root/libfftss/r4_sse2_1_n.c

/* [<][>][^][v][top][bottom][index][help] */

DEFINITIONS

This source file includes following definitions.
  1. kern_r4_sse2_1_n_f
  2. kern_r4_sse2_1_n_b
  3. fftss_r4_sse2_1_n_f
  4. fftss_r4_sse2_1_n_b
/*
 * Copyright (C) 2002-2007 Akira Nukada. All rights reserved.
 * Copyright (C) 2002-2007 The SSI Project. All rights reserved. 
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions are met:
 *     * Redistributions of source code must retain the above copyright
 *       notice, this list of conditions and the following disclaimer.
 *     * Redistributions in binary form must reproduce the above copyright
 *       notice, this list of conditions and the following disclaimer in the
 *       documentation and/or other materials provided with the distribution.
 *     * Neither the name of the project nor the names of its contributors 
 *       may be used to endorse or promote products derived from this software 
 *       without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE SSI PROJECT ``AS IS'' AND ANY EXPRESS 
 * OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED 
 * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
 * DISCLAIMED. IN NO EVENT SHALL THE SSI PROJECT BE LIABLE FOR ANY DIRECT,
 * INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
 * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
 * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
 *  STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 
 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF
 * THE POSSIBILITY OF SUCH DAMAGE.
 *
 */
#include "libfftss.h"

#include <emmintrin.h>

extern unsigned int fftss_sse2_cmz[];
extern unsigned int fftss_sse2_czm[];

static inline void 
kern_r4_sse2_1_n_f(double *restrict o0,
                   double *restrict o1,
                   double *restrict o2,
                   double *restrict o3,
                   double *i0, 
                   double *i1, 
                   double *i2, 
                   double *i3, 
                   double *w,
                   long bsize, long blocks)
{
  long i, j;
  __m128d zm, mz;
  zm = _mm_loadu_pd((double *)fftss_sse2_czm);
  mz = _mm_loadu_pd((double *)fftss_sse2_cmz);

  for (i = 0; i < blocks; i++) {
#if 0
    double wr1, wi1;
    double wr2, wi2;
    double wr3, wi3;

    wr1 = w[2 * i * bsize]; wi1 = w[2 * i * bsize + 1];
    wr2 = w[4 * i * bsize]; wi2 = w[4 * i * bsize + 1];
    wr3 = w[6 * i * bsize]; wi3 = w[6 * i * bsize + 1];
#endif

    __m128d w1, w2, w3;
    __m128d w1p, w2p, w3p;
    
    w1 = _mm_load_pd(w + 2 * i * bsize);
    w2 = _mm_load_pd(w + 4 * i * bsize);
    w3 = _mm_load_pd(w + 6 * i * bsize);
#if 1
    w1p = _mm_unpacklo_pd(w1, w1);
    w2p = _mm_unpacklo_pd(w2, w2);
    w3p = _mm_unpacklo_pd(w3, w3);
    w1 = _mm_unpackhi_pd(w1, w1);
    w2 = _mm_unpackhi_pd(w2, w2);
    w3 = _mm_unpackhi_pd(w3, w3);
#endif
/*
        w1 = _mm_xor_pd(w1, _mm_set_pd(-0.0, 0.0));
    w2 = _mm_xor_pd(w2, _mm_set_pd(-0.0, 0.0));
    w3 = _mm_xor_pd(w3, _mm_set_pd(-0.0, 0.0));
*/
        w1 = _mm_xor_pd(w1, zm);
        w2 = _mm_xor_pd(w2, zm);
        w3 = _mm_xor_pd(w3, zm);


    for (j = 0; j < bsize; j++) {
#if 0
      double pr1, pi1;
      double pr2, pi2;
      double pr3, pi3;
      double qr0, qi0;
      double qr1, qi1;
      double qr2, qi2;
      double qr3, qi3;
      
      pr1 = i1[2 * j] * wr1 + i1[2 * j + 1] * wi1;
      pi1 = i1[2 * j + 1] * wr1 - i1[2 * j] * wi1;
      pr2 = i2[2 * j] * wr2 + i2[2 * j + 1] * wi2;
      pi2 = i2[2 * j + 1] * wr2 - i2[2 * j] * wi2;
      pr3 = i3[2 * j] * wr3 + i3[2 * j + 1] * wi3;
      pi3 = i3[2 * j + 1] * wr3 - i3[2 * j] * wi3;
      
      qr0 = i0[2 * j] + pr2; qr2 = i0[2 * j] - pr2;
      qi0 = i0[2 * j + 1] + pi2; qi2 = i0[2 * j + 1] - pi2; 
      qr1 = pr1 + pr3; qr3 = pr1 - pr3;
      qi1 = pi1 + pi3; qi3 = pi1 - pi3;
      
