When I want to copy data from my host to the GPU, what is the difference between the two approaches:
1. clEnqueueWriteBuffer
2. clCreateBuffer with CL_MEM_COPY_HOST_PTR and non-null 3rd parameter (0th based index)