apps/codecs/demac/libdemac/vector_math32_armv4.h

   1 /*
   2
   3 libdemac - A Monkey's Audio decoder
   4
   5 $Id$
   6
   7 Copyright (C) Dave Chapman 2007
   8
   9 ARMv4 vector math copyright (C) 2008 Jens Arnold
  10
  11 This program is free software; you can redistribute it and/or modify
  12 it under the terms of the GNU General Public License as published by
  13 the Free Software Foundation; either version 2 of the License, or
  14 (at your option) any later version.
  15
  16 This program is distributed in the hope that it will be useful,
  17 but WITHOUT ANY WARRANTY; without even the implied warranty of
  18 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  19 GNU General Public License for more details.
  20
  21 You should have received a copy of the GNU General Public License
  22 along with this program; if not, write to the Free Software
  23 Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110, USA
  24
  25 */
  26
  27 static inline void vector_add(int32_t* v1, int32_t* v2)
  28 {
  29 #if ORDER > 32
  30     int cnt = ORDER>>5;
  31 #endif
  32
  33 #if ORDER > 16
  34 #define ADD_SUB_BLOCKS "8"
  35 #else
  36 #define ADD_SUB_BLOCKS "4"
  37 #endif
  38
  39     asm volatile (
  40     "1:                              \n"
  41         ".rept " ADD_SUB_BLOCKS     "\n"
  42         "ldmia   %[v1],  {r0-r3}     \n"
  43         "ldmia   %[v2]!, {r4-r7}     \n"
  44         "add     r0, r0, r4          \n"
  45         "add     r1, r1, r5          \n"
  46         "add     r2, r2, r6          \n"
  47         "add     r3, r3, r7          \n"
  48         "stmia   %[v1]!, {r0-r3}     \n"
  49         ".endr                       \n"
  50 #if ORDER > 32
  51         "subs    %[cnt], %[cnt], #1  \n"
  52         "bne     1b                  \n"
  53 #endif
  54         : /* outputs */
  55 #if ORDER > 32
  56         [cnt]"+r"(cnt),
  57 #endif
  58         [v1] "+r"(v1),
  59         [v2] "+r"(v2)
  60         : /* inputs */
  61         : /* clobbers */
  62         "r0", "r1", "r2", "r3", "r4",
  63         "r5", "r6", "r7", "memory"
  64     );
  65 }
  66
  67 static inline void vector_sub(int32_t* v1, int32_t* v2)
  68 {
  69 #if ORDER > 32
  70     int cnt = ORDER>>5;
  71 #endif
  72
  73     asm volatile (
  74     "1:                              \n"
  75         ".rept " ADD_SUB_BLOCKS     "\n"
  76         "ldmia   %[v1],  {r0-r3}     \n"
  77         "ldmia   %[v2]!, {r4-r7}     \n"
  78         "sub     r0, r0, r4          \n"
  79         "sub     r1, r1, r5          \n"
  80         "sub     r2, r2, r6          \n"
  81         "sub     r3, r3, r7          \n"
  82         "stmia   %[v1]!, {r0-r3}     \n"
  83         ".endr                       \n"
  84 #if ORDER > 32
  85         "subs    %[cnt], %[cnt], #1  \n"
  86         "bne     1b                  \n"
  87 #endif
  88         : /* outputs */
  89 #if ORDER > 32
  90         [cnt]"+r"(cnt),
  91 #endif
  92         [v1] "+r"(v1),
  93         [v2] "+r"(v2)
  94         : /* inputs */
  95         : /* clobbers */
  96         "r0", "r1", "r2", "r3", "r4",
  97         "r5", "r6", "r7", "memory"
  98     );
  99 }
 100
 101 static inline int32_t scalarproduct(int32_t* v1, int32_t* v2)
 102 {
 103     int res;
 104 #if ORDER > 32
 105     int cnt = ORDER>>5;
 106 #endif
 107
 108     asm volatile (
 109 #if ORDER > 16
 110 #if ORDER > 32
 111         "mov     %[res], #0              \n"
 112 #endif
 113         "ldmia   %[v2]!, {r6-r7}         \n"
 114     "1:                                  \n"
 115         "ldmia   %[v1]!, {r0,r1,r3-r5}   \n"
 116 #if ORDER > 32
