|
808
|
1 /*
|
|
|
2 * Alpha optimized DSP utils
|
|
|
3 * Copyright (c) 2002 Falk Hueffner <falk@debian.org>
|
|
|
4 *
|
|
|
5 * This file is part of FFmpeg.
|
|
|
6 *
|
|
|
7 * FFmpeg is free software; you can redistribute it and/or
|
|
|
8 * modify it under the terms of the GNU Lesser General Public
|
|
|
9 * License as published by the Free Software Foundation; either
|
|
|
10 * version 2.1 of the License, or (at your option) any later version.
|
|
|
11 *
|
|
|
12 * FFmpeg is distributed in the hope that it will be useful,
|
|
|
13 * but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
|
|
15 * Lesser General Public License for more details.
|
|
|
16 *
|
|
|
17 * You should have received a copy of the GNU Lesser General Public
|
|
|
18 * License along with FFmpeg; if not, write to the Free Software
|
|
|
19 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
|
|
|
20 */
|
|
|
21
|
|
|
22 /*
|
|
|
23 * These functions are scheduled for pca56. They should work
|
|
|
24 * reasonably on ev6, though.
|
|
|
25 */
|
|
|
26
|
|
|
27 #include "regdef.h"
|
|
|
28
|
|
|
29 /* Some nicer register names. */
|
|
|
30 #define ta t10
|
|
|
31 #define tb t11
|
|
|
32 #define tc t12
|
|
|
33 #define td AT
|
|
|
34 /* Danger: these overlap with the argument list and the return value */
|
|
|
35 #define te a5
|
|
|
36 #define tf a4
|
|
|
37 #define tg a3
|
|
|
38 #define th v0
|
|
|
39
|
|
|
40 .set noat
|
|
|
41 .set noreorder
|
|
|
42 .arch pca56
|
|
|
43 .text
|
|
|
44
|
|
|
45 /************************************************************************
|
|
|
46 * void put_pixels_axp_asm(uint8_t *block, const uint8_t *pixels,
|
|
|
47 * int line_size, int h)
|
|
|
48 */
|
|
|
49 .align 6
|
|
|
50 .globl put_pixels_axp_asm
|
|
|
51 .ent put_pixels_axp_asm
|
|
|
52 put_pixels_axp_asm:
|
|
|
53 .frame sp, 0, ra
|
|
|
54 .prologue 0
|
|
|
55
|
|
|
56 #ifdef HAVE_GPROF
|
|
|
57 lda AT, _mcount
|
|
|
58 jsr AT, (AT), _mcount
|
|
|
59 #endif
|
|
|
60
|
|
|
61 and a1, 7, t0
|
|
|
62 beq t0, $aligned
|
|
|
63
|
|
|
64 .align 4
|
|
|
65 $unaligned:
|
|
|
66 ldq_u t0, 0(a1)
|
|
|
67 ldq_u t1, 8(a1)
|
|
|
68 addq a1, a2, a1
|
|
|
69 nop
|
|
|
70
|
|
|
71 ldq_u t2, 0(a1)
|
|
|
72 ldq_u t3, 8(a1)
|
|
|
73 addq a1, a2, a1
|
|
|
74 nop
|
|
|
75
|
|
|
76 ldq_u t4, 0(a1)
|
|
|
77 ldq_u t5, 8(a1)
|
|
|
78 addq a1, a2, a1
|
|
|
79 nop
|
|
|
80
|
|
|
81 ldq_u t6, 0(a1)
|
|
|
82 ldq_u t7, 8(a1)
|
|
|
83 extql t0, a1, t0
|
|
|
84 addq a1, a2, a1
|
|
|
85
|
|
|
86 extqh t1, a1, t1
|
|
|
87 addq a0, a2, t8
|
|
|
88 extql t2, a1, t2
|
|
|
89 addq t8, a2, t9
