Jingning Han 82d504b50f Use aligned buffer operations in 8x8/16x16 2D-DCT
This reduces 16x16 2D-DCT runtime from 865 cycles to 837 cycles.

Change-Id: I137758b81cd127b936175284310e81378db64552
2013-06-24 19:56:23 -07:00
..