static void t3fv_4(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
INT m;
R *x;
x = ri;
for (m = mb, W = W + (mb * ((TWVL / VL) * 4)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 4), MAKE_VOLATILE_STRIDE(rs)) {
V T2, T3, T4;
T2 = LDW(&(W[0]));
T3 = LDW(&(W[TWVL * 2]));
T4 = VZMULJ(T2, T3);
{
V T1, Tb, T6, T9, Ta, T5, T8;
T1 = LD(&(x[0]), ms, &(x[0]));
Ta = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
Tb = VZMULJ(T3, Ta);
T5 = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
T6 = VZMULJ(T4, T5);
T8 = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
T9 = VZMULJ(T2, T8);
{
V T7, Tc, Td, Te;
T7 = VSUB(T1, T6);
Tc = VBYI(VSUB(T9, Tb));
ST(&(x[WS(rs, 1)]), VSUB(T7, Tc), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 3)]), VADD(T7, Tc), ms, &(x[WS(rs, 1)]));
Td = VADD(T1, T6);
Te = VADD(T9, Tb);
ST(&(x[WS(rs, 2)]), VSUB(Td, Te), ms, &(x[0]));
ST(&(x[0]), VADD(Td, Te), ms, &(x[0]));
}
}
}
}
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(16, 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)]));
}
}
VLEAVE();
}
static void t1sv_2(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
{
INT m;
for (m = mb, W = W + (mb * 2); m < me; m = m + (2 * VL), ri = ri + ((2 * VL) * ms), ii = ii + ((2 * VL) * ms), W = W + ((2 * VL) * 2), MAKE_VOLATILE_STRIDE(4, rs)) {
V T1, T8, T6, T7;
T1 = LD(&(ri[0]), ms, &(ri[0]));
T8 = LD(&(ii[0]), ms, &(ii[0]));
{
V T3, T5, T2, T4;
T3 = LD(&(ri[WS(rs, 1)]), ms, &(ri[WS(rs, 1)]));
T5 = LD(&(ii[WS(rs, 1)]), ms, &(ii[WS(rs, 1)]));
T2 = LDW(&(W[0]));
T4 = LDW(&(W[TWVL * 1]));
T6 = VFMA(T2, T3, VMUL(T4, T5));
T7 = VFNMS(T4, T3, VMUL(T2, T5));
}
ST(&(ri[WS(rs, 1)]), VSUB(T1, T6), ms, &(ri[WS(rs, 1)]));
ST(&(ii[WS(rs, 1)]), VSUB(T8, T7), ms, &(ii[WS(rs, 1)]));
ST(&(ri[0]), VADD(T1, T6), ms, &(ri[0]));
ST(&(ii[0]), VADD(T7, T8), ms, &(ii[0]));
}
}
VLEAVE();
}
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 T4, Tc, T9, Te, T1, T3, T2, Tb, T6, T8, T7, T5, Td, Tg, Th;
V Ta, Tf, Tk, Tl, Ti, Tj;
T1 = LD(&(Rp[0]), ms, &(Rp[0]));
T2 = LD(&(Rm[0]), -ms, &(Rm[0]));
T3 = VCONJ(T2);
T4 = VADD(T1, T3);
Tb = LDW(&(W[0]));
Tc = VZMULIJ(Tb, VSUB(T3, T1));
T6 = LD(&(Rp[WS(rs, 1)]), ms, &(Rp[WS(rs, 1)]));
T7 = LD(&(Rm[WS(rs, 1)]), -ms, &(Rm[WS(rs, 1)]));
T8 = VCONJ(T7);
T5 = LDW(&(W[TWVL * 2]));
T9 = VZMULJ(T5, VADD(T6, T8));
Td = LDW(&(W[TWVL * 4]));
Te = VZMULIJ(Td, VSUB(T8, T6));
Ta = VSUB(T4, T9);
Tf = VBYI(VSUB(Tc, Te));
Tg = VMUL(LDK(KP500000000), VSUB(Ta, Tf));
Th = VCONJ(VMUL(LDK(KP500000000), VADD(Ta, Tf)));
ST(&(Rp[WS(rs, 1)]), Tg, ms, &(Rp[WS(rs, 1)]));
ST(&(Rm[0]), Th, -ms, &(Rm[0]));
Ti = VADD(T4, T9);
Tj = VADD(Tc, Te);
Tk = VCONJ(VMUL(LDK(KP500000000), VSUB(Ti, Tj)));
Tl = VMUL(LDK(KP500000000), VADD(Ti, Tj));
ST(&(Rm[WS(rs, 1)]), Tk, -ms, &(Rm[WS(rs, 1)]));
ST(&(Rp[0]), Tl, ms, &(Rp[0]));
}
}
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 const R *t1sv_2(R *ri, R *ii, const R *W, stride ios, INT m, INT dist)
{
INT i;
for (i = m; i > 0; i = i - (2 * VL), ri = ri + ((2 * VL) * dist), ii = ii + ((2 * VL) * dist), W = W + ((2 * VL) * 2), MAKE_VOLATILE_STRIDE(ios)) {
V T1, Ta, T3, T6, T2, T5;
T1 = LD(&(ri[0]), dist, &(ri[0]));
Ta = LD(&(ii[0]), dist, &(ii[0]));
T3 = LD(&(ri[WS(ios, 1)]), dist, &(ri[WS(ios, 1)]));
T6 = LD(&(ii[WS(ios, 1)]), dist, &(ii[WS(ios, 1)]));
T2 = LDW(&(W[0]));
T5 = LDW(&(W[TWVL * 1]));
{
V T8, T4, T9, T7;
T8 = VMUL(T2, T6);
T4 = VMUL(T2, T3);
T9 = VFNMS(T5, T3, T8);
T7 = VFMA(T5, T6, T4);
ST(&(ii[0]), VADD(T9, Ta), dist, &(ii[0]));
ST(&(ii[WS(ios, 1)]), VSUB(Ta, T9), dist, &(ii[WS(ios, 1)]));
ST(&(ri[0]), VADD(T1, T7), dist, &(ri[0]));
ST(&(ri[WS(ios, 1)]), VSUB(T1, T7), dist, &(ri[WS(ios, 1)]));
}
}
return W;
}
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 T5, Tc, T9, Td, T2, T4, T3, T6, T8, T7, Tj, Ti, Th, Tk, Tl;
V Ta, Te, T1, Tb, Tf, Tg;
T2 = LD(&(Rp[0]), ms, &(Rp[0]));
T3 = LD(&(Rm[WS(rs, 1)]), -ms, &(Rm[WS(rs, 1)]));
T4 = VCONJ(T3);
T5 = VSUB(T2, T4);
Tc = VADD(T2, T4);
T6 = LD(&(Rp[WS(rs, 1)]), ms, &(Rp[WS(rs, 1)]));
T7 = LD(&(Rm[0]), -ms, &(Rm[0]));
T8 = VCONJ(T7);
T9 = VBYI(VSUB(T6, T8));
Td = VADD(T6, T8);
Tj = VADD(Tc, Td);
Th = LDW(&(W[0]));
Ti = VZMULI(Th, VADD(T5, T9));
Tk = VADD(Ti, Tj);
ST(&(Rp[0]), Tk, ms, &(Rp[0]));
Tl = VCONJ(VSUB(Tj, Ti));
ST(&(Rm[0]), Tl, -ms, &(Rm[0]));
T1 = LDW(&(W[TWVL * 4]));
Ta = VZMULI(T1, VSUB(T5, T9));
Tb = LDW(&(W[TWVL * 2]));
Te = VZMUL(Tb, VSUB(Tc, Td));
Tf = VADD(Ta, Te);
ST(&(Rp[WS(rs, 1)]), Tf, ms, &(Rp[WS(rs, 1)]));
Tg = VCONJ(VSUB(Te, Ta));
ST(&(Rm[WS(rs, 1)]), Tg, -ms, &(Rm[WS(rs, 1)]));
}
}
static void hc2cfdftv_6(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
{
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 Ta, Tu, Tn, Tw, Ti, Tv, T1, T8, Tg, Tf, T7, T3, Te, T6, T2;
V T4, T9, T5, Tk, Tm, Tj, Tl, Tc, Th, Tb, Td, Tr, Tp, Tq, To;
V Tt, Ts, TA, Ty, Tz, Tx, TC, TB;
T1 = LD(&(Rp[0]), ms, &(Rp[0]));
T8 = LD(&(Rp[WS(rs, 1)]), ms, &(Rp[WS(rs, 1)]));
Tg = LD(&(Rp[WS(rs, 2)]), ms, &(Rp[0]));
Te = LD(&(Rm[WS(rs, 2)]), -ms, &(Rm[0]));
Tf = VCONJ(Te);
T6 = LD(&(Rm[WS(rs, 1)]), -ms, &(Rm[WS(rs, 1)]));
T7 = VCONJ(T6);
T2 = LD(&(Rm[0]), -ms, &(Rm[0]));
T3 = VCONJ(T2);
T4 = VADD(T1, T3);
T5 = LDW(&(W[TWVL * 4]));
T9 = VZMULIJ(T5, VSUB(T7, T8));
Ta = VADD(T4, T9);
Tu = VSUB(T4, T9);
Tj = LDW(&(W[0]));
Tk = VZMULIJ(Tj, VSUB(T3, T1));
Tl = LDW(&(W[TWVL * 6]));
Tm = VZMULJ(Tl, VADD(Tf, Tg));
Tn = VADD(Tk, Tm);
Tw = VSUB(Tm, Tk);
Tb = LDW(&(W[TWVL * 2]));
Tc = VZMULJ(Tb, VADD(T7, T8));
Td = LDW(&(W[TWVL * 8]));
Th = VZMULIJ(Td, VSUB(Tf, Tg));
Ti = VADD(Tc, Th);
Tv = VSUB(Tc, Th);
Tr = VMUL(LDK(KP500000000), VBYI(VMUL(LDK(KP866025403), VSUB(Tn, Ti))));
To = VADD(Ti, Tn);
Tp = VMUL(LDK(KP500000000), VADD(Ta, To));
