src/gromacs/gmxlib/nonbonded/nb_kernel_avx_128_fma_double/nb_kernel_ElecEw_VdwCSTab_GeomW4P1_avx_128_fma_double.cpp

   1 /*
   2  * This file is part of the GROMACS molecular simulation package.
   3  *
   4  * Copyright (c) 2012,2013,2014,2015,2017,2018, by the GROMACS development team, led by
   5  * Mark Abraham, David van der Spoel, Berk Hess, and Erik Lindahl,
   6  * and including many others, as listed in the AUTHORS file in the
   7  * top-level source directory and at http://www.gromacs.org.
   8  *
   9  * GROMACS is free software; you can redistribute it and/or
  10  * modify it under the terms of the GNU Lesser General Public License
  11  * as published by the Free Software Foundation; either version 2.1
  12  * of the License, or (at your option) any later version.
  13  *
  14  * GROMACS is distributed in the hope that it will be useful,
  15  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  16  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  17  * Lesser General Public License for more details.
  18  *
  19  * You should have received a copy of the GNU Lesser General Public
  20  * License along with GROMACS; if not, see
  21  * http://www.gnu.org/licenses, or write to the Free Software Foundation,
  22  * Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA.
  23  *
  24  * If you want to redistribute modifications to GROMACS, please
  25  * consider that scientific software is very special. Version
  26  * control is crucial - bugs must be traceable. We will be happy to
  27  * consider code for inclusion in the official distribution, but
  28  * derived work must not be called official GROMACS. Details are found
  29  * in the README & COPYING files - if they are missing, get the
  30  * official version at http://www.gromacs.org.
  31  *
  32  * To help us fund GROMACS development, we humbly ask that you cite
  33  * the research papers on the package. Check out http://www.gromacs.org.
  34  */
  35 /*
  36  * Note: this file was generated by the GROMACS avx_128_fma_double kernel generator.
  37  */
  38 #include "gmxpre.h"
  39
  40 #include "config.h"
  41
  42 #include <math.h>
  43
  44 #include "../nb_kernel.h"
  45 #include "gromacs/gmxlib/nrnb.h"
  46
  47 #include "kernelutil_x86_avx_128_fma_double.h"
  48
  49 /*
  50  * Gromacs nonbonded kernel:   nb_kernel_ElecEw_VdwCSTab_GeomW4P1_VF_avx_128_fma_double
  51  * Electrostatics interaction: Ewald
  52  * VdW interaction:            CubicSplineTable
  53  * Geometry:                   Water4-Particle
  54  * Calculate force/pot:        PotentialAndForce
  55  */
  56 void
  57 nb_kernel_ElecEw_VdwCSTab_GeomW4P1_VF_avx_128_fma_double
  58                     (t_nblist                    * gmx_restrict       nlist,
  59                      rvec                        * gmx_restrict          xx,
  60                      rvec                        * gmx_restrict          ff,
  61                      struct t_forcerec           * gmx_restrict          fr,
  62                      t_mdatoms                   * gmx_restrict     mdatoms,
  63                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
  64                      t_nrnb                      * gmx_restrict        nrnb)
  65 {
  66     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
  67      * just 0 for non-waters.
  68      * Suffixes A,B refer to j loop unrolling done with SSE double precision, e.g. for the two different
  69      * jnr indices corresponding to data put in the four positions in the SIMD register.
  70      */
  71     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
  72     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
  73     int              jnrA,jnrB;
  74     int              j_coord_offsetA,j_coord_offsetB;
  75     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  76     real             rcutoff_scalar;
  77     real             *shiftvec,*fshift,*x,*f;
  78     __m128d          tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
  79     int              vdwioffset0;
  80     __m128d          ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  81     int              vdwioffset1;
  82     __m128d          ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
  83     int              vdwioffset2;
  84     __m128d          ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
  85     int              vdwioffset3;
  86     __m128d          ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
  87     int              vdwjidx0A,vdwjidx0B;
  88     __m128d          jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  89     __m128d          dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
  90     __m128d          dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
  91     __m128d          dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
  92     __m128d          dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30;
  93     __m128d          velec,felec,velecsum,facel,crf,krf,krf2;
  94     real             *charge;
  95     int              nvdwtype;
  96     __m128d          rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
  97     int              *vdwtype;
  98     real             *vdwparam;
  99     __m128d          one_sixth   = _mm_set1_pd(1.0/6.0);
 100     __m128d          one_twelfth = _mm_set1_pd(1.0/12.0);
 101     __m128i          vfitab;
 102     __m128i          ifour       = _mm_set1_epi32(4);
 103     __m128d          rt,vfeps,vftabscale,Y,F,G,H,Heps,Fp,VV,FF,twovfeps;
 104     real             *vftab;
 105     __m128i          ewitab;
 106     __m128d          ewtabscale,eweps,twoeweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
 107     real             *ewtab;
 108     __m128d          dummy_mask,cutoff_mask;
 109     __m128d          signbit   = gmx_mm_castsi128_pd( _mm_set_epi32(0x80000000,0x00000000,0x80000000,0x00000000) );
 110     __m128d          one     = _mm_set1_pd(1.0);
 111     __m128d          two     = _mm_set1_pd(2.0);
 112     x                = xx[0];
 113     f                = ff[0];
 114
 115     nri              = nlist->nri;
 116     iinr             = nlist->iinr;
 117     jindex           = nlist->jindex;
 118     jjnr             = nlist->jjnr;
 119     shiftidx         = nlist->shift;
 120     gid              = nlist->gid;
 121     shiftvec         = fr->shift_vec[0];
 122     fshift           = fr->fshift[0];
 123     facel            = _mm_set1_pd(fr->ic->epsfac);
 124     charge           = mdatoms->chargeA;
 125     nvdwtype         = fr->ntype;
 126     vdwparam         = fr->nbfp;
 127     vdwtype          = mdatoms->typeA;
 128
 129     vftab            = kernel_data->table_vdw->data;
 130     vftabscale       = _mm_set1_pd(kernel_data->table_vdw->scale);
 131
 132     sh_ewald         = _mm_set1_pd(fr->ic->sh_ewald);
 133     ewtab            = fr->ic->tabq_coul_FDV0;
 134     ewtabscale       = _mm_set1_pd(fr->ic->tabq_scale);
 135     ewtabhalfspace   = _mm_set1_pd(0.5/fr->ic->tabq_scale);
 136
 137     /* Setup water-specific parameters */
 138     inr              = nlist->iinr[0];
 139     iq1              = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+1]));
 140     iq2              = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+2]));
 141     iq3              = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+3]));
 142     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 143
 144     /* Avoid stupid compiler warnings */
 145     jnrA = jnrB = 0;
 146     j_coord_offsetA = 0;
 147     j_coord_offsetB = 0;
 148
 149     outeriter        = 0;
 150     inneriter        = 0;
 151
 152     /* Start outer loop over neighborlists */
 153     for(iidx=0; iidx<nri; iidx++)
 154     {
 155         /* Load shift vector for this list */
 156         i_shift_offset   = DIM*shiftidx[iidx];
 157
 158         /* Load limits for loop over neighbors */
 159         j_index_start    = jindex[iidx];
 160         j_index_end      = jindex[iidx+1];
 161
 162         /* Get outer coordinate index */
 163         inr              = iinr[iidx];
 164         i_coord_offset   = DIM*inr;
 165
 166         /* Load i particle coords and add shift vector */
 167         gmx_mm_load_shift_and_4rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset,
