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
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
|
@/******************************************************************************
@ *
@ * Copyright (C) 2015 The Android Open Source Project
@ *
@ * Licensed under the Apache License, Version 2.0 (the "License");
@ * you may not use this file except in compliance with the License.
@ * You may obtain a copy of the License at:
@ *
@ * http://www.apache.org/licenses/LICENSE-2.0
@ *
@ * Unless required by applicable law or agreed to in writing, software
@ * distributed under the License is distributed on an "AS IS" BASIS,
@ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
@ * See the License for the specific language governing permissions and
@ * limitations under the License.
@ *
@ *****************************************************************************
@ * Originally developed and contributed by Ittiam Systems Pvt. Ltd, Bangalore
@*/
@**
@******************************************************************************
@* @file
@* ih264_inter_pred_luma_vert_a9q.s
@*
@* @brief
@* Contains function definitions for inter prediction interpolation.
@*
@* @author
@* Ittiam
@*
@* @par List of Functions:
@*
@* - ih264_inter_pred_luma_vert_a9q()
@*
@* @remarks
@* None
@*
@*******************************************************************************
@*
@* All the functions here are replicated from ih264_inter_pred_filters.c
@
@**
@**
@**
@ *******************************************************************************
@ *
@ * @brief
@ * Interprediction luma filter for vertical input
@ *
@ * @par Description:
@ * Applies a 6 tap vertcal filter.The output is clipped to 8 bits
@ * sec 8.4.2.2.1 titled "Luma sample interpolation process"
@ *
@ * @param[in] pu1_src
@ * UWORD8 pointer to the source
@ *
@ * @param[out] pu1_dst
@ * UWORD8 pointer to the destination
@ *
@ * @param[in] src_strd
@ * integer source stride
@ *
@ * @param[in] dst_strd
@ * integer destination stride
@ *
@ * @param[in] ht
@ * integer height of the array
@ *
@ * @param[in] wd
@ * integer width of the array
@ *
@ * @returns
@ *
@ * @remarks
@ * None
@ *
@ *******************************************************************************
@void ih264_inter_pred_luma_vert (
@ UWORD8 *pu1_src,
@ UWORD8 *pu1_dst,
@ WORD32 src_strd,
@ WORD32 dst_strd,
@ WORD32 ht,
@ WORD32 wd )
@**************Variables Vs Registers*****************************************
@ r0 => *pu1_src
@ r1 => *pu1_dst
@ r2 => src_strd
@ r3 => dst_strd
@ r5 => ht
@ r6 => wd
.text
.p2align 2
.global ih264_inter_pred_luma_vert_a9q
ih264_inter_pred_luma_vert_a9q:
stmfd sp!, {r4-r12, r14} @store register values to stack
vstmdb sp!, {d8-d15} @push neon registers to stack
ldr r5, [sp, #104] @Loads ht