Tq = VFNMS(LDK(KP250000000), To, VMUL(LDK(KP500000000), Ta));
ST(&(Rp[0]), Tp, ms, &(Rp[0]));
Tt = VCONJ(VADD(Tq, Tr));
ST(&(Rm[WS(rs, 1)]), Tt, -ms, &(Rm[WS(rs, 1)]));
Ts = VSUB(Tq, Tr);
ST(&(Rp[WS(rs, 2)]), Ts, ms, &(Rp[0]));
TA = VMUL(LDK(KP500000000), VBYI(VMUL(LDK(KP866025403), VSUB(Tw, Tv))));
Tx = VADD(Tv, Tw);
Ty = VCONJ(VMUL(LDK(KP500000000), VADD(Tu, Tx)));
Tz = VFNMS(LDK(KP250000000), Tx, VMUL(LDK(KP500000000), Tu));
ST(&(Rm[WS(rs, 2)]), Ty, -ms, &(Rm[0]));
TC = VADD(Tz, TA);
ST(&(Rp[WS(rs, 1)]), TC, ms, &(Rp[WS(rs, 1)]));
TB = VCONJ(VSUB(Tz, TA));
ST(&(Rm[0]), TB, -ms, &(Rm[0]));
}
}
VLEAVE();
}
static void t2sv_4(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
{
INT m;
for (m = mb, W = W + (mb * 4); m < me; m = m + (2 * VL), ri = ri + ((2 * VL) * ms), ii = ii + ((2 * VL) * ms), W = W + ((2 * VL) * 4), MAKE_VOLATILE_STRIDE(8, rs)) {
V T2, T4, T3, T5, T6, T8;
T2 = LDW(&(W[0]));
T4 = LDW(&(W[TWVL * 1]));
T3 = LDW(&(W[TWVL * 2]));
T5 = LDW(&(W[TWVL * 3]));
T6 = VFMA(T2, T3, VMUL(T4, T5));
T8 = VFNMS(T4, T3, VMUL(T2, T5));
{
V T1, Tp, Ta, To, Te, Tk, Th, Tl, T7, T9;
T1 = LD(&(ri[0]), ms, &(ri[0]));
Tp = LD(&(ii[0]), ms, &(ii[0]));
T7 = LD(&(ri[WS(rs, 2)]), ms, &(ri[0]));
T9 = LD(&(ii[WS(rs, 2)]), ms, &(ii[0]));
Ta = VFMA(T6, T7, VMUL(T8, T9));
To = VFNMS(T8, T7, VMUL(T6, T9));
{
V Tc, Td, Tf, Tg;
Tc = LD(&(ri[WS(rs, 1)]), ms, &(ri[WS(rs, 1)]));
Td = LD(&(ii[WS(rs, 1)]), ms, &(ii[WS(rs, 1)]));
Te = VFMA(T2, Tc, VMUL(T4, Td));
Tk = VFNMS(T4, Tc, VMUL(T2, Td));
Tf = LD(&(ri[WS(rs, 3)]), ms, &(ri[WS(rs, 1)]));
Tg = LD(&(ii[WS(rs, 3)]), ms, &(ii[WS(rs, 1)]));
Th = VFMA(T3, Tf, VMUL(T5, Tg));
Tl = VFNMS(T5, Tf, VMUL(T3, Tg));
}
{
V Tb, Ti, Tn, Tq;
Tb = VADD(T1, Ta);
Ti = VADD(Te, Th);
ST(&(ri[WS(rs, 2)]), VSUB(Tb, Ti), ms, &(ri[0]));
ST(&(ri[0]), VADD(Tb, Ti), ms, &(ri[0]));
Tn = VADD(Tk, Tl);
Tq = VADD(To, Tp);
ST(&(ii[0]), VADD(Tn, Tq), ms, &(ii[0]));
ST(&(ii[WS(rs, 2)]), VSUB(Tq, Tn), ms, &(ii[0]));
}
{
V Tj, Tm, Tr, Ts;
Tj = VSUB(T1, Ta);
Tm = VSUB(Tk, Tl);
ST(&(ri[WS(rs, 3)]), VSUB(Tj, Tm), ms, &(ri[WS(rs, 1)]));
ST(&(ri[WS(rs, 1)]), VADD(Tj, Tm), ms, &(ri[WS(rs, 1)]));
Tr = VSUB(Tp, To);
Ts = VSUB(Te, Th);
ST(&(ii[WS(rs, 1)]), VSUB(Tr, Ts), ms, &(ii[WS(rs, 1)]));
ST(&(ii[WS(rs, 3)]), VADD(Ts, Tr), ms, &(ii[WS(rs, 1)]));
}
}
}
}
VLEAVE();
}
static void t3fv_5(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
DVK(KP587785252, +0.587785252292473129168705954639072768597652438);
DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
{
INT m;
R *x;
x = ri;
for (m = mb, W = W + (mb * ((TWVL / VL) * 4)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 4), MAKE_VOLATILE_STRIDE(5, rs)) {
V T1, T4, T5, T9;
T1 = LDW(&(W[0]));
T4 = LDW(&(W[TWVL * 2]));
T5 = VZMUL(T1, T4);
T9 = VZMULJ(T1, T4);
{
V Tg, Tk, Tl, T8, Te, Th;
Tg = LD(&(x[0]), ms, &(x[0]));
{
V T3, Td, T7, Tb;
{
V T2, Tc, T6, Ta;
T2 = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
T3 = VZMULJ(T1, T2);
Tc = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
Td = VZMULJ(T4, Tc);
T6 = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
T7 = VZMULJ(T5, T6);
Ta = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
Tb = VZMULJ(T9, Ta);
}
Tk = VSUB(T3, T7);
Tl = VSUB(Tb, Td);
T8 = VADD(T3, T7);
Te = VADD(Tb, Td);
Th = VADD(T8, Te);
}
ST(&(x[0]), VADD(Tg, Th), ms, &(x[0]));
{
V Tm, Tn, Tj, To, Tf, Ti;
Tm = VBYI(VFMA(LDK(KP951056516), Tk, VMUL(LDK(KP587785252), Tl)));
Tn = VBYI(VFNMS(LDK(KP587785252), Tk, VMUL(LDK(KP951056516), Tl)));
Tf = VMUL(LDK(KP559016994), VSUB(T8, Te));
Ti = VFNMS(LDK(KP250000000), Th, Tg);
Tj = VADD(Tf, Ti);
To = VSUB(Ti, Tf);
ST(&(x[WS(rs, 1)]), VSUB(Tj, Tm), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 3)]), VSUB(To, Tn), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 4)]), VADD(Tm, Tj), ms, &(x[0]));
ST(&(x[WS(rs, 2)]), VADD(Tn, To), ms, &(x[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 T5, Th, Te, Ts, Tk, Tm, T2, T4, T3, T6, Tc, T8, Tb, T7, Ta;
V T9, Td, Ti, Tj, TA, Tf, Tn, Tv, Tt, Tz, T1, Tl, Tg, Tu, Tr;
V Tq, Ty, To, Tp, TC, TB, Tx, Tw;
T2 = LD(&(Rp[0]), ms, &(Rp[0]));
T3 = LD(&(Rm[WS(rs, 2)]), -ms, &(Rm[0]));
T4 = VCONJ(T3);
T5 = VSUB(T2, T4);
Th = VADD(T2, T4);
T6 = LD(&(Rp[WS(rs, 2)]), ms, &(Rp[0]));
Tc = LD(&(Rp[WS(rs, 1)]), ms, &(Rp[WS(rs, 1)]));
T7 = LD(&(Rm[0]), -ms, &(Rm[0]));
T8 = VCONJ(T7);
Ta = LD(&(Rm[WS(rs, 1)]), -ms, &(Rm[WS(rs, 1)]));
Tb = VCONJ(Ta);
T9 = VSUB(T6, T8);
Td = VSUB(Tb, Tc);
Te = VADD(T9, Td);
Ts = VBYI(VMUL(LDK(KP866025403), VSUB(T9, Td)));
Ti = VADD(T6, T8);
Tj = VADD(Tb, Tc);
Tk = VADD(Ti, Tj);
Tm = VBYI(VMUL(LDK(KP866025403), VSUB(Ti, Tj)));
TA = VADD(Th, Tk);
T1 = LDW(&(W[TWVL * 4]));
Tf = VZMULI(T1, VADD(T5, Te));
Tl = VFNMS(LDK(KP500000000), Tk, Th);
Tg = LDW(&(W[TWVL * 2]));
Tn = VZMUL(Tg, VSUB(Tl, Tm));
Tu = LDW(&(W[TWVL * 6]));
Tv = VZMUL(Tu, VADD(Tm, Tl));
Tr = VFNMS(LDK(KP500000000), Te, T5);
Tq = LDW(&(W[TWVL * 8]));
Tt = VZMULI(Tq, VSUB(Tr, Ts));
Ty = LDW(&(W[0]));
Tz = VZMULI(Ty, VADD(Ts, Tr));
To = VADD(Tf, Tn);
ST(&(Rp[WS(rs, 1)]), To, ms, &(Rp[WS(rs, 1)]));
Tp = VCONJ(VSUB(Tn, Tf));
ST(&(Rm[WS(rs, 1)]), Tp, -ms, &(Rm[WS(rs, 1)]));
TC = VCONJ(VSUB(TA, Tz));
ST(&(Rm[0]), TC, -ms, &(Rm[0]));
TB = VADD(Tz, TA);
ST(&(Rp[0]), TB, ms, &(Rp[0]));
Tx = VCONJ(VSUB(Tv, Tt));
ST(&(Rm[WS(rs, 2)]), Tx, -ms, &(Rm[0]));
Tw = VADD(Tt, Tv);
ST(&(Rp[WS(rs, 2)]), Tw, ms, &(Rp[0]));
}
}
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 t3fv_5(R *ri, R *ii, 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;
R *x;
x = ri;
for (m = mb, W = W + (mb * ((TWVL / VL) * 4)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 4), MAKE_VOLATILE_STRIDE(5, rs)) {
V T2, T5, T1, T3, Td, T7, Tb;
T2 = LDW(&(W[0]));
T5 = LDW(&(W[TWVL * 2]));
T1 = LD(&(x[0]), ms, &(x[0]));
