you should go to gamedev.com because this forum is for pure opengl coding in blitzmax and glsl is another kind of language very direfent from plain opengl or blitz.
I'm introducing myself into glsl but i've found huge diferences in runtime from a graphic card to another (even with full opengl2.0 support), i've found diferent behaviours in diferent aplications too, wich makes me wonder if I will someday be able to compile something who will run somewhere else than in my computer...
I've found diferences from RenderDonkey to ShaderDesigner (in RD it simply don't do parallax mapping, since the same code in SD works well). It's quite strange.
Another thing I noticed is that Shader Designer's example shows a compile error in a GeForceFX (i don't remember the FX number sorry) since in my ATI radeon 9800 pro a large number of examples work but fractal kind don't.
In RenderDonkey the same code compile well in my ATI but the effect don't work at all (parallax mapping).
I haven't coded shaders inside BMX right now, and i'm finding my way into doing it but i have noticed one important thing: it seems it's not the same using ARB extensions than using actual OpengGL2.0 commands instead.
It shouldn't be this way because the modif between the ARB and the final commands are very litle, however, it seems that the drivers of the graphic cards do have great diferences from ARB to final ones...
This kind of things makes one turning mad because you actually don't know if it doesn't compile because of a coding error or because of an implementation deficience...
(just this line makes my fragment shader turn absolutly mad: float A=gl_FragCoord.x)
However, ATI drivers seems to be more tolerant than geforce ones.
Good luck
I'm introducing myself into glsl but i've found huge diferences in runtime from a graphic card to another (even with full opengl2.0 support), i've found diferent behaviours in diferent aplications too, wich makes me wonder if I will someday be able to compile something who will run somewhere else than in my computer...
I've found diferences from RenderDonkey to ShaderDesigner (in RD it simply don't do parallax mapping, since the same code in SD works well). It's quite strange.
Another thing I noticed is that Shader Designer's example shows a compile error in a GeForceFX (i don't remember the FX number sorry) since in my ATI radeon 9800 pro a large number of examples work but fractal kind don't.
In RenderDonkey the same code compile well in my ATI but the effect don't work at all (parallax mapping).
I haven't coded shaders inside BMX right now, and i'm finding my way into doing it but i have noticed one important thing: it seems it's not the same using ARB extensions than using actual OpengGL2.0 commands instead.
It shouldn't be this way because the modif between the ARB and the final commands are very litle, however, it seems that the drivers of the graphic cards do have great diferences from ARB to final ones...
This kind of things makes one turning mad because you actually don't know if it doesn't compile because of a coding error or because of an implementation deficience...
(just this line makes my fragment shader turn absolutly mad: float A=gl_FragCoord.x)
However, ATI drivers seems to be more tolerant than geforce ones.
Good luck