Unfortunately not.
In the early stages of development, we had to choose between CUDA and OpenCL, since doing R&D plus maintaining two rendering cores was out of the question. We opted for CUDA back then and that is the path we have followed ever since.