static void hc2cfdftv_4(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
INT m;
for (m = mb, W = W + ((mb - 1) * ((TWVL / VL) * 6)); m < me; m = m + VL, Rp = Rp + (VL * ms), Ip = Ip + (VL * ms), Rm = Rm - (VL * ms), Im = Im - (VL * ms), W = W + (TWVL * 6), MAKE_VOLATILE_STRIDE(rs)) {
V T1, T2, Tb, T5, T6, T4, T9, T3, Tc, T7, Ta, Tg, T8, Td, Th;
V Tf, Te, Ti, Tj;
T1 = LD(&(Rp[0]), ms, &(Rp[0]));
T2 = LD(&(Rm[0]), -ms, &(Rm[0]));
Tb = LDW(&(W[0]));
T5 = LD(&(Rp[WS(rs, 1)]), ms, &(Rp[WS(rs, 1)]));
T6 = LD(&(Rm[WS(rs, 1)]), -ms, &(Rm[WS(rs, 1)]));
T4 = LDW(&(W[TWVL * 2]));
T9 = LDW(&(W[TWVL * 4]));
T3 = VFMACONJ(T2, T1);
Tc = VZMULIJ(Tb, VFNMSCONJ(T2, T1));
T7 = VZMULJ(T4, VFMACONJ(T6, T5));
Ta = VZMULIJ(T9, VFNMSCONJ(T6, T5));
Tg = VADD(T3, T7);
T8 = VSUB(T3, T7);
Td = VSUB(Ta, Tc);
Th = VADD(Tc, Ta);
Tf = VCONJ(VMUL(LDK(KP500000000), VFMAI(Td, T8)));
Te = VMUL(LDK(KP500000000), VFNMSI(Td, T8));
Ti = VMUL(LDK(KP500000000), VSUB(Tg, Th));
Tj = VCONJ(VMUL(LDK(KP500000000), VADD(Th, Tg)));
ST(&(Rm[0]), Tf, -ms, &(Rm[0]));
ST(&(Rp[WS(rs, 1)]), Te, ms, &(Rp[WS(rs, 1)]));
ST(&(Rp[0]), Ti, ms, &(Rp[0]));
ST(&(Rm[WS(rs, 1)]), Tj, -ms, &(Rm[WS(rs, 1)]));
}
}
static void hc2cbdftv_4(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms)
{
{
INT m;
for (m = mb, W = W + ((mb - 1) * ((TWVL / VL) * 6)); m < me; m = m + VL, Rp = Rp + (VL * ms), Ip = Ip + (VL * ms), Rm = Rm - (VL * ms), Im = Im - (VL * ms), W = W + (TWVL * 6), MAKE_VOLATILE_STRIDE(rs)) {
V T2, T3, T5, T6, Tf, T1, T9, Ta, T4, Tb, T7, Tc, Th, T8, Tg;
V Te, Td, Ti, Tj;
T2 = LD(&(Rp[0]), ms, &(Rp[0]));
T3 = LD(&(Rm[WS(rs, 1)]), -ms, &(Rm[WS(rs, 1)]));
T5 = LD(&(Rp[WS(rs, 1)]), ms, &(Rp[WS(rs, 1)]));
T6 = LD(&(Rm[0]), -ms, &(Rm[0]));
Tf = LDW(&(W[0]));
T1 = LDW(&(W[TWVL * 4]));
T9 = LDW(&(W[TWVL * 2]));
Ta = VFMACONJ(T3, T2);
T4 = VFNMSCONJ(T3, T2);
Tb = VFMACONJ(T6, T5);
T7 = VFNMSCONJ(T6, T5);
Tc = VZMUL(T9, VSUB(Ta, Tb));
Th = VADD(Ta, Tb);
T8 = VZMULI(T1, VFNMSI(T7, T4));
Tg = VZMULI(Tf, VFMAI(T7, T4));
Te = VCONJ(VSUB(Tc, T8));
Td = VADD(T8, Tc);
Ti = VADD(Tg, Th);
Tj = VCONJ(VSUB(Th, Tg));
ST(&(Rm[WS(rs, 1)]), Te, -ms, &(Rm[WS(rs, 1)]));
ST(&(Rp[WS(rs, 1)]), Td, ms, &(Rp[WS(rs, 1)]));
ST(&(Rp[0]), Ti, ms, &(Rp[0]));
ST(&(Rm[0]), Tj, -ms, &(Rm[0]));
}
}
VLEAVE();
}
static void hc2cfdftv_6(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
{
INT m;
for (m = mb, W = W + ((mb - 1) * ((TWVL / VL) * 10)); m < me; m = m + VL, Rp = Rp + (VL * ms), Ip = Ip + (VL * ms), Rm = Rm - (VL * ms), Im = Im - (VL * ms), W = W + (TWVL * 10), MAKE_VOLATILE_STRIDE(24, rs)) {
V T5, T6, T3, Tj, T4, T9, Te, Th, T1, T2, Ti, Tc, Td, Tb, Tg;
V T7, Ta, Tt, Tk, Tr, T8, Ts, Tf, Tx, Tu, To, Tl, Tw, Tv, Tn;
V Tm, Tz, Ty, Tp, Tq;
T1 = LD(&(Rp[0]), ms, &(Rp[0]));
T2 = LD(&(Rm[0]), -ms, &(Rm[0]));
Ti = LDW(&(W[0]));
Tc = LD(&(Rp[WS(rs, 2)]), ms, &(Rp[0]));
Td = LD(&(Rm[WS(rs, 2)]), -ms, &(Rm[0]));
Tb = LDW(&(W[TWVL * 8]));
Tg = LDW(&(W[TWVL * 6]));
T5 = LD(&(Rp[WS(rs, 1)]), ms, &(Rp[WS(rs, 1)]));
T6 = LD(&(Rm[WS(rs, 1)]), -ms, &(Rm[WS(rs, 1)]));
T3 = VFMACONJ(T2, T1);
Tj = VZMULIJ(Ti, VFNMSCONJ(T2, T1));
T4 = LDW(&(W[TWVL * 4]));
T9 = LDW(&(W[TWVL * 2]));
Te = VZMULIJ(Tb, VFNMSCONJ(Td, Tc));
Th = VZMULJ(Tg, VFMACONJ(Td, Tc));
T7 = VZMULIJ(T4, VFNMSCONJ(T6, T5));
Ta = VZMULJ(T9, VFMACONJ(T6, T5));
Tt = VADD(Tj, Th);
Tk = VSUB(Th, Tj);
Tr = VADD(T3, T7);
T8 = VSUB(T3, T7);
Ts = VADD(Ta, Te);
Tf = VSUB(Ta, Te);
Tx = VMUL(LDK(KP866025403), VSUB(Tt, Ts));
Tu = VADD(Ts, Tt);
To = VMUL(LDK(KP866025403), VSUB(Tk, Tf));
Tl = VADD(Tf, Tk);