 117         "mla     %[res], r6, r0, %[res]  \n"
 118 #else
 119         "mul     %[res], r6, r0          \n"
 120 #endif
 121         "mla     %[res], r7, r1, %[res]  \n"
 122         "ldmia   %[v2]!, {r0-r2,r6-r8}   \n"
 123         "mla     %[res], r0, r3, %[res]  \n"
 124         "mla     %[res], r1, r4, %[res]  \n"
 125         "mla     %[res], r2, r5, %[res]  \n"
 126         "ldmia   %[v1]!, {r0-r4}         \n"
 127         "mla     %[res], r6, r0, %[res]  \n"
 128         "mla     %[res], r7, r1, %[res]  \n"
 129         "mla     %[res], r8, r2, %[res]  \n"
 130         "ldmia   %[v2]!, {r0,r1,r6-r8}   \n"
 131         "mla     %[res], r0, r3, %[res]  \n"
 132         "mla     %[res], r1, r4, %[res]  \n"
 133         "ldmia   %[v1]!, {r0-r5}         \n"
 134         "mla     %[res], r6, r0, %[res]  \n"
 135         "mla     %[res], r7, r1, %[res]  \n"
 136         "mla     %[res], r8, r2, %[res]  \n"
 137         "ldmia   %[v2]!, {r0-r2,r6,r7}   \n"
 138         "mla     %[res], r0, r3, %[res]  \n"
 139         "mla     %[res], r1, r4, %[res]  \n"
 140         "mla     %[res], r2, r5, %[res]  \n"
 141         "ldmia   %[v1]!, {r0,r1,r3-r5}   \n"
 142         "mla     %[res], r6, r0, %[res]  \n"
 143         "mla     %[res], r7, r1, %[res]  \n"
 144         "ldmia   %[v2]!, {r0-r2,r6-r8}   \n"
 145         "mla     %[res], r0, r3, %[res]  \n"
 146         "mla     %[res], r1, r4, %[res]  \n"
 147         "mla     %[res], r2, r5, %[res]  \n"
 148         "ldmia   %[v1]!, {r0-r4}         \n"
 149         "mla     %[res], r6, r0, %[res]  \n"
 150         "mla     %[res], r7, r1, %[res]  \n"
 151         "mla     %[res], r8, r2, %[res]  \n"
 152         "ldmia   %[v2]!, {r0,r1,r6-r8}   \n"
 153         "mla     %[res], r0, r3, %[res]  \n"
 154         "mla     %[res], r1, r4, %[res]  \n"
 155         "ldmia   %[v1]!, {r0-r5}         \n"
 156         "mla     %[res], r6, r0, %[res]  \n"
 157         "mla     %[res], r7, r1, %[res]  \n"
 158         "mla     %[res], r8, r2, %[res]  \n"
 159 #if ORDER > 32
 160         "ldmia   %[v2]!, {r0-r2,r6,r7}   \n"
 161 #else
 162         "ldmia   %[v2]!, {r0-r2}         \n"
 163 #endif
 164         "mla     %[res], r0, r3, %[res]  \n"
 165         "mla     %[res], r1, r4, %[res]  \n"
 166         "mla     %[res], r2, r5, %[res]  \n"
 167 #if ORDER > 32
 168         "subs    %[cnt], %[cnt], #1      \n"
 169         "bne     1b                      \n"
 170 #endif
 171
 172 #else /* ORDER <= 16 */
 173         "ldmia   %[v1]!, {r0-r3}         \n"
 174         "ldmia   %[v2]!, {r4-r7}         \n"
 175         "mul     %[res], r4, r0          \n"
 176         "mla     %[res], r5, r1, %[res]  \n"
 177         "mla     %[res], r6, r2, %[res]  \n"
 178         "mla     %[res], r7, r3, %[res]  \n"
 179
 180         ".rept   3                       \n"
 181         "ldmia   %[v1]!, {r0-r3}         \n"
 182         "ldmia   %[v2]!, {r4-r7}         \n"
 183         "mla     %[res], r4, r0, %[res]  \n"
 184         "mla     %[res], r5, r1, %[res]  \n"
 185         "mla     %[res], r6, r2, %[res]  \n"
 186         "mla     %[res], r7, r3, %[res]  \n"
 187         ".endr                           \n"
 188 #endif /* ORDER <= 16 */
 189         : /* outputs */
 190 #if ORDER > 32
 191         [cnt]"+r"(cnt),
 192 #endif
 193         [v1] "+r"(v1),
 194         [v2] "+r"(v2),
 195         [res]"=r"(res)
 196         : /* inputs */
 197         : /* clobbers */
 198         "r0", "r1", "r2", "r3",
 199         "r4", "r5", "r6", "r7"
 200 #if ORDER > 16
 201         ,"r8"
 202 #endif
 203     );
 204     return res;
 205 }