etc/furnace
0
0
Fork 0
mirror of https://github.com/tildearrow/furnace.git synced 2024-11-24 13:35:11 +00:00
Code Issues Projects Releases Packages Wiki Activity
furnace/extern/fftw/dft/simd/common/n2fv_12.c
tildearrow 54e93db207 GUI: try using FFTW for per-chan osc wave center 
not reliable yet
2022-05-31 03:24:29 -05:00

302 lines
9.3 KiB
C
Raw Blame History

/*
* 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 10:45:12 EDT 2021 */
#include "dft/codelet-dft.h"
#if defined(ARCH_PREFERS_FMA) || defined(ISA_EXTENSION_PREFERS_FMA)
/* Generated by: ../../../genfft/gen_notw_c.native -fma -simd -compact -variables 4 -pipeline-latency 8 -n 12 -name n2fv_12 -with-ostride 2 -include dft/simd/n2f.h -store-multiple 2 */
/*
* This function contains 48 FP additions, 20 FP multiplications,
* (or, 30 additions, 2 multiplications, 18 fused multiply/add),
* 33 stack variables, 2 constants, and 30 memory accesses
*/
#include "dft/simd/n2f.h"
static void n2fv_12(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
{
DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
{
INT i;
const R *xi;
R *xo;
xi = ri;
xo = ro;
for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs), MAKE_VOLATILE_STRIDE(24, is), MAKE_VOLATILE_STRIDE(24, os)) {
V T5, Ta, TG, TF, TB, Tt, Ti, Tm, TJ, TI, TA, Tp;
{
V T1, T6, T4, Tr, T9, Ts;
T1 = LD(&(xi[0]), ivs, &(xi[0]));
T6 = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
{
V T2, T3, T7, T8;
T2 = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
T3 = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
T4 = VADD(T2, T3);
Tr = VSUB(T3, T2);
T7 = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
T8 = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
T9 = VADD(T7, T8);
Ts = VSUB(T8, T7);
}
T5 = VFNMS(LDK(KP500000000), T4, T1);
Ta = VFNMS(LDK(KP500000000), T9, T6);
TG = VADD(T6, T9);
TF = VADD(T1, T4);
TB = VADD(Tr, Ts);
Tt = VSUB(Tr, Ts);
}
{
V Tk, Tn, Te, Tl, Th, To;
Tk = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
Tn = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
{
V Tc, Td, Tf, Tg;
Tc = LD(&(xi[WS(is, 11)]), ivs, &(xi[WS(is, 1)]));
Td = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
Te = VSUB(Tc, Td);
Tl = VADD(Td, Tc);
Tf = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
Tg = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
Th = VSUB(Tf, Tg);
To = VADD(Tf, Tg);
}
Ti = VADD(Te, Th);
Tm = VFNMS(LDK(KP500000000), Tl, Tk);
TJ = VADD(Tn, To);
TI = VADD(Tk, Tl);
TA = VSUB(Te, Th);
Tp = VFNMS(LDK(KP500000000), To, Tn);
}
{
V TN, TO, TP, TQ, TT, TU;
{
V TH, TK, TL, TM;
TH = VSUB(TF, TG);
