t1bv_64.c

fftw-3.0.1

/*
 * Copyright (c) 2003 Matteo Frigo
 * Copyright (c) 2003 Massachusetts Institute of Technology
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
 *
 */

/* This file was automatically generated --- DO NOT EDIT */
/* Generated on Sat Jul  5 21:44:08 EDT 2003 */

#include "codelet-dft.h"

/* Generated by: /homee/stevenj/cvs/fftw3.0.1/genfft/gen_twiddle_c -simd -compact -variables 4 -n 64 -name t1bv_64 -include t1b.h -sign 1 */

/*
 * This function contains 519 FP additions, 250 FP multiplications,
 * (or, 467 additions, 198 multiplications, 52 fused multiply/add),
 * 107 stack variables, and 128 memory accesses
 */
/*
 * Generator Id's : 
 * $Id: algsimp.ml,v 1.7 2003/03/15 20:29:42 stevenj Exp $
 * $Id: fft.ml,v 1.2 2003/03/15 20:29:42 stevenj Exp $
 * $Id: gen_twiddle_c.ml,v 1.7 2003/04/16 19:51:27 athena Exp $
 */

#include "t1b.h"

static const R *t1bv_64(R *ri, R *ii, const R *W, stride ios, int m, int dist)
{
     DVK(KP290284677, +0.290284677254462367636192375817395274691476278);
     DVK(KP956940335, +0.956940335732208864935797886980269969482849206);
     DVK(KP471396736, +0.471396736825997648556387625905254377657460319);
     DVK(KP881921264, +0.881921264348355029712756863660388349508442621);
     DVK(KP634393284, +0.634393284163645498215171613225493370675687095);
     DVK(KP773010453, +0.773010453362736960810906609758469800971041293);
     DVK(KP098017140, +0.098017140329560601994195563888641845861136673);
     DVK(KP995184726, +0.995184726672196886244836953109479921575474869);
     DVK(KP195090322, +0.195090322016128267848284868477022240927691618);
     DVK(KP980785280, +0.980785280403230449126182236134239036973933731);
     DVK(KP555570233, +0.555570233019602224742830813948532874374937191);
     DVK(KP831469612, +0.831469612302545237078788377617905756738560812);
     DVK(KP382683432, +0.382683432365089771728459984030398866761344562);
     DVK(KP923879532, +0.923879532511286756128183189396788286822416626);
     DVK(KP707106781, +0.707106781186547524400844362104849039284835938);
     int i;
     R *x;
     x = ii;
     BEGIN_SIMD();
     for (i = m; i > 0; i = i - VL, x = x + (VL * dist), W = W + (TWVL * 126)) {
	  V Tg, T4B, T6v, T7G, T3r, T4w, T5q, T7F, T5Y, T62, T28, T4d, T2g, T4a, T7g;
	  V T7Y, T6f, T6j, T2Z, T4k, T37, T4h, T7n, T81, T7w, T7x, T7y, T5M, T6q, T1k;
	  V T4s, T1r, T4t, T7t, T7u, T7v, T5F, T6p, TV, T4p, T12, T4q, T7A, T7B, TD;
	  V T4x, T3k, T4C, T5x, T6s, T1R, T4b, T7j, T7Z, T2j, T4e, T5V, T63, T2I, T4i;
	  V T7q, T82, T3a, T4l, T6c, T6k;
	  {
	       V T1, T3, T3p, T3n, Tb, Td, Te, T6, T8, T9, T2, T3o, T3m;
	       T1 = LD(&(x[0]), dist, &(x[0]));
	       T2 = LD(&(x[WS(ios, 32)]), dist, &(x[0]));
	       T3 = BYTW(&(W[TWVL * 62]), T2);
