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
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
|
;
; Copyright (c) 2010 The WebM project authors. All Rights Reserved.
;
; Use of this source code is governed by a BSD-style license
; that can be found in the LICENSE file in the root of the source
; tree. An additional intellectual property rights grant can be found
; in the file PATENTS. All contributing project authors may
; be found in the AUTHORS file in the root of the source tree.
;
.globl vp8_sad16x16_ppc
.globl vp8_sad16x8_ppc
.globl vp8_sad8x16_ppc
.globl vp8_sad8x8_ppc
.globl vp8_sad4x4_ppc
.macro load_aligned_16 V R O
lvsl v3, 0, \R ;# permutate value for alignment
lvx v1, 0, \R
lvx v2, \O, \R
vperm \V, v1, v2, v3
.endm
.macro prologue
mfspr r11, 256 ;# get old VRSAVE
oris r12, r11, 0xffc0
mtspr 256, r12 ;# set VRSAVE
stwu r1, -32(r1) ;# create space on the stack
li r10, 16 ;# load offset and loop counter
vspltisw v8, 0 ;# zero out total to start
.endm
.macro epilogue
addi r1, r1, 32 ;# recover stack
mtspr 256, r11 ;# reset old VRSAVE
.endm
.macro SAD_16
;# v6 = abs (v4 - v5)
vsububs v6, v4, v5
vsububs v7, v5, v4
vor v6, v6, v7
;# v8 += abs (v4 - v5)
vsum4ubs v8, v6, v8
.endm
.macro sad_16_loop loop_label
lvsl v3, 0, r5 ;# only needs to be done once per block
;# preload a line of data before getting into the loop
lvx v4, 0, r3
lvx v1, 0, r5
lvx v2, r10, r5
add r5, r5, r6
add r3, r3, r4
vperm v5, v1, v2, v3
.align 4
\loop_label:
;# compute difference on first row
vsububs v6, v4, v5
vsububs v7, v5, v4
;# load up next set of data
lvx v9, 0, r3
lvx v1, 0, r5
lvx v2, r10, r5
;# perform abs() of difference
vor v6, v6, v7
add r3, r3, r4
;# add to the running tally
vsum4ubs v8, v6, v8
;# now onto the next line
vperm v5, v1, v2, v3
add r5, r5, r6
lvx v4, 0, r3
;# compute difference on second row
vsububs v6, v9, v5
lvx v1, 0, r5
vsububs v7, v5, v9
lvx v2, r10, r5
vor v6, v6, v7
add r3, r3, r4
vsum4ubs v8, v6, v8
vperm v5, v1, v2, v3
add r5, r5, r6
bdnz \loop_label
vspltisw v7, 0
vsumsws v8, v8, v7
stvx v8, 0, r1
lwz r3, 12(r1)
.endm
.macro sad_8_loop loop_label
.align 4
\loop_label:
;# only one of the inputs should need to be aligned.
load_aligned_16 v4, r3, r10
load_aligned_16 v5, r5, r10
;# move onto the next line
add r3, r3, r4
add r5, r5, r6
;# only one of the inputs should need to be aligned.
load_aligned_16 v6, r3, r10
load_aligned_16 v7, r5, r10
;# move onto the next line
add r3, r3, r4
add r5, r5, r6
vmrghb v4, v4, v6
vmrghb v5, v5, v7
SAD_16
bdnz \loop_label
vspltisw v7, 0
vsumsws v8, v8, v7
stvx v8, 0, r1
lwz r3, 12(r1)
.endm
.align 2
;# r3 unsigned char *src_ptr
;# r4 int src_stride
;# r5 unsigned char *ref_ptr
;# r6 int ref_stride
;#
;# r3 return value
vp8_sad16x16_ppc:
prologue
li r9, 8
mtctr r9
sad_16_loop sad16x16_loop
epilogue
blr
.align 2
;# r3 unsigned char *src_ptr
;# r4 int src_stride
;# r5 unsigned char *ref_ptr
;# r6 int ref_stride
;#
;# r3 return value
vp8_sad16x8_ppc:
prologue
li r9, 4
mtctr r9
sad_16_loop sad16x8_loop
epilogue
blr
.align 2
;# r3 unsigned char *src_ptr
;# r4 int src_stride
;# r5 unsigned char *ref_ptr
;# r6 int ref_stride
;#
;# r3 return value
vp8_sad8x16_ppc:
prologue
li r9, 8
mtctr r9
sad_8_loop sad8x16_loop
epilogue
blr
.align 2
;# r3 unsigned char *src_ptr
;# r4 int src_stride
;# r5 unsigned char *ref_ptr
;# r6 int ref_stride
;#
;# r3 return value
vp8_sad8x8_ppc:
prologue
li r9, 4
mtctr r9
sad_8_loop sad8x8_loop
epilogue
blr
.macro transfer_4x4 I P
lwz r0, 0(\I)
add \I, \I, \P
lwz r7, 0(\I)
add \I, \I, \P
lwz r8, 0(\I)
add \I, \I, \P
lwz r9, 0(\I)
stw r0, 0(r1)
stw r7, 4(r1)
stw r8, 8(r1)
stw r9, 12(r1)
.endm
.align 2
;# r3 unsigned char *src_ptr
;# r4 int src_stride
;# r5 unsigned char *ref_ptr
;# r6 int ref_stride
;#
;# r3 return value
vp8_sad4x4_ppc:
prologue
transfer_4x4 r3, r4
lvx v4, 0, r1
transfer_4x4 r5, r6
lvx v5, 0, r1
vspltisw v8, 0 ;# zero out total to start
;# v6 = abs (v4 - v5)
vsububs v6, v4, v5
vsububs v7, v5, v4
vor v6, v6, v7
;# v8 += abs (v4 - v5)
vsum4ubs v7, v6, v8
vsumsws v7, v7, v8
stvx v7, 0, r1
lwz r3, 12(r1)
epilogue
blr
|