The CUDALink documentation suggests that it does not support the NVidia Maxwell architecture. Specifically, I am interested in the GTX 980 Ti GPU. If CUDALink currently does not support this, are there plans in the future to support Maxwell and the upcoming Pascal architectures?