 168                                                  &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
 169
 170         fix0             = _mm_setzero_pd();
 171         fiy0             = _mm_setzero_pd();
 172         fiz0             = _mm_setzero_pd();
 173         fix1             = _mm_setzero_pd();
 174         fiy1             = _mm_setzero_pd();
 175         fiz1             = _mm_setzero_pd();
 176         fix2             = _mm_setzero_pd();
 177         fiy2             = _mm_setzero_pd();
 178         fiz2             = _mm_setzero_pd();
 179         fix3             = _mm_setzero_pd();
 180         fiy3             = _mm_setzero_pd();
 181         fiz3             = _mm_setzero_pd();
 182
 183         /* Reset potential sums */
 184         velecsum         = _mm_setzero_pd();
 185         vvdwsum          = _mm_setzero_pd();
 186
 187         /* Start inner kernel loop */
 188         for(jidx=j_index_start; jidx<j_index_end-1; jidx+=2)
 189         {
 190
 191             /* Get j neighbor index, and coordinate index */
 192             jnrA             = jjnr[jidx];
 193             jnrB             = jjnr[jidx+1];
 194             j_coord_offsetA  = DIM*jnrA;
 195             j_coord_offsetB  = DIM*jnrB;
 196
 197             /* load j atom coordinates */
 198             gmx_mm_load_1rvec_2ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
 199                                               &jx0,&jy0,&jz0);
 200
 201             /* Calculate displacement vector */
 202             dx00             = _mm_sub_pd(ix0,jx0);
 203             dy00             = _mm_sub_pd(iy0,jy0);
 204             dz00             = _mm_sub_pd(iz0,jz0);
 205             dx10             = _mm_sub_pd(ix1,jx0);
 206             dy10             = _mm_sub_pd(iy1,jy0);
 207             dz10             = _mm_sub_pd(iz1,jz0);
 208             dx20             = _mm_sub_pd(ix2,jx0);
 209             dy20             = _mm_sub_pd(iy2,jy0);
 210             dz20             = _mm_sub_pd(iz2,jz0);
 211             dx30             = _mm_sub_pd(ix3,jx0);
 212             dy30             = _mm_sub_pd(iy3,jy0);
 213             dz30             = _mm_sub_pd(iz3,jz0);
 214
 215             /* Calculate squared distance and things based on it */
 216             rsq00            = gmx_mm_calc_rsq_pd(dx00,dy00,dz00);
 217             rsq10            = gmx_mm_calc_rsq_pd(dx10,dy10,dz10);
 218             rsq20            = gmx_mm_calc_rsq_pd(dx20,dy20,dz20);
 219             rsq30            = gmx_mm_calc_rsq_pd(dx30,dy30,dz30);
 220
 221             rinv00           = avx128fma_invsqrt_d(rsq00);
 222             rinv10           = avx128fma_invsqrt_d(rsq10);
 223             rinv20           = avx128fma_invsqrt_d(rsq20);
 224             rinv30           = avx128fma_invsqrt_d(rsq30);
 225
 226             rinvsq10         = _mm_mul_pd(rinv10,rinv10);
 227             rinvsq20         = _mm_mul_pd(rinv20,rinv20);
 228             rinvsq30         = _mm_mul_pd(rinv30,rinv30);
 229
 230             /* Load parameters for j particles */
 231             jq0              = gmx_mm_load_2real_swizzle_pd(charge+jnrA+0,charge+jnrB+0);
 232             vdwjidx0A        = 2*vdwtype[jnrA+0];
 233             vdwjidx0B        = 2*vdwtype[jnrB+0];
 234
 235             fjx0             = _mm_setzero_pd();
 236             fjy0             = _mm_setzero_pd();
 237             fjz0             = _mm_setzero_pd();
 238
 239             /**************************
 240              * CALCULATE INTERACTIONS *
 241              **************************/
 242
 243             r00              = _mm_mul_pd(rsq00,rinv00);
 244
 245             /* Compute parameters for interactions between i and j atoms */
 246             gmx_mm_load_2pair_swizzle_pd(vdwparam+vdwioffset0+vdwjidx0A,
 247                                          vdwparam+vdwioffset0+vdwjidx0B,&c6_00,&c12_00);
 248
 249             /* Calculate table index by multiplying r with table scale and truncate to integer */
 250             rt               = _mm_mul_pd(r00,vftabscale);
 251             vfitab           = _mm_cvttpd_epi32(rt);
 252 #ifdef __XOP__
 253             vfeps            = _mm_frcz_pd(rt);
 254 #else
 255             vfeps            = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
 256 #endif
 257             twovfeps         = _mm_add_pd(vfeps,vfeps);
 258             vfitab           = _mm_slli_epi32(vfitab,3);
 259
 260             /* CUBIC SPLINE TABLE DISPERSION */
 261             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 262             F                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) );
 263             GMX_MM_TRANSPOSE2_PD(Y,F);
 264             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
 265             H                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) +2);
 266             GMX_MM_TRANSPOSE2_PD(G,H);
 267             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(H,vfeps,G),F);
 268             VV               = _mm_macc_pd(vfeps,Fp,Y);
 269             vvdw6            = _mm_mul_pd(c6_00,VV);
 270             FF               = _mm_macc_pd(vfeps,_mm_macc_pd(twovfeps,H,G),Fp);
 271             fvdw6            = _mm_mul_pd(c6_00,FF);
 272
 273             /* CUBIC SPLINE TABLE REPULSION */
 274             vfitab           = _mm_add_epi32(vfitab,ifour);
 275             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 276             F                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) );
 277             GMX_MM_TRANSPOSE2_PD(Y,F);
 278             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
 279             H                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) +2);
 280             GMX_MM_TRANSPOSE2_PD(G,H);
 281             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(H,vfeps,G),F);
 282             VV               = _mm_macc_pd(vfeps,Fp,Y);
 283             vvdw12           = _mm_mul_pd(c12_00,VV);
 284             FF               = _mm_macc_pd(vfeps,_mm_macc_pd(twovfeps,H,G),Fp);
 285             fvdw12           = _mm_mul_pd(c12_00,FF);
 286             vvdw             = _mm_add_pd(vvdw12,vvdw6);
 287             fvdw             = _mm_xor_pd(signbit,_mm_mul_pd(_mm_add_pd(fvdw6,fvdw12),_mm_mul_pd(vftabscale,rinv00)));
 288
 289             /* Update potential sum for this i atom from the interaction with this j atom. */
 290             vvdwsum          = _mm_add_pd(vvdwsum,vvdw);
 291
 292             fscal            = fvdw;
 293
 294             /* Update vectorial force */
 295             fix0             = _mm_macc_pd(dx00,fscal,fix0);
 296             fiy0             = _mm_macc_pd(dy00,fscal,fiy0);
 297             fiz0             = _mm_macc_pd(dz00,fscal,fiz0);
 298
 299             fjx0             = _mm_macc_pd(dx00,fscal,fjx0);
 300             fjy0             = _mm_macc_pd(dy00,fscal,fjy0);
 301             fjz0             = _mm_macc_pd(dz00,fscal,fjz0);
 302
 303             /**************************
 304              * CALCULATE INTERACTIONS *
 305              **************************/
 306
 307             r10              = _mm_mul_pd(rsq10,rinv10);
 308
 309             /* Compute parameters for interactions between i and j atoms */
 310             qq10             = _mm_mul_pd(iq1,jq0);
 311
 312             /* EWALD ELECTROSTATICS */
 313
 314             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 315             ewrt             = _mm_mul_pd(r10,ewtabscale);
 316             ewitab           = _mm_cvttpd_epi32(ewrt);
 317 #ifdef __XOP__
 318             eweps            = _mm_frcz_pd(ewrt);
 319 #else
 320             eweps            = _mm_sub_pd(ewrt,_mm_round_pd(ewrt, _MM_FROUND_FLOOR));
 321 #endif
 322             twoeweps         = _mm_add_pd(eweps,eweps);
 323             ewitab           = _mm_slli_epi32(ewitab,2);
 324             ewtabF           = _mm_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
 325             ewtabD           = _mm_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
 326             GMX_MM_TRANSPOSE2_PD(ewtabF,ewtabD);
 327             ewtabV           = _mm_load_sd( ewtab + _mm_extract_epi32(ewitab,0) +2);
