static void r2cfII_20(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
{
DK(KP572061402, +0.572061402817684297600072783580302076536153377);
DK(KP218508012, +0.218508012224410535399650602527877556893735408);
DK(KP309016994, +0.309016994374947424102293417182819058860154590);
DK(KP809016994, +0.809016994374947424102293417182819058860154590);
DK(KP559016994, +0.559016994374947424102293417182819058860154590);
DK(KP951056516, +0.951056516295153572116439333379382143405698634);
DK(KP587785252, +0.587785252292473129168705954639072768597652438);
DK(KP250000000, +0.250000000000000000000000000000000000000000000);
DK(KP176776695, +0.176776695296636881100211090526212259821208984);
DK(KP395284707, +0.395284707521047416499861693054089816714944392);
DK(KP672498511, +0.672498511963957326960058968885748755876783111);
DK(KP415626937, +0.415626937777453428589967464113135184222253485);
DK(KP707106781, +0.707106781186547524400844362104849039284835938);
{
INT i;
for (i = v; i > 0; i = i - 1, R0 = R0 + ivs, R1 = R1 + ivs, Cr = Cr + ovs, Ci = Ci + ovs, MAKE_VOLATILE_STRIDE(rs), MAKE_VOLATILE_STRIDE(csr), MAKE_VOLATILE_STRIDE(csi)) {
E T8, TD, Tm, TN, T9, TC, TY, TE, Te, TF, Tl, TK, T12, TL, Tk;
E TM, T1, T6, Tq, T1l, T1c, Tp, T1f, T1e, T1d, Ty, TW, T1g, T1m, Tx;
E Tu;
T8 = R1[WS(rs, 2)];
TD = KP707106781 * T8;
Tm = R1[WS(rs, 7)];
TN = KP707106781 * Tm;
{
E Ta, TA, Td, TB, Tb, Tc;
T9 = R1[WS(rs, 6)];
Ta = R1[WS(rs, 8)];
TA = T9 + Ta;
Tb = R1[0];
Tc = R1[WS(rs, 4)];
Td = Tb + Tc;
TB = Tb - Tc;
TC = FMA(KP415626937, TA, KP672498511 * TB);
TY = FNMS(KP415626937, TB, KP672498511 * TA);
TE = KP395284707 * (Ta - Td);
Te = Ta + Td;
TF = KP176776695 * Te;
}
{
E Tg, TJ, Tj, TI, Th, Ti;
Tg = R1[WS(rs, 1)];
Tl = R1[WS(rs, 3)];
TJ = Tg + Tl;
Th = R1[WS(rs, 5)];
Ti = R1[WS(rs, 9)];
Tj = Th + Ti;
TI = Th - Ti;
TK = FNMS(KP415626937, TJ, KP672498511 * TI);
T12 = FMA(KP415626937, TI, KP672498511 * TJ);
TL = KP395284707 * (Tg - Tj);
Tk = Tg + Tj;
TM = KP176776695 * Tk;
}
{
E T2, T5, T3, T4, T1a, T1b;
T1 = R0[0];
T2 = R0[WS(rs, 6)];
T5 = R0[WS(rs, 8)];
T3 = R0[WS(rs, 2)];
T4 = R0[WS(rs, 4)];
T1a = T4 + T2;
T1b = T5 + T3;
T6 = T2 + T3 - (T4 + T5);
Tq = FMA(KP250000000, T6, T1);
T1l = FNMS(KP951056516, T1b, KP587785252 * T1a);
T1c = FMA(KP951056516, T1a, KP587785252 * T1b);
Tp = KP559016994 * (T5 + T2 - (T4 + T3));
}
T1f = R0[WS(rs, 5)];
{
E Tv, Tw, Ts, Tt;
Tv = R0[WS(rs, 9)];
Tw = R0[WS(rs, 1)];
Tx = Tv - Tw;
T1e = Tv + Tw;
Ts = R0[WS(rs, 3)];
Tt = R0[WS(rs, 7)];
Tu = Ts - Tt;
T1d = Ts + Tt;
}
Ty = FMA(KP951056516, Tu, KP587785252 * Tx);
TW = FNMS(KP951056516, Tx, KP587785252 * Tu);
T1g = FMA(KP809016994, T1d, KP309016994 * T1e) + T1f;
T1m = FNMS(KP809016994, T1e, T1f) - (KP309016994 * T1d);
{
E T7, T1r, To, T1q, Tf, Tn;
T7 = T1 - T6;
T1r = T1e + T1f - T1d;
Tf = T8 + (T9 - Te);
Tn = (Tk - Tl) - Tm;
To = KP707106781 * (Tf + Tn);
T1q = KP707106781 * (Tf - Tn);
Cr[WS(csr, 2)] = T7 - To;
Ci[WS(csi, 2)] = T1q - T1r;
Cr[WS(csr, 7)] = T7 + To;
Ci[WS(csi, 7)] = T1q + T1r;
}
{
E T1h, T1j, TX, T15, T10, T16, T13, T17, TV, TZ, T11;
T1h = T1c - T1g;
T1j = T1c + T1g;
TV = Tq - Tp;
TX = TV - TW;
T15 = TV + TW;
TZ = FMA(KP218508012, T9, TD) + TF - TE;
T10 = TY + TZ;
T16 = TZ - TY;
T11 = FNMS(KP218508012, Tl, TL) - (TM + TN);
T13 = T11 - T12;
T17 = T11 + T12;
{
E T14, T19, T18, T1i;
T14 = T10 + T13;
Cr[WS(csr, 5)] = TX - T14;
Cr[WS(csr, 4)] = TX + T14;
T19 = T17 - T16;
Ci[WS(csi, 5)] = T19 - T1h;
Ci[WS(csi, 4)] = T19 + T1h;
T18 = T16 + T17;
Cr[WS(csr, 9)] = T15 - T18;
Cr[0] = T15 + T18;
T1i = T13 - T10;
Ci[0] = T1i - T1j;
Ci[WS(csi, 9)] = T1i + T1j;
}
}
{
E T1n, T1p, Tz, TR, TH, TS, TP, TT, Tr, TG, TO;
T1n = T1l + T1m;
T1p = T1m - T1l;
Tr = Tp + Tq;
Tz = Tr + Ty;
TR = Tr - Ty;
TG = TD + TE + FNMS(KP572061402, T9, TF);
TH = TC + TG;
TS = TC - TG;
TO = TL + TM + FNMS(KP572061402, Tl, TN);
TP = TK - TO;
TT = TK + TO;
{
E TQ, T1o, TU, T1k;
TQ = TH + TP;
Cr[WS(csr, 6)] = Tz - TQ;
Cr[WS(csr, 3)] = Tz + TQ;
T1o = TT - TS;
Ci[WS(csi, 6)] = T1o - T1p;
Ci[WS(csi, 3)] = T1o + T1p;
TU = TS + TT;
Cr[WS(csr, 8)] = TR - TU;
Cr[WS(csr, 1)] = TR + TU;
T1k = TP - TH;
Ci[WS(csi, 8)] = T1k - T1n;
Ci[WS(csi, 1)] = T1k + T1n;
}
}
}
}
}
static void hc2cfdft_8(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms)
{
DK(KP353553390, +0.353553390593273762200422181052424519642417969);
DK(KP500000000, +0.500000000000000000000000000000000000000000000);
{
INT m;
for (m = mb, W = W + ((mb - 1) * 14); m < me; m = m + 1, Rp = Rp + ms, Ip = Ip + ms, Rm = Rm - ms, Im = Im - ms, W = W + 14, MAKE_VOLATILE_STRIDE(32, rs)) {
E Tv, TX, Ts, TY, TE, T1a, TJ, T19, T1l, T1m, T9, T10, Ti, T11, TP;
E T16, TU, T17, T1i, T1j;
{
E Tt, Tu, TD, Tz, TA, TB, Tn, TI, Tr, TG, Tk, To;
Tt = Ip[0];
Tu = Im[0];
TD = Tt + Tu;
Tz = Rm[0];
TA = Rp[0];
TB = Tz - TA;
{
E Tl, Tm, Tp, Tq;
Tl = Ip[WS(rs, 2)];
Tm = Im[WS(rs, 2)];
Tn = Tl - Tm;
TI = Tl + Tm;
Tp = Rp[WS(rs, 2)];
Tq = Rm[WS(rs, 2)];
Tr = Tp + Tq;
TG = Tp - Tq;
}
Tv = Tt - Tu;
TX = TA + Tz;
Tk = W[6];
To = W[7];
Ts = FNMS(To, Tr, Tk * Tn);
TY = FMA(Tk, Tr, To * Tn);
{
E Ty, TC, TF, TH;
Ty = W[0];
TC = W[1];
TE = FNMS(TC, TD, Ty * TB);
T1a = FMA(TC, TB, Ty * TD);
TF = W[8];
TH = W[9];
TJ = FMA(TF, TG, TH * TI);
T19 = FNMS(TH, TG, TF * TI);
}
T1l = TJ + TE;
T1m = T1a - T19;
}
{
E T4, TO, T8, TM, Td, TT, Th, TR;
{
E T2, T3, T6, T7;
T2 = Ip[WS(rs, 1)];
T3 = Im[WS(rs, 1)];
T4 = T2 - T3;
TO = T2 + T3;
T6 = Rp[WS(rs, 1)];
T7 = Rm[WS(rs, 1)];
T8 = T6 + T7;
TM = T6 - T7;
}
{
E Tb, Tc, Tf, Tg;
Tb = Ip[WS(rs, 3)];
Tc = Im[WS(rs, 3)];
Td = Tb - Tc;
TT = Tb + Tc;
Tf = Rp[WS(rs, 3)];
Tg = Rm[WS(rs, 3)];
Th = Tf + Tg;
TR = Tf - Tg;
}
{
E T1, T5, Ta, Te;
T1 = W[2];
T5 = W[3];
T9 = FNMS(T5, T8, T1 * T4);
T10 = FMA(T1, T8, T5 * T4);
Ta = W[10];
Te = W[11];
Ti = FNMS(Te, Th, Ta * Td);
T11 = FMA(Ta, Th, Te * Td);
{
E TL, TN, TQ, TS;
TL = W[4];
TN = W[5];
TP = FMA(TL, TM, TN * TO);
T16 = FNMS(TN, TM, TL * TO);
TQ = W[12];
TS = W[13];
TU = FMA(TQ, TR, TS * TT);
T17 = FNMS(TS, TR, TQ * TT);
}
T1i = T17 - T16;
T1j = TP - TU;
}
}
{
E T1h, T1t, T1w, T1y, T1o, T1s, T1r, T1x;
{
E T1f, T1g, T1u, T1v;
T1f = Tv - Ts;
T1g = T10 - T11;
T1h = KP500000000 * (T1f - T1g);
T1t = KP500000000 * (T1g + T1f);
T1u = T1i - T1j;
T1v = T1l + T1m;
T1w = KP353553390 * (T1u - T1v);
T1y = KP353553390 * (T1u + T1v);
}
{
E T1k, T1n, T1p, T1q;
T1k = T1i + T1j;
T1n = T1l - T1m;
T1o = KP353553390 * (T1k + T1n);
T1s = KP353553390 * (T1n - T1k);
T1p = TX - TY;
T1q = T9 - Ti;
T1r = KP500000000 * (T1p - T1q);
T1x = KP500000000 * (T1p + T1q);
}
Ip[WS(rs, 1)] = T1h + T1o;
Rp[WS(rs, 1)] = T1x + T1y;
Im[WS(rs, 2)] = T1o - T1h;
Rm[WS(rs, 2)] = T1x - T1y;
Rm[0] = T1r - T1s;
Im[0] = T1w - T1t;
Rp[WS(rs, 3)] = T1r + T1s;
Ip[WS(rs, 3)] = T1t + T1w;
}
{
E Tx, T15, T1c, T1e, TW, T14, T13, T1d;
{
E Tj, Tw, T18, T1b;
Tj = T9 + Ti;
Tw = Ts + Tv;
Tx = Tj + Tw;
T15 = Tw - Tj;
T18 = T16 + T17;
T1b = T19 + T1a;
T1c = T18 - T1b;
T1e = T18 + T1b;
}
{
E TK, TV, TZ, T12;
TK = TE - TJ;
TV = TP + TU;
TW = TK - TV;
T14 = TV + TK;
TZ = TX + TY;
T12 = T10 + T11;
T13 = TZ - T12;
T1d = TZ + T12;
}
Ip[0] = KP500000000 * (Tx + TW);
Rp[0] = KP500000000 * (T1d + T1e);
Im[WS(rs, 3)] = KP500000000 * (TW - Tx);
Rm[WS(rs, 3)] = KP500000000 * (T1d - T1e);
Rm[WS(rs, 1)] = KP500000000 * (T13 - T14);
Im[WS(rs, 1)] = KP500000000 * (T1c - T15);
Rp[WS(rs, 2)] = KP500000000 * (T13 + T14);
Ip[WS(rs, 2)] = KP500000000 * (T15 + T1c);
}
}
}
}
static void hc2r_9(const R *ri, const R *ii, R *O, stride ris, stride iis, stride os, int v, int ivs, int ovs)
{
DK(KP984807753, +0.984807753012208059366743024589523013670643252);
DK(KP173648177, +0.173648177666930348851716626769314796000375677);
DK(KP300767466, +0.300767466360870593278543795225003852144476517);
DK(KP1_705737063, +1.705737063904886419256501927880148143872040591);
DK(KP642787609, +0.642787609686539326322643409907263432907559884);
DK(KP766044443, +0.766044443118978035202392650555416673935832457);
DK(KP1_326827896, +1.326827896337876792410842639271782594433726619);
DK(KP1_113340798, +1.113340798452838732905825904094046265936583811);
DK(KP500000000, +0.500000000000000000000000000000000000000000000);
DK(KP866025403, +0.866025403784438646763723170752936183471402627);
DK(KP2_000000000, +2.000000000000000000000000000000000000000000000);
DK(KP1_732050807, +1.732050807568877293527446341505872366942805254);
int i;
for (i = v; i > 0; i = i - 1, ri = ri + ivs, ii = ii + ivs, O = O + ovs) {
E T3, Tq, Tc, Tk, Tj, T8, Tm, Ts, Th, Tr, Tw, Tx;
{
E Tb, T1, T2, T9, Ta;
Ta = ii[WS(iis, 3)];
Tb = KP1_732050807 * Ta;
T1 = ri[0];
T2 = ri[WS(ris, 3)];
T9 = T1 - T2;
T3 = FMA(KP2_000000000, T2, T1);
Tq = T9 + Tb;
Tc = T9 - Tb;
}
{
E T4, T7, Ti, Tg, Tl, Td;
T4 = ri[WS(ris, 1)];
Tk = ii[WS(iis, 1)];
{
E T5, T6, Te, Tf;
T5 = ri[WS(ris, 4)];
T6 = ri[WS(ris, 2)];
T7 = T5 + T6;
Ti = KP866025403 * (T5 - T6);
Te = ii[WS(iis, 4)];
Tf = ii[WS(iis, 2)];
Tg = KP866025403 * (Te + Tf);
Tj = Tf - Te;
}
T8 = T4 + T7;
Tl = FMA(KP500000000, Tj, Tk);
Tm = Ti + Tl;
Ts = Tl - Ti;
Td = FNMS(KP500000000, T7, T4);
Th = Td - Tg;
Tr = Td + Tg;
}
O[0] = FMA(KP2_000000000, T8, T3);
Tw = T3 - T8;
Tx = KP1_732050807 * (Tk - Tj);
O[WS(os, 3)] = Tw - Tx;
O[WS(os, 6)] = Tw + Tx;
{
E Tp, Tn, To, Tv, Tt, Tu;
Tp = FMA(KP1_113340798, Th, KP1_326827896 * Tm);
Tn = FNMS(KP642787609, Tm, KP766044443 * Th);
To = Tc - Tn;
O[WS(os, 1)] = FMA(KP2_000000000, Tn, Tc);
O[WS(os, 7)] = To + Tp;
O[WS(os, 4)] = To - Tp;
Tv = FMA(KP1_705737063, Tr, KP300767466 * Ts);
Tt = FNMS(KP984807753, Ts, KP173648177 * Tr);
Tu = Tq - Tt;
O[WS(os, 2)] = FMA(KP2_000000000, Tt, Tq);
O[WS(os, 8)] = Tu + Tv;
O[WS(os, 5)] = Tu - Tv;
}
}
}
static void hb2_16(R *cr, R *ci, const R *W, stride rs, INT mb, INT me, INT ms)
{
DK(KP382683432, +0.382683432365089771728459984030398866761344562);
DK(KP923879532, +0.923879532511286756128183189396788286822416626);
DK(KP707106781, +0.707106781186547524400844362104849039284835938);
{
INT m;
for (m = mb, W = W + ((mb - 1) * 8); m < me; m = m + 1, cr = cr + ms, ci = ci - ms, W = W + 8, MAKE_VOLATILE_STRIDE(32, rs)) {
E Tv, Ty, T1l, T1n, T1p, T1t, T27, T25, Tz, Tw, TB, T21, T1P, T1H, T1X;
E T17, T1L, T1N, T1v, T1w, T1x, T1B, T2F, T2T, T2b, T2R, T3j, T3x, T35, T3t;
{
E TA, T1J, T15, T1G, Tx, T1K, T16, T1F;
{
E T1m, T1s, T1o, T1r;
Tv = W[0];
Ty = W[1];
T1l = W[2];
T1n = W[3];
T1m = Tv * T1l;
T1s = Ty * T1l;
T1o = Ty * T1n;
T1r = Tv * T1n;
T1p = T1m + T1o;
T1t = T1r - T1s;
T27 = T1r + T1s;
T25 = T1m - T1o;
Tz = W[5];
TA = Ty * Tz;
T1J = T1l * Tz;
T15 = Tv * Tz;
T1G = T1n * Tz;
Tw = W[4];
Tx = Tv * Tw;
T1K = T1n * Tw;
T16 = Ty * Tw;
T1F = T1l * Tw;
}
TB = Tx - TA;
T21 = T1J + T1K;
T1P = T15 - T16;
T1H = T1F + T1G;
T1X = T1F - T1G;
T17 = T15 + T16;
T1L = T1J - T1K;
T1N = Tx + TA;
T1v = W[6];
T1w = W[7];
T1x = FMA(Tv, T1v, Ty * T1w);
T1B = FNMS(Ty, T1v, Tv * T1w);
{
E T2D, T2E, T29, T2a;
T2D = T25 * Tz;
T2E = T27 * Tw;
T2F = T2D + T2E;
T2T = T2D - T2E;
T29 = T25 * Tw;
T2a = T27 * Tz;
T2b = T29 - T2a;
T2R = T29 + T2a;
}
{
E T3h, T3i, T33, T34;
T3h = T1p * Tz;
T3i = T1t * Tw;
T3j = T3h + T3i;
T3x = T3h - T3i;
T33 = T1p * Tw;
T34 = T1t * Tz;
T35 = T33 - T34;
T3t = T33 + T34;
}
}
{
E T7, T36, T3k, TC, T1f, T2e, T2I, T1Q, Te, TJ, T1R, T18, T2L, T37, T2l;
E T3l, Tm, T1T, TT, T1h, T2A, T2N, T3b, T3n, Tt, T1U, T12, T1i, T2t, T2O;
E T3e, T3o;
{
E T3, T2c, T1e, T2d, T6, T2G, T1b, T2H;
{
E T1, T2, T1c, T1d;
T1 = cr[0];
T2 = ci[WS(rs, 7)];
T3 = T1 + T2;
T2c = T1 - T2;
T1c = ci[WS(rs, 11)];
T1d = cr[WS(rs, 12)];
T1e = T1c - T1d;
T2d = T1c + T1d;
}
{
E T4, T5, T19, T1a;
T4 = cr[WS(rs, 4)];
T5 = ci[WS(rs, 3)];
T6 = T4 + T5;
T2G = T4 - T5;
T19 = ci[WS(rs, 15)];
T1a = cr[WS(rs, 8)];
T1b = T19 - T1a;
T2H = T19 + T1a;
}
T7 = T3 + T6;
T36 = T2c + T2d;
T3k = T2H - T2G;
TC = T3 - T6;
T1f = T1b - T1e;
T2e = T2c - T2d;
T2I = T2G + T2H;
T1Q = T1b + T1e;
}
{
E Ta, T2f, TI, T2g, Td, T2i, TF, T2j;
{
E T8, T9, TG, TH;
T8 = cr[WS(rs, 2)];
T9 = ci[WS(rs, 5)];
Ta = T8 + T9;
T2f = T8 - T9;
TG = ci[WS(rs, 13)];
TH = cr[WS(rs, 10)];
TI = TG - TH;
T2g = TG + TH;
}
{
E Tb, Tc, TD, TE;
Tb = ci[WS(rs, 1)];
Tc = cr[WS(rs, 6)];
Td = Tb + Tc;
T2i = Tb - Tc;
TD = ci[WS(rs, 9)];
TE = cr[WS(rs, 14)];
TF = TD - TE;
T2j = TD + TE;
}
Te = Ta + Td;
TJ = TF - TI;
T1R = TI + TF;
T18 = Ta - Td;
{
E T2J, T2K, T2h, T2k;
T2J = T2f + T2g;
T2K = T2i + T2j;
T2L = KP707106781 * (T2J - T2K);
T37 = KP707106781 * (T2J + T2K);
T2h = T2f - T2g;
T2k = T2i - T2j;
T2l = KP707106781 * (T2h + T2k);
T3l = KP707106781 * (T2h - T2k);
}
}
{
E Ti, T2x, TR, T2y, Tl, T2u, TO, T2v, TL, TS;
{
E Tg, Th, TP, TQ;
Tg = cr[WS(rs, 1)];
Th = ci[WS(rs, 6)];
Ti = Tg + Th;
T2x = Tg - Th;
TP = ci[WS(rs, 10)];
TQ = cr[WS(rs, 13)];
TR = TP - TQ;
T2y = TP + TQ;
}
{
E Tj, Tk, TM, TN;
Tj = cr[WS(rs, 5)];
Tk = ci[WS(rs, 2)];
Tl = Tj + Tk;
T2u = Tj - Tk;
TM = ci[WS(rs, 14)];
TN = cr[WS(rs, 9)];
TO = TM - TN;
T2v = TM + TN;
}
Tm = Ti + Tl;
T1T = TO + TR;
TL = Ti - Tl;
TS = TO - TR;
TT = TL - TS;
T1h = TL + TS;
{
E T2w, T2z, T39, T3a;
T2w = T2u + T2v;
T2z = T2x - T2y;
T2A = FMA(KP923879532, T2w, KP382683432 * T2z);
T2N = FNMS(KP382683432, T2w, KP923879532 * T2z);
T39 = T2x + T2y;
T3a = T2v - T2u;
T3b = FNMS(KP923879532, T3a, KP382683432 * T39);
T3n = FMA(KP382683432, T3a, KP923879532 * T39);
}
}
{
E Tp, T2q, T10, T2r, Ts, T2n, TX, T2o, TU, T11;
{
E Tn, To, TY, TZ;
Tn = ci[0];
To = cr[WS(rs, 7)];
Tp = Tn + To;
T2q = Tn - To;
TY = ci[WS(rs, 12)];
TZ = cr[WS(rs, 11)];
T10 = TY - TZ;
T2r = TY + TZ;
}
{
E Tq, Tr, TV, TW;
Tq = cr[WS(rs, 3)];
Tr = ci[WS(rs, 4)];
Ts = Tq + Tr;
T2n = Tq - Tr;
TV = ci[WS(rs, 8)];
TW = cr[WS(rs, 15)];
TX = TV - TW;
T2o = TV + TW;
}
Tt = Tp + Ts;
T1U = TX + T10;
TU = Tp - Ts;
T11 = TX - T10;
T12 = TU + T11;
T1i = T11 - TU;
{
E T2p, T2s, T3c, T3d;
T2p = T2n - T2o;
T2s = T2q - T2r;
T2t = FNMS(KP382683432, T2s, KP923879532 * T2p);
T2O = FMA(KP382683432, T2p, KP923879532 * T2s);
T3c = T2q + T2r;
T3d = T2n + T2o;
T3e = FNMS(KP923879532, T3d, KP382683432 * T3c);
T3o = FMA(KP382683432, T3d, KP923879532 * T3c);
}
}
{
E Tf, Tu, T1O, T1S, T1V, T1W;
Tf = T7 + Te;
Tu = Tm + Tt;
T1O = Tf - Tu;
T1S = T1Q + T1R;
T1V = T1T + T1U;
T1W = T1S - T1V;
cr[0] = Tf + Tu;
ci[0] = T1S + T1V;
cr[WS(rs, 8)] = FNMS(T1P, T1W, T1N * T1O);
ci[WS(rs, 8)] = FMA(T1P, T1O, T1N * T1W);
}
{
E T3g, T3r, T3q, T3s;
{
E T38, T3f, T3m, T3p;
T38 = T36 - T37;
T3f = T3b + T3e;
T3g = T38 - T3f;
T3r = T38 + T3f;
T3m = T3k + T3l;
T3p = T3n - T3o;
T3q = T3m - T3p;
T3s = T3m + T3p;
}
cr[WS(rs, 11)] = FNMS(T3j, T3q, T35 * T3g);
ci[WS(rs, 11)] = FMA(T3j, T3g, T35 * T3q);
cr[WS(rs, 3)] = FNMS(T1n, T3s, T1l * T3r);
ci[WS(rs, 3)] = FMA(T1n, T3r, T1l * T3s);
}
{
E T3w, T3B, T3A, T3C;
{
E T3u, T3v, T3y, T3z;
T3u = T36 + T37;
T3v = T3n + T3o;
T3w = T3u - T3v;
T3B = T3u + T3v;
T3y = T3k - T3l;
T3z = T3b - T3e;
T3A = T3y + T3z;
T3C = T3y - T3z;
}
cr[WS(rs, 7)] = FNMS(T3x, T3A, T3t * T3w);
ci[WS(rs, 7)] = FMA(T3t, T3A, T3x * T3w);
cr[WS(rs, 15)] = FNMS(T1w, T3C, T1v * T3B);
ci[WS(rs, 15)] = FMA(T1v, T3C, T1w * T3B);
}
{
E T14, T1q, T1k, T1u;
{
E TK, T13, T1g, T1j;
TK = TC + TJ;
T13 = KP707106781 * (TT + T12);
T14 = TK - T13;
T1q = TK + T13;
T1g = T18 + T1f;
T1j = KP707106781 * (T1h + T1i);
T1k = T1g - T1j;
T1u = T1g + T1j;
}
cr[WS(rs, 10)] = FNMS(T17, T1k, TB * T14);
ci[WS(rs, 10)] = FMA(T17, T14, TB * T1k);
cr[WS(rs, 2)] = FNMS(T1t, T1u, T1p * T1q);
ci[WS(rs, 2)] = FMA(T1t, T1q, T1p * T1u);
}
{
E T1A, T1I, T1E, T1M;
{
E T1y, T1z, T1C, T1D;
T1y = TC - TJ;
T1z = KP707106781 * (T1i - T1h);
T1A = T1y - T1z;
T1I = T1y + T1z;
T1C = T1f - T18;
T1D = KP707106781 * (TT - T12);
T1E = T1C - T1D;
T1M = T1C + T1D;
}
cr[WS(rs, 14)] = FNMS(T1B, T1E, T1x * T1A);
ci[WS(rs, 14)] = FMA(T1x, T1E, T1B * T1A);
cr[WS(rs, 6)] = FNMS(T1L, T1M, T1H * T1I);
ci[WS(rs, 6)] = FMA(T1H, T1M, T1L * T1I);
}
{
E T2C, T2S, T2Q, T2U;
{
E T2m, T2B, T2M, T2P;
T2m = T2e - T2l;
T2B = T2t - T2A;
T2C = T2m - T2B;
T2S = T2m + T2B;
T2M = T2I - T2L;
T2P = T2N - T2O;
T2Q = T2M - T2P;
T2U = T2M + T2P;
}
cr[WS(rs, 13)] = FNMS(T2F, T2Q, T2b * T2C);
ci[WS(rs, 13)] = FMA(T2F, T2C, T2b * T2Q);
cr[WS(rs, 5)] = FNMS(T2T, T2U, T2R * T2S);
ci[WS(rs, 5)] = FMA(T2T, T2S, T2R * T2U);
}
{
E T2X, T31, T30, T32;
{
E T2V, T2W, T2Y, T2Z;
T2V = T2e + T2l;
T2W = T2N + T2O;
T2X = T2V - T2W;
T31 = T2V + T2W;
T2Y = T2I + T2L;
T2Z = T2A + T2t;
T30 = T2Y - T2Z;
T32 = T2Y + T2Z;
}
cr[WS(rs, 9)] = FNMS(Tz, T30, Tw * T2X);
ci[WS(rs, 9)] = FMA(Tw, T30, Tz * T2X);
cr[WS(rs, 1)] = FNMS(Ty, T32, Tv * T31);
ci[WS(rs, 1)] = FMA(Tv, T32, Ty * T31);
}
{
E T20, T26, T24, T28;
{
E T1Y, T1Z, T22, T23;
T1Y = T7 - Te;
T1Z = T1U - T1T;
T20 = T1Y - T1Z;
T26 = T1Y + T1Z;
T22 = T1Q - T1R;
T23 = Tm - Tt;
T24 = T22 - T23;
T28 = T23 + T22;
}
cr[WS(rs, 12)] = FNMS(T21, T24, T1X * T20);
ci[WS(rs, 12)] = FMA(T1X, T24, T21 * T20);
cr[WS(rs, 4)] = FNMS(T27, T28, T25 * T26);
ci[WS(rs, 4)] = FMA(T25, T28, T27 * T26);
}
}
}
}
}
static void q1_3(float *rio, float *iio, const float *W, stride rs, stride vs, INT mb, INT me, INT ms)
{
DK(KP866025403, +0.866025403784438646763723170752936183471402627);
DK(KP500000000, +0.500000000000000000000000000000000000000000000);
INT m;
for (m = mb, W = W + (mb * 4); m < me; m = m + 1, rio = rio + ms, iio = iio + ms, W = W + 4, MAKE_VOLATILE_STRIDE(rs), MAKE_VOLATILE_STRIDE(vs)) {
E T1, T4, T6, Tc, Td, Te, T9, Tf, Tl, To, Tq, Tw, Tx, Ty, Tt;
E Tz, TR, TS, TN, TT, TF, TI, TK, TQ;
{
E T2, T3, Tr, Ts;
T1 = rio[0];
T2 = rio[WS(rs, 1)];
T3 = rio[WS(rs, 2)];
T4 = T2 + T3;
T6 = FNMS(KP500000000, T4, T1);
Tc = KP866025403 * (T3 - T2);
{
E T7, T8, Tm, Tn;
Td = iio[0];
T7 = iio[WS(rs, 1)];
T8 = iio[WS(rs, 2)];
Te = T7 + T8;
T9 = KP866025403 * (T7 - T8);
Tf = FNMS(KP500000000, Te, Td);
Tl = rio[WS(vs, 1)];
Tm = rio[WS(vs, 1) + WS(rs, 1)];
Tn = rio[WS(vs, 1) + WS(rs, 2)];
To = Tm + Tn;
Tq = FNMS(KP500000000, To, Tl);
Tw = KP866025403 * (Tn - Tm);
}
Tx = iio[WS(vs, 1)];
Tr = iio[WS(vs, 1) + WS(rs, 1)];
Ts = iio[WS(vs, 1) + WS(rs, 2)];
Ty = Tr + Ts;
Tt = KP866025403 * (Tr - Ts);
Tz = FNMS(KP500000000, Ty, Tx);
{
E TL, TM, TG, TH;
TR = iio[WS(vs, 2)];
TL = iio[WS(vs, 2) + WS(rs, 1)];
TM = iio[WS(vs, 2) + WS(rs, 2)];
TS = TL + TM;
TN = KP866025403 * (TL - TM);
TT = FNMS(KP500000000, TS, TR);
TF = rio[WS(vs, 2)];
TG = rio[WS(vs, 2) + WS(rs, 1)];
TH = rio[WS(vs, 2) + WS(rs, 2)];
TI = TG + TH;
TK = FNMS(KP500000000, TI, TF);
TQ = KP866025403 * (TH - TG);
}
}
rio[0] = T1 + T4;
iio[0] = Td + Te;
rio[WS(rs, 1)] = Tl + To;
iio[WS(rs, 1)] = Tx + Ty;
iio[WS(rs, 2)] = TR + TS;
rio[WS(rs, 2)] = TF + TI;
{
E Ta, Tg, T5, Tb;
Ta = T6 + T9;
Tg = Tc + Tf;
T5 = W[0];
Tb = W[1];
rio[WS(vs, 1)] = FMA(T5, Ta, Tb * Tg);
iio[WS(vs, 1)] = FNMS(Tb, Ta, T5 * Tg);
}
{
E TW, TY, TV, TX;
TW = TK - TN;
TY = TT - TQ;
TV = W[2];
TX = W[3];
rio[WS(vs, 2) + WS(rs, 2)] = FMA(TV, TW, TX * TY);
iio[WS(vs, 2) + WS(rs, 2)] = FNMS(TX, TW, TV * TY);
}
{
E TC, TE, TB, TD;
TC = Tq - Tt;
TE = Tz - Tw;
TB = W[2];
TD = W[3];
rio[WS(vs, 2) + WS(rs, 1)] = FMA(TB, TC, TD * TE);
iio[WS(vs, 2) + WS(rs, 1)] = FNMS(TD, TC, TB * TE);
}
{
E Tu, TA, Tp, Tv;
Tu = Tq + Tt;
TA = Tw + Tz;
Tp = W[0];
Tv = W[1];
rio[WS(vs, 1) + WS(rs, 1)] = FMA(Tp, Tu, Tv * TA);
iio[WS(vs, 1) + WS(rs, 1)] = FNMS(Tv, Tu, Tp * TA);
}
{
E TO, TU, TJ, TP;
TO = TK + TN;
TU = TQ + TT;
TJ = W[0];
TP = W[1];
rio[WS(vs, 1) + WS(rs, 2)] = FMA(TJ, TO, TP * TU);
iio[WS(vs, 1) + WS(rs, 2)] = FNMS(TP, TO, TJ * TU);
}
{
E Ti, Tk, Th, Tj;
Ti = T6 - T9;
Tk = Tf - Tc;
Th = W[2];
Tj = W[3];
rio[WS(vs, 2)] = FMA(Th, Ti, Tj * Tk);
iio[WS(vs, 2)] = FNMS(Tj, Ti, Th * Tk);
}
}
}
static void hc2r_15(const R *ri, const R *ii, R *O, stride ris, stride iis, stride os, INT v, INT ivs, INT ovs)
