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
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
|
/*
* ARM NEON optimised MDCT
* Copyright (c) 2009 Mans Rullgard <mans@mansr.com>
*
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* FFmpeg is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include "asm.S"
.fpu neon
.text
function ff_imdct_half_neon, export=1
push {r4-r8,lr}
mov r12, #1
ldr lr, [r0, #4] @ nbits
ldr r4, [r0, #8] @ tcos
ldr r5, [r0, #12] @ tsin
ldr r3, [r0, #24] @ revtab
lsl r12, r12, lr @ n = 1 << nbits
lsr lr, r12, #2 @ n4 = n >> 2
add r7, r2, r12, lsl #1
mov r12, #-16
sub r7, r7, #16
vld2.32 {d16-d17},[r7,:128],r12 @ d16=x,n1 d17=x,n0
vld2.32 {d0-d1}, [r2,:128]! @ d0 =m0,x d1 =m1,x
vrev64.32 d17, d17
vld1.32 {d2}, [r4,:64]! @ d2=c0,c1
vmul.f32 d6, d17, d2
vld1.32 {d3}, [r5,:64]! @ d3=s0,s1
vmul.f32 d7, d0, d2
1:
subs lr, lr, #2
ldr r6, [r3], #4
vmul.f32 d4, d0, d3
vmul.f32 d5, d17, d3
vsub.f32 d4, d6, d4
vadd.f32 d5, d5, d7
uxtah r8, r1, r6, ror #16
uxtah r6, r1, r6
beq 1f
vld2.32 {d16-d17},[r7,:128],r12
vld2.32 {d0-d1}, [r2,:128]!
vrev64.32 d17, d17
vld1.32 {d2}, [r4,:64]!
vmul.f32 d6, d17, d2
vld1.32 {d3}, [r5,:64]!
vmul.f32 d7, d0, d2
vst2.32 {d4[0],d5[0]}, [r6,:64]
vst2.32 {d4[1],d5[1]}, [r8,:64]
b 1b
1:
vst2.32 {d4[0],d5[0]}, [r6,:64]
vst2.32 {d4[1],d5[1]}, [r8,:64]
mov r4, r0
mov r6, r1
add r0, r0, #16
bl ff_fft_calc_neon
mov r12, #1
ldr lr, [r4, #4] @ nbits
ldr r5, [r4, #12] @ tsin
ldr r4, [r4, #8] @ tcos
lsl r12, r12, lr @ n = 1 << nbits
lsr lr, r12, #3 @ n8 = n >> 3
add r4, r4, lr, lsl #2
add r5, r5, lr, lsl #2
add r6, r6, lr, lsl #3
sub r1, r4, #8
sub r2, r5, #8
sub r3, r6, #16
mov r7, #-16
mov r12, #-8
mov r8, r6
mov r0, r3
vld2.32 {d0-d1}, [r3,:128], r7 @ d0 =i1,r1 d1 =i0,r0
vld2.32 {d20-d21},[r6,:128]! @ d20=i2,r2 d21=i3,r3
vld1.32 {d18}, [r2,:64], r12 @ d18=s1,s0
1:
subs lr, lr, #2
vmul.f32 d7, d0, d18
vld1.32 {d19}, [r5,:64]! @ d19=s2,s3
vmul.f32 d4, d1, d18
vld1.32 {d16}, [r1,:64], r12 @ d16=c1,c0
vmul.f32 d5, d21, d19
vld1.32 {d17}, [r4,:64]! @ d17=c2,c3
vmul.f32 d6, d20, d19
vmul.f32 d22, d1, d16
vmul.f32 d23, d21, d17
vmul.f32 d24, d0, d16
vmul.f32 d25, d20, d17
vadd.f32 d7, d7, d22
vadd.f32 d6, d6, d23
vsub.f32 d4, d4, d24
vsub.f32 d5, d5, d25
beq 1f
vld2.32 {d0-d1}, [r3,:128], r7
vld2.32 {d20-d21},[r6,:128]!
vld1.32 {d18}, [r2,:64], r12
vrev64.32 q3, q3
vst2.32 {d4,d6}, [r0,:128], r7
vst2.32 {d5,d7}, [r8,:128]!
b 1b
1:
vrev64.32 q3, q3
vst2.32 {d4,d6}, [r0,:128]
vst2.32 {d5,d7}, [r8,:128]
pop {r4-r8,pc}
.endfunc
function ff_imdct_calc_neon, export=1
push {r4-r6,lr}
ldr r3, [r0, #4]
mov r4, #1
mov r5, r1
lsl r4, r4, r3
add r1, r1, r4
bl ff_imdct_half_neon
add r0, r5, r4, lsl #2
add r1, r5, r4, lsl #1
sub r0, r0, #8
sub r2, r1, #16
mov r3, #-16
mov r6, #-8
vmov.i32 d30, #1<<31
1:
vld1.32 {d0-d1}, [r2,:128], r3
pld [r0, #-16]
vrev64.32 q0, q0
vld1.32 {d2-d3}, [r1,:128]!
veor d4, d1, d30
pld [r2, #-16]
vrev64.32 q1, q1
veor d5, d0, d30
vst1.32 {d2}, [r0,:64], r6
vst1.32 {d3}, [r0,:64], r6
vst1.32 {d4-d5}, [r5,:128]!
subs r4, r4, #16
bgt 1b
pop {r4-r6,pc}
.endfunc
|