|
|
|
90
|
|
|
91 extqh t3, a1, t3
|
|
|
92 addq t9, a2, ta
|
|
|
93 extql t4, a1, t4
|
|
|
94 or t0, t1, t0
|
|
|
95
|
|
|
96 extqh t5, a1, t5
|
|
|
97 or t2, t3, t2
|
|
|
98 extql t6, a1, t6
|
|
|
99 or t4, t5, t4
|
|
|
100
|
|
|
101 extqh t7, a1, t7
|
|
|
102 or t6, t7, t6
|
|
|
103 stq t0, 0(a0)
|
|
|
104 stq t2, 0(t8)
|
|
|
105
|
|
|
106 stq t4, 0(t9)
|
|
|
107 subq a3, 4, a3
|
|
|
108 stq t6, 0(ta)
|
|
|
109 addq ta, a2, a0
|
|
|
110
|
|
|
111 bne a3, $unaligned
|
|
|
112 ret
|
|
|
113
|
|
|
114 .align 4
|
|
|
115 $aligned:
|
|
|
116 ldq t0, 0(a1)
|
|
|
117 addq a1, a2, a1
|
|
|
118 ldq t1, 0(a1)
|
|
|
119 addq a1, a2, a1
|
|
|
120
|
|
|
121 ldq t2, 0(a1)
|
|
|
122 addq a1, a2, a1
|
|
|
123 ldq t3, 0(a1)
|
|
|
124
|
|
|
125 addq a0, a2, t4
|
|
|
126 addq a1, a2, a1
|
|
|
127 addq t4, a2, t5
|
|
|
128 subq a3, 4, a3
|
|
|
129
|
|
|
130 stq t0, 0(a0)
|
|
|
131 addq t5, a2, t6
|
|
|
132 stq t1, 0(t4)
|
|
|
133 addq t6, a2, a0
|
|
|
134
|
|
|
135 stq t2, 0(t5)
|
|
|
136 stq t3, 0(t6)
|
|
|
137
|
|
|
138 bne a3, $aligned
|
|
|
139 ret
|
|
|
140 .end put_pixels_axp_asm
|
|
|
141
|
|
|
142 /************************************************************************
|
|
|
143 * void put_pixels_clamped_mvi_asm(const DCTELEM *block, uint8_t *pixels,
|
|
|
144 * int line_size)
|
|
|
145 */
|
|
|
146 .align 6
|
|
|
147 .globl put_pixels_clamped_mvi_asm
|
|
|
148 .ent put_pixels_clamped_mvi_asm
|
|
|
149 put_pixels_clamped_mvi_asm:
|
|
|
150 .frame sp, 0, ra
|
|
|
151 .prologue 0
|
|
|
152
|
|
|
153 #ifdef HAVE_GPROF
|
|
|
154 lda AT, _mcount
|
|
|
155 jsr AT, (AT), _mcount
|
|
|
156 #endif
|
|
|
157
|
|
|
158 lda t8, -1
|
|
|
159 lda t9, 8 # loop counter
|
|
|
160 zap t8, 0xaa, t8 # 00ff00ff00ff00ff
|
|
|
161
|
|
|
162 .align 4
|
|
|
163 1: ldq t0, 0(a0)
|
|
|
164 ldq t1, 8(a0)
|
|
|
165 ldq t2, 16(a0)
|
|
|
166 ldq t3, 24(a0)
|
|
|
167
|
|
|
168 maxsw4 t0, zero, t0
|
|
|
169 subq t9, 2, t9
|
|
|
170 maxsw4 t1, zero, t1
|
|
|
171 lda a0, 32(a0)
|
|
|
172
|
|
|
173 maxsw4 t2, zero, t2
|
|
|
174 addq a1, a2, ta
|
|
|
175 maxsw4 t3, zero, t3
|
|
|
176 minsw4 t0, t8, t0
|
|
|
177
|
|
|
178 minsw4 t1, t8, t1
|
|
|
179 minsw4 t2, t8, t2
|
|
|
180 minsw4 t3, t8, t3
|
|
|
181 pkwb t0, t0
|
|
|
182
|
|
|
183 pkwb t1, t1
|
|
|
184 pkwb t2, t2
|
|
|
185 pkwb t3, t3
|
|
|
186 stl t0, 0(a1)
|
|
|
187
|
|
|
188 stl t1, 4(a1)
|
|
|
189 addq ta, a2, a1
|
|
|
190 stl t2, 0(ta)
|
|
|
191 stl t3, 4(ta)
|
|
|
192
|
|
|
193 bne t9, 1b
|
|
|
194 ret
|
|
|
195 .end put_pixels_clamped_mvi_asm
|
|
|
196
|
|
|