{
DK(KP559016994, +0.559016994374947424102293417182819058860154590);
DK(KP1_902113032, +1.902113032590307144232878666758764286811397268);
DK(KP250000000, +0.250000000000000000000000000000000000000000000);
DK(KP866025403, +0.866025403784438646763723170752936183471402627);
DK(KP1_118033988, +1.118033988749894848204586834365638117720309180);
DK(KP618033988, +0.618033988749894848204586834365638117720309180);
DK(KP500000000, +0.500000000000000000000000000000000000000000000);
DK(KP1_732050807, +1.732050807568877293527446341505872366942805254);
DK(KP2_000000000, +2.000000000000000000000000000000000000000000000);
INT i;
for (i = v; i > 0; i = i - 1, ri = ri + ivs, ii = ii + ivs, O = O + ovs, MAKE_VOLATILE_STRIDE(ris), MAKE_VOLATILE_STRIDE(iis), MAKE_VOLATILE_STRIDE(os)) {
E TL, Tz, TM, TK;
{
E T3, Th, Tt, TD, TI, TH, TY, TC, TZ, Tu, Tm, Tv, Tr, Te, TW;
E Tg, T1, T2, T12, T10, TV;
Tg = ii[WS(iis, 5)];
T1 = ri[0];
T2 = ri[WS(ris, 5)];
{
E T4, TA, T9, TF, T7, Tj, Tc, Tk, TG, Tq, Tf, Tl, TB;
T4 = ri[WS(ris, 3)];
TA = ii[WS(iis, 3)];
T9 = ri[WS(ris, 6)];
Tf = T1 - T2;
T3 = FMA(KP2_000000000, T2, T1);
TF = ii[WS(iis, 6)];
{
E Ta, Tb, T5, T6, To, Tp;
T5 = ri[WS(ris, 7)];
T6 = ri[WS(ris, 2)];
Th = FMA(KP1_732050807, Tg, Tf);
Tt = FNMS(KP1_732050807, Tg, Tf);
Ta = ri[WS(ris, 4)];
TD = T5 - T6;
T7 = T5 + T6;
Tb = ri[WS(ris, 1)];
To = ii[WS(iis, 4)];
Tp = ii[WS(iis, 1)];
Tj = ii[WS(iis, 7)];
Tc = Ta + Tb;
TI = Ta - Tb;
Tk = ii[WS(iis, 2)];
TG = Tp - To;
Tq = To + Tp;
}
Tl = Tj - Tk;
TB = Tj + Tk;
TH = FNMS(KP500000000, TG, TF);
TY = TG + TF;
TC = FMA(KP500000000, TB, TA);
TZ = TA - TB;
{
E Ti, T8, Td, Tn;
Ti = FNMS(KP2_000000000, T4, T7);
T8 = T4 + T7;
Td = T9 + Tc;
Tn = FNMS(KP2_000000000, T9, Tc);
Tu = FNMS(KP1_732050807, Tl, Ti);
Tm = FMA(KP1_732050807, Tl, Ti);
Tv = FNMS(KP1_732050807, Tq, Tn);
Tr = FMA(KP1_732050807, Tq, Tn);
Te = T8 + Td;
TW = T8 - Td;
}
}
T12 = FMA(KP618033988, TY, TZ);
T10 = FNMS(KP618033988, TZ, TY);
TV = FNMS(KP500000000, Te, T3);
O[0] = FMA(KP2_000000000, Te, T3);
{
E TJ, TE, TT, TP, TU, TS, Ty, Tw, Tx;
{
E TO, Ts, TQ, TN, TR, T11, TX;
TO = Tr - Tm;
Ts = Tm + Tr;
T11 = FMA(KP1_118033988, TW, TV);
TX = FNMS(KP1_118033988, TW, TV);
TQ = FNMS(KP866025403, TI, TH);
TJ = FMA(KP866025403, TI, TH);
TN = FMA(KP250000000, Ts, Th);
O[WS(os, 6)] = FNMS(KP1_902113032, T12, T11);
O[WS(os, 9)] = FMA(KP1_902113032, T12, T11);
O[WS(os, 12)] = FMA(KP1_902113032, T10, TX);
O[WS(os, 3)] = FNMS(KP1_902113032, T10, TX);
TR = FNMS(KP866025403, TD, TC);
TE = FMA(KP866025403, TD, TC);
O[WS(os, 5)] = Th - Ts;
TT = FMA(KP559016994, TO, TN);
TP = FNMS(KP559016994, TO, TN);
TU = FMA(KP618033988, TQ, TR);
TS = FNMS(KP618033988, TR, TQ);
}
Ty = Tv - Tu;
Tw = Tu + Tv;
O[WS(os, 14)] = FMA(KP1_902113032, TU, TT);
O[WS(os, 11)] = FNMS(KP1_902113032, TU, TT);
O[WS(os, 2)] = FMA(KP1_902113032, TS, TP);
O[WS(os, 8)] = FNMS(KP1_902113032, TS, TP);
Tx = FMA(KP250000000, Tw, Tt);
O[WS(os, 10)] = Tt - Tw;
TL = FNMS(KP559016994, Ty, Tx);
Tz = FMA(KP559016994, Ty, Tx);
TM = FNMS(KP618033988, TE, TJ);
TK = FMA(KP618033988, TJ, TE);
}
}
O[WS(os, 7)] = FMA(KP1_902113032, TM, TL);
O[WS(os, 13)] = FNMS(KP1_902113032, TM, TL);
O[WS(os, 4)] = FMA(KP1_902113032, TK, Tz);
O[WS(os, 1)] = FNMS(KP1_902113032, TK, Tz);
}
}
static void r2cbIII_32(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
{
DK(KP1_913880671, +1.913880671464417729871595773960539938965698411);
DK(KP580569354, +0.580569354508924735272384751634790549382952557);
DK(KP942793473, +0.942793473651995297112775251810508755314920638);
DK(KP1_763842528, +1.763842528696710059425513727320776699016885241);
DK(KP1_546020906, +1.546020906725473921621813219516939601942082586);
DK(KP1_268786568, +1.268786568327290996430343226450986741351374190);
DK(KP196034280, +0.196034280659121203988391127777283691722273346);
DK(KP1_990369453, +1.990369453344393772489673906218959843150949737);
DK(KP765366864, +0.765366864730179543456919968060797733522689125);
DK(KP1_847759065, +1.847759065022573512256366378793576573644833252);
DK(KP1_961570560, +1.961570560806460898252364472268478073947867462);
DK(KP390180644, +0.390180644032256535696569736954044481855383236);
DK(KP1_111140466, +1.111140466039204449485661627897065748749874382);
DK(KP1_662939224, +1.662939224605090474157576755235811513477121624);
DK(KP1_414213562, +1.414213562373095048801688724209698078569671875);
DK(KP2_000000000, +2.000000000000000000000000000000000000000000000);
DK(KP382683432, +0.382683432365089771728459984030398866761344562);
DK(KP923879532, +0.923879532511286756128183189396788286822416626);
DK(KP707106781, +0.707106781186547524400844362104849039284835938);
{
INT i;
for (i = v; i > 0; i = i - 1, R0 = R0 + ovs, R1 = R1 + ovs, Cr = Cr + ivs, Ci = Ci + ivs, MAKE_VOLATILE_STRIDE(128, rs), MAKE_VOLATILE_STRIDE(128, csr), MAKE_VOLATILE_STRIDE(128, csi)) {
E T7, T2i, T2F, Tz, T1k, T1I, T1Z, T1x, Te, T22, T2E, T2j, T1f, T1y, TK;
E T1J, Tm, T2B, TW, T1a, T1C, T1L, T28, T2l, Tt, T2A, T17, T1b, T1F, T1M;
E T2d, T2m;
{
E T3, Tv, T1j, T2h, T6, T1g, Ty, T2g;
{
E T1, T2, T1h, T1i;
T1 = Cr[0];
T2 = Cr[WS(csr, 15)];
T3 = T1 + T2;
Tv = T1 - T2;
T1h = Ci[0];
T1i = Ci[WS(csi, 15)];
T1j = T1h + T1i;
T2h = T1i - T1h;
}
{
E T4, T5, Tw, Tx;
T4 = Cr[WS(csr, 8)];
T5 = Cr[WS(csr, 7)];
T6 = T4 + T5;
T1g = T4 - T5;
Tw = Ci[WS(csi, 8)];
Tx = Ci[WS(csi, 7)];
Ty = Tw + Tx;
T2g = Tw - Tx;
}
T7 = T3 + T6;
T2i = T2g + T2h;
T2F = T2h - T2g;
Tz = Tv - Ty;
T1k = T1g + T1j;
T1I = T1g - T1j;
T1Z = T3 - T6;
T1x = Tv + Ty;
}
{
E Ta, TA, TD, T21, Td, TF, TI, T20;
{
E T8, T9, TB, TC;
T8 = Cr[WS(csr, 4)];
T9 = Cr[WS(csr, 11)];
Ta = T8 + T9;
TA = T8 - T9;
TB = Ci[WS(csi, 4)];
TC = Ci[WS(csi, 11)];
TD = TB + TC;
T21 = TB - TC;
}
{
E Tb, Tc, TG, TH;
Tb = Cr[WS(csr, 3)];
Tc = Cr[WS(csr, 12)];
Td = Tb + Tc;
TF = Tb - Tc;
TG = Ci[WS(csi, 3)];
TH = Ci[WS(csi, 12)];
TI = TG + TH;
T20 = TH - TG;
}
Te = Ta + Td;
T22 = T20 - T21;
T2E = T21 + T20;
T2j = Ta - Td;
{
E T1d, T1e, TE, TJ;
T1d = TA + TD;
T1e = TF + TI;
T1f = KP707106781 * (T1d - T1e);
T1y = KP707106781 * (T1d + T1e);
TE = TA - TD;
TJ = TF - TI;
TK = KP707106781 * (TE + TJ);
T1J = KP707106781 * (TE - TJ);
}
}
{
E Ti, TM, TU, T25, Tl, TR, TP, T26, TQ, TV;
{
E Tg, Th, TS, TT;
Tg = Cr[WS(csr, 2)];
Th = Cr[WS(csr, 13)];
Ti = Tg + Th;
TM = Tg - Th;
TS = Ci[WS(csi, 2)];
TT = Ci[WS(csi, 13)];
TU = TS + TT;
T25 = TS - TT;
}
{
E Tj, Tk, TN, TO;
Tj = Cr[WS(csr, 10)];
Tk = Cr[WS(csr, 5)];
Tl = Tj + Tk;
TR = Tj - Tk;
TN = Ci[WS(csi, 10)];
TO = Ci[WS(csi, 5)];
TP = TN + TO;
T26 = TN - TO;
}
Tm = Ti + Tl;
T2B = T26 + T25;
TQ = TM - TP;
TV = TR + TU;
TW = FNMS(KP382683432, TV, KP923879532 * TQ);
T1a = FMA(KP382683432, TQ, KP923879532 * TV);
{
E T1A, T1B, T24, T27;
T1A = TM + TP;
T1B = TU - TR;
T1C = FNMS(KP923879532, T1B, KP382683432 * T1A);
T1L = FMA(KP923879532, T1A, KP382683432 * T1B);
T24 = Ti - Tl;
T27 = T25 - T26;
T28 = T24 - T27;
T2l = T24 + T27;
}
}
{
E Tp, TX, T15, T2a, Ts, T12, T10, T2b, T11, T16;
{
E Tn, To, T13, T14;
Tn = Cr[WS(csr, 1)];
To = Cr[WS(csr, 14)];
Tp = Tn + To;
TX = Tn - To;
T13 = Ci[WS(csi, 1)];
T14 = Ci[WS(csi, 14)];
T15 = T13 + T14;
T2a = T14 - T13;
}
{
E Tq, Tr, TY, TZ;
Tq = Cr[WS(csr, 6)];
Tr = Cr[WS(csr, 9)];
Ts = Tq + Tr;
T12 = Tq - Tr;
TY = Ci[WS(csi, 6)];
TZ = Ci[WS(csi, 9)];
T10 = TY + TZ;
T2b = TY - TZ;
}
Tt = Tp + Ts;
T2A = T2b + T2a;
T11 = TX - T10;
T16 = T12 - T15;
T17 = FMA(KP923879532, T11, KP382683432 * T16);
T1b = FNMS(KP382683432, T11, KP923879532 * T16);
{
E T1D, T1E, T29, T2c;
T1D = TX + T10;
T1E = T12 + T15;
T1F = FNMS(KP923879532, T1E, KP382683432 * T1D);
T1M = FMA(KP923879532, T1D, KP382683432 * T1E);
T29 = Tp - Ts;
T2c = T2a - T2b;
T2d = T29 + T2c;
T2m = T2c - T29;
}
}
{
E Tf, Tu, T2L, T2M, T2N, T2O;
Tf = T7 + Te;
Tu = Tm + Tt;
T2L = Tf - Tu;
T2M = T2B + T2A;
T2N = T2F - T2E;
T2O = T2M + T2N;
R0[0] = KP2_000000000 * (Tf + Tu);
R0[WS(rs, 8)] = KP2_000000000 * (T2N - T2M);
R0[WS(rs, 4)] = KP1_414213562 * (T2L + T2O);
R0[WS(rs, 12)] = KP1_414213562 * (T2O - T2L);
}
{
E T2t, T2x, T2w, T2y;
{
E T2r, T2s, T2u, T2v;
T2r = T1Z - T22;
T2s = KP707106781 * (T2m - T2l);
T2t = T2r + T2s;
T2x = T2r - T2s;
T2u = T2j + T2i;
T2v = KP707106781 * (T28 - T2d);
T2w = T2u - T2v;
T2y = T2v + T2u;
}
R0[WS(rs, 3)] = FMA(KP1_662939224, T2t, KP1_111140466 * T2w);
R0[WS(rs, 15)] = FNMS(KP1_961570560, T2x, KP390180644 * T2y);
R0[WS(rs, 11)] = FNMS(KP1_111140466, T2t, KP1_662939224 * T2w);
R0[WS(rs, 7)] = FMA(KP390180644, T2x, KP1_961570560 * T2y);
}
{
E T2D, T2J, T2I, T2K;
{
E T2z, T2C, T2G, T2H;
T2z = T7 - Te;
T2C = T2A - T2B;
T2D = T2z + T2C;
T2J = T2z - T2C;
T2G = T2E + T2F;
T2H = Tm - Tt;
T2I = T2G - T2H;
T2K = T2H + T2G;
}
R0[WS(rs, 2)] = FMA(KP1_847759065, T2D, KP765366864 * T2I);
R0[WS(rs, 14)] = FNMS(KP1_847759065, T2J, KP765366864 * T2K);
R0[WS(rs, 10)] = FNMS(KP765366864, T2D, KP1_847759065 * T2I);
R0[WS(rs, 6)] = FMA(KP765366864, T2J, KP1_847759065 * T2K);
}
{
E T19, T1n, T1m, T1o;
{
E TL, T18, T1c, T1l;
TL = Tz + TK;
T18 = TW + T17;
T19 = TL + T18;
T1n = TL - T18;
T1c = T1a + T1b;
T1l = T1f + T1k;
T1m = T1c + T1l;
T1o = T1c - T1l;
}
R1[0] = FNMS(KP196034280, T1m, KP1_990369453 * T19);
R1[WS(rs, 12)] = FNMS(KP1_546020906, T1n, KP1_268786568 * T1o);
R1[WS(rs, 8)] = -(FMA(KP196034280, T19, KP1_990369453 * T1m));
R1[WS(rs, 4)] = FMA(KP1_268786568, T1n, KP1_546020906 * T1o);
}
{
E T1r, T1v, T1u, T1w;
{
E T1p, T1q, T1s, T1t;
T1p = Tz - TK;
T1q = T1b - T1a;
T1r = T1p + T1q;
T1v = T1p - T1q;
T1s = T1f - T1k;
T1t = TW - T17;
T1u = T1s - T1t;
T1w = T1t + T1s;
}
R1[WS(rs, 2)] = FMA(KP1_763842528, T1r, KP942793473 * T1u);
R1[WS(rs, 14)] = FNMS(KP1_913880671, T1v, KP580569354 * T1w);
R1[WS(rs, 10)] = FNMS(KP942793473, T1r, KP1_763842528 * T1u);
R1[WS(rs, 6)] = FMA(KP580569354, T1v, KP1_913880671 * T1w);
}
{
E T1T, T1X, T1W, T1Y;
{
E T1R, T1S, T1U, T1V;
T1R = T1x + T1y;
T1S = T1L + T1M;
T1T = T1R - T1S;
T1X = T1R + T1S;
T1U = T1J + T1I;
T1V = T1C - T1F;
T1W = T1U - T1V;
T1Y = T1V + T1U;
}
R1[WS(rs, 3)] = FMA(KP1_546020906, T1T, KP1_268786568 * T1W);
R1[WS(rs, 15)] = FNMS(KP1_990369453, T1X, KP196034280 * T1Y);
R1[WS(rs, 11)] = FNMS(KP1_268786568, T1T, KP1_546020906 * T1W);
R1[WS(rs, 7)] = FMA(KP196034280, T1X, KP1_990369453 * T1Y);
}
{
E T2f, T2p, T2o, T2q;
{
E T23, T2e, T2k, T2n;
T23 = T1Z + T22;
T2e = KP707106781 * (T28 + T2d);
T2f = T23 + T2e;
T2p = T23 - T2e;
T2k = T2i - T2j;
T2n = KP707106781 * (T2l + T2m);
T2o = T2k - T2n;
T2q = T2n + T2k;
}
R0[WS(rs, 1)] = FMA(KP1_961570560, T2f, KP390180644 * T2o);
R0[WS(rs, 13)] = FNMS(KP1_662939224, T2p, KP1_111140466 * T2q);
R0[WS(rs, 9)] = FNMS(KP390180644, T2f, KP1_961570560 * T2o);
R0[WS(rs, 5)] = FMA(KP1_111140466, T2p, KP1_662939224 * T2q);
}
{
E T1H, T1P, T1O, T1Q;
{
E T1z, T1G, T1K, T1N;
T1z = T1x - T1y;
T1G = T1C + T1F;
T1H = T1z + T1G;
T1P = T1z - T1G;
T1K = T1I - T1J;
T1N = T1L - T1M;
T1O = T1K - T1N;
T1Q = T1N + T1K;
}
R1[WS(rs, 1)] = FMA(KP1_913880671, T1H, KP580569354 * T1O);
R1[WS(rs, 13)] = FNMS(KP1_763842528, T1P, KP942793473 * T1Q);
R1[WS(rs, 9)] = FNMS(KP580569354, T1H, KP1_913880671 * T1O);
R1[WS(rs, 5)] = FMA(KP942793473, T1P, KP1_763842528 * T1Q);
}
}
}
}
static void r2cb_20(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
{
DK(KP1_118033988, +1.118033988749894848204586834365638117720309180);
DK(KP500000000, +0.500000000000000000000000000000000000000000000);
DK(KP1_902113032, +1.902113032590307144232878666758764286811397268);
DK(KP1_175570504, +1.175570504584946258337411909278145537195304875);
DK(KP2_000000000, +2.000000000000000000000000000000000000000000000);
INT i;
for (i = v; i > 0; i = i - 1, R0 = R0 + ovs, R1 = R1 + ovs, Cr = Cr + ivs, Ci = Ci + ivs, MAKE_VOLATILE_STRIDE(rs), MAKE_VOLATILE_STRIDE(csr), MAKE_VOLATILE_STRIDE(csi)) {
E T6, TF, Tm, Tt, TQ, T1n, T1f, T12, T1m, TV, T13, T1c, Td, Tk, Tl;
E Ty, TD, TE, Tn, To, Tp, TG, TH, TI;
{
E T5, Ts, T3, Tq;
{
E T4, Tr, T1, T2;
T4 = Cr[WS(csr, 5)];
T5 = KP2_000000000 * T4;
Tr = Ci[WS(csi, 5)];
Ts = KP2_000000000 * Tr;
T1 = Cr[0];
T2 = Cr[WS(csr, 10)];
T3 = T1 + T2;
Tq = T1 - T2;
}
T6 = T3 - T5;
TF = Tq - Ts;
Tm = T3 + T5;
Tt = Tq + Ts;
}
{
E T9, Tu, TO, T1b, Tc, T1a, Tx, TP, Tg, Tz, TT, T1e, Tj, T1d, TC;
E TU;
{
E T7, T8, TM, TN;
T7 = Cr[WS(csr, 4)];
T8 = Cr[WS(csr, 6)];
T9 = T7 + T8;
Tu = T7 - T8;
TM = Ci[WS(csi, 4)];
TN = Ci[WS(csi, 6)];
TO = TM - TN;
T1b = TM + TN;
}
{
E Ta, Tb, Tv, Tw;
Ta = Cr[WS(csr, 9)];
Tb = Cr[WS(csr, 1)];
Tc = Ta + Tb;
T1a = Ta - Tb;
Tv = Ci[WS(csi, 9)];
Tw = Ci[WS(csi, 1)];
Tx = Tv + Tw;
TP = Tv - Tw;
}
{
E Te, Tf, TR, TS;
Te = Cr[WS(csr, 8)];
Tf = Cr[WS(csr, 2)];
Tg = Te + Tf;
Tz = Te - Tf;
TR = Ci[WS(csi, 8)];
TS = Ci[WS(csi, 2)];
TT = TR - TS;
T1e = TR + TS;
}
{
E Th, Ti, TA, TB;
Th = Cr[WS(csr, 7)];
Ti = Cr[WS(csr, 3)];
Tj = Th + Ti;
T1d = Th - Ti;
TA = Ci[WS(csi, 7)];
TB = Ci[WS(csi, 3)];
TC = TA + TB;
TU = TB - TA;
}
TQ = TO - TP;
T1n = T1e - T1d;
T1f = T1d + T1e;
T12 = TP + TO;
T1m = T1b - T1a;
TV = TT - TU;
T13 = TU + TT;
T1c = T1a + T1b;
Td = T9 - Tc;
Tk = Tg - Tj;
Tl = Td + Tk;
Ty = Tu + Tx;
TD = Tz - TC;
TE = Ty + TD;
Tn = T9 + Tc;
To = Tg + Tj;
Tp = Tn + To;
TG = Tu - Tx;
TH = Tz + TC;
TI = TG + TH;
}
R0[WS(rs, 5)] = FMA(KP2_000000000, Tl, T6);
R1[WS(rs, 7)] = FMA(KP2_000000000, TE, Tt);
R1[WS(rs, 2)] = FMA(KP2_000000000, TI, TF);
R0[0] = FMA(KP2_000000000, Tp, Tm);
{
E TW, TY, TL, TX, TJ, TK;
TW = FNMS(KP1_902113032, TV, KP1_175570504 * TQ);
TY = FMA(KP1_902113032, TQ, KP1_175570504 * TV);
TJ = FNMS(KP500000000, Tl, T6);
TK = KP1_118033988 * (Td - Tk);
TL = TJ - TK;
TX = TK + TJ;
R0[WS(rs, 1)] = TL - TW;
R0[WS(rs, 7)] = TX + TY;
R0[WS(rs, 9)] = TL + TW;
R0[WS(rs, 3)] = TX - TY;
}
{
E T1g, T1i, T19, T1h, T17, T18;
T1g = FNMS(KP1_902113032, T1f, KP1_175570504 * T1c);
T1i = FMA(KP1_902113032, T1c, KP1_175570504 * T1f);
T17 = FNMS(KP500000000, TI, TF);
T18 = KP1_118033988 * (TG - TH);
T19 = T17 - T18;
T1h = T18 + T17;
R1[WS(rs, 8)] = T19 - T1g;
R1[WS(rs, 4)] = T1h + T1i;
R1[WS(rs, 6)] = T19 + T1g;
R1[0] = T1h - T1i;
}
{
E T1o, T1q, T1l, T1p, T1j, T1k;
T1o = FNMS(KP1_902113032, T1n, KP1_175570504 * T1m);
T1q = FMA(KP1_902113032, T1m, KP1_175570504 * T1n);
T1j = FNMS(KP500000000, TE, Tt);
T1k = KP1_118033988 * (Ty - TD);
T1l = T1j - T1k;
T1p = T1k + T1j;
R1[WS(rs, 3)] = T1l - T1o;
R1[WS(rs, 9)] = T1p + T1q;
R1[WS(rs, 1)] = T1l + T1o;
R1[WS(rs, 5)] = T1p - T1q;
}
{
E T14, T16, T11, T15, TZ, T10;
T14 = FNMS(KP1_902113032, T13, KP1_175570504 * T12);
T16 = FMA(KP1_902113032, T12, KP1_175570504 * T13);
TZ = FNMS(KP500000000, Tp, Tm);
T10 = KP1_118033988 * (Tn - To);
T11 = TZ - T10;
T15 = T10 + TZ;
R0[WS(rs, 6)] = T11 - T14;
R0[WS(rs, 2)] = T15 + T16;
R0[WS(rs, 4)] = T11 + T14;
R0[WS(rs, 8)] = T15 - T16;
}
}
}
static void r2cb_20(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
{
DK(KP1_902113032, +1.902113032590307144232878666758764286811397268);
DK(KP1_118033988, +1.118033988749894848204586834365638117720309180);
DK(KP500000000, +0.500000000000000000000000000000000000000000000);
DK(KP618033988, +0.618033988749894848204586834365638117720309180);
DK(KP2_000000000, +2.000000000000000000000000000000000000000000000);
INT i;
for (i = v; i > 0; i = i - 1, R0 = R0 + ovs, R1 = R1 + ovs, Cr = Cr + ivs, Ci = Ci + ivs, MAKE_VOLATILE_STRIDE(rs), MAKE_VOLATILE_STRIDE(csr), MAKE_VOLATILE_STRIDE(csi)) {
E TY, T1o, T1m, T14, T12, TX, T1n, T1j, TZ, T13;
{
E Tr, TD, Tl, T5, T1a, T1l, T1d, T1k, TT, T10, TO, T11, TE, TF, Tk;
E TI, TC, T1i, To, TG, T16;
{
E T4, Tq, T1, T2;
T4 = Cr[WS(csr, 5)];
Tq = Ci[WS(csi, 5)];
T1 = Cr[0];
T2 = Cr[WS(csr, 10)];
{
E Ts, T8, T19, TR, T18, Tb, TS, Tv, Tx, Tf, Ty, T1c, TM, T1b, Ti;
E Tz, Tt, Tu, TN, TA;
{
E TP, TQ, T9, Ta;
{
E T6, T7, Tp, T3;
T6 = Cr[WS(csr, 4)];
T7 = Cr[WS(csr, 6)];
TP = Ci[WS(csi, 4)];
Tp = T1 - T2;
T3 = T1 + T2;
Ts = T6 - T7;
T8 = T6 + T7;
Tr = FMA(KP2_000000000, Tq, Tp);
TD = FNMS(KP2_000000000, Tq, Tp);
Tl = FMA(KP2_000000000, T4, T3);
T5 = FNMS(KP2_000000000, T4, T3);
TQ = Ci[WS(csi, 6)];
}
T9 = Cr[WS(csr, 9)];
Ta = Cr[WS(csr, 1)];
Tt = Ci[WS(csi, 9)];
T19 = TP + TQ;
TR = TP - TQ;
T18 = T9 - Ta;
Tb = T9 + Ta;
Tu = Ci[WS(csi, 1)];
}
{
E TK, TL, Td, Te, Tg, Th;
Td = Cr[WS(csr, 8)];
Te = Cr[WS(csr, 2)];
TK = Ci[WS(csi, 8)];
TS = Tt - Tu;
Tv = Tt + Tu;
Tx = Td - Te;
Tf = Td + Te;
TL = Ci[WS(csi, 2)];
Tg = Cr[WS(csr, 7)];
Th = Cr[WS(csr, 3)];
Ty = Ci[WS(csi, 7)];
T1c = TK + TL;
TM = TK - TL;
T1b = Tg - Th;
Ti = Tg + Th;
Tz = Ci[WS(csi, 3)];
}
T1a = T18 + T19;
T1l = T19 - T18;
T1d = T1b + T1c;
T1k = T1c - T1b;
TT = TR - TS;
T10 = TS + TR;
TN = Tz - Ty;
TA = Ty + Tz;
TO = TM - TN;
T11 = TN + TM;
{
E Tm, Tc, Tj, Tn, Tw, TB;
Tm = T8 + Tb;
Tc = T8 - Tb;
Tj = Tf - Ti;
Tn = Tf + Ti;
TE = Ts - Tv;
Tw = Ts + Tv;
TB = Tx - TA;
TF = Tx + TA;
Tk = Tc + Tj;
TI = Tc - Tj;
TC = Tw + TB;
T1i = Tw - TB;
TY = Tm - Tn;
To = Tm + Tn;
}
}
}
R0[WS(rs, 5)] = FMA(KP2_000000000, Tk, T5);
R1[WS(rs, 7)] = FMA(KP2_000000000, TC, Tr);
TG = TE + TF;
T16 = TE - TF;
R0[0] = FMA(KP2_000000000, To, Tl);
{
E TU, TW, T1g, T1e, T15, TV, TJ, TH, T1h, T1f, T17;
TU = FNMS(KP618033988, TT, TO);
TW = FMA(KP618033988, TO, TT);
R1[WS(rs, 2)] = FMA(KP2_000000000, TG, TD);
TH = FNMS(KP500000000, Tk, T5);
T1g = FNMS(KP618033988, T1a, T1d);
T1e = FMA(KP618033988, T1d, T1a);
T15 = FNMS(KP500000000, TG, TD);
TV = FMA(KP1_118033988, TI, TH);
TJ = FNMS(KP1_118033988, TI, TH);
T1o = FMA(KP618033988, T1k, T1l);
T1m = FNMS(KP618033988, T1l, T1k);
R0[WS(rs, 3)] = FNMS(KP1_902113032, TW, TV);
R0[WS(rs, 7)] = FMA(KP1_902113032, TW, TV);
R0[WS(rs, 1)] = FMA(KP1_902113032, TU, TJ);
R0[WS(rs, 9)] = FNMS(KP1_902113032, TU, TJ);
T1f = FNMS(KP1_118033988, T16, T15);
T17 = FMA(KP1_118033988, T16, T15);
T1h = FNMS(KP500000000, TC, Tr);
R1[WS(rs, 6)] = FNMS(KP1_902113032, T1g, T1f);
R1[WS(rs, 8)] = FMA(KP1_902113032, T1g, T1f);
R1[WS(rs, 4)] = FMA(KP1_902113032, T1e, T17);
R1[0] = FNMS(KP1_902113032, T1e, T17);
T14 = FNMS(KP618033988, T10, T11);
T12 = FMA(KP618033988, T11, T10);
TX = FNMS(KP500000000, To, Tl);
T1n = FMA(KP1_118033988, T1i, T1h);
T1j = FNMS(KP1_118033988, T1i, T1h);
}
}
R1[WS(rs, 5)] = FNMS(KP1_902113032, T1o, T1n);
R1[WS(rs, 9)] = FMA(KP1_902113032, T1o, T1n);
R1[WS(rs, 3)] = FMA(KP1_902113032, T1m, T1j);
R1[WS(rs, 1)] = FNMS(KP1_902113032, T1m, T1j);
TZ = FMA(KP1_118033988, TY, TX);
T13 = FNMS(KP1_118033988, TY, TX);
R0[WS(rs, 4)] = FNMS(KP1_902113032, T14, T13);
R0[WS(rs, 6)] = FMA(KP1_902113032, T14, T13);
R0[WS(rs, 2)] = FMA(KP1_902113032, T12, TZ);
R0[WS(rs, 8)] = FNMS(KP1_902113032, T12, TZ);
}
}
static void hc2cbdft_8(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms)
{
DK(KP707106781, +0.707106781186547524400844362104849039284835938);
INT m;
for (m = mb, W = W + ((mb - 1) * 14); m < me; m = m + 1, Rp = Rp + ms, Ip = Ip + ms, Rm = Rm - ms, Im = Im - ms, W = W + 14, MAKE_VOLATILE_STRIDE(rs)) {
E T7, T1d, T1h, Tl, TG, T14, T19, TO, Te, TL, T18, T15, TB, T1e, Tw;
E T1i;
{
E T3, TC, Tk, TM, T6, Th, TF, TN;
{
E T1, T2, Ti, Tj;
T1 = Rp[0];
T2 = Rm[WS(rs, 3)];
T3 = T1 + T2;
TC = T1 - T2;
Ti = Ip[0];
Tj = Im[WS(rs, 3)];
Tk = Ti + Tj;
TM = Ti - Tj;
}
{
E T4, T5, TD, TE;
T4 = Rp[WS(rs, 2)];
T5 = Rm[WS(rs, 1)];
T6 = T4 + T5;
Th = T4 - T5;
TD = Ip[WS(rs, 2)];
TE = Im[WS(rs, 1)];
TF = TD + TE;
TN = TD - TE;
}
T7 = T3 + T6;
T1d = Tk - Th;
T1h = TC + TF;
Tl = Th + Tk;
TG = TC - TF;
T14 = T3 - T6;
T19 = TM - TN;
TO = TM + TN;
}
{
E Ta, Tm, Tp, TJ, Td, Tr, Tu, TK;
{
E T8, T9, Tn, To;
T8 = Rp[WS(rs, 1)];
T9 = Rm[WS(rs, 2)];
Ta = T8 + T9;
Tm = T8 - T9;
Tn = Ip[WS(rs, 1)];
To = Im[WS(rs, 2)];
