1 /*
2 * Loongson SIMD optimized xvid idct
3 *
4 * Copyright (c) 2015 Loongson Technology Corporation Limited
5 * Copyright (c) 2015 Zhou Xiaoyong <zhouxiaoyong@loongson.cn>
6 *
7 * This file is part of FFmpeg.
8 *
9 * FFmpeg is free software; you can redistribute it and/or
10 * modify it under the terms of the GNU Lesser General Public
11 * License as published by the Free Software Foundation; either
12 * version 2.1 of the License, or (at your option) any later version.
13 *
14 * FFmpeg is distributed in the hope that it will be useful,
15 * but WITHOUT ANY WARRANTY; without even the implied warranty of
16 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
17 * Lesser General Public License for more details.
18 *
19 * You should have received a copy of the GNU Lesser General Public
20 * License along with FFmpeg; if not, write to the Free Software
21 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
22 */
23
24 #include "libavutil/mem_internal.h"
25
26 #include "idctdsp_mips.h"
27 #include "xvididct_mips.h"
28
29 #define BITS_INV_ACC 5 // 4 or 5 for IEEE
30 #define SHIFT_INV_ROW (16 - BITS_INV_ACC) //11
31 #define SHIFT_INV_COL (1 + BITS_INV_ACC) //6
32 #define RND_INV_ROW (1024 * (6 - BITS_INV_ACC))
33 #define RND_INV_COL (16 * (BITS_INV_ACC - 3))
34 #define RND_INV_CORR (RND_INV_COL - 1)
35
36 #define BITS_FRW_ACC 3 // 2 or 3 for accuracy
37 #define SHIFT_FRW_COL BITS_FRW_ACC
38 #define SHIFT_FRW_ROW (BITS_FRW_ACC + 17)
39 #define RND_FRW_ROW (262144*(BITS_FRW_ACC - 1))
40
41 DECLARE_ALIGNED(8, static const int16_t, tg_1_16)[4*4] = {
42 13036, 13036, 13036, 13036, // tg * (2<<16) + 0.5
43 27146, 27146, 27146, 27146, // tg * (2<<16) + 0.5
44 -21746,-21746,-21746,-21746, // tg * (2<<16) + 0.5
45 23170, 23170, 23170, 23170 // cos * (2<<15) + 0.5
46 };
47
48 DECLARE_ALIGNED(8, static const int32_t, rounder_0)[2*8] = {
49 65536,65536,
50 3597, 3597,
51 2260, 2260,
52 1203, 1203,
53 0, 0,
54 120, 120,
55 512, 512,
56 512, 512
57 };
58
59 DECLARE_ALIGNED(8, static const int16_t, tab_i_04_mmi)[32*4] = {
60 16384, 21407, 16384, 8867, // w05 w04 w01 w00
61 16384, 8867,-16384,-21407, // w07 w06 w03 w02
62 16384, -8867, 16384,-21407, // w13 w12 w09 w08
63 -16384, 21407, 16384, -8867, // w15 w14 w11 w10
64 22725, 19266, 19266, -4520, // w21 w20 w17 w16
65 12873, 4520,-22725,-12873, // w23 w22 w19 w18
66 12873,-22725, 4520,-12873, // w29 w28 w25 w24
67 4520, 19266, 19266,-22725, // w31 w30 w27 w26
68
69 22725, 29692, 22725, 12299, // w05 w04 w01 w00
70 22725, 12299,-22725,-29692, // w07 w06 w03 w02
71 22725,-12299, 22725,-29692, // w13 w12 w09 w08
72 -22725, 29692, 22725,-12299, // w15 w14 w11 w10
73 31521, 26722, 26722, -6270, // w21 w20 w17 w16
74 17855, 6270,-31521,-17855, // w23 w22 w19 w18
