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
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
|
;*****************************************************************************
;* x86-optimized functions for v360 filter
;*
;* 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 "libavutil/x86/x86util.asm"
%if HAVE_AVX2_EXTERNAL
SECTION_RODATA
pb_mask: db 0,4,8,12,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1
pw_mask: db 0,1,4, 5, 8, 9,12,13,-1,-1,-1,-1,-1,-1,-1,-1
pd_255: times 4 dd 255
pd_65535: times 4 dd 65535
SECTION .text
; void ff_remap2_8bit_line_avx2(uint8_t *dst, int width, const uint8_t *src, ptrdiff_t in_linesize,
; const uint16_t *u, const uint16_t *v, const int16_t *ker);
INIT_YMM avx2
cglobal remap1_8bit_line, 6, 7, 6, dst, width, src, in_linesize, u, v, x
movsxdifnidn widthq, widthd
xor xq, xq
movd xm0, in_linesized
pcmpeqw m4, m4
VBROADCASTI128 m3, [pb_mask]
vpbroadcastd m0, xm0
.loop:
pmovsxwd m1, [vq + xq * 2]
pmovsxwd m2, [uq + xq * 2]
pmulld m1, m0
paddd m1, m2
mova m2, m4
vpgatherdd m5, [srcq + m1], m2
pshufb m1, m5, m3
vextracti128 xm2, m1, 1
movd [dstq+xq], xm1
movd [dstq+xq+4], xm2
add xq, mmsize / 4
cmp xq, widthq
jl .loop
RET
INIT_YMM avx2
cglobal remap1_16bit_line, 6, 7, 6, dst, width, src, in_linesize, u, v, x
movsxdifnidn widthq, widthd
xor xq, xq
movd xm0, in_linesized
pcmpeqw m4, m4
VBROADCASTI128 m3, [pw_mask]
vpbroadcastd m0, xm0
.loop:
pmovsxwd m1, [vq + xq * 2]
pmovsxwd m2, [uq + xq * 2]
pslld m2, 0x1
pmulld m1, m0
paddd m1, m2
mova m2, m4
vpgatherdd m5, [srcq + m1], m2
pshufb m1, m5, m3
vextracti128 xm2, m1, 1
movq [dstq+xq*2], xm1
movq [dstq+xq*2+8], xm2
add xq, mmsize / 4
cmp xq, widthq
jl .loop
RET
INIT_YMM avx2
cglobal remap2_8bit_line, 7, 8, 8, dst, width, src, in_linesize, u, v, ker, x
movsxdifnidn widthq, widthd
movd xm0, in_linesized
%if ARCH_X86_32
DEFINE_ARGS dst, width, src, x, u, v, ker
%endif
xor xq, xq
pcmpeqw m7, m7
vpbroadcastd m0, xm0
vpbroadcastd m6, [pd_255]
.loop:
pmovsxwd m1, [kerq + xq * 8]
pmovsxwd m2, [vq + xq * 8]
pmovsxwd m3, [uq + xq * 8]
pmulld m4, m2, m0
paddd m4, m3
mova m3, m7
vpgatherdd m2, [srcq + m4], m3
pand m2, m6
pmulld m2, m1
phaddd m2, m2
phaddd m1, m2, m2
psrld m1, m1, 0xe
vextracti128 xm2, m1, 1
pextrb [dstq+xq], xm1, 0
pextrb [dstq+xq+1], xm2, 0
add xq, mmsize / 16
cmp xq, widthq
jl .loop
RET
INIT_YMM avx2
cglobal remap2_16bit_line, 7, 8, 8, dst, width, src, in_linesize, u, v, ker, x
movsxdifnidn widthq, widthd
movd xm0, in_linesized
%if ARCH_X86_32
DEFINE_ARGS dst, width, src, x, u, v, ker
%endif
xor xq, xq
pcmpeqw m7, m7
vpbroadcastd m0, xm0
vpbroadcastd m6, [pd_65535]
.loop:
pmovsxwd m1, [kerq + xq * 8]
pmovsxwd m2, [vq + xq * 8]
pmovsxwd m3, [uq + xq * 8]
pslld m3, 0x1
pmulld m4, m2, m0
paddd m4, m3
mova m3, m7
vpgatherdd m2, [srcq + m4], m3
pand m2, m6
pmulld m2, m1
phaddd m2, m2
phaddd m1, m2, m2
psrld m1, m1, 0xe
vextracti128 xm2, m1, 1
pextrw [dstq+xq*2], xm1, 0
pextrw [dstq+xq*2+2], xm2, 0
add xq, mmsize / 16
cmp xq, widthq
jl .loop
RET
%if ARCH_X86_64
INIT_YMM avx2
cglobal remap3_8bit_line, 7, 11, 8, dst, width, src, in_linesize, u, v, ker, x, y, tmp, z
movsxdifnidn widthq, widthd
xor zq, zq
xor yq, yq
xor xq, xq
movd xm0, in_linesized
pcmpeqw m7, m7
vpbroadcastd m0, xm0
vpbroadcastd m6, [pd_255]
.loop:
pmovsxwd m1, [kerq + yq]
pmovsxwd m2, [vq + yq]
pmovsxwd m3, [uq + yq]
pmulld m4, m2, m0
paddd m4, m3
mova m3, m7
vpgatherdd m2, [srcq + m4], m3
pand m2, m6
pmulld m2, m1
HADDD m2, m1
movzx tmpq, word [vq + yq + 16]
imul tmpq, in_linesizeq
movzx zq, word [uq + yq + 16]
add tmpq, zq
movzx zq, byte [srcq + tmpq]
movzx tmpq, word [kerq + yq + 16]
imul zd, tmpd
movd xm1, zd
paddd m2, m1
psrld m2, m2, 0xe
packuswb m2, m2
pextrb [dstq+xq], xm2, 0
add xq, 1
add yq, 18
cmp xq, widthq
jl .loop
RET
INIT_YMM avx2
cglobal remap4_8bit_line, 7, 9, 11, dst, width, src, in_linesize, u, v, ker, x, y
movsxdifnidn widthq, widthd
xor yq, yq
xor xq, xq
movd xm0, in_linesized
pcmpeqw m7, m7
vpbroadcastd m0, xm0
vpbroadcastd m6, [pd_255]
.loop:
pmovsxwd m1, [kerq + yq]
pmovsxwd m5, [kerq + yq + 16]
pmovsxwd m2, [vq + yq]
pmovsxwd m8, [vq + yq + 16]
pmovsxwd m3, [uq + yq]
pmovsxwd m9, [uq + yq + 16]
pmulld m4, m2, m0
pmulld m10, m8, m0
paddd m4, m3
paddd m10, m9
mova m3, m7
vpgatherdd m2, [srcq + m4], m3
mova m3, m7
vpgatherdd m4, [srcq + m10], m3
pand m2, m6
pand m4, m6
pmulld m2, m1
pmulld m4, m5
paddd m2, m4
HADDD m2, m1
psrld m2, m2, 0xe
packuswb m2, m2
pextrb [dstq+xq], xm2, 0
add xq, 1
add yq, 32
cmp xq, widthq
jl .loop
RET
%endif
%endif
|
