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
|
;******************************************************************************
;* MMX optimized DSP utils
;* Copyright (c) 2008 Loren Merritt
;*
;* 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 "libavutil/x86/x86util.asm"
SECTION_RODATA
pb_bswap32: db 3, 2, 1, 0, 7, 6, 5, 4, 11, 10, 9, 8, 15, 14, 13, 12
SECTION_TEXT
%macro SCALARPRODUCT 0
; int ff_scalarproduct_int16(int16_t *v1, int16_t *v2, int order)
cglobal scalarproduct_int16, 3,3,3, v1, v2, order
shl orderq, 1
add v1q, orderq
add v2q, orderq
neg orderq
pxor m2, m2
.loop:
movu m0, [v1q + orderq]
movu m1, [v1q + orderq + mmsize]
pmaddwd m0, [v2q + orderq]
pmaddwd m1, [v2q + orderq + mmsize]
paddd m2, m0
paddd m2, m1
add orderq, mmsize*2
jl .loop
%if mmsize == 16
movhlps m0, m2
paddd m2, m0
pshuflw m0, m2, 0x4e
%else
pshufw m0, m2, 0x4e
%endif
paddd m2, m0
movd eax, m2
RET
%endmacro
INIT_MMX mmxext
SCALARPRODUCT
INIT_XMM sse2
SCALARPRODUCT
;-----------------------------------------------------------------------------
; void ff_vector_clip_int32(int32_t *dst, const int32_t *src, int32_t min,
; int32_t max, unsigned int len)
;-----------------------------------------------------------------------------
; %1 = number of xmm registers used
; %2 = number of inline load/process/store loops per asm loop
; %3 = process 4*mmsize (%3=0) or 8*mmsize (%3=1) bytes per loop
; %4 = CLIPD function takes min/max as float instead of int (CLIPD_SSE2)
; %5 = suffix
%macro VECTOR_CLIP_INT32 4-5
cglobal vector_clip_int32%5, 5,5,%1, dst, src, min, max, len
%if %4
cvtsi2ss m4, minm
cvtsi2ss m5, maxm
%else
movd m4, minm
movd m5, maxm
%endif
SPLATD m4
SPLATD m5
.loop:
%assign %%i 1
%rep %2
mova m0, [srcq+mmsize*0*%%i]
mova m1, [srcq+mmsize*1*%%i]
mova m2, [srcq+mmsize*2*%%i]
mova m3, [srcq+mmsize*3*%%i]
%if %3
mova m7, [srcq+mmsize*4*%%i]
mova m8, [srcq+mmsize*5*%%i]
mova m9, [srcq+mmsize*6*%%i]
mova m10, [srcq+mmsize*7*%%i]
%endif
CLIPD m0, m4, m5, m6
CLIPD m1, m4, m5, m6
CLIPD m2, m4, m5, m6
CLIPD m3, m4, m5, m6
%if %3
CLIPD m7, m4, m5, m6
CLIPD m8, m4, m5, m6
CLIPD m9, m4, m5, m6
CLIPD m10, m4, m5, m6
%endif
mova [dstq+mmsize*0*%%i], m0
mova [dstq+mmsize*1*%%i], m1
mova [dstq+mmsize*2*%%i], m2
mova [dstq+mmsize*3*%%i], m3
%if %3
mova [dstq+mmsize*4*%%i], m7
mova [dstq+mmsize*5*%%i], m8
mova [dstq+mmsize*6*%%i], m9
mova [dstq+mmsize*7*%%i], m10
%endif
%assign %%i %%i+1
%endrep
add srcq, mmsize*4*(%2+%3)
add dstq, mmsize*4*(%2+%3)
sub lend, mmsize*(%2+%3)
jg .loop
REP_RET
%endmacro
INIT_MMX mmx
%define CLIPD CLIPD_MMX
VECTOR_CLIP_INT32 0, 1, 0, 0
INIT_XMM sse2
VECTOR_CLIP_INT32 6, 1, 0, 0, _int
%define CLIPD CLIPD_SSE2
VECTOR_CLIP_INT32 6, 2, 0, 1
INIT_XMM sse4
%define CLIPD CLIPD_SSE41
%ifdef m8
VECTOR_CLIP_INT32 11, 1, 1, 0
%else
VECTOR_CLIP_INT32 6, 1, 0, 0
%endif
; %1 = aligned/unaligned
%macro BSWAP_LOOPS 1
mov r3, r2
sar r2, 3
jz .left4_%1
.loop8_%1:
mov%1 m0, [r1 + 0]
mov%1 m1, [r1 + 16]
%if cpuflag(ssse3)
pshufb m0, m2
pshufb m1, m2
mov%1 [r0 + 0], m0
mov%1 [r0 + 16], m1
%else
pshuflw m0, m0, 10110001b
pshuflw m1, m1, 10110001b
pshufhw m0, m0, 10110001b
pshufhw m1, m1, 10110001b
mova m2, m0
mova m3, m1
psllw m0, 8
psllw m1, 8
psrlw m2, 8
psrlw m3, 8
por m2, m0
por m3, m1
mov%1 [r0 + 0], m2
mov%1 [r0 + 16], m3
%endif
add r0, 32
add r1, 32
dec r2
jnz .loop8_%1
.left4_%1:
mov r2, r3
and r3, 4
jz .left
mov%1 m0, [r1]
%if cpuflag(ssse3)
pshufb m0, m2
mov%1 [r0], m0
%else
pshuflw m0, m0, 10110001b
pshufhw m0, m0, 10110001b
mova m2, m0
psllw m0, 8
psrlw m2, 8
por m2, m0
mov%1 [r0], m2
%endif
add r1, 16
add r0, 16
%endmacro
; void ff_bswap_buf(uint32_t *dst, const uint32_t *src, int w);
%macro BSWAP32_BUF 0
%if cpuflag(ssse3)
cglobal bswap32_buf, 3,4,3
mov r3, r1
mova m2, [pb_bswap32]
%else
cglobal bswap32_buf, 3,4,5
mov r3, r1
%endif
and r3, 15
jz .start_align
BSWAP_LOOPS u
jmp .left
.start_align:
BSWAP_LOOPS a
.left:
%if cpuflag(ssse3)
mov r3, r2
and r2, 2
jz .left1
movq m0, [r1]
pshufb m0, m2
movq [r0], m0
add r1, 8
add r0, 8
.left1:
and r3, 1
jz .end
mov r2d, [r1]
bswap r2d
mov [r0], r2d
%else
and r2, 3
jz .end
.loop2:
mov r3d, [r1]
bswap r3d
mov [r0], r3d
add r1, 4
add r0, 4
dec r2
jnz .loop2
%endif
.end:
RET
%endmacro
INIT_XMM sse2
BSWAP32_BUF
INIT_XMM ssse3
BSWAP32_BUF
|