TK = VSUB(TI, TJ);
TN = VFNMSI(TK, TH);
STM2(&(xo[18]), TN, ovs, &(xo[2]));
TO = VFMAI(TK, TH);
STM2(&(xo[6]), TO, ovs, &(xo[2]));
TL = VADD(TF, TG);
TM = VADD(TI, TJ);
TP = VSUB(TL, TM);
STM2(&(xo[12]), TP, ovs, &(xo[0]));
TQ = VADD(TL, TM);
STM2(&(xo[0]), TQ, ovs, &(xo[0]));
}
{
V Tj, Tv, Tu, Tw, Tb, Tq, TR, TS;
Tb = VSUB(T5, Ta);
Tj = VFMA(LDK(KP866025403), Ti, Tb);
Tv = VFNMS(LDK(KP866025403), Ti, Tb);
Tq = VSUB(Tm, Tp);
Tu = VFNMS(LDK(KP866025403), Tt, Tq);
Tw = VFMA(LDK(KP866025403), Tt, Tq);
TR = VFNMSI(Tu, Tj);
STM2(&(xo[2]), TR, ovs, &(xo[2]));
STN2(&(xo[0]), TQ, TR, ovs);
TS = VFMAI(Tw, Tv);
STM2(&(xo[14]), TS, ovs, &(xo[2]));
STN2(&(xo[12]), TP, TS, ovs);
TT = VFMAI(Tu, Tj);
STM2(&(xo[22]), TT, ovs, &(xo[2]));
TU = VFNMSI(Tw, Tv);
STM2(&(xo[10]), TU, ovs, &(xo[2]));
}
{
V TC, TE, Tz, TD, Tx, Ty;
TC = VMUL(LDK(KP866025403), VSUB(TA, TB));
TE = VMUL(LDK(KP866025403), VADD(TB, TA));
Tx = VADD(T5, Ta);
Ty = VADD(Tm, Tp);
Tz = VSUB(Tx, Ty);
TD = VADD(Tx, Ty);
{
V TV, TW, TX, TY;
TV = VFMAI(TC, Tz);
STM2(&(xo[4]), TV, ovs, &(xo[0]));
STN2(&(xo[4]), TV, TO, ovs);
TW = VFNMSI(TE, TD);
STM2(&(xo[16]), TW, ovs, &(xo[0]));
STN2(&(xo[16]), TW, TN, ovs);
TX = VFNMSI(TC, Tz);
STM2(&(xo[20]), TX, ovs, &(xo[0]));
STN2(&(xo[20]), TX, TT, ovs);
TY = VFMAI(TE, TD);
STM2(&(xo[8]), TY, ovs, &(xo[0]));
STN2(&(xo[8]), TY, TU, ovs);
}
}
}
}
}
VLEAVE();
}
static const kdft_desc desc = { 12, XSIMD_STRING("n2fv_12"), { 30, 2, 18, 0 }, &GENUS, 0, 2, 0, 0 };
void XSIMD(codelet_n2fv_12) (planner *p) { X(kdft_register) (p, n2fv_12, &desc);
}
#else
/* Generated by: ../../../genfft/gen_notw_c.native -simd -compact -variables 4 -pipeline-latency 8 -n 12 -name n2fv_12 -with-ostride 2 -include dft/simd/n2f.h -store-multiple 2 */
/*
* This function contains 48 FP additions, 8 FP multiplications,
* (or, 44 additions, 4 multiplications, 4 fused multiply/add),
* 33 stack variables, 2 constants, and 30 memory accesses
*/
#include "dft/simd/n2f.h"
static void n2fv_12(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
{
DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
{
INT i;
const R *xi;
R *xo;
xi = ri;
xo = ro;
for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs), MAKE_VOLATILE_STRIDE(24, is), MAKE_VOLATILE_STRIDE(24, os)) {
V T5, Ta, TJ, Ty, Tq, Tp, Tg, Tl, TI, TA, Tz, Tu;
{
V T1, T6, T4, Tw, T9, Tx;
T1 = LD(&(xi[0]), ivs, &(xi[0]));
T6 = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
{
V T2, T3, T7, T8;
T2 = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
T3 = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
T4 = VADD(T2, T3);
Tw = VSUB(T3, T2);
