I guess CUDA got a lot more traction and there isn't much of a software base written for OpenCL. Kind of what happened with Unix and Windows - You could write code for Unix and it'd (compile and) run on 20 different OSs, or write it for Windows, and it'd run on one second-tier OS that managed to capture almost all of the desktop market.
I remember Apple did support OpenCL a long time ago, but I don't think they still do.
I guess CUDA got a lot more traction and there isn't much of a software base written for OpenCL. Kind of what happened with Unix and Windows - You could write code for Unix and it'd (compile and) run on 20 different OSs, or write it for Windows, and it'd run on one second-tier OS that managed to capture almost all of the desktop market.
I remember Apple did support OpenCL a long time ago, but I don't think they still do.