Tw = VFNMS(LDK(KP500000000), Tu, Tr);
Tv = VCONJ(VMUL(LDK(KP500000000), VADD(Tr, Tu)));
Tn = VFNMS(LDK(KP500000000), Tl, T8);
Tm = VMUL(LDK(KP500000000), VADD(T8, Tl));
Tz = VMUL(LDK(KP500000000), VFMAI(Tx, Tw));
Ty = VCONJ(VMUL(LDK(KP500000000), VFNMSI(Tx, Tw)));
ST(&(Rm[WS(rs, 2)]), Tv, -ms, &(Rm[0]));
Tp = VMUL(LDK(KP500000000), VFNMSI(To, Tn));
Tq = VCONJ(VMUL(LDK(KP500000000), VFMAI(To, Tn)));
ST(&(Rp[0]), Tm, ms, &(Rp[0]));
ST(&(Rp[WS(rs, 1)]), Tz, ms, &(Rp[WS(rs, 1)]));
ST(&(Rm[0]), Ty, -ms, &(Rm[0]));
ST(&(Rm[WS(rs, 1)]), Tq, -ms, &(Rm[WS(rs, 1)]));
ST(&(Rp[WS(rs, 2)]), Tp, ms, &(Rp[0]));
}
}
VLEAVE();
}
static void hc2cbdftv_6(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
INT m;
for (m = mb, W = W + ((mb - 1) * ((TWVL / VL) * 10)); m < me; m = m + VL, Rp = Rp + (VL * ms), Ip = Ip + (VL * ms), Rm = Rm - (VL * ms), Im = Im - (VL * ms), W = W + (TWVL * 10), MAKE_VOLATILE_STRIDE(rs)) {
V Tv, Tn, Tr, Te, T4, Tg, Ta, Tf, T7, T1, Td, T2, T3, T8, T9;
V T5, T6, Th, Tj, Tb, Tp, Tx, Ti, Tc, To, Tk, Ts, Tq, Tw, Tm;
V Tl, Tu, Tt, Tz, Ty;
T2 = LD(&(Rp[0]), ms, &(Rp[0]));
T3 = LD(&(Rm[WS(rs, 2)]), -ms, &(Rm[0]));
T8 = LD(&(Rp[WS(rs, 1)]), ms, &(Rp[WS(rs, 1)]));
T9 = LD(&(Rm[WS(rs, 1)]), -ms, &(Rm[WS(rs, 1)]));
T5 = LD(&(Rp[WS(rs, 2)]), ms, &(Rp[0]));
T6 = LD(&(Rm[0]), -ms, &(Rm[0]));
Tv = LDW(&(W[0]));
Tn = LDW(&(W[TWVL * 8]));
Tr = LDW(&(W[TWVL * 6]));
Te = VFMACONJ(T3, T2);
T4 = VFNMSCONJ(T3, T2);
Tg = VFMACONJ(T9, T8);
Ta = VFMSCONJ(T9, T8);
Tf = VFMACONJ(T6, T5);
T7 = VFNMSCONJ(T6, T5);
T1 = LDW(&(W[TWVL * 4]));
Td = LDW(&(W[TWVL * 2]));
Th = VADD(Tf, Tg);
Tj = VMUL(LDK(KP866025403), VSUB(Tf, Tg));
Tb = VADD(T7, Ta);
Tp = VMUL(LDK(KP866025403), VSUB(T7, Ta));
Tx = VADD(Te, Th);
Ti = VFNMS(LDK(KP500000000), Th, Te);
Tc = VZMULI(T1, VADD(T4, Tb));
To = VFNMS(LDK(KP500000000), Tb, T4);
Tk = VZMUL(Td, VFNMSI(Tj, Ti));
Ts = VZMUL(Tr, VFMAI(Tj, Ti));
Tq = VZMULI(Tn, VFNMSI(Tp, To));
Tw = VZMULI(Tv, VFMAI(Tp, To));
Tm = VCONJ(VSUB(Tk, Tc));
Tl = VADD(Tc, Tk);
Tu = VCONJ(VSUB(Ts, Tq));
Tt = VADD(Tq, Ts);
Tz = VCONJ(VSUB(Tx, Tw));
Ty = VADD(Tw, Tx);
ST(&(Rm[WS(rs, 1)]), Tm, -ms, &(Rm[WS(rs, 1)]));
ST(&(Rp[WS(rs, 1)]), Tl, ms, &(Rp[WS(rs, 1)]));
ST(&(Rm[WS(rs, 2)]), Tu, -ms, &(Rm[0]));
ST(&(Rp[WS(rs, 2)]), Tt, ms, &(Rp[0]));
ST(&(Rm[0]), Tz, -ms, &(Rm[0]));
ST(&(Rp[0]), Ty, ms, &(Rp[0]));
}
}
static void hc2cfdftv_2(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
INT m;
for (m = mb, W = W + ((mb - 1) * ((TWVL / VL) * 2)); m < me; m = m + VL, Rp = Rp + (VL * ms), Ip = Ip + (VL * ms), Rm = Rm - (VL * ms), Im = Im - (VL * ms), W = W + (TWVL * 2), MAKE_VOLATILE_STRIDE(rs)) {
V T1, T2, T4, T3, T5, T7, T6;
T1 = LD(&(Rp[0]), ms, &(Rp[0]));
T2 = LD(&(Rm[0]), -ms, &(Rm[0]));
T4 = LDW(&(W[0]));
T3 = VFMACONJ(T2, T1);
T5 = VZMULIJ(T4, VFNMSCONJ(T2, T1));
T7 = VCONJ(VMUL(LDK(KP500000000), VADD(T3, T5)));
T6 = VMUL(LDK(KP500000000), VSUB(T3, T5));
ST(&(Rm[0]), T7, -ms, &(Rm[0]));
ST(&(Rp[0]), T6, ms, &(Rp[0]));
}
}
static void hc2cbdftv_2(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms)
{
{
INT m;
for (m = mb, W = W + ((mb - 1) * ((TWVL / VL) * 2)); m < me; m = m + VL, Rp = Rp + (VL * ms), Ip = Ip + (VL * ms), Rm = Rm - (VL * ms), Im = Im - (VL * ms), W = W + (TWVL * 2), MAKE_VOLATILE_STRIDE(8, rs)) {
V T2, T3, T1, T5, T4, T7, T6;
T2 = LD(&(Rp[0]), ms, &(Rp[0]));
T3 = LD(&(Rm[0]), -ms, &(Rm[0]));
T1 = LDW(&(W[0]));
T5 = VFMACONJ(T3, T2);
T4 = VZMULI(T1, VFNMSCONJ(T3, T2));
T7 = VCONJ(VSUB(T5, T4));
T6 = VADD(T4, T5);
ST(&(Rm[0]), T7, -ms, &(Rm[0]));
