Chris@19: /* Chris@19: * Copyright (c) 2003, 2007-14 Matteo Frigo Chris@19: * Copyright (c) 2003, 2007-14 Massachusetts Institute of Technology Chris@19: * Chris@19: * This program is free software; you can redistribute it and/or modify Chris@19: * it under the terms of the GNU General Public License as published by Chris@19: * the Free Software Foundation; either version 2 of the License, or Chris@19: * (at your option) any later version. Chris@19: * Chris@19: * This program is distributed in the hope that it will be useful, Chris@19: * but WITHOUT ANY WARRANTY; without even the implied warranty of Chris@19: * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the Chris@19: * GNU General Public License for more details. Chris@19: * Chris@19: * You should have received a copy of the GNU General Public License Chris@19: * along with this program; if not, write to the Free Software Chris@19: * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA Chris@19: * Chris@19: */ Chris@19: Chris@19: /* This file was automatically generated --- DO NOT EDIT */ Chris@19: /* Generated on Tue Mar 4 13:46:50 EST 2014 */ Chris@19: Chris@19: #include "codelet-dft.h" Chris@19: Chris@19: #ifdef HAVE_FMA Chris@19: Chris@19: /* Generated by: ../../../genfft/gen_notw_c.native -fma -reorder-insns -schedule-for-pipeline -simd -compact -variables 4 -pipeline-latency 8 -sign 1 -n 6 -name n1bv_6 -include n1b.h */ Chris@19: Chris@19: /* Chris@19: * This function contains 18 FP additions, 8 FP multiplications, Chris@19: * (or, 12 additions, 2 multiplications, 6 fused multiply/add), Chris@19: * 23 stack variables, 2 constants, and 12 memory accesses Chris@19: */ Chris@19: #include "n1b.h" Chris@19: Chris@19: static void n1bv_6(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs) Chris@19: { Chris@19: DVK(KP500000000, +0.500000000000000000000000000000000000000000000); Chris@19: DVK(KP866025403, +0.866025403784438646763723170752936183471402627); Chris@19: { Chris@19: INT i; Chris@19: const R *xi; Chris@19: R *xo; Chris@19: xi = ii; Chris@19: xo = io; Chris@19: for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs), MAKE_VOLATILE_STRIDE(12, is), MAKE_VOLATILE_STRIDE(12, os)) { Chris@19: V T1, T2, T4, T5, T7, T8; Chris@19: T1 = LD(&(xi[0]), ivs, &(xi[0])); Chris@19: T2 = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)])); Chris@19: T4 = LD(&(xi[WS(is, 2)]), ivs, &(xi[0])); Chris@19: T5 = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)])); Chris@19: T7 = LD(&(xi[WS(is, 4)]), ivs, &(xi[0])); Chris@19: T8 = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)])); Chris@19: { Chris@19: V T3, Td, T6, Te, T9, Tf; Chris@19: T3 = VSUB(T1, T2); Chris@19: Td = VADD(T1, T2); Chris@19: T6 = VSUB(T4, T5); Chris@19: Te = VADD(T4, T5); Chris@19: T9 = VSUB(T7, T8); Chris@19: Tf = VADD(T7, T8); Chris@19: { Chris@19: V Tg, Ti, Ta, Tc, Th, Tb; Chris@19: Tg = VADD(Te, Tf); Chris@19: Ti = VMUL(LDK(KP866025403), VSUB(Te, Tf)); Chris@19: Ta = VADD(T6, T9); Chris@19: Tc = VMUL(LDK(KP866025403), VSUB(T6, T9)); Chris@19: Th = VFNMS(LDK(KP500000000), Tg, Td); Chris@19: ST(&(xo[0]), VADD(Td, Tg), ovs, &(xo[0])); Chris@19: Tb = VFNMS(LDK(KP500000000), Ta, T3); Chris@19: ST(&(xo[WS(os, 3)]), VADD(T3, Ta), ovs, &(xo[WS(os, 1)])); Chris@19: