OLD | NEW |
1 ; | 1 ; |
2 ; Copyright (c) 2013 The WebM project authors. All Rights Reserved. | 2 ; Copyright (c) 2013 The WebM project authors. All Rights Reserved. |
3 ; | 3 ; |
4 ; Use of this source code is governed by a BSD-style license | 4 ; Use of this source code is governed by a BSD-style license |
5 ; that can be found in the LICENSE file in the root of the source | 5 ; that can be found in the LICENSE file in the root of the source |
6 ; tree. An additional intellectual property rights grant can be found | 6 ; tree. An additional intellectual property rights grant can be found |
7 ; in the file PATENTS. All contributing project authors may | 7 ; in the file PATENTS. All contributing project authors may |
8 ; be found in the AUTHORS file in the root of the source tree. | 8 ; be found in the AUTHORS file in the root of the source tree. |
9 ; | 9 ; |
10 | 10 |
11 EXPORT |vp9_idct4x4_16_add_neon| | 11 EXPORT |vpx_idct4x4_16_add_neon| |
12 ARM | 12 ARM |
13 REQUIRE8 | 13 REQUIRE8 |
14 PRESERVE8 | 14 PRESERVE8 |
15 | 15 |
16 AREA ||.text||, CODE, READONLY, ALIGN=2 | 16 AREA ||.text||, CODE, READONLY, ALIGN=2 |
17 | 17 |
18 AREA Block, CODE, READONLY ; name this block of code | 18 AREA Block, CODE, READONLY ; name this block of code |
19 ;void vp9_idct4x4_16_add_neon(int16_t *input, uint8_t *dest, int dest_stride) | 19 ;void vpx_idct4x4_16_add_neon(int16_t *input, uint8_t *dest, int dest_stride) |
20 ; | 20 ; |
21 ; r0 int16_t input | 21 ; r0 int16_t input |
22 ; r1 uint8_t *dest | 22 ; r1 uint8_t *dest |
23 ; r2 int dest_stride) | 23 ; r2 int dest_stride) |
24 | 24 |
25 |vp9_idct4x4_16_add_neon| PROC | 25 |vpx_idct4x4_16_add_neon| PROC |
26 | 26 |
27 ; The 2D transform is done with two passes which are actually pretty | 27 ; The 2D transform is done with two passes which are actually pretty |
28 ; similar. We first transform the rows. This is done by transposing | 28 ; similar. We first transform the rows. This is done by transposing |
29 ; the inputs, doing an SIMD column transform (the columns are the | 29 ; the inputs, doing an SIMD column transform (the columns are the |
30 ; transposed rows) and then transpose the results (so that it goes back | 30 ; transposed rows) and then transpose the results (so that it goes back |
31 ; in normal/row positions). Then, we transform the columns by doing | 31 ; in normal/row positions). Then, we transform the columns by doing |
32 ; another SIMD column transform. | 32 ; another SIMD column transform. |
33 ; So, two passes of a transpose followed by a column transform. | 33 ; So, two passes of a transpose followed by a column transform. |
34 | 34 |
35 ; load the inputs into q8-q9, d16-d19 | 35 ; load the inputs into q8-q9, d16-d19 |
(...skipping 142 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... |
178 vqmovun.s16 d27, q9 | 178 vqmovun.s16 d27, q9 |
179 | 179 |
180 ; do the stores in reverse order with negative post-increment, by changing | 180 ; do the stores in reverse order with negative post-increment, by changing |
181 ; the sign of the stride | 181 ; the sign of the stride |
182 rsb r2, r2, #0 | 182 rsb r2, r2, #0 |
183 vst1.32 {d27[0]}, [r1], r2 | 183 vst1.32 {d27[0]}, [r1], r2 |
184 vst1.32 {d27[1]}, [r1], r2 | 184 vst1.32 {d27[1]}, [r1], r2 |
185 vst1.32 {d26[1]}, [r1], r2 | 185 vst1.32 {d26[1]}, [r1], r2 |
186 vst1.32 {d26[0]}, [r1] ; no post-increment | 186 vst1.32 {d26[0]}, [r1] ; no post-increment |
187 bx lr | 187 bx lr |
188 ENDP ; |vp9_idct4x4_16_add_neon| | 188 ENDP ; |vpx_idct4x4_16_add_neon| |
189 | 189 |
190 END | 190 END |
OLD | NEW |