ST(&(Rp[0]), T6, ms, &(Rp[0]));
}
}
VLEAVE();
}
static void hc2cbdftv_10(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
DVK(KP618033988, +0.618033988749894848204586834365638117720309180);
DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
INT m;
for (m = mb, W = W + ((mb - 1) * ((TWVL / VL) * 18)); m < me; m = m + VL, Rp = Rp + (VL * ms), Ip = Ip + (VL * ms), Rm = Rm - (VL * ms), Im = Im - (VL * ms), W = W + (TWVL * 18), MAKE_VOLATILE_STRIDE(rs)) {
V Ts, T4, TR, T1, TZ, TD, Ty, Tn, Ti, TT, T11, TJ, T15, Tr, TN;
V TE, Tv, To, Tb, T8, Tw, Te, Tx, Th, Tt, T7, T9, T2, T3, Tc;
V Td, Tf, Tg, T5, T6, Tu, Ta;
T2 = LD(&(Rp[0]), ms, &(Rp[0]));
T3 = LD(&(Rm[WS(rs, 4)]), -ms, &(Rm[0]));
Tc = LD(&(Rp[WS(rs, 4)]), ms, &(Rp[0]));
Td = LD(&(Rm[0]), -ms, &(Rm[0]));
Tf = LD(&(Rp[WS(rs, 1)]), ms, &(Rp[WS(rs, 1)]));
Tg = LD(&(Rm[WS(rs, 3)]), -ms, &(Rm[WS(rs, 1)]));
T5 = LD(&(Rp[WS(rs, 2)]), ms, &(Rp[0]));
T6 = LD(&(Rm[WS(rs, 2)]), -ms, &(Rm[0]));
T8 = LD(&(Rp[WS(rs, 3)]), ms, &(Rp[WS(rs, 1)]));
Ts = VFMACONJ(T3, T2);
T4 = VFNMSCONJ(T3, T2);
Tw = VFMACONJ(Td, Tc);
Te = VFNMSCONJ(Td, Tc);
Tx = VFMACONJ(Tg, Tf);
Th = VFMSCONJ(Tg, Tf);
Tt = VFMACONJ(T6, T5);
T7 = VFNMSCONJ(T6, T5);
T9 = LD(&(Rm[WS(rs, 1)]), -ms, &(Rm[WS(rs, 1)]));
TR = LDW(&(W[TWVL * 8]));
T1 = LDW(&(W[TWVL * 4]));
TZ = LDW(&(W[TWVL * 12]));
TD = VSUB(Tw, Tx);
Ty = VADD(Tw, Tx);
Tn = VSUB(Te, Th);
Ti = VADD(Te, Th);
Tu = VFMACONJ(T9, T8);
Ta = VFMSCONJ(T9, T8);
TT = LDW(&(W[TWVL * 6]));
T11 = LDW(&(W[TWVL * 10]));
TJ = LDW(&(W[TWVL * 16]));
T15 = LDW(&(W[0]));
Tr = LDW(&(W[TWVL * 2]));
TN = LDW(&(W[TWVL * 14]));
TE = VSUB(Tt, Tu);
Tv = VADD(Tt, Tu);
To = VSUB(T7, Ta);
Tb = VADD(T7, Ta);
{
V TV, TF, Tz, TB, TL, Tp, Tj, Tl, T17, TA, TS, Tk, TC, TU, TK;
V Tm, TO, TG, T12, TW, T16, TM, T10, Tq, TX, TY, T18, T19, TQ, TP;
V T13, T14, TI, TH;
TV = VMUL(LDK(KP951056516), VFMA(LDK(KP618033988), TD, TE));
TF = VMUL(LDK(KP951056516), VFNMS(LDK(KP618033988), TE, TD));
Tz = VADD(Tv, Ty);
TB = VSUB(Tv, Ty);
TL = VMUL(LDK(KP951056516), VFMA(LDK(KP618033988), Tn, To));
Tp = VMUL(LDK(KP951056516), VFNMS(LDK(KP618033988), To, Tn));
Tj = VADD(Tb, Ti);
Tl = VSUB(Tb, Ti);
T17 = VADD(Ts, Tz);
TA = VFNMS(LDK(KP250000000), Tz, Ts);
TS = VZMULI(TR, VADD(T4, Tj));
Tk = VFNMS(LDK(KP250000000), Tj, T4);
TC = VFNMS(LDK(KP559016994), TB, TA);
TU = VFMA(LDK(KP559016994), TB, TA);
TK = VFMA(LDK(KP559016994), Tl, Tk);
Tm = VFNMS(LDK(KP559016994), Tl, Tk);
TO = VZMUL(TN, VFMAI(TF, TC));
TG = VZMUL(Tr, VFNMSI(TF, TC));
T12 = VZMUL(T11, VFMAI(TV, TU));
TW = VZMUL(TT, VFNMSI(TV, TU));
T16 = VZMULI(T15, VFMAI(TL, TK));
TM = VZMULI(TJ, VFNMSI(TL, TK));
T10 = VZMULI(TZ, VFNMSI(Tp, Tm));
Tq = VZMULI(T1, VFMAI(Tp, Tm));
TX = VADD(TS, TW);
TY = VCONJ(VSUB(TW, TS));
T18 = VADD(T16, T17);
T19 = VCONJ(VSUB(T17, T16));
TQ = VCONJ(VSUB(TO, TM));
TP = VADD(TM, TO);
T13 = VADD(T10, T12);
T14 = VCONJ(VSUB(T12, T10));
TI = VCONJ(VSUB(TG, Tq));
TH = VADD(Tq, TG);
ST(&(Rp[WS(rs, 2)]), TX, ms, &(Rp[0]));
ST(&(Rm[WS(rs, 2)]), TY, -ms, &(Rm[0]));
ST(&(Rp[0]), T18, ms, &(Rp[0]));
ST(&(Rm[0]), T19, -ms, &(Rm[0]));
ST(&(Rm[WS(rs, 4)]), TQ, -ms, &(Rm[0]));
ST(&(Rp[WS(rs, 4)]), TP, ms, &(Rp[0]));
ST(&(Rp[WS(rs, 3)]), T13, ms, &(Rp[WS(rs, 1)]));
ST(&(Rm[WS(rs, 3)]), T14, -ms, &(Rm[WS(rs, 1)]));
ST(&(Rm[WS(rs, 1)]), TI, -ms, &(Rm[WS(rs, 1)]));
ST(&(Rp[WS(rs, 1)]), TH, ms, &(Rp[WS(rs, 1)]));
}
}
}
static void hc2cfdftv_8(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
DVK(KP707106781, +0.707106781186547524400844362104849039284835938);
{
INT m;
