/* * Copyright (c) 2003, 2007-8 Matteo Frigo * Copyright (c) 2003, 2007-8 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 Sun Jul 12 06:45:13 EDT 2009 */ #include "codelet-rdft.h" #ifdef HAVE_FMA /* Generated by: ../../../genfft/gen_hc2cdft -fma -reorder-insns -schedule-for-pipeline -compact -variables 4 -pipeline-latency 4 -n 6 -dit -name hc2cfdft_6 -include hc2cf.h */ /* * This function contains 58 FP additions, 44 FP multiplications, * (or, 36 additions, 22 multiplications, 22 fused multiply/add), * 42 stack variables, 2 constants, and 24 memory accesses */ #include "hc2cf.h" static void hc2cfdft_6(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms) { DK(KP866025403, +0.866025403784438646763723170752936183471402627); DK(KP500000000, +0.500000000000000000000000000000000000000000000); INT m; for (m = mb, W = W + ((mb - 1) * 10); m < me; m = m + 1, Rp = Rp + ms, Ip = Ip + ms, Rm = Rm - ms, Im = Im - ms, W = W + 10, MAKE_VOLATILE_STRIDE(rs)) { E TP, TT, TN, TM, TY, T13; { E T3, TQ, TJ, T12, Tu, TB, TX, T10, Tj, Tf, Ti, Td, Th, TU, TS; { E TC, TI, TF, TH, TA, Tw, TZ; { E T1, T2, TD, TE; T1 = Ip[0]; T2 = Im[0]; TD = Rm[0]; TE = Rp[0]; TC = W[0]; T3 = T1 - T2; TI = T1 + T2; TQ = TE + TD; TF = TD - TE; TH = W[1]; } { E Tr, To, Ts, Tl, Tq; { E Tm, Tn, TG, T11; Tm = Rm[WS(rs, 2)]; Tn = Rp[WS(rs, 2)]; TG = TC * TF; T11 = TH * TF; Tr = Ip[WS(rs, 2)]; TA = Tn + Tm; To = Tm - Tn; TJ = FNMS(TH, TI, TG); T12 = FMA(TC, TI, T11); Ts = Im[WS(rs, 2)]; } Tl = W[8]; Tq = W[9]; { E Tz, Ty, TW, Tx, Tt, Tp; Tw = W[6]; Tx = Tr - Ts; Tt = Tr + Ts; Tp = Tl * To; Tz = W[7]; Ty = Tw * Tx; TW = Tl * Tt; Tu = FNMS(Tq, Tt, Tp); TZ = Tz * Tx; TB = FNMS(Tz, TA, Ty); TX = FMA(Tq, To, TW); } } { E T5, T6, Ta, Tb; T5 = Ip[WS(rs, 1)]; T10 = FMA(Tw, TA, TZ); T6 = Im[WS(rs, 1)]; Ta = Rp[WS(rs, 1)]; Tb = Rm[WS(rs, 1)]; { E T4, Tg, T7, Tc, T9, T8, TR; T4 = W[5]; Tg = T5 - T6; T7 = T5 + T6; Tj = Ta + Tb; Tc = Ta - Tb; T9 = W[4]; T8 = T4 * T7; Tf = W[2]; Ti = W[3]; TR = T9 * T7; Td = FMA(T9, Tc, T8); Th = Tf * Tg; TU = Ti * Tg; TS = FNMS(T4, Tc, TR); } } } { E Te, T1d, TK, Tv, T1a, T1b, Tk, TV; TP = Td + T3; Te = T3 - Td; Tk = FNMS(Ti, Tj, Th); TV = FMA(Tf, Tj, TU); T1d = TQ + TS; TT = TQ - TS; TN = TJ - TB; TK = TB + TJ; Tv = Tk + Tu; TM = Tu - Tk; TY = TV - TX; T1a = TV + TX; T1b = T10 + T12; T13 = T10 - T12; { E T1g, TL, T1e, T1c, T19, T1f; T1g = Tv - TK; TL = Tv + TK; T1e = T1a + T1b; T1c = T1a - T1b; T19 = FNMS(KP500000000, TL, Te); Ip[0] = KP500000000 * (Te + TL); T1f = FNMS(KP500000000, T1e, T1d); Rp[0] = KP500000000 * (T1d + T1e); Im[WS(rs, 1)] = -(KP500000000 * (FNMS(KP866025403, T1c, T19))); Ip[WS(rs, 2)] = KP500000000 * (FMA(KP866025403, T1c, T19)); Rm[WS(rs, 1)] = KP500000000 * (FMA(KP866025403, T1g, T1f)); Rp[WS(rs, 2)] = KP500000000 * (FNMS(KP866025403, T1g, T1f)); } } } { E TO, T16, T14, T18, T17, T15; TO = TM + TN; T16 = TN - TM; T14 = TY + T13; T18 = T13 - TY; T17 = FMA(KP500000000, TO, TP); Im[WS(rs, 2)] = KP500000000 * (TO - TP); T15 = FNMS(KP500000000, T14, TT); Rm[WS(rs, 2)] = KP500000000 * (TT + T14); Im[0] = -(KP500000000 * (FNMS(KP866025403, T18, T17))); Ip[WS(rs, 1)] = KP500000000 * (FMA(KP866025403, T18, T17)); Rm[0] = KP500000000 * (FNMS(KP866025403, T16, T15)); Rp[WS(rs, 1)] = KP500000000 * (FMA(KP866025403, T16, T15)); } } } static const tw_instr