• Main Page
  • Related Pages
  • Modules
  • Data Structures
  • Files
  • File List
  • Globals

libavcodec/ppc/idct_altivec.c

Go to the documentation of this file.
00001 /*
00002  * Copyright (c) 2001 Michel Lespinasse
00003  *
00004  * This file is part of FFmpeg.
00005  *
00006  * FFmpeg is free software; you can redistribute it and/or
00007  * modify it under the terms of the GNU Lesser General Public
00008  * License as published by the Free Software Foundation; either
00009  * version 2.1 of the License, or (at your option) any later version.
00010  *
00011  * FFmpeg is distributed in the hope that it will be useful,
00012  * but WITHOUT ANY WARRANTY; without even the implied warranty of
00013  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
00014  * Lesser General Public License for more details.
00015  *
00016  * You should have received a copy of the GNU Lesser General Public
00017  * License along with FFmpeg; if not, write to the Free Software
00018  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
00019  */
00020 
00021 /*
00022  * NOTE: This code is based on GPL code from the libmpeg2 project.  The
00023  * author, Michel Lespinasses, has given explicit permission to release
00024  * under LGPL as part of ffmpeg.
00025  */
00026 
00027 /*
00028  * FFMpeg integration by Dieter Shirley
00029  *
00030  * This file is a direct copy of the altivec idct module from the libmpeg2
00031  * project.  I've deleted all of the libmpeg2 specific code, renamed the functions and
00032  * re-ordered the function parameters.  The only change to the IDCT function
00033  * itself was to factor out the partial transposition, and to perform a full
00034  * transpose at the end of the function.
00035  */
00036 
00037 
00038 #include <stdlib.h>                                      /* malloc(), free() */
00039 #include <string.h>
00040 #include "libavcodec/dsputil.h"
00041 
00042 #include "gcc_fixes.h"
00043 #include "types_altivec.h"
00044 #include "dsputil_ppc.h"
00045 
00046 #define IDCT_HALF                                       \
00047     /* 1st stage */                                     \
00048     t1 = vec_mradds (a1, vx7, vx1 );                    \
00049     t8 = vec_mradds (a1, vx1, vec_subs (zero, vx7));    \
00050     t7 = vec_mradds (a2, vx5, vx3);                     \
00051     t3 = vec_mradds (ma2, vx3, vx5);                    \
00052                                                         \
00053     /* 2nd stage */                                     \
00054     t5 = vec_adds (vx0, vx4);                           \
00055     t0 = vec_subs (vx0, vx4);                           \
00056     t2 = vec_mradds (a0, vx6, vx2);                     \
00057     t4 = vec_mradds (a0, vx2, vec_subs (zero, vx6));    \
00058     t6 = vec_adds (t8, t3);                             \
00059     t3 = vec_subs (t8, t3);                             \
00060     t8 = vec_subs (t1, t7);                             \
00061     t1 = vec_adds (t1, t7);                             \
00062                                                         \
00063     /* 3rd stage */                                     \
00064     t7 = vec_adds (t5, t2);                             \
00065     t2 = vec_subs (t5, t2);                             \
00066     t5 = vec_adds (t0, t4);                             \
00067     t0 = vec_subs (t0, t4);                             \
00068     t4 = vec_subs (t8, t3);                             \
00069     t3 = vec_adds (t8, t3);                             \
00070                                                         \
00071     /* 4th stage */                                     \
00072     vy0 = vec_adds (t7, t1);                            \
00073     vy7 = vec_subs (t7, t1);                            \
00074     vy1 = vec_mradds (c4, t3, t5);                      \
00075     vy6 = vec_mradds (mc4, t3, t5);                     \
00076     vy2 = vec_mradds (c4, t4, t0);                      \
00077     vy5 = vec_mradds (mc4, t4, t0);                     \
00078     vy3 = vec_adds (t2, t6);                            \
00079     vy4 = vec_subs (t2, t6);
00080 
00081 