Tp = Tn + To;
TJ = Tn - To;
}
{
E Tb, Tc, Ts, Tt;
Tb = Rm[0];
Tc = Rp[WS(rs, 3)];
Td = Tb + Tc;
Tr = Tb - Tc;
Ts = Im[0];
Tt = Ip[WS(rs, 3)];
Tu = Ts + Tt;
TK = Tt - Ts;
}
Te = Ta + Td;
TL = TJ + TK;
T18 = Ta - Td;
T15 = TK - TJ;
{
E Tz, TA, Tq, Tv;
Tz = Tm - Tp;
TA = Tr - Tu;
TB = KP707106781 * (Tz + TA);
T1e = KP707106781 * (Tz - TA);
Tq = Tm + Tp;
Tv = Tr + Tu;
Tw = KP707106781 * (Tq - Tv);
T1i = KP707106781 * (Tq + Tv);
}
}
{
E Tf, TP, TI, TQ;
Tf = T7 + Te;
TP = TL + TO;
{
E Tx, TH, Tg, Ty;
Tx = Tl + Tw;
TH = TB + TG;
Tg = W[0];
Ty = W[1];
TI = FMA(Tg, Tx, Ty * TH);
TQ = FNMS(Ty, Tx, Tg * TH);
}
Rp[0] = Tf - TI;
Ip[0] = TP + TQ;
Rm[0] = Tf + TI;
Im[0] = TQ - TP;
}
{
E T1r, T1x, T1w, T1y;
{
E T1o, T1q, T1n, T1p;
T1o = T14 - T15;
T1q = T19 - T18;
T1n = W[10];
T1p = W[11];
T1r = FNMS(T1p, T1q, T1n * T1o);
T1x = FMA(T1p, T1o, T1n * T1q);
}
{
E T1t, T1v, T1s, T1u;
T1t = T1d - T1e;
T1v = T1i + T1h;
T1s = W[12];
T1u = W[13];
T1w = FMA(T1s, T1t, T1u * T1v);
T1y = FNMS(T1u, T1t, T1s * T1v);
}
Rp[WS(rs, 3)] = T1r - T1w;
Ip[WS(rs, 3)] = T1x + T1y;
Rm[WS(rs, 3)] = T1r + T1w;
Im[WS(rs, 3)] = T1y - T1x;
}
{
E TV, T11, T10, T12;
{
E TS, TU, TR, TT;
TS = T7 - Te;
TU = TO - TL;
TR = W[6];
TT = W[7];
TV = FNMS(TT, TU, TR * TS);
T11 = FMA(TT, TS, TR * TU);
}
{
E TX, TZ, TW, TY;
TX = Tl - Tw;
TZ = TG - TB;
TW = W[8];
TY = W[9];
T10 = FMA(TW, TX, TY * TZ);
T12 = FNMS(TY, TX, TW * TZ);
}
Rp[WS(rs, 2)] = TV - T10;
Ip[WS(rs, 2)] = T11 + T12;
Rm[WS(rs, 2)] = TV + T10;
Im[WS(rs, 2)] = T12 - T11;
}
{
E T1b, T1l, T1k, T1m;
{
E T16, T1a, T13, T17;
T16 = T14 + T15;
T1a = T18 + T19;
T13 = W[2];
T17 = W[3];
T1b = FNMS(T17, T1a, T13 * T16);
T1l = FMA(T17, T16, T13 * T1a);
}
{
E T1f, T1j, T1c, T1g;
T1f = T1d + T1e;
T1j = T1h - T1i;
T1c = W[4];
T1g = W[5];
T1k = FMA(T1c, T1f, T1g * T1j);
T1m = FNMS(T1g, T1f, T1c * T1j);
}
Rp[WS(rs, 1)] = T1b - T1k;
Ip[WS(rs, 1)] = T1l + T1m;
Rm[WS(rs, 1)] = T1b + T1k;
Im[WS(rs, 1)] = T1m - T1l;
}
}
}
static void hc2cbdft_8(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms)
{
DK(KP707106781, +0.707106781186547524400844362104849039284835938);
INT m;
for (m = mb, W = W + ((mb - 1) * 14); m < me; m = m + 1, Rp = Rp + ms, Ip = Ip + ms, Rm = Rm - ms, Im = Im - ms, W = W + 14, MAKE_VOLATILE_STRIDE(rs)) {
E T1m, T1r, T1i, T1u, T1o, T1v, T1n, T1w, T1s;
{
E T1k, Tl, T1p, TE, TP, T1g, TM, T1b, T1f, T1a, TU, Tf, T1l, TH, Tw;
E T1q;
{
E TA, T3, TN, Tk, Th, T6, TO, TD, Tb, Tm, Ta, TK, Tp, Tc, Ts;
E Tt;
{
E T4, T5, TB, TC;
{
E T1, T2, Ti, Tj;
T1 = Rp[0];
T2 = Rm[WS(rs, 3)];
Ti = Ip[0];
Tj = Im[WS(rs, 3)];
T4 = Rp[WS(rs, 2)];
TA = T1 - T2;
T3 = T1 + T2;
TN = Ti - Tj;
Tk = Ti + Tj;
T5 = Rm[WS(rs, 1)];
TB = Ip[WS(rs, 2)];
TC = Im[WS(rs, 1)];
}
{
E T8, T9, Tn, To;
T8 = Rp[WS(rs, 1)];
Th = T4 - T5;
T6 = T4 + T5;
TO = TB - TC;
TD = TB + TC;
T9 = Rm[WS(rs, 2)];
Tn = Ip[WS(rs, 1)];
To = Im[WS(rs, 2)];
Tb = Rm[0];
Tm = T8 - T9;
Ta = T8 + T9;
TK = Tn - To;
Tp = Tn + To;
Tc = Rp[WS(rs, 3)];
Ts = Im[0];
Tt = Ip[WS(rs, 3)];
}
}
{
E Tr, Td, Tu, TL, Te, T7;
T1k = Tk - Th;
Tl = Th + Tk;
Tr = Tb - Tc;
Td = Tb + Tc;
TL = Tt - Ts;
Tu = Ts + Tt;
T1p = TA + TD;
TE = TA - TD;
TP = TN + TO;
T1g = TN - TO;
TM = TK + TL;
T1b = TL - TK;
T1f = Ta - Td;
Te = Ta + Td;
T1a = T3 - T6;
T7 = T3 + T6;
{
E Tq, TF, TG, Tv;
Tq = Tm + Tp;
TF = Tm - Tp;
TG = Tr - Tu;
Tv = Tr + Tu;
TU = T7 - Te;
Tf = T7 + Te;
T1l = TF - TG;
TH = TF + TG;
Tw = Tq - Tv;
T1q = Tq + Tv;
}
}
}
{
E TX, T10, T1c, T13, T1h, T1E, T1H, T1C, T1K, T1G, T1L, T1F;
{
E TQ, Tx, T1y, TI, Tg, Tz;
TX = TP - TM;
TQ = TM + TP;
Tx = FMA(KP707106781, Tw, Tl);
T10 = FNMS(KP707106781, Tw, Tl);
T1c = T1a + T1b;
T1y = T1a - T1b;
T13 = FNMS(KP707106781, TH, TE);
TI = FMA(KP707106781, TH, TE);
Tg = W[0];
Tz = W[1];
{
E T1B, T1A, T1x, T1J, T1z, T1D;
{
E TR, Ty, TS, TJ;
T1B = T1g - T1f;
T1h = T1f + T1g;
T1A = W[11];
TR = Tg * TI;
Ty = Tg * Tx;
T1x = W[10];
T1J = T1A * T1y;
TS = FNMS(Tz, Tx, TR);
TJ = FMA(Tz, TI, Ty);
T1z = T1x * T1y;
T1m = FMA(KP707106781, T1l, T1k);
T1E = FNMS(KP707106781, T1l, T1k);
Im[0] = TS - TQ;
Ip[0] = TQ + TS;
Rm[0] = Tf + TJ;
Rp[0] = Tf - TJ;
T1H = FMA(KP707106781, T1q, T1p);
T1r = FNMS(KP707106781, T1q, T1p);
T1D = W[12];
}
T1C = FNMS(T1A, T1B, T1z);
T1K = FMA(T1x, T1B, T1J);
T1G = W[13];
T1L = T1D * T1H;
T1F = T1D * T1E;
}
}
{
E TY, T16, T12, T17, T11;
{
E TW, TT, T15, TV, TZ, T1M, T1I;
TW = W[7];
T1M = FNMS(T1G, T1E, T1L);
T1I = FMA(T1G, T1H, T1F);
TT = W[6];
T15 = TW * TU;
Im[WS(rs, 3)] = T1M - T1K;
Ip[WS(rs, 3)] = T1K + T1M;
Rm[WS(rs, 3)] = T1C + T1I;
Rp[WS(rs, 3)] = T1C - T1I;
TV = TT * TU;
TZ = W[8];
TY = FNMS(TW, TX, TV);
T16 = FMA(TT, TX, T15);
T12 = W[9];
T17 = TZ * T13;
T11 = TZ * T10;
}
{
E T1e, T19, T1t, T1d, T1j, T18, T14;
T1e = W[3];
T18 = FNMS(T12, T10, T17);
T14 = FMA(T12, T13, T11);
T19 = W[2];
T1t = T1e * T1c;
Im[WS(rs, 2)] = T18 - T16;
Ip[WS(rs, 2)] = T16 + T18;
Rm[WS(rs, 2)] = TY + T14;
Rp[WS(rs, 2)] = TY - T14;
T1d = T19 * T1c;
T1j = W[4];
T1i = FNMS(T1e, T1h, T1d);
T1u = FMA(T19, T1h, T1t);
T1o = W[5];
T1v = T1j * T1r;
T1n = T1j * T1m;
}
}
}
}
T1w = FNMS(T1o, T1m, T1v);
T1s = FMA(T1o, T1r, T1n);
Im[WS(rs, 1)] = T1w - T1u;
Ip[WS(rs, 1)] = T1u + T1w;
Rm[WS(rs, 1)] = T1i + T1s;
Rp[WS(rs, 1)] = T1i - T1s;
}
}
static void t2_8(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DK(KP707106781, +0.707106781186547524400844362104849039284835938);
{
INT m;
for (m = mb, W = W + (mb * 6); m < me; m = m + 1, ri = ri + ms, ii = ii + ms, W = W + 6, MAKE_VOLATILE_STRIDE(16, rs)) {
E TS, T1m, TJ, T1l, T1k, Tw, T1w, T1u;
{
E T2, T3, Tl, Tn, T5, T4, Tm, Tr, T6;
T2 = W[0];
T3 = W[2];
Tl = W[4];
Tn = W[5];
T5 = W[1];
T4 = T2 * T3;
Tm = T2 * Tl;
Tr = T2 * Tn;
T6 = W[3];
{
E T1, T1s, TG, Td, T1r, Tu, TY, Tk, TW, T18, T1d, TD, TH, TA, T13;
E TE, T14;
{
E To, Ts, Tf, T7, T8, Ti, Tb, T9, Tc, TC, Ta, TF, TB, Tg, Th;
E Tj;
T1 = ri[0];
To = FMA(T5, Tn, Tm);
Ts = FNMS(T5, Tl, Tr);
Tf = FMA(T5, T6, T4);
T7 = FNMS(T5, T6, T4);
Ta = T2 * T6;
T1s = ii[0];
T8 = ri[WS(rs, 4)];
TF = Tf * Tn;
TB = Tf * Tl;
Ti = FNMS(T5, T3, Ta);
Tb = FMA(T5, T3, Ta);
T9 = T7 * T8;
Tc = ii[WS(rs, 4)];
TG = FNMS(Ti, Tl, TF);
TC = FMA(Ti, Tn, TB);
{
E Tp, T1q, Tt, Tq, TX;
Tp = ri[WS(rs, 6)];
Td = FMA(Tb, Tc, T9);
T1q = T7 * Tc;
Tt = ii[WS(rs, 6)];
Tq = To * Tp;
Tg = ri[WS(rs, 2)];
T1r = FNMS(Tb, T8, T1q);
TX = To * Tt;
Tu = FMA(Ts, Tt, Tq);
Th = Tf * Tg;
Tj = ii[WS(rs, 2)];
TY = FNMS(Ts, Tp, TX);
}
{
E TO, TQ, TN, TP, T1a, T1b;
{
E TK, TM, TL, T19, TV;
TK = ri[WS(rs, 7)];
TM = ii[WS(rs, 7)];
Tk = FMA(Ti, Tj, Th);
TV = Tf * Tj;
TL = Tl * TK;
T19 = Tl * TM;
TO = ri[WS(rs, 3)];
TW = FNMS(Ti, Tg, TV);
TQ = ii[WS(rs, 3)];
TN = FMA(Tn, TM, TL);
TP = T3 * TO;
T1a = FNMS(Tn, TK, T19);
T1b = T3 * TQ;
}
{
E Tx, Tz, Ty, T12, T1c, TR;
Tx = ri[WS(rs, 1)];
TR = FMA(T6, TQ, TP);
Tz = ii[WS(rs, 1)];
T1c = FNMS(T6, TO, T1b);
Ty = T2 * Tx;
T18 = TN - TR;
TS = TN + TR;
T12 = T2 * Tz;
T1d = T1a - T1c;
T1m = T1a + T1c;
TD = ri[WS(rs, 5)];
TH = ii[WS(rs, 5)];
TA = FMA(T5, Tz, Ty);
T13 = FNMS(T5, Tx, T12);
TE = TC * TD;
T14 = TC * TH;
}
}
}
{
E Te, T1p, T1t, Tv;
{
E T1g, T10, T1z, T1B, T1A, T1j, T1C, T1f;
{
E T1x, T11, T16, T1y;
{
E TU, TZ, TI, T15;
Te = T1 + Td;
TU = T1 - Td;
TZ = TW - TY;
T1p = TW + TY;
TI = FMA(TG, TH, TE);
T15 = FNMS(TG, TD, T14);
T1t = T1r + T1s;
T1x = T1s - T1r;
T1g = TU - TZ;
T10 = TU + TZ;
T11 = TA - TI;
TJ = TA + TI;
T1l = T13 + T15;
T16 = T13 - T15;
T1y = Tk - Tu;
Tv = Tk + Tu;
}
{
E T1i, T1e, T17, T1h;
T1i = T18 + T1d;
T1e = T18 - T1d;
T17 = T11 + T16;
T1h = T16 - T11;
T1z = T1x - T1y;
T1B = T1y + T1x;
T1A = T1h + T1i;
T1j = T1h - T1i;
T1C = T1e - T17;
T1f = T17 + T1e;
}
}
ri[WS(rs, 7)] = FNMS(KP707106781, T1j, T1g);
ii[WS(rs, 7)] = FNMS(KP707106781, T1C, T1B);
ri[WS(rs, 1)] = FMA(KP707106781, T1f, T10);
ri[WS(rs, 5)] = FNMS(KP707106781, T1f, T10);
ii[WS(rs, 1)] = FMA(KP707106781, T1A, T1z);
ii[WS(rs, 5)] = FNMS(KP707106781, T1A, T1z);
ri[WS(rs, 3)] = FMA(KP707106781, T1j, T1g);
ii[WS(rs, 3)] = FMA(KP707106781, T1C, T1B);
}
T1k = Te - Tv;
Tw = Te + Tv;
T1w = T1t - T1p;
T1u = T1p + T1t;
}
}
}
{
E TT, T1v, T1n, T1o;
TT = TJ + TS;
T1v = TS - TJ;
T1n = T1l - T1m;
T1o = T1l + T1m;
ii[WS(rs, 2)] = T1v + T1w;
ii[WS(rs, 6)] = T1w - T1v;
ri[0] = Tw + TT;
ri[WS(rs, 4)] = Tw - TT;
ii[0] = T1o + T1u;
ii[WS(rs, 4)] = T1u - T1o;
ri[WS(rs, 2)] = T1k + T1n;
ri[WS(rs, 6)] = T1k - T1n;
}
}
}
}
static void t2_8(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DK(KP707106781, +0.707106781186547524400844362104849039284835938);
{
INT m;
for (m = mb, W = W + (mb * 6); m < me; m = m + 1, ri = ri + ms, ii = ii + ms, W = W + 6, MAKE_VOLATILE_STRIDE(16, rs)) {
E T2, T5, T3, T6, T8, Tc, Tg, Ti, Tl, Tm, Tn, Tz, Tp, Tx;
{
E T4, Tb, T7, Ta;
T2 = W[0];
T5 = W[1];
T3 = W[2];
T6 = W[3];
T4 = T2 * T3;
Tb = T5 * T3;
T7 = T5 * T6;
Ta = T2 * T6;
T8 = T4 - T7;
Tc = Ta + Tb;
Tg = T4 + T7;
Ti = Ta - Tb;
Tl = W[4];
Tm = W[5];
Tn = FMA(T2, Tl, T5 * Tm);
Tz = FNMS(Ti, Tl, Tg * Tm);
Tp = FNMS(T5, Tl, T2 * Tm);
Tx = FMA(Tg, Tl, Ti * Tm);
}
{
E Tf, T1i, TL, T1d, TJ, T17, TV, TY, Ts, T1j, TO, T1a, TC, T16, TQ;
E TT;
{
E T1, T1c, Te, T1b, T9, Td;
T1 = ri[0];
T1c = ii[0];
T9 = ri[WS(rs, 4)];
Td = ii[WS(rs, 4)];
Te = FMA(T8, T9, Tc * Td);
T1b = FNMS(Tc, T9, T8 * Td);
Tf = T1 + Te;
T1i = T1c - T1b;
TL = T1 - Te;
T1d = T1b + T1c;
}
{
E TF, TW, TI, TX;
{
E TD, TE, TG, TH;
TD = ri[WS(rs, 7)];
TE = ii[WS(rs, 7)];
TF = FMA(Tl, TD, Tm * TE);
TW = FNMS(Tm, TD, Tl * TE);
TG = ri[WS(rs, 3)];
TH = ii[WS(rs, 3)];
TI = FMA(T3, TG, T6 * TH);
TX = FNMS(T6, TG, T3 * TH);
}
TJ = TF + TI;
T17 = TW + TX;
TV = TF - TI;
TY = TW - TX;
}
{
E Tk, TM, Tr, TN;
{
E Th, Tj, To, Tq;
Th = ri[WS(rs, 2)];
Tj = ii[WS(rs, 2)];
Tk = FMA(Tg, Th, Ti * Tj);
TM = FNMS(Ti, Th, Tg * Tj);
To = ri[WS(rs, 6)];
Tq = ii[WS(rs, 6)];
Tr = FMA(Tn, To, Tp * Tq);
TN = FNMS(Tp, To, Tn * Tq);
}
Ts = Tk + Tr;
T1j = Tk - Tr;
TO = TM - TN;
T1a = TM + TN;
}
{
E Tw, TR, TB, TS;
{
E Tu, Tv, Ty, TA;
Tu = ri[WS(rs, 1)];
Tv = ii[WS(rs, 1)];
Tw = FMA(T2, Tu, T5 * Tv);
TR = FNMS(T5, Tu, T2 * Tv);
Ty = ri[WS(rs, 5)];
TA = ii[WS(rs, 5)];
TB = FMA(Tx, Ty, Tz * TA);
TS = FNMS(Tz, Ty, Tx * TA);
}
TC = Tw + TB;
T16 = TR + TS;
TQ = Tw - TB;
TT = TR - TS;
}
{
E Tt, TK, T1f, T1g;
Tt = Tf + Ts;
TK = TC + TJ;
ri[WS(rs, 4)] = Tt - TK;
ri[0] = Tt + TK;
{
E T19, T1e, T15, T18;
T19 = T16 + T17;
T1e = T1a + T1d;
ii[0] = T19 + T1e;
ii[WS(rs, 4)] = T1e - T19;
T15 = Tf - Ts;
T18 = T16 - T17;
ri[WS(rs, 6)] = T15 - T18;
ri[WS(rs, 2)] = T15 + T18;
}
T1f = TJ - TC;
T1g = T1d - T1a;
ii[WS(rs, 2)] = T1f + T1g;
ii[WS(rs, 6)] = T1g - T1f;
{
E T11, T1k, T14, T1h, T12, T13;
T11 = TL - TO;
T1k = T1i - T1j;
T12 = TT - TQ;
T13 = TV + TY;
T14 = KP707106781 * (T12 - T13);
T1h = KP707106781 * (T12 + T13);
ri[WS(rs, 7)] = T11 - T14;
ii[WS(rs, 5)] = T1k - T1h;
ri[WS(rs, 3)] = T11 + T14;
ii[WS(rs, 1)] = T1h + T1k;
}
{
E TP, T1m, T10, T1l, TU, TZ;
TP = TL + TO;
T1m = T1j + T1i;
TU = TQ + TT;
TZ = TV - TY;
T10 = KP707106781 * (TU + TZ);
T1l = KP707106781 * (TZ - TU);
ri[WS(rs, 5)] = TP - T10;
ii[WS(rs, 7)] = T1m - T1l;
ri[WS(rs, 1)] = TP + T10;
ii[WS(rs, 3)] = T1l + T1m;
}
}
}
}
}
}
static void r2cfII_20(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
{
DK(KP707106781, +0.707106781186547524400844362104849039284835938);
DK(KP951056516, +0.951056516295153572116439333379382143405698634);
DK(KP559016994, +0.559016994374947424102293417182819058860154590);
DK(KP690983005, +0.690983005625052575897706582817180941139845410);
DK(KP552786404, +0.552786404500042060718165266253744752911876328);
DK(KP447213595, +0.447213595499957939281834733746255247088123672);
DK(KP809016994, +0.809016994374947424102293417182819058860154590);
DK(KP250000000, +0.250000000000000000000000000000000000000000000);
DK(KP618033988, +0.618033988749894848204586834365638117720309180);
DK(KP381966011, +0.381966011250105151795413165634361882279690820);
{
INT i;
for (i = v; i > 0; i = i - 1, R0 = R0 + ivs, R1 = R1 + ivs, Cr = Cr + ovs, Ci = Ci + ovs, MAKE_VOLATILE_STRIDE(rs), MAKE_VOLATILE_STRIDE(csr), MAKE_VOLATILE_STRIDE(csi)) {
E Tv, TK, TN, Th, T1l, T1n, Ts, TH;
{
E Ti, T1d, T1f, T1e, T1g, T1p, TS, Tg, To, T8, T7, T19, T1r, T1k, Tx;
E Tp, TX, Ty, TF, Tr, TV, Tz, TA, TI;
{
E Ta, Tb, Td, Te;
Ti = R1[WS(rs, 2)];
T1d = R0[WS(rs, 5)];
Ta = R0[WS(rs, 9)];
Tb = R0[WS(rs, 1)];
Td = R0[WS(rs, 3)];
Te = R0[WS(rs, 7)];
{
E T1, T2, T5, T3, T4, T1i, Tc, Tf;
T1 = R0[0];
T1f = Ta + Tb;
Tc = Ta - Tb;
T1e = Td + Te;
Tf = Td - Te;
T2 = R0[WS(rs, 4)];
T5 = R0[WS(rs, 6)];
T1g = FMA(KP381966011, T1f, T1e);
T1p = FMA(KP381966011, T1e, T1f);
TS = FMA(KP618033988, Tc, Tf);
Tg = FNMS(KP618033988, Tf, Tc);
T3 = R0[WS(rs, 8)];
T4 = R0[WS(rs, 2)];
T1i = T2 + T5;
{
E Tj, Tu, Tm, Tt, Tn, Tq, TU;
Tj = R1[WS(rs, 8)];
To = R1[WS(rs, 6)];
{
E T6, T1j, Tk, Tl;
T6 = T2 + T3 - T4 - T5;
T8 = (T3 + T5 - T2) - T4;
T1j = T3 + T4;
Tk = R1[0];
Tl = R1[WS(rs, 4)];
T7 = FNMS(KP250000000, T6, T1);
T19 = T1 + T6;
T1r = FNMS(KP618033988, T1i, T1j);
T1k = FMA(KP618033988, T1j, T1i);
Tu = Tk - Tl;
Tm = Tk + Tl;
}
Tt = To + Tj;
Tx = R1[WS(rs, 7)];
Tn = Tj - Tm;
Tp = Tj + Tm;
Tv = FNMS(KP618033988, Tu, Tt);
TX = FMA(KP618033988, Tt, Tu);
Tq = FMA(KP809016994, Tp, To);
TU = FMA(KP447213595, Tp, Tn);
Ty = R1[WS(rs, 1)];
TF = R1[WS(rs, 3)];
Tr = FNMS(KP552786404, Tq, Tn);
TV = FNMS(KP690983005, TU, To);
Tz = R1[WS(rs, 5)];
TA = R1[WS(rs, 9)];
TI = TF + Ty;
}
}
}
{
E T1w, TJ, TB, T1a;
T1w = T1f + T1d - T1e;
TJ = Tz - TA;
TB = Tz + TA;
T1a = Ti + To - Tp;
{
E T9, T12, TT, T15, TG, TD, T1s, T1u, TW, T11, T10, T1h;
{
E TE, TC, TR, T1b;
T9 = FNMS(KP559016994, T8, T7);
TR = FMA(KP559016994, T8, T7);
TK = FMA(KP618033988, TJ, TI);
T12 = FNMS(KP618033988, TI, TJ);
TE = Ty - TB;
TC = Ty + TB;
TT = FMA(KP951056516, TS, TR);
T15 = FNMS(KP951056516, TS, TR);
TG = FNMS(KP552786404, TF, TE);
T1b = TC - TF - Tx;
{
E TZ, T1q, T1c, T1x;
TZ = FMA(KP447213595, TC, TE);
TD = FMA(KP250000000, TC, Tx);
T1q = FNMS(KP809016994, T1p, T1d);
T1c = T1a + T1b;
T1x = T1a - T1b;
T10 = FNMS(KP690983005, TZ, TF);
T1s = FNMS(KP951056516, T1r, T1q);
T1u = FMA(KP951056516, T1r, T1q);
Ci[WS(csi, 7)] = FMA(KP707106781, T1x, T1w);
Ci[WS(csi, 2)] = FMS(KP707106781, T1x, T1w);
Cr[WS(csr, 7)] = FMA(KP707106781, T1c, T19);
Cr[WS(csr, 2)] = FNMS(KP707106781, T1c, T19);
}
}
TW = FNMS(KP809016994, TV, Ti);
T11 = FNMS(KP809016994, T10, Tx);
T1h = FMA(KP809016994, T1g, T1d);
{
E T17, TY, T16, T13;
T17 = FNMS(KP951056516, TX, TW);
TY = FMA(KP951056516, TX, TW);
T16 = FMA(KP951056516, T12, T11);
T13 = FNMS(KP951056516, T12, T11);
TN = FMA(KP951056516, Tg, T9);
Th = FNMS(KP951056516, Tg, T9);
{
E T18, T1v, T1t, T14;
T18 = T16 - T17;
T1v = T17 + T16;
T1t = TY + T13;
T14 = TY - T13;
Cr[WS(csr, 1)] = FMA(KP707106781, T18, T15);
Cr[WS(csr, 8)] = FNMS(KP707106781, T18, T15);
Ci[WS(csi, 3)] = FMA(KP707106781, T1v, T1u);
Ci[WS(csi, 6)] = FMS(KP707106781, T1v, T1u);
Ci[WS(csi, 1)] = FNMS(KP707106781, T1t, T1s);
Ci[WS(csi, 8)] = -(FMA(KP707106781, T1t, T1s));
Cr[WS(csr, 3)] = FMA(KP707106781, T14, TT);
Cr[WS(csr, 6)] = FNMS(KP707106781, T14, TT);
T1l = FMA(KP951056516, T1k, T1h);
T1n = FNMS(KP951056516, T1k, T1h);
}
}
Ts = FNMS(KP559016994, Tr, Ti);
TH = FNMS(KP559016994, TG, TD);
}
}
}
{
E TO, Tw, TP, TL;
TO = FMA(KP951056516, Tv, Ts);
Tw = FNMS(KP951056516, Tv, Ts);
TP = FMA(KP951056516, TK, TH);
TL = FNMS(KP951056516, TK, TH);
{
E TQ, T1m, T1o, TM;
TQ = TO - TP;
T1m = TO + TP;
T1o = Tw + TL;
TM = Tw - TL;
Cr[WS(csr, 4)] = FMA(KP707106781, TQ, TN);
Cr[WS(csr, 5)] = FNMS(KP707106781, TQ, TN);
Ci[WS(csi, 9)] = FNMS(KP707106781, T1m, T1l);
Ci[0] = -(FMA(KP707106781, T1m, T1l));
Ci[WS(csi, 5)] = FNMS(KP707106781, T1o, T1n);
Ci[WS(csi, 4)] = -(FMA(KP707106781, T1o, T1n));
Cr[0] = FMA(KP707106781, TM, Th);
Cr[WS(csr, 9)] = FNMS(KP707106781, TM, Th);
}
}
}
}
}
static void t1_8(float *ri, float *ii, const float *W, stride rs, INT mb, INT me, INT ms)
{
DK(KP707106781, +0.707106781186547524400844362104849039284835938);
INT m;
for (m = mb, W = W + (mb * 14); m < me; m = m + 1, ri = ri + ms, ii = ii + ms, W = W + 14, MAKE_VOLATILE_STRIDE(rs)) {
E T1g, T1f, T1e, Tm, T1q, T1o, T1p, TN, T1h, T1i;
{
E T1, T1m, T1l, T7, TS, Tk, TQ, Te, To, Tr, T17, TM, T12, Tu, TW;
E Tp, Tx, Tt, Tq, Tw;
{
E T3, T6, T2, T5;
T1 = ri[0];
T1m = ii[0];
T3 = ri[WS(rs, 4)];
T6 = ii[WS(rs, 4)];
T2 = W[6];
T5 = W[7];
{
E Ta, Td, T9, Tc;
{
E Tg, Tj, Ti, TR, Th, T1k, T4, Tf;
Tg = ri[WS(rs, 6)];
Tj = ii[WS(rs, 6)];
T1k = T2 * T6;
T4 = T2 * T3;
Tf = W[10];
Ti = W[11];
T1l = FNMS(T5, T3, T1k);
T7 = FMA(T5, T6, T4);
TR = Tf * Tj;
Th = Tf * Tg;
Ta = ri[WS(rs, 2)];
Td = ii[WS(rs, 2)];
TS = FNMS(Ti, Tg, TR);
Tk = FMA(Ti, Tj, Th);
T9 = W[2];
Tc = W[3];
}
{
E TB, TE, TH, T13, TC, TK, TG, TD, TJ, TP, Tb, TA, Tn;
TB = ri[WS(rs, 7)];
TE = ii[WS(rs, 7)];
TP = T9 * Td;
Tb = T9 * Ta;
TA = W[12];
TH = ri[WS(rs, 3)];
TQ = FNMS(Tc, Ta, TP);
Te = FMA(Tc, Td, Tb);
T13 = TA * TE;
TC = TA * TB;
TK = ii[WS(rs, 3)];
TG = W[4];
TD = W[13];
TJ = W[5];
{
E T14, TF, T16, TL, T15, TI;
To = ri[WS(rs, 1)];
T15 = TG * TK;
TI = TG * TH;
T14 = FNMS(TD, TB, T13);
TF = FMA(TD, TE, TC);
T16 = FNMS(TJ, TH, T15);
TL = FMA(TJ, TK, TI);
Tr = ii[WS(rs, 1)];
Tn = W[0];
T17 = T14 - T16;
T1g = T14 + T16;
TM = TF + TL;
T12 = TF - TL;
}
Tu = ri[WS(rs, 5)];
TW = Tn * Tr;
Tp = Tn * To;
Tx = ii[WS(rs, 5)];
Tt = W[8];
Tq = W[1];
Tw = W[9];
}
}
}
{
E T8, T1j, T1n, Tz, T1a, TU, Tl, T1b, T1c, T1v, T1t, T1w, T19, T1u, T1d;
{
E T1r, T10, TV, T1s, T11, T18;
{
E TO, TX, Ts, TZ, Ty, TT, TY, Tv;
T8 = T1 + T7;
TO = T1 - T7;
TY = Tt * Tx;
Tv = Tt * Tu;
TX = FNMS(Tq, To, TW);
Ts = FMA(Tq, Tr, Tp);
TZ = FNMS(Tw, Tu, TY);
Ty = FMA(Tw, Tx, Tv);
TT = TQ - TS;
T1j = TQ + TS;
T1n = T1l + T1m;
T1r = T1m - T1l;
T10 = TX - TZ;
T1f = TX + TZ;
Tz = Ts + Ty;
TV = Ts - Ty;
T1a = TO - TT;
TU = TO + TT;
T1s = Te - Tk;
Tl = Te + Tk;
}
T1b = T10 - TV;
T11 = TV + T10;
T18 = T12 - T17;
T1c = T12 + T17;
T1v = T1s + T1r;
T1t = T1r - T1s;
T1w = T18 - T11;
T19 = T11 + T18;
}
ii[WS(rs, 3)] = FMA(KP707106781, T1w, T1v);
ii[WS(rs, 7)] = FNMS(KP707106781, T1w, T1v);
ri[WS(rs, 1)] = FMA(KP707106781, T19, TU);
ri[WS(rs, 5)] = FNMS(KP707106781, T19, TU);
T1u = T1b + T1c;
T1d = T1b - T1c;
ii[WS(rs, 1)] = FMA(KP707106781, T1u, T1t);
ii[WS(rs, 5)] = FNMS(KP707106781, T1u, T1t);
ri[WS(rs, 3)] = FMA(KP707106781, T1d, T1a);
ri[WS(rs, 7)] = FNMS(KP707106781, T1d, T1a);
T1e = T8 - Tl;
Tm = T8 + Tl;
T1q = T1n - T1j;
T1o = T1j + T1n;
T1p = TM - Tz;
TN = Tz + TM;
}
}
ii[WS(rs, 2)] = T1p + T1q;
ii[WS(rs, 6)] = T1q - T1p;
ri[0] = Tm + TN;
ri[WS(rs, 4)] = Tm - TN;
T1h = T1f - T1g;
T1i = T1f + T1g;
ii[0] = T1i + T1o;
ii[WS(rs, 4)] = T1o - T1i;
ri[WS(rs, 2)] = T1e + T1h;
ri[WS(rs, 6)] = T1e - T1h;
}
}
static void hc2cfdft2_16(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms)
{
DK(KP923879532, +0.923879532511286756128183189396788286822416626);
DK(KP414213562, +0.414213562373095048801688724209698078569671875);
DK(KP707106781, +0.707106781186547524400844362104849039284835938);
DK(KP500000000, +0.500000000000000000000000000000000000000000000);
{
INT m;
for (m = mb, W = W + ((mb - 1) * 8); m < me; m = m + 1, Rp = Rp + ms, Ip = Ip + ms, Rm = Rm - ms, Im = Im - ms, W = W + 8, MAKE_VOLATILE_STRIDE(64, rs)) {
E T4p, T4o, T4n, T4s;
{
E T1, T2, Tw, Ty, Th, T3, Tx, TE, Ti, TK, Tj, T4, T5;
T1 = W[0];
T2 = W[2];
Tw = W[6];
Ty = W[7];
Th = W[4];
T3 = T1 * T2;
Tx = T1 * Tw;
TE = T1 * Ty;
Ti = T1 * Th;
TK = T2 * Th;
Tj = W[5];
T4 = W[1];
T5 = W[3];
{
E T1v, T2q, T1s, T2s, T38, T3T, T1Y, T3P, T17, T1h, T2x, T2v, T33, T3Q, T3S;
E T1N, Tv, T3A, T2E, T3B, T3L, T2c, T3I, T2S, TW, T3E, T3J, T2n, T3D, T2J;