sub r0, r0, r2, lsl #1 @pu1_src-2*src_strd
ldr r6, [sp, #108] @Loads wd
vmov.u16 q11, #20 @ Filter coeff 0x14 into Q11
subs r12, r6, #8 @if wd=8 branch to loop_8
vmov.u16 q12, #5 @ Filter coeff 0x5 into Q12
beq loop_8
subs r12, r6, #4 @if wd=4 branch to loop_4
beq loop_4
loop_16: @when wd=16
vld1.u32 {q0}, [r0], r2 @ Vector load from src[0_0]
vld1.u32 {q1}, [r0], r2 @ Vector load from src[1_0]
vld1.u32 {q2}, [r0], r2 @ Vector load from src[2_0]
vld1.u32 {q3}, [r0], r2 @ Vector load from src[3_0]
vld1.u32 {q4}, [r0], r2 @ Vector load from src[4_0]
vaddl.u8 q6, d4, d6 @ temp1 = src[2_0] + src[3_0]
vld1.u32 {q5}, [r0], r2 @ Vector load from src[5_0]
vaddl.u8 q7, d0, d10 @ temp = src[0_0] + src[5_0]
vaddl.u8 q8, d2, d8 @ temp2 = src[1_0] + src[4_0]
vmla.u16 q7, q6, q11 @ temp += temp1 * 20
vaddl.u8 q10, d1, d11 @ temp4 = src[0_8] + src[5_8]
vaddl.u8 q9, d5, d7 @ temp3 = src[2_8] + src[3_8]
vmla.u16 q10, q9, q11 @ temp4 += temp3 * 20
vld1.u32 {q0}, [r0], r2
vaddl.u8 q13, d3, d9 @ temp5 = src[1_8] + src[4_8]
vaddl.u8 q6, d6, d8
vmls.u16 q7, q8, q12 @ temp -= temp2 * 5
vaddl.u8 q8, d2, d0
vaddl.u8 q9, d4, d10
vmla.u16 q8, q6, q11
vmls.u16 q10, q13, q12 @ temp4 -= temp5 * 5
vaddl.u8 q13, d5, d11
vaddl.u8 q6, d7, d9
vqrshrun.s16 d30, q7, #5 @ dst[0_0] = CLIP_U8((temp +16) >> 5)
vaddl.u8 q7, d3, d1
vld1.u32 {q1}, [r0], r2
vmla.u16 q7, q6, q11
vmls.u16 q8, q9, q12
vqrshrun.s16 d31, q10, #5 @ dst[0_8] = CLIP_U8((temp4 +16) >> 5)
vaddl.u8 q9, d4, d2
vaddl.u8 q6, d8, d10
vst1.u32 {q15}, [r1], r3 @ Vector store to dst[0_0]
vmla.u16 q9, q6, q11
vaddl.u8 q10, d6, d0
vmls.u16 q7, q13, q12
vqrshrun.s16 d30, q8, #5
vaddl.u8 q6, d9, d11
vaddl.u8 q8, d5, d3
vaddl.u8 q13, d7, d1
vmla.u16 q8, q6, q11
vmls.u16 q9, q10, q12
vld1.u32 {q2}, [r0], r2
vqrshrun.s16 d31, q7, #5
vaddl.u8 q6, d10, d0
vaddl.u8 q7, d6, d4
vaddl.u8 q10, d8, d2
vmla.u16 q7, q6, q11
vmls.u16 q8, q13, q12
vst1.u32 {q15}, [r1], r3 @store row 1
vqrshrun.s16 d30, q9, #5
vaddl.u8 q9, d7, d5
vaddl.u8 q6, d11, d1
vmla.u16 q9, q6, q11
vaddl.u8 q13, d9, d3
vmls.u16 q7, q10, q12
vqrshrun.s16 d31, q8, #5
vmls.u16 q9, q13, q12
vaddl.u8 q6, d0, d2 @ temp1 = src[2_0] + src[3_0]
vst1.u32 {q15}, [r1], r3 @store row 2
vaddl.u8 q8, d10, d4 @ temp2 = src[1_0] + src[4_0]
vaddl.u8 q10, d9, d7 @ temp4 = src[0_8] + src[5_8]
vqrshrun.s16 d30, q7, #5
vaddl.u8 q13, d5, d11 @ temp5 = src[1_8] + src[4_8]
vaddl.u8 q7, d8, d6 @ temp = src[0_0] + src[5_0]
vqrshrun.s16 d31, q9, #5
vmla.u16 q7, q6, q11 @ temp += temp1 * 20
vaddl.u8 q9, d1, d3 @ temp3 = src[2_8] + src[3_8]
vst1.u32 {q15}, [r1], r3 @store row 3
subs r5, r5, #4 @ 4 rows processed, decrement by 4
subne r0, r0 , r2, lsl #2
subne r0, r0, r2
beq end_func @ Branch if height==4
b loop_16 @ looping if height = 8 or 16
loop_8:
@ Processing row0 and row1
vld1.u32 d0, [r0], r2 @ Vector load from src[0_0]