00082 #define IDCT                                                            \
00083     vec_s16 vx0, vx1, vx2, vx3, vx4, vx5, vx6, vx7;                \
00084     vec_s16 vy0, vy1, vy2, vy3, vy4, vy5, vy6, vy7;                \
00085     vec_s16 a0, a1, a2, ma2, c4, mc4, zero, bias;                  \
00086     vec_s16 t0, t1, t2, t3, t4, t5, t6, t7, t8;                    \
00087     vec_u16 shift;                                                 \
00088                                                                         \
00089     c4 = vec_splat (constants[0], 0);                                   \
00090     a0 = vec_splat (constants[0], 1);                                   \
00091     a1 = vec_splat (constants[0], 2);                                   \
00092     a2 = vec_splat (constants[0], 3);                                   \
00093     mc4 = vec_splat (constants[0], 4);                                  \
00094     ma2 = vec_splat (constants[0], 5);                                  \
00095     bias = (vec_s16)vec_splat ((vec_s32)constants[0], 3);     \
00096                                                                         \
00097     zero = vec_splat_s16 (0);                                           \
00098     shift = vec_splat_u16 (4);                                          \
00099                                                                         \
00100     vx0 = vec_mradds (vec_sl (block[0], shift), constants[1], zero);    \
00101     vx1 = vec_mradds (vec_sl (block[1], shift), constants[2], zero);    \
00102     vx2 = vec_mradds (vec_sl (block[2], shift), constants[3], zero);    \
00103     vx3 = vec_mradds (vec_sl (block[3], shift), constants[4], zero);    \
00104     vx4 = vec_mradds (vec_sl (block[4], shift), constants[1], zero);    \
00105     vx5 = vec_mradds (vec_sl (block[5], shift), constants[4], zero);    \
00106     vx6 = vec_mradds (vec_sl (block[6], shift), constants[3], zero);    \
00107     vx7 = vec_mradds (vec_sl (block[7], shift), constants[2], zero);    \
00108                                                                         \
00109     IDCT_HALF                                                           \
00110                                                                         \
00111     vx0 = vec_mergeh (vy0, vy4);                                        \
00112     vx1 = vec_mergel (vy0, vy4);                                        \
00113     vx2 = vec_mergeh (vy1, vy5);                                        \
00114     vx3 = vec_mergel (vy1, vy5);                                        \
00115     vx4 = vec_mergeh (vy2, vy6);                                        \
00116     vx5 = vec_mergel (vy2, vy6);                                        \
00117     vx6 = vec_mergeh (vy3, vy7);                                        \
00118     vx7 = vec_mergel (vy3, vy7);                                        \
00119                                                                         \
00120     vy0 = vec_mergeh (vx0, vx4);                                        \
00121     vy1 = vec_mergel (vx0, vx4);                                        \
00122     vy2 = vec_mergeh (vx1, vx5);                                        \
00123     vy3 = vec_mergel (vx1, vx5);                                        \
00124     vy4 = vec_mergeh (vx2, vx6);                                        \
00125     vy5 = vec_mergel (vx2, vx6);                                        \
00126     vy6 = vec_mergeh (vx3, vx7);                                        \
00127     vy7 = vec_mergel (vx3, vx7);                                        \
00128                                                                         \
00129     vx0 = vec_adds (vec_mergeh (vy0, vy4), bias);                       \
00130     vx1 = vec_mergel (vy0, vy4);                                        \
00131     vx2 = vec_mergeh (vy1, vy5);                                        \
00132     vx3 = vec_mergel (vy1, vy5);                                        \
00133     vx4 = vec_mergeh (vy2, vy6);                                        \
00134     vx5 = vec_mergel (vy2, vy6);                                        \