E T3M, T2X;
{
E TF, Tk, Tz, TL, T6, TR, Tq, Tc, T2h, T25, T2k, T29, T1G, T1M, T2P;
E T2R;
{
E T18, TY, T1d, T13, T1H, T1A, T1K, T1E, T37, T1R, T35, T1X;
{
E T1j, T1o, T1W, T1p, T1m, T1Q, T1U, T1q;
{
E T1k, T1l, T1S, T1T;
{
E T1t, T28, T24, T1D, T1z, T1u, TQ, Tp, Tb;
T1t = Ip[0];
TQ = T2 * Tj;
Tp = T1 * Tj;
TF = FNMS(T4, Tw, TE);
T1j = FMA(T4, Tj, Ti);
Tk = FNMS(T4, Tj, Ti);
Tz = FMA(T4, Ty, Tx);
T18 = FNMS(T5, Tj, TK);
TL = FMA(T5, Tj, TK);
TY = FNMS(T4, T5, T3);
T6 = FMA(T4, T5, T3);
Tb = T1 * T5;
TR = FNMS(T5, Th, TQ);
T1d = FMA(T5, Th, TQ);
Tq = FMA(T4, Th, Tp);
T1o = FNMS(T4, Th, Tp);
T28 = T6 * Tj;
T24 = T6 * Th;
T1D = TY * Tj;
T1z = TY * Th;
Tc = FNMS(T4, T2, Tb);
T13 = FMA(T4, T2, Tb);
T1u = Im[0];
T1k = Ip[WS(rs, 4)];
T2h = FMA(Tc, Tj, T24);
T25 = FNMS(Tc, Tj, T24);
T2k = FNMS(Tc, Th, T28);
T29 = FMA(Tc, Th, T28);
T1H = FNMS(T13, Tj, T1z);
T1A = FMA(T13, Tj, T1z);
T1K = FMA(T13, Th, T1D);
T1E = FNMS(T13, Th, T1D);
T1W = T1t + T1u;
T1v = T1t - T1u;
T1l = Im[WS(rs, 4)];
}
T1S = Rm[0];
T1T = Rp[0];
T1p = Rp[WS(rs, 4)];
T1m = T1k - T1l;
T1Q = T1k + T1l;
T2q = T1T + T1S;
T1U = T1S - T1T;
T1q = Rm[WS(rs, 4)];
}
{
E T36, T1V, T1O, T1r, T1n, T1P, T34, T2r;
T36 = T4 * T1U;
T1V = T1 * T1U;
T1O = T1q - T1p;
T1r = T1p + T1q;
T1n = T1j * T1m;
T37 = FMA(T1, T1W, T36);
T2r = T1j * T1r;
T1P = Th * T1O;
T34 = Tj * T1O;
T1s = FNMS(T1o, T1r, T1n);
T2s = FMA(T1o, T1m, T2r);
T1R = FNMS(Tj, T1Q, T1P);
T35 = FMA(Th, T1Q, T34);
T1X = FNMS(T4, T1W, T1V);
}
}
{
E T1F, T11, T1e, T16, T1L, T1b, T1f, T1C, T2Z;
{
E T14, T15, TZ, T10, T19, T1a, T1B;
TZ = Ip[WS(rs, 2)];
T10 = Im[WS(rs, 2)];
T38 = T35 + T37;
T3T = T37 - T35;
T1Y = T1R + T1X;
T3P = T1X - T1R;
T1F = TZ + T10;
T11 = TZ - T10;
T14 = Rp[WS(rs, 2)];
T15 = Rm[WS(rs, 2)];
T19 = Ip[WS(rs, 6)];
T1a = Im[WS(rs, 6)];
T1e = Rp[WS(rs, 6)];
T16 = T14 + T15;
T1B = T15 - T14;
T1L = T19 + T1a;
T1b = T19 - T1a;
T1f = Rm[WS(rs, 6)];
T1C = T1A * T1B;
T2Z = T1E * T1B;
}
{
E T1J, T31, T2u, T30, T32;
{
E T12, T1g, T1I, T1c, T2w;
T12 = TY * T11;
T1g = T1e + T1f;
T1I = T1f - T1e;
T1c = T18 * T1b;
T17 = FNMS(T13, T16, T12);
T2w = T18 * T1g;
T1J = T1H * T1I;
T31 = T1K * T1I;
T1h = FNMS(T1d, T1g, T1c);
T2x = FMA(T1d, T1b, T2w);
}
T2u = TY * T16;
T30 = FMA(T1A, T1F, T2Z);
T32 = FMA(T1H, T1L, T31);
T1G = FNMS(T1E, T1F, T1C);
T2v = FMA(T13, T11, T2u);
T1M = FNMS(T1K, T1L, T1J);
T33 = T30 + T32;
T3Q = T30 - T32;
}
}
}
{
E Tl, T22, T9, T20, Tf, T2O, Ta, T21, T2A, Tm, Tr, Ts;
{
E T7, T8, Td, Te;
T7 = Ip[WS(rs, 1)];
T3S = T1G - T1M;
T1N = T1G + T1M;
T8 = Im[WS(rs, 1)];
Td = Rp[WS(rs, 1)];
Te = Rm[WS(rs, 1)];
Tl = Ip[WS(rs, 5)];
T22 = T7 + T8;
T9 = T7 - T8;
T20 = Td - Te;
Tf = Td + Te;
T2O = T2 * T22;
Ta = T6 * T9;
T21 = T2 * T20;
T2A = T6 * Tf;
Tm = Im[WS(rs, 5)];
Tr = Rp[WS(rs, 5)];
Ts = Rm[WS(rs, 5)];
}
{
E Tg, T2a, Tn, T26, T2Q, T27, T2C, T2B, Tu, Tt, To, T23, T2D, T2b;
Tg = FNMS(Tc, Tf, Ta);
T2a = Tl + Tm;
Tn = Tl - Tm;
T26 = Tr - Ts;
Tt = Tr + Ts;
T2Q = T25 * T2a;
To = Tk * Tn;
T27 = T25 * T26;
T2C = Tk * Tt;
T2B = FMA(Tc, T9, T2A);
Tu = FNMS(Tq, Tt, To);
T23 = FMA(T5, T22, T21);
T2D = FMA(Tq, Tn, T2C);
T2b = FMA(T29, T2a, T27);
Tv = Tg + Tu;
T3A = Tg - Tu;
T2P = FNMS(T5, T20, T2O);
T2E = T2B + T2D;
T3B = T2B - T2D;
T3L = T2b - T23;
T2c = T23 + T2b;
T2R = FNMS(T29, T26, T2Q);
}
}
{
E T2f, TC, T2T, TD, T2d, TI, TS, T2e, T2F, T2l, TO, TT;
{
E TG, TH, TA, TB, TM, TN;
TA = Ip[WS(rs, 7)];
TB = Im[WS(rs, 7)];
TG = Rp[WS(rs, 7)];
T3I = T2R - T2P;
T2S = T2P + T2R;
T2f = TA + TB;
TC = TA - TB;
TH = Rm[WS(rs, 7)];
TM = Ip[WS(rs, 3)];
T2T = Tw * T2f;
TD = Tz * TC;
T2d = TG - TH;
TI = TG + TH;
TN = Im[WS(rs, 3)];
TS = Rp[WS(rs, 3)];
T2e = Tw * T2d;
T2F = Tz * TI;
T2l = TM + TN;
TO = TM - TN;
TT = Rm[WS(rs, 3)];
}
{
E TJ, T2V, TP, T2i, TU, T2G;
TJ = FNMS(TF, TI, TD);
T2V = T2h * T2l;
TP = TL * TO;
T2i = TS - TT;
TU = TS + TT;
T2G = FMA(TF, TC, T2F);
{
E T2g, T2j, TV, T2H;
T2g = FMA(Ty, T2f, T2e);
T2j = T2h * T2i;
TV = FNMS(TR, TU, TP);
T2H = TL * TU;
{
E T2U, T2m, T2I, T2W;
T2U = FNMS(Ty, T2d, T2T);
T2m = FMA(T2k, T2l, T2j);
TW = TJ + TV;
T3E = TJ - TV;
T2I = FMA(TR, TO, T2H);
T2W = FNMS(T2k, T2i, T2V);
T3J = T2m - T2g;
T2n = T2g + T2m;
T3D = T2G - T2I;
T2J = T2G + T2I;
T3M = T2U - T2W;
T2X = T2U + T2W;
}
}
}
}
}
{
E T3Y, T3x, T3X, T3y, T3r, T3q, T3p, T3u;
{
E T2Y, T3o, TX, T3s, T3i, T39, T3t, T3l, T3e, T1x, T2M, T2p, T3d, T2K, T2t;
E T2y;
{
E T2o, T1Z, T3j, T3k, T1i, T1w, T3g, T3h;
T2Y = T2S + T2X;
T3g = T2X - T2S;
T3h = T2c - T2n;
T2o = T2c + T2n;
T1Z = T1N + T1Y;
T3j = T1Y - T1N;
T3o = Tv - TW;
TX = Tv + TW;
T3s = T3g - T3h;
T3i = T3g + T3h;
T3k = T38 - T33;
T39 = T33 + T38;
T3Y = T17 - T1h;
T1i = T17 + T1h;
T1w = T1s + T1v;
T3x = T1v - T1s;
T3t = T3j + T3k;
T3l = T3j - T3k;
T3e = T1w - T1i;
T1x = T1i + T1w;
T2M = T2o + T1Z;
T2p = T1Z - T2o;
T3d = T2J - T2E;
T2K = T2E + T2J;
T3X = T2q - T2s;
T2t = T2q + T2s;
T2y = T2v + T2x;
T3y = T2v - T2x;
}
{
E T2N, T3c, T3a, T3n, T3b, T2L, T2z, T1y;
T2N = T1x - TX;
T1y = TX + T1x;
T3c = T2Y + T39;
T3a = T2Y - T39;
T3n = T2t - T2y;
T2z = T2t + T2y;
Ip[0] = KP500000000 * (T1y + T2p);
Im[WS(rs, 7)] = KP500000000 * (T2p - T1y);
T3b = T2z + T2K;
T2L = T2z - T2K;
{
E T3f, T3m, T3v, T3w;
T3r = T3e - T3d;
T3f = T3d + T3e;
Im[WS(rs, 3)] = KP500000000 * (T3a - T2N);
Ip[WS(rs, 4)] = KP500000000 * (T2N + T3a);
Rp[WS(rs, 4)] = KP500000000 * (T2L + T2M);
Rm[WS(rs, 3)] = KP500000000 * (T2L - T2M);
Rp[0] = KP500000000 * (T3b + T3c);
Rm[WS(rs, 7)] = KP500000000 * (T3b - T3c);
T3m = T3i + T3l;
T3q = T3l - T3i;
T3p = T3n - T3o;
T3v = T3n + T3o;
T3w = T3s + T3t;
T3u = T3s - T3t;
Im[WS(rs, 5)] = -(KP500000000 * (FNMS(KP707106781, T3m, T3f)));
Ip[WS(rs, 2)] = KP500000000 * (FMA(KP707106781, T3m, T3f));
Rp[WS(rs, 2)] = KP500000000 * (FMA(KP707106781, T3w, T3v));
Rm[WS(rs, 5)] = KP500000000 * (FNMS(KP707106781, T3w, T3v));
}
}
}
{
E T3R, T4b, T3z, T4q, T4g, T3U, T40, T41, T4r, T4j, T4m, T3G, T46, T3O, T4l;
E T3Z, T4c;
{
E T3K, T3N, T4h, T4i, T3C, T3F, T4e, T4f;
Rp[WS(rs, 6)] = KP500000000 * (FMA(KP707106781, T3q, T3p));
Rm[WS(rs, 1)] = KP500000000 * (FNMS(KP707106781, T3q, T3p));
Im[WS(rs, 1)] = -(KP500000000 * (FNMS(KP707106781, T3u, T3r)));
Ip[WS(rs, 6)] = KP500000000 * (FMA(KP707106781, T3u, T3r));
T3K = T3I + T3J;
T4e = T3I - T3J;
T4f = T3M - T3L;
T3N = T3L + T3M;
T3R = T3P - T3Q;
T4h = T3Q + T3P;
T4b = T3y + T3x;
T3z = T3x - T3y;
T4q = FNMS(KP414213562, T4e, T4f);
T4g = FMA(KP414213562, T4f, T4e);
T4i = T3T - T3S;
T3U = T3S + T3T;
T40 = T3B + T3A;
T3C = T3A - T3B;
T3F = T3D + T3E;
T41 = T3D - T3E;
T4r = FNMS(KP414213562, T4h, T4i);
T4j = FMA(KP414213562, T4i, T4h);
T4m = T3C - T3F;
T3G = T3C + T3F;
T46 = FNMS(KP414213562, T3K, T3N);
T3O = FMA(KP414213562, T3N, T3K);
T4l = T3X - T3Y;
T3Z = T3X + T3Y;
}
{
E T45, T3H, T42, T47, T3V;
T45 = FNMS(KP707106781, T3G, T3z);
T3H = FMA(KP707106781, T3G, T3z);
T4c = T41 - T40;
T42 = T40 + T41;
T47 = FMA(KP414213562, T3R, T3U);
T3V = FNMS(KP414213562, T3U, T3R);
{
E T49, T43, T48, T4a, T44, T3W;
T49 = FMA(KP707106781, T42, T3Z);
T43 = FNMS(KP707106781, T42, T3Z);
T48 = T46 - T47;
T4a = T46 + T47;
T44 = T3V - T3O;
T3W = T3O + T3V;
Rp[WS(rs, 1)] = KP500000000 * (FMA(KP923879532, T4a, T49));
Rm[WS(rs, 6)] = KP500000000 * (FNMS(KP923879532, T4a, T49));
Rp[WS(rs, 5)] = KP500000000 * (FMA(KP923879532, T44, T43));
Rm[WS(rs, 2)] = KP500000000 * (FNMS(KP923879532, T44, T43));
Im[WS(rs, 6)] = -(KP500000000 * (FNMS(KP923879532, T3W, T3H)));
Ip[WS(rs, 1)] = KP500000000 * (FMA(KP923879532, T3W, T3H));
Ip[WS(rs, 5)] = KP500000000 * (FMA(KP923879532, T48, T45));
Im[WS(rs, 2)] = -(KP500000000 * (FNMS(KP923879532, T48, T45)));
}
}
{
E T4d, T4k, T4t, T4u;
T4p = FMA(KP707106781, T4c, T4b);
T4d = FNMS(KP707106781, T4c, T4b);
T4k = T4g - T4j;
T4o = T4g + T4j;
T4n = FMA(KP707106781, T4m, T4l);
T4t = FNMS(KP707106781, T4m, T4l);
T4u = T4q + T4r;
T4s = T4q - T4r;
Im[0] = -(KP500000000 * (FNMS(KP923879532, T4k, T4d)));
Ip[WS(rs, 7)] = KP500000000 * (FMA(KP923879532, T4k, T4d));
Rm[0] = KP500000000 * (FMA(KP923879532, T4u, T4t));
Rp[WS(rs, 7)] = KP500000000 * (FNMS(KP923879532, T4u, T4t));
}
}
}
}
}
Rp[WS(rs, 3)] = KP500000000 * (FMA(KP923879532, T4o, T4n));
Rm[WS(rs, 4)] = KP500000000 * (FNMS(KP923879532, T4o, T4n));
Im[WS(rs, 4)] = -(KP500000000 * (FNMS(KP923879532, T4s, T4p)));
Ip[WS(rs, 3)] = KP500000000 * (FMA(KP923879532, T4s, T4p));
}
}
}
static void hc2r_15(const R *ri, const R *ii, R *O, stride ris, stride iis, stride os, INT v, INT ivs, INT ovs)
{
DK(KP1_118033988, +1.118033988749894848204586834365638117720309180);
DK(KP1_902113032, +1.902113032590307144232878666758764286811397268);
DK(KP1_175570504, +1.175570504584946258337411909278145537195304875);
DK(KP500000000, +0.500000000000000000000000000000000000000000000);
DK(KP866025403, +0.866025403784438646763723170752936183471402627);
DK(KP2_000000000, +2.000000000000000000000000000000000000000000000);
DK(KP1_732050807, +1.732050807568877293527446341505872366942805254);
INT i;
for (i = v; i > 0; i = i - 1, ri = ri + ivs, ii = ii + ivs, O = O + ovs, MAKE_VOLATILE_STRIDE(ris), MAKE_VOLATILE_STRIDE(iis), MAKE_VOLATILE_STRIDE(os)) {
E T3, Tu, Ti, TB, TZ, T10, TE, TG, TJ, Tn, Tv, Ts, Tw, T8, Td;
E Te;
{
E Th, T1, T2, Tf, Tg;
Tg = ii[WS(iis, 5)];
Th = KP1_732050807 * Tg;
T1 = ri[0];
T2 = ri[WS(ris, 5)];
Tf = T1 - T2;
T3 = FMA(KP2_000000000, T2, T1);
Tu = Tf - Th;
Ti = Tf + Th;
}
{
E T4, TD, T9, TI, T5, T6, T7, Ta, Tb, Tc, Tr, TH, Tm, TC, Tj;
E To;
T4 = ri[WS(ris, 3)];
TD = ii[WS(iis, 3)];
T9 = ri[WS(ris, 6)];
TI = ii[WS(iis, 6)];
T5 = ri[WS(ris, 7)];
T6 = ri[WS(ris, 2)];
T7 = T5 + T6;
Ta = ri[WS(ris, 4)];
Tb = ri[WS(ris, 1)];
Tc = Ta + Tb;
{
E Tp, Tq, Tk, Tl;
Tp = ii[WS(iis, 4)];
Tq = ii[WS(iis, 1)];
Tr = KP866025403 * (Tp + Tq);
TH = Tp - Tq;
Tk = ii[WS(iis, 7)];
Tl = ii[WS(iis, 2)];
Tm = KP866025403 * (Tk - Tl);
TC = Tk + Tl;
}
TB = KP866025403 * (T5 - T6);
TZ = TD - TC;
T10 = TI - TH;
TE = FMA(KP500000000, TC, TD);
TG = KP866025403 * (Ta - Tb);
TJ = FMA(KP500000000, TH, TI);
Tj = FNMS(KP500000000, T7, T4);
Tn = Tj - Tm;
Tv = Tj + Tm;
To = FNMS(KP500000000, Tc, T9);
Ts = To - Tr;
Tw = To + Tr;
T8 = T4 + T7;
Td = T9 + Tc;
Te = T8 + Td;
}
O[0] = FMA(KP2_000000000, Te, T3);
{
E T11, T13, TY, T12, TW, TX;
T11 = FNMS(KP1_902113032, T10, KP1_175570504 * TZ);
T13 = FMA(KP1_902113032, TZ, KP1_175570504 * T10);
TW = FNMS(KP500000000, Te, T3);
TX = KP1_118033988 * (T8 - Td);
TY = TW - TX;
T12 = TX + TW;
O[WS(os, 12)] = TY - T11;
O[WS(os, 9)] = T12 + T13;
O[WS(os, 3)] = TY + T11;
O[WS(os, 6)] = T12 - T13;
}
{
E TP, Tt, TO, TT, TV, TR, TS, TU, TQ;
TP = KP1_118033988 * (Tn - Ts);
Tt = Tn + Ts;
TO = FNMS(KP500000000, Tt, Ti);
TR = TE - TB;
TS = TJ - TG;
TT = FNMS(KP1_902113032, TS, KP1_175570504 * TR);
TV = FMA(KP1_902113032, TR, KP1_175570504 * TS);
O[WS(os, 5)] = FMA(KP2_000000000, Tt, Ti);
TU = TP + TO;
O[WS(os, 11)] = TU - TV;
O[WS(os, 14)] = TU + TV;
TQ = TO - TP;
O[WS(os, 2)] = TQ - TT;
O[WS(os, 8)] = TQ + TT;
}
{
E Tz, Tx, Ty, TL, TN, TF, TK, TM, TA;
Tz = KP1_118033988 * (Tv - Tw);
Tx = Tv + Tw;
Ty = FNMS(KP500000000, Tx, Tu);
TF = TB + TE;
TK = TG + TJ;
TL = FNMS(KP1_902113032, TK, KP1_175570504 * TF);
TN = FMA(KP1_902113032, TF, KP1_175570504 * TK);
O[WS(os, 10)] = FMA(KP2_000000000, Tx, Tu);
TM = Tz + Ty;
O[WS(os, 1)] = TM - TN;
O[WS(os, 4)] = TM + TN;
TA = Ty - Tz;
O[WS(os, 7)] = TA - TL;
O[WS(os, 13)] = TA + TL;
}
}
}
static void hc2cfdft2_16(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms)
{
DK(KP461939766, +0.461939766255643378064091594698394143411208313);
DK(KP191341716, +0.191341716182544885864229992015199433380672281);
DK(KP353553390, +0.353553390593273762200422181052424519642417969);
DK(KP500000000, +0.500000000000000000000000000000000000000000000);
{
INT m;
for (m = mb, W = W + ((mb - 1) * 8); m < me; m = m + 1, Rp = Rp + ms, Ip = Ip + ms, Rm = Rm - ms, Im = Im - ms, W = W + 8, MAKE_VOLATILE_STRIDE(64, rs)) {
E T1, T4, T2, T5, T7, Td, T12, TY, Tk, Ti, Tm, T1l, T1b, TL, T1h;
E Ts, TR, T17, Ty, Tz, TA, TE, T1L, T1Q, T1H, T1O, T24, T2d, T20, T2b;
{
E Tl, TP, Tq, TK, Tj, TQ, Tr, TJ;
{
E T3, Tc, T6, Tb;
T1 = W[0];
T4 = W[1];
T2 = W[2];
T5 = W[3];
T3 = T1 * T2;
Tc = T4 * T2;
T6 = T4 * T5;
Tb = T1 * T5;
T7 = T3 + T6;
Td = Tb - Tc;
T12 = Tb + Tc;
TY = T3 - T6;
Tk = W[5];
Tl = T4 * Tk;
TP = T2 * Tk;
Tq = T1 * Tk;
TK = T5 * Tk;
Ti = W[4];
Tj = T1 * Ti;
TQ = T5 * Ti;
Tr = T4 * Ti;
TJ = T2 * Ti;
}
Tm = Tj - Tl;
T1l = Tq - Tr;
T1b = TP + TQ;
TL = TJ + TK;
T1h = Tj + Tl;
Ts = Tq + Tr;
TR = TP - TQ;
T17 = TJ - TK;
Ty = W[6];
Tz = W[7];
TA = FMA(T1, Ty, T4 * Tz);
TE = FNMS(T4, Ty, T1 * Tz);
{
E T1J, T1K, T1F, T1G;
T1J = TY * Tk;
T1K = T12 * Ti;
T1L = T1J - T1K;
T1Q = T1J + T1K;
T1F = TY * Ti;
T1G = T12 * Tk;
T1H = T1F + T1G;
T1O = T1F - T1G;
}
{
E T22, T23, T1Y, T1Z;
T22 = T7 * Tk;
T23 = Td * Ti;
T24 = T22 + T23;
T2d = T22 - T23;
T1Y = T7 * Ti;
T1Z = Td * Tk;
T20 = T1Y - T1Z;
T2b = T1Y + T1Z;
}
}
{
E T1t, T3i, T2l, T3B, T1E, T3t, T2M, T3x, T1g, T3C, T2J, T3u, T1T, T3w, T2o;
E T3j, Tx, T3b, T2C, T3q, T27, T3m, T2s, T3c, TW, T3f, T2F, T3n, T2g, T3p;
E T2v, T3e;
{
E T1k, T1C, T1o, T1B, T1s, T1z, T1y, T2j, T1p, T2k;
{
E T1i, T1j, T1m, T1n;
T1i = Ip[WS(rs, 4)];
T1j = Im[WS(rs, 4)];
T1k = T1i - T1j;
T1C = T1i + T1j;
T1m = Rp[WS(rs, 4)];
T1n = Rm[WS(rs, 4)];
T1o = T1m + T1n;
T1B = T1m - T1n;
}
{
E T1q, T1r, T1w, T1x;
T1q = Ip[0];
T1r = Im[0];
T1s = T1q - T1r;
T1z = T1q + T1r;
T1w = Rm[0];
T1x = Rp[0];
T1y = T1w - T1x;
T2j = T1x + T1w;
}
T1p = FNMS(T1l, T1o, T1h * T1k);
T1t = T1p + T1s;
T3i = T1s - T1p;
T2k = FMA(T1h, T1o, T1l * T1k);
T2l = T2j + T2k;
T3B = T2j - T2k;
{
E T1A, T1D, T2K, T2L;
T1A = FNMS(T4, T1z, T1 * T1y);
T1D = FMA(Ti, T1B, Tk * T1C);
T1E = T1A - T1D;
T3t = T1D + T1A;
T2K = FNMS(Tk, T1B, Ti * T1C);
T2L = FMA(T4, T1y, T1 * T1z);
T2M = T2K + T2L;
T3x = T2L - T2K;
}
}
{
E T11, T1M, T15, T1I, T1a, T1R, T1e, T1P;
{
E TZ, T10, T13, T14;
TZ = Ip[WS(rs, 2)];
T10 = Im[WS(rs, 2)];
T11 = TZ - T10;
T1M = TZ + T10;
T13 = Rp[WS(rs, 2)];
T14 = Rm[WS(rs, 2)];
T15 = T13 + T14;
T1I = T13 - T14;
}
{
E T18, T19, T1c, T1d;
T18 = Ip[WS(rs, 6)];
T19 = Im[WS(rs, 6)];
T1a = T18 - T19;
T1R = T18 + T19;
T1c = Rp[WS(rs, 6)];
T1d = Rm[WS(rs, 6)];
T1e = T1c + T1d;
T1P = T1c - T1d;
}
{
E T16, T1f, T2H, T2I;
T16 = FNMS(T12, T15, TY * T11);
T1f = FNMS(T1b, T1e, T17 * T1a);
T1g = T16 + T1f;
T3C = T16 - T1f;
T2H = FNMS(T1L, T1I, T1H * T1M);
T2I = FNMS(T1Q, T1P, T1O * T1R);
T2J = T2H + T2I;
T3u = T2H - T2I;
}
{
E T1N, T1S, T2m, T2n;
T1N = FMA(T1H, T1I, T1L * T1M);
T1S = FMA(T1O, T1P, T1Q * T1R);
T1T = T1N + T1S;
T3w = T1S - T1N;
T2m = FMA(TY, T15, T12 * T11);
T2n = FMA(T17, T1e, T1b * T1a);
T2o = T2m + T2n;
T3j = T2m - T2n;
}
}
{
E Ta, T1W, Tg, T1V, Tp, T25, Tv, T21;
{
E T8, T9, Te, Tf;
T8 = Ip[WS(rs, 1)];
T9 = Im[WS(rs, 1)];
Ta = T8 - T9;
T1W = T8 + T9;
Te = Rp[WS(rs, 1)];
Tf = Rm[WS(rs, 1)];
Tg = Te + Tf;
T1V = Te - Tf;
}
{
E Tn, To, Tt, Tu;
Tn = Ip[WS(rs, 5)];
To = Im[WS(rs, 5)];
Tp = Tn - To;
T25 = Tn + To;
Tt = Rp[WS(rs, 5)];
Tu = Rm[WS(rs, 5)];
Tv = Tt + Tu;
T21 = Tt - Tu;
}
{
E Th, Tw, T2A, T2B;
Th = FNMS(Td, Tg, T7 * Ta);
Tw = FNMS(Ts, Tv, Tm * Tp);
Tx = Th + Tw;
T3b = Th - Tw;
T2A = FNMS(T5, T1V, T2 * T1W);
T2B = FNMS(T24, T21, T20 * T25);
T2C = T2A + T2B;
T3q = T2A - T2B;
}
{
E T1X, T26, T2q, T2r;
T1X = FMA(T2, T1V, T5 * T1W);
T26 = FMA(T20, T21, T24 * T25);
T27 = T1X + T26;
T3m = T26 - T1X;
T2q = FMA(T7, Tg, Td * Ta);
T2r = FMA(Tm, Tv, Ts * Tp);
T2s = T2q + T2r;
T3c = T2q - T2r;
}
}
{
E TD, T29, TH, T28, TO, T2e, TU, T2c;
{
E TB, TC, TF, TG;
TB = Ip[WS(rs, 7)];
TC = Im[WS(rs, 7)];
TD = TB - TC;
T29 = TB + TC;
TF = Rp[WS(rs, 7)];
TG = Rm[WS(rs, 7)];
TH = TF + TG;
T28 = TF - TG;
}
{
E TM, TN, TS, TT;
TM = Ip[WS(rs, 3)];
TN = Im[WS(rs, 3)];
TO = TM - TN;
T2e = TM + TN;
TS = Rp[WS(rs, 3)];
TT = Rm[WS(rs, 3)];
TU = TS + TT;
T2c = TS - TT;
}
{
E TI, TV, T2D, T2E;
TI = FNMS(TE, TH, TA * TD);
TV = FNMS(TR, TU, TL * TO);
TW = TI + TV;
T3f = TI - TV;
T2D = FNMS(Tz, T28, Ty * T29);
T2E = FNMS(T2d, T2c, T2b * T2e);
T2F = T2D + T2E;
T3n = T2D - T2E;
}
{
E T2a, T2f, T2t, T2u;
T2a = FMA(Ty, T28, Tz * T29);
T2f = FMA(T2b, T2c, T2d * T2e);
T2g = T2a + T2f;
T3p = T2f - T2a;
T2t = FMA(TA, TH, TE * TD);
T2u = FMA(TL, TU, TR * TO);
T2v = T2t + T2u;
T3e = T2t - T2u;
}
}
{
E T1v, T2z, T2O, T2Q, T2i, T2y, T2x, T2P;
{
E TX, T1u, T2G, T2N;
TX = Tx + TW;
T1u = T1g + T1t;
T1v = TX + T1u;
T2z = T1u - TX;
T2G = T2C + T2F;
T2N = T2J + T2M;
T2O = T2G - T2N;
T2Q = T2G + T2N;
}
{
E T1U, T2h, T2p, T2w;
T1U = T1E - T1T;
T2h = T27 + T2g;
T2i = T1U - T2h;
T2y = T2h + T1U;
T2p = T2l + T2o;
T2w = T2s + T2v;
T2x = T2p - T2w;
T2P = T2p + T2w;
}
Ip[0] = KP500000000 * (T1v + T2i);
Rp[0] = KP500000000 * (T2P + T2Q);
Im[WS(rs, 7)] = KP500000000 * (T2i - T1v);
Rm[WS(rs, 7)] = KP500000000 * (T2P - T2Q);
Rm[WS(rs, 3)] = KP500000000 * (T2x - T2y);
Im[WS(rs, 3)] = KP500000000 * (T2O - T2z);
Rp[WS(rs, 4)] = KP500000000 * (T2x + T2y);
Ip[WS(rs, 4)] = KP500000000 * (T2z + T2O);
}
{
E T2T, T35, T33, T39, T2W, T36, T2Z, T37;
{
E T2R, T2S, T31, T32;
T2R = T2v - T2s;
T2S = T1t - T1g;
T2T = KP500000000 * (T2R + T2S);
T35 = KP500000000 * (T2S - T2R);
T31 = T2l - T2o;
T32 = Tx - TW;
T33 = KP500000000 * (T31 - T32);
T39 = KP500000000 * (T31 + T32);
}
{
E T2U, T2V, T2X, T2Y;
T2U = T2F - T2C;
T2V = T27 - T2g;
T2W = T2U + T2V;
T36 = T2U - T2V;
T2X = T1T + T1E;
T2Y = T2M - T2J;
T2Z = T2X - T2Y;
T37 = T2X + T2Y;
}
{
E T30, T3a, T34, T38;
T30 = KP353553390 * (T2W + T2Z);
Ip[WS(rs, 2)] = T2T + T30;
Im[WS(rs, 5)] = T30 - T2T;
T3a = KP353553390 * (T36 + T37);
Rm[WS(rs, 5)] = T39 - T3a;
Rp[WS(rs, 2)] = T39 + T3a;
T34 = KP353553390 * (T2Z - T2W);
Rm[WS(rs, 1)] = T33 - T34;
Rp[WS(rs, 6)] = T33 + T34;
T38 = KP353553390 * (T36 - T37);
Ip[WS(rs, 6)] = T35 + T38;
Im[WS(rs, 1)] = T38 - T35;
}
}
{
E T3k, T3Q, T3Z, T3D, T3h, T40, T3X, T45, T3G, T3P, T3s, T3K, T3U, T44, T3z;
E T3L;
{
E T3d, T3g, T3o, T3r;
T3k = KP500000000 * (T3i - T3j);
T3Q = KP500000000 * (T3j + T3i);
T3Z = KP500000000 * (T3B - T3C);
T3D = KP500000000 * (T3B + T3C);
T3d = T3b - T3c;
T3g = T3e + T3f;
T3h = KP353553390 * (T3d + T3g);
T40 = KP353553390 * (T3d - T3g);
{
E T3V, T3W, T3E, T3F;
T3V = T3u + T3t;
T3W = T3x - T3w;
T3X = FNMS(KP461939766, T3W, KP191341716 * T3V);
T45 = FMA(KP461939766, T3V, KP191341716 * T3W);
T3E = T3c + T3b;
T3F = T3e - T3f;
T3G = KP353553390 * (T3E + T3F);
T3P = KP353553390 * (T3F - T3E);
}
T3o = T3m + T3n;
T3r = T3p - T3q;
T3s = FMA(KP191341716, T3o, KP461939766 * T3r);
T3K = FNMS(KP191341716, T3r, KP461939766 * T3o);
{
E T3S, T3T, T3v, T3y;
T3S = T3n - T3m;
T3T = T3q + T3p;
T3U = FMA(KP461939766, T3S, KP191341716 * T3T);
T44 = FNMS(KP461939766, T3T, KP191341716 * T3S);
T3v = T3t - T3u;
T3y = T3w + T3x;
T3z = FNMS(KP191341716, T3y, KP461939766 * T3v);
T3L = FMA(KP191341716, T3v, KP461939766 * T3y);
}
}
{
E T3l, T3A, T3N, T3O;
T3l = T3h + T3k;
T3A = T3s + T3z;
Ip[WS(rs, 1)] = T3l + T3A;
Im[WS(rs, 6)] = T3A - T3l;
T3N = T3D + T3G;
T3O = T3K + T3L;
Rm[WS(rs, 6)] = T3N - T3O;
Rp[WS(rs, 1)] = T3N + T3O;
}
{
E T3H, T3I, T3J, T3M;
T3H = T3D - T3G;
T3I = T3z - T3s;
Rm[WS(rs, 2)] = T3H - T3I;
Rp[WS(rs, 5)] = T3H + T3I;
T3J = T3k - T3h;
T3M = T3K - T3L;
Ip[WS(rs, 5)] = T3J + T3M;
Im[WS(rs, 2)] = T3M - T3J;
}
{
E T3R, T3Y, T47, T48;
T3R = T3P + T3Q;
T3Y = T3U + T3X;
Ip[WS(rs, 3)] = T3R + T3Y;
Im[WS(rs, 4)] = T3Y - T3R;
T47 = T3Z + T40;
T48 = T44 + T45;
Rm[WS(rs, 4)] = T47 - T48;
Rp[WS(rs, 3)] = T47 + T48;
}
{
E T41, T42, T43, T46;
T41 = T3Z - T40;
T42 = T3X - T3U;
Rm[0] = T41 - T42;
Rp[WS(rs, 7)] = T41 + T42;
T43 = T3Q - T3P;
T46 = T44 - T45;
Ip[WS(rs, 7)] = T43 + T46;
