On Thu, Apr 9, 2009 at 1:07 PM, Callum Lerwick <seg@xxxxxxxxxx> wrote:
On Thu, 2009-04-09 at 14:40 +0200, Rudolf Kastl wrote:Yes, near as I can tell, Cg is really primarily used as a code
> its like that since ages but actually the generated glsl code is not
> really pretty in alot test cases i did.
obfuscation method. Something that is fundamentally at odds with open
source. So I have no idea why any self respecting open source project
would even want to touch it.
Also, its rumored that it favors nvidia. Nvidia's drivers compile Cg to
native Nvidia binary code, and output sub-par GLSL/HLSL on everything
else.
Crystal Space uses Cg. WHY? They don't even use Direct3D. The only
justification for Cg instead of GLSL is if you're wanting to share
shaders across OpenGL and Direct3D, in which case writing in HLSL and
using ATI's tool is a non-proprietary solution. (Now how about a
GLSL2HLSL compiler...)
--
fedora-devel-list mailing list
fedora-devel-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/fedora-devel-list
Maybe because it makes it easy to make shaders to both GLSL and HLSL? For open source projects like PCSX2 which need shaders a lot, it is absolutely necessary that the SL output is equivalent on OGL and D3D.
-- fedora-devel-list mailing list fedora-devel-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/fedora-devel-list