twinstr[] = { {TW_FULL, 1, 6}, {TW_NEXT, 1, 0} }; static const hc2c_desc desc = { 6, "hc2cfdft_6", twinstr, &GENUS, {36, 22, 22, 0} }; void X(codelet_hc2cfdft_6) (planner *p) { X(khc2c_register) (p, hc2cfdft_6, &desc, HC2C_VIA_DFT); } #else /* HAVE_FMA */ /* Generated by: ../../../genfft/gen_hc2cdft -compact -variables 4 -pipeline-latency 4 -n 6 -dit -name hc2cfdft_6 -include hc2cf.h */ /* * This function contains 58 FP additions, 36 FP multiplications, * (or, 44 additions, 22 multiplications, 14 fused multiply/add), * 40 stack variables, 3 constants, and 24 memory accesses */ #include "hc2cf.h" static void hc2cfdft_6(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms) { DK(KP250000000, +0.250000000000000000000000000000000000000000000); DK(KP500000000, +0.500000000000000000000000000000000000000000000); DK(KP433012701, +0.433012701892219323381861585376468091735701313); INT m; for (m = mb, W = W + ((mb - 1) * 10); m < me; m = m + 1, Rp = Rp + ms, Ip = Ip + ms, Rm = Rm - ms, Im = Im - ms, W = W + 10, MAKE_VOLATILE_STRIDE(rs)) { E T3, TM, Tc, TN, Ts, T10, TI, TR, TF, T11, TH, TU; { E T1, T2, TD, Tz, TA, TB, T7, Tf, Tb, Th, Tq, Tw, Tm, Tu, T4; E T8; { E T5, T6, T9, Ta; T1 = Ip[0]; T2 = Im[0]; TD = T1 + T2; Tz = Rm[0]; TA = Rp[0]; TB = Tz - TA; T5 = Ip[WS(rs, 1)]; T6 = Im[WS(rs, 1)]; T7 = T5 + T6; Tf = T5 - T6; T9 = Rp[WS(rs, 1)]; Ta = Rm[WS(rs, 1)]; Tb = T9 - Ta; Th = T9 + Ta; { E To, Tp, Tk, Tl; To = Rp[WS(rs, 2)]; Tp = Rm[WS(rs, 2)]; Tq = To - Tp; Tw = To + Tp; Tk = Ip[WS(rs, 2)]; Tl = Im[WS(rs, 2)]; Tm = Tk + Tl; Tu = Tk - Tl; } } T3 = T1 - T2; TM = TA + Tz; T4 = W[5]; T8 = W[4]; Tc = FMA(T4, T7, T8 * Tb); TN = FNMS(T4, Tb, T8 * T7); { E Ti, TP, Tr, TQ; { E Te, Tg, Tj, Tn; Te = W[2]; Tg = W[3]; Ti = FNMS(Tg, Th, Te * Tf); TP = FMA(Tg, Tf, Te * Th); Tj = W[9]; Tn = W[8]; Tr = FMA(Tj, Tm, Tn * Tq); TQ = FNMS(Tj, Tq, Tn * Tm); } Ts = Ti - Tr; T10 = TP + TQ; TI = Ti + Tr; TR = TP - TQ; } { E Tx, TS, TE, TT; { E Tt, Tv, Ty, TC; Tt = W[6]; Tv = W[7]; Tx = FNMS(Tv, Tw, Tt * Tu); TS = FMA(Tv, Tu, Tt * Tw); Ty = W[0]; TC = W[1]; TE = FNMS(TC, TD, Ty * TB); TT = FMA(TC, TB, Ty * TD); } TF = Tx + TE; T11 = TS + TT; TH = TE - Tx; TU = TS - TT; } } { E T12, Td, TG, TZ; T12 = KP433012701 * (T10 - T11); Td = T3 - Tc; TG = Ts + TF; TZ = FNMS(KP250000000, TG, KP500000000 * Td); Ip[0] = KP500000000 * (Td + TG); Im[WS(rs, 1)] = T12 - TZ; Ip[WS(rs, 2)] = TZ + T12; } { E T16, T13, T14, T15; T16 = KP433012701 * (Ts - TF); T13 = TM + TN; T14 = T10 + T11; T15 = FNMS(KP250000000, T14, KP500000000 * T13); Rp[WS(rs, 2)] = T15 - T16; Rp[0] = KP500000000 * (T13 + T14); Rm[WS(rs, 1)] = T16 + T15; } { E TY, TJ, TK, TX; TY = KP433012701 * (TU - TR); TJ = TH - TI; TK = Tc + T3; TX = FMA(KP500000000, TK, KP250000000 * TJ); Im[WS(rs, 2)] = KP500000000 * (TJ - TK); Im[0] = TY - TX; Ip[WS(rs, 1)] = TX + TY; } { E TL, TO, TV, TW; TL = KP433012701 * (TI + TH); TO = TM - TN; TV = TR + TU; TW = FNMS(KP250000000, TV, KP500000000 * TO); Rp[WS(rs, 1)] = TL + TW; Rm[WS(rs, 2)] = KP500000000 * (TO + TV); Rm[0] = TW - TL; } } } static const tw_instr twinstr[] = { {TW_FULL, 1, 6}, {TW_NEXT, 1, 0} }; static const hc2c_desc desc = { 6, "hc2cfdft_6", twinstr, &GENUS, {44, 22, 14, 0} }; void X(codelet_hc2cfdft_6) (planner *p) { X(khc2c_register) (p, hc2cfdft_6, &desc, HC2C_VIA_DFT); } #endif /* HAVE_FMA */