1. 08 Jan, 2018 1 commit
    • Kyle Siefring's avatar
      Use alignr in lbd convolve_2d/convolve_x AVX2 · 94dfe5aa
      Kyle Siefring authored
      Results measured on Haswell.
      
      For convolve_2d
      Width Height Improvement
          4      4        8.3%
          8      8        5.6%
         64     64         13%
          4     16         13%
         32      8          9%
      
      For convolve_x
      Width Height Improvement
          4      4          0%
          8      8          0%
         64     64         21%
          4     16          0%
         32      8         29%
      
      Change-Id: Ic13215d61f48400b3e31d5ffea3af11c0b3f0ed7
      94dfe5aa
  2. 23 Dec, 2017 1 commit
    • Yunqing Wang's avatar
      Add optimized convolve functions for single reference case · 94e3fe3b
      Yunqing Wang authored
      Added optimized convolve functions for single reference case, so that no
      separate post rounding is needed and the result is written to the
      destination buffer directly. Duplicate code will be cleaned up later.
      
      Change-Id: Iffc0cc6e135b8b6f45a95c314d63368f5aa35f34
      94e3fe3b
  3. 20 Dec, 2017 1 commit
    • Yunqing Wang's avatar
      Modify convolve function prototype · 940c22a2
      Yunqing Wang authored
      This patch modified convolve function prototype. The temp buffer used in
      convolve function is already available in conv_params, which doesn't need
      to be passed in. Instead, pass in the destination buffer so that the
      result can be written in the destination directly for single ref case.
      More patch will be followed.
      
      Change-Id: Ib28dc3ba5783a1034c70570d78fa8c8af7cbed7c
      940c22a2
  4. 13 Dec, 2017 1 commit
  5. 28 Nov, 2017 2 commits