	       T3o = LD(&(x[WS(ios, 48)]), dist, &(x[0]));
	       T3p = BYTW(&(W[TWVL * 94]), T3o);
	       T3m = LD(&(x[WS(ios, 16)]), dist, &(x[0]));
	       T3n = BYTW(&(W[TWVL * 30]), T3m);
	       {
		    V Ta, Tc, T5, T7;
		    Ta = LD(&(x[WS(ios, 56)]), dist, &(x[0]));
		    Tb = BYTW(&(W[TWVL * 110]), Ta);
		    Tc = LD(&(x[WS(ios, 24)]), dist, &(x[0]));
		    Td = BYTW(&(W[TWVL * 46]), Tc);
		    Te = VSUB(Tb, Td);
		    T5 = LD(&(x[WS(ios, 8)]), dist, &(x[0]));
		    T6 = BYTW(&(W[TWVL * 14]), T5);
		    T7 = LD(&(x[WS(ios, 40)]), dist, &(x[0]));
		    T8 = BYTW(&(W[TWVL * 78]), T7);
		    T9 = VSUB(T6, T8);
	       }
	       {
		    V T4, Tf, T6t, T6u;
		    T4 = VSUB(T1, T3);
		    Tf = VMUL(LDK(KP707106781), VADD(T9, Te));
		    Tg = VSUB(T4, Tf);
		    T4B = VADD(T4, Tf);
		    T6t = VADD(T6, T8);
		    T6u = VADD(Tb, Td);
		    T6v = VSUB(T6t, T6u);
		    T7G = VADD(T6t, T6u);
	       }
	       {
		    V T3l, T3q, T5o, T5p;
		    T3l = VMUL(LDK(KP707106781), VSUB(T9, Te));
		    T3q = VSUB(T3n, T3p);
		    T3r = VSUB(T3l, T3q);
		    T4w = VADD(T3q, T3l);
		    T5o = VADD(T1, T3);
		    T5p = VADD(T3n, T3p);
		    T5q = VSUB(T5o, T5p);
		    T7F = VADD(T5o, T5p);
	       }
	  }
	  {
	       V T24, T26, T61, T2b, T2d, T60, T1W, T5W, T21, T5X, T22, T27;
	       {
		    V T23, T25, T2a, T2c;
		    T23 = LD(&(x[WS(ios, 17)]), dist, &(x[WS(ios, 1)]));
		    T24 = BYTW(&(W[TWVL * 32]), T23);
		    T25 = LD(&(x[WS(ios, 49)]), dist, &(x[WS(ios, 1)]));
		    T26 = BYTW(&(W[TWVL * 96]), T25);
		    T61 = VADD(T24, T26);
		    T2a = LD(&(x[WS(ios, 1)]), dist, &(x[WS(ios, 1)]));
		    T2b = BYTW(&(W[0]), T2a);
		    T2c = LD(&(x[WS(ios, 33)]), dist, &(x[WS(ios, 1)]));
		    T2d = BYTW(&(W[TWVL * 64]), T2c);
		    T60 = VADD(T2b, T2d);
	       }
	       {
		    V T1T, T1V, T1S, T1U;
		    T1S = LD(&(x[WS(ios, 9)]), dist, &(x[WS(ios, 1)]));
		    T1T = BYTW(&(W[TWVL * 16]), T1S);
		    T1U = LD(&(x[WS(ios, 41)]), dist, &(x[WS(ios, 1)]));
		    T1V = BYTW(&(W[TWVL * 80]), T1U);
		    T1W = VSUB(T1T, T1V);
		    T5W = VADD(T1T, T1V);
	       }
	       {
		    V T1Y, T20, T1X, T1Z;
		    T1X = LD(&(x[WS(ios, 57)]), dist, &(x[WS(ios, 1)]));
		    T1Y = BYTW(&(W[TWVL * 112]), T1X);
		    T1Z = LD(&(x[WS(ios, 25)]), dist, &(x[WS(ios, 1)]));
		    T20 = BYTW(&(W[TWVL * 48]), T1Z);
		    T21 = VSUB(T1Y, T20);
		    T5X = VADD(T1Y, T20);
	       }
	       T5Y = VSUB(T5W, T5X);
	       T62 = VSUB(T60, T61);
	       T22 = VMUL(LDK(KP707106781), VSUB(T1W, T21));
	       T27 = VSUB(T24, T26);
	       T28 = VSUB(T22, T27);
	       T4d = VADD(T27, T22);
	       {
		    V T2e, T2f, T7e, T7f;
		    T2e = VSUB(T2b, T2d);
		    T2f = VMUL(LDK(KP707106781), VADD(T1W, T21));
		    T2g = VSUB(T2e, T2f);
		    T4a = VADD(T2e, T2f);
		    T7e = VADD(T60, T61);
