OpenGl Transformations
Author
I was wondering if there is anyone out there who''s built a high poly game and who uses OpenGls matrix stack operations to do all the vertex transforms. Are these Operations fast enough? Or should I use my own routines?
Hi REF_Cracker,
I''ve heard many conflicting views on this topic, and to tell the truth, I don''t really know who to believe.
Some FAQ''s say "make your own matrices and use glMultMatrix", and some (notably, some I found linked from opengl.org) say "use the OpenGL matrix operation functions rather than constructing your own and using glMultMatrix". I tend to think using the functions provided is best (I could be wrong)... I mean, hey, someone has been paid to optimise the damn things, so why should I try and rewrite them? (and cut myself out of any available hardware T and L in the process)
Anyho, I use the functions provided by OpenGL, and they''ve worked just dandy fast enough for me. If there''s someone else who knows for sure what''s better, feel free to denounce this as a load of bollocks
-------------
squirrels are a remarkable source of protein...
I''ve heard many conflicting views on this topic, and to tell the truth, I don''t really know who to believe.
Some FAQ''s say "make your own matrices and use glMultMatrix", and some (notably, some I found linked from opengl.org) say "use the OpenGL matrix operation functions rather than constructing your own and using glMultMatrix". I tend to think using the functions provided is best (I could be wrong)... I mean, hey, someone has been paid to optimise the damn things, so why should I try and rewrite them? (and cut myself out of any available hardware T and L in the process)
Anyho, I use the functions provided by OpenGL, and they''ve worked just dandy fast enough for me. If there''s someone else who knows for sure what''s better, feel free to denounce this as a load of bollocks
-------------
squirrels are a remarkable source of protein...
From my limited knowledge..
the OpenGL transformations are fast unless you need to get the values from the matrix ie. glTranslate, glRotate, as they have to be retrieve from hardware..(Goes through the system bus and all the hardware bottlenecks)
in which case it is better to do your own matrix class and calculate..
the OpenGL transformations are fast unless you need to get the values from the matrix ie. glTranslate, glRotate, as they have to be retrieve from hardware..(Goes through the system bus and all the hardware bottlenecks)
in which case it is better to do your own matrix class and calculate..
Author
Seems a though people are finding the openGl implementation of transformations good enough. Anyone else on the subject?
There''s absolutely nothing wrong with OpenGL transformations, if you use them only for what they are for.
However, if you need intermediate results, or have to do funky shearing stuff, or whatever, you might be better off writing your own functions.
( Hierarchical systems come to mind )
#pragma DWIM // Do What I Mean!
~ Mad Keith ~
**I use Software Mode**
However, if you need intermediate results, or have to do funky shearing stuff, or whatever, you might be better off writing your own functions.
( Hierarchical systems come to mind )
#pragma DWIM // Do What I Mean!
~ Mad Keith ~
**I use Software Mode**
Author
Good this makes me happy to hear, because I was worried about having to design a program to use both my own functions and those that would use transformation acceleration if present.
This topic is closed to new replies.
Advertisement
Popular Topics
Advertisement
