src/gromacs/gmxlib/nonbonded/nb_kernel_avx_256_double/nb_kernel_ElecEw_VdwLJEw_GeomW4P1_avx_256_double.c

   1 /*
   2  * This file is part of the GROMACS molecular simulation package.
   3  *
   4  * Copyright (c) 2012,2013,2014, 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_256_double kernel generator.
  37  */
  38 #include "config.h"
  39
  40 #include <math.h>
  41
  42 #include "../nb_kernel.h"
  43 #include "gromacs/legacyheaders/types/simple.h"
  44 #include "gromacs/math/vec.h"
  45 #include "gromacs/legacyheaders/nrnb.h"
  46
  47 #include "gromacs/simd/math_x86_avx_256_double.h"
  48 #include "kernelutil_x86_avx_256_double.h"
  49
  50 /*
  51  * Gromacs nonbonded kernel:   nb_kernel_ElecEw_VdwLJEw_GeomW4P1_VF_avx_256_double
  52  * Electrostatics interaction: Ewald
  53  * VdW interaction:            LJEwald
  54  * Geometry:                   Water4-Particle
  55  * Calculate force/pot:        PotentialAndForce
  56  */
  57 void
  58 nb_kernel_ElecEw_VdwLJEw_GeomW4P1_VF_avx_256_double
  59                     (t_nblist                    * gmx_restrict       nlist,
  60                      rvec                        * gmx_restrict          xx,
  61                      rvec                        * gmx_restrict          ff,
  62                      t_forcerec                  * gmx_restrict          fr,
  63                      t_mdatoms                   * gmx_restrict     mdatoms,
  64                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
  65                      t_nrnb                      * gmx_restrict        nrnb)
  66 {
  67     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
  68      * just 0 for non-waters.
  69      * Suffixes A,B,C,D refer to j loop unrolling done with AVX, e.g. for the four different
  70      * jnr indices corresponding to data put in the four positions in the SIMD register.
  71      */
  72     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
  73     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
  74     int              jnrA,jnrB,jnrC,jnrD;
  75     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
  76     int              jnrlistE,jnrlistF,jnrlistG,jnrlistH;
  77     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
  78     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  79     real             rcutoff_scalar;
  80     real             *shiftvec,*fshift,*x,*f;
  81     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD;
  82     real             scratch[4*DIM];
  83     __m256d          tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
  84     real *           vdwioffsetptr0;
  85     real *           vdwgridioffsetptr0;
  86     __m256d          ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  87     real *           vdwioffsetptr1;
  88     real *           vdwgridioffsetptr1;
  89     __m256d          ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
  90     real *           vdwioffsetptr2;
  91     real *           vdwgridioffsetptr2;
  92     __m256d          ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
  93     real *           vdwioffsetptr3;
  94     real *           vdwgridioffsetptr3;
  95     __m256d          ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
  96     int              vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D;
  97     __m256d          jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  98     __m256d          dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
  99     __m256d          dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
 100     __m256d          dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
 101     __m256d          dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30;
 102     __m256d          velec,felec,velecsum,facel,crf,krf,krf2;
 103     real             *charge;
 104     int              nvdwtype;
 105     __m256d          rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
 106     int              *vdwtype;
 107     real             *vdwparam;
 108     __m256d          one_sixth   = _mm256_set1_pd(1.0/6.0);
 109     __m256d          one_twelfth = _mm256_set1_pd(1.0/12.0);
 110     __m256d           c6grid_00;
 111     __m256d           c6grid_10;
 112     __m256d           c6grid_20;
 113     __m256d           c6grid_30;
 114     real             *vdwgridparam;
 115     __m256d           ewclj,ewclj2,ewclj6,ewcljrsq,poly,exponent,f6A,f6B,sh_lj_ewald;
