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
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
|
;******************************************************************************
;* x86 optimized channel mixing
;* Copyright (c) 2012 Justin Ruggles <justin.ruggles@gmail.com>
;*
;* This file is part of Libav.
;*
;* Libav 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.
;*
;* Libav 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 Libav; if not, write to the Free Software
;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
;******************************************************************************
%include "x86inc.asm"
%include "x86util.asm"
%include "util.asm"
SECTION_TEXT
;-----------------------------------------------------------------------------
; void ff_mix_2_to_1_fltp_flt(float **src, float **matrix, int len,
; int out_ch, int in_ch);
;-----------------------------------------------------------------------------
%macro MIX_2_TO_1_FLTP_FLT 0
cglobal mix_2_to_1_fltp_flt, 3,4,6, src, matrix, len, src1
mov src1q, [srcq+gprsize]
mov srcq, [srcq ]
sub src1q, srcq
mov matrixq, [matrixq ]
VBROADCASTSS m4, [matrixq ]
VBROADCASTSS m5, [matrixq+4]
ALIGN 16
.loop:
mulps m0, m4, [srcq ]
mulps m1, m5, [srcq+src1q ]
mulps m2, m4, [srcq+ mmsize]
mulps m3, m5, [srcq+src1q+mmsize]
addps m0, m0, m1
addps m2, m2, m3
mova [srcq ], m0
mova [srcq+mmsize], m2
add srcq, mmsize*2
sub lend, mmsize*2/4
jg .loop
%if mmsize == 32
vzeroupper
RET
%else
REP_RET
%endif
%endmacro
INIT_XMM sse
MIX_2_TO_1_FLTP_FLT
%if HAVE_AVX
INIT_YMM avx
MIX_2_TO_1_FLTP_FLT
%endif
;-----------------------------------------------------------------------------
; void ff_mix_2_to_1_s16p_flt(int16_t **src, float **matrix, int len,
; int out_ch, int in_ch);
;-----------------------------------------------------------------------------
%macro MIX_2_TO_1_S16P_FLT 0
cglobal mix_2_to_1_s16p_flt, 3,4,6, src, matrix, len, src1
mov src1q, [srcq+gprsize]
mov srcq, [srcq]
sub src1q, srcq
mov matrixq, [matrixq ]
VBROADCASTSS m4, [matrixq ]
VBROADCASTSS m5, [matrixq+4]
ALIGN 16
.loop:
mova m0, [srcq ]
mova m2, [srcq+src1q]
S16_TO_S32_SX 0, 1
S16_TO_S32_SX 2, 3
cvtdq2ps m0, m0
cvtdq2ps m1, m1
cvtdq2ps m2, m2
cvtdq2ps m3, m3
mulps m0, m4
mulps m1, m4
mulps m2, m5
mulps m3, m5
addps m0, m2
addps m1, m3
cvtps2dq m0, m0
cvtps2dq m1, m1
packssdw m0, m1
mova [srcq], m0
add srcq, mmsize
sub lend, mmsize/2
jg .loop
REP_RET
%endmacro
INIT_XMM sse2
MIX_2_TO_1_S16P_FLT
INIT_XMM sse4
MIX_2_TO_1_S16P_FLT
;-----------------------------------------------------------------------------
; void ff_mix_2_to_1_s16p_q8(int16_t **src, int16_t **matrix, int len,
; int out_ch, int in_ch);
;-----------------------------------------------------------------------------
INIT_XMM sse2
cglobal mix_2_to_1_s16p_q8, 3,4,6, src, matrix, len, src1
mov src1q, [srcq+gprsize]
mov srcq, [srcq]
sub src1q, srcq
mov matrixq, [matrixq]
movd m4, [matrixq]
movd m5, [matrixq]
SPLATW m4, m4, 0
SPLATW m5, m5, 1
pxor m0, m0
punpcklwd m4, m0
punpcklwd m5, m0
ALIGN 16
.loop:
mova m0, [srcq ]
mova m2, [srcq+src1q]
punpckhwd m1, m0, m0
punpcklwd m0, m0
punpckhwd m3, m2, m2
punpcklwd m2, m2
pmaddwd m0, m4
pmaddwd m1, m4
pmaddwd m2, m5
pmaddwd m3, m5
paddd m0, m2
paddd m1, m3
psrad m0, 8
psrad m1, 8
packssdw m0, m1
mova [srcq], m0
add srcq, mmsize
sub lend, mmsize/2
jg .loop
REP_RET
;-----------------------------------------------------------------------------
; void ff_mix_1_to_2_fltp_flt(float **src, float **matrix, int len,
; int out_ch, int in_ch);
;-----------------------------------------------------------------------------
%macro MIX_1_TO_2_FLTP_FLT 0
cglobal mix_1_to_2_fltp_flt, 3,5,4, src0, matrix0, len, src1, matrix1
mov src1q, [src0q+gprsize]
mov src0q, [src0q]
sub src1q, src0q
mov matrix1q, [matrix0q+gprsize]
mov matrix0q, [matrix0q]
VBROADCASTSS m2, [matrix0q]
VBROADCASTSS m3, [matrix1q]
ALIGN 16
.loop:
mova m0, [src0q]
mulps m1, m0, m3
mulps m0, m0, m2
mova [src0q ], m0
mova [src0q+src1q], m1
add src0q, mmsize
sub lend, mmsize/4
jg .loop
%if mmsize == 32
vzeroupper
RET
%else
REP_RET
%endif
%endmacro
INIT_XMM sse
MIX_1_TO_2_FLTP_FLT
%if HAVE_AVX
INIT_YMM avx
MIX_1_TO_2_FLTP_FLT
%endif
;-----------------------------------------------------------------------------
; void ff_mix_1_to_2_s16p_flt(int16_t **src, float **matrix, int len,
; int out_ch, int in_ch);
;-----------------------------------------------------------------------------
%macro MIX_1_TO_2_S16P_FLT 0
cglobal mix_1_to_2_s16p_flt, 3,5,6, src0, matrix0, len, src1, matrix1
mov src1q, [src0q+gprsize]
mov src0q, [src0q]
sub src1q, src0q
mov matrix1q, [matrix0q+gprsize]
mov matrix0q, [matrix0q]
VBROADCASTSS m4, [matrix0q]
VBROADCASTSS m5, [matrix1q]
ALIGN 16
.loop:
mova m0, [src0q]
S16_TO_S32_SX 0, 2
cvtdq2ps m0, m0
cvtdq2ps m2, m2
mulps m1, m0, m5
mulps m0, m0, m4
mulps m3, m2, m5
mulps m2, m2, m4
cvtps2dq m0, m0
cvtps2dq m1, m1
cvtps2dq m2, m2
cvtps2dq m3, m3
packssdw m0, m2
packssdw m1, m3
mova [src0q ], m0
mova [src0q+src1q], m1
add src0q, mmsize
sub lend, mmsize/2
jg .loop
REP_RET
%endmacro
INIT_XMM sse2
MIX_1_TO_2_S16P_FLT
INIT_XMM sse4
MIX_1_TO_2_S16P_FLT
%if HAVE_AVX
INIT_XMM avx
MIX_1_TO_2_S16P_FLT
%endif
|