for (m = mb, W = W + ((mb - 1) * ((TWVL / VL) * 14)); m < me; m = m + VL, Rp = Rp + (VL * ms), Ip = Ip + (VL * ms), Rm = Rm - (VL * ms), Im = Im - (VL * ms), W = W + (TWVL * 14), MAKE_VOLATILE_STRIDE(32, rs)) {
V T3, Tc, Tl, Ts, Tf, Tg, Te, Tp, T7, Ta, T1, T2, Tb, Tj, Tk;
V Ti, Tr, T5, T6, T4, T9, Th, Tq, TC, T8, Td, TF, Tm, TG, TD;
V Tt, Tu, Tn, TH, TL, TE, TK, Tz, Tv, Ty, To, TJ, TI, TN, TM;
V TB, TA, Tx, Tw;
T1 = LD(&(Rp[0]), ms, &(Rp[0]));
T2 = LD(&(Rm[0]), -ms, &(Rm[0]));
Tb = LDW(&(W[0]));
Tj = LD(&(Rp[WS(rs, 3)]), ms, &(Rp[WS(rs, 1)]));
Tk = LD(&(Rm[WS(rs, 3)]), -ms, &(Rm[WS(rs, 1)]));
Ti = LDW(&(W[TWVL * 12]));
Tr = LDW(&(W[TWVL * 10]));
T5 = LD(&(Rp[WS(rs, 2)]), ms, &(Rp[0]));
T6 = LD(&(Rm[WS(rs, 2)]), -ms, &(Rm[0]));
T3 = VFMACONJ(T2, T1);
Tc = VZMULIJ(Tb, VFNMSCONJ(T2, T1));
T4 = LDW(&(W[TWVL * 6]));
T9 = LDW(&(W[TWVL * 8]));
Tl = VZMULIJ(Ti, VFNMSCONJ(Tk, Tj));
Ts = VZMULJ(Tr, VFMACONJ(Tk, Tj));
Tf = LD(&(Rp[WS(rs, 1)]), ms, &(Rp[WS(rs, 1)]));
Tg = LD(&(Rm[WS(rs, 1)]), -ms, &(Rm[WS(rs, 1)]));
Te = LDW(&(W[TWVL * 4]));
Tp = LDW(&(W[TWVL * 2]));
T7 = VZMULJ(T4, VFMACONJ(T6, T5));
Ta = VZMULIJ(T9, VFNMSCONJ(T6, T5));
Th = VZMULIJ(Te, VFNMSCONJ(Tg, Tf));
Tq = VZMULJ(Tp, VFMACONJ(Tg, Tf));
TC = VADD(T3, T7);
T8 = VSUB(T3, T7);
Td = VSUB(Ta, Tc);
TF = VADD(Tc, Ta);
Tm = VSUB(Th, Tl);
TG = VADD(Th, Tl);
TD = VADD(Tq, Ts);
Tt = VSUB(Tq, Ts);
Tu = VSUB(Tm, Td);
Tn = VADD(Td, Tm);
TH = VSUB(TF, TG);
TL = VADD(TF, TG);
TE = VSUB(TC, TD);
TK = VADD(TC, TD);
Tz = VFMA(LDK(KP707106781), Tu, Tt);
Tv = VFNMS(LDK(KP707106781), Tu, Tt);
Ty = VFNMS(LDK(KP707106781), Tn, T8);
To = VFMA(LDK(KP707106781), Tn, T8);
TJ = VCONJ(VMUL(LDK(KP500000000), VFNMSI(TH, TE)));
TI = VMUL(LDK(KP500000000), VFMAI(TH, TE));
TN = VCONJ(VMUL(LDK(KP500000000), VADD(TL, TK)));
TM = VMUL(LDK(KP500000000), VSUB(TK, TL));
TB = VMUL(LDK(KP500000000), VFMAI(Tz, Ty));
TA = VCONJ(VMUL(LDK(KP500000000), VFNMSI(Tz, Ty)));
Tx = VCONJ(VMUL(LDK(KP500000000), VFMAI(Tv, To)));
Tw = VMUL(LDK(KP500000000), VFNMSI(Tv, To));
ST(&(Rm[WS(rs, 1)]), TJ, -ms, &(Rm[WS(rs, 1)]));
ST(&(Rp[WS(rs, 2)]), TI, ms, &(Rp[0]));
ST(&(Rm[WS(rs, 3)]), TN, -ms, &(Rm[WS(rs, 1)]));
ST(&(Rp[0]), TM, ms, &(Rp[0]));
ST(&(Rp[WS(rs, 3)]), TB, ms, &(Rp[WS(rs, 1)]));
ST(&(Rm[WS(rs, 2)]), TA, -ms, &(Rm[0]));
ST(&(Rm[0]), Tx, -ms, &(Rm[0]));
ST(&(Rp[WS(rs, 1)]), Tw, ms, &(Rp[WS(rs, 1)]));
}
}
VLEAVE();
}
static void hc2cfdftv_12(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
INT m;
for (m = mb, W = W + ((mb - 1) * ((TWVL / VL) * 22)); m < me; m = m + VL, Rp = Rp + (VL * ms), Ip = Ip + (VL * ms), Rm = Rm - (VL * ms), Im = Im - (VL * ms), W = W + (TWVL * 22), MAKE_VOLATILE_STRIDE(rs)) {
V T3, T7, TH, TE, Th, TC, Tq, T11, TU, Tx, Tb, Tz, Tu, Tw, Tp;
V Tl, T9, Ta, T8, Ty, Tn, To, Tm, TG, T1, T2, Tt, T5, T6, T4;
V Tv, Tj, Tk, Ti, TD, Tf, Tg, Te, TB, TT, TF, TR, Tr;
T1 = LD(&(Rp[0]), ms, &(Rp[0]));
T2 = LD(&(Rm[0]), -ms, &(Rm[0]));
Tt = LDW(&(W[0]));
T5 = LD(&(Rp[WS(rs, 2)]), ms, &(Rp[0]));
T6 = LD(&(Rm[WS(rs, 2)]), -ms, &(Rm[0]));
T4 = LDW(&(W[TWVL * 6]));
Tv = LDW(&(W[TWVL * 8]));
Tn = LD(&(Rp[WS(rs, 1)]), ms, &(Rp[WS(rs, 1)]));
To = LD(&(Rm[WS(rs, 1)]), -ms, &(Rm[WS(rs, 1)]));
T3 = VFMACONJ(T2, T1);
Tu = VZMULIJ(Tt, VFNMSCONJ(T2, T1));
Tm = LDW(&(W[TWVL * 2]));
TG = LDW(&(W[TWVL * 4]));
T7 = VZMULJ(T4, VFMACONJ(T6, T5));
Tw = VZMULIJ(Tv, VFNMSCONJ(T6, T5));
Tj = LD(&(Rp[WS(rs, 5)]), ms, &(Rp[WS(rs, 1)]));
Tk = LD(&(Rm[WS(rs, 5)]), -ms, &(Rm[WS(rs, 1)]));
Ti = LDW(&(W[TWVL * 18]));
TD = LDW(&(W[TWVL * 20]));
