useful is it to use only one (be it using cairo or or not) when
> looking for cairo hotspots.

Who said anything about cairo?-) I just wanted to see GTK+ vs GTK+ and
for that, using the builtin engine makes IMO sense.

> At least for Maemo using the builtin theme engine makes little sense as
> the theme engine is likely to remain bitmap based, and as such have
> pretty different impact from cairo.

And again, I'm not (only) looking at Cairo. And yes, the maemo-gtk is
of course not sensible _to use_ with the builtin theme, but shouldn't
it be faster than regular 2.6 anyway due to other optimizations?

Last night I didn't have time but now I'm running the test with
sapwood enabled. We'll see how it affects the numbers.

I'm planning on running it with other theme engines too of course.

And hints for some other feasible (gtk/pango/cairo/glib) performance
test progs would be appreciated, I know of the theme torturer thing
but is there others?

--
Kalle Vahlman, zuh iki fi
Powered by http://movial.fi
Interesting stuff at http://syslog.movial.fi