vp8/common/arm/neon/reconb_neon.asm

   1 ;
   2 ;  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
   3 ;
   4 ;  Use of this source code is governed by a BSD-style license
   5 ;  that can be found in the LICENSE file in the root of the source
   6 ;  tree. An additional intellectual property rights grant can be found
   7 ;  in the file PATENTS.  All contributing project authors may
   8 ;  be found in the AUTHORS file in the root of the source tree.
   9 ;
  10
  11
  12     EXPORT  |vp8_recon_b_neon|
  13     ARM
  14     REQUIRE8
  15     PRESERVE8
  16
  17     AREA ||.text||, CODE, READONLY, ALIGN=2
  18
  19 ; r0    unsigned char  *pred_ptr,
  20 ; r1    short *diff_ptr,
  21 ; r2    unsigned char *dst_ptr,
  22 ; r3    int stride
  23
  24 |vp8_recon_b_neon| PROC
  25     mov             r12, #16
  26
  27     vld1.u8         {d28}, [r0], r12    ;load 4 data/line from pred_ptr
  28     vld1.16         {q10, q11}, [r1]!   ;load data from diff_ptr
  29     vld1.u8         {d29}, [r0], r12
  30     vld1.16         {q11, q12}, [r1]!
  31     vld1.u8         {d30}, [r0], r12
  32     vld1.16         {q12, q13}, [r1]!
  33     vld1.u8         {d31}, [r0], r12
  34     vld1.16         {q13}, [r1]
  35
  36     vmovl.u8        q0, d28             ;modify Pred data from 8 bits to 16 bits
  37     vmovl.u8        q1, d29             ;Pred data in d0, d2, d4, d6
  38     vmovl.u8        q2, d30
  39     vmovl.u8        q3, d31
  40
  41     vadd.s16        d0, d0, d20         ;add Diff data and Pred data together
  42     vadd.s16        d2, d2, d22
  43     vadd.s16        d4, d4, d24
  44     vadd.s16        d6, d6, d26
  45
  46     vqmovun.s16     d0, q0              ;CLAMP() saturation
  47     vqmovun.s16     d1, q1
  48     vqmovun.s16     d2, q2
  49     vqmovun.s16     d3, q3
  50     add             r1, r2, r3
  51
  52     vst1.32         {d0[0]}, [r2]       ;store result
  53     vst1.32         {d1[0]}, [r1], r3
  54     add             r2, r1, r3
  55     vst1.32         {d2[0]}, [r1]
  56     vst1.32         {d3[0]}, [r2], r3
  57
  58     bx             lr
  59
  60     ENDP
  61     END