267 lines
6.8 KiB
C
267 lines
6.8 KiB
C
|
/*
|
||
|
|
* Copyright (c) 2003, 2007-14 Matteo Frigo
|
||
|
|
* Copyright (c) 2003, 2007-14 Massachusetts Institute of Technology
|
||
|
|
*
|
||
|
|
* This program is free software; you can redistribute it and/or modify
|
||
|
|
* it under the terms of the GNU General Public License as published by
|
||
|
|
* the Free Software Foundation; either version 2 of the License, or
|
||
|
|
* (at your option) any later version.
|
||
|
|
*
|
||
|
|
* This program is distributed in the hope that it will be useful,
|
||
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
||
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
||
|
|
* GNU General Public License for more details.
|
||
|
|
*
|
||
|
|
* You should have received a copy of the GNU General Public License
|
||
|
|
* along with this program; if not, write to the Free Software
|
||
|
|
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
|
||
|
|
*
|
||
|
|
*/
|
||
|
|
|
||
|
|
/* This file was automatically generated --- DO NOT EDIT */
|
||
|
|
/* Generated on Tue Sep 14 10:44:24 EDT 2021 */
|
||
|
|
|
||
|
|
#include "dft/codelet-dft.h"
|
||
|
|
|
||
|
|
#if defined(ARCH_PREFERS_FMA) || defined(ISA_EXTENSION_PREFERS_FMA)
|
||
|
|
|
||
|
|
/* Generated by: ../../../genfft/gen_notw.native -fma -compact -variables 4 -pipeline-latency 4 -n 8 -name n1_8 -include dft/scalar/n.h */
|
||
|
|
|
||
|
|
/*
|
||
|
|
* This function contains 52 FP additions, 8 FP multiplications,
|
||
|
|
* (or, 44 additions, 0 multiplications, 8 fused multiply/add),
|
||
|
|
* 28 stack variables, 1 constants, and 32 memory accesses
|
||
|
|
*/
|
||
|
|
#include "dft/scalar/n.h"
|
||
|
|
|
||
|
|
static void n1_8(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
|
||
|
|
{
|
||
|
|
DK(KP707106781, +0.707106781186547524400844362104849039284835938);
|
||
|
|
{
|
||
|
|
INT i;
|
||
|
|
for (i = v; i > 0; i = i - 1, ri = ri + ivs, ii = ii + ivs, ro = ro + ovs, io = io + ovs, MAKE_VOLATILE_STRIDE(32, is), MAKE_VOLATILE_STRIDE(32, os)) {
|
||
|
|
E T3, Tn, Ti, TC, T6, TB, Tl, To, Td, TN, Tz, TH, Ta, TM, Tu;
|
||
|
|
E TG;
|
||
|
|
{
|
||
|
|
E T1, T2, Tj, Tk;
|
||
|
|
T1 = ri[0];
|
||
|
|
T2 = ri[WS(is, 4)];
|
||
|
|
T3 = T1 + T2;
|
||
|
|
Tn = T1 - T2;
|
||
|
|
{
|
||
|
|
E Tg, Th, T4, T5;
|
||
|
|
Tg = ii[0];
|
||
|
|
Th = ii[WS(is, 4)];
|
||
|
|
Ti = Tg + Th;
|
||
|
|
TC = Tg - Th;
|
||
|
|
T4 = ri[WS(is, 2)];
|
||
|
|
T5 = ri[WS(is, 6)];
|
||
|
|
T6 = T4 + T5;
|
||
|
|
TB = T4 - T5;
|
||
|
|
}
|
||
|
|
Tj = ii[WS(is, 2)];
|
||
|
|
Tk = ii[WS(is, 6)];
|
||
|
|
Tl = Tj + Tk;
|
||
|
|
To = Tj - Tk;
|
||
|
|
{
|
||
|
|
E Tb, Tc, Tv, Tw, Tx, Ty;
|
||
|
|
Tb = ri[WS(is, 7)];
|
||
|
|
Tc = ri[WS(is, 3)];