 328             ewtabFn          = _mm_load_sd( ewtab + _mm_extract_epi32(ewitab,1) +2);
 329             GMX_MM_TRANSPOSE2_PD(ewtabV,ewtabFn);
 330             felec            = _mm_macc_pd(eweps,ewtabD,ewtabF);
 331             velec            = _mm_nmacc_pd(_mm_mul_pd(ewtabhalfspace,eweps) ,_mm_add_pd(ewtabF,felec), ewtabV);
 332             velec            = _mm_mul_pd(qq10,_mm_sub_pd(rinv10,velec));
 333             felec            = _mm_mul_pd(_mm_mul_pd(qq10,rinv10),_mm_sub_pd(rinvsq10,felec));
 334
 335             /* Update potential sum for this i atom from the interaction with this j atom. */
 336             velecsum         = _mm_add_pd(velecsum,velec);
 337
 338             fscal            = felec;
 339
 340             /* Update vectorial force */
 341             fix1             = _mm_macc_pd(dx10,fscal,fix1);
 342             fiy1             = _mm_macc_pd(dy10,fscal,fiy1);
 343             fiz1             = _mm_macc_pd(dz10,fscal,fiz1);
 344
 345             fjx0             = _mm_macc_pd(dx10,fscal,fjx0);
 346             fjy0             = _mm_macc_pd(dy10,fscal,fjy0);
 347             fjz0             = _mm_macc_pd(dz10,fscal,fjz0);
 348
 349             /**************************
 350              * CALCULATE INTERACTIONS *
 351              **************************/
 352
 353             r20              = _mm_mul_pd(rsq20,rinv20);
 354
 355             /* Compute parameters for interactions between i and j atoms */
 356             qq20             = _mm_mul_pd(iq2,jq0);
 357
 358             /* EWALD ELECTROSTATICS */
 359
 360             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 361             ewrt             = _mm_mul_pd(r20,ewtabscale);
 362             ewitab           = _mm_cvttpd_epi32(ewrt);
 363 #ifdef __XOP__
 364             eweps            = _mm_frcz_pd(ewrt);
 365 #else
 366             eweps            = _mm_sub_pd(ewrt,_mm_round_pd(ewrt, _MM_FROUND_FLOOR));
 367 #endif
 368             twoeweps         = _mm_add_pd(eweps,eweps);
 369             ewitab           = _mm_slli_epi32(ewitab,2);
 370             ewtabF           = _mm_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
 371             ewtabD           = _mm_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
 372             GMX_MM_TRANSPOSE2_PD(ewtabF,ewtabD);
 373             ewtabV           = _mm_load_sd( ewtab + _mm_extract_epi32(ewitab,0) +2);
 374             ewtabFn          = _mm_load_sd( ewtab + _mm_extract_epi32(ewitab,1) +2);
 375             GMX_MM_TRANSPOSE2_PD(ewtabV,ewtabFn);
 376             felec            = _mm_macc_pd(eweps,ewtabD,ewtabF);
 377             velec            = _mm_nmacc_pd(_mm_mul_pd(ewtabhalfspace,eweps) ,_mm_add_pd(ewtabF,felec), ewtabV);
 378             velec            = _mm_mul_pd(qq20,_mm_sub_pd(rinv20,velec));
 379             felec            = _mm_mul_pd(_mm_mul_pd(qq20,rinv20),_mm_sub_pd(rinvsq20,felec));
 380
 381             /* Update potential sum for this i atom from the interaction with this j atom. */
 382             velecsum         = _mm_add_pd(velecsum,velec);
 383
 384             fscal            = felec;
 385
 386             /* Update vectorial force */
 387             fix2             = _mm_macc_pd(dx20,fscal,fix2);
 388             fiy2             = _mm_macc_pd(dy20,fscal,fiy2);
 389             fiz2             = _mm_macc_pd(dz20,fscal,fiz2);
 390
 391             fjx0             = _mm_macc_pd(dx20,fscal,fjx0);
 392             fjy0             = _mm_macc_pd(dy20,fscal,fjy0);
 393             fjz0             = _mm_macc_pd(dz20,fscal,fjz0);
 394
 395             /**************************
 396              * CALCULATE INTERACTIONS *
 397              **************************/
 398
 399             r30              = _mm_mul_pd(rsq30,rinv30);
 400
 401             /* Compute parameters for interactions between i and j atoms */
 402             qq30             = _mm_mul_pd(iq3,jq0);
 403
 404             /* EWALD ELECTROSTATICS */
 405
 406             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 407             ewrt             = _mm_mul_pd(r30,ewtabscale);
 408             ewitab           = _mm_cvttpd_epi32(ewrt);
 409 #ifdef __XOP__
 410             eweps            = _mm_frcz_pd(ewrt);
 411 #else
 412             eweps            = _mm_sub_pd(ewrt,_mm_round_pd(ewrt, _MM_FROUND_FLOOR));
 413 #endif
 414             twoeweps         = _mm_add_pd(eweps,eweps);
 415             ewitab           = _mm_slli_epi32(ewitab,2);
 416             ewtabF           = _mm_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
 417             ewtabD           = _mm_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
 418             GMX_MM_TRANSPOSE2_PD(ewtabF,ewtabD);
 419             ewtabV           = _mm_load_sd( ewtab + _mm_extract_epi32(ewitab,0) +2);
 420             ewtabFn          = _mm_load_sd( ewtab + _mm_extract_epi32(ewitab,1) +2);
 421             GMX_MM_TRANSPOSE2_PD(ewtabV,ewtabFn);
 422             felec            = _mm_macc_pd(eweps,ewtabD,ewtabF);
 423             velec            = _mm_nmacc_pd(_mm_mul_pd(ewtabhalfspace,eweps) ,_mm_add_pd(ewtabF,felec), ewtabV);
 424             velec            = _mm_mul_pd(qq30,_mm_sub_pd(rinv30,velec));
 425             felec            = _mm_mul_pd(_mm_mul_pd(qq30,rinv30),_mm_sub_pd(rinvsq30,felec));
 426
 427             /* Update potential sum for this i atom from the interaction with this j atom. */
 428             velecsum         = _mm_add_pd(velecsum,velec);
 429
 430             fscal            = felec;
 431
 432             /* Update vectorial force */
 433             fix3             = _mm_macc_pd(dx30,fscal,fix3);
 434             fiy3             = _mm_macc_pd(dy30,fscal,fiy3);
 435             fiz3             = _mm_macc_pd(dz30,fscal,fiz3);
 436
 437             fjx0             = _mm_macc_pd(dx30,fscal,fjx0);
 438             fjy0             = _mm_macc_pd(dy30,fscal,fjy0);
 439             fjz0             = _mm_macc_pd(dz30,fscal,fjz0);
 440
 441             gmx_mm_decrement_1rvec_2ptr_swizzle_pd(f+j_coord_offsetA,f+j_coord_offsetB,fjx0,fjy0,fjz0);
 442
 443             /* Inner loop uses 194 flops */
 444         }
 445
 446         if(jidx<j_index_end)
 447         {
 448
 449             jnrA             = jjnr[jidx];
 450             j_coord_offsetA  = DIM*jnrA;
 451
 452             /* load j atom coordinates */
 453             gmx_mm_load_1rvec_1ptr_swizzle_pd(x+j_coord_offsetA,
 454                                               &jx0,&jy0,&jz0);
 455
 456             /* Calculate displacement vector */
 457             dx00             = _mm_sub_pd(ix0,jx0);
 458             dy00             = _mm_sub_pd(iy0,jy0);
 459             dz00             = _mm_sub_pd(iz0,jz0);
 460             dx10             = _mm_sub_pd(ix1,jx0);
 461             dy10             = _mm_sub_pd(iy1,jy0);
 462             dz10             = _mm_sub_pd(iz1,jz0);
 463             dx20             = _mm_sub_pd(ix2,jx0);
 464             dy20             = _mm_sub_pd(iy2,jy0);
 465             dz20             = _mm_sub_pd(iz2,jz0);
 466             dx30             = _mm_sub_pd(ix3,jx0);
 467             dy30             = _mm_sub_pd(iy3,jy0);
 468             dz30             = _mm_sub_pd(iz3,jz0);
 469
 470             /* Calculate squared distance and things based on it */
 471             rsq00            = gmx_mm_calc_rsq_pd(dx00,dy00,dz00);
 472             rsq10            = gmx_mm_calc_rsq_pd(dx10,dy10,dz10);
 473             rsq20            = gmx_mm_calc_rsq_pd(dx20,dy20,dz20);
 474             rsq30            = gmx_mm_calc_rsq_pd(dx30,dy30,dz30);
 475
 476             rinv00           = avx128fma_invsqrt_d(rsq00);
 477             rinv10           = avx128fma_invsqrt_d(rsq10);
 478             rinv20           = avx128fma_invsqrt_d(rsq20);
 479             rinv30           = avx128fma_invsqrt_d(rsq30);
 480
 481             rinvsq10         = _mm_mul_pd(rinv10,rinv10);
 482             rinvsq20         = _mm_mul_pd(rinv20,rinv20);
 483             rinvsq30         = _mm_mul_pd(rinv30,rinv30);
 484
 485             /* Load parameters for j particles */
 486             jq0              = _mm_load_sd(charge+jnrA+0);
