Chris@42: /*
Chris@42:  * Copyright (c) 2003, 2007-14 Matteo Frigo
Chris@42:  * Copyright (c) 2003, 2007-14 Massachusetts Institute of Technology
Chris@42:  *
Chris@42:  * This program is free software; you can redistribute it and/or modify
Chris@42:  * it under the terms of the GNU General Public License as published by
Chris@42:  * the Free Software Foundation; either version 2 of the License, or
Chris@42:  * (at your option) any later version.
Chris@42:  *
Chris@42:  * This program is distributed in the hope that it will be useful,
Chris@42:  * but WITHOUT ANY WARRANTY; without even the implied warranty of
Chris@42:  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
Chris@42:  * GNU General Public License for more details.
Chris@42:  *
Chris@42:  * You should have received a copy of the GNU General Public License
Chris@42:  * along with this program; if not, write to the Free Software
Chris@42:  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA
Chris@42:  *
Chris@42:  */
Chris@42: 
Chris@42: /* This file was automatically generated --- DO NOT EDIT */
Chris@42: /* Generated on Sat Jul 30 16:52:42 EDT 2016 */
Chris@42: 
Chris@42: #include "codelet-rdft.h"
Chris@42: 
Chris@42: #ifdef HAVE_FMA
Chris@42: 
Chris@42: /* Generated by: ../../../genfft/gen_hc2cdft_c.native -fma -reorder-insns -schedule-for-pipeline -simd -compact -variables 4 -pipeline-latency 8 -trivial-stores -variables 32 -no-generate-bytw -n 2 -dif -sign 1 -name hc2cbdftv_2 -include hc2cbv.h */
Chris@42: 
Chris@42: /*
Chris@42:  * This function contains 5 FP additions, 4 FP multiplications,
Chris@42:  * (or, 3 additions, 2 multiplications, 2 fused multiply/add),
Chris@42:  * 8 stack variables, 0 constants, and 4 memory accesses
Chris@42:  */
Chris@42: #include "hc2cbv.h"
Chris@42: 
Chris@42: static void hc2cbdftv_2(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms)
Chris@42: {
Chris@42:      {
Chris@42: 	  INT m;
Chris@42: 	  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)) {
Chris@42: 	       V T2, T3, T1, T5, T4, T7, T6;
Chris@42: 	       T2 = LD(&(Rp[0]), ms, &(Rp[0]));
Chris@42: 	       T3 = LD(&(Rm[0]), -ms, &(Rm[0]));
Chris@42: 	       T1 = LDW(&(W[0]));
Chris@42: 	       T5 = VFMACONJ(T3, T2);
Chris@42: 	       T4 = VZMULI(T1, VFNMSCONJ(T3, T2));
Chris@42: 	       T7 = VCONJ(VSUB(T5, T4));
Chris@42: 	       T6 = VADD(T4, T5);
Chris@42: 	       ST(&(Rm[0]), T7, -ms, &(Rm[0]));
Chris@42: 	       ST(&(Rp[0]), T6, ms, &(Rp[0]));
Chris@42: 	  }
Chris@42:      }
Chris@42:      VLEAVE();
Chris@42: }
Chris@42: 
Chris@42: static const tw_instr twinstr[] = {
Chris@42:      VTW(1, 1),
Chris@42:      {TW_NEXT, VL, 0}
Chris@42: };
Chris@42: 
Chris@42: static const hc2c_desc desc = { 2, XSIMD_STRING("hc2cbdftv_2"), twinstr, &GENUS, {3, 2, 2, 0} };
Chris@42: 
Chris@42: void XSIMD(codelet_hc2cbdftv_2) (planner *p) {
Chris@42:      X(khc2c_register) (p, hc2cbdftv_2, &desc, HC2C_VIA_DFT);
Chris@42: }
Chris@42: #else				/* HAVE_FMA */
Chris@42: 
Chris@42: /* Generated by: ../../../genfft/gen_hc2cdft_c.native -simd -compact -variables 4 -pipeline-latency 8 -trivial-stores -variables 32 -no-generate-bytw -n 2 -dif -sign 1 -name hc2cbdftv_2 -include hc2cbv.h */
Chris@42: 
Chris@42: /*
Chris@42:  * This function contains 5 FP additions, 2 FP multiplications,
Chris@42:  * (or, 5 additions, 2 multiplications, 0 fused multiply/add),
Chris@42:  * 9 stack variables, 0 constants, and 4 memory accesses
Chris@42:  */
Chris@42: #include "hc2cbv.h"
Chris@42: 
Chris@42: static void hc2cbdftv_2(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms)
Chris@42: {
Chris@42:      {
Chris@42: 	  INT m;
Chris@42: 	  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)) {
Chris@42: 	       V T6, T5, T2, T4, T3, T1, T7, T8;
Chris@42: 	       T2 = LD(&(Rp[0]), ms, &(Rp[0]));
Chris@42: 	       T3 = LD(&(Rm[0]), -ms, &(Rm[0]));
Chris@42: 	       T4 = VCONJ(T3);
Chris@42: 	       T6 = VADD(T2, T4);
Chris@42: 	       T1 = LDW(&(W[0]));
Chris@42: 	       T5 = VZMULI(T1, VSUB(T2, T4));
Chris@42: 	       T7 = VADD(T5, T6);
Chris@42: 	       ST(&(Rp[0]), T7, ms, &(Rp[0]));
Chris@42: 	       T8 = VCONJ(VSUB(T6, T5));
Chris@42: 	       ST(&(Rm[0]), T8, -ms, &(Rm[0]));
Chris@42: 	  }
Chris@42:      }
Chris@42:      VLEAVE();
Chris@42: }
Chris@42: 
Chris@42: static const tw_instr twinstr[] = {
Chris@42:      VTW(1, 1),
Chris@42:      {TW_NEXT, VL, 0}
Chris@42: };
Chris@42: 
Chris@42: static const hc2c_desc desc = { 2, XSIMD_STRING("hc2cbdftv_2"), twinstr, &GENUS, {5, 2, 0, 0} };
Chris@42: 
Chris@42: void XSIMD(codelet_hc2cbdftv_2) (planner *p) {
Chris@42:      X(khc2c_register) (p, hc2cbdftv_2, &desc, HC2C_VIA_DFT);
Chris@42: }
Chris@42: #endif				/* HAVE_FMA */