T3 = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
Td = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
T7 = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
Tb = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
{
V Ta, T6, T4, Te, Tc, T8;
Ta = VZMULJ(T2, T5);
T6 = VZMUL(T2, T5);
T4 = VZMULJ(T2, T3);
Te = VZMULJ(T5, Td);
Tc = VZMULJ(Ta, Tb);
T8 = VZMULJ(T6, T7);
{
V Tf, Tl, T9, Tk;
Tf = VADD(Tc, Te);
Tl = VSUB(Tc, Te);
T9 = VADD(T4, T8);
Tk = VSUB(T4, T8);
{
V Ti, Tg, To, Tm, Th, Tn, Tj;
Ti = VSUB(T9, Tf);
Tg = VADD(T9, Tf);
To = VMUL(LDK(KP951056516), VFNMS(LDK(KP618033988), Tk, Tl));
Tm = VMUL(LDK(KP951056516), VFMA(LDK(KP618033988), Tl, Tk));
Th = VFNMS(LDK(KP250000000), Tg, T1);
ST(&(x[0]), VADD(T1, Tg), ms, &(x[0]));
Tn = VFNMS(LDK(KP559016994), Ti, Th);
Tj = VFMA(LDK(KP559016994), Ti, Th);
ST(&(x[WS(rs, 2)]), VFMAI(To, Tn), ms, &(x[0]));
ST(&(x[WS(rs, 3)]), VFNMSI(To, Tn), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 4)]), VFMAI(Tm, Tj), ms, &(x[0]));
ST(&(x[WS(rs, 1)]), VFNMSI(Tm, Tj), ms, &(x[WS(rs, 1)]));
}
}
}
}
}
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 const R *t1sv_4(R *ri, R *ii, const R *W, stride ios, INT m, INT dist)
{
INT i;
for (i = m; i > 0; i = i - (2 * VL), ri = ri + ((2 * VL) * dist), ii = ii + ((2 * VL) * dist), W = W + ((2 * VL) * 6), MAKE_VOLATILE_STRIDE(ios)) {
V T1, Tv, T3, T6, T5, Ta, Td, Tc, Tg, Tj, Tt, T4, Tf, Ti, Tn;
V Tb, T2, T9;
T1 = LD(&(ri[0]), dist, &(ri[0]));
Tv = LD(&(ii[0]), dist, &(ii[0]));
T3 = LD(&(ri[WS(ios, 2)]), dist, &(ri[0]));
T6 = LD(&(ii[WS(ios, 2)]), dist, &(ii[0]));
T2 = LDW(&(W[TWVL * 2]));
T5 = LDW(&(W[TWVL * 3]));
Ta = LD(&(ri[WS(ios, 1)]), dist, &(ri[WS(ios, 1)]));
Td = LD(&(ii[WS(ios, 1)]), dist, &(ii[WS(ios, 1)]));
T9 = LDW(&(W[0]));
Tc = LDW(&(W[TWVL * 1]));
Tg = LD(&(ri[WS(ios, 3)]), dist, &(ri[WS(ios, 1)]));
Tj = LD(&(ii[WS(ios, 3)]), dist, &(ii[WS(ios, 1)]));
Tt = VMUL(T2, T6);
T4 = VMUL(T2, T3);
Tf = LDW(&(W[TWVL * 4]));
Ti = LDW(&(W[TWVL * 5]));
Tn = VMUL(T9, Td);
Tb = VMUL(T9, Ta);
{
V Tu, T7, Tp, Th, To, Te;
Tu = VFNMS(T5, T3, Tt);
T7 = VFMA(T5, T6, T4);
Tp = VMUL(Tf, Tj);
Th = VMUL(Tf, Tg);
To = VFNMS(Tc, Ta, Tn);
Te = VFMA(Tc, Td, Tb);
{
V Tw, Tx, T8, Tm, Tq, Tk;
Tw = VADD(Tu, Tv);
Tx = VSUB(Tv, Tu);
T8 = VADD(T1, T7);
Tm = VSUB(T1, T7);
Tq = VFNMS(Ti, Tg, Tp);
Tk = VFMA(Ti, Tj, Th);
{
V Ts, Tr, Tl, Ty;
Ts = VADD(To, Tq);
Tr = VSUB(To, Tq);
Tl = VADD(Te, Tk);
Ty = VSUB(Te, Tk);
ST(&(ri[WS(ios, 1)]), VADD(Tm, Tr), dist, &(ri[WS(ios, 1)]));
ST(&(ri[WS(ios, 3)]), VSUB(Tm, Tr), dist, &(ri[WS(ios, 1)]));
ST(&(ii[WS(ios, 2)]), VSUB(Tw, Ts), dist, &(ii[0]));
ST(&(ii[0]), VADD(Ts, Tw), dist, &(ii[0]));
ST(&(ii[WS(ios, 3)]), VADD(Ty, Tx), dist, &(ii[WS(ios, 1)]));
ST(&(ii[WS(ios, 1)]), VSUB(Tx, Ty), dist, &(ii[WS(ios, 1)]));
ST(&(ri[0]), VADD(T8, Tl), dist, &(ri[0]));
ST(&(ri[WS(ios, 2)]), VSUB(T8, Tl), dist, &(ri[0]));
}
}
}
}
return W;
}
static const R *t1sv_4(R *ri, R *ii, const R *W, stride ios, INT m, INT dist)
{
INT i;
for (i = m; i > 0; i = i - (2 * VL), ri = ri + ((2 * VL) * dist), ii = ii + ((2 * VL) * dist), W = W + ((2 * VL) * 6), MAKE_VOLATILE_STRIDE(ios)) {
V T1, Tp, T6, To, Tc, Tk, Th, Tl;
T1 = LD(&(ri[0]), dist, &(ri[0]));
Tp = LD(&(ii[0]), dist, &(ii[0]));
{
V T3, T5, T2, T4;
T3 = LD(&(ri[WS(ios, 2)]), dist, &(ri[0]));
T5 = LD(&(ii[WS(ios, 2)]), dist, &(ii[0]));
T2 = LDW(&(W[TWVL * 2]));
T4 = LDW(&(W[TWVL * 3]));
T6 = VFMA(T2, T3, VMUL(T4, T5));
To = VFNMS(T4, T3, VMUL(T2, T5));
}
{
V T9, Tb, T8, Ta;
T9 = LD(&(ri[WS(ios, 1)]), dist, &(ri[WS(ios, 1)]));
Tb = LD(&(ii[WS(ios, 1)]), dist, &(ii[WS(ios, 1)]));
T8 = LDW(&(W[0]));
Ta = LDW(&(W[TWVL * 1]));
Tc = VFMA(T8, T9, VMUL(Ta, Tb));
Tk = VFNMS(Ta, T9, VMUL(T8, Tb));
}
{
V Te, Tg, Td, Tf;
Te = LD(&(ri[WS(ios, 3)]), dist, &(ri[WS(ios, 1)]));
Tg = LD(&(ii[WS(ios, 3)]), dist, &(ii[WS(ios, 1)]));
Td = LDW(&(W[TWVL * 4]));
Tf = LDW(&(W[TWVL * 5]));
Th = VFMA(Td, Te, VMUL(Tf, Tg));
Tl = VFNMS(Tf, Te, VMUL(Td, Tg));
}
{
V T7, Ti, Tn, Tq;
T7 = VADD(T1, T6);
Ti = VADD(Tc, Th);
ST(&(ri[WS(ios, 2)]), VSUB(T7, Ti), dist, &(ri[0]));
ST(&(ri[0]), VADD(T7, Ti), dist, &(ri[0]));
Tn = VADD(Tk, Tl);
Tq = VADD(To, Tp);
ST(&(ii[0]), VADD(Tn, Tq), dist, &(ii[0]));
ST(&(ii[WS(ios, 2)]), VSUB(Tq, Tn), dist, &(ii[0]));
}
{
V Tj, Tm, Tr, Ts;
Tj = VSUB(T1, T6);
Tm = VSUB(Tk, Tl);
ST(&(ri[WS(ios, 3)]), VSUB(Tj, Tm), dist, &(ri[WS(ios, 1)]));
ST(&(ri[WS(ios, 1)]), VADD(Tj, Tm), dist, &(ri[WS(ios, 1)]));
Tr = VSUB(Tp, To);
Ts = VSUB(Tc, Th);
ST(&(ii[WS(ios, 1)]), VSUB(Tr, Ts), dist, &(ii[WS(ios, 1)]));
ST(&(ii[WS(ios, 3)]), VADD(Ts, Tr), dist, &(ii[WS(ios, 1)]));
}
}
return W;
}
static void t1sv_2(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
INT m;
for (m = mb, W = W + (mb * 2); m < me; m = m + (2 * VL), ri = ri + ((2 * VL) * ms), ii = ii + ((2 * VL) * ms), W = W + ((2 * VL) * 2), MAKE_VOLATILE_STRIDE(rs)) {
V T1, Ta, T3, T6, T2, T5;
T1 = LD(&(ri[0]), ms, &(ri[0]));
Ta = LD(&(ii[0]), ms, &(ii[0]));
T3 = LD(&(ri[WS(rs, 1)]), ms, &(ri[WS(rs, 1)]));
T6 = LD(&(ii[WS(rs, 1)]), ms, &(ii[WS(rs, 1)]));
T2 = LDW(&(W[0]));
T5 = LDW(&(W[TWVL * 1]));
{
V T8, T4, T9, T7;
T8 = VMUL(T2, T6);
T4 = VMUL(T2, T3);
T9 = VFNMS(T5, T3, T8);
T7 = VFMA(T5, T6, T4);
ST(&(ii[0]), VADD(T9, Ta), ms, &(ii[0]));
ST(&(ii[WS(rs, 1)]), VSUB(Ta, T9), ms, &(ii[WS(rs, 1)]));
ST(&(ri[0]), VADD(T1, T7), ms, &(ri[0]));
ST(&(ri[WS(rs, 1)]), VSUB(T1, T7), ms, &(ri[WS(rs, 1)]));
}
}
}
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 t3bv_10(R *ri, R *ii, 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;
R *x;
x = ii;
for (m = mb, W = W + (mb * ((TWVL / VL) * 6)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 6), MAKE_VOLATILE_STRIDE(10, rs)) {