 487             vdwjidx0A        = 2*vdwtype[jnrA+0];
 488
 489             fjx0             = _mm_setzero_pd();
 490             fjy0             = _mm_setzero_pd();
 491             fjz0             = _mm_setzero_pd();
 492
 493             /**************************
 494              * CALCULATE INTERACTIONS *
 495              **************************/
 496
 497             r00              = _mm_mul_pd(rsq00,rinv00);
 498
 499             /* Compute parameters for interactions between i and j atoms */
 500             gmx_mm_load_1pair_swizzle_pd(vdwparam+vdwioffset0+vdwjidx0A,&c6_00,&c12_00);
 501
 502             /* Calculate table index by multiplying r with table scale and truncate to integer */
 503             rt               = _mm_mul_pd(r00,vftabscale);
 504             vfitab           = _mm_cvttpd_epi32(rt);
 505 #ifdef __XOP__
 506             vfeps            = _mm_frcz_pd(rt);
 507 #else
 508             vfeps            = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
 509 #endif
 510             twovfeps         = _mm_add_pd(vfeps,vfeps);
 511             vfitab           = _mm_slli_epi32(vfitab,3);
 512
 513             /* CUBIC SPLINE TABLE DISPERSION */
 514             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 515             F                = _mm_setzero_pd();
 516             GMX_MM_TRANSPOSE2_PD(Y,F);
 517             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
 518             H                = _mm_setzero_pd();
 519             GMX_MM_TRANSPOSE2_PD(G,H);
 520             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(H,vfeps,G),F);
 521             VV               = _mm_macc_pd(vfeps,Fp,Y);
 522             vvdw6            = _mm_mul_pd(c6_00,VV);
 523             FF               = _mm_macc_pd(vfeps,_mm_macc_pd(twovfeps,H,G),Fp);
 524             fvdw6            = _mm_mul_pd(c6_00,FF);
 525
 526             /* CUBIC SPLINE TABLE REPULSION */
 527             vfitab           = _mm_add_epi32(vfitab,ifour);
 528             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 529             F                = _mm_setzero_pd();
 530             GMX_MM_TRANSPOSE2_PD(Y,F);
 531             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
 532             H                = _mm_setzero_pd();
 533             GMX_MM_TRANSPOSE2_PD(G,H);
 534             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(H,vfeps,G),F);
 535             VV               = _mm_macc_pd(vfeps,Fp,Y);
 536             vvdw12           = _mm_mul_pd(c12_00,VV);
 537             FF               = _mm_macc_pd(vfeps,_mm_macc_pd(twovfeps,H,G),Fp);
 538             fvdw12           = _mm_mul_pd(c12_00,FF);
 539             vvdw             = _mm_add_pd(vvdw12,vvdw6);
 540             fvdw             = _mm_xor_pd(signbit,_mm_mul_pd(_mm_add_pd(fvdw6,fvdw12),_mm_mul_pd(vftabscale,rinv00)));
 541
 542             /* Update potential sum for this i atom from the interaction with this j atom. */
 543             vvdw             = _mm_unpacklo_pd(vvdw,_mm_setzero_pd());
 544             vvdwsum          = _mm_add_pd(vvdwsum,vvdw);
 545
 546             fscal            = fvdw;
 547
 548             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
 549
 550             /* Update vectorial force */
 551             fix0             = _mm_macc_pd(dx00,fscal,fix0);
 552             fiy0             = _mm_macc_pd(dy00,fscal,fiy0);
 553             fiz0             = _mm_macc_pd(dz00,fscal,fiz0);
 554
 555             fjx0             = _mm_macc_pd(dx00,fscal,fjx0);
 556             fjy0             = _mm_macc_pd(dy00,fscal,fjy0);
 557             fjz0             = _mm_macc_pd(dz00,fscal,fjz0);
 558
 559             /**************************
 560              * CALCULATE INTERACTIONS *
 561              **************************/
 562
 563             r10              = _mm_mul_pd(rsq10,rinv10);
 564
 565             /* Compute parameters for interactions between i and j atoms */
 566             qq10             = _mm_mul_pd(iq1,jq0);
 567
 568             /* EWALD ELECTROSTATICS */
 569
 570             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 571             ewrt             = _mm_mul_pd(r10,ewtabscale);
 572             ewitab           = _mm_cvttpd_epi32(ewrt);
 573 #ifdef __XOP__
 574             eweps            = _mm_frcz_pd(ewrt);
 575 #else
 576             eweps            = _mm_sub_pd(ewrt,_mm_round_pd(ewrt, _MM_FROUND_FLOOR));
 577 #endif
 578             twoeweps         = _mm_add_pd(eweps,eweps);
 579             ewitab           = _mm_slli_epi32(ewitab,2);
 580             ewtabF           = _mm_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
 581             ewtabD           = _mm_setzero_pd();
 582             GMX_MM_TRANSPOSE2_PD(ewtabF,ewtabD);
 583             ewtabV           = _mm_load_sd( ewtab + _mm_extract_epi32(ewitab,0) +2);
 584             ewtabFn          = _mm_setzero_pd();
 585             GMX_MM_TRANSPOSE2_PD(ewtabV,ewtabFn);
 586             felec            = _mm_macc_pd(eweps,ewtabD,ewtabF);
 587             velec            = _mm_nmacc_pd(_mm_mul_pd(ewtabhalfspace,eweps) ,_mm_add_pd(ewtabF,felec), ewtabV);
 588             velec            = _mm_mul_pd(qq10,_mm_sub_pd(rinv10,velec));
 589             felec            = _mm_mul_pd(_mm_mul_pd(qq10,rinv10),_mm_sub_pd(rinvsq10,felec));
 590
 591             /* Update potential sum for this i atom from the interaction with this j atom. */
 592             velec            = _mm_unpacklo_pd(velec,_mm_setzero_pd());
 593             velecsum         = _mm_add_pd(velecsum,velec);
 594
 595             fscal            = felec;
 596
 597             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
 598
 599             /* Update vectorial force */
 600             fix1             = _mm_macc_pd(dx10,fscal,fix1);
 601             fiy1             = _mm_macc_pd(dy10,fscal,fiy1);
 602             fiz1             = _mm_macc_pd(dz10,fscal,fiz1);
 603
 604             fjx0             = _mm_macc_pd(dx10,fscal,fjx0);
 605             fjy0             = _mm_macc_pd(dy10,fscal,fjy0);
 606             fjz0             = _mm_macc_pd(dz10,fscal,fjz0);
 607
 608             /**************************
 609              * CALCULATE INTERACTIONS *
 610              **************************/
 611
 612             r20              = _mm_mul_pd(rsq20,rinv20);
 613
 614             /* Compute parameters for interactions between i and j atoms */
 615             qq20             = _mm_mul_pd(iq2,jq0);
 616
 617             /* EWALD ELECTROSTATICS */
 618
 619             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 620             ewrt             = _mm_mul_pd(r20,ewtabscale);
 621             ewitab           = _mm_cvttpd_epi32(ewrt);
 622 #ifdef __XOP__
 623             eweps            = _mm_frcz_pd(ewrt);
 624 #else
 625             eweps            = _mm_sub_pd(ewrt,_mm_round_pd(ewrt, _MM_FROUND_FLOOR));
 626 #endif
 627             twoeweps         = _mm_add_pd(eweps,eweps);
 628             ewitab           = _mm_slli_epi32(ewitab,2);
 629             ewtabF           = _mm_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
 630             ewtabD           = _mm_setzero_pd();
 631             GMX_MM_TRANSPOSE2_PD(ewtabF,ewtabD);
 632             ewtabV           = _mm_load_sd( ewtab + _mm_extract_epi32(ewitab,0) +2);
 633             ewtabFn          = _mm_setzero_pd();
 634             GMX_MM_TRANSPOSE2_PD(ewtabV,ewtabFn);
 635             felec            = _mm_macc_pd(eweps,ewtabD,ewtabF);
 636             velec            = _mm_nmacc_pd(_mm_mul_pd(ewtabhalfspace,eweps) ,_mm_add_pd(ewtabF,felec), ewtabV);
 637             velec            = _mm_mul_pd(qq20,_mm_sub_pd(rinv20,velec));
 638             felec            = _mm_mul_pd(_mm_mul_pd(qq20,rinv20),_mm_sub_pd(rinvsq20,felec));
 639
 640             /* Update potential sum for this i atom from the interaction with this j atom. */
 641             velec            = _mm_unpacklo_pd(velec,_mm_setzero_pd());
 642             velecsum         = _mm_add_pd(velecsum,velec);
 643
 644             fscal            = felec;
 645
 646             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
 647
 648             /* Update vectorial force */
