2 * Copyright (c) 2002 Brian Foley
3 * Copyright (c) 2002 Dieter Shirley
4 * Copyright (c) 2003-2004 Romain Dolbeau <romain@dolbeau.org>
6 * This file is part of FFmpeg.
8 * FFmpeg is free software; you can redistribute it and/or
9 * modify it under the terms of the GNU Lesser General Public
10 * License as published by the Free Software Foundation; either
11 * version 2.1 of the License, or (at your option) any later version.
13 * FFmpeg is distributed in the hope that it will be useful,
14 * but WITHOUT ANY WARRANTY; without even the implied warranty of
15 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
16 * Lesser General Public License for more details.
18 * You should have received a copy of the GNU Lesser General Public
19 * License along with FFmpeg; if not, write to the Free Software
20 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
23 #ifndef DSPUTIL_ALTIVEC_H
24 #define DSPUTIL_ALTIVEC_H
26 #include "dsputil_ppc.h"
30 extern int has_altivec(void);
32 void put_pixels16_altivec(uint8_t *block, const uint8_t *pixels, int line_size, int h);
34 void avg_pixels16_altivec(uint8_t *block, const uint8_t *pixels, int line_size, int h);
36 // used to build registers permutation vectors (vcprm)
37 // the 's' are for words in the _s_econd vector
38 #define WORD_0 0x00,0x01,0x02,0x03
39 #define WORD_1 0x04,0x05,0x06,0x07
40 #define WORD_2 0x08,0x09,0x0a,0x0b
41 #define WORD_3 0x0c,0x0d,0x0e,0x0f
42 #define WORD_s0 0x10,0x11,0x12,0x13
43 #define WORD_s1 0x14,0x15,0x16,0x17
44 #define WORD_s2 0x18,0x19,0x1a,0x1b
45 #define WORD_s3 0x1c,0x1d,0x1e,0x1f
48 #define vcprm(a,b,c,d) (const vector unsigned char)(WORD_ ## a, WORD_ ## b, WORD_ ## c, WORD_ ## d)
50 #define vcprm(a,b,c,d) (const vector unsigned char){WORD_ ## a, WORD_ ## b, WORD_ ## c, WORD_ ## d}
53 // vcprmle is used to keep the same index as in the SSE version.
54 // it's the same as vcprm, with the index inversed
55 // ('le' is Little Endian)
56 #define vcprmle(a,b,c,d) vcprm(d,c,b,a)
58 // used to build inverse/identity vectors (vcii)
59 // n is _n_egative, p is _p_ositive
65 #define vcii(a,b,c,d) (const vector float)(FLOAT_ ## a, FLOAT_ ## b, FLOAT_ ## c, FLOAT_ ## d)
67 #define vcii(a,b,c,d) (const vector float){FLOAT_ ## a, FLOAT_ ## b, FLOAT_ ## c, FLOAT_ ## d}
70 // Transpose 8x8 matrix of 16-bit elements (in-place)
71 #define TRANSPOSE8(a,b,c,d,e,f,g,h) \
73 vector signed short A1, B1, C1, D1, E1, F1, G1, H1; \
74 vector signed short A2, B2, C2, D2, E2, F2, G2, H2; \
76 A1 = vec_mergeh (a, e); \
77 B1 = vec_mergel (a, e); \
78 C1 = vec_mergeh (b, f); \
79 D1 = vec_mergel (b, f); \
80 E1 = vec_mergeh (c, g); \
81 F1 = vec_mergel (c, g); \
82 G1 = vec_mergeh (d, h); \
83 H1 = vec_mergel (d, h); \
85 A2 = vec_mergeh (A1, E1); \
86 B2 = vec_mergel (A1, E1); \
87 C2 = vec_mergeh (B1, F1); \
88 D2 = vec_mergel (B1, F1); \
89 E2 = vec_mergeh (C1, G1); \
90 F2 = vec_mergel (C1, G1); \
91 G2 = vec_mergeh (D1, H1); \
92 H2 = vec_mergel (D1, H1); \
94 a = vec_mergeh (A2, E2); \
95 b = vec_mergel (A2, E2); \
96 c = vec_mergeh (B2, F2); \
97 d = vec_mergel (B2, F2); \
98 e = vec_mergeh (C2, G2); \
99 f = vec_mergel (C2, G2); \
100 g = vec_mergeh (D2, H2); \
101 h = vec_mergel (D2, H2); \
105 /** \brief loads unaligned vector \a *src with offset \a offset
107 static inline vector unsigned char unaligned_load(int offset, uint8_t *src)
109 register vector unsigned char first = vec_ld(offset, src);
110 register vector unsigned char second = vec_ld(offset+15, src);
111 register vector unsigned char mask = vec_lvsl(offset, src);
112 return vec_perm(first, second, mask);
115 #endif /* HAVE_ALTIVEC */
117 #endif /* DSPUTIL_ALTIVEC_H */