Is it possible to set the size of the CUDA block from the underlying kernel launch of Thrust function templates such as thrust::transform or thrust::sort<
thrust::transform
thrust::sort<