 649             fix2             = _mm_macc_pd(dx20,fscal,fix2);
 650             fiy2             = _mm_macc_pd(dy20,fscal,fiy2);
 651             fiz2             = _mm_macc_pd(dz20,fscal,fiz2);
 652
 653             fjx0             = _mm_macc_pd(dx20,fscal,fjx0);
 654             fjy0             = _mm_macc_pd(dy20,fscal,fjy0);
 655             fjz0             = _mm_macc_pd(dz20,fscal,fjz0);
 656
 657             /**************************
 658              * CALCULATE INTERACTIONS *
 659              **************************/
 660
 661             r30              = _mm_mul_pd(rsq30,rinv30);
 662
 663             /* Compute parameters for interactions between i and j atoms */
 664             qq30             = _mm_mul_pd(iq3,jq0);
 665
 666             /* EWALD ELECTROSTATICS */
 667
 668             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 669             ewrt             = _mm_mul_pd(r30,ewtabscale);
 670             ewitab           = _mm_cvttpd_epi32(ewrt);
 671 #ifdef __XOP__
 672             eweps            = _mm_frcz_pd(ewrt);
 673 #else
 674             eweps            = _mm_sub_pd(ewrt,_mm_round_pd(ewrt, _MM_FROUND_FLOOR));
 675 #endif
 676             twoeweps         = _mm_add_pd(eweps,eweps);
 677             ewitab           = _mm_slli_epi32(ewitab,2);
 678             ewtabF           = _mm_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
 679             ewtabD           = _mm_setzero_pd();
 680             GMX_MM_TRANSPOSE2_PD(ewtabF,ewtabD);
 681             ewtabV           = _mm_load_sd( ewtab + _mm_extract_epi32(ewitab,0) +2);
 682             ewtabFn          = _mm_setzero_pd();
 683             GMX_MM_TRANSPOSE2_PD(ewtabV,ewtabFn);
 684             felec            = _mm_macc_pd(eweps,ewtabD,ewtabF);
 685             velec            = _mm_nmacc_pd(_mm_mul_pd(ewtabhalfspace,eweps) ,_mm_add_pd(ewtabF,felec), ewtabV);
 686             velec            = _mm_mul_pd(qq30,_mm_sub_pd(rinv30,velec));
 687             felec            = _mm_mul_pd(_mm_mul_pd(qq30,rinv30),_mm_sub_pd(rinvsq30,felec));
 688
 689             /* Update potential sum for this i atom from the interaction with this j atom. */
 690             velec            = _mm_unpacklo_pd(velec,_mm_setzero_pd());
 691             velecsum         = _mm_add_pd(velecsum,velec);
 692
 693             fscal            = felec;
 694
 695             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
 696
 697             /* Update vectorial force */
 698             fix3             = _mm_macc_pd(dx30,fscal,fix3);
 699             fiy3             = _mm_macc_pd(dy30,fscal,fiy3);
 700             fiz3             = _mm_macc_pd(dz30,fscal,fiz3);
 701
 702             fjx0             = _mm_macc_pd(dx30,fscal,fjx0);
 703             fjy0             = _mm_macc_pd(dy30,fscal,fjy0);
 704             fjz0             = _mm_macc_pd(dz30,fscal,fjz0);
 705
 706             gmx_mm_decrement_1rvec_1ptr_swizzle_pd(f+j_coord_offsetA,fjx0,fjy0,fjz0);
 707
 708             /* Inner loop uses 194 flops */
 709         }
 710
 711         /* End of innermost loop */
 712
 713         gmx_mm_update_iforce_4atom_swizzle_pd(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
 714                                               f+i_coord_offset,fshift+i_shift_offset);
 715
 716         ggid                        = gid[iidx];
 717         /* Update potential energies */
 718         gmx_mm_update_1pot_pd(velecsum,kernel_data->energygrp_elec+ggid);
 719         gmx_mm_update_1pot_pd(vvdwsum,kernel_data->energygrp_vdw+ggid);
 720
 721         /* Increment number of inner iterations */
 722         inneriter                  += j_index_end - j_index_start;
 723
 724         /* Outer loop uses 26 flops */
 725     }
 726
 727     /* Increment number of outer iterations */
 728     outeriter        += nri;
 729
 730     /* Update outer/inner flops */
 731
 732     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_VF,outeriter*26 + inneriter*194);
 733 }
 734 /*
 735  * Gromacs nonbonded kernel:   nb_kernel_ElecEw_VdwCSTab_GeomW4P1_F_avx_128_fma_double
 736  * Electrostatics interaction: Ewald
 737  * VdW interaction:            CubicSplineTable
 738  * Geometry:                   Water4-Particle
 739  * Calculate force/pot:        Force
 740  */
 741 void
 742 nb_kernel_ElecEw_VdwCSTab_GeomW4P1_F_avx_128_fma_double
 743                     (t_nblist                    * gmx_restrict       nlist,
 744                      rvec                        * gmx_restrict          xx,
 745                      rvec                        * gmx_restrict          ff,
 746                      struct t_forcerec           * gmx_restrict          fr,
 747                      t_mdatoms                   * gmx_restrict     mdatoms,
 748                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 749                      t_nrnb                      * gmx_restrict        nrnb)
 750 {
 751     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
 752      * just 0 for non-waters.
 753      * Suffixes A,B refer to j loop unrolling done with SSE double precision, e.g. for the two different
 754      * jnr indices corresponding to data put in the four positions in the SIMD register.
 755      */
 756     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
 757     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
 758     int              jnrA,jnrB;
 759     int              j_coord_offsetA,j_coord_offsetB;
 760     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 761     real             rcutoff_scalar;
 762     real             *shiftvec,*fshift,*x,*f;
 763     __m128d          tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
 764     int              vdwioffset0;
 765     __m128d          ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 766     int              vdwioffset1;
 767     __m128d          ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
 768     int              vdwioffset2;
 769     __m128d          ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
 770     int              vdwioffset3;
 771     __m128d          ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
 772     int              vdwjidx0A,vdwjidx0B;
 773     __m128d          jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 774     __m128d          dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
 775     __m128d          dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
 776     __m128d          dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
 777     __m128d          dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30;
 778     __m128d          velec,felec,velecsum,facel,crf,krf,krf2;
 779     real             *charge;
 780     int              nvdwtype;
 781     __m128d          rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
 782     int              *vdwtype;
 783     real             *vdwparam;
 784     __m128d          one_sixth   = _mm_set1_pd(1.0/6.0);
 785     __m128d          one_twelfth = _mm_set1_pd(1.0/12.0);
 786     __m128i          vfitab;
 787     __m128i          ifour       = _mm_set1_epi32(4);
 788     __m128d          rt,vfeps,vftabscale,Y,F,G,H,Heps,Fp,VV,FF,twovfeps;
 789     real             *vftab;
 790     __m128i          ewitab;
 791     __m128d          ewtabscale,eweps,twoeweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
 792     real             *ewtab;
 793     __m128d          dummy_mask,cutoff_mask;
 794     __m128d          signbit   = gmx_mm_castsi128_pd( _mm_set_epi32(0x80000000,0x00000000,0x80000000,0x00000000) );
 795     __m128d          one     = _mm_set1_pd(1.0);
 796     __m128d          two     = _mm_set1_pd(2.0);
 797     x                = xx[0];
 798     f                = ff[0];
 799
 800     nri              = nlist->nri;
 801     iinr             = nlist->iinr;
 802     jindex           = nlist->jindex;
 803     jjnr             = nlist->jjnr;
 804     shiftidx         = nlist->shift;
 805     gid              = nlist->gid;
 806     shiftvec         = fr->shift_vec[0];
 807     fshift           = fr->fshift[0];
 808     facel            = _mm_set1_pd(fr->ic->epsfac);
 809     charge           = mdatoms->chargeA;
 810     nvdwtype         = fr->ntype;