Im[0] = T46 - T43;
}
}
}
}
}
}
static void r2cbIII_32(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
{
DK(KP534511135, +0.534511135950791641089685961295362908582039528);
DK(KP1_763842528, +1.763842528696710059425513727320776699016885241);
DK(KP303346683, +0.303346683607342391675883946941299872384187453);
DK(KP1_913880671, +1.913880671464417729871595773960539938965698411);
DK(KP098491403, +0.098491403357164253077197521291327432293052451);
DK(KP1_990369453, +1.990369453344393772489673906218959843150949737);
DK(KP820678790, +0.820678790828660330972281985331011598767386482);
DK(KP1_546020906, +1.546020906725473921621813219516939601942082586);
DK(KP1_847759065, +1.847759065022573512256366378793576573644833252);
DK(KP923879532, +0.923879532511286756128183189396788286822416626);
DK(KP668178637, +0.668178637919298919997757686523080761552472251);
DK(KP1_662939224, +1.662939224605090474157576755235811513477121624);
DK(KP198912367, +0.198912367379658006911597622644676228597850501);
DK(KP1_961570560, +1.961570560806460898252364472268478073947867462);
DK(KP707106781, +0.707106781186547524400844362104849039284835938);
DK(KP1_414213562, +1.414213562373095048801688724209698078569671875);
DK(KP2_000000000, +2.000000000000000000000000000000000000000000000);
DK(KP414213562, +0.414213562373095048801688724209698078569671875);
{
INT i;
for (i = v; i > 0; i = i - 1, R0 = R0 + ovs, R1 = R1 + ovs, Cr = Cr + ivs, Ci = Ci + ivs, MAKE_VOLATILE_STRIDE(128, rs), MAKE_VOLATILE_STRIDE(128, csr), MAKE_VOLATILE_STRIDE(128, csi)) {
E T1N, T1K, T1Q, T1H, T1O, T1P;
{
E T1I, T1e, T1Z, T7, T2E, T2i, T1x, Tz, Te, T2j, T22, T2F, T1h, T1y, TK;
E T1J, Tm, T2B, TX, Tp, T2m, T28, T1M, T1C, T1k, TW, TY, T2a, T14, T15;
E Ts, TZ;
{
E TE, T1g, TJ, T1f;
{
E T4, Tv, T3, T2g, T1d, T5, Tw, Tx;
{
E T1, T2, T1b, T1c;
T1 = Cr[0];
T2 = Cr[WS(csr, 15)];
T1b = Ci[0];
T1c = Ci[WS(csi, 15)];
T4 = Cr[WS(csr, 8)];
Tv = T1 - T2;
T3 = T1 + T2;
T2g = T1c - T1b;
T1d = T1b + T1c;
T5 = Cr[WS(csr, 7)];
Tw = Ci[WS(csi, 8)];
Tx = Ci[WS(csi, 7)];
}
{
E Tb, TA, Ta, T20, TD, Tc, TG, TH;
{
E T8, T9, TB, TC;
T8 = Cr[WS(csr, 4)];
{
E T1a, T6, T2h, Ty;
T1a = T4 - T5;
T6 = T4 + T5;
T2h = Tx - Tw;
Ty = Tw + Tx;
T1I = T1a - T1d;
T1e = T1a + T1d;
T1Z = T3 - T6;
T7 = T3 + T6;
T2E = T2h + T2g;
T2i = T2g - T2h;
T1x = Tv + Ty;
Tz = Tv - Ty;
T9 = Cr[WS(csr, 11)];
}
TB = Ci[WS(csi, 4)];
TC = Ci[WS(csi, 11)];
Tb = Cr[WS(csr, 3)];
TA = T8 - T9;
Ta = T8 + T9;
T20 = TC - TB;
TD = TB + TC;
Tc = Cr[WS(csr, 12)];
TG = Ci[WS(csi, 3)];
TH = Ci[WS(csi, 12)];
}
{
E TF, Td, T21, TI;
TE = TA - TD;
T1g = TA + TD;
TF = Tb - Tc;
Td = Tb + Tc;
T21 = TG - TH;
TI = TG + TH;
Te = Ta + Td;
T2j = Ta - Td;
T22 = T20 - T21;
T2F = T20 + T21;
TJ = TF - TI;
T1f = TF + TI;
}
}
}
{
E TM, Ti, TN, T25, TU, TR, Tl, TO;
{
E TS, TT, Tg, Th, Tj, Tk;
Tg = Cr[WS(csr, 2)];
Th = Cr[WS(csr, 13)];
T1h = T1f - T1g;
T1y = T1g + T1f;
TK = TE + TJ;
T1J = TE - TJ;
TM = Tg - Th;
Ti = Tg + Th;
TS = Ci[WS(csi, 2)];
TT = Ci[WS(csi, 13)];
Tj = Cr[WS(csr, 10)];
Tk = Cr[WS(csr, 5)];
TN = Ci[WS(csi, 10)];
T25 = TS - TT;
TU = TS + TT;
TR = Tj - Tk;
Tl = Tj + Tk;
TO = Ci[WS(csi, 5)];
}
{
E T12, T13, Tq, Tr;
{
E Tn, T1A, TV, T24, T26, TP, To, T27, T1B, TQ;
Tn = Cr[WS(csr, 1)];
T1A = TR - TU;
TV = TR + TU;
T24 = Ti - Tl;
Tm = Ti + Tl;
T26 = TN - TO;
TP = TN + TO;
To = Cr[WS(csr, 14)];
T12 = Ci[WS(csi, 1)];
T27 = T25 - T26;
T2B = T26 + T25;
T1B = TM + TP;
TQ = TM - TP;
TX = Tn - To;
Tp = Tn + To;
T2m = T24 + T27;
T28 = T24 - T27;
T1M = FNMS(KP414213562, T1A, T1B);
T1C = FMA(KP414213562, T1B, T1A);
T1k = FMA(KP414213562, TQ, TV);
TW = FNMS(KP414213562, TV, TQ);
T13 = Ci[WS(csi, 14)];
}
Tq = Cr[WS(csr, 6)];
Tr = Cr[WS(csr, 9)];
TY = Ci[WS(csi, 6)];
T2a = T13 - T12;
T14 = T12 + T13;
T15 = Tq - Tr;
Ts = Tq + Tr;
TZ = Ci[WS(csi, 9)];
}
}
}
{
E T1L, T1F, T23, T2n, T2k, T2e, T1p, T1t, T1s, T1i, T1o, T19, T1l, T1q;
{
E T2z, T2G, T2H, T2C, T1j, T17, T2r, T2s, T2u, T2v, T2K, T2D;
{
E T2L, T2d, T2l, T2O;
{
E Tf, T2N, Tu, T2M;
{
E T1D, T16, T29, Tt, T2b, T10;
T2z = T7 - Te;
Tf = T7 + Te;
T1D = T15 + T14;
T16 = T14 - T15;
T29 = Tp - Ts;
Tt = Tp + Ts;
T2b = TY - TZ;
T10 = TY + TZ;
T2N = T2F + T2E;
T2G = T2E - T2F;
T2H = Tm - Tt;
Tu = Tm + Tt;
{
E T2c, T2A, T1E, T11;
T2c = T2a - T2b;
T2A = T2b + T2a;
T1E = TX + T10;
T11 = TX - T10;
T2L = Tf - Tu;
T2d = T29 + T2c;
T2l = T29 - T2c;
T2C = T2A - T2B;
T2M = T2B + T2A;
T1L = FMA(KP414213562, T1D, T1E);
T1F = FNMS(KP414213562, T1E, T1D);
T1j = FMA(KP414213562, T11, T16);
T17 = FNMS(KP414213562, T16, T11);
T2O = T2M + T2N;
}
}
R0[0] = KP2_000000000 * (Tf + Tu);
R0[WS(rs, 8)] = KP2_000000000 * (T2N - T2M);
}
T23 = T1Z + T22;
T2r = T1Z - T22;
R0[WS(rs, 12)] = KP1_414213562 * (T2O - T2L);
R0[WS(rs, 4)] = KP1_414213562 * (T2L + T2O);
T2s = T2m + T2l;
T2n = T2l - T2m;
T2k = T2i - T2j;
T2u = T2j + T2i;
T2v = T28 - T2d;
T2e = T28 + T2d;
}
{
E T2y, T2t, T2x, T2w;
T2y = FMA(KP707106781, T2s, T2r);
T2t = FNMS(KP707106781, T2s, T2r);
T2x = FMA(KP707106781, T2v, T2u);
T2w = FNMS(KP707106781, T2v, T2u);
R0[WS(rs, 7)] = KP1_961570560 * (FMA(KP198912367, T2y, T2x));
R0[WS(rs, 15)] = -(KP1_961570560 * (FNMS(KP198912367, T2x, T2y)));
R0[WS(rs, 11)] = KP1_662939224 * (FNMS(KP668178637, T2t, T2w));
R0[WS(rs, 3)] = KP1_662939224 * (FMA(KP668178637, T2w, T2t));
T2K = T2z - T2C;
T2D = T2z + T2C;
}
{
E TL, T18, T2J, T2I;
T1p = FNMS(KP707106781, TK, Tz);
TL = FMA(KP707106781, TK, Tz);
T18 = TW + T17;
T1t = TW - T17;
T1s = FMA(KP707106781, T1h, T1e);
T1i = FNMS(KP707106781, T1h, T1e);
T2J = T2H + T2G;
T2I = T2G - T2H;
T1o = FNMS(KP923879532, T18, TL);
T19 = FMA(KP923879532, T18, TL);
R0[WS(rs, 6)] = KP1_847759065 * (FMA(KP414213562, T2K, T2J));
R0[WS(rs, 14)] = -(KP1_847759065 * (FNMS(KP414213562, T2J, T2K)));
R0[WS(rs, 10)] = KP1_847759065 * (FNMS(KP414213562, T2D, T2I));
R0[WS(rs, 2)] = KP1_847759065 * (FMA(KP414213562, T2I, T2D));
T1l = T1j - T1k;
T1q = T1k + T1j;
}
}
{
E T1z, T1U, T1Y, T1T, T1V, T1G;
{
E T1w, T1r, T1n, T1m;
T1n = FMA(KP923879532, T1l, T1i);
T1m = FNMS(KP923879532, T1l, T1i);
T1w = FMA(KP923879532, T1q, T1p);
T1r = FNMS(KP923879532, T1q, T1p);
R1[WS(rs, 4)] = -(KP1_546020906 * (FNMS(KP820678790, T1o, T1n)));
R1[WS(rs, 12)] = -(KP1_546020906 * (FMA(KP820678790, T1n, T1o)));
R1[WS(rs, 8)] = -(KP1_990369453 * (FMA(KP098491403, T19, T1m)));
R1[0] = KP1_990369453 * (FNMS(KP098491403, T1m, T19));
{
E T1R, T1S, T1v, T1u;
T1z = FNMS(KP707106781, T1y, T1x);
T1R = FMA(KP707106781, T1y, T1x);
T1S = T1M + T1L;
T1N = T1L - T1M;
T1K = FNMS(KP707106781, T1J, T1I);
T1U = FMA(KP707106781, T1J, T1I);
T1v = FNMS(KP923879532, T1t, T1s);
T1u = FMA(KP923879532, T1t, T1s);
T1Y = FMA(KP923879532, T1S, T1R);
T1T = FNMS(KP923879532, T1S, T1R);
R1[WS(rs, 6)] = -(KP1_913880671 * (FNMS(KP303346683, T1w, T1v)));
R1[WS(rs, 14)] = -(KP1_913880671 * (FMA(KP303346683, T1v, T1w)));
R1[WS(rs, 10)] = -(KP1_763842528 * (FMA(KP534511135, T1r, T1u)));
R1[WS(rs, 2)] = KP1_763842528 * (FNMS(KP534511135, T1u, T1r));
T1V = T1C + T1F;
T1G = T1C - T1F;
}
}
{
E T2q, T2f, T1X, T1W, T2p, T2o;
T1X = FMA(KP923879532, T1V, T1U);
T1W = FNMS(KP923879532, T1V, T1U);
T2q = FNMS(KP707106781, T2e, T23);
T2f = FMA(KP707106781, T2e, T23);
R1[WS(rs, 7)] = KP1_990369453 * (FMA(KP098491403, T1Y, T1X));
R1[WS(rs, 15)] = -(KP1_990369453 * (FNMS(KP098491403, T1X, T1Y)));
R1[WS(rs, 11)] = KP1_546020906 * (FNMS(KP820678790, T1T, T1W));
R1[WS(rs, 3)] = KP1_546020906 * (FMA(KP820678790, T1W, T1T));
T2p = FNMS(KP707106781, T2n, T2k);
T2o = FMA(KP707106781, T2n, T2k);
T1Q = FNMS(KP923879532, T1G, T1z);
T1H = FMA(KP923879532, T1G, T1z);
R0[WS(rs, 5)] = KP1_662939224 * (FMA(KP668178637, T2q, T2p));
R0[WS(rs, 13)] = -(KP1_662939224 * (FNMS(KP668178637, T2p, T2q)));
R0[WS(rs, 9)] = KP1_961570560 * (FNMS(KP198912367, T2f, T2o));
R0[WS(rs, 1)] = KP1_961570560 * (FMA(KP198912367, T2o, T2f));
}
}
}
}
T1O = FMA(KP923879532, T1N, T1K);
T1P = FNMS(KP923879532, T1N, T1K);
R1[WS(rs, 5)] = KP1_763842528 * (FMA(KP534511135, T1Q, T1P));
R1[WS(rs, 13)] = -(KP1_763842528 * (FNMS(KP534511135, T1P, T1Q)));
R1[WS(rs, 9)] = KP1_913880671 * (FNMS(KP303346683, T1H, T1O));
R1[WS(rs, 1)] = KP1_913880671 * (FMA(KP303346683, T1O, T1H));
}
}
}
static void hc2cfdft_4(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms)
{
DK(KP500000000, +0.500000000000000000000000000000000000000000000);
{
INT m;
for (m = mb, W = W + ((mb - 1) * 6); m < me; m = m + 1, Rp = Rp + ms, Ip = Ip + ms, Rm = Rm - ms, Im = Im - ms, W = W + 6, MAKE_VOLATILE_STRIDE(16, rs)) {
E Tc, Tr, Tk, Tx, T9, Ts, Tp, Tw;
{
E Ta, Tb, Tj, Tf, Tg, Th, Te, Ti;
Ta = Ip[0];
Tb = Im[0];
Tj = Ta + Tb;
Tf = Rm[0];
Tg = Rp[0];
Th = Tf - Tg;
Tc = Ta - Tb;
Tr = Tg + Tf;
Te = W[0];
Ti = W[1];
Tk = FNMS(Ti, Tj, Te * Th);
Tx = FMA(Ti, Th, Te * Tj);
}
{
E T4, To, T8, Tm;
{
E T2, T3, T6, T7;
T2 = Ip[WS(rs, 1)];
T3 = Im[WS(rs, 1)];
T4 = T2 - T3;
To = T2 + T3;
T6 = Rp[WS(rs, 1)];
T7 = Rm[WS(rs, 1)];
T8 = T6 + T7;
Tm = T6 - T7;
}
{
E T1, T5, Tl, Tn;
T1 = W[2];
T5 = W[3];
T9 = FNMS(T5, T8, T1 * T4);
Ts = FMA(T1, T8, T5 * T4);
Tl = W[4];
Tn = W[5];
Tp = FMA(Tl, Tm, Tn * To);
Tw = FNMS(Tn, Tm, Tl * To);
}
}
{
E Td, Tq, Tz, TA;
Td = T9 + Tc;
Tq = Tk - Tp;
Ip[0] = KP500000000 * (Td + Tq);
Im[WS(rs, 1)] = KP500000000 * (Tq - Td);
Tz = Tr + Ts;
TA = Tw + Tx;
Rm[WS(rs, 1)] = KP500000000 * (Tz - TA);
Rp[0] = KP500000000 * (Tz + TA);
}
{
E Tt, Tu, Tv, Ty;
Tt = Tr - Ts;
Tu = Tp + Tk;
Rm[0] = KP500000000 * (Tt - Tu);
Rp[WS(rs, 1)] = KP500000000 * (Tt + Tu);
Tv = Tc - T9;
Ty = Tw - Tx;
Ip[WS(rs, 1)] = KP500000000 * (Tv + Ty);
Im[0] = KP500000000 * (Ty - Tv);
}
}
}
}
static void hb2_16(R *cr, R *ci, const R *W, stride rs, INT mb, INT me, INT ms)
{
DK(KP923879532, +0.923879532511286756128183189396788286822416626);
DK(KP707106781, +0.707106781186547524400844362104849039284835938);
DK(KP414213562, +0.414213562373095048801688724209698078569671875);
{
INT m;
for (m = mb, W = W + ((mb - 1) * 8); m < me; m = m + 1, cr = cr + ms, ci = ci - ms, W = W + 8, MAKE_VOLATILE_STRIDE(32, rs)) {
E Tv, Tw, T2z, T2C, TB, TF, Ty, Tz, T1V, TA, T2G, T3Q, T3C, T3g, T3L;
E T30, T3m, T3z, T3w, T3s, T1X, T1Y, T2u, T2c, T2p, TE, TG, T1G, T1o, T1D;
{
E T3f, T3l, T2F, T3r, T2Z, T3v, TD, Tx;
Tv = W[0];
Tw = W[2];
Tx = Tv * Tw;
T2z = W[6];
T3f = Tv * T2z;
T2C = W[7];
T3l = Tv * T2C;
TB = W[4];
T2F = Tv * TB;
T3r = Tw * TB;
TF = W[5];
T2Z = Tv * TF;
T3v = Tw * TF;
Ty = W[1];
Tz = W[3];
TD = Tv * Tz;
T1V = FMA(Ty, Tz, Tx);
TA = FNMS(Ty, Tz, Tx);
T2G = FNMS(Ty, TF, T2F);
T3Q = FMA(Tz, TB, T3v);
T3C = FNMS(Ty, TB, T2Z);
T3g = FMA(Ty, T2C, T3f);
T3L = FNMS(Tz, TF, T3r);
T30 = FMA(Ty, TB, T2Z);
T3m = FNMS(Ty, T2z, T3l);
T3z = FMA(Ty, TF, T2F);
T3w = FNMS(Tz, TB, T3v);
T3s = FMA(Tz, TF, T3r);
{
E T1W, T2b, TC, T1n;
T1W = T1V * TB;
T2b = T1V * TF;
T1X = FNMS(Ty, Tw, TD);
T1Y = FNMS(T1X, TF, T1W);
T2u = FNMS(T1X, TB, T2b);
T2c = FMA(T1X, TB, T2b);
T2p = FMA(T1X, TF, T1W);
TC = TA * TB;
T1n = TA * TF;
TE = FMA(Ty, Tw, TD);
TG = FNMS(TE, TF, TC);
T1G = FNMS(TE, TB, T1n);
T1o = FMA(TE, TB, T1n);
T1D = FMA(TE, TF, TC);
}
}
{
E TL, T1Z, T2d, T1t, T31, T34, T3n, T3D, T3E, T3R, T1w, T20, Tf, T3M, T2L;
E T3h, TW, T2e, T3G, T3H, T3N, T2Q, T36, T2V, T37, Tu, T3S, T18, T1z, T24;
E T2g, T27, T2h, T1j, T1y;
{
E T3, TH, TU, T2I, T1s, T32, T6, T1p, Ta, TM, TK, T33, TP, T2J, Td;
E TR;
{
E T1, T2, TS, TT;
T1 = cr[0];
T2 = ci[WS(rs, 7)];
T3 = T1 + T2;
TH = T1 - T2;
TS = ci[WS(rs, 9)];
TT = cr[WS(rs, 14)];
TU = TS + TT;
T2I = TS - TT;
}
{
E T1q, T1r, T4, T5;
T1q = ci[WS(rs, 15)];
T1r = cr[WS(rs, 8)];
T1s = T1q + T1r;
T32 = T1q - T1r;
T4 = cr[WS(rs, 4)];
T5 = ci[WS(rs, 3)];
T6 = T4 + T5;
T1p = T4 - T5;
}
{
E T8, T9, TI, TJ;
T8 = cr[WS(rs, 2)];
T9 = ci[WS(rs, 5)];
Ta = T8 + T9;
TM = T8 - T9;
TI = ci[WS(rs, 11)];
TJ = cr[WS(rs, 12)];
TK = TI + TJ;
T33 = TI - TJ;
}
{
E TN, TO, Tb, Tc;
TN = ci[WS(rs, 13)];
TO = cr[WS(rs, 10)];
TP = TN + TO;
T2J = TN - TO;
Tb = ci[WS(rs, 1)];
Tc = cr[WS(rs, 6)];
Td = Tb + Tc;
TR = Tb - Tc;
}
TL = TH - TK;
T1Z = TH + TK;
T2d = T1s - T1p;
T1t = T1p + T1s;
T31 = Ta - Td;
T34 = T32 - T33;
T3n = T34 - T31;
{
E T1u, T1v, T7, Te;
T3D = T32 + T33;
T3E = T2J + T2I;
T3R = T3D - T3E;
T1u = TM + TP;
T1v = TR + TU;
T1w = T1u - T1v;
T20 = T1u + T1v;
T7 = T3 + T6;
Te = Ta + Td;
Tf = T7 + Te;
T3M = T7 - Te;
{
E T2H, T2K, TQ, TV;
T2H = T3 - T6;
T2K = T2I - T2J;
T2L = T2H + T2K;
T3h = T2H - T2K;
TQ = TM - TP;
TV = TR - TU;
TW = TQ + TV;
T2e = TQ - TV;
}
}
}
{
E Ti, T1e, T1c, T2N, T1h, T2O, Tl, T19, Tp, T13, T11, T2S, T16, T2T, Ts;
E TY, T2M, T2P;
{
E Tg, Th, T1a, T1b;
Tg = cr[WS(rs, 1)];
Th = ci[WS(rs, 6)];
Ti = Tg + Th;
T1e = Tg - Th;
T1a = ci[WS(rs, 14)];
T1b = cr[WS(rs, 9)];
T1c = T1a + T1b;
T2N = T1a - T1b;
}
{
E T1f, T1g, Tj, Tk;
T1f = ci[WS(rs, 10)];
T1g = cr[WS(rs, 13)];
T1h = T1f + T1g;
T2O = T1f - T1g;
Tj = cr[WS(rs, 5)];
Tk = ci[WS(rs, 2)];
Tl = Tj + Tk;
T19 = Tj - Tk;
}
{
E Tn, To, TZ, T10;
Tn = ci[0];
To = cr[WS(rs, 7)];
Tp = Tn + To;
T13 = Tn - To;
TZ = ci[WS(rs, 8)];
T10 = cr[WS(rs, 15)];
T11 = TZ + T10;
T2S = TZ - T10;
}
{
E T14, T15, Tq, Tr;
T14 = ci[WS(rs, 12)];
T15 = cr[WS(rs, 11)];
T16 = T14 + T15;
T2T = T14 - T15;
Tq = cr[WS(rs, 3)];
Tr = ci[WS(rs, 4)];
Ts = Tq + Tr;
TY = Tq - Tr;
}
T3G = T2N + T2O;
T3H = T2S + T2T;
T3N = T3H - T3G;
T2M = Ti - Tl;
T2P = T2N - T2O;
T2Q = T2M - T2P;
T36 = T2M + T2P;
{
E T2R, T2U, Tm, Tt;
T2R = Tp - Ts;
T2U = T2S - T2T;
T2V = T2R + T2U;
T37 = T2U - T2R;
Tm = Ti + Tl;
Tt = Tp + Ts;
Tu = Tm + Tt;
T3S = Tm - Tt;
}
{
E T12, T17, T22, T23;
T12 = TY - T11;
T17 = T13 - T16;
T18 = FNMS(KP414213562, T17, T12);
T1z = FMA(KP414213562, T12, T17);
T22 = T1c - T19;
T23 = T1e + T1h;
T24 = FNMS(KP414213562, T23, T22);
T2g = FMA(KP414213562, T22, T23);
}
{
E T25, T26, T1d, T1i;
T25 = TY + T11;
T26 = T13 + T16;
T27 = FNMS(KP414213562, T26, T25);
T2h = FMA(KP414213562, T25, T26);
T1d = T19 + T1c;
T1i = T1e - T1h;
T1j = FMA(KP414213562, T1i, T1d);
T1y = FNMS(KP414213562, T1d, T1i);
}
}
cr[0] = Tf + Tu;
{
E T3B, T3K, T3F, T3I, T3J, T3A;
T3A = Tf - Tu;
T3B = T3z * T3A;
T3K = T3C * T3A;
T3F = T3D + T3E;
T3I = T3G + T3H;
T3J = T3F - T3I;
ci[0] = T3F + T3I;
ci[WS(rs, 8)] = FMA(T3z, T3J, T3K);
cr[WS(rs, 8)] = FNMS(T3C, T3J, T3B);
}
{
E T3O, T3P, T3T, T3U;
T3O = T3M - T3N;
T3P = T3L * T3O;
T3T = T3R - T3S;
T3U = T3L * T3T;
cr[WS(rs, 12)] = FNMS(T3Q, T3T, T3P);
ci[WS(rs, 12)] = FMA(T3Q, T3O, T3U);
}
{
E T3V, T3W, T3X, T3Y;
T3V = T3M + T3N;
T3W = TA * T3V;
T3X = T3S + T3R;
T3Y = TA * T3X;
cr[WS(rs, 4)] = FNMS(TE, T3X, T3W);
ci[WS(rs, 4)] = FMA(TE, T3V, T3Y);
}
{
E T3j, T3t, T3p, T3x, T3i, T3o;
T3i = T37 - T36;
T3j = FNMS(KP707106781, T3i, T3h);
T3t = FMA(KP707106781, T3i, T3h);
T3o = T2Q - T2V;
T3p = FNMS(KP707106781, T3o, T3n);
T3x = FMA(KP707106781, T3o, T3n);
{
E T3k, T3q, T3u, T3y;
T3k = T3g * T3j;
cr[WS(rs, 14)] = FNMS(T3m, T3p, T3k);
T3q = T3g * T3p;
ci[WS(rs, 14)] = FMA(T3m, T3j, T3q);
T3u = T3s * T3t;
cr[WS(rs, 6)] = FNMS(T3w, T3x, T3u);
T3y = T3s * T3x;
ci[WS(rs, 6)] = FMA(T3w, T3t, T3y);
}
}
{
E T2X, T3b, T39, T3d, T2W, T35, T38;
T2W = T2Q + T2V;
T2X = FNMS(KP707106781, T2W, T2L);
T3b = FMA(KP707106781, T2W, T2L);
T35 = T31 + T34;
T38 = T36 + T37;
T39 = FNMS(KP707106781, T38, T35);
T3d = FMA(KP707106781, T38, T35);
{
E T2Y, T3a, T3c, T3e;
T2Y = T2G * T2X;
cr[WS(rs, 10)] = FNMS(T30, T39, T2Y);
T3a = T30 * T2X;
ci[WS(rs, 10)] = FMA(T2G, T39, T3a);
T3c = T1V * T3b;
cr[WS(rs, 2)] = FNMS(T1X, T3d, T3c);
T3e = T1X * T3b;
ci[WS(rs, 2)] = FMA(T1V, T3d, T3e);
}
}
{
E T29, T2l, T2j, T2n;
{
E T21, T28, T2f, T2i;
T21 = FNMS(KP707106781, T20, T1Z);
T28 = T24 + T27;
T29 = FMA(KP923879532, T28, T21);
T2l = FNMS(KP923879532, T28, T21);
T2f = FMA(KP707106781, T2e, T2d);
T2i = T2g - T2h;
T2j = FNMS(KP923879532, T2i, T2f);
T2n = FMA(KP923879532, T2i, T2f);
}
{
E T2a, T2k, T2m, T2o;
T2a = T1Y * T29;
cr[WS(rs, 11)] = FNMS(T2c, T2j, T2a);
T2k = T2c * T29;
ci[WS(rs, 11)] = FMA(T1Y, T2j, T2k);
T2m = Tw * T2l;
cr[WS(rs, 3)] = FNMS(Tz, T2n, T2m);
T2o = Tz * T2l;
ci[WS(rs, 3)] = FMA(Tw, T2n, T2o);
}
}
{
E T1l, T1E, T1B, T1H;
{
E TX, T1k, T1x, T1A;
TX = FNMS(KP707106781, TW, TL);
T1k = T18 - T1j;
T1l = FNMS(KP923879532, T1k, TX);
T1E = FMA(KP923879532, T1k, TX);
T1x = FNMS(KP707106781, T1w, T1t);
T1A = T1y - T1z;
T1B = FNMS(KP923879532, T1A, T1x);
T1H = FMA(KP923879532, T1A, T1x);
}
{
E T1m, T1C, T1F, T1I;
T1m = TG * T1l;
cr[WS(rs, 13)] = FNMS(T1o, T1B, T1m);
T1C = T1o * T1l;
ci[WS(rs, 13)] = FMA(TG, T1B, T1C);
T1F = T1D * T1E;
cr[WS(rs, 5)] = FNMS(T1G, T1H, T1F);
T1I = T1G * T1E;
ci[WS(rs, 5)] = FMA(T1D, T1H, T1I);
}
}
{
E T2s, T2A, T2x, T2D;
{
E T2q, T2r, T2v, T2w;
T2q = FMA(KP707106781, T20, T1Z);
T2r = T2g + T2h;
T2s = FNMS(KP923879532, T2r, T2q);
T2A = FMA(KP923879532, T2r, T2q);
T2v = FNMS(KP707106781, T2e, T2d);
T2w = T27 - T24;
T2x = FMA(KP923879532, T2w, T2v);
T2D = FNMS(KP923879532, T2w, T2v);
}
{
E T2t, T2y, T2B, T2E;
T2t = T2p * T2s;
cr[WS(rs, 7)] = FNMS(T2u, T2x, T2t);
T2y = T2p * T2x;
ci[WS(rs, 7)] = FMA(T2u, T2s, T2y);
T2B = T2z * T2A;
cr[WS(rs, 15)] = FNMS(T2C, T2D, T2B);
T2E = T2z * T2D;
ci[WS(rs, 15)] = FMA(T2C, T2A, T2E);
}
}
{
E T1L, T1R, T1P, T1T;
{
E T1J, T1K, T1N, T1O;
T1J = FMA(KP707106781, TW, TL);
T1K = T1y + T1z;
T1L = FNMS(KP923879532, T1K, T1J);
T1R = FMA(KP923879532, T1K, T1J);
T1N = FMA(KP707106781, T1w, T1t);
T1O = T1j + T18;
T1P = FNMS(KP923879532, T1O, T1N);
T1T = FMA(KP923879532, T1O, T1N);
}
{
E T1M, T1Q, T1S, T1U;
T1M = TB * T1L;
cr[WS(rs, 9)] = FNMS(TF, T1P, T1M);
T1Q = TB * T1P;
ci[WS(rs, 9)] = FMA(TF, T1L, T1Q);
T1S = Tv * T1R;
cr[WS(rs, 1)] = FNMS(Ty, T1T, T1S);
T1U = Tv * T1T;
ci[WS(rs, 1)] = FMA(Ty, T1R, T1U);
}
}
}
}
}
}
static void hc2cfdft_4(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms)
{
DK(KP500000000, +0.500000000000000000000000000000000000000000000);
{
INT m;
for (m = mb, W = W + ((mb - 1) * 6); m < me; m = m + 1, Rp = Rp + ms, Ip = Ip + ms, Rm = Rm - ms, Im = Im - ms, W = W + 6, MAKE_VOLATILE_STRIDE(16, rs)) {
E Td, Tu, Tr, T4, Tm, To, T9, T5, TA, Tp, Tv, TD, T6, Tq;
{
E Tk, Tl, Tf, TC, Tj, T7, T8, T1, Tn, Tb, Tc;
Tb = Ip[0];
Tc = Im[0];
{
E Ti, Tg, Th, T2, T3;
Tg = Rm[0];
Th = Rp[0];
Tk = W[1];
Tl = Tb + Tc;
Td = Tb - Tc;
Tu = Th + Tg;
Ti = Tg - Th;
Tf = W[0];
T2 = Ip[WS(rs, 1)];
T3 = Im[WS(rs, 1)];
TC = Tk * Ti;
Tj = Tf * Ti;
T7 = Rp[WS(rs, 1)];
Tr = T2 + T3;
T4 = T2 - T3;
T8 = Rm[WS(rs, 1)];
T1 = W[2];
Tn = W[4];
}
Tm = FNMS(Tk, Tl, Tj);
To = T7 - T8;
T9 = T7 + T8;
T5 = T1 * T4;
TA = Tn * Tr;
Tp = Tn * To;
Tv = T1 * T9;
TD = FMA(Tf, Tl, TC);
T6 = W[3];
Tq = W[5];
}
{
E Tw, Ta, TB, Ts;
Tw = FMA(T6, T4, Tv);
Ta = FNMS(T6, T9, T5);
TB = FNMS(Tq, To, TA);
Ts = FMA(Tq, Tr, Tp);
{
E TF, Tx, Te, Tz;
TF = Tu + Tw;
Tx = Tu - Tw;
Te = Ta + Td;
Tz = Td - Ta;
{
E TG, TE, Tt, Ty;
TG = TB + TD;
TE = TB - TD;
Tt = Tm - Ts;
Ty = Ts + Tm;
Im[0] = KP500000000 * (TE - Tz);
Ip[WS(rs, 1)] = KP500000000 * (Tz + TE);
Rp[0] = KP500000000 * (TF + TG);
Rm[WS(rs, 1)] = KP500000000 * (TF - TG);
Rp[WS(rs, 1)] = KP500000000 * (Tx + Ty);
Rm[0] = KP500000000 * (Tx - Ty);
Im[WS(rs, 1)] = KP500000000 * (Tt - Te);
Ip[0] = KP500000000 * (Te + Tt);
}
}
}
}
}
}
static void r2hc_11(const R *I, R *ro, R *io, stride is, stride ros, stride ios, INT v, INT ivs, INT ovs)
{
DK(KP959492973, +0.959492973614497389890368057066327699062454848);
DK(KP876768831, +0.876768831002589333891339807079336796764054852);
DK(KP918985947, +0.918985947228994779780736114132655398124909697);
DK(KP989821441, +0.989821441880932732376092037776718787376519372);
DK(KP778434453, +0.778434453334651800608337670740821884709317477);
DK(KP830830026, +0.830830026003772851058548298459246407048009821);
