1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
|
;
; Copyright (c) 2013 The WebM project authors. All Rights Reserved.
;
; Use of this source code is governed by a BSD-style license and patent
; grant that can be found in the LICENSE file in the root of the source
; tree. All contributing project authors may be found in the AUTHORS
; file in the root of the source tree.
;
EXPORT |vp9_idct8x8_1_add_neon|
ARM
REQUIRE8
PRESERVE8
AREA ||.text||, CODE, READONLY, ALIGN=2
;void vp9_idct8x8_1_add_neon(int16_t *input, uint8_t *dest,
; int dest_stride)
;
; r0 int16_t input
; r1 uint8_t *dest
; r2 int dest_stride)
|vp9_idct8x8_1_add_neon| PROC
ldrsh r0, [r0]
; generate cospi_16_64 = 11585
mov r12, #0x2d00
add r12, #0x41
; out = dct_const_round_shift(input[0] * cospi_16_64)
mul r0, r0, r12 ; input[0] * cospi_16_64
add r0, r0, #0x2000 ; +(1 << ((DCT_CONST_BITS) - 1))
asr r0, r0, #14 ; >> DCT_CONST_BITS
; out = dct_const_round_shift(out * cospi_16_64)
mul r0, r0, r12 ; out * cospi_16_64
mov r12, r1 ; save dest
add r0, r0, #0x2000 ; +(1 << ((DCT_CONST_BITS) - 1))
asr r0, r0, #14 ; >> DCT_CONST_BITS
; a1 = ROUND_POWER_OF_TWO(out, 5)
add r0, r0, #16 ; + (1 <<((5) - 1))
asr r0, r0, #5 ; >> 5
vdup.s16 q0, r0 ; duplicate a1
; load destination data
vld1.64 {d2}, [r1], r2
vld1.64 {d3}, [r1], r2
vld1.64 {d4}, [r1], r2
vld1.64 {d5}, [r1], r2
vld1.64 {d6}, [r1], r2
vld1.64 {d7}, [r1], r2
vld1.64 {d16}, [r1], r2
vld1.64 {d17}, [r1]
vaddw.u8 q9, q0, d2 ; dest[x] + a1
vaddw.u8 q10, q0, d3 ; dest[x] + a1
vaddw.u8 q11, q0, d4 ; dest[x] + a1
vaddw.u8 q12, q0, d5 ; dest[x] + a1
vqmovun.s16 d2, q9 ; clip_pixel
vqmovun.s16 d3, q10 ; clip_pixel
vqmovun.s16 d30, q11 ; clip_pixel
vqmovun.s16 d31, q12 ; clip_pixel
vst1.64 {d2}, [r12], r2
vst1.64 {d3}, [r12], r2
vst1.64 {d30}, [r12], r2
vst1.64 {d31}, [r12], r2
vaddw.u8 q9, q0, d6 ; dest[x] + a1
vaddw.u8 q10, q0, d7 ; dest[x] + a1
vaddw.u8 q11, q0, d16 ; dest[x] + a1
vaddw.u8 q12, q0, d17 ; dest[x] + a1
vqmovun.s16 d2, q9 ; clip_pixel
vqmovun.s16 d3, q10 ; clip_pixel
vqmovun.s16 d30, q11 ; clip_pixel
vqmovun.s16 d31, q12 ; clip_pixel
vst1.64 {d2}, [r12], r2
vst1.64 {d3}, [r12], r2
vst1.64 {d30}, [r12], r2
vst1.64 {d31}, [r12], r2
bx lr
ENDP ; |vp9_idct8x8_1_add_neon|
END
|
