vp8/encoder/arm/neon/sad8_neon.asm

   1 ;
   2 ;  Copyright (c) 2010 The VP8 project authors. All Rights Reserved.
   3 ;
   4 ;  Use of this source code is governed by a BSD-style license and patent
   5 ;  grant that can be found in the LICENSE file in the root of the source
   6 ;  tree. All contributing project authors may be found in the AUTHORS
   7 ;  file in the root of the source tree.
   8 ;
   9
  10
  11     EXPORT  |vp8_sad8x8_neon|
  12     EXPORT  |vp8_sad8x16_neon|
  13     EXPORT  |vp8_sad4x4_neon|
  14
  15     ARM
  16     REQUIRE8
  17     PRESERVE8
  18
  19     AREA ||.text||, CODE, READONLY, ALIGN=2
  20 ; unsigned int vp8_sad8x8_c(
  21 ;    unsigned char *src_ptr,
  22 ;    int  src_stride,
  23 ;    unsigned char *ref_ptr,
  24 ;    int  ref_stride)
  25
  26 |vp8_sad8x8_neon| PROC
  27     vld1.8          {d0}, [r0], r1
  28     vld1.8          {d8}, [r2], r3
  29
  30     vld1.8          {d2}, [r0], r1
  31     vld1.8          {d10}, [r2], r3
  32
  33     vabdl.u8        q12, d0, d8
  34
  35     vld1.8          {d4}, [r0], r1
  36     vld1.8          {d12}, [r2], r3
  37
  38     vabal.u8        q12, d2, d10
  39
  40     vld1.8          {d6}, [r0], r1
  41     vld1.8          {d14}, [r2], r3
  42
  43     vabal.u8        q12, d4, d12
  44
  45     vld1.8          {d0}, [r0], r1
  46     vld1.8          {d8}, [r2], r3
  47
  48     vabal.u8        q12, d6, d14
  49
  50     vld1.8          {d2}, [r0], r1
  51     vld1.8          {d10}, [r2], r3
  52
  53     vabal.u8        q12, d0, d8
  54
  55     vld1.8          {d4}, [r0], r1
  56     vld1.8          {d12}, [r2], r3
  57
  58     vabal.u8        q12, d2, d10
  59
  60     vld1.8          {d6}, [r0], r1
  61     vld1.8          {d14}, [r2], r3
  62
  63     vabal.u8        q12, d4, d12
  64     vabal.u8        q12, d6, d14
  65
  66     vpaddl.u16      q1, q12
  67     vpaddl.u32      q0, q1
  68     vadd.u32        d0, d0, d1
  69
  70     vmov.32         r0, d0[0]
  71
  72     bx              lr
  73
  74     ENDP
  75
  76 ;============================
  77 ;unsigned int vp8_sad8x16_c(
  78 ;    unsigned char *src_ptr,
  79 ;    int  src_stride,
  80 ;    unsigned char *ref_ptr,
  81 ;    int  ref_stride)
  82
  83 |vp8_sad8x16_neon| PROC
  84     vld1.8          {d0}, [r0], r1
  85     vld1.8          {d8}, [r2], r3
  86
  87     vld1.8          {d2}, [r0], r1
  88     vld1.8          {d10}, [r2], r3
  89
  90     vabdl.u8        q12, d0, d8
  91
  92     vld1.8          {d4}, [r0], r1
  93     vld1.8          {d12}, [r2], r3
  94
  95     vabal.u8        q12, d2, d10
  96
  97     vld1.8          {d6}, [r0], r1
  98     vld1.8          {d14}, [r2], r3
  99
 100     vabal.u8        q12, d4, d12
 101
 102     vld1.8          {d0}, [r0], r1
 103     vld1.8          {d8}, [r2], r3
 104
 105     vabal.u8        q12, d6, d14
 106
 107     vld1.8          {d2}, [r0], r1
 108     vld1.8          {d10}, [r2], r3
 109
 110     vabal.u8        q12, d0, d8
 111
 112     vld1.8          {d4}, [r0], r1
 113     vld1.8          {d12}, [r2], r3
 114
 115     vabal.u8        q12, d2, d10
 116
 117     vld1.8          {d6}, [r0], r1
 118     vld1.8          {d14}, [r2], r3
 119
 120     vabal.u8        q12, d4, d12
 121
 122     vld1.8          {d0}, [r0], r1
 123     vld1.8          {d8}, [r2], r3
 124
 125     vabal.u8        q12, d6, d14
 126
 127     vld1.8          {d2}, [r0], r1
 128     vld1.8          {d10}, [r2], r3
 129
 130     vabal.u8        q12, d0, d8
 131
 132     vld1.8          {d4}, [r0], r1
 133     vld1.8          {d12}, [r2], r3
 134
 135     vabal.u8        q12, d2, d10
 136
 137     vld1.8          {d6}, [r0], r1
 138     vld1.8          {d14}, [r2], r3
 139
 140     vabal.u8        q12, d4, d12
 141
 142     vld1.8          {d0}, [r0], r1
 143     vld1.8          {d8}, [r2], r3
 144
 145     vabal.u8        q12, d6, d14
 146
 147     vld1.8          {d2}, [r0], r1
 148     vld1.8          {d10}, [r2], r3
 149
 150     vabal.u8        q12, d0, d8
 151
 152     vld1.8          {d4}, [r0], r1
 153     vld1.8          {d12}, [r2], r3
 154
 155     vabal.u8        q12, d2, d10
 156
 157     vld1.8          {d6}, [r0], r1
 158     vld1.8          {d14}, [r2], r3
 159
 160     vabal.u8        q12, d4, d12
 161     vabal.u8        q12, d6, d14
 162
 163     vpaddl.u16      q1, q12
 164     vpaddl.u32      q0, q1
 165     vadd.u32        d0, d0, d1
 166
 167     vmov.32         r0, d0[0]
 168
 169     bx              lr
 170
 171     ENDP
 172
 173 ;===========================
 174 ;unsigned int vp8_sad4x4_c(
 175 ;    unsigned char *src_ptr,
 176 ;    int  src_stride,
 177 ;    unsigned char *ref_ptr,
 178 ;    int  ref_stride)
 179
 180 |vp8_sad4x4_neon| PROC
 181     vld1.8          {d0}, [r0], r1
 182     vld1.8          {d8}, [r2], r3
 183
 184     vld1.8          {d2}, [r0], r1
 185     vld1.8          {d10}, [r2], r3
 186
 187     vabdl.u8        q12, d0, d8
 188
 189     vld1.8          {d4}, [r0], r1
 190     vld1.8          {d12}, [r2], r3
 191
 192     vabal.u8        q12, d2, d10
 193
 194     vld1.8          {d6}, [r0], r1
 195     vld1.8          {d14}, [r2], r3
 196
 197     vabal.u8        q12, d4, d12
 198     vabal.u8        q12, d6, d14
 199
 200     vpaddl.u16      d1, d24
 201     vpaddl.u32      d0, d1
 202     vmov.32         r0, d0[0]
 203
 204     bx              lr
 205
 206     ENDP
 207
 208     END