V T1, T7, Th, Tx, Tr, Td, Tp, T6, Tv, Tc, Te, Ti, Tl, T2, T3;
V T5;
T2 = LDW(&(W[0]));
T3 = LDW(&(W[TWVL * 2]));
T5 = LDW(&(W[TWVL * 4]));
T1 = LD(&(x[0]), ms, &(x[0]));
T7 = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
{
V To, Tw, Tq, Tu, Ta, T4, Tt, Tk, Tb;
To = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
Tw = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
Tq = LD(&(x[WS(rs, 9)]), ms, &(x[WS(rs, 1)]));
Tu = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
Ta = VZMULJ(T2, T3);
T4 = VZMUL(T2, T3);
Th = VZMULJ(T2, T5);
Tt = VZMULJ(T3, T5);
Tb = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
Tx = VZMUL(T2, Tw);
Tr = VZMUL(T5, Tq);
Tk = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
Td = VZMULJ(Ta, T5);
Tp = VZMUL(T4, To);
T6 = VZMULJ(T4, T5);
Tv = VZMUL(Tt, Tu);
Tc = VZMUL(Ta, Tb);
Te = LD(&(x[WS(rs, 7)]), ms, &(x[WS(rs, 1)]));
Ti = LD(&(x[WS(rs, 8)]), ms, &(x[0]));
Tl = VZMUL(T3, Tk);
}
{
V TN, Ts, T8, Ty, TO, Tf, Tj;
TN = VADD(Tp, Tr);
Ts = VSUB(Tp, Tr);
T8 = VZMUL(T6, T7);
Ty = VSUB(Tv, Tx);
TO = VADD(Tv, Tx);
Tf = VZMUL(Td, Te);
Tj = VZMUL(Th, Ti);
{
V T9, TJ, TP, TU, Tz, TF, Tg, TK, Tm, TL;
T9 = VSUB(T1, T8);
TJ = VADD(T1, T8);
TP = VADD(TN, TO);
TU = VSUB(TN, TO);
Tz = VADD(Ts, Ty);
TF = VSUB(Ts, Ty);
Tg = VSUB(Tc, Tf);
TK = VADD(Tc, Tf);
Tm = VSUB(Tj, Tl);
TL = VADD(Tj, Tl);
{
V TM, TV, Tn, TE;
TM = VADD(TK, TL);
TV = VSUB(TK, TL);
Tn = VADD(Tg, Tm);
TE = VSUB(Tg, Tm);
{
V TW, TY, TS, TQ, TG, TI, TC, TA, TR, TB;
TW = VMUL(LDK(KP951056516), VFNMS(LDK(KP618033988), TV, TU));
TY = VMUL(LDK(KP951056516), VFMA(LDK(KP618033988), TU, TV));
TS = VSUB(TM, TP);
TQ = VADD(TM, TP);
TG = VMUL(LDK(KP951056516), VFMA(LDK(KP618033988), TF, TE));
TI = VMUL(LDK(KP951056516), VFNMS(LDK(KP618033988), TE, TF));
TC = VSUB(Tn, Tz);
TA = VADD(Tn, Tz);
ST(&(x[0]), VADD(TJ, TQ), ms, &(x[0]));
TR = VFNMS(LDK(KP250000000), TQ, TJ);
ST(&(x[WS(rs, 5)]), VADD(T9, TA), ms, &(x[WS(rs, 1)]));
TB = VFNMS(LDK(KP250000000), TA, T9);
{
V TX, TT, TH, TD;
TX = VFMA(LDK(KP559016994), TS, TR);
TT = VFNMS(LDK(KP559016994), TS, TR);
TH = VFNMS(LDK(KP559016994), TC, TB);
TD = VFMA(LDK(KP559016994), TC, TB);
ST(&(x[WS(rs, 8)]), VFMAI(TW, TT), ms, &(x[0]));
ST(&(x[WS(rs, 2)]), VFNMSI(TW, TT), ms, &(x[0]));
ST(&(x[WS(rs, 6)]), VFMAI(TY, TX), ms, &(x[0]));
ST(&(x[WS(rs, 4)]), VFNMSI(TY, TX), ms, &(x[0]));
ST(&(x[WS(rs, 9)]), VFNMSI(TG, TD), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 1)]), VFMAI(TG, TD), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 7)]), VFNMSI(TI, TH), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 3)]), VFMAI(TI, TH), ms, &(x[WS(rs, 1)]));
}
}
}
}
}
}
}
VLEAVE();
}
static void t3bv_10(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP587785252, +0.587785252292473129168705954639072768597652438);
DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
{
INT m;
R *x;
x = ii;
for (m = mb, W = W + (mb * ((TWVL / VL) * 6)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 6), MAKE_VOLATILE_STRIDE(10, rs)) {
V T1, T2, T3, Ti, T6, T7, TA, Tb, To;
T1 = LDW(&(W[0]));
T2 = LDW(&(W[TWVL * 2]));
T3 = VZMULJ(T1, T2);
Ti = VZMUL(T1, T2);
T6 = LDW(&(W[TWVL * 4]));
T7 = VZMULJ(T3, T6);
TA = VZMULJ(Ti, T6);
Tb = VZMULJ(T1, T6);
To = VZMULJ(T2, T6);
{
V TD, TQ, Tn, Tt, Tx, TM, TN, TS, Ta, Tg, Tw, TJ, TK, TR, Tz;
V TC, TB;
Tz = LD(&(x[0]), ms, &(x[0]));
TB = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
TC = VZMUL(TA, TB);
TD = VSUB(Tz, TC);
TQ = VADD(Tz, TC);
{
V Tk, Ts, Tm, Tq;
{
V Tj, Tr, Tl, Tp;
Tj = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
Tk = VZMUL(Ti, Tj);
Tr = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
Ts = VZMUL(T1, Tr);
Tl = LD(&(x[WS(rs, 9)]), ms, &(x[WS(rs, 1)]));
Tm = VZMUL(T6, Tl);
Tp = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
Tq = VZMUL(To, Tp);
}
Tn = VSUB(Tk, Tm);
Tt = VSUB(Tq, Ts);
Tx = VADD(Tn, Tt);
TM = VADD(Tk, Tm);
TN = VADD(Tq, Ts);
TS = VADD(TM, TN);
}
{
V T5, Tf, T9, Td;
{
V T4, Te, T8, Tc;
T4 = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
T5 = VZMUL(T3, T4);
Te = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
Tf = VZMUL(T2, Te);
T8 = LD(&(x[WS(rs, 7)]), ms, &(x[WS(rs, 1)]));
T9 = VZMUL(T7, T8);
Tc = LD(&(x[WS(rs, 8)]), ms, &(x[0]));
Td = VZMUL(Tb, Tc);
}
Ta = VSUB(T5, T9);
Tg = VSUB(Td, Tf);
Tw = VADD(Ta, Tg);
TJ = VADD(T5, T9);
TK = VADD(Td, Tf);
TR = VADD(TJ, TK);
}
{
V Ty, TE, TF, Tv, TI, Th, Tu, TH, TG;
Ty = VMUL(LDK(KP559016994), VSUB(Tw, Tx));
TE = VADD(Tw, Tx);
TF = VFNMS(LDK(KP250000000), TE, TD);
Th = VSUB(Ta, Tg);
Tu = VSUB(Tn, Tt);
Tv = VBYI(VFMA(LDK(KP951056516), Th, VMUL(LDK(KP587785252), Tu)));
TI = VBYI(VFNMS(LDK(KP951056516), Tu, VMUL(LDK(KP587785252), Th)));
ST(&(x[WS(rs, 5)]), VADD(TD, TE), ms, &(x[WS(rs, 1)]));
TH = VSUB(TF, Ty);
ST(&(x[WS(rs, 3)]), VSUB(TH, TI), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 7)]), VADD(TI, TH), ms, &(x[WS(rs, 1)]));
TG = VADD(Ty, TF);
ST(&(x[WS(rs, 1)]), VADD(Tv, TG), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 9)]), VSUB(TG, Tv), ms, &(x[WS(rs, 1)]));
}
{
V TV, TT, TU, TP, TY, TL, TO, TX, TW;
TV = VMUL(LDK(KP559016994), VSUB(TR, TS));
TT = VADD(TR, TS);
TU = VFNMS(LDK(KP250000000), TT, TQ);
TL = VSUB(TJ, TK);
TO = VSUB(TM, TN);
TP = VBYI(VFNMS(LDK(KP951056516), TO, VMUL(LDK(KP587785252), TL)));
TY = VBYI(VFMA(LDK(KP951056516), TL, VMUL(LDK(KP587785252), TO)));
ST(&(x[0]), VADD(TQ, TT), ms, &(x[0]));
TX = VADD(TV, TU);
ST(&(x[WS(rs, 4)]), VSUB(TX, TY), ms, &(x[0]));
ST(&(x[WS(rs, 6)]), VADD(TY, TX), ms, &(x[0]));
TW = VSUB(TU, TV);
ST(&(x[WS(rs, 2)]), VADD(TP, TW), ms, &(x[0]));
ST(&(x[WS(rs, 8)]), VSUB(TW, TP), ms, &(x[0]));
}
}
}
}
VLEAVE();
}
static void t3fv_20(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
