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 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 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 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 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 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_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_16(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms)
{
DVK(KP382683432, +0.382683432365089771728459984030398866761344562);
DVK(KP923879532, +0.923879532511286756128183189396788286822416626);
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 Tf, T16, TZ, T1C, TI, T1a, TV, T1D, T1F, T1G, Ty, T19, TC, T17, TS;
V T10;
{
V T2, TD, T4, TF, Tc, Tb, Td, T6, T8, T9, T3, TE, Ta, T7, T5;
V Te, TX, TY, TG, TH, TT, TU, Tj, TM, Tw, TQ, Tn, TN, Ts, TP;
V Tg, Ti, Th, Tt, Tv, Tu, Tk, Tm, Tl, Tr, Tq, Tp, To, Tx, TA;
V TB, TO, TR;
T2 = LD(&(Rp[0]), ms, &(Rp[0]));
TD = LD(&(Rp[WS(rs, 4)]), ms, &(Rp[0]));
T3 = LD(&(Rm[WS(rs, 7)]), -ms, &(Rm[WS(rs, 1)]));
T4 = VCONJ(T3);
TE = LD(&(Rm[WS(rs, 3)]), -ms, &(Rm[WS(rs, 1)]));
TF = VCONJ(TE);
Tc = LD(&(Rp[WS(rs, 6)]), ms, &(Rp[0]));
Ta = LD(&(Rm[WS(rs, 1)]), -ms, &(Rm[WS(rs, 1)]));
Tb = VCONJ(Ta);
Td = VSUB(Tb, Tc);
T6 = LD(&(Rp[WS(rs, 2)]), ms, &(Rp[0]));
T7 = LD(&(Rm[WS(rs, 5)]), -ms, &(Rm[WS(rs, 1)]));
T8 = VCONJ(T7);
T9 = VSUB(T6, T8);
T5 = VSUB(T2, T4);
Te = VMUL(LDK(KP707106781), VADD(T9, Td));
Tf = VADD(T5, Te);
T16 = VSUB(T5, Te);
TX = VADD(T2, T4);
TY = VADD(TD, TF);
TZ = VSUB(TX, TY);
T1C = VADD(TX, TY);
TG = VSUB(TD, TF);
TH = VMUL(LDK(KP707106781), VSUB(T9, Td));
TI = VADD(TG, TH);
T1a = VSUB(TH, TG);
TT = VADD(T6, T8);
TU = VADD(Tb, Tc);
TV = VSUB(TT, TU);
T1D = VADD(TT, TU);
Tg = LD(&(Rp[WS(rs, 1)]), ms, &(Rp[WS(rs, 1)]));
Th = LD(&(Rm[WS(rs, 6)]), -ms, &(Rm[0]));
Ti = VCONJ(Th);
Tj = VSUB(Tg, Ti);
TM = VADD(Tg, Ti);
Tt = LD(&(Rp[WS(rs, 3)]), ms, &(Rp[WS(rs, 1)]));
Tu = LD(&(Rm[WS(rs, 4)]), -ms, &(Rm[0]));
Tv = VCONJ(Tu);
Tw = VSUB(Tt, Tv);
TQ = VADD(Tt, Tv);
Tk = LD(&(Rp[WS(rs, 5)]), ms, &(Rp[WS(rs, 1)]));
Tl = LD(&(Rm[WS(rs, 2)]), -ms, &(Rm[0]));
Tm = VCONJ(Tl);
Tn = VSUB(Tk, Tm);
TN = VADD(Tk, Tm);
Tr = LD(&(Rp[WS(rs, 7)]), ms, &(Rp[WS(rs, 1)]));
Tp = LD(&(Rm[0]), -ms, &(Rm[0]));
Tq = VCONJ(Tp);
Ts = VSUB(Tq, Tr);
TP = VADD(Tq, Tr);
T1F = VADD(TM, TN);
T1G = VADD(TP, TQ);
To = VFNMS(LDK(KP382683432), Tn, VMUL(LDK(KP923879532), Tj));
Tx = VFMA(LDK(KP923879532), Ts, VMUL(LDK(KP382683432), Tw));
Ty = VADD(To, Tx);
T19 = VSUB(To, Tx);
TA = VFMA(LDK(KP382683432), Tj, VMUL(LDK(KP923879532), Tn));