|
||
|
|
Tv = Tb - Tc;
|
||
|
|
Tw = ii[WS(is, 7)];
|
||
|
|
Tx = ii[WS(is, 3)];
|
||
|
|
Ty = Tw - Tx;
|
||
|
|
Td = Tb + Tc;
|
||
|
|
TN = Tw + Tx;
|
||
|
|
Tz = Tv - Ty;
|
||
|
|
TH = Tv + Ty;
|
||
|
|
}
|
||
|
|
{
|
||
|
|
E T8, T9, Tq, Tr, Ts, Tt;
|
||
|
|
T8 = ri[WS(is, 1)];
|
||
|
|
T9 = ri[WS(is, 5)];
|
||
|
|
Tq = T8 - T9;
|
||
|
|
Tr = ii[WS(is, 1)];
|
||
|
|
Ts = ii[WS(is, 5)];
|
||
|
|
Tt = Tr - Ts;
|
||
|
|
Ta = T8 + T9;
|
||
|
|
TM = Tr + Ts;
|
||
|
|
Tu = Tq + Tt;
|
||
|
|
TG = Tt - Tq;
|
||
|
|
}
|
||
|
|
}
|
||
|
|
{
|
||
|
|
E T7, Te, TP, TQ;
|
||
|
|
T7 = T3 + T6;
|
||
|
|
Te = Ta + Td;
|
||
|
|
ro[WS(os, 4)] = T7 - Te;
|
||
|
|
ro[0] = T7 + Te;
|
||
|
|
TP = Ti + Tl;
|
||
|
|
TQ = TM + TN;
|
||
|
|
io[WS(os, 4)] = TP - TQ;
|
||
|
|
io[0] = TP + TQ;
|
||
|
|
}
|
||
|
|
{
|
||
|
|
E Tf, Tm, TL, TO;
|
||
|
|
Tf = Td - Ta;
|
||
|
|
Tm = Ti - Tl;
|
||
|
|
io[WS(os, 2)] = Tf + Tm;
|
||
|
|
io[WS(os, 6)] = Tm - Tf;
|
||
|
|
TL = T3 - T6;
|
||
|
|
TO = TM - TN;
|
||
|
|
ro[WS(os, 6)] = TL - TO;
|
||
|
|
ro[WS(os, 2)] = TL + TO;
|
||
|
|
}
|
||
|
|
{
|
||
|
|
E Tp, TA, TJ, TK;
|
||
|
|
Tp = Tn + To;
|
||
|
|
TA = Tu + Tz;
|
||
|
|
ro[WS(os, 5)] = FNMS(KP707106781, TA, Tp);
|
||
|
|
ro[WS(os, 1)] = FMA(KP707106781, TA, Tp);
|
||
|
|
TJ = TC - TB;
|
||
|
|
TK = TG + TH;
|
||
|
|
io[WS(os, 5)] = FNMS(KP707106781, TK, TJ);
|
||
|
|
io[WS(os, 1)] = FMA(KP707106781, TK, TJ);
|
||
|
|
}
|
||
|
|
{
|
||
|
|
E TD, TE, TF, TI;
|
||
|
|
TD = TB + TC;
|
||
|
|
TE = Tz - Tu;
|
||
|
|
io[WS(os, 7)] = FNMS(KP707106781, TE, TD);
|
||
|
|
io[WS(os, 3)] = FMA(KP707106781, TE, TD);
|
||
|
|
TF = Tn - To;
|
||
|
|
TI = TG - TH;
|
||
|
|
ro[WS(os, 7)] = FNMS(KP707106781, TI, TF);
|
||
|
|
ro[WS(os, 3)] = FMA(KP707106781, TI, TF);
|
||
|
|
}
|
||
|
|
}
|
||
|
|
}
|
||
|
|
}
|
||
|
|
|
||
|
|
static const kdft_desc desc = { 8, "n1_8", { 44, 0, 8, 0 }, &GENUS, 0, 0, 0, 0 };
|
||
|
|
|
||
|
|
void X(codelet_n1_8) (planner *p) { X(kdft_register) (p, n1_8, &desc);
|
||
|
|
}
|
||
|
|
|
||
|
|
#else
|
||
|
|
|
||
|
|
/* Generated by: ../../../genfft/gen_notw.native -compact -variables 4 -pipeline-latency 4 -n 8 -name n1_8 -include dft/scalar/n.h */
|
||
|
|
|
||
|
|
/*
|
||
|
|
* This function contains 52 FP additions, 4 FP multiplications,
|
||
|
|
* (or, 52 additions, 4 multiplications, 0 fused multiply/add),
|
||
|
|
* 28 stack variables, 1 constants, and 32 memory accesses
|
||
|
|
*/
|
||
|
|
#include "dft/scalar/n.h"
|
||
|
|
|
||
|
|
static void n1_8(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
|
||
|
|
{
|
||
|
|
DK(KP707106781, +0.707106781186547524400844362104849039284835938);
|
||
|
|
{
|
||
|
|
INT i;
|
||
|
|
