Re: [Gnash-dev] opengl->agg?

[Udo Giacomozzi]

Hello Martin,

Tuesday, May 8, 2007, 1:35:27 PM, you wrote:
MG> And it'll encourage people to make the AGG rendering faster :)

Making AGG faster is not an easy task. It already uses the fastest
known methods (to me) to render the scene. What could be improved are
those vector calculations. That includes world matrix transformations
and global scaling.

Global scaling could be merged with apply_matrix_to_path(), that is
scaling the matrix itself before applying it. This would avoid two
float multiplications per anchor.

Another way would be to cache already transformed points for later use
(improvement for graphic objects that don't change).

However, both of them will only improve performance for very complex
(lots of lines/curves) graphics and probably not by a huge degree
(except if you have no MMU).

Equally difficult would be to automatically create bitmap caches of
already rendered characters.


I see more potential in improving invalidated bounds calculation since
AGG currently is very busy doing useless work (see gotoAndPlay thread).


Idp