TB = VFNMS(LDK(KP382683432), Ts, VMUL(LDK(KP923879532), Tw));
TC = VADD(TA, TB);
T17 = VSUB(TA, TB);
TO = VSUB(TM, TN);
TR = VSUB(TP, TQ);
TS = VMUL(LDK(KP707106781), VSUB(TO, TR));
T10 = VMUL(LDK(KP707106781), VADD(TO, TR));
}
{
V T21, T1W, T1u, T20, T1I, T1O, TK, T1S, T12, T1e, T1k, T1A, T1o, T1w, T1c;
V T1M, T1U, T1V, T1T, T1s, T1t, T1r, T1Z, T1E, T1H, T1B, T1N, Tz, TJ, T1;
V T1R, TW, T11, TL, T1d, T1i, T1j, T1h, T1z, T1m, T1n, T1l, T1v, T18, T1b;
V T15, T1L, T13, T1g, T1X, T23, T14, T1f, T1Y, T22, T1p, T1y, T1J, T1Q, T1q;
V T1x, T1K, T1P;
T1U = VADD(T1C, T1D);
T1V = VADD(T1F, T1G);
T21 = VADD(T1U, T1V);
T1T = LDW(&(W[TWVL * 14]));
T1W = VZMUL(T1T, VSUB(T1U, T1V));
T1s = VADD(Tf, Ty);
T1t = VBYI(VADD(TI, TC));
T1r = LDW(&(W[TWVL * 28]));
T1u = VZMULI(T1r, VSUB(T1s, T1t));
T1Z = LDW(&(W[0]));
T20 = VZMULI(T1Z, VADD(T1s, T1t));
T1E = VSUB(T1C, T1D);
T1H = VBYI(VSUB(T1F, T1G));
T1B = LDW(&(W[TWVL * 22]));
T1I = VZMUL(T1B, VSUB(T1E, T1H));
T1N = LDW(&(W[TWVL * 6]));
T1O = VZMUL(T1N, VADD(T1E, T1H));
Tz = VSUB(Tf, Ty);
TJ = VBYI(VSUB(TC, TI));
T1 = LDW(&(W[TWVL * 12]));
TK = VZMULI(T1, VADD(Tz, TJ));
T1R = LDW(&(W[TWVL * 16]));
T1S = VZMULI(T1R, VSUB(Tz, TJ));
TW = VBYI(VSUB(TS, TV));
T11 = VSUB(TZ, T10);
TL = LDW(&(W[TWVL * 10]));
T12 = VZMUL(TL, VADD(TW, T11));
T1d = LDW(&(W[TWVL * 18]));
T1e = VZMUL(T1d, VSUB(T11, TW));
T1i = VBYI(VADD(T1a, T19));
T1j = VADD(T16, T17);
T1h = LDW(&(W[TWVL * 4]));
T1k = VZMULI(T1h, VADD(T1i, T1j));
T1z = LDW(&(W[TWVL * 24]));
T1A = VZMULI(T1z, VSUB(T1j, T1i));
T1m = VBYI(VADD(TV, TS));
T1n = VADD(TZ, T10);
T1l = LDW(&(W[TWVL * 2]));
T1o = VZMUL(T1l, VADD(T1m, T1n));
T1v = LDW(&(W[TWVL * 26]));
T1w = VZMUL(T1v, VSUB(T1n, T1m));
T18 = VSUB(T16, T17);
T1b = VBYI(VSUB(T19, T1a));
T15 = LDW(&(W[TWVL * 20]));
T1c = VZMULI(T15, VSUB(T18, T1b));
T1L = LDW(&(W[TWVL * 8]));
T1M = VZMULI(T1L, VADD(T1b, T18));
T13 = VADD(TK, T12);
ST(&(Rp[WS(rs, 3)]), T13, ms, &(Rp[WS(rs, 1)]));
T1g = VCONJ(VSUB(T1e, T1c));
ST(&(Rm[WS(rs, 5)]), T1g, -ms, &(Rm[WS(rs, 1)]));
T1X = VADD(T1S, T1W);
ST(&(Rp[WS(rs, 4)]), T1X, ms, &(Rp[0]));
T23 = VCONJ(VSUB(T21, T20));
ST(&(Rm[0]), T23, -ms, &(Rm[0]));
T14 = VCONJ(VSUB(T12, TK));
ST(&(Rm[WS(rs, 3)]), T14, -ms, &(Rm[WS(rs, 1)]));
T1f = VADD(T1c, T1e);
ST(&(Rp[WS(rs, 5)]), T1f, ms, &(Rp[WS(rs, 1)]));
T1Y = VCONJ(VSUB(T1W, T1S));
ST(&(Rm[WS(rs, 4)]), T1Y, -ms, &(Rm[0]));
T22 = VADD(T20, T21);
ST(&(Rp[0]), T22, ms, &(Rp[0]));
T1p = VADD(T1k, T1o);
ST(&(Rp[WS(rs, 1)]), T1p, ms, &(Rp[WS(rs, 1)]));
T1y = VCONJ(VSUB(T1w, T1u));