75 17855,-31521, 6270,-17855, // w29 w28 w25 w24
76 6270, 26722, 26722,-31521, // w31 w30 w27 w26
77
78 21407, 27969, 21407, 11585, // w05 w04 w01 w00
79 21407, 11585,-21407,-27969, // w07 w06 w03 w02
80 21407,-11585, 21407,-27969, // w13 w12 w09 w08
81 -21407, 27969, 21407,-11585, // w15 w14 w11 w10
82 29692, 25172, 25172, -5906, // w21 w20 w17 w16
83 16819, 5906,-29692,-16819, // w23 w22 w19 w18
84 16819,-29692, 5906,-16819, // w29 w28 w25 w24
85 5906, 25172, 25172,-29692, // w31 w30 w27 w26
86
87 19266, 25172, 19266, 10426, // w05 w04 w01 w00
88 19266, 10426,-19266,-25172, // w07 w06 w03 w02
89 19266,-10426, 19266,-25172, // w13 w12 w09 w08
90 -19266, 25172, 19266,-10426, // w15 w14 w11 w10
91 26722, 22654, 22654, -5315, // w21 w20 w17 w16
92 15137, 5315,-26722,-15137, // w23 w22 w19 w18
93 15137,-26722, 5315,-15137, // w29 w28 w25 w24
94 5315, 22654, 22654,-26722, // w31 w30 w27 w26
95 };
96
97 #define DCT_8_INV_ROW_MMI(A1,A2,A3,A4) \
98 "dli $10, 0x88 \n\t" \
99 "ldc1 $f4, "#A1" \n\t" /* 0; x3 x2 x1 x0 */\
100 "dmtc1 $10, $f16 \n\t" \
101 "ldc1 $f10, 8+"#A1" \n\t" /* 1; x7 x6 x5 x4 */\
102 "ldc1 $f6, "#A3" \n\t" /* 3; w05 w04 w01 w00 */\
103 "pshufh $f0, $f4, $f16 \n\t" /* x2 x0 x2 x0 */\
104 "ldc1 $f8, 8+"#A3" \n\t" /* 4; w07 w06 w03 w02 */\
105 "ldc1 $f12, 32+"#A3" \n\t" /* 6; w21 w20 w17 w16 */\
106 "pmaddhw $f6, $f6, $f0 \n\t" /* x2*w05+x0*w04 x2*w01+x0*w00 */\
107 "dli $10, 0xdd \n\t" \
108 "pshufh $f2, $f10, $f16 \n\t" /* x6 x4 x6 x4 */\
109 "dmtc1 $10, $f16 \n\t" \
110 "pmaddhw $f8, $f8, $f2 \n\t" /* x6*w07+x4*w06 x6*w03+x4*w02 */\
111 "ldc1 $f14, 40+"#A3" \n\t" /* 7; w23 w22 w19 w18 */\
112 "pshufh $f4, $f4, $f16 \n\t" /* x3 x1 x3 x1 */\
113 "pmaddhw $f12, $f12, $f4 \n\t" /* x3*w21+x1*w20 x3*w17+x1*w16 */\
114 "pshufh $f10, $f10, $f16 \n\t" /* x7 x5 x7 x5 */\
115 "ldc1 $f18, "#A4" \n\t" \
116 "pmaddhw $f14, $f14, $f10 \n\t" /* x7*w23+x5*w22 x7*w19+x5*w18 */\
117 "paddw $f6, $f6, $f18 \n\t" /* +%4 */\
118 "ldc1 $f16, 16+"#A3" \n\t" \
119 "pmaddhw $f0, $f0, $f16 \n\t" /* x2*w13+x0*w12 x2*w09+x0*w08 */\
120 "ldc1 $f16, 24+"#A3" \n\t" \
121 "paddw $f6, $f6, $f8 \n\t" /* 4; a1=sum(even1) a0=sum(even0) */\
122 "pmaddhw $f2, $f2, $f16 \n\t" /* x6*w15+x4*w14 x6*w11+x4*w10 */\
123 "ldc1 $f16, 48+"#A3" \n\t" \
124 "pmaddhw $f4, $f4, $f16 \n\t" /* x3*w29+x1*w28 x3*w25+x1*w24 */\
125 "ldc1 $f16, 56+"#A3" \n\t" \
126 "paddw $f12, $f12, $f14 \n\t" /* 7; b1=sum(odd1) b0=sum(odd0) */\
127 "dli $10, 11 \n\t" \
128 "pmaddhw $f10, $f10, $f16 \n\t" /* x7*w31+x5*w30 x7*w27+x5*w26 */\
129 "dmtc1 $10, $f16 \n\t" \
130 "psubw $f8, $f6, $f12 \n\t" /* 6; a1-b1 a0-b0 */\
131 "paddw $f6, $f6, $f12 \n\t" /* a1+b1 a0+b0 */\
132 "paddw $f0, $f0, $f18 \n\t" /* +%4 */\
