Hi SLeo, You won't better performance with that: all compute units are dedicated to OpenCL computation in this demo... And the demo is using CL+GL interoperability: the OpenCL kernel is writting in an image that is in fact an OpenGL texture. (the one displayed on the screen)
but it don't use OpenCL + OpenGL interoperability, that's why it is slow
ReplyDeleteHi SLeo,
ReplyDeleteYou won't better performance with that: all compute units are dedicated to OpenCL computation in this demo...
And the demo is using CL+GL interoperability: the OpenCL kernel is writting in an image that is in fact an OpenGL texture. (the one displayed on the screen)