Imagination PowerVR SDK Blog

Simulating SVGA vs glDrawTex


I have some legacy code running on a OMAP35x EVM which can me mostly tricked into SVGA mode by calling glOrthof() with a 600,800 resolution, though of course the eglQuerySurface() will return 480,640.  I say “mostly” since glDrawTexiOES() pixmaps are not affected by scale, etc.

Are there any other tricks which might help with scaling pixmaps as well?  I’m investigating just drawing texture on a rectangle, but I assume that glDrawTex() is faster.

BTW, I’ll be using the 3530 TV out for SVGA, not the LCD, and am wondering if the PVR drivers care or would return different surface dimensions based on video mode, or if SGX always returns VGA.



Its your framebuffer driver's properties that the SGX queries and uses for its output. For example, we routinely validate the SGX output on DVI and the LCD, on the EVM.

What would be your usecase ?




There’s a video DAC on the DSS bus on our hardware which gives analog video out.  It sounds like it will be up to the video driver to publish the right resolution, and I assume that the GLES surface will have the right width, height.