[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
(mgp-users 00128) Re: embarassingly slow tfont handling...
- To: mgp-users@mew.org
- Subject: (mgp-users 00128) Re: embarassingly slow tfont handling...
- From: Luigi Rizzo <luigi@labinfo.iet.unipi.it>
- Date: Tue, 23 Mar 1999 13:39:22 +0900
- In-reply-to: <> from "Atsushi Onoe" at Mar 23, 99 10:00:04 am
- Posted: Tue, 23 Mar 1999 03:24:22 +0100 (MET)
- Reply-to: mgp-users@mew.org
> > On a PPro200, a 6-line paragraph on 640x480 might take some
> > 2-3 seconds to render. What is worse, rendering time becomes
> > worse and worse as time goes by, e.g. you switch back and forth
> > on the same pages and it takes longer and longer.
>
> I cannot reproduce the problem (takes longer and longer) on current snap,
> though it is slow.
i was seeing this on:
mgp-snap-981221 + local mods (unrelated to rendering or X in general)
freetype-1.1
FreeBSD 2.2.6 and 3.1 (don't think it really matters)
XFree86 3.2 (on 226) and 3.3something (on 3.1)
1024x768 16bpp truecolor w/ S3 and SVGA servers
test document: something with the following background
%default 1 center, fore "black", back "white", bimage "power1.gif", vgap 50, font "rm3"
(where power1.gif is basically the "powerlogo", a 171x64 image
with the freebsd daemon)
the font definition is
%deffont "rm3" xfont "helvetica-medium-r", tfont "ARIAL.TTF", fore "white" size 4
the offending text is typically a long paragraph like
%tab i2 font "rm2", prefix " ", icon arc "red" 50
&i2 bla bla bla this and that and more and again \
bla bla bla this and that and more and again \
bla bla bla this and that and more and again \
bla bla bla this and that and more and again \
bla bla bla this and that and more and again \
bla bla bla this and that and more and again
and to really see that it takes longer and longer it may help to put
a couple of gettimeofday() around the while() loop in
draw.c::draw_string()
putting a
if (back == -1) back = 0 ;
near the beginning of tfont.c::tfc_image() makes thing faster.
The two most evident differences between the two test environment are
i) the use of a true gif image rather than a 'bgrad one' and ii) the X
server.
I think that the fault is somehow in the X server, maybe the gif image
helps trashing its behaviour.
cheers
luigi
-----------------------------------+-------------------------------------
Luigi RIZZO .
EMAIL: luigi@iet.unipi.it . Dip. di Ing. dell'Informazione
HTTP://www.iet.unipi.it/~luigi/ . Universita` di Pisa
TEL/FAX: +39-050-568.533/522 . via Diotisalvi 2, 56126 PISA (Italy)
-----------------------------------+-------------------------------------