parallel processing - Is the distinction between vertex and pixel shader necessary or even beneficial? -
from i've been able get, both vertex , pixel shader operations boil down passing data , doing lot of same every available unit. surely, vertex , pixel shaders in different parts of classical graphical pipeline, wouldn't better have more abstraction , able create arbitrary order of arbitrary data , parallel operations it? guess such abstraction can applied emulate opencl, compute shaders , whatever general or specialized compute api.
specialization helps drivers perform optimally , simplifies application code. pixel-shaders occur after rasterization. it's great not have worry rasterizing. use cuda or opencl in graphics agnostic way. alt. yes. it's coming.
Comments
Post a Comment