This operation is accelerated on systems supporting AMD's SSE 4A. This is a limited subset of SSE4 with their own operations.