DVK(KP618033988, +0.618033988749894848204586834365638117720309180);
DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
{
INT m;
R *x;
x = ri;
for (m = mb, W = W + (mb * ((TWVL / VL) * 8)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 8), MAKE_VOLATILE_STRIDE(20, rs)) {
V T1k, T1w, T1r, T1z, T1o, T1y, T1v, T1h;
{
V T2, T8, T3, Td;
T2 = LDW(&(W[0]));
T8 = LDW(&(W[TWVL * 2]));
T3 = LDW(&(W[TWVL * 4]));
Td = LDW(&(W[TWVL * 6]));
{
V T7, TM, T1F, T23, T1p, Tp, T1j, T27, T1P, T1I, T1i, T1L, T28, T1S, T1q;
V TE, T1n, T1d, T26, T2e;
{
V T1, TK, T5, TH;
T1 = LD(&(x[0]), ms, &(x[0]));
TK = LD(&(x[WS(rs, 15)]), ms, &(x[WS(rs, 1)]));
T5 = LD(&(x[WS(rs, 10)]), ms, &(x[0]));
TH = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
{
V TA, Tx, TU, T1O, T14, Th, T1G, T1R, T1b, T1J, To, Ts, TV, Tv, TO;
V TQ, TT, Ty, TB;
{
V Tq, Tt, T17, T1a, Tk, Tn;
{
V Tl, Ti, T15, T18, TZ, Tc, T6, Tb, Tf, T10, T12, TL;
{
V TJ, Ta, T9, T4;
Ta = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
TA = VZMULJ(T2, T8);
T9 = VZMUL(T2, T8);
Tx = VZMUL(T8, T3);
Tl = VZMULJ(T8, T3);
T4 = VZMUL(T2, T3);
Tq = VZMULJ(T2, T3);
Tt = VZMULJ(T2, Td);
Ti = VZMULJ(T8, Td);
T15 = VZMULJ(TA, Td);
T18 = VZMULJ(TA, T3);
TU = VZMUL(TA, T3);
TJ = VZMULJ(T9, Td);
TZ = VZMUL(T9, T3);
Tc = VZMULJ(T9, T3);
T6 = VZMULJ(T4, T5);
Tb = VZMULJ(T9, Ta);
Tf = LD(&(x[WS(rs, 14)]), ms, &(x[0]));
T10 = LD(&(x[WS(rs, 13)]), ms, &(x[WS(rs, 1)]));
T12 = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
TL = VZMULJ(TJ, TK);
}
{
V T1D, T11, T13, T19, T1E, Tg, T16, TI, Te, Tj, Tm;
T16 = LD(&(x[WS(rs, 17)]), ms, &(x[WS(rs, 1)]));
TI = VZMULJ(Tc, TH);
Te = VZMULJ(Tc, Td);
T7 = VSUB(T1, T6);
T1D = VADD(T1, T6);
T11 = VZMULJ(TZ, T10);
T13 = VZMULJ(T8, T12);
T19 = LD(&(x[WS(rs, 7)]), ms, &(x[WS(rs, 1)]));
T17 = VZMULJ(T15, T16);
TM = VSUB(TI, TL);
T1E = VADD(TI, TL);
Tg = VZMULJ(Te, Tf);
Tj = LD(&(x[WS(rs, 16)]), ms, &(x[0]));
Tm = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
T1O = VADD(T11, T13);
T14 = VSUB(T11, T13);
T1a = VZMULJ(T18, T19);
T1F = VSUB(T1D, T1E);
T23 = VADD(T1D, T1E);
Th = VSUB(Tb, Tg);
T1G = VADD(Tb, Tg);
Tk = VZMULJ(Ti, Tj);
Tn = VZMULJ(Tl, Tm);
}
}
{
V Tr, Tu, TN, TP, TS;
Tr = LD(&(x[WS(rs, 8)]), ms, &(x[0]));
T1R = VADD(T17, T1a);
T1b = VSUB(T17, T1a);
Tu = LD(&(x[WS(rs, 18)]), ms, &(x[0]));
TN = LD(&(x[WS(rs, 9)]), ms, &(x[WS(rs, 1)]));
TP = LD(&(x[WS(rs, 19)]), ms, &(x[WS(rs, 1)]));
TS = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
T1J = VADD(Tk, Tn);
To = VSUB(Tk, Tn);
Ts = VZMULJ(Tq, Tr);
TV = LD(&(x[WS(rs, 11)]), ms, &(x[WS(rs, 1)]));
Tv = VZMULJ(Tt, Tu);
TO = VZMULJ(T3, TN);
TQ = VZMULJ(Td, TP);
TT = VZMULJ(T2, TS);
Ty = LD(&(x[WS(rs, 12)]), ms, &(x[0]));
TB = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
}
}
{
V T1N, Tw, T1H, TR, Tz, TC, T1c, TX, T1K, TW;
T1p = VSUB(Th, To);
Tp = VADD(Th, To);
TW = VZMULJ(TU, TV);
T1N = VADD(Ts, Tv);
Tw = VSUB(Ts, Tv);
T1H = VADD(TO, TQ);
TR = VSUB(TO, TQ);
Tz = VZMULJ(Tx, Ty);
TC = VZMULJ(TA, TB);
T1j = VSUB(T1b, T14);
T1c = VADD(T14, T1b);
TX = VSUB(TT, TW);
T1K = VADD(TT, TW);
T27 = VADD(T1N, T1O);
T1P = VSUB(T1N, T1O);
{
V TD, T1Q, T24, TY, T25;
TD = VSUB(Tz, TC);
T1Q = VADD(Tz, TC);
T1I = VSUB(T1G, T1H);
T24 = VADD(T1G, T1H);
TY = VADD(TR, TX);
T1i = VSUB(TX, TR);
T25 = VADD(T1J, T1K);
T1L = VSUB(T1J, T1K);
T28 = VADD(T1Q, T1R);
T1S = VSUB(T1Q, T1R);
T1q = VSUB(Tw, TD);
TE = VADD(Tw, TD);
T1n = VSUB(T1c, TY);
T1d = VADD(TY, T1c);
T26 = VADD(T24, T25);
T2e = VSUB(T24, T25);
}
}
}
}
{
V T1M, T1Z, T1Y, T1T, T29, T2f, T1g, TF, T1m, T1e;
T1M = VADD(T1I, T1L);
T1Z = VSUB(T1I, T1L);
T1Y = VSUB(T1P, T1S);
T1T = VADD(T1P, T1S);
T29 = VADD(T27, T28);
T2f = VSUB(T27, T28);
T1g = VSUB(Tp, TE);
TF = VADD(Tp, TE);
T1m = VFNMS(LDK(KP250000000), T1d, TM);
T1e = VADD(TM, T1d);
{
V T1W, T2c, T1f, T2i, T2g, T22, T20, T1V, T2b, T1U, T2a, TG;
T1k = VFMA(LDK(KP618033988), T1j, T1i);
T1w = VFNMS(LDK(KP618033988), T1i, T1j);
T1W = VSUB(T1M, T1T);
T1U = VADD(T1M, T1T);
T2c = VSUB(T26, T29);
T2a = VADD(T26, T29);
T1f = VFNMS(LDK(KP250000000), TF, T7);
TG = VADD(T7, TF);
T2i = VMUL(LDK(KP951056516), VFNMS(LDK(KP618033988), T2e, T2f));
T2g = VMUL(LDK(KP951056516), VFMA(LDK(KP618033988), T2f, T2e));
T22 = VMUL(LDK(KP951056516), VFMA(LDK(KP618033988), T1Y, T1Z));
T20 = VMUL(LDK(KP951056516), VFNMS(LDK(KP618033988), T1Z, T1Y));
ST(&(x[WS(rs, 10)]), VADD(T1F, T1U), ms, &(x[0]));
T1V = VFNMS(LDK(KP250000000), T1U, T1F);
ST(&(x[0]), VADD(T23, T2a), ms, &(x[0]));
T2b = VFNMS(LDK(KP250000000), T2a, T23);
ST(&(x[WS(rs, 15)]), VFMAI(T1e, TG), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 5)]), VFNMSI(T1e, TG), ms, &(x[WS(rs, 1)]));
T1r = VFMA(LDK(KP618033988), T1q, T1p);
T1z = VFNMS(LDK(KP618033988), T1p, T1q);
{
V T21, T1X, T2h, T2d;
T21 = VFMA(LDK(KP559016994), T1W, T1V);
T1X = VFNMS(LDK(KP559016994), T1W, T1V);
T2h = VFNMS(LDK(KP559016994), T2c, T2b);
T2d = VFMA(LDK(KP559016994), T2c, T2b);
ST(&(x[WS(rs, 18)]), VFNMSI(T20, T1X), ms, &(x[0]));
ST(&(x[WS(rs, 2)]), VFMAI(T20, T1X), ms, &(x[0]));
ST(&(x[WS(rs, 14)]), VFMAI(T22, T21), ms, &(x[0]));
ST(&(x[WS(rs, 6)]), VFNMSI(T22, T21), ms, &(x[0]));
ST(&(x[WS(rs, 16)]), VFNMSI(T2g, T2d), ms, &(x[0]));
ST(&(x[WS(rs, 4)]), VFMAI(T2g, T2d), ms, &(x[0]));
ST(&(x[WS(rs, 12)]), VFMAI(T2i, T2h), ms, &(x[0]));
ST(&(x[WS(rs, 8)]), VFNMSI(T2i, T2h), ms, &(x[0]));
T1o = VFNMS(LDK(KP559016994), T1n, T1m);
T1y = VFMA(LDK(KP559016994), T1n, T1m);
T1v = VFNMS(LDK(KP559016994), T1g, T1f);
T1h = VFMA(LDK(KP559016994), T1g, T1f);
}
}
}
}
}
{
V T1C, T1A, T1s, T1u, T1l, T1t, T1B, T1x;
T1C = VFMA(LDK(KP951056516), T1z, T1y);
T1A = VFNMS(LDK(KP951056516), T1z, T1y);
T1s = VFMA(LDK(KP951056516), T1r, T1o);
T1u = VFNMS(LDK(KP951056516), T1r, T1o);
T1l = VFMA(LDK(KP951056516), T1k, T1h);
