I'm building a large (200x200 or larger) heightfield mesh. While parsing,
the render progress dialog shows the token count, then stops (Mac OS
version - the Windows version shows this info in the status bar at the
bottom). After a long pause, often longer than the initial verbose stage,
the render window is opened and rendering commences. I'm guessing that the
parser is going through the actual SDL during the verbose stage, and then
doing some kind of internal reorganisation during the pause. Can anyone
confirm this for me, or explain the actual processes?

What I want to do is to speed up the parsing (rendering is pretty quick),
and I have the distinct feeling that making my SDL more efficient will only
speed up the initial verbose stage.

Bill

Post a reply to this message

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Bill Pragnell wrote:

> I'm guessing that the
> parser is going through the actual SDL during the verbose stage, and then
> doing some kind of internal reorganisation during the pause. Can anyone
> confirm this for me, or explain the actual processes?
> 
	If that's what's happening, you should get faster results with a
mesh2 instead of a mesh. The reason being that POV uses the same
structure as the mesh2 for its internal representation, so there
should be less reorganizing needed...

		Jerome
- --
+------------------------- Jerome M. BERGER ---------------------+
|    mailto:jeb### [at] freefr      | ICQ:    238062172            |
|    http://jeberger.free.fr/     | Jabber: jeb### [at] jabberfr   |
+---------------------------------+------------------------------+
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2.2 (GNU/Linux)

iD8DBQFEcfP1d0kWM4JG3k8RAvp1AJ9W7zHim2jsOvbsp00xK8Uaan4amQCeNz+v
QWEjn1oaXupds+swmg8Z/Mo=
=0OcP
-----END PGP SIGNATURE-----

Post a reply to this message

=?ISO-8859-1?Q?=22J=E9r=F4me_M=2E_Berger=22?= <jeb### [at] freefr> wrote:
>  If that's what's happening, you should get faster results with a
> mesh2 instead of a mesh. The reason being that POV uses the same
> structure as the mesh2 for its internal representation, so there
> should be less reorganizing needed...

I've now implemented an optimized mesh2 instead of a mesh. An interesting
and highly useful exercise, although mesh2 is a little strange if you're
used to mesh. But it's well worth the effort - the speed increase in
parsing is about a factor of 10 for my scene, and it appears to render
faster, too.

Example pic and code to follow in p.b.i shortly...

Bill

Post a reply to this message