)]}'
{
  "commit": "511d218c60b9b6c1ab9383db746815e907af0359",
  "tree": "a7cbf64477adac2433384293d88d08f27c373fec",
  "parents": [
    "a622ed554f7072268e4c8d0b8f26d2e8865c2b3b"
  ],
  "author": {
    "name": "levytamar82",
    "email": "levytamar82@gmail.com",
    "time": "Thu Nov 21 15:49:29 2013 -0700"
  },
  "committer": {
    "name": "levytamar82",
    "email": "levytamar82@gmail.com",
    "time": "Thu Jan 09 12:27:51 2014 -0700"
  },
  "message": "SSSE3 convolution optimization\n\nOptimizing all SSSE3 assembly for convolution:\n1. vp9_filter_block1d4_h8_sse2\n2. vp9_filter_block1d8_h8_sse2\n3. vp9_filter_block1d16_h8_sse2\n4. vp9_filter_block1d4_v8_sse2\n5. vp9_filter_block1d8_v8_sse2\n6. vp9_filter_block1d16_v8_sse2\nmy optimization include:\n-processing 2x8 elements in one 128 bit register instead of processing\n8 elements in one 128 bit register.\n-removing unecessary loads.\nThis optimization gives between 2.4% user level gain for 480p input\nand 1.6% user level gain for 720p.\nThis Optimization done only for 64bit.\n\nChange-Id: Icb586dc0c938b56699864fcee6c52fd43b36b969\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "106e6d426f7d1469362930caf018b93db307e60c",
      "old_mode": 33188,
      "old_path": "vp9/common/x86/vp9_asm_stubs.c",
      "new_id": "b4ab4c3061cf8c35c1bf76a6640301c8d0854f3a",
      "new_mode": 33188,
      "new_path": "vp9/common/x86/vp9_asm_stubs.c"
    },
    {
      "type": "add",
      "old_id": "0000000000000000000000000000000000000000",
      "old_mode": 0,
      "old_path": "/dev/null",
      "new_id": "303fced3ba3763b1067ffa60cdf7e641ad9e4847",
      "new_mode": 33188,
      "new_path": "vp9/common/x86/vp9_subpixel_8t_intrin_ssse3.c"
    },
    {
      "type": "modify",
      "old_id": "eefbd1ac9ae4e2d5cc42a9935dd02bb61b027d19",
      "old_mode": 33188,
      "old_path": "vp9/vp9_common.mk",
      "new_id": "8a8b63d97913d9933ca3f5686dc29e19b6d189c4",
      "new_mode": 33188,
      "new_path": "vp9/vp9_common.mk"
    }
  ]
}