 811     vdwparam         = fr->nbfp;
 812     vdwtype          = mdatoms->typeA;
 813
 814     vftab            = kernel_data->table_vdw->data;
 815     vftabscale       = _mm_set1_pd(kernel_data->table_vdw->scale);
 816
 817     sh_ewald         = _mm_set1_pd(fr->ic->sh_ewald);
 818     ewtab            = fr->ic->tabq_coul_F;
 819     ewtabscale       = _mm_set1_pd(fr->ic->tabq_scale);
 820     ewtabhalfspace   = _mm_set1_pd(0.5/fr->ic->tabq_scale);
 821
 822     /* Setup water-specific parameters */
 823     inr              = nlist->iinr[0];
 824     iq1              = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+1]));
 825     iq2              = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+2]));
 826     iq3              = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+3]));
 827     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 828
 829     /* Avoid stupid compiler warnings */
 830     jnrA = jnrB = 0;
 831     j_coord_offsetA = 0;
 832     j_coord_offsetB = 0;
 833
 834     outeriter        = 0;
 835     inneriter        = 0;
 836
 837     /* Start outer loop over neighborlists */
 838     for(iidx=0; iidx<nri; iidx++)
 839     {
 840         /* Load shift vector for this list */
 841         i_shift_offset   = DIM*shiftidx[iidx];
 842
 843         /* Load limits for loop over neighbors */
 844         j_index_start    = jindex[iidx];
 845         j_index_end      = jindex[iidx+1];
 846
 847         /* Get outer coordinate index */
 848         inr              = iinr[iidx];
 849         i_coord_offset   = DIM*inr;
 850
 851         /* Load i particle coords and add shift vector */
 852         gmx_mm_load_shift_and_4rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset,
 853                                                  &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
 854
 855         fix0             = _mm_setzero_pd();
 856         fiy0             = _mm_setzero_pd();
 857         fiz0             = _mm_setzero_pd();
 858         fix1             = _mm_setzero_pd();
 859         fiy1             = _mm_setzero_pd();
 860         fiz1             = _mm_setzero_pd();
 861         fix2             = _mm_setzero_pd();
 862         fiy2             = _mm_setzero_pd();
 863         fiz2             = _mm_setzero_pd();
 864         fix3             = _mm_setzero_pd();
 865         fiy3             = _mm_setzero_pd();
 866         fiz3             = _mm_setzero_pd();
 867
 868         /* Start inner kernel loop */
 869         for(jidx=j_index_start; jidx<j_index_end-1; jidx+=2)
 870         {
 871
 872             /* Get j neighbor index, and coordinate index */
 873             jnrA             = jjnr[jidx];
 874             jnrB             = jjnr[jidx+1];
 875             j_coord_offsetA  = DIM*jnrA;
 876             j_coord_offsetB  = DIM*jnrB;
 877
 878             /* load j atom coordinates */
 879             gmx_mm_load_1rvec_2ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
 880                                               &jx0,&jy0,&jz0);
 881
 882             /* Calculate displacement vector */
 883             dx00             = _mm_sub_pd(ix0,jx0);
 884             dy00             = _mm_sub_pd(iy0,jy0);
 885             dz00             = _mm_sub_pd(iz0,jz0);
 886             dx10             = _mm_sub_pd(ix1,jx0);
 887             dy10             = _mm_sub_pd(iy1,jy0);
 888             dz10             = _mm_sub_pd(iz1,jz0);
 889             dx20             = _mm_sub_pd(ix2,jx0);
 890             dy20             = _mm_sub_pd(iy2,jy0);
 891             dz20             = _mm_sub_pd(iz2,jz0);
 892             dx30             = _mm_sub_pd(ix3,jx0);
 893             dy30             = _mm_sub_pd(iy3,jy0);
 894             dz30             = _mm_sub_pd(iz3,jz0);
 895
 896             /* Calculate squared distance and things based on it */
 897             rsq00            = gmx_mm_calc_rsq_pd(dx00,dy00,dz00);
 898             rsq10            = gmx_mm_calc_rsq_pd(dx10,dy10,dz10);
 899             rsq20            = gmx_mm_calc_rsq_pd(dx20,dy20,dz20);
 900             rsq30            = gmx_mm_calc_rsq_pd(dx30,dy30,dz30);
 901
 902             rinv00           = avx128fma_invsqrt_d(rsq00);
 903             rinv10           = avx128fma_invsqrt_d(rsq10);
 904             rinv20           = avx128fma_invsqrt_d(rsq20);
 905             rinv30           = avx128fma_invsqrt_d(rsq30);
 906
 907             rinvsq10         = _mm_mul_pd(rinv10,rinv10);
 908             rinvsq20         = _mm_mul_pd(rinv20,rinv20);
 909             rinvsq30         = _mm_mul_pd(rinv30,rinv30);
 910
 911             /* Load parameters for j particles */
 912             jq0              = gmx_mm_load_2real_swizzle_pd(charge+jnrA+0,charge+jnrB+0);
 913             vdwjidx0A        = 2*vdwtype[jnrA+0];
 914             vdwjidx0B        = 2*vdwtype[jnrB+0];
 915
 916             fjx0             = _mm_setzero_pd();
 917             fjy0             = _mm_setzero_pd();
 918             fjz0             = _mm_setzero_pd();
 919
 920             /**************************
 921              * CALCULATE INTERACTIONS *
 922              **************************/
 923
 924             r00              = _mm_mul_pd(rsq00,rinv00);
 925
 926             /* Compute parameters for interactions between i and j atoms */
 927             gmx_mm_load_2pair_swizzle_pd(vdwparam+vdwioffset0+vdwjidx0A,
 928                                          vdwparam+vdwioffset0+vdwjidx0B,&c6_00,&c12_00);
 929
 930             /* Calculate table index by multiplying r with table scale and truncate to integer */
 931             rt               = _mm_mul_pd(r00,vftabscale);
 932             vfitab           = _mm_cvttpd_epi32(rt);
 933 #ifdef __XOP__
 934             vfeps            = _mm_frcz_pd(rt);
 935 #else
 936             vfeps            = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
 937 #endif
 938             twovfeps         = _mm_add_pd(vfeps,vfeps);
 939             vfitab           = _mm_slli_epi32(vfitab,3);
 940
 941             /* CUBIC SPLINE TABLE DISPERSION */
 942             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 943             F                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) );
 944             GMX_MM_TRANSPOSE2_PD(Y,F);
 945             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
 946             H                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) +2);
 947             GMX_MM_TRANSPOSE2_PD(G,H);
 948             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(H,vfeps,G),F);
 949             FF               = _mm_macc_pd(vfeps,_mm_macc_pd(twovfeps,H,G),Fp);
 950             fvdw6            = _mm_mul_pd(c6_00,FF);
 951
 952             /* CUBIC SPLINE TABLE REPULSION */
 953             vfitab           = _mm_add_epi32(vfitab,ifour);
 954             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
 955             F                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) );
 956             GMX_MM_TRANSPOSE2_PD(Y,F);
 957             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
 958             H                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) +2);
 959             GMX_MM_TRANSPOSE2_PD(G,H);
 960             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(H,vfeps,G),F);
 961             FF               = _mm_macc_pd(vfeps,_mm_macc_pd(twovfeps,H,G),Fp);
 962             fvdw12           = _mm_mul_pd(c12_00,FF);
 963             fvdw             = _mm_xor_pd(signbit,_mm_mul_pd(_mm_add_pd(fvdw6,fvdw12),_mm_mul_pd(vftabscale,rinv00)));
 964
 965             fscal            = fvdw;
 966
 967             /* Update vectorial force */
 968             fix0             = _mm_macc_pd(dx00,fscal,fix0);
 969             fiy0             = _mm_macc_pd(dy00,fscal,fiy0);
 970             fiz0             = _mm_macc_pd(dz00,fscal,fiz0);
 971
 972             fjx0             = _mm_macc_pd(dx00,fscal,fjx0);
 973             fjy0             = _mm_macc_pd(dy00,fscal,fjy0);
 974             fjz0             = _mm_macc_pd(dz00,fscal,fjz0);
 975
 976             /**************************
 977              * CALCULATE INTERACTIONS *
