Alc/mixer_sse2.c

   1 /**
   2  * OpenAL cross platform audio library
   3  * Copyright (C) 1999-2014 by authors.
   4  * This library is free software; you can redistribute it and/or
   5  *  modify it under the terms of the GNU Library General Public
   6  *  License as published by the Free Software Foundation; either
   7  *  version 2 of the License, or (at your option) any later version.
   8  *
   9  * This library 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  *  Library General Public License for more details.
  13  *
  14  * You should have received a copy of the GNU Library General Public
  15  *  License along with this library; if not, write to the
  16  *  Free Software Foundation, Inc., 59 Temple Place - Suite 330,
  17  *  Boston, MA  02111-1307, USA.
  18  * Or go to http://www.gnu.org/copyleft/lgpl.html
  19  */
  20
  21 #include "config.h"
  22
  23 #include <xmmintrin.h>
  24 #include <emmintrin.h>
  25
  26 #include "alu.h"
  27 #include "mixer_defs.h"
  28
  29
  30 static inline void InitiatePositionArrays(ALuint frac, ALuint increment,
  31                                           ALuint *frac_arr, ALuint *pos_arr)
  32 {
  33     ALuint frac_tmp;
  34
  35     pos_arr[0] = 0;
  36     frac_arr[0] = frac;
  37
  38     frac_tmp = frac_arr[0] + increment;
  39     pos_arr[1] = pos_arr[0] + (frac_tmp>>FRACTIONBITS);
  40     frac_arr[1] = frac_tmp & FRACTIONMASK;
  41
  42     frac_tmp = frac_arr[1] + increment;
  43     pos_arr[2] = pos_arr[1] + (frac_tmp>>FRACTIONBITS);
  44     frac_arr[2] = frac_tmp & FRACTIONMASK;
  45
  46     frac_tmp = frac_arr[2] + increment;
  47     pos_arr[3] = pos_arr[2] + (frac_tmp>>FRACTIONBITS);
  48     frac_arr[3] = frac_tmp & FRACTIONMASK;
  49 }
  50
  51 const ALfloat *Resample_lerp32_SSE2(const ALfloat *src, ALuint frac, ALuint increment,
  52                                     ALfloat *restrict dst, ALuint numsamples)
  53 {
  54     const __m128i increment4 = _mm_set1_epi32(increment*4);
  55     const __m128 fracOne4 = _mm_set1_ps(1.0f/FRACTIONONE);
  56     const __m128i fracMask4 = _mm_set1_epi32(FRACTIONMASK);
  57     alignas(16) union { ALuint i[4]; float f[4]; } pos_;
  58     alignas(16) union { ALuint i[4]; float f[4]; } frac_;
  59     __m128i frac4, pos4;
  60     ALuint pos;
  61     ALuint i;
  62
  63     InitiatePositionArrays(frac, increment, frac_.i, pos_.i);
  64
  65     frac4 = _mm_castps_si128(_mm_load_ps(frac_.f));
  66     pos4 = _mm_castps_si128(_mm_load_ps(pos_.f));
  67
  68     for(i = 0;numsamples-i > 3;i += 4)
  69     {
  70         const __m128 val1 = _mm_setr_ps(src[pos_.i[0]], src[pos_.i[1]], src[pos_.i[2]], src[pos_.i[3]]);
  71         const __m128 val2 = _mm_setr_ps(src[pos_.i[0]+1], src[pos_.i[1]+1], src[pos_.i[2]+1], src[pos_.i[3]+1]);
  72
  73         /* val1 + (val2-val1)*mu */
  74         const __m128 r0 = _mm_sub_ps(val2, val1);
  75         const __m128 mu = _mm_mul_ps(_mm_cvtepi32_ps(frac4), fracOne4);
  76         const __m128 out = _mm_add_ps(val1, _mm_mul_ps(mu, r0));
  77
  78         _mm_store_ps(&dst[i], out);
  79
  80         frac4 = _mm_add_epi32(frac4, increment4);
  81         pos4 = _mm_add_epi32(pos4, _mm_srli_epi32(frac4, FRACTIONBITS));
  82         frac4 = _mm_and_si128(frac4, fracMask4);
  83
  84         _mm_store_ps(pos_.f, _mm_castsi128_ps(pos4));
  85     }
  86
  87     pos = pos_.i[0];
  88     frac = _mm_cvtsi128_si32(frac4);
  89
  90     for(;i < numsamples;i++)
  91     {
  92         dst[i] = lerp(src[pos], src[pos+1], frac * (1.0f/FRACTIONONE));
  93
  94         frac += increment;
  95         pos  += frac>>FRACTIONBITS;
  96         frac &= FRACTIONMASK;
  97     }
  98     return dst;
  99 }