T1t = VFNMS(LDK(KP951056516), T1k, T1h);
T1B = VFMA(LDK(KP951056516), T1w, T1v);
T1x = VFNMS(LDK(KP951056516), T1w, T1v);
ST(&(x[WS(rs, 11)]), VFMAI(T1u, T1t), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 9)]), VFNMSI(T1u, T1t), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 19)]), VFMAI(T1s, T1l), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 1)]), VFNMSI(T1s, T1l), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 3)]), VFMAI(T1A, T1x), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 17)]), VFNMSI(T1A, T1x), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 7)]), VFMAI(T1C, T1B), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 13)]), VFNMSI(T1C, T1B), ms, &(x[WS(rs, 1)]));
}
}
}
VLEAVE();
}
static void t3fv_20(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP587785252, +0.587785252292473129168705954639072768597652438);
DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
{
INT m;
R *x;
x = ri;
for (m = mb, W = W + (mb * ((TWVL / VL) * 8)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 8), MAKE_VOLATILE_STRIDE(20, rs)) {
V T2, T8, T9, TA, T3, Tc, T4, TZ, T18, Tl, Tq, Tx, TU, Td, Te;
V T15, Ti, Tt, TJ;
T2 = LDW(&(W[0]));
T8 = LDW(&(W[TWVL * 2]));
T9 = VZMUL(T2, T8);
TA = VZMULJ(T2, T8);
T3 = LDW(&(W[TWVL * 4]));
Tc = VZMULJ(T9, T3);
T4 = VZMUL(T2, T3);
TZ = VZMUL(T9, T3);
T18 = VZMULJ(TA, T3);
Tl = VZMULJ(T8, T3);
Tq = VZMULJ(T2, T3);
Tx = VZMUL(T8, T3);
TU = VZMUL(TA, T3);
Td = LDW(&(W[TWVL * 6]));
Te = VZMULJ(Tc, Td);
T15 = VZMULJ(TA, Td);
Ti = VZMULJ(T8, Td);
Tt = VZMULJ(T2, Td);
TJ = VZMULJ(T9, Td);
{
V T7, TM, T1U, T2d, T1i, T1p, T1q, T1j, Tp, TE, TF, T26, T27, T2b, T1M;
V T1P, T1V, TY, T1c, T1d, T23, T24, T2a, T1F, T1I, T1W, TG, T1e;
{
V T1, TL, T6, TI, TK, T5, TH, T1S, T1T;
T1 = LD(&(x[0]), ms, &(x[0]));
TK = LD(&(x[WS(rs, 15)]), ms, &(x[WS(rs, 1)]));
TL = VZMULJ(TJ, TK);
T5 = LD(&(x[WS(rs, 10)]), ms, &(x[0]));
T6 = VZMULJ(T4, T5);
TH = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
TI = VZMULJ(Tc, TH);
T7 = VSUB(T1, T6);
TM = VSUB(TI, TL);
T1S = VADD(T1, T6);
T1T = VADD(TI, TL);
T1U = VSUB(T1S, T1T);
T2d = VADD(T1S, T1T);
}
{
V Th, T1K, T14, T1E, T1b, T1H, To, T1N, Tw, T1D, TR, T1L, TX, T1O, TD;
V T1G;
{
V Tb, Tg, Ta, Tf;
Ta = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
Tb = VZMULJ(T9, Ta);
Tf = LD(&(x[WS(rs, 14)]), ms, &(x[0]));
Tg = VZMULJ(Te, Tf);
Th = VSUB(Tb, Tg);
T1K = VADD(Tb, Tg);
}
{
V T11, T13, T10, T12;
T10 = LD(&(x[WS(rs, 13)]), ms, &(x[WS(rs, 1)]));
T11 = VZMULJ(TZ, T10);
T12 = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
T13 = VZMULJ(T8, T12);
T14 = VSUB(T11, T13);
T1E = VADD(T11, T13);
}
{
V T17, T1a, T16, T19;
T16 = LD(&(x[WS(rs, 17)]), ms, &(x[WS(rs, 1)]));
T17 = VZMULJ(T15, T16);
T19 = LD(&(x[WS(rs, 7)]), ms, &(x[WS(rs, 1)]));
T1a = VZMULJ(T18, T19);
T1b = VSUB(T17, T1a);
T1H = VADD(T17, T1a);
}
{
V Tk, Tn, Tj, Tm;
Tj = LD(&(x[WS(rs, 16)]), ms, &(x[0]));
Tk = VZMULJ(Ti, Tj);
Tm = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
Tn = VZMULJ(Tl, Tm);
To = VSUB(Tk, Tn);
T1N = VADD(Tk, Tn);
}
{
V Ts, Tv, Tr, Tu;
Tr = LD(&(x[WS(rs, 8)]), ms, &(x[0]));
Ts = VZMULJ(Tq, Tr);
Tu = LD(&(x[WS(rs, 18)]), ms, &(x[0]));
Tv = VZMULJ(Tt, Tu);
Tw = VSUB(Ts, Tv);
T1D = VADD(Ts, Tv);
}
{
V TO, TQ, TN, TP;
TN = LD(&(x[WS(rs, 9)]), ms, &(x[WS(rs, 1)]));
TO = VZMULJ(T3, TN);
TP = LD(&(x[WS(rs, 19)]), ms, &(x[WS(rs, 1)]));
TQ = VZMULJ(Td, TP);
TR = VSUB(TO, TQ);
T1L = VADD(TO, TQ);
}
{
V TT, TW, TS, TV;
TS = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
TT = VZMULJ(T2, TS);
TV = LD(&(x[WS(rs, 11)]), ms, &(x[WS(rs, 1)]));
TW = VZMULJ(TU, TV);
TX = VSUB(TT, TW);
T1O = VADD(TT, TW);
}
{
V Tz, TC, Ty, TB;
Ty = LD(&(x[WS(rs, 12)]), ms, &(x[0]));
Tz = VZMULJ(Tx, Ty);
TB = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
TC = VZMULJ(TA, TB);
TD = VSUB(Tz, TC);
T1G = VADD(Tz, TC);
}
T1i = VSUB(TX, TR);
T1p = VSUB(Th, To);
T1q = VSUB(Tw, TD);
T1j = VSUB(T1b, T14);
Tp = VADD(Th, To);
TE = VADD(Tw, TD);
TF = VADD(Tp, TE);
T26 = VADD(T1D, T1E);
T27 = VADD(T1G, T1H);
T2b = VADD(T26, T27);
T1M = VSUB(T1K, T1L);
T1P = VSUB(T1N, T1O);
T1V = VADD(T1M, T1P);
TY = VADD(TR, TX);
T1c = VADD(T14, T1b);
T1d = VADD(TY, T1c);
T23 = VADD(T1K, T1L);
T24 = VADD(T1N, T1O);
T2a = VADD(T23, T24);
T1F = VSUB(T1D, T1E);
T1I = VSUB(T1G, T1H);
T1W = VADD(T1F, T1I);
}
TG = VADD(T7, TF);
T1e = VBYI(VADD(TM, T1d));
ST(&(x[WS(rs, 5)]), VSUB(TG, T1e), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 15)]), VADD(TG, T1e), ms, &(x[WS(rs, 1)]));
{
V T2c, T2e, T2f, T29, T2i, T25, T28, T2h, T2g;
T2c = VMUL(LDK(KP559016994), VSUB(T2a, T2b));
T2e = VADD(T2a, T2b);
T2f = VFNMS(LDK(KP250000000), T2e, T2d);
T25 = VSUB(T23, T24);
T28 = VSUB(T26, T27);
T29 = VBYI(VFMA(LDK(KP951056516), T25, VMUL(LDK(KP587785252), T28)));
T2i = VBYI(VFNMS(LDK(KP587785252), T25, VMUL(LDK(KP951056516), T28)));
ST(&(x[0]), VADD(T2d, T2e), ms, &(x[0]));
T2h = VSUB(T2f, T2c);
ST(&(x[WS(rs, 8)]), VSUB(T2h, T2i), ms, &(x[0]));
ST(&(x[WS(rs, 12)]), VADD(T2i, T2h), ms, &(x[0]));
T2g = VADD(T2c, T2f);
ST(&(x[WS(rs, 4)]), VADD(T29, T2g), ms, &(x[0]));
ST(&(x[WS(rs, 16)]), VSUB(T2g, T29), ms, &(x[0]));
}
{
V T1Z, T1X, T1Y, T1R, T22, T1J, T1Q, T21, T20;
T1Z = VMUL(LDK(KP559016994), VSUB(T1V, T1W));
T1X = VADD(T1V, T1W);
T1Y = VFNMS(LDK(KP250000000), T1X, T1U);
T1J = VSUB(T1F, T1I);
T1Q = VSUB(T1M, T1P);
T1R = VBYI(VFNMS(LDK(KP587785252), T1Q, VMUL(LDK(KP951056516), T1J)));
T22 = VBYI(VFMA(LDK(KP951056516), T1Q, VMUL(LDK(KP587785252), T1J)));
ST(&(x[WS(rs, 10)]), VADD(T1U, T1X), ms, &(x[0]));
T21 = VADD(T1Z, T1Y);
ST(&(x[WS(rs, 6)]), VSUB(T21, T22), ms, &(x[0]));
ST(&(x[WS(rs, 14)]), VADD(T22, T21), ms, &(x[0]));
T20 = VSUB(T1Y, T1Z);
ST(&(x[WS(rs, 2)]), VADD(T1R, T20), ms, &(x[0]));