DK(KP715370323, +0.715370323453429719112414662767260662417897278);
DK(KP634356270, +0.634356270682424498893150776899916060542806975);
DK(KP342584725, +0.342584725681637509502641509861112333758894680);
DK(KP521108558, +0.521108558113202722944698153526659300680427422);
INT i;
for (i = v; i > 0; i = i - 1, I = I + ivs, ro = ro + ovs, io = io + ovs, MAKE_VOLATILE_STRIDE(is), MAKE_VOLATILE_STRIDE(ros), MAKE_VOLATILE_STRIDE(ios)) {
E T1, Tg, TF, TB, TI, TL, Tz, TA;
{
E T4, TC, TE, T7, TD, Ta, TS, TG, TJ, Td, TP, TM, Ty, Tq, Th;
E Tt, Tl;
T1 = I[0];
{
E Tb, Tc, Tx, Tp;
{
E T2, T3, Te, Tf;
T2 = I[WS(is, 1)];
T3 = I[WS(is, 10)];
Te = I[WS(is, 5)];
Tf = I[WS(is, 6)];
{
E T5, T6, T8, T9;
T5 = I[WS(is, 2)];
T4 = T2 + T3;
TC = T3 - T2;
Tg = Te + Tf;
TE = Tf - Te;
T6 = I[WS(is, 9)];
T8 = I[WS(is, 3)];
T9 = I[WS(is, 8)];
Tb = I[WS(is, 4)];
T7 = T5 + T6;
TD = T5 - T6;
Ta = T8 + T9;
TF = T9 - T8;
Tc = I[WS(is, 7)];
}
}
TS = FMA(KP521108558, TC, TD);
TG = FMA(KP521108558, TF, TE);
TJ = FMA(KP521108558, TE, TC);
Td = Tb + Tc;
TB = Tb - Tc;
Tx = FNMS(KP342584725, Ta, T7);
Tp = FNMS(KP342584725, T4, Ta);
TP = FNMS(KP521108558, TB, TF);
TM = FNMS(KP521108558, TD, TB);
Ty = FNMS(KP634356270, Tx, Td);
Tq = FNMS(KP634356270, Tp, Tg);
Th = FNMS(KP342584725, Tg, Td);
Tt = FNMS(KP342584725, Td, T4);
Tl = FNMS(KP342584725, T7, Tg);
}
{
E Tu, Ts, TN, Tv;
{
E Tm, TU, Tj, Ti, TT;
TT = FMA(KP715370323, TS, TF);
Ti = FNMS(KP634356270, Th, Ta);
Tu = FNMS(KP634356270, Tt, T7);
Tm = FNMS(KP634356270, Tl, T4);
TU = FMA(KP830830026, TT, TB);
Tj = FNMS(KP778434453, Ti, T7);
{
E Tk, TR, To, Tn, TQ, Tr;
TQ = FMA(KP715370323, TP, TC);
Tn = FNMS(KP778434453, Tm, Ta);
io[WS(ios, 5)] = KP989821441 * (FMA(KP918985947, TU, TE));
Tk = FNMS(KP876768831, Tj, T4);
TR = FNMS(KP830830026, TQ, TE);
To = FNMS(KP876768831, Tn, Td);
Tr = FNMS(KP778434453, Tq, Td);
ro[WS(ros, 5)] = FNMS(KP959492973, Tk, T1);
io[WS(ios, 4)] = KP989821441 * (FNMS(KP918985947, TR, TD));
ro[WS(ros, 4)] = FNMS(KP959492973, To, T1);
Ts = FNMS(KP876768831, Tr, T7);
}
}
TN = FNMS(KP715370323, TM, TE);
Tv = FNMS(KP778434453, Tu, Tg);
ro[0] = T1 + T4 + T7 + Ta + Td + Tg;
ro[WS(ros, 3)] = FNMS(KP959492973, Ts, T1);
{
E TO, Tw, TH, TK;
TO = FNMS(KP830830026, TN, TF);
Tw = FNMS(KP876768831, Tv, Ta);
TH = FMA(KP715370323, TG, TD);
TK = FNMS(KP715370323, TJ, TB);
io[WS(ios, 3)] = KP989821441 * (FNMS(KP918985947, TO, TC));
ro[WS(ros, 2)] = FNMS(KP959492973, Tw, T1);
TI = FNMS(KP830830026, TH, TC);
TL = FMA(KP830830026, TK, TD);
Tz = FNMS(KP778434453, Ty, T4);
}
}
}
io[WS(ios, 2)] = KP989821441 * (FMA(KP918985947, TI, TB));
io[WS(ios, 1)] = KP989821441 * (FNMS(KP918985947, TL, TF));
TA = FNMS(KP876768831, Tz, Tg);
ro[WS(ros, 1)] = FNMS(KP959492973, TA, T1);
}
}
static void hf_25(R *cr, R *ci, const R *W, stride rs, INT mb, INT me, INT ms)
{
DK(KP949179823, +0.949179823508441261575555465843363271711583843);
DK(KP860541664, +0.860541664367944677098261680920518816412804187);
DK(KP621716863, +0.621716863012209892444754556304102309693593202);
DK(KP614372930, +0.614372930789563808870829930444362096004872855);
DK(KP557913902, +0.557913902031834264187699648465567037992437152);
DK(KP249506682, +0.249506682107067890488084201715862638334226305);
DK(KP560319534, +0.560319534973832390111614715371676131169633784);
DK(KP681693190, +0.681693190061530575150324149145440022633095390);
DK(KP906616052, +0.906616052148196230441134447086066874408359177);
DK(KP968479752, +0.968479752739016373193524836781420152702090879);
DK(KP845997307, +0.845997307939530944175097360758058292389769300);
DK(KP998026728, +0.998026728428271561952336806863450553336905220);
DK(KP994076283, +0.994076283785401014123185814696322018529298887);
DK(KP734762448, +0.734762448793050413546343770063151342619912334);
DK(KP772036680, +0.772036680810363904029489473607579825330539880);
DK(KP062914667, +0.062914667253649757225485955897349402364686947);
DK(KP833417178, +0.833417178328688677408962550243238843138996060);
DK(KP921177326, +0.921177326965143320250447435415066029359282231);
DK(KP541454447, +0.541454447536312777046285590082819509052033189);
DK(KP803003575, +0.803003575438660414833440593570376004635464850);
DK(KP943557151, +0.943557151597354104399655195398983005179443399);
DK(KP554608978, +0.554608978404018097464974850792216217022558774);
DK(KP242145790, +0.242145790282157779872542093866183953459003101);
DK(KP559154169, +0.559154169276087864842202529084232643714075927);
DK(KP683113946, +0.683113946453479238701949862233725244439656928);
DK(KP248028675, +0.248028675328619457762448260696444630363259177);
DK(KP968583161, +0.968583161128631119490168375464735813836012403);
DK(KP525970792, +0.525970792408939708442463226536226366643874659);
DK(KP726211448, +0.726211448929902658173535992263577167607493062);
DK(KP904730450, +0.904730450839922351881287709692877908104763647);
DK(KP831864738, +0.831864738706457140726048799369896829771167132);
DK(KP871714437, +0.871714437527667770979999223229522602943903653);
DK(KP549754652, +0.549754652192770074288023275540779861653779767);
DK(KP992114701, +0.992114701314477831049793042785778521453036709);
DK(KP939062505, +0.939062505817492352556001843133229685779824606);
DK(KP256756360, +0.256756360367726783319498520922669048172391148);
DK(KP851038619, +0.851038619207379630836264138867114231259902550);
DK(KP912575812, +0.912575812670962425556968549836277086778922727);
DK(KP912018591, +0.912018591466481957908415381764119056233607330);
DK(KP634619297, +0.634619297544148100711287640319130485732531031);
DK(KP470564281, +0.470564281212251493087595091036643380879947982);
DK(KP827271945, +0.827271945972475634034355757144307982555673741);
DK(KP126329378, +0.126329378446108174786050455341811215027378105);
DK(KP951056516, +0.951056516295153572116439333379382143405698634);
DK(KP559016994, +0.559016994374947424102293417182819058860154590);
DK(KP250000000, +0.250000000000000000000000000000000000000000000);
DK(KP618033988, +0.618033988749894848204586834365638117720309180);
{
INT m;
for (m = mb, W = W + ((mb - 1) * 48); m < me; m = m + 1, cr = cr + ms, ci = ci - ms, W = W + 48, MAKE_VOLATILE_STRIDE(50, rs)) {
E T7i, T6o, T6m, T7o, T7m, T7h, T6n, T6f, T7j, T7n;
{
E T6W, T5G, T3Y, T3M, T7q, T70, T6V, T7P, Tt, T3L, T5T, T45, T5Q, T4c, T3G;
E T2G, T5P, T49, T5S, T42, T65, T4H, T68, T4A, T2Z, T11, T67, T4x, T64, T4E;
E T5Y, T4W, T61, T4P, T3d, T1z, T60, T4M, T5X, T4T, T3g, T1G, T3q, T4q, T4j;
E T26, T3i, T1M, T3k, T1S;
{
E T3u, T2e, T3E, T44, T4b, T2E, T3w, T2k, T3y, T2q;
{
E T1, T6R, T3P, T7, T3W, Tq, T9, Tc, Tb, T3U, Tk, T3Q, Ta;
{
E T3, T6, T2, T5;
T1 = cr[0];
T6R = ci[0];
T3 = cr[WS(rs, 5)];
T6 = ci[WS(rs, 5)];
T2 = W[8];
T5 = W[9];
{
E Tm, Tp, To, T3V, Tn, T3O, T4, Tl;
Tm = cr[WS(rs, 15)];
Tp = ci[WS(rs, 15)];
T3O = T2 * T6;
T4 = T2 * T3;
Tl = W[28];
To = W[29];
T3P = FNMS(T5, T3, T3O);
T7 = FMA(T5, T6, T4);
T3V = Tl * Tp;
Tn = Tl * Tm;
{
E Tg, Tj, Tf, Ti, T3T, Th, T8;
Tg = cr[WS(rs, 10)];
Tj = ci[WS(rs, 10)];
T3W = FNMS(To, Tm, T3V);
Tq = FMA(To, Tp, Tn);
Tf = W[18];
Ti = W[19];
T9 = cr[WS(rs, 20)];
Tc = ci[WS(rs, 20)];
T3T = Tf * Tj;
Th = Tf * Tg;
T8 = W[38];
Tb = W[39];
T3U = FNMS(Ti, Tg, T3T);
Tk = FMA(Ti, Tj, Th);
T3Q = T8 * Tc;
Ta = T8 * T9;
}
}
}
{
E T6T, T3X, T6Y, Tr, T3R, Td;
T6T = T3U + T3W;
T3X = T3U - T3W;
T6Y = Tk - Tq;
Tr = Tk + Tq;
T3R = FNMS(Tb, T9, T3Q);
Td = FMA(Tb, Tc, Ta);
{
E T3S, T6Z, Te, T6U, T6S, Ts;
T3S = T3P - T3R;
T6S = T3P + T3R;
T6Z = T7 - Td;
Te = T7 + Td;
T6W = T6S - T6T;
T6U = T6S + T6T;
T5G = FNMS(KP618033988, T3S, T3X);
T3Y = FMA(KP618033988, T3X, T3S);
T3M = Te - Tr;
Ts = Te + Tr;
T7q = FMA(KP618033988, T6Y, T6Z);
T70 = FNMS(KP618033988, T6Z, T6Y);
T6V = FNMS(KP250000000, T6U, T6R);
T7P = T6U + T6R;
Tt = T1 + Ts;
T3L = FNMS(KP250000000, Ts, T1);
}
}
}
{
E T2g, T2j, T2m, T3v, T2h, T2p, T2l, T2i, T2o, T3x, T2n;
{
E T2a, T2d, T29, T2c;
T2a = cr[WS(rs, 3)];
T2d = ci[WS(rs, 3)];
T29 = W[4];
T2c = W[5];
{
E T2t, T2w, T2z, T3A, T2u, T2C, T2y, T2v, T2B, T3t, T2b, T2s, T2f;
T2t = cr[WS(rs, 13)];
T2w = ci[WS(rs, 13)];
T3t = T29 * T2d;
T2b = T29 * T2a;
T2s = W[24];
T2z = cr[WS(rs, 18)];
T3u = FNMS(T2c, T2a, T3t);
T2e = FMA(T2c, T2d, T2b);
T3A = T2s * T2w;
T2u = T2s * T2t;
T2C = ci[WS(rs, 18)];
T2y = W[34];
T2v = W[25];
T2B = W[35];
{
E T3B, T2x, T3D, T2D, T3C, T2A;
T2g = cr[WS(rs, 8)];
T3C = T2y * T2C;
T2A = T2y * T2z;
T3B = FNMS(T2v, T2t, T3A);
T2x = FMA(T2v, T2w, T2u);
T3D = FNMS(T2B, T2z, T3C);
T2D = FMA(T2B, T2C, T2A);
T2j = ci[WS(rs, 8)];
T2f = W[14];
T3E = T3B + T3D;
T44 = T3D - T3B;
T4b = T2x - T2D;
T2E = T2x + T2D;
}
T2m = cr[WS(rs, 23)];
T3v = T2f * T2j;
T2h = T2f * T2g;
T2p = ci[WS(rs, 23)];
T2l = W[44];
T2i = W[15];
T2o = W[45];
}
}
T3x = T2l * T2p;
T2n = T2l * T2m;
T3w = FNMS(T2i, T2g, T3v);
T2k = FMA(T2i, T2j, T2h);
T3y = FNMS(T2o, T2m, T3x);
T2q = FMA(T2o, T2p, T2n);
}
{
E T2N, Tz, T2X, T4G, T4z, TZ, T2P, TF, T2R, TL;
{
E TB, TE, TH, T2O, TC, TK, TG, TD, TJ, T2Q, TI;
{
E Tv, Ty, Tu, Tx;
{
E T48, T41, T47, T40, T43, T3z;
Tv = cr[WS(rs, 1)];
T43 = T3y - T3w;
T3z = T3w + T3y;
{
E T4a, T2r, T3F, T2F;
T4a = T2k - T2q;
T2r = T2k + T2q;
T5T = FNMS(KP618033988, T43, T44);
T45 = FMA(KP618033988, T44, T43);
T3F = T3z + T3E;
T48 = T3E - T3z;
T5Q = FNMS(KP618033988, T4a, T4b);
T4c = FMA(KP618033988, T4b, T4a);
T2F = T2r + T2E;
T41 = T2E - T2r;
T3G = T3u + T3F;
T47 = FNMS(KP250000000, T3F, T3u);
T2G = T2e + T2F;
T40 = FNMS(KP250000000, T2F, T2e);
Ty = ci[WS(rs, 1)];
}
T5P = FMA(KP559016994, T48, T47);
T49 = FNMS(KP559016994, T48, T47);
T5S = FMA(KP559016994, T41, T40);
T42 = FNMS(KP559016994, T41, T40);
Tu = W[0];
}
Tx = W[1];
{
E TO, TR, TU, T2T, TP, TX, TT, TQ, TW, T2M, Tw, TN, TA;
TO = cr[WS(rs, 11)];
TR = ci[WS(rs, 11)];
T2M = Tu * Ty;
Tw = Tu * Tv;
TN = W[20];
TU = cr[WS(rs, 16)];
T2N = FNMS(Tx, Tv, T2M);
Tz = FMA(Tx, Ty, Tw);
T2T = TN * TR;
TP = TN * TO;
TX = ci[WS(rs, 16)];
TT = W[30];
TQ = W[21];
TW = W[31];
{
E T2U, TS, T2W, TY, T2V, TV;
TB = cr[WS(rs, 6)];
T2V = TT * TX;
TV = TT * TU;
T2U = FNMS(TQ, TO, T2T);
TS = FMA(TQ, TR, TP);
T2W = FNMS(TW, TU, T2V);
TY = FMA(TW, TX, TV);
TE = ci[WS(rs, 6)];
TA = W[10];
T2X = T2U + T2W;
T4G = T2W - T2U;
T4z = TY - TS;
TZ = TS + TY;
}
TH = cr[WS(rs, 21)];
T2O = TA * TE;
TC = TA * TB;
TK = ci[WS(rs, 21)];
TG = W[40];
TD = W[11];
TJ = W[41];
}
}
T2Q = TG * TK;
TI = TG * TH;
T2P = FNMS(TD, TB, T2O);
TF = FMA(TD, TE, TC);
T2R = FNMS(TJ, TH, T2Q);
TL = FMA(TJ, TK, TI);
}
{
E T31, T17, T3b, T4V, T4O, T1x, T33, T1d, T35, T1j;
{
E T19, T1c, T1f, T32, T1a, T1i, T1e, T1b, T1h, T34, T1g;
{
E T13, T16, T12, T15;
{
E T4w, T4D, T4v, T4C, T4F, T2S;
T13 = cr[WS(rs, 4)];
T4F = T2P - T2R;
T2S = T2P + T2R;
{
E T4y, TM, T2Y, T10;
T4y = TL - TF;
TM = TF + TL;
T65 = FMA(KP618033988, T4F, T4G);
T4H = FNMS(KP618033988, T4G, T4F);
T2Y = T2S + T2X;
T4w = T2S - T2X;
T68 = FNMS(KP618033988, T4y, T4z);
T4A = FMA(KP618033988, T4z, T4y);
T10 = TM + TZ;
T4D = TM - TZ;
T2Z = T2N + T2Y;
T4v = FNMS(KP250000000, T2Y, T2N);
T11 = Tz + T10;
T4C = FNMS(KP250000000, T10, Tz);
T16 = ci[WS(rs, 4)];
}
T67 = FNMS(KP559016994, T4w, T4v);
T4x = FMA(KP559016994, T4w, T4v);
T64 = FNMS(KP559016994, T4D, T4C);
T4E = FMA(KP559016994, T4D, T4C);
T12 = W[6];
}
T15 = W[7];
{
E T1m, T1p, T1s, T37, T1n, T1v, T1r, T1o, T1u, T30, T14, T1l, T18;
T1m = cr[WS(rs, 14)];
T1p = ci[WS(rs, 14)];
T30 = T12 * T16;
T14 = T12 * T13;
T1l = W[26];
T1s = cr[WS(rs, 19)];
T31 = FNMS(T15, T13, T30);
T17 = FMA(T15, T16, T14);
T37 = T1l * T1p;
T1n = T1l * T1m;
T1v = ci[WS(rs, 19)];
T1r = W[36];
T1o = W[27];
T1u = W[37];
{
E T38, T1q, T3a, T1w, T39, T1t;
T19 = cr[WS(rs, 9)];
T39 = T1r * T1v;
T1t = T1r * T1s;
T38 = FNMS(T1o, T1m, T37);
T1q = FMA(T1o, T1p, T1n);
T3a = FNMS(T1u, T1s, T39);
T1w = FMA(T1u, T1v, T1t);
T1c = ci[WS(rs, 9)];
T18 = W[16];
T3b = T38 + T3a;
T4V = T3a - T38;
T4O = T1w - T1q;
T1x = T1q + T1w;
}
T1f = cr[WS(rs, 24)];
T32 = T18 * T1c;
T1a = T18 * T19;
T1i = ci[WS(rs, 24)];
T1e = W[46];
T1b = W[17];
T1h = W[47];
}
}
T34 = T1e * T1i;
T1g = T1e * T1f;
T33 = FNMS(T1b, T19, T32);
T1d = FMA(T1b, T1c, T1a);
T35 = FNMS(T1h, T1f, T34);
T1j = FMA(T1h, T1i, T1g);
}
{
E T1I, T1L, T1O, T3h, T1J, T1R, T1N, T1K, T1Q, T3j, T1P;
{
E T1C, T1F, T1B, T1E;
{
E T4L, T4S, T4K, T4R, T4U, T36;
T1C = cr[WS(rs, 2)];
T4U = T35 - T33;
T36 = T33 + T35;
{
E T4N, T1k, T3c, T1y;
T4N = T1j - T1d;
T1k = T1d + T1j;
T5Y = FNMS(KP618033988, T4U, T4V);
T4W = FMA(KP618033988, T4V, T4U);
T3c = T36 + T3b;
T4L = T3b - T36;
T61 = FNMS(KP618033988, T4N, T4O);
T4P = FMA(KP618033988, T4O, T4N);
T1y = T1k + T1x;
T4S = T1k - T1x;
T3d = T31 + T3c;
T4K = FNMS(KP250000000, T3c, T31);
T1z = T17 + T1y;
T4R = FNMS(KP250000000, T1y, T17);
T1F = ci[WS(rs, 2)];
}
T60 = FMA(KP559016994, T4L, T4K);
T4M = FNMS(KP559016994, T4L, T4K);
T5X = FNMS(KP559016994, T4S, T4R);
T4T = FMA(KP559016994, T4S, T4R);
T1B = W[2];
}
T1E = W[3];
{
E T1V, T1Y, T21, T3m, T1W, T24, T20, T1X, T23, T3f, T1D, T1U, T1H;
T1V = cr[WS(rs, 12)];
T1Y = ci[WS(rs, 12)];
T3f = T1B * T1F;
T1D = T1B * T1C;
T1U = W[22];
T21 = cr[WS(rs, 17)];
T3g = FNMS(T1E, T1C, T3f);
T1G = FMA(T1E, T1F, T1D);
T3m = T1U * T1Y;
T1W = T1U * T1V;
T24 = ci[WS(rs, 17)];
T20 = W[32];
T1X = W[23];
T23 = W[33];
{
E T3n, T1Z, T3p, T25, T3o, T22;
T1I = cr[WS(rs, 7)];
T3o = T20 * T24;
T22 = T20 * T21;
T3n = FNMS(T1X, T1V, T3m);
T1Z = FMA(T1X, T1Y, T1W);
T3p = FNMS(T23, T21, T3o);
T25 = FMA(T23, T24, T22);
T1L = ci[WS(rs, 7)];
T1H = W[12];
T3q = T3n + T3p;
T4q = T3n - T3p;
T4j = T25 - T1Z;
T26 = T1Z + T25;
}
T1O = cr[WS(rs, 22)];
T3h = T1H * T1L;
T1J = T1H * T1I;
T1R = ci[WS(rs, 22)];
T1N = W[42];
T1K = W[13];
T1Q = W[43];
}
}
T3j = T1N * T1R;
T1P = T1N * T1O;
T3i = FNMS(T1K, T1I, T3h);
T1M = FMA(T1K, T1L, T1J);
T3k = FNMS(T1Q, T1O, T3j);
T1S = FMA(T1Q, T1R, T1P);
}
}
}
}
{
E T7Q, T5M, T5J, T7R, T5I, T5L, T7X, T7W, T5F, T6X, T5u, T7M, T7O, T5C, T5E;
E T5t, T7J, T7N;
{
E T4r, T4k, T4h, T4o, T3K, T3I, T1A, T2H, T28;
{
E T3e, T4g, T4n, T4f, T4m, T3H, T4p, T3l;
T7Q = T2Z + T3d;
T3e = T2Z - T3d;
T4p = T3k - T3i;
T3l = T3i + T3k;
{
E T4i, T1T, T3r, T27, T3s;
T4i = T1S - T1M;
T1T = T1M + T1S;
T5M = FMA(KP618033988, T4p, T4q);
T4r = FNMS(KP618033988, T4q, T4p);
T3r = T3l + T3q;
T4g = T3q - T3l;
T5J = FNMS(KP618033988, T4i, T4j);
T4k = FMA(KP618033988, T4j, T4i);
T27 = T1T + T26;
T4n = T26 - T1T;
T3s = T3g + T3r;
T4f = FNMS(KP250000000, T3r, T3g);
T28 = T1G + T27;
T4m = FNMS(KP250000000, T27, T1G);
T3H = T3s - T3G;
T7R = T3s + T3G;
}
T5I = FMA(KP559016994, T4g, T4f);
T4h = FNMS(KP559016994, T4g, T4f);
T5L = FMA(KP559016994, T4n, T4m);
T4o = FNMS(KP559016994, T4n, T4m);
T3K = FNMS(KP618033988, T3e, T3H);
T3I = FMA(KP618033988, T3H, T3e);
}
T1A = T11 + T1z;
T7X = T1z - T11;
T7W = T28 - T2G;
T2H = T28 + T2G;
{
E T3Z, T5d, T7r, T7D, T5h, T5i, T5m, T5l, T59, T7K, T56, T7L, T7I, T7G, T52;
E T50, T5w, T5g, T5q, T5A, T3N, T7p;
T3N = FMA(KP559016994, T3M, T3L);
T5F = FNMS(KP559016994, T3M, T3L);
T6X = FNMS(KP559016994, T6W, T6V);
T7p = FMA(KP559016994, T6W, T6V);
{
E T5o, T5p, T57, T4e, T4Y, T55, T4l, T4s, T4B, T5f, T5e, T4I;
{
E T46, T2K, T2J, T4d, T2I;
T46 = FMA(KP951056516, T45, T42);
T5o = FNMS(KP951056516, T45, T42);
T2I = T1A + T2H;
T2K = T1A - T2H;
T3Z = FNMS(KP951056516, T3Y, T3N);
T5d = FMA(KP951056516, T3Y, T3N);
T7r = FNMS(KP951056516, T7q, T7p);
T7D = FMA(KP951056516, T7q, T7p);
cr[0] = Tt + T2I;
T2J = FNMS(KP250000000, T2I, Tt);
T5p = FNMS(KP951056516, T4c, T49);
T4d = FMA(KP951056516, T4c, T49);
{
E T4Q, T4X, T2L, T3J;
T4Q = FNMS(KP951056516, T4P, T4M);
T5h = FMA(KP951056516, T4P, T4M);
T5i = FNMS(KP951056516, T4W, T4T);
T4X = FMA(KP951056516, T4W, T4T);
T2L = FMA(KP559016994, T2K, T2J);
T3J = FNMS(KP559016994, T2K, T2J);
T57 = FMA(KP126329378, T46, T4d);
T4e = FNMS(KP126329378, T4d, T46);
cr[WS(rs, 5)] = FMA(KP951056516, T3I, T2L);
ci[WS(rs, 4)] = FNMS(KP951056516, T3I, T2L);
ci[WS(rs, 9)] = FMA(KP951056516, T3K, T3J);
cr[WS(rs, 10)] = FNMS(KP951056516, T3K, T3J);
T4Y = FMA(KP827271945, T4X, T4Q);
T55 = FNMS(KP827271945, T4Q, T4X);
}
}
T4l = FNMS(KP951056516, T4k, T4h);
T5m = FMA(KP951056516, T4k, T4h);
T5l = FNMS(KP951056516, T4r, T4o);
T4s = FMA(KP951056516, T4r, T4o);
T4B = FNMS(KP951056516, T4A, T4x);
T5f = FMA(KP951056516, T4A, T4x);
T5e = FMA(KP951056516, T4H, T4E);
T4I = FNMS(KP951056516, T4H, T4E);
{
E T4u, T4Z, T4t, T58;
T4t = FNMS(KP470564281, T4s, T4l);
T58 = FMA(KP470564281, T4l, T4s);
{
E T4J, T54, T7E, T7F;
T4J = FMA(KP634619297, T4I, T4B);
T54 = FNMS(KP634619297, T4B, T4I);
T59 = FNMS(KP912018591, T58, T57);
T7E = FMA(KP912018591, T58, T57);
T7K = FMA(KP912018591, T4t, T4e);
T4u = FNMS(KP912018591, T4t, T4e);
T56 = FMA(KP912575812, T55, T54);
T7F = FNMS(KP912575812, T55, T54);
T7L = FMA(KP912575812, T4Y, T4J);
T4Z = FNMS(KP912575812, T4Y, T4J);
T7I = FNMS(KP851038619, T7F, T7E);
T7G = FMA(KP851038619, T7F, T7E);
}
T52 = FMA(KP851038619, T4Z, T4u);
T50 = FNMS(KP851038619, T4Z, T4u);
}
T5w = FNMS(KP256756360, T5e, T5f);
T5g = FMA(KP256756360, T5f, T5e);
T5q = FMA(KP939062505, T5p, T5o);
T5A = FNMS(KP939062505, T5o, T5p);
}
{
E T5y, T7z, T5B, T7y, T7w, T7u, T5s;
{
E T5k, T5r, T5j, T5x;
cr[WS(rs, 4)] = FNMS(KP992114701, T50, T3Z);
T5j = FMA(KP634619297, T5i, T5h);
T5x = FNMS(KP634619297, T5h, T5i);
{
E T5n, T5z, T7s, T7t;
T5n = FMA(KP549754652, T5m, T5l);
T5z = FNMS(KP549754652, T5l, T5m);
T5y = FMA(KP871714437, T5x, T5w);
T7s = FNMS(KP871714437, T5x, T5w);
T7z = FNMS(KP871714437, T5j, T5g);
T5k = FMA(KP871714437, T5j, T5g);
T5B = FNMS(KP831864738, T5A, T5z);
T7t = FMA(KP831864738, T5A, T5z);
T7y = FNMS(KP831864738, T5q, T5n);
T5r = FMA(KP831864738, T5q, T5n);
T7w = FNMS(KP904730450, T7t, T7s);
T7u = FMA(KP904730450, T7t, T7s);
}
ci[WS(rs, 20)] = FNMS(KP992114701, T7G, T7D);
T5u = FNMS(KP904730450, T5r, T5k);
T5s = FMA(KP904730450, T5r, T5k);
}
{
E T5a, T5c, T7A, T7C, T7v, T53, T5b, T51, T7H, T7x, T7B;
T5a = FNMS(KP726211448, T59, T56);
T5c = FMA(KP525970792, T56, T59);
ci[WS(rs, 23)] = FMA(KP968583161, T7u, T7r);
cr[WS(rs, 1)] = FMA(KP968583161, T5s, T5d);
T51 = FMA(KP248028675, T50, T3Z);
T7A = FNMS(KP683113946, T7z, T7y);
T7C = FMA(KP559154169, T7y, T7z);
T7v = FNMS(KP242145790, T7u, T7r);
T53 = FMA(KP554608978, T52, T51);
T5b = FNMS(KP554608978, T52, T51);
T7M = FNMS(KP525970792, T7L, T7K);
T7O = FMA(KP726211448, T7K, T7L);
ci[WS(rs, 10)] = FNMS(KP943557151, T5c, T5b);
ci[WS(rs, 5)] = FMA(KP943557151, T5c, T5b);
ci[0] = FMA(KP803003575, T5a, T53);
cr[WS(rs, 9)] = FNMS(KP803003575, T5a, T53);
T7x = FNMS(KP541454447, T7w, T7v);
T7B = FMA(KP541454447, T7w, T7v);
T7H = FMA(KP248028675, T7G, T7D);
cr[WS(rs, 21)] = -(FMA(KP921177326, T7C, T7B));
ci[WS(rs, 18)] = FNMS(KP921177326, T7C, T7B);
ci[WS(rs, 13)] = FMA(KP833417178, T7A, T7x);
cr[WS(rs, 16)] = FMS(KP833417178, T7A, T7x);
T5C = FMA(KP559154169, T5B, T5y);
T5E = FNMS(KP683113946, T5y, T5B);
T5t = FNMS(KP242145790, T5s, T5d);
T7J = FNMS(KP554608978, T7I, T7H);
T7N = FMA(KP554608978, T7I, T7H);
}
}
}
}
{
E T7Y, T80, T5v, T5D;
cr[WS(rs, 24)] = -(FMA(KP803003575, T7O, T7N));
ci[WS(rs, 15)] = FNMS(KP803003575, T7O, T7N);
cr[WS(rs, 19)] = FMS(KP943557151, T7M, T7J);
cr[WS(rs, 14)] = -(FMA(KP943557151, T7M, T7J));
T5v = FMA(KP541454447, T5u, T5t);
T5D = FNMS(KP541454447, T5u, T5t);
cr[WS(rs, 11)] = FNMS(KP833417178, T5E, T5D);
ci[WS(rs, 8)] = FMA(KP833417178, T5E, T5D);
cr[WS(rs, 6)] = FMA(KP921177326, T5C, T5v);
ci[WS(rs, 3)] = FNMS(KP921177326, T5C, T5v);
T7Y = FMA(KP618033988, T7X, T7W);
T80 = FNMS(KP618033988, T7W, T7X);
{
E T6t, T6p, T5H, T7d, T71, T6u, T6y, T6x, T6l, T7k, T6i, T7l, T7g, T6c, T6e;
E T6s, T6L, T6J, T6C;
