GSLANG 1.1 compliance of catalyst 4.7?

[relmas]

the title says it all?

as usual, there is no precise info in the release notes. Has anyone experimented with the new built in variables?

Thank you.

[Sunray]

No I haven't, and point sprites still dosen't work... :

[Sunray]

Yay! 3DLabs GLSL Parse Test reports 99% score with Cat4.7 compared to 94% with Cat4.6.

http://sunray.cplusplus.se/dump/Cat4.7.html
http://sunray.cplusplus.se/dump/Cat4.4.html (same result as 4.6)

[relmas]

so what's that one percent failure due to? I couldn't figure out the content of the vertex shader that fails the test.is it a specific instruction or a more complex construct?

[Corrail]

glGetString(GL_SHADING_LANGUAGE_VERSION_ARB) only returns "1.00", so there's no full GLSL 1.1 support. But some features seems to be implemented, like GL_SAMPLER_2D types when using glGetActiveUniform and gl_FragData.

[Corrail]

I just tested the Cat 4.7 with the new GLSL Parser and got only 97%:
http://mitarbeiter.kjk.cc/rudolf/catalyst_4_7.html

[eyebex]

@corrail: it seems you're using a newer version of the glsl parser test program (1.4 is current) which includes some new "problematic" shaders. i also get 97% with this version on my ati 9800 pro running catalyst 4.7 drivers.

edit: to be more specific, the compiler in this driver maily fails with glsl code following the new 1.10 specs, so it's not really surprising.

[Corrail]

Yes, I tested this with GLSL Parser v1.4. I also took a look at the shaders which don't work on Catalyst 4.7. These shaders mainly use #extension and #version statements.

[azazello]

may be we could ask somebody to place this 3dlabs conformation test for different drivers and hardware (ie Cat & R3x0, Cat & R4x0, Det 6x & NV4x, etc )?

may be it can be Tom Nuydens on his Delphi3d.net resource or rigth here, in opengl.org?