		    T7f = VADD(T5W, T5X);
		    T7g = VSUB(T7e, T7f);
		    T7Y = VADD(T7e, T7f);
	       }
	  }
	  {
	       V T2V, T2X, T6i, T32, T34, T6h, T2N, T6d, T2S, T6e, T2T, T2Y;
	       {
		    V T2U, T2W, T31, T33;
		    T2U = LD(&(x[WS(ios, 15)]), dist, &(x[WS(ios, 1)]));
		    T2V = BYTW(&(W[TWVL * 28]), T2U);
		    T2W = LD(&(x[WS(ios, 47)]), dist, &(x[WS(ios, 1)]));
		    T2X = BYTW(&(W[TWVL * 92]), T2W);
		    T6i = VADD(T2V, T2X);
		    T31 = LD(&(x[WS(ios, 63)]), dist, &(x[WS(ios, 1)]));
		    T32 = BYTW(&(W[TWVL * 124]), T31);
		    T33 = LD(&(x[WS(ios, 31)]), dist, &(x[WS(ios, 1)]));
		    T34 = BYTW(&(W[TWVL * 60]), T33);
		    T6h = VADD(T32, T34);
	       }
	       {
		    V T2K, T2M, T2J, T2L;
		    T2J = LD(&(x[WS(ios, 7)]), dist, &(x[WS(ios, 1)]));
		    T2K = BYTW(&(W[TWVL * 12]), T2J);
		    T2L = LD(&(x[WS(ios, 39)]), dist, &(x[WS(ios, 1)]));
		    T2M = BYTW(&(W[TWVL * 76]), T2L);
		    T2N = VSUB(T2K, T2M);
		    T6d = VADD(T2K, T2M);
	       }
	       {
		    V T2P, T2R, T2O, T2Q;
		    T2O = LD(&(x[WS(ios, 55)]), dist, &(x[WS(ios, 1)]));
		    T2P = BYTW(&(W[TWVL * 108]), T2O);
		    T2Q = LD(&(x[WS(ios, 23)]), dist, &(x[WS(ios, 1)]));
		    T2R = BYTW(&(W[TWVL * 44]), T2Q);
		    T2S = VSUB(T2P, T2R);
		    T6e = VADD(T2P, T2R);
	       }
	       T6f = VSUB(T6d, T6e);
	       T6j = VSUB(T6h, T6i);
	       T2T = VMUL(LDK(KP707106781), VSUB(T2N, T2S));
	       T2Y = VSUB(T2V, T2X);
	       T2Z = VSUB(T2T, T2Y);
	       T4k = VADD(T2Y, T2T);
	       {
		    V T35, T36, T7l, T7m;
		    T35 = VSUB(T32, T34);
		    T36 = VMUL(LDK(KP707106781), VADD(T2N, T2S));
		    T37 = VSUB(T35, T36);
		    T4h = VADD(T35, T36);
		    T7l = VADD(T6h, T6i);
		    T7m = VADD(T6d, T6e);
		    T7n = VSUB(T7l, T7m);
		    T81 = VADD(T7l, T7m);
	       }
	  }
	  {
	       V T1g, T1i, T5K, T1m, T1o, T5J, T18, T5G, T1d, T5H, T5I, T5L;
	       {
		    V T1f, T1h, T1l, T1n;
		    T1f = LD(&(x[WS(ios, 14)]), dist, &(x[0]));
		    T1g = BYTW(&(W[TWVL * 26]), T1f);
		    T1h = LD(&(x[WS(ios, 46)]), dist, &(x[0]));
		    T1i = BYTW(&(W[TWVL * 90]), T1h);
		    T5K = VADD(T1g, T1i);
		    T1l = LD(&(x[WS(ios, 62)]), dist, &(x[0]));
		    T1m = BYTW(&(W[TWVL * 122]), T1l);
		    T1n = LD(&(x[WS(ios, 30)]), dist, &(x[0]));
		    T1o = BYTW(&(W[TWVL * 58]), T1n);
		    T5J = VADD(T1m, T1o);
	       }
	       {
		    V T15, T17, T14, T16;
		    T14 = LD(&(x[WS(ios, 6)]), dist, &(x[0]));
		    T15 = BYTW(&(W[TWVL * 10]), T14);
		    T16 = LD(&(x[WS(ios, 38)]), dist, &(x[0]));
		    T17 = BYTW(&(W[TWVL * 74]), T16);
		    T18 = VSUB(T15, T17);
		    T5G = VADD(T15, T17);
	       }
	       {
		    V T1a, T1c, T19, T1b;
		    T19 = LD(&(x[WS(ios, 54)]), dist, &(x[0]));
		    T1a = BYTW(&(W[TWVL * 106]), T19);
