On 28.03.2011 00:06, Antony Martin wrote: > Well then I'll try to know what the DR can do exactly in order to make a > reasonable timeline. Who could know that?
As a start, build CS and run “deferreddemo”. > I was thinking about it too. In my engine I tried to support both Cg and GLSL > but since such a feature wasn't the main purpose of my engine I gave up when > I > encoutered the first issue: Cg distinguishes vertex and pixel shaders (like > DX > afaik) and allow to switch only one of them It is right Cg distinguishes between VPs and FPs, but “only one of them” is definitely wrong - you can have a VP and an FP active at the same time. > whilst GLSL requires them to be > bound together into one "program". Interestingly, Cg also allows to link multiple programs. (Actually something necessary with the GLSL backend.) > I actually never tried to find out how much > expensive is the glLinkProgram() function, Once for each program, I'd guess. > or how often it is needed to switch > only one shader, maybe it is negligible. Also since I'm not familiar with CS > I > don't know how much time it would take to add support for GLSL shaders. CS currently also distinguishes between vertex and fragment programs. So in theory you could do stuff like having a Cg vertex program and a ARB fragment program, or a fixed function fragment setup. However, in practice that mixing isn't really used. Supporting GLSL with that “split” approach is significant hassle, so one of the first steps would be to change the shader subsystem to use “unified” shader programs (i.e. VP+FP wrapped into one ‘shader program’ object) - should actually be relatively easy. -f.r.
signature.asc
Description: OpenPGP digital signature
------------------------------------------------------------------------------ Enable your software for Intel(R) Active Management Technology to meet the growing manageability and security demands of your customers. Businesses are taking advantage of Intel(R) vPro (TM) technology - will your software be a part of the solution? Download the Intel(R) Manageability Checker today! http://p.sf.net/sfu/intel-dev2devmar
_______________________________________________ Crystal-main mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/crystal-main Unsubscribe: mailto:[email protected]?subject=unsubscribe
