libavcodec/ppc/util_altivec.h

   1 /*
   2  * This file is part of FFmpeg.
   3  *
   4  * FFmpeg is free software; you can redistribute it and/or
   5  * modify it under the terms of the GNU Lesser General Public
   6  * License as published by the Free Software Foundation; either
   7  * version 2.1 of the License, or (at your option) any later version.
   8  *
   9  * FFmpeg is distributed in the hope that it will be useful,
  10  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  11  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  12  * Lesser General Public License for more details.
  13  *
  14  * You should have received a copy of the GNU Lesser General Public
  15  * License along with FFmpeg; if not, write to the Free Software
  16  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
  17  */
  18
  19 /**
  20  * @file util_altivec.h
  21  * Contains misc utility macros and inline functions
  22  */
  23
  24 #ifndef FFMPEG_UTIL_ALTIVEC_H
  25 #define FFMPEG_UTIL_ALTIVEC_H
  26
  27 #include <stdint.h>
  28
  29 #include "config.h"
  30
  31 #ifdef HAVE_ALTIVEC_H
  32 #include <altivec.h>
  33 #endif
  34
  35 // used to build registers permutation vectors (vcprm)
  36 // the 's' are for words in the _s_econd vector
  37 #define WORD_0 0x00,0x01,0x02,0x03
  38 #define WORD_1 0x04,0x05,0x06,0x07
  39 #define WORD_2 0x08,0x09,0x0a,0x0b
  40 #define WORD_3 0x0c,0x0d,0x0e,0x0f
  41 #define WORD_s0 0x10,0x11,0x12,0x13
  42 #define WORD_s1 0x14,0x15,0x16,0x17
  43 #define WORD_s2 0x18,0x19,0x1a,0x1b
  44 #define WORD_s3 0x1c,0x1d,0x1e,0x1f
  45
  46 #define vcprm(a,b,c,d) (const vector unsigned char){WORD_ ## a, WORD_ ## b, WORD_ ## c, WORD_ ## d}
  47 #define vcii(a,b,c,d) (const vector float){FLOAT_ ## a, FLOAT_ ## b, FLOAT_ ## c, FLOAT_ ## d}
  48
  49 // vcprmle is used to keep the same index as in the SSE version.
  50 // it's the same as vcprm, with the index inversed
  51 // ('le' is Little Endian)
  52 #define vcprmle(a,b,c,d) vcprm(d,c,b,a)
  53
  54 // used to build inverse/identity vectors (vcii)
  55 // n is _n_egative, p is _p_ositive
  56 #define FLOAT_n -1.
  57 #define FLOAT_p 1.
  58
  59
  60 // Transpose 8x8 matrix of 16-bit elements (in-place)
  61 #define TRANSPOSE8(a,b,c,d,e,f,g,h) \
  62 do { \
  63     vector signed short A1, B1, C1, D1, E1, F1, G1, H1; \
  64     vector signed short A2, B2, C2, D2, E2, F2, G2, H2; \
  65  \
  66     A1 = vec_mergeh (a, e); \
  67     B1 = vec_mergel (a, e); \
  68     C1 = vec_mergeh (b, f); \
  69     D1 = vec_mergel (b, f); \
  70     E1 = vec_mergeh (c, g); \
  71     F1 = vec_mergel (c, g); \
  72     G1 = vec_mergeh (d, h); \
  73     H1 = vec_mergel (d, h); \
  74  \
  75     A2 = vec_mergeh (A1, E1); \
  76     B2 = vec_mergel (A1, E1); \
  77     C2 = vec_mergeh (B1, F1); \
  78     D2 = vec_mergel (B1, F1); \
  79     E2 = vec_mergeh (C1, G1); \
  80     F2 = vec_mergel (C1, G1); \
  81     G2 = vec_mergeh (D1, H1); \
  82     H2 = vec_mergel (D1, H1); \
  83  \
  84     a = vec_mergeh (A2, E2); \
  85     b = vec_mergel (A2, E2); \
  86     c = vec_mergeh (B2, F2); \
  87     d = vec_mergel (B2, F2); \
  88     e = vec_mergeh (C2, G2); \
  89     f = vec_mergel (C2, G2); \
  90     g = vec_mergeh (D2, H2); \
  91     h = vec_mergel (D2, H2); \
  92 } while (0)
  93
  94
  95 /** \brief loads unaligned vector \a *src with offset \a offset
  96     and returns it */
  97 static inline vector unsigned char unaligned_load(int offset, uint8_t *src)
  98 {
  99     register vector unsigned char first = vec_ld(offset, src);
 100     register vector unsigned char second = vec_ld(offset+15, src);
 101     register vector unsigned char mask = vec_lvsl(offset, src);
 102     return vec_perm(first, second, mask);
 103 }
 104
 105 #endif /* FFMPEG_UTIL_ALTIVEC_H */