Tp = VZMULJ(Tm, VFMACONJ(To, Tn));
TH = VZMULIJ(TG, VFNMSCONJ(To, Tn));
Tf = LD(&(Rp[WS(rs, 3)]), ms, &(Rp[WS(rs, 1)]));
Tg = LD(&(Rm[WS(rs, 3)]), -ms, &(Rm[WS(rs, 1)]));
Te = LDW(&(W[TWVL * 10]));
TB = LDW(&(W[TWVL * 12]));
Tl = VZMULJ(Ti, VFMACONJ(Tk, Tj));
TE = VZMULIJ(TD, VFNMSCONJ(Tk, Tj));
T9 = LD(&(Rp[WS(rs, 4)]), ms, &(Rp[0]));
Ta = LD(&(Rm[WS(rs, 4)]), -ms, &(Rm[0]));
T8 = LDW(&(W[TWVL * 14]));
Ty = LDW(&(W[TWVL * 16]));
Th = VZMULJ(Te, VFMACONJ(Tg, Tf));
TC = VZMULIJ(TB, VFNMSCONJ(Tg, Tf));
Tq = VADD(Tl, Tp);
T11 = VSUB(Tp, Tl);
TU = VSUB(Tu, Tw);
Tx = VADD(Tu, Tw);
Tb = VZMULJ(T8, VFMACONJ(Ta, T9));
Tz = VZMULIJ(Ty, VFNMSCONJ(Ta, T9));
TT = VSUB(TC, TE);
TF = VADD(TC, TE);
TR = VFNMS(LDK(KP500000000), Tq, Th);
Tr = VADD(Th, Tq);
{
V TX, TA, T1d, TV, TY, TI, T1e, T12, TQ, Td, T10, Tc, T1a, TN, TJ;
V T1j, T1f, T1b, TS, TM, Ts, T17, T13, TZ, T1i, T1c, T16, TW, TP, TO;
V TL, TK, T1k, T1l, T1h, T1g, T18, T19, T15, T14;
T10 = VSUB(Tb, T7);
Tc = VADD(T7, Tb);
TX = VFNMS(LDK(KP500000000), Tx, Tz);
TA = VADD(Tx, Tz);
T1d = VADD(TU, TT);
TV = VSUB(TT, TU);
TY = VFNMS(LDK(KP500000000), TF, TH);
TI = VADD(TF, TH);
T1e = VADD(T10, T11);
T12 = VSUB(T10, T11);
TQ = VFNMS(LDK(KP500000000), Tc, T3);
Td = VADD(T3, Tc);
T1a = VADD(TX, TY);
TZ = VSUB(TX, TY);
TN = VADD(TA, TI);
TJ = VSUB(TA, TI);
T1j = VMUL(LDK(KP866025403), VADD(T1d, T1e));
T1f = VMUL(LDK(KP866025403), VSUB(T1d, T1e));
T1b = VADD(TQ, TR);
TS = VSUB(TQ, TR);
TM = VADD(Td, Tr);
Ts = VSUB(Td, Tr);
T17 = VFMA(LDK(KP866025403), T12, TZ);
T13 = VFNMS(LDK(KP866025403), T12, TZ);
T1i = VSUB(T1b, T1a);
T1c = VADD(T1a, T1b);
T16 = VFNMS(LDK(KP866025403), TV, TS);
TW = VFMA(LDK(KP866025403), TV, TS);
TP = VCONJ(VMUL(LDK(KP500000000), VADD(TN, TM)));
TO = VMUL(LDK(KP500000000), VSUB(TM, TN));
TL = VCONJ(VMUL(LDK(KP500000000), VFNMSI(TJ, Ts)));
TK = VMUL(LDK(KP500000000), VFMAI(TJ, Ts));
T1k = VCONJ(VMUL(LDK(KP500000000), VFNMSI(T1j, T1i)));
T1l = VMUL(LDK(KP500000000), VFMAI(T1j, T1i));
T1h = VMUL(LDK(KP500000000), VFMAI(T1f, T1c));
T1g = VCONJ(VMUL(LDK(KP500000000), VFNMSI(T1f, T1c)));
T18 = VMUL(LDK(KP500000000), VFNMSI(T17, T16));
T19 = VCONJ(VMUL(LDK(KP500000000), VFMAI(T17, T16)));
T15 = VCONJ(VMUL(LDK(KP500000000), VFMAI(T13, TW)));
T14 = VMUL(LDK(KP500000000), VFNMSI(T13, TW));
ST(&(Rm[WS(rs, 5)]), TP, -ms, &(Rm[WS(rs, 1)]));
ST(&(Rp[0]), TO, ms, &(Rp[0]));
ST(&(Rm[WS(rs, 2)]), TL, -ms, &(Rm[0]));
ST(&(Rp[WS(rs, 3)]), TK, ms, &(Rp[WS(rs, 1)]));
ST(&(Rm[WS(rs, 3)]), T1k, -ms, &(Rm[WS(rs, 1)]));
ST(&(Rp[WS(rs, 4)]), T1l, ms, &(Rp[0]));
ST(&(Rp[WS(rs, 2)]), T1h, ms, &(Rp[0]));
ST(&(Rm[WS(rs, 1)]), T1g, -ms, &(Rm[WS(rs, 1)]));
ST(&(Rp[WS(rs, 5)]), T18, ms, &(Rp[WS(rs, 1)]));
ST(&(Rm[WS(rs, 4)]), T19, -ms, &(Rm[0]));
ST(&(Rm[0]), T15, -ms, &(Rm[0]));
ST(&(Rp[WS(rs, 1)]), T14, ms, &(Rp[WS(rs, 1)]));
}
}
}
static void hc2cbdftv_16(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP923879532, +0.923879532511286756128183189396788286822416626);
DVK(KP414213562, +0.414213562373095048801688724209698078569671875);
DVK(KP707106781, +0.707106781186547524400844362104849039284835938);
{
INT m;
for (m = mb, W = W + ((mb - 1) * ((TWVL / VL) * 30)); m < me; m = m + VL, Rp = Rp + (VL * ms), Ip = Ip + (VL * ms), Rm = Rm - (VL * ms), Im = Im - (VL * ms), W = W + (TWVL * 30), MAKE_VOLATILE_STRIDE(64, rs)) {
V T8, Tv, TE, T1t, TP, T1w, T10, T1p, Tn, Tw, T13, T1q, TL, T1x, TS;
V T1u;
{
V T4, TA, Tu, TC, T7, TN, Tr, TB, T2, T3, Ts, Tt, T5, T6, Tp;
V Tq, TD, TO, TY, TZ, Tb, TF, Tl, TJ, Te, TG, Ti, TI, T9, Ta;