 978              **************************/
 979
 980             r10              = _mm_mul_pd(rsq10,rinv10);
 981
 982             /* Compute parameters for interactions between i and j atoms */
 983             qq10             = _mm_mul_pd(iq1,jq0);
 984
 985             /* EWALD ELECTROSTATICS */
 986
 987             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 988             ewrt             = _mm_mul_pd(r10,ewtabscale);
 989             ewitab           = _mm_cvttpd_epi32(ewrt);
 990 #ifdef __XOP__
 991             eweps            = _mm_frcz_pd(ewrt);
 992 #else
 993             eweps            = _mm_sub_pd(ewrt,_mm_round_pd(ewrt, _MM_FROUND_FLOOR));
 994 #endif
 995             twoeweps         = _mm_add_pd(eweps,eweps);
 996             gmx_mm_load_2pair_swizzle_pd(ewtab+_mm_extract_epi32(ewitab,0),ewtab+_mm_extract_epi32(ewitab,1),
 997                                          &ewtabF,&ewtabFn);
 998             felec            = _mm_macc_pd(eweps,ewtabFn,_mm_mul_pd( _mm_sub_pd(one,eweps),ewtabF));
 999             felec            = _mm_mul_pd(_mm_mul_pd(qq10,rinv10),_mm_sub_pd(rinvsq10,felec));
1000
1001             fscal            = felec;
1002
1003             /* Update vectorial force */
1004             fix1             = _mm_macc_pd(dx10,fscal,fix1);
1005             fiy1             = _mm_macc_pd(dy10,fscal,fiy1);
1006             fiz1             = _mm_macc_pd(dz10,fscal,fiz1);
1007
1008             fjx0             = _mm_macc_pd(dx10,fscal,fjx0);
1009             fjy0             = _mm_macc_pd(dy10,fscal,fjy0);
1010             fjz0             = _mm_macc_pd(dz10,fscal,fjz0);
1011
1012             /**************************
1013              * CALCULATE INTERACTIONS *
1014              **************************/
1015
1016             r20              = _mm_mul_pd(rsq20,rinv20);
1017
1018             /* Compute parameters for interactions between i and j atoms */
1019             qq20             = _mm_mul_pd(iq2,jq0);
1020
1021             /* EWALD ELECTROSTATICS */
1022
1023             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1024             ewrt             = _mm_mul_pd(r20,ewtabscale);
1025             ewitab           = _mm_cvttpd_epi32(ewrt);
1026 #ifdef __XOP__
1027             eweps            = _mm_frcz_pd(ewrt);
1028 #else
1029             eweps            = _mm_sub_pd(ewrt,_mm_round_pd(ewrt, _MM_FROUND_FLOOR));
1030 #endif
1031             twoeweps         = _mm_add_pd(eweps,eweps);
1032             gmx_mm_load_2pair_swizzle_pd(ewtab+_mm_extract_epi32(ewitab,0),ewtab+_mm_extract_epi32(ewitab,1),
1033                                          &ewtabF,&ewtabFn);
1034             felec            = _mm_macc_pd(eweps,ewtabFn,_mm_mul_pd( _mm_sub_pd(one,eweps),ewtabF));
1035             felec            = _mm_mul_pd(_mm_mul_pd(qq20,rinv20),_mm_sub_pd(rinvsq20,felec));
1036
1037             fscal            = felec;
1038
1039             /* Update vectorial force */
1040             fix2             = _mm_macc_pd(dx20,fscal,fix2);
1041             fiy2             = _mm_macc_pd(dy20,fscal,fiy2);
1042             fiz2             = _mm_macc_pd(dz20,fscal,fiz2);
1043
1044             fjx0             = _mm_macc_pd(dx20,fscal,fjx0);
1045             fjy0             = _mm_macc_pd(dy20,fscal,fjy0);
1046             fjz0             = _mm_macc_pd(dz20,fscal,fjz0);
1047
1048             /**************************
1049              * CALCULATE INTERACTIONS *
1050              **************************/
1051
1052             r30              = _mm_mul_pd(rsq30,rinv30);
1053
1054             /* Compute parameters for interactions between i and j atoms */
1055             qq30             = _mm_mul_pd(iq3,jq0);
1056
1057             /* EWALD ELECTROSTATICS */
1058
1059             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1060             ewrt             = _mm_mul_pd(r30,ewtabscale);
1061             ewitab           = _mm_cvttpd_epi32(ewrt);
1062 #ifdef __XOP__
1063             eweps            = _mm_frcz_pd(ewrt);
1064 #else
1065             eweps            = _mm_sub_pd(ewrt,_mm_round_pd(ewrt, _MM_FROUND_FLOOR));
1066 #endif
1067             twoeweps         = _mm_add_pd(eweps,eweps);
1068             gmx_mm_load_2pair_swizzle_pd(ewtab+_mm_extract_epi32(ewitab,0),ewtab+_mm_extract_epi32(ewitab,1),
1069                                          &ewtabF,&ewtabFn);
1070             felec            = _mm_macc_pd(eweps,ewtabFn,_mm_mul_pd( _mm_sub_pd(one,eweps),ewtabF));
1071             felec            = _mm_mul_pd(_mm_mul_pd(qq30,rinv30),_mm_sub_pd(rinvsq30,felec));
1072
1073             fscal            = felec;
1074
1075             /* Update vectorial force */
1076             fix3             = _mm_macc_pd(dx30,fscal,fix3);
1077             fiy3             = _mm_macc_pd(dy30,fscal,fiy3);
1078             fiz3             = _mm_macc_pd(dz30,fscal,fiz3);
1079
1080             fjx0             = _mm_macc_pd(dx30,fscal,fjx0);
1081             fjy0             = _mm_macc_pd(dy30,fscal,fjy0);
1082             fjz0             = _mm_macc_pd(dz30,fscal,fjz0);
1083
1084             gmx_mm_decrement_1rvec_2ptr_swizzle_pd(f+j_coord_offsetA,f+j_coord_offsetB,fjx0,fjy0,fjz0);
1085
1086             /* Inner loop uses 171 flops */
1087         }
1088
1089         if(jidx<j_index_end)
1090         {
1091
1092             jnrA             = jjnr[jidx];
1093             j_coord_offsetA  = DIM*jnrA;
1094
1095             /* load j atom coordinates */
1096             gmx_mm_load_1rvec_1ptr_swizzle_pd(x+j_coord_offsetA,
1097                                               &jx0,&jy0,&jz0);
1098
1099             /* Calculate displacement vector */
1100             dx00             = _mm_sub_pd(ix0,jx0);
1101             dy00             = _mm_sub_pd(iy0,jy0);
1102             dz00             = _mm_sub_pd(iz0,jz0);
1103             dx10             = _mm_sub_pd(ix1,jx0);
1104             dy10             = _mm_sub_pd(iy1,jy0);
1105             dz10             = _mm_sub_pd(iz1,jz0);
1106             dx20             = _mm_sub_pd(ix2,jx0);
1107             dy20             = _mm_sub_pd(iy2,jy0);
1108             dz20             = _mm_sub_pd(iz2,jz0);
1109             dx30             = _mm_sub_pd(ix3,jx0);
1110             dy30             = _mm_sub_pd(iy3,jy0);
1111             dz30             = _mm_sub_pd(iz3,jz0);
1112
1113             /* Calculate squared distance and things based on it */
1114             rsq00            = gmx_mm_calc_rsq_pd(dx00,dy00,dz00);
1115             rsq10            = gmx_mm_calc_rsq_pd(dx10,dy10,dz10);
1116             rsq20            = gmx_mm_calc_rsq_pd(dx20,dy20,dz20);
1117             rsq30            = gmx_mm_calc_rsq_pd(dx30,dy30,dz30);
1118
1119             rinv00           = avx128fma_invsqrt_d(rsq00);
1120             rinv10           = avx128fma_invsqrt_d(rsq10);
1121             rinv20           = avx128fma_invsqrt_d(rsq20);
1122             rinv30           = avx128fma_invsqrt_d(rsq30);
1123
1124             rinvsq10         = _mm_mul_pd(rinv10,rinv10);
1125             rinvsq20         = _mm_mul_pd(rinv20,rinv20);
1126             rinvsq30         = _mm_mul_pd(rinv30,rinv30);
1127
1128             /* Load parameters for j particles */
1129             jq0              = _mm_load_sd(charge+jnrA+0);
1130             vdwjidx0A        = 2*vdwtype[jnrA+0];
1131
1132             fjx0             = _mm_setzero_pd();
1133             fjy0             = _mm_setzero_pd();
1134             fjz0             = _mm_setzero_pd();
1135
1136             /**************************
1137              * CALCULATE INTERACTIONS *
1138              **************************/
1139
1140             r00              = _mm_mul_pd(rsq00,rinv00);
1141
1142             /* Compute parameters for interactions between i and j atoms */
1143             gmx_mm_load_1pair_swizzle_pd(vdwparam+vdwioffset0+vdwjidx0A,&c6_00,&c12_00);
1144
1145             /* Calculate table index by multiplying r with table scale and truncate to integer */
1146             rt               = _mm_mul_pd(r00,vftabscale);
1147             vfitab           = _mm_cvttpd_epi32(rt);
1148 #ifdef __XOP__
1149             vfeps            = _mm_frcz_pd(rt);
1150 #else
1151             vfeps            = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