      o0[2 * j] = qr0 + qr1; o0[2 * j + 1] = qi0 + qi1;
      o2[2 * j] = qr0 - qr1; o2[2 * j + 1] = qi0 - qi1;
      o1[2 * j] = qr2 + qi3; o1[2 * j + 1] = qi2 - qr3;
      o3[2 * j] = qr2 - qi3; o3[2 * j + 1] = qi2 + qr3;
#endif
      __m128d p0, p1, p2, p3;
      __m128d d0, d1, d2, d3;

      d0 = _mm_load_pd(i0 + 2 * j);
      d1 = _mm_load_pd(i1 + 2 * j);
      d2 = _mm_load_pd(i2 + 2 * j);
      d3 = _mm_load_pd(i3 + 2 * j);

      p1 = _mm_shuffle_pd(d1, d1, 1);
      p1 = _mm_mul_pd(p1, w1); 
      d1 = _mm_mul_pd(d1, w1p);
      d1 = _mm_add_pd(d1, p1);

      p2 = _mm_shuffle_pd(d2, d2, 1);
      p2 = _mm_mul_pd(p2, w2); 
      d2 = _mm_mul_pd(d2, w2p);
      d2 = _mm_add_pd(d2, p2);

      p3 = _mm_shuffle_pd(d3, d3, 1);
      p3 = _mm_mul_pd(p3, w3); 
      d3 = _mm_mul_pd(d3, w3p);
      d3 = _mm_add_pd(d3, p3);

      p0 = _mm_add_pd(d0, d2);
      p2 = _mm_sub_pd(d0, d2);
      p1 = _mm_add_pd(d1, d3);
      p3 = _mm_sub_pd(d1, d3);
/*      p3 = _mm_xor_pd(p3, _mm_set_sd(-0.0));  */
          p3 = _mm_xor_pd(p3, mz);
      p3 = _mm_shuffle_pd(p3, p3, 1);
      
      _mm_store_pd(o0 + 2 * j, _mm_add_pd(p0, p1));
      _mm_store_pd(o2 + 2 * j, _mm_sub_pd(p0, p1));
      _mm_store_pd(o1 + 2 * j, _mm_add_pd(p2, p3));
      _mm_store_pd(o3 + 2 * j, _mm_sub_pd(p2, p3));

    }
    i0 += 8 * bsize; i1 += 8 * bsize;
    i2 += 8 * bsize; i3 += 8 * bsize;
    o0 += 2 * bsize; o1 += 2 * bsize;
    o2 += 2 * bsize; o3 += 2 * bsize;
  }
}

static inline void 
kern_r4_sse2_1_n_b(double *restrict o0,
                   double *restrict o1,
                   double *restrict o2,
                   double *restrict o3,
                   double *i0, 
                   double *i1, 
                   double *i2, 
                   double *i3, 
                   double *w,
                   long bsize, long blocks)
{
  long i, j;
  __m128d zm, mz;
  zm = _mm_loadu_pd((double *)fftss_sse2_czm);
  mz = _mm_loadu_pd((double *)fftss_sse2_cmz);

  for (i = 0; i < blocks; i++) {
#if 0
    double wr1, wi1;
    double wr2, wi2;
    double wr3, wi3;
    
    wr1 = w[2 * i * bsize]; wi1 = w[2 * i * bsize + 1];
    wr2 = w[4 * i * bsize]; wi2 = w[4 * i * bsize + 1];
    wr3 = w[6 * i * bsize]; wi3 = w[6 * i * bsize + 1];

#endif
    __m128d w1, w2, w3;
    __m128d w1p, w2p, w3p;
    
    w1 = _mm_load_pd(w + 2 * i * bsize);
    w2 = _mm_load_pd(w + 4 * i * bsize);
    w3 = _mm_load_pd(w + 6 * i * bsize);
#if 1
    w1p = _mm_unpacklo_pd(w1, w1);
    w2p = _mm_unpacklo_pd(w2, w2);
    w3p = _mm_unpacklo_pd(w3, w3);
    w1 = _mm_unpackhi_pd(w1, w1);
    w2 = _mm_unpackhi_pd(w2, w2);
    w3 = _mm_unpackhi_pd(w3, w3);
#endif
        /*
    w1 = _mm_xor_pd(w1, _mm_set_sd(-0.0));
    w2 = _mm_xor_pd(w2, _mm_set_sd(-0.0));
    w3 = _mm_xor_pd(w3, _mm_set_sd(-0.0));
    */
        w1 = _mm_xor_pd(w1, mz);
        w2 = _mm_xor_pd(w2, mz);
        w3 = _mm_xor_pd(w3, mz);
        