V Tj, Tk, Tc, Td, Tg, Th, Tf, Tm, T11, T12, TH, TK, TQ, TR;
T2 = LD(&(Rp[0]), ms, &(Rp[0]));
T3 = LD(&(Rm[WS(rs, 7)]), -ms, &(Rm[WS(rs, 1)]));
T4 = VFMACONJ(T3, T2);
TA = VFNMSCONJ(T3, T2);
Ts = LD(&(Rp[WS(rs, 6)]), ms, &(Rp[0]));
Tt = LD(&(Rm[WS(rs, 1)]), -ms, &(Rm[WS(rs, 1)]));
Tu = VFMACONJ(Tt, Ts);
TC = VFMSCONJ(Tt, Ts);
T5 = LD(&(Rp[WS(rs, 4)]), ms, &(Rp[0]));
T6 = LD(&(Rm[WS(rs, 3)]), -ms, &(Rm[WS(rs, 1)]));
T7 = VFMACONJ(T6, T5);
TN = VFNMSCONJ(T6, T5);
Tp = LD(&(Rp[WS(rs, 2)]), ms, &(Rp[0]));
Tq = LD(&(Rm[WS(rs, 5)]), -ms, &(Rm[WS(rs, 1)]));
Tr = VFMACONJ(Tq, Tp);
TB = VFNMSCONJ(Tq, Tp);
T8 = VSUB(T4, T7);
Tv = VSUB(Tr, Tu);
TD = VADD(TB, TC);
TE = VFMA(LDK(KP707106781), TD, TA);
T1t = VFNMS(LDK(KP707106781), TD, TA);
TO = VSUB(TB, TC);
TP = VFMA(LDK(KP707106781), TO, TN);
T1w = VFNMS(LDK(KP707106781), TO, TN);
TY = VADD(T4, T7);
TZ = VADD(Tr, Tu);
T10 = VADD(TY, TZ);
T1p = VSUB(TY, TZ);
T9 = LD(&(Rp[WS(rs, 1)]), ms, &(Rp[WS(rs, 1)]));
Ta = LD(&(Rm[WS(rs, 6)]), -ms, &(Rm[0]));
Tb = VFMACONJ(Ta, T9);
TF = VFNMSCONJ(Ta, T9);
Tj = LD(&(Rp[WS(rs, 3)]), ms, &(Rp[WS(rs, 1)]));
Tk = LD(&(Rm[WS(rs, 4)]), -ms, &(Rm[0]));
Tl = VFMACONJ(Tk, Tj);
TJ = VFNMSCONJ(Tk, Tj);
Tc = LD(&(Rp[WS(rs, 5)]), ms, &(Rp[WS(rs, 1)]));
Td = LD(&(Rm[WS(rs, 2)]), -ms, &(Rm[0]));
Te = VFMACONJ(Td, Tc);
TG = VFNMSCONJ(Td, Tc);
Tg = LD(&(Rp[WS(rs, 7)]), ms, &(Rp[WS(rs, 1)]));
Th = LD(&(Rm[0]), -ms, &(Rm[0]));
Ti = VFMACONJ(Th, Tg);
TI = VFMSCONJ(Th, Tg);
Tf = VSUB(Tb, Te);
Tm = VSUB(Ti, Tl);
Tn = VADD(Tf, Tm);
Tw = VSUB(Tf, Tm);
T11 = VADD(Tb, Te);
T12 = VADD(Ti, Tl);
T13 = VADD(T11, T12);
T1q = VSUB(T11, T12);
TH = VFNMS(LDK(KP414213562), TG, TF);
TK = VFMA(LDK(KP414213562), TJ, TI);
TL = VADD(TH, TK);
T1x = VSUB(TH, TK);
TQ = VFMA(LDK(KP414213562), TF, TG);
TR = VFNMS(LDK(KP414213562), TI, TJ);
TS = VADD(TQ, TR);
T1u = VSUB(TQ, TR);
}
{
V T1j, T1R, T1c, T1J, T1g, T1l, T1N, T1T, T1Q, T1a, T1b, T19, T1I, T1e, T1f;
V T1d, T1k, T1L, T1M, T1K, T1S, T1h, T1U, T1V, T1i, T1m, T1O, T1P, T1n, T14;
V T1r, Ty, T1D, TU, T16, T1z, T1F, TX, T1o, To, Tx, T1, T1C, TM, TT;
V Tz, T15, T1v, T1y, T1s, T1E, TV, T1G, T1H, TW, T17, T1A, T1B, T18;
T1j = VADD(T10, T13);
T1Q = LDW(&(W[TWVL * 22]));
T1R = VZMUL(T1Q, VFNMSI(T1q, T1p));
T1a = VFMA(LDK(KP707106781), Tn, T8);
T1b = VFMA(LDK(KP707106781), Tw, Tv);
T19 = LDW(&(W[TWVL * 26]));
T1c = VZMUL(T19, VFNMSI(T1b, T1a));
T1I = LDW(&(W[TWVL * 2]));
T1J = VZMUL(T1I, VFMAI(T1b, T1a));
T1e = VFMA(LDK(KP923879532), TL, TE);
T1f = VFMA(LDK(KP923879532), TS, TP);
T1d = LDW(&(W[TWVL * 28]));
T1g = VZMULI(T1d, VFNMSI(T1f, T1e));
T1k = LDW(&(W[0]));
T1l = VZMULI(T1k, VFMAI(T1f, T1e));
T1L = VFMA(LDK(KP923879532), T1u, T1t);
T1M = VFNMS(LDK(KP923879532), T1x, T1w);
T1K = LDW(&(W[TWVL * 4]));
T1N = VZMULI(T1K, VFNMSI(T1M, T1L));
T1S = LDW(&(W[TWVL * 24]));
T1T = VZMULI(T1S, VFMAI(T1M, T1L));
T1h = VCONJ(VSUB(T1c, T1g));
ST(&(Rm[WS(rs, 7)]), T1h, -ms, &(Rm[WS(rs, 1)]));
T1U = VCONJ(VSUB(T1R, T1T));
ST(&(Rm[WS(rs, 6)]), T1U, -ms, &(Rm[0]));
T1V = VADD(T1R, T1T);
ST(&(Rp[WS(rs, 6)]), T1V, ms, &(Rp[0]));
T1i = VADD(T1c, T1g);
ST(&(Rp[WS(rs, 7)]), T1i, ms, &(Rp[WS(rs, 1)]));
T1m = VCONJ(VSUB(T1j, T1l));
ST(&(Rm[0]), T1m, -ms, &(Rm[0]));
T1O = VCONJ(VSUB(T1J, T1N));
ST(&(Rm[WS(rs, 1)]), T1O, -ms, &(Rm[WS(rs, 1)]));
T1P = VADD(T1J, T1N);
ST(&(Rp[WS(rs, 1)]), T1P, ms, &(Rp[WS(rs, 1)]));
T1n = VADD(T1j, T1l);
ST(&(Rp[0]), T1n, ms, &(Rp[0]));
TX = LDW(&(W[TWVL * 14]));
T14 = VZMUL(TX, VSUB(T10, T13));
T1o = LDW(&(W[TWVL * 6]));
T1r = VZMUL(T1o, VFMAI(T1q, T1p));
To = VFNMS(LDK(KP707106781), Tn, T8);
