/* * 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 06:37:34 EDT 2009 */ #include "codelet-dft.h" #ifdef HAVE_FMA /* Generated by: ../../../genfft/gen_twiddle -fma -reorder-insns -schedule-for-pipeline -compact -variables 4 -pipeline-latency 4 -n 9 -name t1_9 -include t.h */ /* * This function contains 96 FP additions, 88 FP multiplications, * (or, 24 additions, 16 multiplications, 72 fused multiply/add), * 72 stack variables, 10 constants, and 36 memory accesses */ #include "t.h" static void t1_9(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms) { DK(KP954188894, +0.954188894138671133499268364187245676532219158); DK(KP852868531, +0.852868531952443209628250963940074071936020296); DK(KP363970234, +0.363970234266202361351047882776834043890471784); DK(KP492403876, +0.492403876506104029683371512294761506835321626); DK(KP984807753, +0.984807753012208059366743024589523013670643252); DK(KP777861913, +0.777861913430206160028177977318626690410586096); DK(KP839099631, +0.839099631177280011763127298123181364687434283); DK(KP176326980, +0.176326980708464973471090386868618986121633062); DK(KP866025403, +0.866025403784438646763723170752936183471402627); DK(KP500000000, +0.500000000000000000000000000000000000000000000); INT m; for (m = mb, W = W + (mb * 16); m < me; m = m + 1, ri = ri + ms, ii = ii + ms, W = W + 16, MAKE_VOLATILE_STRIDE(rs)) { E T1K, T24, T1H, T23; { E T1, T1R, T1Q, T10, T1W, Te, TB, T1l, T1r, T1q, T1M, TE, T1g, Tz, T12; E TC, TH, TK, T17, TR, TG, TJ, TD; T1 = ri[0]; T1R = ii[0]; { E T9, Tc, TY, Ta, Tb, TX, T7; { E T3, T6, T8, TW, T4, T2, T5; T3 = ri[WS(rs, 3)]; T6 = ii[WS(rs, 3)]; T2 = W[4]; T9 = ri[WS(rs, 6)]; Tc = ii[WS(rs, 6)]; T8 = W[10]; TW = T2 * T6; T4 = T2 * T3; T5 = W[5]; TY = T8 * Tc; Ta = T8 * T9; Tb = W[11]; TX = FNMS(T5, T3, TW); T7 = FMA(T5, T6, T4); } { E Th, Tk, Ti, T1n, Tn, Tq, Tp, T1i, Tx, T1j, To, Tj, TZ, Td, Tg; E TA, Tl, Ty; Th = ri[WS(rs, 1)]; TZ = FNMS(Tb, T9, TY); Td = FMA(Tb, Tc, Ta); Tk = ii[WS(rs, 1)]; Tg = W[0]; T1Q = TX + TZ; T10 = TX - TZ; T1W = Td - T7; Te = T7 + Td; Ti = Tg * Th; T1n = Tg * Tk; { E Tt, Tw, Ts, Tv, T1h, Tu, Tm; Tt = ri[WS(rs, 7)]; Tw = ii[WS(rs, 7)]; Ts = W[12]; Tv = W[13]; Tn = ri[WS(rs, 4)]; Tq = ii[WS(rs, 4)]; T1h = Ts * Tw; Tu = Ts * Tt; Tm = W[6]; Tp = W[7]; T1i = FNMS(Tv, Tt, T1h); Tx = FMA(Tv, Tw, Tu); T1j = Tm * Tq; To = Tm * Tn; } Tj = W[1]; TB = ri[WS(rs, 2)]; { E T1k, Tr, T1o, T1p; T1k = FNMS(Tp, Tn, T1j); Tr = FMA(Tp, Tq, To); T1o = FNMS(Tj, Th, T1n); Tl = FMA(Tj, Tk, Ti); T1p = T1k + T1i; T1l = T1i - T1k; Ty = Tr + Tx; T1r = Tr - Tx; T1q = FNMS(KP500000000, T1p, T1o); T1M = T1o + T1p; TE = ii[WS(rs, 2)]; } T1g = FNMS(KP500000000, Ty, Tl); Tz = Tl + Ty; TA = W[2]; { E TN, TQ, TP, T16, TO, TM; TN = ri[WS(rs, 8)]; TQ = ii[WS(rs, 8)]; TM = W[14]; T12 = TA * TE; TC = TA * TB; TP = W[15]; T16 = TM * TQ; TO = TM * TN; TH = ri[WS(rs, 5)]; TK = ii[WS(rs, 5)]; T17 = FNMS(TP, TN, T16); TR = FMA(TP, TQ, TO); TG = W[8]; TJ = W[9]; } TD = W[3]; } } { E TV, Tf, T1S, T1V, T1d, T1a, T19, T1N, TT, T1c; { E T13, TF, T15, TL, T14, TI, TS, T18; TV = FNMS(KP500000000, Te, T1); Tf = T1 + Te; T14 = TG * TK; TI = TG * TH; T13 = FNMS(TD, TB, T12); TF = FMA(TD, TE, TC); T15 = FNMS(TJ, TH, T14); TL = FMA(TJ, TK, TI); T1S = T1Q + T1R; T1V = FNMS(KP500000000, T1Q, T1R); T18 = T15 + T17; T1d = T15 - T17; TS = TL + TR; T1a = TR - TL; T19 = FNMS(KP500000000, T18, T13); T1N = T13 + T18; TT = TF + TS; T1c = FNMS(KP500000000, TS, TF); } { E T11, T1z, T1E, T1D, T21, T1X, T1I, T1C, T1Y, T1y, T20, T1u, T1U, TU; T1U = TT - Tz; TU = Tz + TT; { E T1P, T1O, T1L, T1T; T1P = T1M + T1N; T1O = T1M - T1N; T11 = FMA(KP866025403, T10, TV); T1z = FNMS(KP866025403, T10, TV); T1L = FNMS(KP500000000, TU, Tf); ri[0] = Tf + TU; T1T = FNMS(KP500000000, T1P, T1S); ii[0] = T1P + T1S; ri[WS(rs, 3)] = FMA(KP866025403, T1O, T1L); ri[WS(rs, 6)] = FNMS(KP866025403, T1O, T1L); ii[WS(rs, 6)] = FNMS(KP866025403, T1U, T1T); ii[WS(rs, 3)] = FMA(KP866025403, T1U, T1T); } { E T1B, T1m, T1w, T1f, T1s, T1A, T1b, T1e, T1x, T1t; T1E = FNMS(KP866025403, T1a, T19); T1b = FMA(KP866025403, T1a, T19); T1e = FMA(KP866025403, T1d, T1c); T1D = FNMS(KP866025403, T1d, T1c); T1B = FMA(KP866025403, T1l, T1g); T1m = FNMS(KP866025403, T1l, T1g); T21 = FNMS(KP866025403, T1W, T1V); T1X = FMA(KP866025403, T1W, T1V); T1w = FNMS(KP176326980, T1b, T1e); T1f = FMA(KP176326980, T1e, T1b); T1s = FNMS(KP866025403, T1r, T1q); T1A = FMA(KP866025403, T1r, T1q); T1x = FNMS(KP839099631, T1m, T1s); T1t = FMA(KP839099631, T1s, T1m); T1I = FNMS(KP176326980, T1A, T1B); T1C = FMA(KP176326980, T1B, T1A); T1Y = FNMS(KP777861913, T1x, T1w); T1y = FMA(KP777861913, T1x, T1w); T20 = FNMS(KP777861913, T1t, T1f); T1u = FMA(KP777861913, T1t, T1f); } { E T22, T1G, T1Z, T1F, T1J, T1v; ii[WS(rs, 1)] = FNMS(KP984807753, T1Y, T1X); T1v = FNMS(KP492403876, T1u, T11); ri[WS(rs, 1)] = FMA(KP984807753, T1u, T11); T1F = FNMS(KP363970234, T1E, T1D); T1J = FMA(KP363970234, T1D, T1E); ri[WS(rs, 4)] = FMA(KP852868531, T1y, T1v); ri[WS(rs, 7)] = FNMS(KP852868531, T1y, T1v); T1K = FNMS(KP954188894, T1J, T1I); T22 = FMA(KP954188894, T1J, T1I); T1G = FNMS(KP954188894, T1F, T1C); T24 = FMA(KP954188894, T1F, T1C); T1Z = FMA(KP492403876, T1Y, T1X); ii[WS(rs, 2)] = FNMS(KP984807753, T22, T21); ri[WS(rs, 2)] = FMA(KP984807753, T1G, T1z); T1H = FNMS(KP492403876, T1G, T1z); ii[WS(rs, 7)] = FNMS(KP852868531, T20, T1Z); ii[WS(rs, 4)] = FMA(KP852868531, T20, T1Z); T23 = FMA(KP492403876, T22, T21); } } } } ri[WS(rs, 