T7 = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
T8 = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
T9 = VADD(T7, T8);
Tx = VSUB(T8, T7);
}
T5 = VADD(T1, T4);
Ta = VADD(T6, T9);
TJ = VADD(Tw, Tx);
Ty = VMUL(LDK(KP866025403), VSUB(Tw, Tx));
Tq = VFNMS(LDK(KP500000000), T9, T6);
Tp = VFNMS(LDK(KP500000000), T4, T1);
}
{
V Tc, Th, Tf, Ts, Tk, Tt;
Tc = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
Th = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
{
V Td, Te, Ti, Tj;
Td = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
Te = LD(&(xi[WS(is, 11)]), ivs, &(xi[WS(is, 1)]));
Tf = VADD(Td, Te);
Ts = VSUB(Te, Td);
Ti = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
Tj = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
Tk = VADD(Ti, Tj);
Tt = VSUB(Tj, Ti);
}
Tg = VADD(Tc, Tf);
Tl = VADD(Th, Tk);
TI = VADD(Ts, Tt);
TA = VFNMS(LDK(KP500000000), Tk, Th);
Tz = VFNMS(LDK(KP500000000), Tf, Tc);
Tu = VMUL(LDK(KP866025403), VSUB(Ts, Tt));
}
{
V TN, TO, TP, TQ, TR, TS;
{
V Tb, Tm, Tn, To;
Tb = VSUB(T5, Ta);
Tm = VBYI(VSUB(Tg, Tl));
TN = VSUB(Tb, Tm);
STM2(&(xo[18]), TN, ovs, &(xo[2]));
TO = VADD(Tb, Tm);
STM2(&(xo[6]), TO, ovs, &(xo[2]));
Tn = VADD(T5, Ta);
To = VADD(Tg, Tl);
TP = VSUB(Tn, To);
STM2(&(xo[12]), TP, ovs, &(xo[0]));
TQ = VADD(Tn, To);
STM2(&(xo[0]), TQ, ovs, &(xo[0]));
}
{
V Tv, TE, TC, TD, Tr, TB, TT, TU;
Tr = VSUB(Tp, Tq);
Tv = VSUB(Tr, Tu);
TE = VADD(Tr, Tu);
TB = VSUB(Tz, TA);
TC = VBYI(VADD(Ty, TB));
TD = VBYI(VSUB(Ty, TB));
TR = VSUB(Tv, TC);
STM2(&(xo[10]), TR, ovs, &(xo[2]));
TS = VSUB(TE, TD);
STM2(&(xo[22]), TS, ovs, &(xo[2]));
TT = VADD(TC, Tv);
STM2(&(xo[14]), TT, ovs, &(xo[2]));
STN2(&(xo[12]), TP, TT, ovs);
TU = VADD(TD, TE);
STM2(&(xo[2]), TU, ovs, &(xo[2]));
STN2(&(xo[0]), TQ, TU, ovs);
}
{
V TK, TM, TH, TL, TF, TG;
TK = VBYI(VMUL(LDK(KP866025403), VSUB(TI, TJ)));
TM = VBYI(VMUL(LDK(KP866025403), VADD(TJ, TI)));
TF = VADD(Tp, Tq);
TG = VADD(Tz, TA);
TH = VSUB(TF, TG);
TL = VADD(TF, TG);
{
V TV, TW, TX, TY;
TV = VSUB(TH, TK);
STM2(&(xo[20]), TV, ovs, &(xo[0]));
STN2(&(xo[20]), TV, TS, ovs);
TW = VADD(TL, TM);
STM2(&(xo[8]), TW, ovs, &(xo[0]));
STN2(&(xo[8]), TW, TR, ovs);
TX = VADD(TH, TK);
STM2(&(xo[4]), TX, ovs, &(xo[0]));
STN2(&(xo[4]), TX, TO, ovs);
TY = VSUB(TL, TM);
STM2(&(xo[16]), TY, ovs, &(xo[0]));
STN2(&(xo[16]), TY, TN, ovs);
}
}
}
}
}
VLEAVE();
}
static const kdft_desc desc = { 12, XSIMD_STRING("n2fv_12"), { 44, 4, 4, 0 }, &GENUS, 0, 2, 0, 0 };
void XSIMD(codelet_n2fv_12) (planner *p) { X(kdft_register) (p, n2fv_12, &desc);
}
#endif
Reference in a new issue View git blame Copy permalink