Tx = VFNMS(LDK(KP707106781), Tw, Tv);
T1 = LDW(&(W[TWVL * 10]));
Ty = VZMUL(T1, VFNMSI(Tx, To));
T1C = LDW(&(W[TWVL * 18]));
T1D = VZMUL(T1C, VFMAI(Tx, To));
TM = VFNMS(LDK(KP923879532), TL, TE);
TT = VFNMS(LDK(KP923879532), TS, TP);
Tz = LDW(&(W[TWVL * 12]));
TU = VZMULI(Tz, VFNMSI(TT, TM));
T15 = LDW(&(W[TWVL * 16]));
T16 = VZMULI(T15, VFMAI(TT, TM));
T1v = VFNMS(LDK(KP923879532), T1u, T1t);
T1y = VFMA(LDK(KP923879532), T1x, T1w);
T1s = LDW(&(W[TWVL * 8]));
T1z = VZMULI(T1s, VFMAI(T1y, T1v));
T1E = LDW(&(W[TWVL * 20]));
T1F = VZMULI(T1E, VFNMSI(T1y, T1v));
TV = VCONJ(VSUB(Ty, TU));
ST(&(Rm[WS(rs, 3)]), TV, -ms, &(Rm[WS(rs, 1)]));
T1G = VCONJ(VSUB(T1D, T1F));
ST(&(Rm[WS(rs, 5)]), T1G, -ms, &(Rm[WS(rs, 1)]));
T1H = VADD(T1D, T1F);
ST(&(Rp[WS(rs, 5)]), T1H, ms, &(Rp[WS(rs, 1)]));
TW = VADD(Ty, TU);
ST(&(Rp[WS(rs, 3)]), TW, ms, &(Rp[WS(rs, 1)]));
T17 = VCONJ(VSUB(T14, T16));
ST(&(Rm[WS(rs, 4)]), T17, -ms, &(Rm[0]));
T1A = VCONJ(VSUB(T1r, T1z));
ST(&(Rm[WS(rs, 2)]), T1A, -ms, &(Rm[0]));
T1B = VADD(T1r, T1z);
ST(&(Rp[WS(rs, 2)]), T1B, ms, &(Rp[0]));
T18 = VADD(T14, T16);
ST(&(Rp[WS(rs, 4)]), T18, ms, &(Rp[0]));
}
}
}
VLEAVE();
}
static void hc2cbdftv_8(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP707106781, +0.707106781186547524400844362104849039284835938);
INT m;
for (m = mb, W = W + ((mb - 1) * ((TWVL / VL) * 14)); m < me; m = m + VL, Rp = Rp + (VL * ms), Ip = Ip + (VL * ms), Rm = Rm - (VL * ms), Im = Im - (VL * ms), W = W + (TWVL * 14), MAKE_VOLATILE_STRIDE(rs)) {
V TJ, T4, Tf, TB, TD, TE, Tm, T1, Tj, TF, Tp, Tb, Tg, Tt, Tx;
V T2, T3, Td, Te, T5, T6, T8, T9, Tn, T7, To, Ta, Tk, Tl, TG;
V TL, Tq, Tc, Tu, Th, Tv, Ty, Tw, TC, Ti, TK, TA, Tz, TI, TH;
V Ts, Tr, TN, TM;
T2 = LD(&(Rp[0]), ms, &(Rp[0]));
T3 = LD(&(Rm[WS(rs, 3)]), -ms, &(Rm[WS(rs, 1)]));
Td = LD(&(Rp[WS(rs, 2)]), ms, &(Rp[0]));
Te = LD(&(Rm[WS(rs, 1)]), -ms, &(Rm[WS(rs, 1)]));
T5 = LD(&(Rp[WS(rs, 1)]), ms, &(Rp[WS(rs, 1)]));
T6 = LD(&(Rm[WS(rs, 2)]), -ms, &(Rm[0]));
T8 = LD(&(Rp[WS(rs, 3)]), ms, &(Rp[WS(rs, 1)]));
T9 = LD(&(Rm[0]), -ms, &(Rm[0]));
TJ = LDW(&(W[0]));
Tk = VFMACONJ(T3, T2);
T4 = VFNMSCONJ(T3, T2);
Tl = VFMACONJ(Te, Td);
Tf = VFNMSCONJ(Te, Td);
Tn = VFMACONJ(T6, T5);
T7 = VFNMSCONJ(T6, T5);
To = VFMACONJ(T9, T8);
Ta = VFMSCONJ(T9, T8);
TB = LDW(&(W[TWVL * 8]));
TD = LDW(&(W[TWVL * 6]));
TE = VADD(Tk, Tl);
Tm = VSUB(Tk, Tl);
T1 = LDW(&(W[TWVL * 12]));
Tj = LDW(&(W[TWVL * 10]));
TF = VADD(Tn, To);
Tp = VSUB(Tn, To);
Tb = VADD(T7, Ta);
Tg = VSUB(T7, Ta);
Tt = LDW(&(W[TWVL * 4]));
Tx = LDW(&(W[TWVL * 2]));
TG = VZMUL(TD, VSUB(TE, TF));
TL = VADD(TE, TF);
Tq = VZMUL(Tj, VFNMSI(Tp, Tm));
Tc = VFMA(LDK(KP707106781), Tb, T4);
Tu = VFNMS(LDK(KP707106781), Tb, T4);
Th = VFMA(LDK(KP707106781), Tg, Tf);
Tv = VFNMS(LDK(KP707106781), Tg, Tf);
Ty = VZMUL(Tx, VFMAI(Tp, Tm));
Tw = VZMULI(Tt, VFNMSI(Tv, Tu));
TC = VZMULI(TB, VFMAI(Tv, Tu));
Ti = VZMULI(T1, VFNMSI(Th, Tc));
TK = VZMULI(TJ, VFMAI(Th, Tc));
TA = VCONJ(VSUB(Ty, Tw));
Tz = VADD(Tw, Ty);
TI = VCONJ(VSUB(TG, TC));
TH = VADD(TC, TG);
Ts = VCONJ(VSUB(Tq, Ti));
Tr = VADD(Ti, Tq);
TN = VCONJ(VSUB(TL, TK));
TM = VADD(TK, TL);
ST(&(Rm[WS(rs, 1)]), TA, -ms, &(Rm[WS(rs, 1)]));
ST(&(Rp[WS(rs, 1)]), Tz, ms, &(Rp[WS(rs, 1)]));
ST(&(Rm[WS(rs, 2)]), TI, -ms, &(Rm[0]));
ST(&(Rp[WS(rs, 2)]), TH, ms, &(Rp[0]));
ST(&(Rm[WS(rs, 3)]), Ts, -ms, &(Rm[WS(rs, 1)]));
ST(&(Rp[WS(rs, 3)]), Tr, ms, &(Rp[WS(rs, 1)]));
ST(&(Rm[0]), TN, -ms, &(Rm[0]));
ST(&(Rp[0]), TM, ms, &(Rp[0]));
}
}
static void hc2cfdftv_10(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