ST(&(x[WS(rs, 18)]), VSUB(T20, T1R), ms, &(x[0]));
}
{
V T1k, T1r, T1z, T1w, T1o, T1y, T1h, T1v;
T1k = VFMA(LDK(KP951056516), T1i, VMUL(LDK(KP587785252), T1j));
T1r = VFMA(LDK(KP951056516), T1p, VMUL(LDK(KP587785252), T1q));
T1z = VFNMS(LDK(KP587785252), T1p, VMUL(LDK(KP951056516), T1q));
T1w = VFNMS(LDK(KP587785252), T1i, VMUL(LDK(KP951056516), T1j));
{
V T1m, T1n, T1f, T1g;
T1m = VFMS(LDK(KP250000000), T1d, TM);
T1n = VMUL(LDK(KP559016994), VSUB(T1c, TY));
T1o = VADD(T1m, T1n);
T1y = VSUB(T1n, T1m);
T1f = VMUL(LDK(KP559016994), VSUB(Tp, TE));
T1g = VFNMS(LDK(KP250000000), TF, T7);
T1h = VADD(T1f, T1g);
T1v = VSUB(T1g, T1f);
}
{
V T1l, T1s, T1B, T1C;
T1l = VADD(T1h, T1k);
T1s = VBYI(VSUB(T1o, T1r));
ST(&(x[WS(rs, 19)]), VSUB(T1l, T1s), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 1)]), VADD(T1l, T1s), ms, &(x[WS(rs, 1)]));
T1B = VADD(T1v, T1w);
T1C = VBYI(VADD(T1z, T1y));
ST(&(x[WS(rs, 13)]), VSUB(T1B, T1C), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 7)]), VADD(T1B, T1C), ms, &(x[WS(rs, 1)]));
}
{
V T1t, T1u, T1x, T1A;
T1t = VSUB(T1h, T1k);
T1u = VBYI(VADD(T1r, T1o));
ST(&(x[WS(rs, 11)]), VSUB(T1t, T1u), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 9)]), VADD(T1t, T1u), ms, &(x[WS(rs, 1)]));
T1x = VSUB(T1v, T1w);
T1A = VBYI(VSUB(T1y, T1z));
ST(&(x[WS(rs, 17)]), VSUB(T1x, T1A), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 3)]), VADD(T1x, T1A), ms, &(x[WS(rs, 1)]));
}
}
}
}
}
VLEAVE();
}
static void hc2cfdftv_8(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP707106781, +0.707106781186547524400844362104849039284835938);
DVK(KP353553390, +0.353553390593273762200422181052424519642417969);
DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
{
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 Ta, TE, Tr, TF, Tl, TK, Tw, TG, T1, T6, T3, T8, T2, T7, T4;
V T9, T5, To, Tq, Tn, Tp, Tc, Th, Te, Tj, Td, Ti, Tf, Tk, Tb;
V Tg, Tt, Tv, Ts, Tu, Ty, Tz, Tm, Tx, TC, TD, TA, TB, TI, TO;
V TL, TP, TH, TJ, TM, TR, TN, TQ;
T1 = LD(&(Rp[0]), ms, &(Rp[0]));
T6 = LD(&(Rp[WS(rs, 2)]), ms, &(Rp[0]));
T2 = LD(&(Rm[0]), -ms, &(Rm[0]));
T3 = VCONJ(T2);
T7 = LD(&(Rm[WS(rs, 2)]), -ms, &(Rm[0]));
T8 = VCONJ(T7);
T4 = VADD(T1, T3);
T5 = LDW(&(W[TWVL * 6]));
T9 = VZMULJ(T5, VADD(T6, T8));
Ta = VADD(T4, T9);
TE = VMUL(LDK(KP500000000), VSUB(T4, T9));
Tn = LDW(&(W[0]));
To = VZMULIJ(Tn, VSUB(T3, T1));
Tp = LDW(&(W[TWVL * 8]));
Tq = VZMULIJ(Tp, VSUB(T8, T6));
Tr = VADD(To, Tq);
TF = VSUB(To, Tq);
Tc = LD(&(Rp[WS(rs, 1)]), ms, &(Rp[WS(rs, 1)]));
Th = LD(&(Rp[WS(rs, 3)]), ms, &(Rp[WS(rs, 1)]));
Td = LD(&(Rm[WS(rs, 1)]), -ms, &(Rm[WS(rs, 1)]));
Te = VCONJ(Td);
Ti = LD(&(Rm[WS(rs, 3)]), -ms, &(Rm[WS(rs, 1)]));
Tj = VCONJ(Ti);
Tb = LDW(&(W[TWVL * 2]));
Tf = VZMULJ(Tb, VADD(Tc, Te));
Tg = LDW(&(W[TWVL * 10]));
Tk = VZMULJ(Tg, VADD(Th, Tj));
Tl = VADD(Tf, Tk);
TK = VSUB(Tf, Tk);
Ts = LDW(&(W[TWVL * 4]));
Tt = VZMULIJ(Ts, VSUB(Te, Tc));
Tu = LDW(&(W[TWVL * 12]));
Tv = VZMULIJ(Tu, VSUB(Tj, Th));
Tw = VADD(Tt, Tv);
TG = VSUB(Tv, Tt);
Tm = VADD(Ta, Tl);
Tx = VADD(Tr, Tw);
Ty = VCONJ(VMUL(LDK(KP500000000), VSUB(Tm, Tx)));
Tz = VMUL(LDK(KP500000000), VADD(Tm, Tx));
ST(&(Rm[WS(rs, 3)]), Ty, -ms, &(Rm[WS(rs, 1)]));
ST(&(Rp[0]), Tz, ms, &(Rp[0]));
TA = VSUB(Ta, Tl);
TB = VBYI(VSUB(Tw, Tr));
TC = VCONJ(VMUL(LDK(KP500000000), VSUB(TA, TB)));
TD = VMUL(LDK(KP500000000), VADD(TA, TB));
ST(&(Rm[WS(rs, 1)]), TC, -ms, &(Rm[WS(rs, 1)]));
ST(&(Rp[WS(rs, 2)]), TD, ms, &(Rp[0]));
TH = VMUL(LDK(KP353553390), VADD(TF, TG));
TI = VADD(TE, TH);
TO = VSUB(TE, TH);
TJ = VMUL(LDK(KP707106781), VSUB(TG, TF));
TL = VMUL(LDK(KP500000000), VBYI(VSUB(TJ, TK)));
TP = VMUL(LDK(KP500000000), VBYI(VADD(TK, TJ)));
TM = VCONJ(VSUB(TI, TL));
ST(&(Rm[0]), TM, -ms, &(Rm[0]));
TR = VADD(TO, TP);
ST(&(Rp[WS(rs, 3)]), TR, ms, &(Rp[WS(rs, 1)]));
TN = VADD(TI, TL);
ST(&(Rp[WS(rs, 1)]), TN, ms, &(Rp[WS(rs, 1)]));
TQ = VCONJ(VSUB(TO, TP));
ST(&(Rm[WS(rs, 2)]), TQ, -ms, &(Rm[0]));
}
}
VLEAVE();
}
static void t3bv_8(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP707106781, +0.707106781186547524400844362104849039284835938);
{
INT m;
R *x;
x = ii;
for (m = mb, W = W + (mb * ((TWVL / VL) * 6)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 6), MAKE_VOLATILE_STRIDE(8, rs)) {
V T2, T3, Tb, T1, T5, Tn, Tq, T8, Td, T4, Ta, Tp, Tg, Ti, T9;
T2 = LDW(&(W[0]));
T3 = LDW(&(W[TWVL * 2]));
Tb = LDW(&(W[TWVL * 4]));
T1 = LD(&(x[0]), ms, &(x[0]));
T5 = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
Tn = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
Tq = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
T8 = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
Td = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
T4 = VZMUL(T2, T3);
Ta = VZMULJ(T2, T3);
Tp = VZMULJ(T2, Tb);
Tg = LD(&(x[WS(rs, 7)]), ms, &(x[WS(rs, 1)]));
Ti = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
T9 = VZMUL(T2, T8);
{
V T6, To, Tc, Tr, Th, Tj;
T6 = VZMUL(T4, T5);
To = VZMUL(Ta, Tn);
Tc = VZMULJ(Ta, Tb);
Tr = VZMUL(Tp, Tq);
Th = VZMUL(Tb, Tg);
Tj = VZMUL(T3, Ti);
{
V Tx, T7, Te, Ts, Ty, Tk, TB;
Tx = VADD(T1, T6);
T7 = VSUB(T1, T6);
Te = VZMUL(Tc, Td);
Ts = VSUB(To, Tr);
Ty = VADD(To, Tr);
Tk = VSUB(Th, Tj);
TB = VADD(Th, Tj);
{
V Tf, TA, Tz, TD;
Tf = VSUB(T9, Te);
TA = VADD(T9, Te);
Tz = VSUB(Tx, Ty);
TD = VADD(Tx, Ty);
{
V TC, TE, Tl, Tt;
TC = VSUB(TA, TB);
TE = VADD(TA, TB);
Tl = VADD(Tf, Tk);
Tt = VSUB(Tf, Tk);
{
V Tu, Tw, Tm, Tv;
ST(&(x[0]), VADD(TD, TE), ms, &(x[0]));
ST(&(x[WS(rs, 4)]), VSUB(TD, TE), ms, &(x[0]));
ST(&(x[WS(rs, 2)]), VFMAI(TC, Tz), ms, &(x[0]));
ST(&(x[WS(rs, 6)]), VFNMSI(TC, Tz), ms, &(x[0]));
Tu = VFNMS(LDK(KP707106781), Tt, Ts);
Tw = VFMA(LDK(KP707106781), Tt, Ts);
Tm = VFNMS(LDK(KP707106781), Tl, T7);