133 "psraw $f6, $f6, $f16 \n\t" /* y1=a1+b1 y0=a0+b0 */\
134 "paddw $f0, $f0, $f2 \n\t" /* 1; a3=sum(even3) a2=sum(even2) */\
135 "paddw $f4, $f4, $f10 \n\t" /* 5; b3=sum(odd3) b2=sum(odd2) */\
136 "psraw $f8, $f8, $f16 \n\t" /* y6=a1-b1 y7=a0-b0 */\
137 "psubw $f14, $f0, $f4 \n\t" /* 2; a3-b3 a2-b2 */\
138 "paddw $f0, $f0, $f4 \n\t" /* a3+b3 a2+b2 */\
139 "psraw $f0, $f0, $f16 \n\t" /* y3=a3+b3 y2=a2+b2 */\
140 "psraw $f14, $f14, $f16 \n\t" /* y4=a3-b3 y5=a2-b2 */\
141 "dli $10, 0xb1 \n\t" \
142 "packsswh $f6, $f6, $f0 \n\t" /* 0; y3 y2 y1 y0 */\
143 "dmtc1 $10, $f16 \n\t" \
144 "packsswh $f14, $f14, $f8 \n\t" /* 4; y6 y7 y4 y5 */\
145 "sdc1 $f6, "#A2" \n\t" /* 3; save y3 y2 y1 y0 */\
146 "pshufh $f14, $f14, $f16 \n\t" /* y7 y6 y5 y4 */\
147 "sdc1 $f14, 8+"#A2" \n\t" /* 7; save y7 y6 y5 y4 */\
148
149
150 #define DCT_8_INV_COL(A1,A2) \
151 "ldc1 $f2, 2*8(%3) \n\t" \
152 "ldc1 $f6, 16*3+"#A1" \n\t" \
153 "ldc1 $f10, 16*5+"#A1" \n\t" \
154 "pmulhh $f0, $f2, $f6 \n\t" /* x3*(tg_3_16-1) */\
155 "ldc1 $f4, 0(%3) \n\t" \
156 "pmulhh $f2, $f2, $f10 \n\t" /* x5*(tg_3_16-1) */\
157 "ldc1 $f14, 16*7+"#A1" \n\t" \
158 "ldc1 $f12, 16*1+"#A1" \n\t" \
159 "pmulhh $f8, $f4, $f14 \n\t" /* x7*tg_1_16 */\
160 "paddsh $f0, $f0, $f6 \n\t" /* x3*tg_3_16 */\
161 "pmulhh $f4, $f4, $f12 \n\t" /* x1*tg_1_16 */\
162 "paddsh $f2, $f2, $f6 \n\t" /* x3+x5*(tg_3_16-1) */\
163 "psubsh $f0, $f0, $f10 \n\t" /* x3*tg_3_16-x5 = tm35 */\
164 "ldc1 $f6, 3*8(%3) \n\t" \
165 "paddsh $f2, $f2, $f10 \n\t" /* x3+x5*tg_3_16 = tp35 */\
166 "paddsh $f8, $f8, $f12 \n\t" /* x1+tg_1_16*x7 = tp17 */\
167 "psubsh $f4, $f4, $f14 \n\t" /* x1*tg_1_16-x7 = tm17 */\
168 "paddsh $f10, $f8, $f2 \n\t" /* tp17+tp35 = b0 */\
169 "psubsh $f12, $f4, $f0 \n\t" /* tm17-tm35 = b3 */\
170 "psubsh $f8, $f8, $f2 \n\t" /* tp17-tp35 = t1 */\
171 "paddsh $f4, $f4, $f0 \n\t" /* tm17+tm35 = t2 */\
172 "ldc1 $f14, 1*8(%3) \n\t" \
173 "sdc1 $f10, 3*16+"#A2" \n\t" /* save b0 */\
174 "paddsh $f2, $f8, $f4 \n\t" /* t1+t2 */\
175 "sdc1 $f12, 5*16+"#A2" \n\t" /* save b3 */\
176 "psubsh $f8, $f8, $f4 \n\t" /* t1-t2 */\
177 "ldc1 $f10, 2*16+"#A1" \n\t" \
178 "ldc1 $f12, 6*16+"#A1" \n\t" \
179 "pmulhh $f0, $f14, $f10 \n\t" /* x2*tg_2_16 */\
180 "pmulhh $f14, $f14, $f12 \n\t" /* x6*tg_2_16 */\
181 "pmulhh $f2, $f2, $f6 \n\t" /* ocos_4_16*(t1+t2) = b1/2 */\
182 "ldc1 $f4, 0*16+"#A1" \n\t" \
183 "pmulhh $f8, $f8, $f6 \n\t" /* ocos_4_16*(t1-t2) = b2/2 */\
184 "psubsh $f0, $f0, $f12 \n\t" /* t2*tg_2_16-x6 = tm26 */\
185 "ldc1 $f12, 4*16+"#A1" \n\t" \
186 "paddsh $f14, $f14, $f10 \n\t" /* x2+x6*tg_2_16 = tp26 */\
187 "psubsh $f6, $f4, $f12 \n\t" /* x0-x4 = tm04 */\
188 "paddsh $f4, $f4, $f12 \n\t" /* x0+x4 = tp04 */\
189 "paddsh $f10, $f4, $f14 \n\t" /* tp04+tp26 = a0 */\