DVK(KP618033988, +0.618033988749894848204586834365638117720309180);
DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
{
INT m;
for (m = mb, W = W + ((mb - 1) * ((TWVL / VL) * 18)); m < me; m = m + VL, Rp = Rp + (VL * ms), Ip = Ip + (VL * ms), Rm = Rm - (VL * ms), Im = Im - (VL * ms), W = W + (TWVL * 18), MAKE_VOLATILE_STRIDE(40, rs)) {
V T5, T6, Tw, Tr, Tc, Tj, Tl, Tm, Tk, Ts, Tg, Ty, T3, T4, T1;
V T2, Tv, Tq, Ta, Tb, T9, Ti, Te, Tf, Td, Tx, Tn, Tt, Th, TQ;
V TT, Tz, T7, TR, To, Tu, TU;
T1 = LD(&(Rp[0]), ms, &(Rp[0]));
T2 = LD(&(Rm[0]), -ms, &(Rm[0]));
Tv = LDW(&(W[0]));
T5 = LD(&(Rp[WS(rs, 2)]), ms, &(Rp[0]));
T6 = LD(&(Rm[WS(rs, 2)]), -ms, &(Rm[0]));
Tq = LDW(&(W[TWVL * 6]));
Ta = LD(&(Rp[WS(rs, 1)]), ms, &(Rp[WS(rs, 1)]));
Tb = LD(&(Rm[WS(rs, 1)]), -ms, &(Rm[WS(rs, 1)]));
T9 = LDW(&(W[TWVL * 2]));
Ti = LDW(&(W[TWVL * 4]));
Tw = VZMULIJ(Tv, VFNMSCONJ(T2, T1));
Te = LD(&(Rp[WS(rs, 3)]), ms, &(Rp[WS(rs, 1)]));
Tf = LD(&(Rm[WS(rs, 3)]), -ms, &(Rm[WS(rs, 1)]));
Tr = VZMULJ(Tq, VFMACONJ(T6, T5));
Td = LDW(&(W[TWVL * 12]));
Tx = LDW(&(W[TWVL * 10]));
Tc = VZMULJ(T9, VFMACONJ(Tb, Ta));
Tj = VZMULIJ(Ti, VFNMSCONJ(Tb, Ta));
Tl = LD(&(Rp[WS(rs, 4)]), ms, &(Rp[0]));
Tm = LD(&(Rm[WS(rs, 4)]), -ms, &(Rm[0]));
Tk = LDW(&(W[TWVL * 14]));
Ts = LDW(&(W[TWVL * 16]));
Tg = VZMULIJ(Td, VFNMSCONJ(Tf, Te));
Ty = VZMULJ(Tx, VFMACONJ(Tf, Te));
T3 = VFMACONJ(T2, T1);
T4 = LDW(&(W[TWVL * 8]));
Tn = VZMULJ(Tk, VFMACONJ(Tm, Tl));
Tt = VZMULIJ(Ts, VFNMSCONJ(Tm, Tl));
Th = VSUB(Tc, Tg);
TQ = VADD(Tc, Tg);
TT = VADD(Tw, Ty);
Tz = VSUB(Tw, Ty);
T7 = VZMULIJ(T4, VFNMSCONJ(T6, T5));
TR = VADD(Tj, Tn);
To = VSUB(Tj, Tn);
Tu = VSUB(Tr, Tt);
TU = VADD(Tr, Tt);
{
V TP, T8, TS, T11, Tp, TH, TA, TG, TV, T12, TE, TB, TM, TI, TZ;
V TW, T17, T13, TD, TC, TY, TX, TL, TF, T10, T16, TN, TO, TK, TJ;
V T18, T19, T15, T14;
TP = VADD(T3, T7);
T8 = VSUB(T3, T7);
TS = VADD(TQ, TR);
T11 = VSUB(TQ, TR);
Tp = VSUB(Th, To);
TH = VADD(Th, To);
TA = VSUB(Tu, Tz);
TG = VADD(Tz, Tu);
TV = VADD(TT, TU);
T12 = VSUB(TU, TT);
TE = VSUB(Tp, TA);
TB = VADD(Tp, TA);
TM = VMUL(LDK(KP951056516), VFMA(LDK(KP618033988), TG, TH));
TI = VMUL(LDK(KP951056516), VFNMS(LDK(KP618033988), TH, TG));
TZ = VSUB(TS, TV);
TW = VADD(TS, TV);
T17 = VMUL(LDK(KP951056516), VFNMS(LDK(KP618033988), T11, T12));
T13 = VMUL(LDK(KP951056516), VFMA(LDK(KP618033988), T12, T11));
TD = VFNMS(LDK(KP250000000), TB, T8);
TC = VMUL(LDK(KP500000000), VADD(T8, TB));
TY = VFNMS(LDK(KP250000000), TW, TP);
TX = VCONJ(VMUL(LDK(KP500000000), VADD(TP, TW)));
TL = VFMA(LDK(KP559016994), TE, TD);
TF = VFNMS(LDK(KP559016994), TE, TD);
ST(&(Rp[0]), TC, ms, &(Rp[0]));
T10 = VFMA(LDK(KP559016994), TZ, TY);
T16 = VFNMS(LDK(KP559016994), TZ, TY);
ST(&(Rm[WS(rs, 4)]), TX, -ms, &(Rm[0]));
TN = VCONJ(VMUL(LDK(KP500000000), VFNMSI(TM, TL)));
TO = VMUL(LDK(KP500000000), VFMAI(TM, TL));
TK = VMUL(LDK(KP500000000), VFMAI(TI, TF));
TJ = VCONJ(VMUL(LDK(KP500000000), VFNMSI(TI, TF)));
T18 = VMUL(LDK(KP500000000), VFNMSI(T17, T16));
T19 = VCONJ(VMUL(LDK(KP500000000), VFMAI(T17, T16)));
T15 = VCONJ(VMUL(LDK(KP500000000), VFMAI(T13, T10)));
T14 = VMUL(LDK(KP500000000), VFNMSI(T13, T10));
ST(&(Rm[WS(rs, 3)]), TN, -ms, &(Rm[WS(rs, 1)]));
ST(&(Rp[WS(rs, 4)]), TO, ms, &(Rp[0]));
ST(&(Rp[WS(rs, 2)]), TK, ms, &(Rp[0]));
ST(&(Rm[WS(rs, 1)]), TJ, -ms, &(Rm[WS(rs, 1)]));
ST(&(Rp[WS(rs, 3)]), T18, ms, &(Rp[WS(rs, 1)]));
ST(&(Rm[WS(rs, 2)]), T19, -ms, &(Rm[0]));
ST(&(Rm[0]), T15, -ms, &(Rm[0]));
ST(&(Rp[WS(rs, 1)]), T14, ms, &(Rp[WS(rs, 1)]));
}
}
}
VLEAVE();
}