		    T1b = LD(&(x[WS(ios, 22)]), dist, &(x[0]));
		    T1c = BYTW(&(W[TWVL * 42]), T1b);
		    T1d = VSUB(T1a, T1c);
		    T5H = VADD(T1a, T1c);
	       }
	       T7w = VADD(T5J, T5K);
	       T7x = VADD(T5G, T5H);
	       T7y = VSUB(T7w, T7x);
	       T5I = VSUB(T5G, T5H);
	       T5L = VSUB(T5J, T5K);
	       T5M = VFNMS(LDK(KP382683432), T5L, VMUL(LDK(KP923879532), T5I));
	       T6q = VFMA(LDK(KP923879532), T5L, VMUL(LDK(KP382683432), T5I));
	       {
		    V T1e, T1j, T1p, T1q;
		    T1e = VMUL(LDK(KP707106781), VSUB(T18, T1d));
		    T1j = VSUB(T1g, T1i);
		    T1k = VSUB(T1e, T1j);
		    T4s = VADD(T1j, T1e);
		    T1p = VSUB(T1m, T1o);
		    T1q = VMUL(LDK(KP707106781), VADD(T18, T1d));
		    T1r = VSUB(T1p, T1q);
		    T4t = VADD(T1p, T1q);
	       }
	  }
	  {
	       V TR, TT, T5A, TX, TZ, T5z, TJ, T5C, TO, T5D, T5B, T5E;
	       {
		    V TQ, TS, TW, TY;
		    TQ = LD(&(x[WS(ios, 18)]), dist, &(x[0]));
		    TR = BYTW(&(W[TWVL * 34]), TQ);
		    TS = LD(&(x[WS(ios, 50)]), dist, &(x[0]));
		    TT = BYTW(&(W[TWVL * 98]), TS);
		    T5A = VADD(TR, TT);
		    TW = LD(&(x[WS(ios, 2)]), dist, &(x[0]));
		    TX = BYTW(&(W[TWVL * 2]), TW);
		    TY = LD(&(x[WS(ios, 34)]), dist, &(x[0]));
		    TZ = BYTW(&(W[TWVL * 66]), TY);
		    T5z = VADD(TX, TZ);
	       }
	       {
		    V TG, TI, TF, TH;
		    TF = LD(&(x[WS(ios, 10)]), dist, &(x[0]));
		    TG = BYTW(&(W[TWVL * 18]), TF);
		    TH = LD(&(x[WS(ios, 42)]), dist, &(x[0]));
		    TI = BYTW(&(W[TWVL * 82]), TH);
		    TJ = VSUB(TG, TI);
		    T5C = VADD(TG, TI);
	       }
	       {
		    V TL, TN, TK, TM;
		    TK = LD(&(x[WS(ios, 58)]), dist, &(x[0]));
		    TL = BYTW(&(W[TWVL * 114]), TK);
		    TM = LD(&(x[WS(ios, 26)]), dist, &(x[0]));
		    TN = BYTW(&(W[TWVL * 50]), TM);
		    TO = VSUB(TL, TN);
		    T5D = VADD(TL, TN);
	       }
	       T7t = VADD(T5z, T5A);
	       T7u = VADD(T5C, T5D);
	       T7v = VSUB(T7t, T7u);
	       T5B = VSUB(T5z, T5A);
	       T5E = VSUB(T5C, T5D);
	       T5F = VFMA(LDK(KP382683432), T5B, VMUL(LDK(KP923879532), T5E));
	       T6p = VFNMS(LDK(KP382683432), T5E, VMUL(LDK(KP923879532), T5B));
	       {
		    V TP, TU, T10, T11;
		    TP = VMUL(LDK(KP707106781), VSUB(TJ, TO));
		    TU = VSUB(TR, TT);
		    TV = VSUB(TP, TU);
		    T4p = VADD(TU, TP);
		    T10 = VSUB(TX, TZ);
		    T11 = VMUL(LDK(KP707106781), VADD(TJ, TO));
		    T12 = VSUB(T10, T11);
		    T4q = VADD(T10, T11);
	       }
	  }
	  {
	       V Tl, T5r, TB, T5u, Tq, T5s, Tw, T5v, Tr, TC;
	       {
		    V Ti, Tk, Th, Tj;
		    Th = LD(&(x[WS(ios, 4)]), dist, &(x[0]));
		    Ti = BYTW(&(W[TWVL * 6]), Th);
		    Tj = LD(&(x[WS(ios, 36)]), dist, &(x[0]));
		    Tk = BYTW(&(W[TWVL * 70]), Tj);
		    Tl = VSUB(Ti, Tk);
		    T5r = VADD(Ti, Tk);
	       }
	       {