include/common/SC_Altivec.h

   1 /*
   2         SuperCollider real time audio synthesis system
   3     Copyright (c) 2003 James McCartney. All rights reserved.
   4         http://www.audiosynth.com
   5
   6     This program is free software; you can redistribute it and/or modify
   7     it under the terms of the GNU General Public License as published by
   8     the Free Software Foundation; either version 2 of the License, or
   9     (at your option) any later version.
  10
  11     This program is distributed in the hope that it will be useful,
  12     but WITHOUT ANY WARRANTY; without even the implied warranty of
  13     MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  14     GNU General Public License for more details.
  15
  16     You should have received a copy of the GNU General Public License
  17     along with this program; if not, write to the Free Software
  18     Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301  USA
  19 */
  20
  21 #ifndef _SC_Altivec_
  22 #define _SC_Altivec_
  23
  24 #if defined(HAS_ALTIVEC) || defined(__ALTIVEC__) && !defined(__APPLE_ALTIVEC__)
  25 # include <altivec.h>
  26 /* From <altivec.h>:
  27    You are allowed to undef these for C++ compatibility. */
  28 # ifdef bool
  29 #  undef bool
  30 # endif
  31 #endif
  32
  33 #if __VEC__
  34
  35 typedef vector signed int vint32;
  36 typedef vector unsigned int vuint32;
  37 typedef vector float vfloat32;
  38
  39 // Since gcc 3.3 vector initializers are surrounded by brackets. <sk>
  40 #if defined(__GNUC__) && ((__GNUC__ >= 4) || ((__GNUC__ == 3) && (__GNUC_MINOR__ >= 3)))
  41 # define vinit(x)       { x, x, x, x }
  42 #else
  43 # define vinit(x)       ( x )
  44 #endif
  45
  46 //#define vload(x) (*vtempptr = (x), vec_splat(vtemp,0))
  47 #define define_vtemp vfloat32 vtemp; float *vtempptr = (float*)&vtemp;
  48 #define define_vones vfloat32 vones = vec_ctf(vec_splat_s32(1),0);
  49 #define define_vzero vfloat32 vzero = (vfloat32)vec_splat_s32(0);
  50 #define vi0123 (vec_unpackh(vec_unpackh((vector signed char)vec_lvsl(0,(int*)0))))
  51 #define v0123 (vec_ctf(vec_unpackh(vec_unpackh((vector signed char)vec_lvsl(0,(int*)0))), 0))
  52 #define v0123_4ths (vec_ctf(vec_unpackh(vec_unpackh((vector signed char)vec_lvsl(0,(int*)0))), 2))
  53 #define vstart(x, vslope) (vec_madd(vslope, v0123_4ths, vload(x)))
  54
  55 inline vint32 vec_not(vint32 arg)
  56 {
  57         return vec_nor(arg, arg);
  58 }
  59
  60 inline vfloat32 vec_not(vfloat32 arg)
  61 {
  62         return vec_nor(arg, arg);
  63 }
  64
  65 inline vfloat32 vec_mul(vfloat32 a, vfloat32 b)
  66 {
  67         define_vzero;
  68         return vec_madd(a, b, vzero);
  69 }
  70
  71 #define vec_2sComp(x) (vec_sub(vec_sub (x, x), x))
  72
  73 #define USEVEC (ft->mAltivecAvailable && !(BUFLENGTH & 3))
  74
  75 typedef union vec_union {
  76         int32           i[4];
  77         float32         f[4];
  78         vint32          vi;
  79         vfloat32        vf;
  80 } vec_union;
  81
  82 inline vfloat32 vload( float f )
  83 {
  84         vec_union temp;
  85         temp.f[0] = f;
  86         return vec_splat( vec_lde( 0, temp.f ), 0 );
  87 }
  88
  89 inline vint32 vload( int32 i )
  90 {
  91         vec_union temp;
  92         temp.i[0] = i;
  93         return vec_splat( vec_lde( 0, temp.i ), 0 );
  94 }
  95
  96 inline vint32 vload( int32 a, int32 b, int32 c, int32 d )
  97 {
  98         vec_union temp;
  99         temp.i[0] = a;
 100         temp.i[1] = b;
 101         temp.i[2] = c;
 102         temp.i[3] = d;
 103         return temp.vi;
 104 }
 105
 106 inline vector float vec_float_1( void )
 107 {
 108         return vec_ctf( vec_splat_u32(1), 0);
 109 }
 110
 111 inline vector float vec_reciprocal( vector float v )
 112 {
 113         vector float reciprocal = vec_re( v );
 114         return vec_madd( reciprocal, vec_nmsub( reciprocal, v, vec_float_1()), reciprocal ); //Newton Rapheson refinement
 115 }
 116
 117
 118 // seed = ((seed & mask) << shift1) ^ (((seed << shift2) ^ seed) >> shift3);
 119
 120 #define define_trshifts \
 121         vuint32 trmask = ((vuint32)(0xFFFFFFFE,0xFFFFFFF8,0xFFFFFFF0,0)); \
 122         vuint32 trshift1 = ((vuint32)(12, 14,  7,  0)); \
 123         vuint32 trshift2 = ((vuint32)(13,  2,  3,  0)); \
 124         vuint32 trshift3 = ((vuint32)(19, 25, 11,  0));
 125
 126 inline vuint32 trands(vuint32 seed, vuint32 trmask, vuint32 shift1, vuint32 shift2, vuint32 shift3)
 127 {
 128         return vec_xor(vec_sl(vec_and(seed, trmask),shift1), vec_sr(vec_xor(vec_sl(seed,shift2),seed),shift3));
 129 }
 130
 131 #endif
 132 #endif
 133
 134