cannam@167: /* cannam@167: * Copyright (c) 2003, 2007-14 Matteo Frigo cannam@167: * Copyright (c) 2003, 2007-14 Massachusetts Institute of Technology cannam@167: * cannam@167: * This program is free software; you can redistribute it and/or modify cannam@167: * it under the terms of the GNU General Public License as published by cannam@167: * the Free Software Foundation; either version 2 of the License, or cannam@167: * (at your option) any later version. cannam@167: * cannam@167: * This program is distributed in the hope that it will be useful, cannam@167: * but WITHOUT ANY WARRANTY; without even the implied warranty of cannam@167: * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the cannam@167: * GNU General Public License for more details. cannam@167: * cannam@167: * You should have received a copy of the GNU General Public License cannam@167: * along with this program; if not, write to the Free Software cannam@167: * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA cannam@167: * cannam@167: */ cannam@167: cannam@167: /* This file was automatically generated --- DO NOT EDIT */ cannam@167: /* Generated on Thu May 24 08:06:42 EDT 2018 */ cannam@167: cannam@167: #include "rdft/codelet-rdft.h" cannam@167: cannam@167: #if defined(ARCH_PREFERS_FMA) || defined(ISA_EXTENSION_PREFERS_FMA) cannam@167: cannam@167: /* Generated by: ../../../genfft/gen_r2cf.native -fma -compact -variables 4 -pipeline-latency 4 -n 8 -name r2cfII_8 -dft-II -include rdft/scalar/r2cfII.h */ cannam@167: cannam@167: /* cannam@167: * This function contains 22 FP additions, 16 FP multiplications, cannam@167: * (or, 6 additions, 0 multiplications, 16 fused multiply/add), cannam@167: * 18 stack variables, 3 constants, and 16 memory accesses cannam@167: */ cannam@167: #include "rdft/scalar/r2cfII.h" cannam@167: cannam@167: static void r2cfII_8(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs) cannam@167: { cannam@167: DK(KP923879532, +0.923879532511286756128183189396788286822416626); cannam@167: DK(KP707106781, +0.707106781186547524400844362104849039284835938); cannam@167: DK(KP414213562, +0.414213562373095048801688724209698078569671875); cannam@167: { cannam@167: INT i; cannam@167: for (i = v; i > 0; i = i - 1, R0 = R0 + ivs, R1 = R1 + ivs, Cr = Cr + ovs, Ci = Ci + ovs, MAKE_VOLATILE_STRIDE(32, rs), MAKE_VOLATILE_STRIDE(32, csr), MAKE_VOLATILE_STRIDE(32, csi)) { cannam@167: E T1, Th, T4, Ti, T8, Te, Tb, Tf, T2, T3; cannam@167: T1 = R0[0]; cannam@167: Th = R0[WS(rs, 2)]; cannam@167: T2 = R0[WS(rs, 1)]; cannam@167: T3 = R0[WS(rs, 3)]; cannam@167: T4 = T2 - T3; cannam@167: Ti = T2 + T3; cannam@167: { cannam@167: E T6, T7, T9, Ta; cannam@167: T6 = R1[0]; cannam@167: T7 = R1[WS(rs, 2)]; cannam@167: T8 = FNMS(KP414213562, T7, T6); cannam@167: Te = FMA(KP414213562, T6, T7); cannam@167: T9 = R1[WS(rs, 3)]; cannam@167: Ta = R1[WS(rs, 1)]; cannam@167: Tb = FMS(KP414213562, Ta, T9); cannam@167: Tf = FMA(KP414213562, T9, Ta); cannam@167: } cannam@167: { cannam@167: E T5, Tc, Tj, Tk; cannam@167: T5 = FMA(KP707106781, T4, T1); cannam@167: Tc = T8 + Tb; cannam@167: Cr[WS(csr, 3)] = FNMS(KP923879532, Tc, T5); cannam@167: Cr[0] = FMA(KP923879532, Tc, T5); cannam@167: Tj = FMA(KP707106781, Ti, Th); cannam@167: Tk = Te + Tf; cannam@167: Ci[0] = -(FMA(KP923879532, Tk, Tj)); cannam@167: Ci[WS(csi, 3)] = FNMS(KP923879532, Tk, Tj); cannam@167: } cannam@167: { cannam@167: E Td, Tg, Tl, Tm; cannam@167: Td = FNMS(KP707106781, T4, T1); cannam@167: Tg = Te - Tf; cannam@167: Cr[WS(csr, 2)] = FNMS(KP923879532, Tg, Td); cannam@167: Cr[WS(csr, 1)] = FMA(KP923879532, Tg, Td); cannam@167: Tl = FNMS(KP707106781, Ti, Th); cannam@167: Tm = Tb - T8; cannam@167: Ci[WS(csi, 2)] = FMS(KP923879532, Tm, Tl); cannam@167: Ci[WS(csi, 1)] = FMA(KP923879532, Tm, Tl); cannam@167: } cannam@167: } cannam@167: } cannam@167: } cannam@167: cannam@167: static const kr2c_desc desc = { 8, "r2cfII_8", {6, 0, 16, 0}, &GENUS }; cannam@167: cannam@167: void X(codelet_r2cfII_8) (planner *p) { cannam@167: X(kr2c_register) (p, r2cfII_8, &desc); cannam@167: } cannam@167: cannam@167: #else cannam@167: cannam@167: /* Generated by: ../../../genfft/gen_r2cf.native -compact -variables 4 -pipeline-latency 4 -n 8 -name r2cfII_8 -dft-II -include rdft/scalar/r2cfII.h */ cannam@167: cannam@167: /* cannam@167: * This function contains 22 FP additions, 10 FP multiplications, cannam@167: * (or, 18 additions, 6 multiplications, 4 fused multiply/add), cannam@167: * 18 stack variables, 3 constants, and 16 memory accesses cannam@167: */ cannam@167: #include "rdft/scalar/r2cfII.h" cannam@167: cannam@167: static void r2cfII_8(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs) cannam@167: { cannam@167: DK(KP382683432, +0.382683432365089771728459984030398866761344562); cannam@167: DK(KP923879532, +0.923879532511286756128183189396788286822416626); cannam@167: DK(KP707106781, +0.707106781186547524400844362104849039284835938); cannam@167: { cannam@167: INT i; cannam@167: for (i = v; i > 0; i = i - 1, R0 = R0 + ivs, R1 = R1 + ivs, Cr = Cr + ovs, Ci = Ci + ovs, MAKE_VOLATILE_STRIDE(32, rs), MAKE_VOLATILE_STRIDE(32, csr), MAKE_VOLATILE_STRIDE(32, csi)) { cannam@167: E T1, Tj, T4, Ti, T8, Te, Tb, Tf, T2, T3; cannam@167: T1 = R0[0]; cannam@167: Tj = R0[WS(rs, 2)]; cannam@167: T2 = R0[WS(rs, 1)]; cannam@167: T3 = R0[WS(rs, 3)]; cannam@167: T4 = KP707106781 * (T2 - T3); cannam@167: Ti = KP707106781 * (T2 + T3); cannam@167: { cannam@167: E T6, T7, T9, Ta; cannam@167: T6 = R1[0]; cannam@167: T7 = R1[WS(rs, 2)]; cannam@167: T8 = FNMS(KP382683432, T7, KP923879532 * T6); cannam@167: Te = FMA(KP382683432, T6, KP923879532 * T7); cannam@167: T9 = R1[WS(rs, 1)]; cannam@167: Ta = R1[WS(rs, 3)]; cannam@167: Tb = FNMS(KP923879532, Ta, KP382683432 * T9); cannam@167: Tf = FMA(KP923879532, T9, KP382683432 * Ta); cannam@167: } cannam@167: { cannam@167: E T5, Tc, Th, Tk; cannam@167: T5 = T1 + T4; cannam@167: Tc = T8 + Tb; cannam@167: Cr[WS(csr, 3)] = T5 - Tc; cannam@167: Cr[0] = T5 + Tc; cannam@167: Th = Te + Tf; cannam@167: Tk = Ti + Tj; cannam@167: Ci[0] = -(Th + Tk); cannam@167: Ci[WS(csi, 3)] = Tk - Th; cannam@167: } cannam@167: { cannam@167: E Td, Tg, Tl, Tm; cannam@167: Td = T1 - T4; cannam@167: Tg = Te - Tf; cannam@167: Cr[WS(csr, 2)] = Td - Tg; cannam@167: Cr[WS(csr, 1)] = Td + Tg; cannam@167: Tl = Tb - T8; cannam@167: Tm = Tj - Ti; cannam@167: Ci[WS(csi, 2)] = Tl - Tm; cannam@167: Ci[WS(csi, 1)] = Tl + Tm; cannam@167: } cannam@167: } cannam@167: } cannam@167: } cannam@167: cannam@167: static const kr2c_desc desc = { 8, "r2cfII_8", {18, 6, 4, 0}, &GENUS }; cannam@167: cannam@167: void X(codelet_r2cfII_8) (planner *p) { cannam@167: X(kr2c_register) (p, r2cfII_8, &desc); cannam@167: } cannam@167: cannam@167: #endif