1152 #endif
1153             twovfeps         = _mm_add_pd(vfeps,vfeps);
1154             vfitab           = _mm_slli_epi32(vfitab,3);
1155
1156             /* CUBIC SPLINE TABLE DISPERSION */
1157             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
1158             F                = _mm_setzero_pd();
1159             GMX_MM_TRANSPOSE2_PD(Y,F);
1160             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
1161             H                = _mm_setzero_pd();
1162             GMX_MM_TRANSPOSE2_PD(G,H);
1163             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(H,vfeps,G),F);
1164             FF               = _mm_macc_pd(vfeps,_mm_macc_pd(twovfeps,H,G),Fp);
1165             fvdw6            = _mm_mul_pd(c6_00,FF);
1166
1167             /* CUBIC SPLINE TABLE REPULSION */
1168             vfitab           = _mm_add_epi32(vfitab,ifour);
1169             Y                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
1170             F                = _mm_setzero_pd();
1171             GMX_MM_TRANSPOSE2_PD(Y,F);
1172             G                = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
1173             H                = _mm_setzero_pd();
1174             GMX_MM_TRANSPOSE2_PD(G,H);
1175             Fp               = _mm_macc_pd(vfeps,_mm_macc_pd(H,vfeps,G),F);
1176             FF               = _mm_macc_pd(vfeps,_mm_macc_pd(twovfeps,H,G),Fp);
1177             fvdw12           = _mm_mul_pd(c12_00,FF);
1178             fvdw             = _mm_xor_pd(signbit,_mm_mul_pd(_mm_add_pd(fvdw6,fvdw12),_mm_mul_pd(vftabscale,rinv00)));
1179
1180             fscal            = fvdw;
1181
1182             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
1183
1184             /* Update vectorial force */
1185             fix0             = _mm_macc_pd(dx00,fscal,fix0);
1186             fiy0             = _mm_macc_pd(dy00,fscal,fiy0);
1187             fiz0             = _mm_macc_pd(dz00,fscal,fiz0);
1188
1189             fjx0             = _mm_macc_pd(dx00,fscal,fjx0);
1190             fjy0             = _mm_macc_pd(dy00,fscal,fjy0);
1191             fjz0             = _mm_macc_pd(dz00,fscal,fjz0);
1192
1193             /**************************
1194              * CALCULATE INTERACTIONS *
1195              **************************/
1196
1197             r10              = _mm_mul_pd(rsq10,rinv10);
1198
1199             /* Compute parameters for interactions between i and j atoms */
1200             qq10             = _mm_mul_pd(iq1,jq0);
1201
1202             /* EWALD ELECTROSTATICS */
1203
1204             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1205             ewrt             = _mm_mul_pd(r10,ewtabscale);
1206             ewitab           = _mm_cvttpd_epi32(ewrt);
1207 #ifdef __XOP__
1208             eweps            = _mm_frcz_pd(ewrt);
1209 #else
1210             eweps            = _mm_sub_pd(ewrt,_mm_round_pd(ewrt, _MM_FROUND_FLOOR));
1211 #endif
1212             twoeweps         = _mm_add_pd(eweps,eweps);
1213             gmx_mm_load_1pair_swizzle_pd(ewtab+_mm_extract_epi32(ewitab,0),&ewtabF,&ewtabFn);
1214             felec            = _mm_macc_pd(eweps,ewtabFn,_mm_mul_pd( _mm_sub_pd(one,eweps),ewtabF));
1215             felec            = _mm_mul_pd(_mm_mul_pd(qq10,rinv10),_mm_sub_pd(rinvsq10,felec));
1216
1217             fscal            = felec;
1218
1219             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
1220
1221             /* Update vectorial force */
1222             fix1             = _mm_macc_pd(dx10,fscal,fix1);
1223             fiy1             = _mm_macc_pd(dy10,fscal,fiy1);
1224             fiz1             = _mm_macc_pd(dz10,fscal,fiz1);
1225
1226             fjx0             = _mm_macc_pd(dx10,fscal,fjx0);
1227             fjy0             = _mm_macc_pd(dy10,fscal,fjy0);
1228             fjz0             = _mm_macc_pd(dz10,fscal,fjz0);
1229
1230             /**************************
1231              * CALCULATE INTERACTIONS *
1232              **************************/
1233
1234             r20              = _mm_mul_pd(rsq20,rinv20);
1235
1236             /* Compute parameters for interactions between i and j atoms */
1237             qq20             = _mm_mul_pd(iq2,jq0);
1238
1239             /* EWALD ELECTROSTATICS */
1240
1241             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1242             ewrt             = _mm_mul_pd(r20,ewtabscale);
1243             ewitab           = _mm_cvttpd_epi32(ewrt);
1244 #ifdef __XOP__
1245             eweps            = _mm_frcz_pd(ewrt);
1246 #else
1247             eweps            = _mm_sub_pd(ewrt,_mm_round_pd(ewrt, _MM_FROUND_FLOOR));
1248 #endif
1249             twoeweps         = _mm_add_pd(eweps,eweps);
1250             gmx_mm_load_1pair_swizzle_pd(ewtab+_mm_extract_epi32(ewitab,0),&ewtabF,&ewtabFn);
1251             felec            = _mm_macc_pd(eweps,ewtabFn,_mm_mul_pd( _mm_sub_pd(one,eweps),ewtabF));
1252             felec            = _mm_mul_pd(_mm_mul_pd(qq20,rinv20),_mm_sub_pd(rinvsq20,felec));
1253
1254             fscal            = felec;
1255
1256             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
1257
1258             /* Update vectorial force */
1259             fix2             = _mm_macc_pd(dx20,fscal,fix2);
1260             fiy2             = _mm_macc_pd(dy20,fscal,fiy2);
1261             fiz2             = _mm_macc_pd(dz20,fscal,fiz2);
1262
1263             fjx0             = _mm_macc_pd(dx20,fscal,fjx0);
1264             fjy0             = _mm_macc_pd(dy20,fscal,fjy0);
1265             fjz0             = _mm_macc_pd(dz20,fscal,fjz0);
1266
1267             /**************************
1268              * CALCULATE INTERACTIONS *
1269              **************************/
1270
1271             r30              = _mm_mul_pd(rsq30,rinv30);
1272
1273             /* Compute parameters for interactions between i and j atoms */
1274             qq30             = _mm_mul_pd(iq3,jq0);
1275
1276             /* EWALD ELECTROSTATICS */
1277
1278             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1279             ewrt             = _mm_mul_pd(r30,ewtabscale);
1280             ewitab           = _mm_cvttpd_epi32(ewrt);
1281 #ifdef __XOP__
1282             eweps            = _mm_frcz_pd(ewrt);
1283 #else
1284             eweps            = _mm_sub_pd(ewrt,_mm_round_pd(ewrt, _MM_FROUND_FLOOR));
1285 #endif
1286             twoeweps         = _mm_add_pd(eweps,eweps);
1287             gmx_mm_load_1pair_swizzle_pd(ewtab+_mm_extract_epi32(ewitab,0),&ewtabF,&ewtabFn);
1288             felec            = _mm_macc_pd(eweps,ewtabFn,_mm_mul_pd( _mm_sub_pd(one,eweps),ewtabF));
1289             felec            = _mm_mul_pd(_mm_mul_pd(qq30,rinv30),_mm_sub_pd(rinvsq30,felec));
1290
1291             fscal            = felec;
1292
1293             fscal            = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
1294
1295             /* Update vectorial force */
1296             fix3             = _mm_macc_pd(dx30,fscal,fix3);
1297             fiy3             = _mm_macc_pd(dy30,fscal,fiy3);
1298             fiz3             = _mm_macc_pd(dz30,fscal,fiz3);
1299
1300             fjx0             = _mm_macc_pd(dx30,fscal,fjx0);
1301             fjy0             = _mm_macc_pd(dy30,fscal,fjy0);
1302             fjz0             = _mm_macc_pd(dz30,fscal,fjz0);
1303
1304             gmx_mm_decrement_1rvec_1ptr_swizzle_pd(f+j_coord_offsetA,fjx0,fjy0,fjz0);
1305
1306             /* Inner loop uses 171 flops */
1307         }
1308
1309         /* End of innermost loop */
1310
1311         gmx_mm_update_iforce_4atom_swizzle_pd(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
1312                                               f+i_coord_offset,fshift+i_shift_offset);
1313
1314         /* Increment number of inner iterations */
1315         inneriter                  += j_index_end - j_index_start;
1316
1317         /* Outer loop uses 24 flops */
1318     }
1319
1320     /* Increment number of outer iterations */
1321     outeriter        += nri;
1322
1323     /* Update outer/inner flops */
1324
1325     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_F,outeriter*24 + inneriter*171);
1326 }