📄 t1fv_8.c
字号:
/* * 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 21:43:17 EDT 2003 */#include "codelet-dft.h"/* Generated by: /homee/stevenj/cvs/fftw3.0.1/genfft/gen_twiddle_c -simd -compact -variables 4 -n 8 -name t1fv_8 -include t1f.h *//* * This function contains 33 FP additions, 16 FP multiplications, * (or, 33 additions, 16 multiplications, 0 fused multiply/add), * 24 stack variables, and 16 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 "t1f.h"static const R *t1fv_8(R *ri, R *ii, const R *W, stride ios, int m, int dist){ DVK(KP707106781, +0.707106781186547524400844362104849039284835938); int i; R *x; x = ri; BEGIN_SIMD(); for (i = m; i > 0; i = i - VL, x = x + (VL * dist), W = W + (TWVL * 14)) { V T4, Tq, Tm, Tr, T9, Tt, Te, Tu, T1, T3, T2; T1 = LD(&(x[0]), dist, &(x[0])); T2 = LD(&(x[WS(ios, 4)]), dist, &(x[0])); T3 = BYTWJ(&(W[TWVL * 6]), T2); T4 = VSUB(T1, T3); Tq = VADD(T1, T3); { V Tj, Tl, Ti, Tk; Ti = LD(&(x[WS(ios, 2)]), dist, &(x[0])); Tj = BYTWJ(&(W[TWVL * 2]), Ti); Tk = LD(&(x[WS(ios, 6)]), dist, &(x[0])); Tl = BYTWJ(&(W[TWVL * 10]), Tk); Tm = VSUB(Tj, Tl); Tr = VADD(Tj, Tl); } { V T6, T8, T5, T7; T5 = LD(&(x[WS(ios, 1)]), dist, &(x[WS(ios, 1)])); T6 = BYTWJ(&(W[0]), T5); T7 = LD(&(x[WS(ios, 5)]), dist, &(x[WS(ios, 1)])); T8 = BYTWJ(&(W[TWVL * 8]), T7); T9 = VSUB(T6, T8); Tt = VADD(T6, T8); } { V Tb, Td, Ta, Tc; Ta = LD(&(x[WS(ios, 7)]), dist, &(x[WS(ios, 1)])); Tb = BYTWJ(&(W[TWVL * 12]), Ta); Tc = LD(&(x[WS(ios, 3)]), dist, &(x[WS(ios, 1)])); Td = BYTWJ(&(W[TWVL * 4]), Tc); Te = VSUB(Tb, Td); Tu = VADD(Tb, Td); } { V Ts, Tv, Tw, Tx; Ts = VADD(Tq, Tr); Tv = VADD(Tt, Tu); ST(&(x[WS(ios, 4)]), VSUB(Ts, Tv), dist, &(x[0])); ST(&(x[0]), VADD(Ts, Tv), dist, &(x[0])); Tw = VSUB(Tq, Tr); Tx = VBYI(VSUB(Tu, Tt)); ST(&(x[WS(ios, 6)]), VSUB(Tw, Tx), dist, &(x[0])); ST(&(x[WS(ios, 2)]), VADD(Tw, Tx), dist, &(x[0])); { V Tg, To, Tn, Tp, Tf, Th; Tf = VMUL(LDK(KP707106781), VADD(T9, Te)); Tg = VADD(T4, Tf); To = VSUB(T4, Tf); Th = VMUL(LDK(KP707106781), VSUB(Te, T9)); Tn = VBYI(VSUB(Th, Tm)); Tp = VBYI(VADD(Tm, Th)); ST(&(x[WS(ios, 7)]), VSUB(Tg, Tn), dist, &(x[WS(ios, 1)])); ST(&(x[WS(ios, 3)]), VADD(To, Tp), dist, &(x[WS(ios, 1)])); ST(&(x[WS(ios, 1)]), VADD(Tg, Tn), dist, &(x[WS(ios, 1)])); ST(&(x[WS(ios, 5)]), VSUB(To, Tp), dist, &(x[WS(ios, 1)])); } } } END_SIMD(); return W;}static const tw_instr twinstr[] = { VTW(1), VTW(2), VTW(3), VTW(4), VTW(5), VTW(6), VTW(7), {TW_NEXT, VL, 0}};static const ct_desc desc = { 8, "t1fv_8", twinstr, {33, 16, 0, 0}, &GENUS, 0, 0, 0 };void X(codelet_t1fv_8) (planner *p) { X(kdft_dit_register) (p, t1fv_8, &desc);}
⌨️ 快捷键说明
复制代码
Ctrl + C
搜索代码
Ctrl + F
全屏模式
F11
切换主题
Ctrl + Shift + D
显示快捷键
?
增大字号
Ctrl + =
减小字号
Ctrl + -