libavcodec/i386/fft_sse.c

   1 /*
   2  * FFT/MDCT transform with SSE optimizations
   3  * Copyright (c) 2002 Fabrice Bellard.
   4  *
   5  * This library is free software; you can redistribute it and/or
   6  * modify it under the terms of the GNU Lesser General Public
   7  * License as published by the Free Software Foundation; either
   8  * version 2 of the License, or (at your option) any later version.
   9  *
  10  * This library is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  13  * Lesser General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU Lesser General Public
  16  * License along with this library; if not, write to the Free Software
  17  * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  18  */
  19 #include "../dsputil.h"
  20 #include <math.h>
  21
  22 #include <xmmintrin.h>
  23
  24 static const float p1p1p1m1[4] __attribute__((aligned(16))) =
  25     { 1.0, 1.0, 1.0, -1.0 };
  26
  27 static const float p1p1m1m1[4] __attribute__((aligned(16))) =
  28     { 1.0, 1.0, -1.0, -1.0 };
  29
  30 #if 0
  31 static void print_v4sf(const char *str, __m128 a)
  32 {
  33     float *p = (float *)&a;
  34     printf("%s: %f %f %f %f\n",
  35            str, p[0], p[1], p[2], p[3]);
  36 }
  37 #endif
  38
  39 /* XXX: handle reverse case */
  40 void fft_calc_sse(FFTContext *s, FFTComplex *z)
  41 {
  42     int ln = s->nbits;
  43     int j, np, np2;
  44     int nblocks, nloops;
  45     register FFTComplex *p, *q;
  46     FFTComplex *cptr, *cptr1;
  47     int k;
  48
  49     np = 1 << ln;
  50
  51     {
  52         __m128 *r, a, b, a1, c1, c2;
  53
  54         r = (__m128 *)&z[0];
  55         c1 = *(__m128 *)p1p1m1m1;
  56         c2 = *(__m128 *)p1p1p1m1;
  57         j = (np >> 2);
  58         do {
  59             a = r[0];
  60             b = _mm_shuffle_ps(a, a, _MM_SHUFFLE(1, 0, 3, 2));
  61             a = _mm_mul_ps(a, c1);
  62             /* do the pass 0 butterfly */
  63             a = _mm_add_ps(a, b);
  64
  65             a1 = r[1];
  66             b = _mm_shuffle_ps(a1, a1, _MM_SHUFFLE(1, 0, 3, 2));
  67             a1 = _mm_mul_ps(a1, c1);
  68             /* do the pass 0 butterfly */
  69             b = _mm_add_ps(a1, b);
  70
  71             /* multiply third by -i */
  72             b = _mm_shuffle_ps(b, b, _MM_SHUFFLE(2, 3, 1, 0));
  73             b = _mm_mul_ps(b, c2);
  74
  75             /* do the pass 1 butterfly */
  76             r[0] = _mm_add_ps(a, b);
  77             r[1] = _mm_sub_ps(a, b);
  78             r += 2;
  79         } while (--j != 0);
  80     }
  81     /* pass 2 .. ln-1 */
  82
  83     nblocks = np >> 3;
  84     nloops = 1 << 2;
  85     np2 = np >> 1;
  86
  87     cptr1 = s->exptab1;
  88     do {
  89         p = z;
  90         q = z + nloops;
  91         j = nblocks;
  92         do {
  93             cptr = cptr1;
  94             k = nloops >> 1;
  95             do {
  96                 __m128 a, b, c, t1, t2;
  97
  98                 a = *(__m128 *)p;
  99                 b = *(__m128 *)q;
 100
 101                 /* complex mul */
 102                 c = *(__m128 *)cptr;
 103                 /*  cre*re cim*re */
 104                 t1 = _mm_mul_ps(c,
 105                                 _mm_shuffle_ps(b, b, _MM_SHUFFLE(2, 2, 0, 0)));
 106                 c = *(__m128 *)(cptr + 2);
 107                 /*  -cim*im cre*im */
 108                 t2 = _mm_mul_ps(c,
 109                                 _mm_shuffle_ps(b, b, _MM_SHUFFLE(3, 3, 1, 1)));
 110                 b = _mm_add_ps(t1, t2);
 111
 112                 /* butterfly */
 113                 *(__m128 *)p = _mm_add_ps(a, b);
 114                 *(__m128 *)q = _mm_sub_ps(a, b);
 115
 116                 p += 2;
 117                 q += 2;
 118                 cptr += 4;
 119             } while (--k);
 120
 121             p += nloops;
 122             q += nloops;
 123         } while (--j);
 124         cptr1 += nloops * 2;
 125         nblocks = nblocks >> 1;
 126         nloops = nloops << 1;
 127     } while (nblocks != 0);
 128 }