190 "psubsh $f12, $f6, $f0 \n\t" /* tm04-tm26 = a2 */\
191 "psubsh $f4, $f4, $f14 \n\t" /* tp04-tp26 = a3 */\
192 "paddsh $f6, $f6, $f0 \n\t" /* tm04+tm26 = a1 */\
193 "paddsh $f2, $f2, $f2 \n\t" /* b1 */\
194 "paddsh $f8, $f8, $f8 \n\t" /* b2 */\
195 "psubsh $f14, $f6, $f2 \n\t" /* a1-b1 */\
196 "dli $10, 6 \n\t" \
197 "paddsh $f6, $f6, $f2 \n\t" /* a1+b1 */\
198 "dmtc1 $10, $f16 \n\t" \
199 "psubsh $f0, $f12, $f8 \n\t" /* a2-b2 */\
200 "paddsh $f12, $f12, $f8 \n\t" /* a2+b2 */\
201 "psrah $f6, $f6, $f16 \n\t" /* dst1 */\
202 "psrah $f12, $f12, $f16 \n\t" /* dst2 */\
203 "ldc1 $f2, 3*16+"#A2" \n\t" /* load b0 */\
204 "psrah $f14, $f14, $f16 \n\t" /* dst6 */\
205 "psrah $f0, $f0, $f16 \n\t" /* dst5 */\
206 "sdc1 $f6, 1*16+"#A2" \n\t" \
207 "psubsh $f8, $f10, $f2 \n\t" /* a0-b0 */\
208 "paddsh $f10, $f10, $f2 \n\t" /* a0+b0 */\
209 "sdc1 $f12, 2*16+"#A2" \n\t" \
210 "ldc1 $f6, 5*16+"#A2" \n\t" /* load b3 */\
211 "psrah $f10, $f10, $f16 \n\t" /* dst0 */\
212 "psrah $f8, $f8, $f16 \n\t" /* dst7 */\
213 "sdc1 $f0, 5*16+"#A2" \n\t" \
214 "psubsh $f12, $f4, $f6 \n\t" /* a3-b3 */\
215 "paddsh $f4, $f4, $f6 \n\t" /* a3+b3 */\
216 "sdc1 $f14, 6*16+"#A2" \n\t" \
217 "sdc1 $f10, 0*16+"#A2" \n\t" \
218 "psrah $f4, $f4, $f16 \n\t" /* dst3 */\
219 "sdc1 $f8, 7*16+"#A2" \n\t" \
220 "psrah $f12, $f12, $f16 \n\t" /* dst4 */\
221 "sdc1 $f4, 3*16+"#A2" \n\t" \
222 "sdc1 $f12, 4*16+"#A2" \n\t" \
223
224
ff_xvid_idct_mmi(int16_t * block)225 void ff_xvid_idct_mmi(int16_t *block)
226 {
227 __asm__ volatile (
228 //# Process each row
229 DCT_8_INV_ROW_MMI(0*16(%0), 0*16(%0), 64*0(%2), 8*0(%1))
230 DCT_8_INV_ROW_MMI(1*16(%0), 1*16(%0), 64*1(%2), 8*1(%1))
231 DCT_8_INV_ROW_MMI(2*16(%0), 2*16(%0), 64*2(%2), 8*2(%1))
232 DCT_8_INV_ROW_MMI(3*16(%0), 3*16(%0), 64*3(%2), 8*3(%1))
233 DCT_8_INV_ROW_MMI(4*16(%0), 4*16(%0), 64*0(%2), 8*4(%1))
234 DCT_8_INV_ROW_MMI(5*16(%0), 5*16(%0), 64*3(%2), 8*5(%1))
235 DCT_8_INV_ROW_MMI(6*16(%0), 6*16(%0), 64*2(%2), 8*6(%1))
236 DCT_8_INV_ROW_MMI(7*16(%0), 7*16(%0), 64*1(%2), 8*7(%1))
237 //# Process the columns (4 at a time)
238 DCT_8_INV_COL(0(%0), 0(%0))
239 DCT_8_INV_COL(8(%0), 8(%0))
240 ::"r"(block),"r"(rounder_0),"r"(tab_i_04_mmi),"r"(tg_1_16)
241 : "$10"
242 );
243 }
244
ff_xvid_idct_put_mmi(uint8_t * dest,ptrdiff_t line_size,int16_t * block)245 void ff_xvid_idct_put_mmi(uint8_t *dest, ptrdiff_t line_size, int16_t *block)
246 {
247 ff_xvid_idct_mmi(block);
248 ff_put_pixels_clamped_mmi(block, dest, line_size);
249 }
250
ff_xvid_idct_add_mmi(uint8_t * dest,ptrdiff_t line_size,int16_t * block)251 void ff_xvid_idct_add_mmi(uint8_t *dest, ptrdiff_t line_size, int16_t *block)
252 {
253 ff_xvid_idct_mmi(block);
254 ff_add_pixels_clamped_mmi(block, dest, line_size);
255 }
256