{
E T6A, T6B, T5O, T6j, T6h, T6a, T6q, T5R, T5U, T6r, T5Z, T62;
{
E T5K, T7U, T7T, T5N, T7S;
T6t = FNMS(KP951056516, T5J, T5I);
T5K = FMA(KP951056516, T5J, T5I);
T7U = T7Q - T7R;
T7S = T7Q + T7R;
T6p = FNMS(KP951056516, T5G, T5F);
T5H = FMA(KP951056516, T5G, T5F);
T7d = FNMS(KP951056516, T70, T6X);
T71 = FMA(KP951056516, T70, T6X);
ci[WS(rs, 24)] = T7S + T7P;
T7T = FNMS(KP250000000, T7S, T7P);
T5N = FMA(KP951056516, T5M, T5L);
T6u = FNMS(KP951056516, T5M, T5L);
{
E T66, T69, T7Z, T7V;
T6A = FMA(KP951056516, T65, T64);
T66 = FNMS(KP951056516, T65, T64);
T69 = FMA(KP951056516, T68, T67);
T6B = FNMS(KP951056516, T68, T67);
T7Z = FMA(KP559016994, T7U, T7T);
T7V = FNMS(KP559016994, T7U, T7T);
T5O = FMA(KP062914667, T5N, T5K);
T6j = FNMS(KP062914667, T5K, T5N);
ci[WS(rs, 14)] = FMA(KP951056516, T7Y, T7V);
cr[WS(rs, 15)] = FMS(KP951056516, T7Y, T7V);
ci[WS(rs, 19)] = FMA(KP951056516, T80, T7Z);
cr[WS(rs, 20)] = FMS(KP951056516, T80, T7Z);
T6h = FNMS(KP939062505, T66, T69);
T6a = FMA(KP939062505, T69, T66);
}
}
T6q = FMA(KP951056516, T5Q, T5P);
T5R = FNMS(KP951056516, T5Q, T5P);
T5U = FNMS(KP951056516, T5T, T5S);
T6r = FMA(KP951056516, T5T, T5S);
T6y = FMA(KP951056516, T5Y, T5X);
T5Z = FNMS(KP951056516, T5Y, T5X);
T62 = FMA(KP951056516, T61, T60);
T6x = FNMS(KP951056516, T61, T60);
{
E T5W, T6b, T6k, T5V;
T6k = FMA(KP827271945, T5R, T5U);
T5V = FNMS(KP827271945, T5U, T5R);
{
E T6g, T63, T7e, T7f;
T6g = FMA(KP126329378, T5Z, T62);
T63 = FNMS(KP126329378, T62, T5Z);
T7e = FMA(KP772036680, T6k, T6j);
T6l = FNMS(KP772036680, T6k, T6j);
T5W = FMA(KP772036680, T5V, T5O);
T7k = FNMS(KP772036680, T5V, T5O);
T7f = FNMS(KP734762448, T6h, T6g);
T6i = FMA(KP734762448, T6h, T6g);
T6b = FNMS(KP734762448, T6a, T63);
T7l = FMA(KP734762448, T6a, T63);
T7g = FMA(KP994076283, T7f, T7e);
T7i = FNMS(KP994076283, T7f, T7e);
}
T6c = FNMS(KP994076283, T6b, T5W);
T6e = FMA(KP994076283, T6b, T5W);
}
T6s = FMA(KP062914667, T6r, T6q);
T6L = FNMS(KP062914667, T6q, T6r);
T6J = FNMS(KP549754652, T6A, T6B);
T6C = FMA(KP549754652, T6B, T6A);
}
{
E T6N, T78, T6K, T79, T74, T76, T6E, T6G;
{
E T6w, T6D, T6M, T6v;
cr[WS(rs, 3)] = FMA(KP998026728, T6c, T5H);
T6M = FNMS(KP634619297, T6t, T6u);
T6v = FMA(KP634619297, T6u, T6t);
{
E T6I, T6z, T72, T73;
T6I = FMA(KP470564281, T6x, T6y);
T6z = FNMS(KP470564281, T6y, T6x);
T72 = FMA(KP845997307, T6M, T6L);
T6N = FNMS(KP845997307, T6M, T6L);
T6w = FMA(KP845997307, T6v, T6s);
T78 = FNMS(KP845997307, T6v, T6s);
T73 = FNMS(KP968479752, T6J, T6I);
T6K = FMA(KP968479752, T6J, T6I);
T6D = FMA(KP968479752, T6C, T6z);
T79 = FNMS(KP968479752, T6C, T6z);
T74 = FMA(KP906616052, T73, T72);
T76 = FNMS(KP906616052, T73, T72);
}
ci[WS(rs, 21)] = FNMS(KP998026728, T7g, T7d);
T6E = FMA(KP906616052, T6D, T6w);
T6G = FNMS(KP906616052, T6D, T6w);
}
{
E T7c, T7a, T6Q, T6O, T6F, T7b, T77, T75, T6d, T6P, T6H;
T7c = FMA(KP681693190, T78, T79);
T7a = FNMS(KP560319534, T79, T78);
ci[WS(rs, 22)] = FNMS(KP998026728, T74, T71);
cr[WS(rs, 2)] = FMA(KP998026728, T6E, T6p);
T75 = FMA(KP249506682, T74, T71);
T6Q = FNMS(KP560319534, T6K, T6N);
T6O = FMA(KP681693190, T6N, T6K);
T6F = FNMS(KP249506682, T6E, T6p);
T7b = FMA(KP557913902, T76, T75);
T77 = FNMS(KP557913902, T76, T75);
T6o = FMA(KP614372930, T6i, T6l);
T6m = FNMS(KP621716863, T6l, T6i);
cr[WS(rs, 22)] = FMS(KP860541664, T7c, T7b);
ci[WS(rs, 17)] = FMA(KP860541664, T7c, T7b);
ci[WS(rs, 12)] = FNMS(KP949179823, T7a, T77);
cr[WS(rs, 17)] = -(FMA(KP949179823, T7a, T77));
T6P = FMA(KP557913902, T6G, T6F);
T6H = FNMS(KP557913902, T6G, T6F);
T6d = FNMS(KP249506682, T6c, T5H);
ci[WS(rs, 7)] = FMA(KP949179823, T6Q, T6P);
cr[WS(rs, 12)] = FNMS(KP949179823, T6Q, T6P);
cr[WS(rs, 7)] = FMA(KP860541664, T6O, T6H);
ci[WS(rs, 2)] = FNMS(KP860541664, T6O, T6H);
T7o = FMA(KP621716863, T7k, T7l);
T7m = FNMS(KP614372930, T7l, T7k);
T7h = FMA(KP249506682, T7g, T7d);
T6n = FMA(KP557913902, T6e, T6d);
T6f = FNMS(KP557913902, T6e, T6d);
}
}
}
}
}
}
ci[WS(rs, 6)] = FNMS(KP949179823, T6o, T6n);
ci[WS(rs, 11)] = FMA(KP949179823, T6o, T6n);
cr[WS(rs, 8)] = FMA(KP943557151, T6m, T6f);
ci[WS(rs, 1)] = FNMS(KP943557151, T6m, T6f);
T7j = FNMS(KP557913902, T7i, T7h);
T7n = FMA(KP557913902, T7i, T7h);
cr[WS(rs, 23)] = -(FMA(KP943557151, T7o, T7n));
ci[WS(rs, 16)] = FNMS(KP943557151, T7o, T7n);
cr[WS(rs, 18)] = FMS(KP949179823, T7m, T7j);
cr[WS(rs, 13)] = -(FMA(KP949179823, T7m, T7j));
}
}
}
static void q1_3(float *rio, float *iio, const float *W, stride rs, stride vs, INT mb, INT me, INT ms)
{
DK(KP866025403, +0.866025403784438646763723170752936183471402627);
DK(KP500000000, +0.500000000000000000000000000000000000000000000);
INT m;
for (m = mb, W = W + (mb * 4); m < me; m = m + 1, rio = rio + ms, iio = iio + ms, W = W + 4, MAKE_VOLATILE_STRIDE(rs), MAKE_VOLATILE_STRIDE(vs)) {
E Tk, Tn, Tm, To, Tl;
{
E T1, Td, T4, Tg, Tp, T9, Te, T6, Tf, TB, TE, Ts, TZ, Tu, Tx;
E TC, TN, TO, TD, TV, T10, TP, Tq, Tr;
{
E T2, T3, T7, T8;
T1 = rio[0];
T2 = rio[WS(rs, 1)];
T3 = rio[WS(rs, 2)];
Td = iio[0];
T7 = iio[WS(rs, 1)];
T8 = iio[WS(rs, 2)];
T4 = T2 + T3;
Tg = T3 - T2;
Tp = rio[WS(vs, 1)];
T9 = T7 - T8;
Te = T7 + T8;
T6 = FNMS(KP500000000, T4, T1);
Tq = rio[WS(vs, 1) + WS(rs, 1)];
Tr = rio[WS(vs, 1) + WS(rs, 2)];
Tf = FNMS(KP500000000, Te, Td);
}
{
E Tv, Tw, TT, TU;
TB = iio[WS(vs, 1)];
Tv = iio[WS(vs, 1) + WS(rs, 1)];
TE = Tr - Tq;
Ts = Tq + Tr;
Tw = iio[WS(vs, 1) + WS(rs, 2)];
TZ = iio[WS(vs, 2)];
TT = iio[WS(vs, 2) + WS(rs, 1)];
Tu = FNMS(KP500000000, Ts, Tp);
Tx = Tv - Tw;
TC = Tv + Tw;
TU = iio[WS(vs, 2) + WS(rs, 2)];
TN = rio[WS(vs, 2)];
TO = rio[WS(vs, 2) + WS(rs, 1)];
TD = FNMS(KP500000000, TC, TB);
TV = TT - TU;
T10 = TT + TU;
TP = rio[WS(vs, 2) + WS(rs, 2)];
}
{
E T11, T12, TS, TQ;
rio[0] = T1 + T4;
iio[0] = Td + Te;
T11 = FNMS(KP500000000, T10, TZ);
T12 = TP - TO;
TQ = TO + TP;
rio[WS(rs, 1)] = Tp + Ts;
iio[WS(rs, 1)] = TB + TC;
iio[WS(rs, 2)] = TZ + T10;
TS = FNMS(KP500000000, TQ, TN);
rio[WS(rs, 2)] = TN + TQ;
{
E TW, T13, Ty, TI, TL, TF, TH, TK;
{
E Ta, Th, T5, Tc;
Tk = FNMS(KP866025403, T9, T6);
Ta = FMA(KP866025403, T9, T6);
Th = FMA(KP866025403, Tg, Tf);
Tn = FNMS(KP866025403, Tg, Tf);
T5 = W[0];
Tc = W[1];
{
E T16, T19, T18, T1a, T17, Ti, Tb, T15;
TW = FMA(KP866025403, TV, TS);
T16 = FNMS(KP866025403, TV, TS);
T19 = FNMS(KP866025403, T12, T11);
T13 = FMA(KP866025403, T12, T11);
Ti = T5 * Th;
Tb = T5 * Ta;
T15 = W[2];
T18 = W[3];
iio[WS(vs, 1)] = FNMS(Tc, Ta, Ti);
rio[WS(vs, 1)] = FMA(Tc, Th, Tb);
T1a = T15 * T19;
T17 = T15 * T16;
Ty = FMA(KP866025403, Tx, Tu);
TI = FNMS(KP866025403, Tx, Tu);
TL = FNMS(KP866025403, TE, TD);
TF = FMA(KP866025403, TE, TD);
iio[WS(vs, 2) + WS(rs, 2)] = FNMS(T18, T16, T1a);
rio[WS(vs, 2) + WS(rs, 2)] = FMA(T18, T19, T17);
TH = W[2];
TK = W[3];
}
}
{
E TA, TG, Tz, TM, TJ, Tt;
TM = TH * TL;
TJ = TH * TI;
Tt = W[0];
TA = W[1];
iio[WS(vs, 2) + WS(rs, 1)] = FNMS(TK, TI, TM);
rio[WS(vs, 2) + WS(rs, 1)] = FMA(TK, TL, TJ);
TG = Tt * TF;
Tz = Tt * Ty;
{
E TR, TY, T14, TX, Tj;
iio[WS(vs, 1) + WS(rs, 1)] = FNMS(TA, Ty, TG);
rio[WS(vs, 1) + WS(rs, 1)] = FMA(TA, TF, Tz);
TR = W[0];
TY = W[1];
T14 = TR * T13;
TX = TR * TW;
Tj = W[2];
Tm = W[3];
iio[WS(vs, 1) + WS(rs, 2)] = FNMS(TY, TW, T14);
rio[WS(vs, 1) + WS(rs, 2)] = FMA(TY, T13, TX);
To = Tj * Tn;
Tl = Tj * Tk;
}
}
}
}
}
iio[WS(vs, 2)] = FNMS(Tm, Tk, To);
rio[WS(vs, 2)] = FMA(Tm, Tn, Tl);
}
}
static void hf_25(R *cr, R *ci, const R *W, stride rs, INT mb, INT me, INT ms)
{
DK(KP998026728, +0.998026728428271561952336806863450553336905220);
DK(KP062790519, +0.062790519529313376076178224565631133122484832);
DK(KP684547105, +0.684547105928688673732283357621209269889519233);
DK(KP728968627, +0.728968627421411523146730319055259111372571664);
DK(KP481753674, +0.481753674101715274987191502872129653528542010);
DK(KP876306680, +0.876306680043863587308115903922062583399064238);
DK(KP248689887, +0.248689887164854788242283746006447968417567406);
DK(KP968583161, +0.968583161128631119490168375464735813836012403);
DK(KP992114701, +0.992114701314477831049793042785778521453036709);
DK(KP125333233, +0.125333233564304245373118759816508793942918247);
DK(KP425779291, +0.425779291565072648862502445744251703979973042);
DK(KP904827052, +0.904827052466019527713668647932697593970413911);
DK(KP637423989, +0.637423989748689710176712811676016195434917298);
DK(KP770513242, +0.770513242775789230803009636396177847271667672);
DK(KP844327925, +0.844327925502015078548558063966681505381659241);
DK(KP535826794, +0.535826794978996618271308767867639978063575346);
DK(KP587785252, +0.587785252292473129168705954639072768597652438);
DK(KP951056516, +0.951056516295153572116439333379382143405698634);
DK(KP250000000, +0.250000000000000000000000000000000000000000000);
DK(KP559016994, +0.559016994374947424102293417182819058860154590);
{
INT m;
for (m = mb, W = W + ((mb - 1) * 48); m < me; m = m + 1, cr = cr + ms, ci = ci - ms, W = W + 48, MAKE_VOLATILE_STRIDE(50, rs)) {
E T1, T6b, T2l, T6g, To, T2m, T6e, T6f, T6a, T6H, T2u, T4I, T2i, T60, T3S;
E T5D, T4r, T58, T3Z, T5C, T4q, T5b, TS, T5W, T2G, T5s, T4g, T4M, T2R, T5t;
E T4h, T4P, T1l, T5X, T37, T5v, T4k, T4T, T3e, T5w, T4j, T4W, T1P, T5Z, T3v;
E T5A, T4o, T54, T3C, T5z, T4n, T51;
{
E T6, T2o, Tb, T2p, Tc, T6c, Th, T2r, Tm, T2s, Tn, T6d;
T1 = cr[0];
T6b = ci[0];
{
E T3, T5, T2, T4;
T3 = cr[WS(rs, 5)];
T5 = ci[WS(rs, 5)];
T2 = W[8];
T4 = W[9];
T6 = FMA(T2, T3, T4 * T5);
T2o = FNMS(T4, T3, T2 * T5);
}
{
E T8, Ta, T7, T9;
T8 = cr[WS(rs, 20)];
Ta = ci[WS(rs, 20)];
T7 = W[38];
T9 = W[39];
Tb = FMA(T7, T8, T9 * Ta);
T2p = FNMS(T9, T8, T7 * Ta);
}
Tc = T6 + Tb;
T6c = T2o + T2p;
{
E Te, Tg, Td, Tf;
Te = cr[WS(rs, 10)];
Tg = ci[WS(rs, 10)];
Td = W[18];
Tf = W[19];
Th = FMA(Td, Te, Tf * Tg);
T2r = FNMS(Tf, Te, Td * Tg);
}
{
E Tj, Tl, Ti, Tk;
Tj = cr[WS(rs, 15)];
Tl = ci[WS(rs, 15)];
Ti = W[28];
Tk = W[29];
Tm = FMA(Ti, Tj, Tk * Tl);
T2s = FNMS(Tk, Tj, Ti * Tl);
}
Tn = Th + Tm;
T6d = T2r + T2s;
T2l = KP559016994 * (Tc - Tn);
T6g = KP559016994 * (T6c - T6d);
To = Tc + Tn;
T2m = FNMS(KP250000000, To, T1);
T6e = T6c + T6d;
T6f = FNMS(KP250000000, T6e, T6b);
{
E T68, T69, T2q, T2t;
T68 = Th - Tm;
T69 = T6 - Tb;
T6a = FNMS(KP587785252, T69, KP951056516 * T68);
T6H = FMA(KP951056516, T69, KP587785252 * T68);
T2q = T2o - T2p;
T2t = T2r - T2s;
T2u = FMA(KP951056516, T2q, KP587785252 * T2t);
T4I = FNMS(KP587785252, T2q, KP951056516 * T2t);
}
}
{
E T1U, T3O, T3E, T3F, T3X, T3W, T3J, T3M, T3P, T25, T2g, T2h;
{
E T1R, T1T, T1Q, T1S;
T1R = cr[WS(rs, 3)];
T1T = ci[WS(rs, 3)];
T1Q = W[4];
T1S = W[5];
T1U = FMA(T1Q, T1R, T1S * T1T);
T3O = FNMS(T1S, T1R, T1Q * T1T);
}
{
E T1Z, T3H, T2f, T3L, T24, T3I, T2a, T3K;
{
E T1W, T1Y, T1V, T1X;
T1W = cr[WS(rs, 8)];
T1Y = ci[WS(rs, 8)];
T1V = W[14];
T1X = W[15];
T1Z = FMA(T1V, T1W, T1X * T1Y);
T3H = FNMS(T1X, T1W, T1V * T1Y);
}
{
E T2c, T2e, T2b, T2d;
T2c = cr[WS(rs, 18)];
T2e = ci[WS(rs, 18)];
T2b = W[34];
T2d = W[35];
T2f = FMA(T2b, T2c, T2d * T2e);
T3L = FNMS(T2d, T2c, T2b * T2e);
}
{
E T21, T23, T20, T22;
T21 = cr[WS(rs, 23)];
T23 = ci[WS(rs, 23)];
T20 = W[44];
T22 = W[45];
T24 = FMA(T20, T21, T22 * T23);
T3I = FNMS(T22, T21, T20 * T23);
}
{
E T27, T29, T26, T28;
T27 = cr[WS(rs, 13)];
T29 = ci[WS(rs, 13)];
T26 = W[24];
T28 = W[25];
T2a = FMA(T26, T27, T28 * T29);
T3K = FNMS(T28, T27, T26 * T29);
}
T3E = T1Z - T24;
T3F = T2a - T2f;
T3X = T3K - T3L;
T3W = T3H - T3I;
T3J = T3H + T3I;
T3M = T3K + T3L;
T3P = T3J + T3M;
T25 = T1Z + T24;
T2g = T2a + T2f;
T2h = T25 + T2g;
}
T2i = T1U + T2h;
T60 = T3O + T3P;
{
E T3G, T57, T3R, T56, T3N, T3Q;
T3G = FMA(KP951056516, T3E, KP587785252 * T3F);
T57 = FNMS(KP587785252, T3E, KP951056516 * T3F);
T3N = KP559016994 * (T3J - T3M);
T3Q = FNMS(KP250000000, T3P, T3O);
T3R = T3N + T3Q;
T56 = T3Q - T3N;
T3S = T3G + T3R;
T5D = T57 + T56;
T4r = T3R - T3G;
T58 = T56 - T57;
}
{
E T3Y, T5a, T3V, T59, T3T, T3U;
T3Y = FMA(KP951056516, T3W, KP587785252 * T3X);
T5a = FNMS(KP587785252, T3W, KP951056516 * T3X);
T3T = KP559016994 * (T25 - T2g);
T3U = FNMS(KP250000000, T2h, T1U);
T3V = T3T + T3U;
T59 = T3U - T3T;
T3Z = T3V - T3Y;
T5C = T59 - T5a;
T4q = T3V + T3Y;
T5b = T59 + T5a;
}
}
{
E Tu, T2N, T2B, T2E, T2I, T2H, T2K, T2L, T2O, TF, TQ, TR;
{
E Tr, Tt, Tq, Ts;
Tr = cr[WS(rs, 1)];
Tt = ci[WS(rs, 1)];
Tq = W[0];
Ts = W[1];
Tu = FMA(Tq, Tr, Ts * Tt);
T2N = FNMS(Ts, Tr, Tq * Tt);
}
{
E Tz, T2z, TP, T2D, TE, T2A, TK, T2C;
{
E Tw, Ty, Tv, Tx;
Tw = cr[WS(rs, 6)];
Ty = ci[WS(rs, 6)];
Tv = W[10];
Tx = W[11];
Tz = FMA(Tv, Tw, Tx * Ty);
T2z = FNMS(Tx, Tw, Tv * Ty);
}
{
E TM, TO, TL, TN;
TM = cr[WS(rs, 16)];
TO = ci[WS(rs, 16)];
TL = W[30];
TN = W[31];
TP = FMA(TL, TM, TN * TO);
T2D = FNMS(TN, TM, TL * TO);
}
{
E TB, TD, TA, TC;
TB = cr[WS(rs, 21)];
TD = ci[WS(rs, 21)];
TA = W[40];
TC = W[41];
TE = FMA(TA, TB, TC * TD);
T2A = FNMS(TC, TB, TA * TD);
}
{
E TH, TJ, TG, TI;
TH = cr[WS(rs, 11)];
TJ = ci[WS(rs, 11)];
TG = W[20];
TI = W[21];
TK = FMA(TG, TH, TI * TJ);
T2C = FNMS(TI, TH, TG * TJ);
}
T2B = T2z - T2A;
T2E = T2C - T2D;
T2I = TK - TP;
T2H = Tz - TE;
T2K = T2z + T2A;
T2L = T2C + T2D;
T2O = T2K + T2L;
TF = Tz + TE;
TQ = TK + TP;
TR = TF + TQ;
}
TS = Tu + TR;
T5W = T2N + T2O;
{
E T2F, T4L, T2y, T4K, T2w, T2x;
T2F = FMA(KP951056516, T2B, KP587785252 * T2E);
T4L = FNMS(KP587785252, T2B, KP951056516 * T2E);
T2w = KP559016994 * (TF - TQ);
T2x = FNMS(KP250000000, TR, Tu);
T2y = T2w + T2x;
T4K = T2x - T2w;
T2G = T2y - T2F;
T5s = T4K - T4L;
T4g = T2y + T2F;
T4M = T4K + T4L;
}
{
E T2J, T4O, T2Q, T4N, T2M, T2P;
T2J = FMA(KP951056516, T2H, KP587785252 * T2I);
T4O = FNMS(KP587785252, T2H, KP951056516 * T2I);
T2M = KP559016994 * (T2K - T2L);
T2P = FNMS(KP250000000, T2O, T2N);
T2Q = T2M + T2P;
T4N = T2P - T2M;
T2R = T2J + T2Q;
T5t = T4O + T4N;
T4h = T2Q - T2J;
T4P = T4N - T4O;
}
}
{
E TX, T33, T2T, T2U, T3c, T3b, T2Y, T31, T34, T18, T1j, T1k;
{
E TU, TW, TT, TV;
TU = cr[WS(rs, 4)];
TW = ci[WS(rs, 4)];
TT = W[6];
TV = W[7];
TX = FMA(TT, TU, TV * TW);
T33 = FNMS(TV, TU, TT * TW);
}
{
E T12, T2W, T1i, T30, T17, T2X, T1d, T2Z;
{
E TZ, T11, TY, T10;
TZ = cr[WS(rs, 9)];
T11 = ci[WS(rs, 9)];
TY = W[16];
T10 = W[17];
T12 = FMA(TY, TZ, T10 * T11);
T2W = FNMS(T10, TZ, TY * T11);
}
{
E T1f, T1h, T1e, T1g;
T1f = cr[WS(rs, 19)];
T1h = ci[WS(rs, 19)];
T1e = W[36];
T1g = W[37];
T1i = FMA(T1e, T1f, T1g * T1h);
T30 = FNMS(T1g, T1f, T1e * T1h);
}
{
E T14, T16, T13, T15;
T14 = cr[WS(rs, 24)];
T16 = ci[WS(rs, 24)];
T13 = W[46];
T15 = W[47];
T17 = FMA(T13, T14, T15 * T16);
T2X = FNMS(T15, T14, T13 * T16);
}
{
E T1a, T1c, T19, T1b;
T1a = cr[WS(rs, 14)];
T1c = ci[WS(rs, 14)];
T19 = W[26];
T1b = W[27];
T1d = FMA(T19, T1a, T1b * T1c);
T2Z = FNMS(T1b, T1a, T19 * T1c);
}
T2T = T17 - T12;
T2U = T1d - T1i;
T3c = T2Z - T30;
T3b = T2W - T2X;
T2Y = T2W + T2X;
T31 = T2Z + T30;
T34 = T2Y + T31;
T18 = T12 + T17;
T1j = T1d + T1i;
T1k = T18 + T1j;
}
T1l = TX + T1k;
T5X = T33 + T34;
{
E T2V, T4S, T36, T4R, T32, T35;
T2V = FNMS(KP587785252, T2U, KP951056516 * T2T);
T4S = FMA(KP587785252, T2T, KP951056516 * T2U);
T32 = KP559016994 * (T2Y - T31);
T35 = FNMS(KP250000000, T34, T33);
T36 = T32 + T35;
T4R = T35 - T32;
T37 = T2V - T36;
T5v = T4S + T4R;
T4k = T2V + T36;
T4T = T4R - T4S;
}
{
E T3d, T4V, T3a, T4U, T38, T39;
T3d = FMA(KP951056516, T3b, KP587785252 * T3c);
T4V = FNMS(KP587785252, T3b, KP951056516 * T3c);
T38 = KP559016994 * (T18 - T1j);
T39 = FNMS(KP250000000, T1k, TX);
T3a = T38 + T39;
T4U = T39 - T38;
T3e = T3a - T3d;
T5w = T4U - T4V;
T4j = T3a + T3d;
T4W = T4U + T4V;
}
}
{
E T1r, T3r, T3h, T3i, T3A, T3z, T3m, T3p, T3s, T1C, T1N, T1O;
{
E T1o, T1q, T1n, T1p;
T1o = cr[WS(rs, 2)];
T1q = ci[WS(rs, 2)];
T1n = W[2];
T1p = W[3];
T1r = FMA(T1n, T1o, T1p * T1q);
T3r = FNMS(T1p, T1o, T1n * T1q);
}
{
E T1w, T3k, T1M, T3o, T1B, T3l, T1H, T3n;
{
E T1t, T1v, T1s, T1u;
T1t = cr[WS(rs, 7)];
T1v = ci[WS(rs, 7)];
T1s = W[12];
T1u = W[13];
T1w = FMA(T1s, T1t, T1u * T1v);
T3k = FNMS(T1u, T1t, T1s * T1v);
}
{
E T1J, T1L, T1I, T1K;
T1J = cr[WS(rs, 17)];
T1L = ci[WS(rs, 17)];
T1I = W[32];
T1K = W[33];
T1M = FMA(T1I, T1J, T1K * T1L);
T3o = FNMS(T1K, T1J, T1I * T1L);
}
{
E T1y, T1A, T1x, T1z;
T1y = cr[WS(rs, 22)];
T1A = ci[WS(rs, 22)];
T1x = W[42];
T1z = W[43];
T1B = FMA(T1x, T1y, T1z * T1A);
T3l = FNMS(T1z, T1y, T1x * T1A);
}
{
E T1E, T1G, T1D, T1F;
T1E = cr[WS(rs, 12)];
T1G = ci[WS(rs, 12)];
T1D = W[22];
T1F = W[23];
T1H = FMA(T1D, T1E, T1F * T1G);
T3n = FNMS(T1F, T1E, T1D * T1G);
}
T3h = T1w - T1B;
T3i = T1H - T1M;
T3A = T3n - T3o;
T3z = T3k - T3l;
T3m = T3k + T3l;
T3p = T3n + T3o;
T3s = T3m + T3p;
T1C = T1w + T1B;
T1N = T1H + T1M;
T1O = T1C + T1N;
}
T1P = T1r + T1O;
T5Z = T3r + T3s;
{
E T3j, T53, T3u, T52, T3q, T3t;
T3j = FMA(KP951056516, T3h, KP587785252 * T3i);
T53 = FNMS(KP587785252, T3h, KP951056516 * T3i);
T3q = KP559016994 * (T3m - T3p);
T3t = FNMS(KP250000000, T3s, T3r);
T3u = T3q + T3t;
T52 = T3t - T3q;
T3v = T3j + T3u;
T5A = T53 + T52;
T4o = T3u - T3j;
T54 = T52 - T53;
}
{
E T3B, T50, T3y, T4Z, T3w, T3x;
T3B = FMA(KP951056516, T3z, KP587785252 * T3A);
T50 = FNMS(KP587785252, T3z, KP951056516 * T3A);
T3w = KP559016994 * (T1C - T1N);
T3x = FNMS(KP250000000, T1O, T1r);
T3y = T3w + T3x;
T4Z = T3x - T3w;
T3C = T3y - T3B;
T5z = T4Z - T50;
T4n = T3y + T3B;
T51 = T4Z + T50;
}
}
{
E T62, T64, Tp, T2k, T5T, T5U, T63, T5V;
{
E T5Y, T61, T1m, T2j;
T5Y = T5W - T5X;
T61 = T5Z - T60;
T62 = FMA(KP951056516, T5Y, KP587785252 * T61);
T64 = FNMS(KP587785252, T5Y, KP951056516 * T61);
Tp = T1 + To;
T1m = TS + T1l;
T2j = T1P + T2i;
T2k = T1m + T2j;
T5T = KP559016994 * (T1m - T2j);
T5U = FNMS(KP250000000, T2k, Tp);
}
cr[0] = Tp + T2k;
T63 = T5U - T5T;
cr[WS(rs, 10)] = T63 - T64;
ci[WS(rs, 9)] = T63 + T64;
T5V = T5T + T5U;
ci[WS(rs, 4)] = T5V - T62;
cr[WS(rs, 5)] = T5V + T62;
}
{
E T2v, T4f, T6I, T6U, T42, T6Z, T43, T6Y, T4A, T6N, T4D, T6L, T4u, T6E, T4v;
E T6D, T48, T6V, T4b, T6T, T2n, T6G;
T2n = T2l + T2m;
T2v = T2n - T2u;
T4f = T2n + T2u;
T6G = T6g + T6f;
T6I = T6G - T6H;
T6U = T6H + T6G;
{
E T2S, T3f, T3g, T3D, T40, T41;
T2S = FMA(KP535826794, T2G, KP844327925 * T2R);
T3f = FNMS(KP637423989, T3e, KP770513242 * T37);
T3g = T2S + T3f;
T3D = FNMS(KP425779291, T3C, KP904827052 * T3v);
T40 = FNMS(KP992114701, T3Z, KP125333233 * T3S);
T41 = T3D + T40;
T42 = T3g + T41;
T6Z = T3D - T40;
T43 = KP559016994 * (T3g - T41);
T6Y = T3f - T2S;
}
{
E T4y, T4z, T6J, T4B, T4C, T6K;
T4y = FNMS(KP248689887, T4g, KP968583161 * T4h);
T4z = FNMS(KP844327925, T4j, KP535826794 * T4k);
T6J = T4y + T4z;
T4B = FNMS(KP481753674, T4n, KP876306680 * T4o);
T4C = FNMS(KP684547105, T4q, KP728968627 * T4r);
T6K = T4B + T4C;
T4A = T4y - T4z;
T6N = KP559016994 * (T6J - T6K);
T4D = T4B - T4C;
T6L = T6J + T6K;
}
{
E T4i, T4l, T4m, T4p, T4s, T4t;
T4i = FMA(KP968583161, T4g, KP248689887 * T4h);
T4l = FMA(KP535826794, T4j, KP844327925 * T4k);
T4m = T4i + T4l;
T4p = FMA(KP876306680, T4n, KP481753674 * T4o);
T4s = FMA(KP728968627, T4q, KP684547105 * T4r);
T4t = T4p + T4s;
T4u = T4m + T4t;
T6E = T4p - T4s;
T4v = KP559016994 * (T4m - T4t);
T6D = T4l - T4i;
}
{
E T46, T47, T6R, T49, T4a, T6S;
T46 = FNMS(KP844327925, T2G, KP535826794 * T2R);
T47 = FMA(KP770513242, T3e, KP637423989 * T37);
T6R = T46 + T47;
T49 = FMA(KP125333233, T3Z, KP992114701 * T3S);
T4a = FMA(KP904827052, T3C, KP425779291 * T3v);
T6S = T4a + T49;
T48 = T46 - T47;
T6V = T6R - T6S;
T4b = T49 - T4a;
T6T = KP559016994 * (T6R + T6S);
}
cr[WS(rs, 4)] = T2v + T42;
ci[WS(rs, 23)] = T6L + T6I;
ci[WS(rs, 20)] = T6V + T6U;
cr[WS(rs, 1)] = T4f + T4u;
{
E T4c, T4e, T45, T4d, T44;
T4c = FMA(KP951056516, T48, KP587785252 * T4b);
