[x265] [arm64] port avg_pp

[chen]

Looks good




2021-07-24 07:04:03，"Pop, Sebastian" <spop at amazon.com> 

Hi,

 

the attached patch ports to arm64 the following kernels:

 

 

         avg_pp[  4x4]  8.50x    8.85            75.21

avg_pp_aligned[  4x4]  8.49x    8.89            75.46

         avg_pp[  8x8]  29.12x   11.61           338.01

avg_pp_aligned[  8x8]  30.20x   11.42           344.78

         avg_pp[  8x4]  27.12x   7.34            199.01

avg_pp_aligned[  8x4]  27.18x   7.40            201.11

         avg_pp[  4x8]  9.63x    14.89           143.37

avg_pp_aligned[  4x8]  10.65x   14.94           159.20

         avg_pp[16x16]  50.41x   22.63           1140.85

avg_pp_aligned[16x16]  49.74x   22.45           1116.51

         avg_pp[ 16x8]  66.87x   11.27           753.83

avg_pp_aligned[ 16x8]  68.10x   11.16           759.76

         avg_pp[ 8x16]  25.07x   22.85           572.83

avg_pp_aligned[ 8x16]  24.71x   22.69           560.73

         avg_pp[ 16x4]  41.45x   7.34            304.42

avg_pp_aligned[ 16x4]  48.04x   7.43            356.89

         avg_pp[16x12]  63.50x   16.99           1078.53

avg_pp_aligned[16x12]  45.91x   16.87           774.56

         avg_pp[ 4x16]  10.80x   26.74           288.84

avg_pp_aligned[ 4x16]  10.90x   26.69           290.97

         avg_pp[12x16]  30.99x   31.28           969.46

avg_pp_aligned[12x16]  26.61x   31.61           841.17

         avg_pp[32x32]  92.92x   55.84           5189.14

avg_pp_aligned[32x32]  71.96x   55.72           4009.62

         avg_pp[32x16]  93.70x   28.91           2709.20

avg_pp_aligned[32x16]  68.55x   29.06           1992.17

         avg_pp[16x32]  65.12x   45.81           2983.30

avg_pp_aligned[16x32]  51.43x   45.51           2340.67

         avg_pp[ 32x8]  93.24x   15.82           1475.04

avg_pp_aligned[ 32x8]  76.66x   15.88           1217.75

         avg_pp[32x24]  70.85x   42.36           3001.17

avg_pp_aligned[32x24]  70.10x   42.46           2976.72

         avg_pp[ 8x32]  31.19x   45.98           1434.10

avg_pp_aligned[ 8x32]  27.80x   45.73           1271.58

         avg_pp[24x32]  50.96x   75.62           3853.13

avg_pp_aligned[24x32]  50.17x   75.71           3798.44

         avg_pp[64x64]  74.94x   221.76          16617.97

avg_pp_aligned[64x64]  71.24x   221.74          15797.84

         avg_pp[64x32]  82.22x   112.25          9229.40

avg_pp_aligned[64x32]  70.60x   112.25          7925.30

         avg_pp[32x64]  79.00x   110.78          8751.21

avg_pp_aligned[32x64]  71.68x   110.70          7934.54

         avg_pp[64x16]  87.17x   57.66           5026.56

avg_pp_aligned[64x16]  68.42x   57.66           3945.34

         avg_pp[64x48]  87.96x   166.85          14676.53

avg_pp_aligned[64x48]  71.82x   166.86          11983.28

         avg_pp[16x64]  48.84x   92.63           4523.80

avg_pp_aligned[16x64]  43.73x   92.32           4037.08

         avg_pp[48x64]  96.16x   143.53          13801.49

avg_pp_aligned[48x64]  83.02x   143.73          11932.26

 

Ok to commit?

 

Thanks,

Sebastian

 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.videolan.org/pipermail/x265-devel/attachments/20210724/25539f2e/attachment-0001.html>