8)] = FMA(KP852868531, T1K, T1H); ri[WS(rs, 5)] = FNMS(KP852868531, T1K, T1H); ii[WS(rs, 8)] = FMA(KP852868531, T24, T23); ii[WS(rs, 5)] = FNMS(KP852868531, T24, T23); } } static const tw_instr twinstr[] = { {TW_FULL, 0, 9}, {TW_NEXT, 1, 0} }; static const ct_desc desc = { 9, "t1_9", twinstr, &GENUS, {24, 16, 72, 0}, 0, 0, 0 }; void X(codelet_t1_9) (planner *p) { X(kdft_dit_register) (p, t1_9, &desc); } #else /* HAVE_FMA */ /* Generated by: ../../../genfft/gen_twiddle -compact -variables 4 -pipeline-latency 4 -n 9 -name t1_9 -include t.h */ /* * This function contains 96 FP additions, 72 FP multiplications, * (or, 60 additions, 36 multiplications, 36 fused multiply/add), * 41 stack variables, 8 constants, and 36 memory accesses */ #include "t.h" static void t1_9(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms) { DK(KP939692620, +0.939692620785908384054109277324731469936208134); DK(KP342020143, +0.342020143325668733044099614682259580763083368); DK(KP984807753, +0.984807753012208059366743024589523013670643252); DK(KP173648177, +0.173648177666930348851716626769314796000375677); DK(KP642787609, +0.642787609686539326322643409907263432907559884); DK(KP766044443, +0.766044443118978035202392650555416673935832457); DK(KP500000000, +0.500000000000000000000000000000000000000000000); DK(KP866025403, +0.866025403784438646763723170752936183471402627); INT m; for (m = mb, W = W + (mb * 16); m < me; m = m + 1, ri = ri + ms, ii = ii + ms, W = W + 16, MAKE_VOLATILE_STRIDE(rs)) { E T1, T1B, TQ, T1G, Tc, TN, T1A, T1H, TL, T1x, T17, T1o, T1c, T1n, Tu; E T1w, TW, T1k, T11, T1l; { E T6, TO, Tb, TP; T1 = ri[0]; T1B = ii[0]; { E T3, T5, T2, T4; T3 = ri[WS(rs, 3)]; T5 = ii[WS(rs, 3)]; T2 = W[4]; T4 = W[5]; T6 = FMA(T2, T3, T4 * T5); TO = FNMS(T4, T3, T2 * T5); } { E T8, Ta, T7, T9; T8 = ri[WS(rs, 6)]; Ta = ii[WS(rs, 6)]; T7 = W[10]; T9 = W[11]; Tb = FMA(T7, T8, T9 * Ta); TP = FNMS(T9, T8, T7 * Ta); } TQ = KP866025403 * (TO - TP); T1G = KP866025403 * (Tb - T6); Tc = T6 + Tb; TN = FNMS(KP500000000, Tc, T1); T1A = TO + TP; T1H = FNMS(KP500000000, T1A, T1B); } { E Tz, T19, TE, T14, TJ, T15, TK, T1a; { E Tw, Ty, Tv, Tx; Tw = ri[WS(rs, 2)]; Ty = ii[WS(rs, 2)]; Tv = W[2]; Tx = W[3]; Tz = FMA(Tv, Tw, Tx * Ty); T19 = FNMS(Tx, Tw, Tv * Ty); } { E TB, TD, TA, TC; TB = ri[WS(rs, 5)]; TD = ii[WS(rs, 5)]; TA = W[8]; TC = W[9]; TE = FMA(TA, TB, TC * TD); T14 = FNMS(TC, TB, TA * TD); } { E TG, TI, TF, TH; TG = ri[WS(rs, 8)]; TI = ii[WS(rs, 8)]; TF = W[14]; TH = W[15]; TJ = FMA(TF, TG, TH * TI); T15 = FNMS(TH, TG, TF * TI); } TK = TE + TJ; T1a = T14 + T15; TL = Tz + TK; T1x = T19 + T1a; { E T13, T16, T18, T1b; T13 = FNMS(KP500000000, TK, Tz); T16 = KP866025403 * (T14 - T15); T17 = T13 + T16; T1o = T13 - T16; T18 = KP866025403 * (TJ - TE); T1b = FNMS(KP500000000, T1a, T19); T1c = T18 + T1b; T1n = T1b - T18; } } { E Ti, TY, Tn, TT, Ts, TU, Tt, TZ; { E Tf, Th, Te, Tg; Tf = ri[WS(rs, 1)]; Th = ii[WS(rs, 1)]; Te = W[0]; Tg = W[1]; Ti = FMA(Te, Tf, Tg * Th); TY = FNMS(Tg, Tf, Te * Th); } { E Tk, Tm, Tj, Tl; Tk = ri[WS(rs, 4)]; Tm = ii[WS(rs, 4)]; Tj = W[6]; Tl = W[7]; Tn = FMA(Tj, Tk, Tl * Tm); TT = FNMS(Tl, Tk, Tj * Tm); } { E Tp, Tr, To, Tq; Tp = ri[WS(rs, 7)]; Tr = ii[WS(rs, 7)]; To = W[12]; Tq = W[13]; Ts = FMA(To, Tp, Tq * Tr); TU = FNMS(Tq, Tp, To * Tr); } Tt = Tn + Ts; TZ = TT + TU; Tu = Ti + Tt; T1w = TY + TZ; { E TS, TV, TX, T10; TS = FNMS(KP500000000, Tt, Ti); TV = KP866025403 * (TT - TU); TW = TS + TV; T1k = TS - TV; TX = KP866025403 * (Ts - Tn); T10 = FNMS(KP500000000, TZ, TY); T11 = TX + T10; T1l = T10 - TX; } } { E T1y, Td, TM, T1v; T1y = KP866025403 * (T1w - T1x); Td = T1 + Tc; TM = Tu + TL; T1v = FNMS(KP500000000, TM, Td); ri[0] = Td + TM; ri[WS(rs, 3)] = T1v + T1y; ri[WS(rs, 6)] = T1v - T1y; } { E T1D, T1z, T1C, T1E; T1D = KP866025403 * (TL - Tu); T1z = T1w + T1x; T1C = T1A + T1B; T1E = FNMS(KP500000000, T1z, T1C); ii[0] = T1z + T1C; ii[WS(rs, 6)] = T1E - T1D; ii[WS(rs, 3)] = T1D + T1E; } { E TR, T1I, T1e, T1J, T1i, T1F, T1f, T1K; TR = TN + TQ; T1I = T1G + T1H; { E T12, T1d, T1g, T1h; T12 = FMA(KP766044443, TW, KP642787609 * T11); T1d = FMA(KP173648177, T17, KP984807753 * T1c); T1e = T12 + T1d; T1J = KP866025403 * (T1d - T12); T1g = FNMS(KP642787609, TW, KP766044443 * T11); T1h = FNMS(KP984807753, T17, KP173648177 * T1c); T1i = KP866025403 * (T1g - T1h); T1F = T1g + T1h; } ri[WS(rs, 1)] = TR + T1e; ii[WS(rs, 1)] = T1F + T1I; T1f = FNMS(KP500000000, T1e, TR); ri[WS(rs, 7)] = T1f - T1i; ri[WS(rs, 4)] = T1f + T1i; T1K = FNMS(KP500000000, T1F, T1I); ii[WS(rs, 4)] = T1J + T1K; ii[WS(rs, 7)] = T1K - T1J; } { E T1j, T1M, T1q, T1N, T1u, T1L, T1r, T1O; T1j = TN - TQ; T1M = T1H - T1G; { E T1m, T1p, T1s, T1t; T1m = FMA(KP173648177, T1k, KP984807753 * T1l); T1p = FNMS(KP939692620, T1o, KP342020143 * T1n); T1q = T1m + T1p; T1N = KP866025403 * (T1p - T1m); T1s = FNMS(KP984807753, T1k, KP173648177 * T1l); T1t = FMA(KP342020143, T1o, KP939692620 * T1n); T1u = KP866025403 * (T1s + T1t); T1L = T1s - T1t; } ri[WS(rs, 2)] = T1j + T1q; ii[WS(rs, 2)] = T1L + T1M; T1r = FNMS(KP500000000, T1q, T1j); ri[WS(rs, 8)] = T1r - T1u; ri[WS(rs, 5)] = T1r + T1u; T1O = FNMS(KP500000000, T1L, T1M); ii[WS(rs, 5)] = T1N + T1O; ii[WS(rs, 8)] = T1O - T1N; } } } static const tw_instr twinstr[] = { {TW_FULL, 0, 9}, {TW_NEXT, 1, 0} }; static const ct_desc desc = { 9, "t1_9", twinstr, &GENUS, {60, 36, 36, 0}, 0, 0, 0 }; void X(codelet_t1_9) (planner *p) { X(kdft_dit_register) (p, t1_9, &desc); } #endif /* HAVE_FMA */