Tv = VFMA(LDK(KP707106781), Tl, T7);
ST(&(x[WS(rs, 1)]), VFMAI(Tw, Tv), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 7)]), VFNMSI(Tw, Tv), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 5)]), VFMAI(Tu, Tm), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 3)]), VFNMSI(Tu, Tm), ms, &(x[WS(rs, 1)]));
}
}
}
}
}
}
}
VLEAVE();
}
static void t3bv_8(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP707106781, +0.707106781186547524400844362104849039284835938);
{
INT m;
R *x;
x = ii;
for (m = mb, W = W + (mb * ((TWVL / VL) * 6)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 6), MAKE_VOLATILE_STRIDE(8, rs)) {
V T1, T4, T5, Tp, T6, T7, Tj;
T1 = LDW(&(W[0]));
T4 = LDW(&(W[TWVL * 2]));
T5 = VZMULJ(T1, T4);
Tp = VZMUL(T1, T4);
T6 = LDW(&(W[TWVL * 4]));
T7 = VZMULJ(T5, T6);
Tj = VZMULJ(T1, T6);
{
V Ts, Tx, Tm, Ty, Ta, TA, Tf, TB, To, Tr, Tq;
To = LD(&(x[0]), ms, &(x[0]));
Tq = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
Tr = VZMUL(Tp, Tq);
Ts = VSUB(To, Tr);
Tx = VADD(To, Tr);
{
V Ti, Tl, Th, Tk;
Th = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
Ti = VZMUL(T5, Th);
Tk = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
Tl = VZMUL(Tj, Tk);
Tm = VSUB(Ti, Tl);
Ty = VADD(Ti, Tl);
}
{
V T3, T9, T2, T8;
T2 = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
T3 = VZMUL(T1, T2);
T8 = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
T9 = VZMUL(T7, T8);
Ta = VSUB(T3, T9);
TA = VADD(T3, T9);
}
{
V Tc, Te, Tb, Td;
Tb = LD(&(x[WS(rs, 7)]), ms, &(x[WS(rs, 1)]));
Tc = VZMUL(T6, Tb);
Td = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
Te = VZMUL(T4, Td);
Tf = VSUB(Tc, Te);
TB = VADD(Tc, Te);
}
{
V Tz, TC, TD, TE;
Tz = VSUB(Tx, Ty);
TC = VBYI(VSUB(TA, TB));
ST(&(x[WS(rs, 6)]), VSUB(Tz, TC), ms, &(x[0]));
ST(&(x[WS(rs, 2)]), VADD(Tz, TC), ms, &(x[0]));
TD = VADD(Tx, Ty);
TE = VADD(TA, TB);
ST(&(x[WS(rs, 4)]), VSUB(TD, TE), ms, &(x[0]));
ST(&(x[0]), VADD(TD, TE), ms, &(x[0]));
{
V Tn, Tv, Tu, Tw, Tg, Tt;
Tg = VMUL(LDK(KP707106781), VSUB(Ta, Tf));
Tn = VBYI(VSUB(Tg, Tm));
Tv = VBYI(VADD(Tm, Tg));
Tt = VMUL(LDK(KP707106781), VADD(Ta, Tf));
Tu = VSUB(Ts, Tt);
Tw = VADD(Ts, Tt);
ST(&(x[WS(rs, 3)]), VADD(Tn, Tu), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 7)]), VSUB(Tw, Tv), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 5)]), VSUB(Tu, Tn), ms, &(x[WS(rs, 1)]));
ST(&(x[WS(rs, 1)]), VADD(Tv, Tw), ms, &(x[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_10(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
DVK(KP587785252, +0.587785252292473129168705954639072768597652438);
DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
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 T5, TE, Ts, Tt, TC, Tz, TH, TJ, To, Tq, T2, T4, T3, T9, Tx;
V Tm, TB, Td, Ty, Ti, TA, T6, T8, T7, Tl, Tk, Tj, Tc, Tb, Ta;
V Tf, Th, Tg, TF, TG, Te, Tn;
T2 = LD(&(Rp[0]), ms, &(Rp[0]));
T3 = LD(&(Rm[WS(rs, 4)]), -ms, &(Rm[0]));
T4 = VCONJ(T3);
T5 = VSUB(T2, T4);
TE = VADD(T2, T4);
T6 = LD(&(Rp[WS(rs, 2)]), ms, &(Rp[0]));
T7 = LD(&(Rm[WS(rs, 2)]), -ms, &(Rm[0]));
T8 = VCONJ(T7);
T9 = VSUB(T6, T8);
Tx = VADD(T6, T8);
Tl = LD(&(Rp[WS(rs, 1)]), ms, &(Rp[WS(rs, 1)]));
Tj = LD(&(Rm[WS(rs, 3)]), -ms, &(Rm[WS(rs, 1)]));
Tk = VCONJ(Tj);
Tm = VSUB(Tk, Tl);
TB = VADD(Tk, Tl);
Tc = LD(&(Rp[WS(rs, 3)]), ms, &(Rp[WS(rs, 1)]));
Ta = LD(&(Rm[WS(rs, 1)]), -ms, &(Rm[WS(rs, 1)]));
Tb = VCONJ(Ta);
Td = VSUB(Tb, Tc);
Ty = VADD(Tb, Tc);
Tf = LD(&(Rp[WS(rs, 4)]), ms, &(Rp[0]));
Tg = LD(&(Rm[0]), -ms, &(Rm[0]));
Th = VCONJ(Tg);
Ti = VSUB(Tf, Th);
TA = VADD(Tf, Th);
Ts = VSUB(T9, Td);
Tt = VSUB(Ti, Tm);
TC = VSUB(TA, TB);
Tz = VSUB(Tx, Ty);
TF = VADD(Tx, Ty);
TG = VADD(TA, TB);
TH = VADD(TF, TG);
TJ = VMUL(LDK(KP559016994), VSUB(TF, TG));
Te = VADD(T9, Td);
Tn = VADD(Ti, Tm);
To = VADD(Te, Tn);
Tq = VMUL(LDK(KP559016994), VSUB(Te, Tn));
{
V T1c, TX, Tv, T1b, TR, T15, TL, T17, TT, T11, TW, Tu, TQ, Tr, TP;
V Tp, T1, T1a, TO, T14, TD, T10, TK, TZ, TI, Tw, T16, TS, TY, TM;
V TU, T1e, TN, T1d, T19, T13, TV, T18, T12;
T1c = VADD(TE, TH);
TW = LDW(&(W[TWVL * 8]));
TX = VZMULI(TW, VADD(T5, To));
Tu = VBYI(VFNMS(LDK(KP951056516), Tt, VMUL(LDK(KP587785252), Ts)));
TQ = VBYI(VFMA(LDK(KP951056516), Ts, VMUL(LDK(KP587785252), Tt)));
Tp = VFNMS(LDK(KP250000000), To, T5);
Tr = VSUB(Tp, Tq);
TP = VADD(Tq, Tp);
T1 = LDW(&(W[TWVL * 4]));
Tv = VZMULI(T1, VSUB(Tr, Tu));
T1a = LDW(&(W[0]));
T1b = VZMULI(T1a, VADD(TQ, TP));
TO = LDW(&(W[TWVL * 16]));
TR = VZMULI(TO, VSUB(TP, TQ));
T14 = LDW(&(W[TWVL * 12]));
T15 = VZMULI(T14, VADD(Tu, Tr));
TD = VBYI(VFNMS(LDK(KP951056516), TC, VMUL(LDK(KP587785252), Tz)));
T10 = VBYI(VFMA(LDK(KP951056516), Tz, VMUL(LDK(KP587785252), TC)));
TI = VFNMS(LDK(KP250000000), TH, TE);
TK = VSUB(TI, TJ);
TZ = VADD(TJ, TI);
Tw = LDW(&(W[TWVL * 2]));
TL = VZMUL(Tw, VADD(TD, TK));
T16 = LDW(&(W[TWVL * 10]));
T17 = VZMUL(T16, VADD(T10, TZ));
TS = LDW(&(W[TWVL * 14]));
TT = VZMUL(TS, VSUB(TK, TD));
TY = LDW(&(W[TWVL * 6]));
T11 = VZMUL(TY, VSUB(TZ, T10));
TM = VADD(Tv, TL);
ST(&(Rp[WS(rs, 1)]), TM, ms, &(Rp[WS(rs, 1)]));
TU = VADD(TR, TT);
ST(&(Rp[WS(rs, 4)]), TU, ms, &(Rp[0]));
T1e = VCONJ(VSUB(T1c, T1b));
ST(&(Rm[0]), T1e, -ms, &(Rm[0]));
TN = VCONJ(VSUB(TL, Tv));
ST(&(Rm[WS(rs, 1)]), TN, -ms, &(Rm[WS(rs, 1)]));
T1d = VADD(T1b, T1c);
ST(&(Rp[0]), T1d, ms, &(Rp[0]));
T19 = VCONJ(VSUB(T17, T15));
ST(&(Rm[WS(rs, 3)]), T19, -ms, &(Rm[WS(rs, 1)]));
T13 = VCONJ(VSUB(T11, TX));
ST(&(Rm[WS(rs, 2)]), T13, -ms, &(Rm[0]));
TV = VCONJ(VSUB(TT, TR));
ST(&(Rm[WS(rs, 4)]), TV, -ms, &(Rm[0]));
T18 = VADD(T15, T17);
ST(&(Rp[WS(rs, 3)]), T18, ms, &(Rp[WS(rs, 1)]));
T12 = VADD(TX, T11);
ST(&(Rp[WS(rs, 2)]), T12, ms, &(Rp[0]));
}
}
}
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();
}
