Page Migration on OpenCL

Hi, does page migration (from Pascal) work in OpenCL, or only with CUDA? It is mentioned here that you need to use cudaMallocManaged() to get unified memory. Is there an OpenCL equivalent?