    for (j = 0; j < bsize; j++) {
#if 0
      double pr1, pi1;
      double pr2, pi2;
      double pr3, pi3;
      double qr0, qi0;
      double qr1, qi1;
      double qr2, qi2;
      double qr3, qi3;
      
      pr1 = i1[2 * j] * wr1 - i1[2 * j + 1] * wi1;
      pi1 = i1[2 * j] * wi1 + i1[2 * j + 1] * wr1;
      pr2 = i2[2 * j] * wr2 - i2[2 * j + 1] * wi2;
      pi2 = i2[2 * j] * wi2 + i2[2 * j + 1] * wr2;
      pr3 = i3[2 * j] * wr3 - i3[2 * j + 1] * wi3;
      pi3 = i3[2 * j] * wi3 + i3[2 * j + 1] * wr3;
      
      qr0 = i0[2 * j] + pr2; qr2 = i0[2 * j] - pr2;
      qi0 = i0[2 * j + 1] + pi2; qi2 = i0[2 * j + 1] - pi2; 
      qr1 = pr1 + pr3; qr3 = pr1 - pr3;
      qi1 = pi1 + pi3; qi3 = pi1 - pi3;
      
      o0[2 * j] = qr0 + qr1; o0[2 * j + 1] = qi0 + qi1;
      o2[2 * j] = qr0 - qr1; o2[2 * j + 1] = qi0 - qi1;
      o1[2 * j] = qr2 - qi3; o1[2 * j + 1] = qi2 + qr3;
      o3[2 * j] = qr2 + qi3; o3[2 * j + 1] = qi2 - qr3;
#endif

      __m128d p0, p1, p2, p3;
      __m128d d0, d1, d2, d3;

      d0 = _mm_load_pd(i0 + 2 * j);
      d1 = _mm_load_pd(i1 + 2 * j);
      d2 = _mm_load_pd(i2 + 2 * j);
      d3 = _mm_load_pd(i3 + 2 * j);

      p1 = _mm_shuffle_pd(d1, d1, 1);
      p1 = _mm_mul_pd(p1, w1); 
      d1 = _mm_mul_pd(d1, w1p);
      d1 = _mm_add_pd(d1, p1);

      p2 = _mm_shuffle_pd(d2, d2, 1);
      p2 = _mm_mul_pd(p2, w2); 
      d2 = _mm_mul_pd(d2, w2p);
      d2 = _mm_add_pd(d2, p2);

      p3 = _mm_shuffle_pd(d3, d3, 1);
      p3 = _mm_mul_pd(p3, w3); 
      d3 = _mm_mul_pd(d3, w3p);
      d3 = _mm_add_pd(d3, p3);

      p0 = _mm_add_pd(d0, d2);
      p2 = _mm_sub_pd(d0, d2);
      p1 = _mm_add_pd(d1, d3);
      p3 = _mm_sub_pd(d1, d3);

/*      p3 = _mm_xor_pd(p3, _mm_set_sd(-0.0)); */
          p3 = _mm_xor_pd(p3, mz);
      p3 = _mm_shuffle_pd(p3, p3, 1);
      
      _mm_store_pd(o0 + 2 * j, _mm_add_pd(p0, p1));
      _mm_store_pd(o2 + 2 * j, _mm_sub_pd(p0, p1));
      _mm_store_pd(o1 + 2 * j, _mm_sub_pd(p2, p3));
      _mm_store_pd(o3 + 2 * j, _mm_add_pd(p2, p3));


    }
    i0 += 8 * bsize; i1 += 8 * bsize;
    i2 += 8 * bsize; i3 += 8 * bsize;
    o0 += 2 * bsize; o1 += 2 * bsize;
    o2 += 2 * bsize; o3 += 2 * bsize;
  }
}

void fftss_r4_sse2_1_n_f(double *in, double *out, double *w,
                   long bsize, long blocks)
{
  kern_r4_sse2_1_n_f(out, out + bsize * blocks * 2, 
                     out + bsize * blocks * 4, out + bsize * blocks * 6,
                     in, in + bsize * 2,
                     in + bsize * 4, in + bsize * 6,
                     w, bsize, blocks);
}

void fftss_r4_sse2_1_n_b(double *in, double *out, double *w,
                   long bsize, long blocks)
{
  kern_r4_sse2_1_n_b(out, out + bsize * blocks * 2, 
                     out + bsize * blocks * 4, out + bsize * blocks * 6,
                     in, in + bsize * 2,
                     in + bsize * 4, in + bsize * 6,
                     w, bsize, blocks);
}
/* [<][>][^][v][top][bottom][index][help] */