197 /************************************************************************
|
|
|
198 * void add_pixels_clamped_mvi_asm(const DCTELEM *block, uint8_t *pixels,
|
|
|
199 * int line_size)
|
|
|
200 */
|
|
|
201 .align 6
|
|
|
202 .globl add_pixels_clamped_mvi_asm
|
|
|
203 .ent add_pixels_clamped_mvi_asm
|
|
|
204 add_pixels_clamped_mvi_asm:
|
|
|
205 .frame sp, 0, ra
|
|
|
206 .prologue 0
|
|
|
207
|
|
|
208 #ifdef HAVE_GPROF
|
|
|
209 lda AT, _mcount
|
|
|
210 jsr AT, (AT), _mcount
|
|
|
211 #endif
|
|
|
212
|
|
|
213 lda t1, -1
|
|
|
214 lda th, 8
|
|
|
215 zap t1, 0x33, tg
|
|
|
216 nop
|
|
|
217
|
|
|
218 srl tg, 1, t0
|
|
|
219 xor tg, t0, tg # 0x8000800080008000
|
|
|
220 zap t1, 0xaa, tf # 0x00ff00ff00ff00ff
|
|
|
221
|
|
|
222 .align 4
|
|
|
223 1: ldl t1, 0(a1) # pix0 (try to hit cache line soon)
|
|
|
224 ldl t4, 4(a1) # pix1
|
|
|
225 addq a1, a2, te # pixels += line_size
|
|
|
226 ldq t0, 0(a0) # shorts0
|
|
|
227
|
|
|
228 ldl t7, 0(te) # pix2 (try to hit cache line soon)
|
|
|
229 ldl ta, 4(te) # pix3
|
|
|
230 ldq t3, 8(a0) # shorts1
|
|
|
231 ldq t6, 16(a0) # shorts2
|
|
|
232
|
|
|
233 ldq t9, 24(a0) # shorts3
|
|
|
234 unpkbw t1, t1 # 0 0 (quarter/op no.)
|
|
|
235 and t0, tg, t2 # 0 1
|
|
|
236 unpkbw t4, t4 # 1 0
|
|
|
237
|
|
|
238 bic t0, tg, t0 # 0 2
|
|
|
239 unpkbw t7, t7 # 2 0
|
|
|
240 and t3, tg, t5 # 1 1
|
|
|
241 addq t0, t1, t0 # 0 3
|
|
|
242
|
|
|
243 xor t0, t2, t0 # 0 4
|
|
|
244 unpkbw ta, ta # 3 0
|
|
|
245 and t6, tg, t8 # 2 1
|
|
|
246 maxsw4 t0, zero, t0 # 0 5
|
|
|
247
|
|
|
248 bic t3, tg, t3 # 1 2
|
|
|
249 bic t6, tg, t6 # 2 2
|
|
|
250 minsw4 t0, tf, t0 # 0 6
|
|
|
251 addq t3, t4, t3 # 1 3
|
|
|
252
|
|
|
253 pkwb t0, t0 # 0 7
|
|
|
254 xor t3, t5, t3 # 1 4
|
|
|
255 maxsw4 t3, zero, t3 # 1 5
|
|
|
256 addq t6, t7, t6 # 2 3
|
|
|
257
|
|
|
258 xor t6, t8, t6 # 2 4
|
|
|
259 and t9, tg, tb # 3 1
|
|
|
260 minsw4 t3, tf, t3 # 1 6
|
|
|
261 bic t9, tg, t9 # 3 2
|
|
|
262
|
|
|
263 maxsw4 t6, zero, t6 # 2 5
|
|
|
264 addq t9, ta, t9 # 3 3
|
|
|
265 stl t0, 0(a1) # 0 8
|
|
|
266 minsw4 t6, tf, t6 # 2 6
|
|
|
267
|
|
|
268 xor t9, tb, t9 # 3 4
|
|
|
269 maxsw4 t9, zero, t9 # 3 5
|
|
|
270 lda a0, 32(a0) # block += 16;
|
|
|
271 pkwb t3, t3 # 1 7
|
|
|
272
|
|
|
273 minsw4 t9, tf, t9 # 3 6
|
|
|
274 subq th, 2, th
|
|
|
275 pkwb t6, t6 # 2 7
|
|
|
276 pkwb t9, t9 # 3 7
|
|
|
277
|
|
|
278 stl t3, 4(a1) # 1 8
|
|
|
279 addq te, a2, a1 # pixels += line_size
|
|
|
280 stl t6, 0(te) # 2 8
|
|
|
281 stl t9, 4(te) # 3 8
|
|
|
282
|
|
|
283 bne th, 1b
|
|
|
284 ret
|
|
|
285 .end add_pixels_clamped_mvi_asm
|