ST(&(Rm[WS(rs, 7)]), T1y, -ms, &(Rm[WS(rs, 1)]));
T1J = VADD(T1A, T1I);
ST(&(Rp[WS(rs, 6)]), T1J, ms, &(Rp[0]));
T1Q = VCONJ(VSUB(T1O, T1M));
ST(&(Rm[WS(rs, 2)]), T1Q, -ms, &(Rm[0]));
T1q = VCONJ(VSUB(T1o, T1k));
ST(&(Rm[WS(rs, 1)]), T1q, -ms, &(Rm[WS(rs, 1)]));
T1x = VADD(T1u, T1w);
ST(&(Rp[WS(rs, 7)]), T1x, ms, &(Rp[WS(rs, 1)]));
T1K = VCONJ(VSUB(T1I, T1A));
ST(&(Rm[WS(rs, 6)]), T1K, -ms, &(Rm[0]));
T1P = VADD(T1M, T1O);
ST(&(Rp[WS(rs, 2)]), T1P, 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 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 T5, Tj, Tq, TI, Te, Tk, Tt, TJ, T2, Tg, T4, Ti, T3, Th, To;
V Tp, T6, Tc, T8, Tb, T7, Ta, T9, Td, Tr, Ts, TP, Tu, Tm, TO;
V Tn, Tf, Tl, T1, TN, Tv, TR, Tw, TQ, TC, TK, TA, TG, TB, TH;
V Ty, Tz, Tx, TF, TD, TM, TE, TL;
T2 = LD(&(Rp[0]), ms, &(Rp[0]));
Tg = LD(&(Rp[WS(rs, 2)]), ms, &(Rp[0]));
T3 = LD(&(Rm[WS(rs, 3)]), -ms, &(Rm[WS(rs, 1)]));
T4 = VCONJ(T3);
Th = LD(&(Rm[WS(rs, 1)]), -ms, &(Rm[WS(rs, 1)]));
Ti = VCONJ(Th);
T5 = VSUB(T2, T4);
Tj = VSUB(Tg, Ti);
To = VADD(T2, T4);
Tp = VADD(Tg, Ti);
Tq = VSUB(To, Tp);
TI = VADD(To, Tp);
T6 = LD(&(Rp[WS(rs, 1)]), ms, &(Rp[WS(rs, 1)]));
Tc = LD(&(Rp[WS(rs, 3)]), ms, &(Rp[WS(rs, 1)]));
T7 = LD(&(Rm[WS(rs, 2)]), -ms, &(Rm[0]));
T8 = VCONJ(T7);
Ta = LD(&(Rm[0]), -ms, &(Rm[0]));
Tb = VCONJ(Ta);
T9 = VSUB(T6, T8);
Td = VSUB(Tb, Tc);
Te = VMUL(LDK(KP707106781), VADD(T9, Td));
Tk = VMUL(LDK(KP707106781), VSUB(T9, Td));
Tr = VADD(T6, T8);
Ts = VADD(Tb, Tc);
Tt = VBYI(VSUB(Tr, Ts));
TJ = VADD(Tr, Ts);
TP = VADD(TI, TJ);
Tn = LDW(&(W[TWVL * 10]));
Tu = VZMUL(Tn, VSUB(Tq, Tt));
Tf = VADD(T5, Te);
Tl = VBYI(VADD(Tj, Tk));
T1 = LDW(&(W[TWVL * 12]));
Tm = VZMULI(T1, VSUB(Tf, Tl));
TN = LDW(&(W[0]));
TO = VZMULI(TN, VADD(Tl, Tf));
Tv = VADD(Tm, Tu);
ST(&(Rp[WS(rs, 3)]), Tv, ms, &(Rp[WS(rs, 1)]));
TR = VCONJ(VSUB(TP, TO));
ST(&(Rm[0]), TR, -ms, &(Rm[0]));
Tw = VCONJ(VSUB(Tu, Tm));
ST(&(Rm[WS(rs, 3)]), Tw, -ms, &(Rm[WS(rs, 1)]));
TQ = VADD(TO, TP);
ST(&(Rp[0]), TQ, ms, &(Rp[0]));
TB = LDW(&(W[TWVL * 2]));
TC = VZMUL(TB, VADD(Tq, Tt));
TH = LDW(&(W[TWVL * 6]));
TK = VZMUL(TH, VSUB(TI, TJ));
Ty = VBYI(VSUB(Tk, Tj));
Tz = VSUB(T5, Te);
Tx = LDW(&(W[TWVL * 4]));
TA = VZMULI(Tx, VADD(Ty, Tz));
TF = LDW(&(W[TWVL * 8]));
TG = VZMULI(TF, VSUB(Tz, Ty));
TD = VADD(TA, TC);
ST(&(Rp[WS(rs, 1)]), TD, ms, &(Rp[WS(rs, 1)]));
TM = VCONJ(VSUB(TK, TG));
ST(&(Rm[WS(rs, 2)]), TM, -ms, &(Rm[0]));
TE = VCONJ(VSUB(TC, TA));
ST(&(Rm[WS(rs, 1)]), TE, -ms, &(Rm[WS(rs, 1)]));
TL = VADD(TG, TK);
ST(&(Rp[WS(rs, 2)]), TL, ms, &(Rp[0]));
}
}