vld1.u32 d1, [r0], r2 @ Vector load from src[1_0]
vld1.u32 d2, [r0], r2 @ Vector load from src[2_0]
vld1.u32 d3, [r0], r2 @ Vector load from src[3_0]
vld1.u32 d4, [r0], r2 @ Vector load from src[4_0]
vld1.u32 d5, [r0], r2 @ Vector load from src[5_0]
vaddl.u8 q3, d2, d3 @ temp1 = src[2_0] + src[3_0]
vaddl.u8 q4, d0, d5 @ temp = src[0_0] + src[5_0]
vaddl.u8 q5, d1, d4 @ temp2 = src[1_0] + src[4_0]
vmla.u16 q4, q3, q11 @ temp += temp1 * 20
vld1.u32 d6, [r0], r2
vaddl.u8 q7, d3, d4
vaddl.u8 q8, d1, d6
vaddl.u8 q9, d2, d5
vmls.u16 q4, q5, q12 @ temp -= temp2 * 5
vmla.u16 q8, q7, q11
vld1.u32 d7, [r0], r2
vaddl.u8 q10, d4, d5
vaddl.u8 q6, d2, d7
vaddl.u8 q5, d3, d6
vmls.u16 q8, q9, q12
vqrshrun.s16 d26, q4, #5 @ dst[0_0] = CLIP_U8( (temp + 16) >> 5)
vmla.u16 q6, q10, q11
vld1.u32 d0, [r0], r2
vaddl.u8 q7, d5, d6
vqrshrun.s16 d27, q8, #5
vaddl.u8 q10, d3, d0
vmls.u16 q6, q5, q12
vst1.u32 d26, [r1], r3 @ Vector store to dst[0_0]
vaddl.u8 q9, d4, d7
vmla.u16 q10, q7, q11
vst1.u32 d27, [r1], r3
vqrshrun.s16 d28, q6, #5
vst1.u32 d28, [r1], r3
vmls.u16 q10, q9, q12
vqrshrun.s16 d29, q10, #5
vst1.u32 d29, [r1], r3 @store row 3
subs r5, r5, #4 @ 4 rows processed, decrement by 4
subne r0, r0 , r2, lsl #2
subne r0, r0, r2
beq end_func @ Branch if height==4
b loop_8 @looping if height == 8 or 16
loop_4:
@ Processing row0 and row1
vld1.u32 d0[0], [r0], r2 @ Vector load from src[0_0]
vld1.u32 d1[0], [r0], r2 @ Vector load from src[1_0]
vld1.u32 d2[0], [r0], r2 @ Vector load from src[2_0]
vld1.u32 d3[0], [r0], r2 @ Vector load from src[3_0]
vld1.u32 d4[0], [r0], r2 @ Vector load from src[4_0]
vld1.u32 d5[0], [r0], r2 @ Vector load from src[5_0]
vaddl.u8 q3, d2, d3 @ temp1 = src[2_0] + src[3_0]
vaddl.u8 q4, d0, d5 @ temp = src[0_0] + src[5_0]
vaddl.u8 q5, d1, d4 @ temp2 = src[1_0] + src[4_0]
vmla.u16 q4, q3, q11 @ temp += temp1 * 20
vld1.u32 d6[0], [r0], r2
vaddl.u8 q7, d3, d4
vaddl.u8 q8, d1, d6
vaddl.u8 q9, d2, d5
vmls.u16 q4, q5, q12 @ temp -= temp2 * 5
vld1.u32 d7[0], [r0], r2
vmla.u16 q8, q7, q11
vaddl.u8 q10, d4, d5
vaddl.u8 q6, d2, d7
vaddl.u8 q5, d3, d6
vmls.u16 q8, q9, q12
vqrshrun.s16 d26, q4, #5 @ dst[0_0] = CLIP_U8( (temp + 16) >> 5)
vmla.u16 q6, q10, q11
vld1.u32 d0[0], [r0], r2
vaddl.u8 q7, d5, d6
vqrshrun.s16 d27, q8, #5
vaddl.u8 q10, d3, d0
vmls.u16 q6, q5, q12
vst1.u32 d26[0], [r1], r3 @ Vector store to dst[0_0]
vaddl.u8 q9, d4, d7
vmla.u16 q10, q7, q11
vst1.u32 d27[0], [r1], r3
vqrshrun.s16 d28, q6, #5
vst1.u32 d28[0], [r1], r3
vmls.u16 q10, q9, q12
vqrshrun.s16 d29, q10, #5
vst1.u32 d29[0], [r1], r3 @store row 3
subs r5, r5, #8
subeq r0, r0, r2, lsl #2
subeq r0, r0, r2
beq loop_4 @ Loop if height==8
end_func:
vldmia sp!, {d8-d15} @ Restore neon registers that were saved
ldmfd sp!, {r4-r12, pc} @Restoring registers from stack
|