While creating a buffer in the host code, ensure the memory pointer is page-aligned.
Runtime (XRT) prefers
page-aligned memory for efficient data transfer to and from the FPGA. If the buffer pointer
is not aligned to a page boundary, XRT performs extra
memcpy to make it aligned. This extra
operation can be avoided if host code memory is aligned to page boundary.
and so on to align the
cl_buffer host pointer to the page