ST(&(xo[WS(os, 4)]), VFMAI(Ti, Th), ovs, &(xo[0])); Chris@19: ST(&(xo[WS(os, 2)]), VFNMSI(Ti, Th), ovs, &(xo[0])); Chris@19: ST(&(xo[WS(os, 5)]), VFNMSI(Tc, Tb), ovs, &(xo[WS(os, 1)])); Chris@19: ST(&(xo[WS(os, 1)]), VFMAI(Tc, Tb), ovs, &(xo[WS(os, 1)])); Chris@19: } Chris@19: } Chris@19: } Chris@19: } Chris@19: VLEAVE(); Chris@19: } Chris@19: Chris@19: static const kdft_desc desc = { 6, XSIMD_STRING("n1bv_6"), {12, 2, 6, 0}, &GENUS, 0, 0, 0, 0 }; Chris@19: Chris@19: void XSIMD(codelet_n1bv_6) (planner *p) { Chris@19: X(kdft_register) (p, n1bv_6, &desc); Chris@19: } Chris@19: Chris@19: #else /* HAVE_FMA */ Chris@19: Chris@19: /* Generated by: ../../../genfft/gen_notw_c.native -simd -compact -variables 4 -pipeline-latency 8 -sign 1 -n 6 -name n1bv_6 -include n1b.h */ Chris@19: Chris@19: /* Chris@19: * This function contains 18 FP additions, 4 FP multiplications, Chris@19: * (or, 16 additions, 2 multiplications, 2 fused multiply/add), Chris@19: * 19 stack variables, 2 constants, and 12 memory accesses Chris@19: */ Chris@19: #include "n1b.h" Chris@19: Chris@19: static void n1bv_6(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs) Chris@19: { Chris@19: DVK(KP500000000, +0.500000000000000000000000000000000000000000000); Chris@19: DVK(KP866025403, +0.866025403784438646763723170752936183471402627); Chris@19: { Chris@19: INT i; Chris@19: const R *xi; Chris@19: R *xo; Chris@19: xi = ii; Chris@19: xo = io; Chris@19: for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs), MAKE_VOLATILE_STRIDE(12, is), MAKE_VOLATILE_STRIDE(12, os)) { Chris@19: V Ta, Td, T3, Te, T6, Tf, Tb, Tg, T8, T9; Chris@19: T8 = LD(&(xi[0]), ivs, &(xi[0])); Chris@19: T9 = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)])); Chris@19: Ta = VSUB(T8, T9); Chris@19: Td = VADD(T8, T9); Chris@19: { Chris@19: V T1, T2, T4, T5; Chris@19: T1 = LD(&(xi[WS(is, 2)]), ivs, &(xi[0])); Chris@19: T2 = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)])); Chris@19: T3 = VSUB(T1, T2); Chris@19: Te = VADD(T1, T2); Chris@19: T4 = LD(&(xi[WS(is, 4)]), ivs, &(xi[0])); Chris@19: T5 = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)])); Chris@19: T6 = VSUB(T4, T5); Chris@19: Tf = VADD(T4, T5); Chris@19: } Chris@19: Tb = VADD(T3, T6); Chris@19: Tg = VADD(Te, Tf); Chris@19: ST(&(xo[WS(os, 3)]), VADD(Ta, Tb), ovs, &(xo[WS(os, 1)])); Chris@19: ST(&(xo[0]), VADD(Td, Tg), ovs, &(xo[0])); Chris@19: { Chris@19: V T7, Tc, Th, Ti; Chris@19: T7 = VBYI(VMUL(LDK(KP866025403), VSUB(T3, T6))); Chris@19: Tc = VFNMS(LDK(KP500000000), Tb, Ta); Chris@19: ST(&(xo[WS(os, 1)]), VADD(T7, Tc), ovs, &(xo[WS(os, 1)])); Chris@19: ST(&(xo[WS(os, 5)]), VSUB(Tc, T7), ovs, &(xo[WS(os, 1)])); Chris@19: Th = VFNMS(LDK(KP500000000), Tg, Td); Chris@19: Ti = VBYI(VMUL(LDK(KP866025403), VSUB(Te, Tf))); Chris@19: ST(&(xo[WS(os, 2)]), VSUB(Th, Ti), ovs, &(xo[0])); Chris@19: ST(&(xo[WS(os, 4)]), VADD(Ti, Th), ovs, &(xo[0])); Chris@19: } Chris@19: } Chris@19: } Chris@19: VLEAVE(); Chris@19: } Chris@19: Chris@19: static const kdft_desc desc = { 6, XSIMD_STRING("n1bv_6"), {16, 2, 2, 0}, &GENUS, 0, 0, 0, 0 }; Chris@19: Chris@19: void XSIMD(codelet_n1bv_6) (planner *p) { Chris@19: X(kdft_register) (p, n1bv_6, &desc); Chris@19: } Chris@19: Chris@19: #endif /* HAVE_FMA */