00135     vx6 = vec_mergeh (vy3, vy7);                                        \
00136     vx7 = vec_mergel (vy3, vy7);                                        \
00137                                                                         \
00138     IDCT_HALF                                                           \
00139                                                                         \
00140     shift = vec_splat_u16 (6);                                          \
00141     vx0 = vec_sra (vy0, shift);                                         \
00142     vx1 = vec_sra (vy1, shift);                                         \
00143     vx2 = vec_sra (vy2, shift);                                         \
00144     vx3 = vec_sra (vy3, shift);                                         \
00145     vx4 = vec_sra (vy4, shift);                                         \
00146     vx5 = vec_sra (vy5, shift);                                         \
00147     vx6 = vec_sra (vy6, shift);                                         \
00148     vx7 = vec_sra (vy7, shift);
00149 
00150 
00151 static const vec_s16 constants[5] = {
00152     {23170, 13573,  6518, 21895, -23170, -21895,    32,    31},
00153     {16384, 22725, 21407, 19266,  16384,  19266, 21407, 22725},
00154     {22725, 31521, 29692, 26722,  22725,  26722, 29692, 31521},
00155     {21407, 29692, 27969, 25172,  21407,  25172, 27969, 29692},
00156     {19266, 26722, 25172, 22654,  19266,  22654, 25172, 26722}
00157 };
00158 
00159 void idct_put_altivec(uint8_t* dest, int stride, vec_s16* block)
00160 {
00161 POWERPC_PERF_DECLARE(altivec_idct_put_num, 1);
00162     vec_u8 tmp;
00163 
00164 #if CONFIG_POWERPC_PERF
00165 POWERPC_PERF_START_COUNT(altivec_idct_put_num, 1);
00166 #endif
00167     IDCT
00168 
00169 #define COPY(dest,src)                                          \
00170     tmp = vec_packsu (src, src);                                \
00171     vec_ste ((vec_u32)tmp, 0, (unsigned int *)dest);       \
00172     vec_ste ((vec_u32)tmp, 4, (unsigned int *)dest);
00173 
00174     COPY (dest, vx0)    dest += stride;
00175     COPY (dest, vx1)    dest += stride;
00176     COPY (dest, vx2)    dest += stride;
00177     COPY (dest, vx3)    dest += stride;
00178     COPY (dest, vx4)    dest += stride;
00179     COPY (dest, vx5)    dest += stride;
00180     COPY (dest, vx6)    dest += stride;
00181     COPY (dest, vx7)
00182 
00183 POWERPC_PERF_STOP_COUNT(altivec_idct_put_num, 1);
00184 }
00185 
00186 void idct_add_altivec(uint8_t* dest, int stride, vec_s16* block)
00187 {
00188 POWERPC_PERF_DECLARE(altivec_idct_add_num, 1);
00189     vec_u8 tmp;
00190     vec_s16 tmp2, tmp3;
00191     vec_u8 perm0;
00192     vec_u8 perm1;
00193     vec_u8 p0, p1, p;
00194 
00195 #if CONFIG_POWERPC_PERF
00196 POWERPC_PERF_START_COUNT(altivec_idct_add_num, 1);
00197 #endif
00198 
00199     IDCT
00200 
00201     p0 = vec_lvsl (0, dest);
00202     p1 = vec_lvsl (stride, dest);
00203     p = vec_splat_u8 (-1);
00204     perm0 = vec_mergeh (p, p0);
00205     perm1 = vec_mergeh (p, p1);
00206 
00207 #define ADD(dest,src,perm)                                              \
00208     /* *(uint64_t *)&tmp = *(uint64_t *)dest; */                        \
00209     tmp = vec_ld (0, dest);                                             \
00210     tmp2 = (vec_s16)vec_perm (tmp, (vec_u8)zero, perm);       \
00211     tmp3 = vec_adds (tmp2, src);                                        \
00212     tmp = vec_packsu (tmp3, tmp3);                                      \
00213     vec_ste ((vec_u32)tmp, 0, (unsigned int *)dest);               \
00214     vec_ste ((vec_u32)tmp, 4, (unsigned int *)dest);
00215 
00216     ADD (dest, vx0, perm0)      dest += stride;
00217     ADD (dest, vx1, perm1)      dest += stride;
00218     ADD (dest, vx2, perm0)      dest += stride;
00219     ADD (dest, vx3, perm1)      dest += stride;
00220     ADD (dest, vx4, perm0)      dest += stride;
00221     ADD (dest, vx5, perm1)      dest += stride;
00222     ADD (dest, vx6, perm0)      dest += stride;
00223     ADD (dest, vx7, perm1)
00224 
00225 POWERPC_PERF_STOP_COUNT(altivec_idct_add_num, 1);
00226 }
00227 

Generated on Sat Feb 16 2013 09:23:13 for ffmpeg by  doxygen 1.7.1