for (i = v; i > 0; i = i - 1, ri = ri + ivs, ii = ii + ivs, ro = ro + ovs, io = io + ovs, MAKE_VOLATILE_STRIDE(32, is), MAKE_VOLATILE_STRIDE(32, os)) {
|
||
|
|
E T3, Tn, Ti, TC, T6, TB, Tl, To, Td, TN, Tz, TH, Ta, TM, Tu;
|
||
|
|
E TG;
|
||
|
|
{
|
||
|
|
E T1, T2, Tj, Tk;
|
||
|
|
T1 = ri[0];
|
||
|
|
T2 = ri[WS(is, 4)];
|
||
|
|
T3 = T1 + T2;
|
||
|
|
Tn = T1 - T2;
|
||
|
|
{
|
||
|
|
E Tg, Th, T4, T5;
|
||
|
|
Tg = ii[0];
|
||
|
|
Th = ii[WS(is, 4)];
|
||
|
|
Ti = Tg + Th;
|
||
|
|
TC = Tg - Th;
|
||
|
|
T4 = ri[WS(is, 2)];
|
||
|
|
T5 = ri[WS(is, 6)];
|
||
|
|
T6 = T4 + T5;
|
||
|
|
TB = T4 - T5;
|
||
|
|
}
|
||
|
|
Tj = ii[WS(is, 2)];
|
||
|
|
Tk = ii[WS(is, 6)];
|
||
|
|
Tl = Tj + Tk;
|
||
|
|
To = Tj - Tk;
|
||
|
|
{
|
||
|
|
E Tb, Tc, Tv, Tw, Tx, Ty;
|
||
|
|
Tb = ri[WS(is, 7)];
|
||
|
|
Tc = ri[WS(is, 3)];
|
||
|
|
Tv = Tb - Tc;
|
||
|
|
Tw = ii[WS(is, 7)];
|
||
|
|
Tx = ii[WS(is, 3)];
|
||
|
|
Ty = Tw - Tx;
|
||
|
|
Td = Tb + Tc;
|
||
|
|
TN = Tw + Tx;
|
||
|
|
Tz = Tv - Ty;
|
||
|
|
TH = Tv + Ty;
|
||
|
|
}
|
||
|
|
{
|
||
|
|
E T8, T9, Tq, Tr, Ts, Tt;
|
||
|
|
T8 = ri[WS(is, 1)];
|
||
|
|
T9 = ri[WS(is, 5)];
|
||
|
|
Tq = T8 - T9;
|
||
|
|
Tr = ii[WS(is, 1)];
|
||
|
|
Ts = ii[WS(is, 5)];
|
||
|
|
Tt = Tr - Ts;
|
||
|
|
Ta = T8 + T9;
|
||
|
|
TM = Tr + Ts;
|
||
|
|
Tu = Tq + Tt;
|
||
|
|
TG = Tt - Tq;
|
||
|
|
}
|
||
|
|
}
|
||
|
|
{
|
||
|
|
E T7, Te, TP, TQ;
|
||
|
|
T7 = T3 + T6;
|
||
|
|
Te = Ta + Td;
|
||
|
|
ro[WS(os, 4)] = T7 - Te;
|
||
|
|
ro[0] = T7 + Te;
|
||
|
|
TP = Ti + Tl;
|
||
|
|
TQ = TM + TN;
|
||
|
|
io[WS(os, 4)] = TP - TQ;
|
||
|
|
io[0] = TP + TQ;
|
||
|
|
}
|
||
|
|
{
|
||
|
|
E Tf, Tm, TL, TO;
|
||
|
|
Tf = Td - Ta;
|
||
|
|
Tm = Ti - Tl;
|
||
|
|
io[WS(os, 2)] = Tf + Tm;
|
||
|
|
io[WS(os, 6)] = Tm - Tf;
|
||
|
|
TL = T3 - T6;
|
||
|
|
TO = TM - TN;
|
||
|
|
ro[WS(os, 6)] = TL - TO;
|
||
|
|
ro[WS(os, 2)] = TL + TO;
|
||
|
|
}
|
||
|
|
{
|
||
|
|
E Tp, TA, TJ, TK;
|
||
|
|
Tp = Tn + To;
|
||
|
|
TA = KP707106781 * (Tu + Tz);
|
||
|
|
ro[WS(os, 5)] = Tp - TA;
|
||
|
|
ro[WS(os, 1)] = Tp + TA;
|
||
|
|
TJ = TC - TB;
|
||
|
|
TK = KP707106781 * (TG + TH);
|
||
|
|
io[WS(os, 5)] = TJ - TK;
|
||
|
|
io[WS(os, 1)] = TJ + TK;
|
||
|
|
}
|
||
|
|
{
|
||
|
|
E TD, TE, TF, TI;
|
||
|
|
TD = TB + TC;
|
||
|
|
TE = KP707106781 * (Tz - Tu);
|
||
|
|
io[WS(os, 7)] = TD - TE;
|
||
|
|
io[WS(os, 3)] = TD + TE;
|
||
|
|
TF = Tn - To;
|
||
|
|
TI = KP707106781 * (TG - TH);
|
||
|
|
ro[WS(os, 7)] = TF - TI;
|
||
|
|
ro[WS(os, 3)] = TF + TI;
|
||
|
|
}
|
||
|
|
}
|
||
|
|
}
|
||
|
|
}
|
||
|
|
|
||
|
|
static const kdft_desc desc = { 8, "n1_8", { 52, 4, 0, 0 }, &GENUS, 0, 0, 0, 0 };
|
||
|
|
|
||
|
|
void X(codelet_n1_8) (planner *p) { X(kdft_register) (p, n1_8, &desc);
|
||
|
|
}
|
||
|
|
|
||
|
|
#endif
|