T4e = FNMS(KP587785252, T48, KP951056516 * T4b);
T44 = FNMS(KP250000000, T42, T2v);
T45 = T43 + T44;
T4d = T44 - T43;
ci[0] = T45 - T4c;
ci[WS(rs, 5)] = T4d + T4e;
cr[WS(rs, 9)] = T45 + T4c;
ci[WS(rs, 10)] = T4d - T4e;
}
{
E T6F, T6P, T6O, T6Q, T6M;
T6F = FMA(KP587785252, T6D, KP951056516 * T6E);
T6P = FNMS(KP587785252, T6E, KP951056516 * T6D);
T6M = FNMS(KP250000000, T6L, T6I);
T6O = T6M - T6N;
T6Q = T6N + T6M;
cr[WS(rs, 16)] = T6F - T6O;
ci[WS(rs, 18)] = T6P + T6Q;
ci[WS(rs, 13)] = T6F + T6O;
cr[WS(rs, 21)] = T6P - T6Q;
}
{
E T70, T71, T6X, T72, T6W;
T70 = FMA(KP587785252, T6Y, KP951056516 * T6Z);
T71 = FNMS(KP587785252, T6Z, KP951056516 * T6Y);
T6W = FNMS(KP250000000, T6V, T6U);
T6X = T6T - T6W;
T72 = T6T + T6W;
cr[WS(rs, 14)] = T6X - T70;
ci[WS(rs, 15)] = T71 + T72;
cr[WS(rs, 19)] = T70 + T6X;
cr[WS(rs, 24)] = T71 - T72;
}
{
E T4E, T4G, T4x, T4F, T4w;
T4E = FMA(KP951056516, T4A, KP587785252 * T4D);
T4G = FNMS(KP587785252, T4A, KP951056516 * T4D);
T4w = FNMS(KP250000000, T4u, T4f);
T4x = T4v + T4w;
T4F = T4w - T4v;
ci[WS(rs, 3)] = T4x - T4E;
ci[WS(rs, 8)] = T4F + T4G;
cr[WS(rs, 6)] = T4x + T4E;
cr[WS(rs, 11)] = T4F - T4G;
}
}
{
E T75, T7d, T76, T79, T7a, T7b, T7e, T7c;
{
E T73, T74, T77, T78;
T73 = T1l - TS;
T74 = T1P - T2i;
T75 = FMA(KP587785252, T73, KP951056516 * T74);
T7d = FNMS(KP587785252, T74, KP951056516 * T73);
T76 = T6e + T6b;
T77 = T5W + T5X;
T78 = T5Z + T60;
T79 = T77 + T78;
T7a = FNMS(KP250000000, T79, T76);
T7b = KP559016994 * (T77 - T78);
}
ci[WS(rs, 24)] = T79 + T76;
T7e = T7b + T7a;
cr[WS(rs, 20)] = T7d - T7e;
ci[WS(rs, 19)] = T7d + T7e;
T7c = T7a - T7b;
cr[WS(rs, 15)] = T75 - T7c;
ci[WS(rs, 14)] = T75 + T7c;
}
{
E T4J, T5r, T6i, T6u, T5e, T6z, T5f, T6y, T5M, T6n, T5P, T6l, T5G, T66, T5H;
E T65, T5k, T6v, T5n, T6t, T4H, T6h;
T4H = T2m - T2l;
T4J = T4H + T4I;
T5r = T4H - T4I;
T6h = T6f - T6g;
T6i = T6a + T6h;
T6u = T6h - T6a;
{
E T4Q, T4X, T4Y, T55, T5c, T5d;
T4Q = FMA(KP728968627, T4M, KP684547105 * T4P);
T4X = FNMS(KP992114701, T4W, KP125333233 * T4T);
T4Y = T4Q + T4X;
T55 = FMA(KP062790519, T51, KP998026728 * T54);
T5c = FNMS(KP637423989, T5b, KP770513242 * T58);
T5d = T55 + T5c;
T5e = T4Y + T5d;
T6z = T55 - T5c;
T5f = KP559016994 * (T4Y - T5d);
T6y = T4X - T4Q;
}
{
E T5K, T5L, T6j, T5N, T5O, T6k;
T5K = FNMS(KP481753674, T5s, KP876306680 * T5t);
T5L = FMA(KP904827052, T5w, KP425779291 * T5v);
T6j = T5K - T5L;
T5N = FNMS(KP844327925, T5z, KP535826794 * T5A);
T5O = FNMS(KP998026728, T5C, KP062790519 * T5D);
T6k = T5N + T5O;
T5M = T5K + T5L;
T6n = KP559016994 * (T6j - T6k);
T5P = T5N - T5O;
T6l = T6j + T6k;
}
{
E T5u, T5x, T5y, T5B, T5E, T5F;
T5u = FMA(KP876306680, T5s, KP481753674 * T5t);
T5x = FNMS(KP425779291, T5w, KP904827052 * T5v);
T5y = T5u + T5x;
T5B = FMA(KP535826794, T5z, KP844327925 * T5A);
T5E = FMA(KP062790519, T5C, KP998026728 * T5D);
T5F = T5B + T5E;
T5G = T5y + T5F;
T66 = T5B - T5E;
T5H = KP559016994 * (T5y - T5F);
T65 = T5x - T5u;
}
{
E T5i, T5j, T6r, T5l, T5m, T6s;
T5i = FNMS(KP684547105, T4M, KP728968627 * T4P);
T5j = FMA(KP125333233, T4W, KP992114701 * T4T);
T6r = T5i - T5j;
T5l = FNMS(KP998026728, T51, KP062790519 * T54);
T5m = FMA(KP770513242, T5b, KP637423989 * T58);
T6s = T5l - T5m;
T5k = T5i + T5j;
T6v = T6r + T6s;
T5n = T5l + T5m;
T6t = KP559016994 * (T6r - T6s);
}
cr[WS(rs, 3)] = T4J + T5e;
ci[WS(rs, 22)] = T6l + T6i;
ci[WS(rs, 21)] = T6v + T6u;
cr[WS(rs, 2)] = T5r + T5G;
{
E T67, T6p, T6o, T6q, T6m;
T67 = FMA(KP587785252, T65, KP951056516 * T66);
T6p = FNMS(KP587785252, T66, KP951056516 * T65);
T6m = FNMS(KP250000000, T6l, T6i);
T6o = T6m - T6n;
T6q = T6n + T6m;
cr[WS(rs, 17)] = T67 - T6o;
ci[WS(rs, 17)] = T6p + T6q;
ci[WS(rs, 12)] = T67 + T6o;
cr[WS(rs, 22)] = T6p - T6q;
}
{
E T5Q, T5S, T5J, T5R, T5I;
T5Q = FMA(KP951056516, T5M, KP587785252 * T5P);
T5S = FNMS(KP587785252, T5M, KP951056516 * T5P);
T5I = FNMS(KP250000000, T5G, T5r);
T5J = T5H + T5I;
T5R = T5I - T5H;
ci[WS(rs, 2)] = T5J - T5Q;
ci[WS(rs, 7)] = T5R + T5S;
cr[WS(rs, 7)] = T5J + T5Q;
cr[WS(rs, 12)] = T5R - T5S;
}
{
E T5o, T5q, T5h, T5p, T5g;
T5o = FMA(KP951056516, T5k, KP587785252 * T5n);
T5q = FNMS(KP587785252, T5k, KP951056516 * T5n);
T5g = FNMS(KP250000000, T5e, T4J);
T5h = T5f + T5g;
T5p = T5g - T5f;
ci[WS(rs, 1)] = T5h - T5o;
ci[WS(rs, 6)] = T5p + T5q;
cr[WS(rs, 8)] = T5h + T5o;
ci[WS(rs, 11)] = T5p - T5q;
}
{
E T6A, T6B, T6x, T6C, T6w;
T6A = FMA(KP587785252, T6y, KP951056516 * T6z);
T6B = FNMS(KP587785252, T6z, KP951056516 * T6y);
T6w = FNMS(KP250000000, T6v, T6u);
T6x = T6t - T6w;
T6C = T6t + T6w;
cr[WS(rs, 13)] = T6x - T6A;
ci[WS(rs, 16)] = T6B + T6C;
cr[WS(rs, 18)] = T6A + T6x;
cr[WS(rs, 23)] = T6B - T6C;
}
}
}
}
}
static void hc2cfdft_8(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms)
{
DK(KP707106781, +0.707106781186547524400844362104849039284835938);
DK(KP500000000, +0.500000000000000000000000000000000000000000000);
{
INT m;
for (m = mb, W = W + ((mb - 1) * 14); m < me; m = m + 1, Rp = Rp + ms, Ip = Ip + ms, Rm = Rm - ms, Im = Im - ms, W = W + 14, MAKE_VOLATILE_STRIDE(32, rs)) {
E T1A, T1w, T1z, T1x, T1H, T1v, T1L, T1F;
{
E Ty, T14, TO, T1o, Tv, TG, T16, T1m, Ta, T19, T1h, TV, T10, TX, TZ;
E Tk, T1i, TY, T1b, TF, TB, T1l;
{
E TH, TN, TK, TM;
{
E Tw, Tx, TI, TJ;
Tw = Ip[0];
Tx = Im[0];
TI = Rm[0];
TJ = Rp[0];
TH = W[0];
Ty = Tw - Tx;
TN = Tw + Tx;
T14 = TJ + TI;
TK = TI - TJ;
TM = W[1];
}
{
E Ts, Tp, Tt, Tm, Tr;
{
E Tn, To, TL, T1n;
Tn = Ip[WS(rs, 2)];
To = Im[WS(rs, 2)];
TL = TH * TK;
T1n = TM * TK;
Ts = Rp[WS(rs, 2)];
TF = Tn + To;
Tp = Tn - To;
TO = FNMS(TM, TN, TL);
T1o = FMA(TH, TN, T1n);
Tt = Rm[WS(rs, 2)];
}
Tm = W[6];
Tr = W[7];
{
E TE, TD, T15, TC, Tu, Tq;
TB = W[8];
TC = Tt - Ts;
Tu = Ts + Tt;
Tq = Tm * Tp;
TE = W[9];
TD = TB * TC;
T15 = Tm * Tu;
Tv = FNMS(Tr, Tu, Tq);
T1l = TE * TC;
TG = FNMS(TE, TF, TD);
T16 = FMA(Tr, Tp, T15);
}
}
}
{
E TU, TR, TT, T1g, TS;
{
E T2, T3, T7, T8;
T2 = Ip[WS(rs, 1)];
T1m = FMA(TB, TF, T1l);
T3 = Im[WS(rs, 1)];
T7 = Rp[WS(rs, 1)];
T8 = Rm[WS(rs, 1)];
{
E T1, T4, T9, T6, T5, TQ, T18;
T1 = W[2];
TU = T2 + T3;
T4 = T2 - T3;
TR = T7 - T8;
T9 = T7 + T8;
T6 = W[3];
T5 = T1 * T4;
TQ = W[4];
T18 = T1 * T9;
TT = W[5];
Ta = FNMS(T6, T9, T5);
T1g = TQ * TU;
TS = TQ * TR;
T19 = FMA(T6, T4, T18);
}
}
{
E Tc, Td, Th, Ti;
Tc = Ip[WS(rs, 3)];
T1h = FNMS(TT, TR, T1g);
TV = FMA(TT, TU, TS);
Td = Im[WS(rs, 3)];
Th = Rp[WS(rs, 3)];
Ti = Rm[WS(rs, 3)];
{
E Tb, Te, Tj, Tg, Tf, TW, T1a;
Tb = W[10];
T10 = Tc + Td;
Te = Tc - Td;
TX = Th - Ti;
Tj = Th + Ti;
Tg = W[11];
Tf = Tb * Te;
TW = W[12];
T1a = Tb * Tj;
TZ = W[13];
Tk = FNMS(Tg, Tj, Tf);
T1i = TW * T10;
TY = TW * TX;
T1b = FMA(Tg, Te, T1a);
}
}
}
{
E T1E, T1t, TA, T1s, T1D, T1u, T1e, T13, T1r, T1d;
{
E TP, T1f, T1q, T12, T17, T1c;
{
E Tl, T11, Tz, T1p, T1k, T1j;
T1E = Ta - Tk;
Tl = Ta + Tk;
T1j = FNMS(TZ, TX, T1i);
T11 = FMA(TZ, T10, TY);
Tz = Tv + Ty;
T1t = Ty - Tv;
T1A = T1o - T1m;
T1p = T1m + T1o;
T1k = T1h + T1j;
T1w = T1j - T1h;
T1z = TO - TG;
TP = TG + TO;
T1f = Tz - Tl;
TA = Tl + Tz;
T1s = T1k + T1p;
T1q = T1k - T1p;
T12 = TV + T11;
T1x = TV - T11;
T1D = T14 - T16;
T17 = T14 + T16;
T1c = T19 + T1b;
T1u = T19 - T1b;
}
Im[WS(rs, 1)] = KP500000000 * (T1q - T1f);
T1e = T12 + TP;
T13 = TP - T12;
T1r = T17 + T1c;
T1d = T17 - T1c;
Ip[WS(rs, 2)] = KP500000000 * (T1f + T1q);
}
Im[WS(rs, 3)] = KP500000000 * (T13 - TA);
Ip[0] = KP500000000 * (TA + T13);
Rm[WS(rs, 3)] = KP500000000 * (T1r - T1s);
Rp[0] = KP500000000 * (T1r + T1s);
Rp[WS(rs, 2)] = KP500000000 * (T1d + T1e);
Rm[WS(rs, 1)] = KP500000000 * (T1d - T1e);
T1H = T1u + T1t;
T1v = T1t - T1u;
T1L = T1D + T1E;
T1F = T1D - T1E;
}
}
{
E T1y, T1I, T1B, T1J;
T1y = T1w + T1x;
T1I = T1w - T1x;
T1B = T1z - T1A;
T1J = T1z + T1A;
{
E T1M, T1K, T1C, T1G;
T1M = T1I + T1J;
T1K = T1I - T1J;
T1C = T1y + T1B;
T1G = T1B - T1y;
Im[0] = -(KP500000000 * (FNMS(KP707106781, T1K, T1H)));
Ip[WS(rs, 3)] = KP500000000 * (FMA(KP707106781, T1K, T1H));
Rp[WS(rs, 1)] = KP500000000 * (FMA(KP707106781, T1M, T1L));
Rm[WS(rs, 2)] = KP500000000 * (FNMS(KP707106781, T1M, T1L));
Rp[WS(rs, 3)] = KP500000000 * (FMA(KP707106781, T1G, T1F));
Rm[0] = KP500000000 * (FNMS(KP707106781, T1G, T1F));
Im[WS(rs, 2)] = -(KP500000000 * (FNMS(KP707106781, T1C, T1v)));
Ip[WS(rs, 1)] = KP500000000 * (FMA(KP707106781, T1C, T1v));
}
}
}
}
}
static void t1_8(float *ri, float *ii, const float *W, stride rs, INT mb, INT me, INT ms)
{
DK(KP707106781, +0.707106781186547524400844362104849039284835938);
INT m;
for (m = mb, W = W + (mb * 14); m < me; m = m + 1, ri = ri + ms, ii = ii + ms, W = W + 14, MAKE_VOLATILE_STRIDE(rs)) {
E T7, T1e, TH, T19, TF, T13, TR, TU, Ti, T1f, TK, T16, Tu, T12, TM;
E TP;
{
E T1, T18, T6, T17;
T1 = ri[0];
T18 = ii[0];
{
E T3, T5, T2, T4;
T3 = ri[WS(rs, 4)];
T5 = ii[WS(rs, 4)];
T2 = W[6];
T4 = W[7];
T6 = FMA(T2, T3, T4 * T5);
T17 = FNMS(T4, T3, T2 * T5);
}
T7 = T1 + T6;
T1e = T18 - T17;
TH = T1 - T6;
T19 = T17 + T18;
}
{
E Tz, TS, TE, TT;
{
E Tw, Ty, Tv, Tx;
Tw = ri[WS(rs, 7)];
Ty = ii[WS(rs, 7)];
Tv = W[12];
Tx = W[13];
Tz = FMA(Tv, Tw, Tx * Ty);
TS = FNMS(Tx, Tw, Tv * Ty);
}
{
E TB, TD, TA, TC;
TB = ri[WS(rs, 3)];
TD = ii[WS(rs, 3)];
TA = W[4];
TC = W[5];
TE = FMA(TA, TB, TC * TD);
TT = FNMS(TC, TB, TA * TD);
}
TF = Tz + TE;
T13 = TS + TT;
TR = Tz - TE;
TU = TS - TT;
}
{
E Tc, TI, Th, TJ;
{
E T9, Tb, T8, Ta;
T9 = ri[WS(rs, 2)];
Tb = ii[WS(rs, 2)];
T8 = W[2];
Ta = W[3];
Tc = FMA(T8, T9, Ta * Tb);
TI = FNMS(Ta, T9, T8 * Tb);
}
{
E Te, Tg, Td, Tf;
Te = ri[WS(rs, 6)];
Tg = ii[WS(rs, 6)];
Td = W[10];
Tf = W[11];
Th = FMA(Td, Te, Tf * Tg);
TJ = FNMS(Tf, Te, Td * Tg);
}
Ti = Tc + Th;
T1f = Tc - Th;
TK = TI - TJ;
T16 = TI + TJ;
}
{
E To, TN, Tt, TO;
{
E Tl, Tn, Tk, Tm;
Tl = ri[WS(rs, 1)];
Tn = ii[WS(rs, 1)];
Tk = W[0];
Tm = W[1];
To = FMA(Tk, Tl, Tm * Tn);
TN = FNMS(Tm, Tl, Tk * Tn);
}
{
E Tq, Ts, Tp, Tr;
Tq = ri[WS(rs, 5)];
Ts = ii[WS(rs, 5)];
Tp = W[8];
Tr = W[9];
Tt = FMA(Tp, Tq, Tr * Ts);
TO = FNMS(Tr, Tq, Tp * Ts);
}
Tu = To + Tt;
T12 = TN + TO;
TM = To - Tt;
TP = TN - TO;
}
{
E Tj, TG, T1b, T1c;
Tj = T7 + Ti;
TG = Tu + TF;
ri[WS(rs, 4)] = Tj - TG;
ri[0] = Tj + TG;
{
E T15, T1a, T11, T14;
T15 = T12 + T13;
T1a = T16 + T19;
ii[0] = T15 + T1a;
ii[WS(rs, 4)] = T1a - T15;
T11 = T7 - Ti;
T14 = T12 - T13;
ri[WS(rs, 6)] = T11 - T14;
ri[WS(rs, 2)] = T11 + T14;
}
T1b = TF - Tu;
T1c = T19 - T16;
ii[WS(rs, 2)] = T1b + T1c;
ii[WS(rs, 6)] = T1c - T1b;
{
E TX, T1g, T10, T1d, TY, TZ;
TX = TH - TK;
T1g = T1e - T1f;
TY = TP - TM;
TZ = TR + TU;
T10 = KP707106781 * (TY - TZ);
T1d = KP707106781 * (TY + TZ);
ri[WS(rs, 7)] = TX - T10;
ii[WS(rs, 5)] = T1g - T1d;
ri[WS(rs, 3)] = TX + T10;
ii[WS(rs, 1)] = T1d + T1g;
}
{
E TL, T1i, TW, T1h, TQ, TV;
TL = TH + TK;
T1i = T1f + T1e;
TQ = TM + TP;
TV = TR - TU;
TW = KP707106781 * (TQ + TV);
T1h = KP707106781 * (TV - TQ);
ri[WS(rs, 5)] = TL - TW;
ii[WS(rs, 7)] = T1i - T1h;
ri[WS(rs, 1)] = TL + TW;
ii[WS(rs, 3)] = T1h + T1i;
}
}
}
}
static void hb_5(R *cr, R *ci, const R *W, stride rs, INT mb, INT me, INT ms)
{
DK(KP250000000, +0.250000000000000000000000000000000000000000000);
DK(KP587785252, +0.587785252292473129168705954639072768597652438);
DK(KP951056516, +0.951056516295153572116439333379382143405698634);
DK(KP559016994, +0.559016994374947424102293417182819058860154590);
INT m;
for (m = mb, W = W + ((mb - 1) * 8); m < me; m = m + 1, cr = cr + ms, ci = ci - ms, W = W + 8, MAKE_VOLATILE_STRIDE(rs)) {
E T1, Tj, TG, Ts, T8, Ti, T9, Tn, TD, Tu, Tg, Tt;
{
E T4, Tq, T7, Tr;
T1 = cr[0];
{
E T2, T3, T5, T6;
T2 = cr[WS(rs, 1)];
T3 = ci[0];
T4 = T2 + T3;
Tq = T2 - T3;
T5 = cr[WS(rs, 2)];
T6 = ci[WS(rs, 1)];
T7 = T5 + T6;
Tr = T5 - T6;
}
Tj = KP559016994 * (T4 - T7);
TG = FMA(KP951056516, Tq, KP587785252 * Tr);
Ts = FNMS(KP951056516, Tr, KP587785252 * Tq);
T8 = T4 + T7;
Ti = FNMS(KP250000000, T8, T1);
}
{
E Tc, Tl, Tf, Tm;
T9 = ci[WS(rs, 4)];
{
E Ta, Tb, Td, Te;
Ta = ci[WS(rs, 3)];
Tb = cr[WS(rs, 4)];
Tc = Ta - Tb;
Tl = Ta + Tb;
Td = ci[WS(rs, 2)];
Te = cr[WS(rs, 3)];
Tf = Td - Te;
Tm = Td + Te;
}
Tn = FNMS(KP951056516, Tm, KP587785252 * Tl);
TD = FMA(KP951056516, Tl, KP587785252 * Tm);
Tu = KP559016994 * (Tc - Tf);
Tg = Tc + Tf;
Tt = FNMS(KP250000000, Tg, T9);
}
cr[0] = T1 + T8;
ci[0] = T9 + Tg;
{
E To, Ty, Tw, TA, Tk, Tv;
Tk = Ti - Tj;
To = Tk - Tn;
Ty = Tk + Tn;
Tv = Tt - Tu;
Tw = Ts + Tv;
TA = Tv - Ts;
{
E Th, Tp, Tx, Tz;
Th = W[2];
Tp = W[3];
cr[WS(rs, 2)] = FNMS(Tp, Tw, Th * To);
ci[WS(rs, 2)] = FMA(Th, Tw, Tp * To);
Tx = W[4];
Tz = W[5];
cr[WS(rs, 3)] = FNMS(Tz, TA, Tx * Ty);
ci[WS(rs, 3)] = FMA(Tx, TA, Tz * Ty);
}
}
{
E TE, TK, TI, TM, TC, TH;
TC = Tj + Ti;
TE = TC - TD;
TK = TC + TD;
TH = Tu + Tt;
TI = TG + TH;
TM = TH - TG;
{
E TB, TF, TJ, TL;
TB = W[0];
TF = W[1];
cr[WS(rs, 1)] = FNMS(TF, TI, TB * TE);
ci[WS(rs, 1)] = FMA(TB, TI, TF * TE);
TJ = W[6];
TL = W[7];
cr[WS(rs, 4)] = FNMS(TL, TM, TJ * TK);
ci[WS(rs, 4)] = FMA(TJ, TM, TL * TK);
}
}
}
}
static void hc2cb_16(float *Rp, float *Ip, float *Rm, float *Im, const float *W, stride rs, INT mb, INT me, INT ms)
{
DK(KP382683432, +0.382683432365089771728459984030398866761344562);
DK(KP923879532, +0.923879532511286756128183189396788286822416626);
DK(KP707106781, +0.707106781186547524400844362104849039284835938);
INT m;
for (m = mb, W = W + ((mb - 1) * 30); m < me; m = m + 1, Rp = Rp + ms, Ip = Ip + ms, Rm = Rm - ms, Im = Im - ms, W = W + 30, MAKE_VOLATILE_STRIDE(rs)) {
E T7, T2K, T2W, Tw, T17, T1S, T2k, T1w, Te, TD, T1x, T10, T2n, T2L, T1Z;
E T2X, Tm, T1z, TN, T19, T2e, T2p, T2P, T2Z, Tt, T1A, TW, T1a, T27, T2q;
E T2S, T30;
{
E T3, T1Q, T13, T2j, T6, T2i, T16, T1R;
{
E T1, T2, T11, T12;
T1 = Rp[0];
T2 = Rm[WS(rs, 7)];
T3 = T1 + T2;
T1Q = T1 - T2;
T11 = Ip[0];
T12 = Im[WS(rs, 7)];
T13 = T11 - T12;
T2j = T11 + T12;
}
{
E T4, T5, T14, T15;
T4 = Rp[WS(rs, 4)];
T5 = Rm[WS(rs, 3)];
T6 = T4 + T5;
T2i = T4 - T5;
T14 = Ip[WS(rs, 4)];
T15 = Im[WS(rs, 3)];
T16 = T14 - T15;
T1R = T14 + T15;
}
T7 = T3 + T6;
T2K = T1Q + T1R;
T2W = T2j - T2i;
Tw = T3 - T6;
T17 = T13 - T16;
T1S = T1Q - T1R;
T2k = T2i + T2j;
T1w = T13 + T16;
}
{
E Ta, T1T, TC, T1U, Td, T1W, Tz, T1X;
{
E T8, T9, TA, TB;
T8 = Rp[WS(rs, 2)];
T9 = Rm[WS(rs, 5)];
Ta = T8 + T9;
T1T = T8 - T9;
TA = Ip[WS(rs, 2)];
TB = Im[WS(rs, 5)];
TC = TA - TB;
T1U = TA + TB;
}
{
E Tb, Tc, Tx, Ty;
Tb = Rm[WS(rs, 1)];
Tc = Rp[WS(rs, 6)];
Td = Tb + Tc;
T1W = Tb - Tc;
Tx = Ip[WS(rs, 6)];
Ty = Im[WS(rs, 1)];
Tz = Tx - Ty;
T1X = Tx + Ty;
}
Te = Ta + Td;
TD = Tz - TC;
T1x = TC + Tz;
T10 = Ta - Td;
{
E T2l, T2m, T1V, T1Y;
T2l = T1T + T1U;
T2m = T1W + T1X;
T2n = KP707106781 * (T2l - T2m);
T2L = KP707106781 * (T2l + T2m);
T1V = T1T - T1U;
T1Y = T1W - T1X;
T1Z = KP707106781 * (T1V + T1Y);
T2X = KP707106781 * (T1V - T1Y);
}
}
{
E Ti, T2b, TI, T29, Tl, T28, TL, T2c, TF, TM;
{
E Tg, Th, TG, TH;
Tg = Rp[WS(rs, 1)];
Th = Rm[WS(rs, 6)];
Ti = Tg + Th;
T2b = Tg - Th;
TG = Ip[WS(rs, 1)];
TH = Im[WS(rs, 6)];
TI = TG - TH;
T29 = TG + TH;
}
{
E Tj, Tk, TJ, TK;
Tj = Rp[WS(rs, 5)];
Tk = Rm[WS(rs, 2)];
Tl = Tj + Tk;
T28 = Tj - Tk;
TJ = Ip[WS(rs, 5)];
TK = Im[WS(rs, 2)];
TL = TJ - TK;
T2c = TJ + TK;
}
Tm = Ti + Tl;
T1z = TI + TL;
TF = Ti - Tl;
TM = TI - TL;
TN = TF - TM;
T19 = TF + TM;
{
E T2a, T2d, T2N, T2O;
T2a = T28 + T29;
T2d = T2b - T2c;
T2e = FMA(KP923879532, T2a, KP382683432 * T2d);
T2p = FNMS(KP382683432, T2a, KP923879532 * T2d);
T2N = T2b + T2c;
T2O = T29 - T28;
T2P = FNMS(KP923879532, T2O, KP382683432 * T2N);
T2Z = FMA(KP382683432, T2O, KP923879532 * T2N);
}
}
{
E Tp, T24, TR, T22, Ts, T21, TU, T25, TO, TV;
{
E Tn, To, TP, TQ;
Tn = Rm[0];
To = Rp[WS(rs, 7)];
Tp = Tn + To;
T24 = Tn - To;
TP = Ip[WS(rs, 7)];
TQ = Im[0];
TR = TP - TQ;
T22 = TP + TQ;
}
{
E Tq, Tr, TS, TT;
Tq = Rp[WS(rs, 3)];
Tr = Rm[WS(rs, 4)];
Ts = Tq + Tr;
T21 = Tq - Tr;
TS = Ip[WS(rs, 3)];
TT = Im[WS(rs, 4)];
TU = TS - TT;
T25 = TS + TT;
}
Tt = Tp + Ts;
T1A = TR + TU;
TO = Tp - Ts;
TV = TR - TU;
TW = TO + TV;
T1a = TV - TO;
{
E T23, T26, T2Q, T2R;
T23 = T21 - T22;
T26 = T24 - T25;
T27 = FNMS(KP382683432, T26, KP923879532 * T23);
T2q = FMA(KP382683432, T23, KP923879532 * T26);
T2Q = T24 + T25;
T2R = T21 + T22;
T2S = FNMS(KP923879532, T2R, KP382683432 * T2Q);
T30 = FMA(KP382683432, T2R, KP923879532 * T2Q);
}
}
{
E Tf, Tu, T1u, T1y, T1B, T1C, T1t, T1v;
Tf = T7 + Te;
Tu = Tm + Tt;
T1u = Tf - Tu;
T1y = T1w + T1x;
T1B = T1z + T1A;
T1C = T1y - T1B;
Rp[0] = Tf + Tu;
Rm[0] = T1y + T1B;
T1t = W[14];
T1v = W[15];
Rp[WS(rs, 4)] = FNMS(T1v, T1C, T1t * T1u);
Rm[WS(rs, 4)] = FMA(T1v, T1u, T1t * T1C);
}
{
E T2U, T34, T32, T36;
{
E T2M, T2T, T2Y, T31;
T2M = T2K - T2L;
T2T = T2P + T2S;
T2U = T2M - T2T;
T34 = T2M + T2T;
T2Y = T2W + T2X;
T31 = T2Z - T30;
T32 = T2Y - T31;
T36 = T2Y + T31;
}
{
E T2J, T2V, T33, T35;
T2J = W[20];
T2V = W[21];
Ip[WS(rs, 5)] = FNMS(T2V, T32, T2J * T2U);
Im[WS(rs, 5)] = FMA(T2V, T2U, T2J * T32);
T33 = W[4];
T35 = W[5];
Ip[WS(rs, 1)] = FNMS(T35, T36, T33 * T34);
Im[WS(rs, 1)] = FMA(T35, T34, T33 * T36);
}
}
{
E T3a, T3g, T3e, T3i;
{
E T38, T39, T3c, T3d;
T38 = T2K + T2L;
T39 = T2Z + T30;
T3a = T38 - T39;
T3g = T38 + T39;
T3c = T2W - T2X;
T3d = T2P - T2S;
T3e = T3c + T3d;
T3i = T3c - T3d;
}
{
E T37, T3b, T3f, T3h;
T37 = W[12];
T3b = W[13];
Ip[WS(rs, 3)] = FNMS(T3b, T3e, T37 * T3a);
Im[WS(rs, 3)] = FMA(T37, T3e, T3b * T3a);
T3f = W[28];
T3h = W[29];
Ip[WS(rs, 7)] = FNMS(T3h, T3i, T3f * T3g);
Im[WS(rs, 7)] = FMA(T3f, T3i, T3h * T3g);
}
}
{
E TY, T1e, T1c, T1g;
{
E TE, TX, T18, T1b;
TE = Tw + TD;
TX = KP707106781 * (TN + TW);
TY = TE - TX;
T1e = TE + TX;
T18 = T10 + T17;
T1b = KP707106781 * (T19 + T1a);
T1c = T18 - T1b;
T1g = T18 + T1b;
}
{
E Tv, TZ, T1d, T1f;
Tv = W[18];
TZ = W[19];
Rp[WS(rs, 5)] = FNMS(TZ, T1c, Tv * TY);
Rm[WS(rs, 5)] = FMA(TZ, TY, Tv * T1c);
T1d = W[2];
T1f = W[3];
Rp[WS(rs, 1)] = FNMS(T1f, T1g, T1d * T1e);
Rm[WS(rs, 1)] = FMA(T1f, T1e, T1d * T1g);
}
}
{
E T1k, T1q, T1o, T1s;
{
E T1i, T1j, T1m, T1n;
T1i = Tw - TD;
T1j = KP707106781 * (T1a - T19);
T1k = T1i - T1j;
T1q = T1i + T1j;
T1m = T17 - T10;
T1n = KP707106781 * (TN - TW);
T1o = T1m - T1n;
T1s = T1m + T1n;
}
{
E T1h, T1l, T1p, T1r;
T1h = W[26];
T1l = W[27];
Rp[WS(rs, 7)] = FNMS(T1l, T1o, T1h * T1k);
Rm[WS(rs, 7)] = FMA(T1h, T1o, T1l * T1k);
T1p = W[10];
T1r = W[11];
Rp[WS(rs, 3)] = FNMS(T1r, T1s, T1p * T1q);
Rm[WS(rs, 3)] = FMA(T1p, T1s, T1r * T1q);
}
}
{
E T2g, T2u, T2s, T2w;
{
E T20, T2f, T2o, T2r;
T20 = T1S - T1Z;
T2f = T27 - T2e;
T2g = T20 - T2f;
T2u = T20 + T2f;
T2o = T2k - T2n;
T2r = T2p - T2q;
T2s = T2o - T2r;
T2w = T2o + T2r;
}
{
E T1P, T2h, T2t, T2v;
T1P = W[24];
T2h = W[25];
Ip[WS(rs, 6)] = FNMS(T2h, T2s, T1P * T2g);
Im[WS(rs, 6)] = FMA(T2h, T2g, T1P * T2s);
T2t = W[8];
T2v = W[9];
Ip[WS(rs, 2)] = FNMS(T2v, T2w, T2t * T2u);
Im[WS(rs, 2)] = FMA(T2v, T2u, T2t * T2w);
}
}
{
E T2A, T2G, T2E, T2I;
{
E T2y, T2z, T2C, T2D;
T2y = T1S + T1Z;
T2z = T2p + T2q;
T2A = T2y - T2z;
T2G = T2y + T2z;
T2C = T2k + T2n;
T2D = T2e + T27;
T2E = T2C - T2D;
T2I = T2C + T2D;
}
{
E T2x, T2B, T2F, T2H;
T2x = W[16];
T2B = W[17];
Ip[WS(rs, 4)] = FNMS(T2B, T2E, T2x * T2A);
Im[WS(rs, 4)] = FMA(T2x, T2E, T2B * T2A);
T2F = W[0];
T2H = W[1];
Ip[0] = FNMS(T2H, T2I, T2F * T2G);
Im[0] = FMA(T2F, T2I, T2H * T2G);
}
}
{
E T1G, T1M, T1K, T1O;
{
E T1E, T1F, T1I, T1J;
T1E = T7 - Te;
T1F = T1A - T1z;
T1G = T1E - T1F;
T1M = T1E + T1F;
T1I = T1w - T1x;
T1J = Tm - Tt;
T1K = T1I - T1J;
T1O = T1J + T1I;
}
{
E T1D, T1H, T1L, T1N;
T1D = W[22];
T1H = W[23];
Rp[WS(rs, 6)] = FNMS(T1H, T1K, T1D * T1G);
Rm[WS(rs, 6)] = FMA(T1D, T1K, T1H * T1G);
T1L = W[6];
T1N = W[7];
Rp[WS(rs, 2)] = FNMS(T1N, T1O, T1L * T1M);
Rm[WS(rs, 2)] = FMA(T1L, T1O, T1N * T1M);
}
}
}
}