 116     __m256d           one_half  = _mm256_set1_pd(0.5);
 117     __m256d           minus_one = _mm256_set1_pd(-1.0);
 118     __m128i          ewitab;
 119     __m256d          ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
 120     __m256d          beta,beta2,beta3,zeta2,pmecorrF,pmecorrV,rinv3;
 121     real             *ewtab;
 122     __m256d          dummy_mask,cutoff_mask;
 123     __m128           tmpmask0,tmpmask1;
 124     __m256d          signbit = _mm256_castsi256_pd( _mm256_set1_epi32(0x80000000) );
 125     __m256d          one     = _mm256_set1_pd(1.0);
 126     __m256d          two     = _mm256_set1_pd(2.0);
 127     x                = xx[0];
 128     f                = ff[0];
 129
 130     nri              = nlist->nri;
 131     iinr             = nlist->iinr;
 132     jindex           = nlist->jindex;
 133     jjnr             = nlist->jjnr;
 134     shiftidx         = nlist->shift;
 135     gid              = nlist->gid;
 136     shiftvec         = fr->shift_vec[0];
 137     fshift           = fr->fshift[0];
 138     facel            = _mm256_set1_pd(fr->epsfac);
 139     charge           = mdatoms->chargeA;
 140     nvdwtype         = fr->ntype;
 141     vdwparam         = fr->nbfp;
 142     vdwtype          = mdatoms->typeA;
 143     vdwgridparam     = fr->ljpme_c6grid;
 144     sh_lj_ewald      = _mm256_set1_pd(fr->ic->sh_lj_ewald);
 145     ewclj            = _mm256_set1_pd(fr->ewaldcoeff_lj);
 146     ewclj2           = _mm256_mul_pd(minus_one,_mm256_mul_pd(ewclj,ewclj));
 147
 148     sh_ewald         = _mm256_set1_pd(fr->ic->sh_ewald);
 149     beta             = _mm256_set1_pd(fr->ic->ewaldcoeff_q);
 150     beta2            = _mm256_mul_pd(beta,beta);
 151     beta3            = _mm256_mul_pd(beta,beta2);
 152
 153     ewtab            = fr->ic->tabq_coul_FDV0;
 154     ewtabscale       = _mm256_set1_pd(fr->ic->tabq_scale);
 155     ewtabhalfspace   = _mm256_set1_pd(0.5/fr->ic->tabq_scale);
 156
 157     /* Setup water-specific parameters */
 158     inr              = nlist->iinr[0];
 159     iq1              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+1]));
 160     iq2              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+2]));
 161     iq3              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+3]));
 162     vdwioffsetptr0   = vdwparam+2*nvdwtype*vdwtype[inr+0];
 163     vdwgridioffsetptr0 = vdwgridparam+2*nvdwtype*vdwtype[inr+0];
 164
 165     /* Avoid stupid compiler warnings */
 166     jnrA = jnrB = jnrC = jnrD = 0;
 167     j_coord_offsetA = 0;
 168     j_coord_offsetB = 0;
 169     j_coord_offsetC = 0;
 170     j_coord_offsetD = 0;
 171
 172     outeriter        = 0;
 173     inneriter        = 0;
 174
 175     for(iidx=0;iidx<4*DIM;iidx++)
 176     {
 177         scratch[iidx] = 0.0;
 178     }
 179
 180     /* Start outer loop over neighborlists */
 181     for(iidx=0; iidx<nri; iidx++)
 182     {
 183         /* Load shift vector for this list */
 184         i_shift_offset   = DIM*shiftidx[iidx];
 185
 186         /* Load limits for loop over neighbors */
 187         j_index_start    = jindex[iidx];
 188         j_index_end      = jindex[iidx+1];
 189
 190         /* Get outer coordinate index */
 191         inr              = iinr[iidx];
 192         i_coord_offset   = DIM*inr;
 193
 194         /* Load i particle coords and add shift vector */
 195         gmx_mm256_load_shift_and_4rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset,
 196                                                     &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
 197
 198         fix0             = _mm256_setzero_pd();
 199         fiy0             = _mm256_setzero_pd();
 200         fiz0             = _mm256_setzero_pd();
 201         fix1             = _mm256_setzero_pd();
 202         fiy1             = _mm256_setzero_pd();
 203         fiz1             = _mm256_setzero_pd();
 204         fix2             = _mm256_setzero_pd();
 205         fiy2             = _mm256_setzero_pd();
 206         fiz2             = _mm256_setzero_pd();
 207         fix3             = _mm256_setzero_pd();
 208         fiy3             = _mm256_setzero_pd();
 209         fiz3             = _mm256_setzero_pd();
 210
 211         /* Reset potential sums */
 212         velecsum         = _mm256_setzero_pd();
 213         vvdwsum          = _mm256_setzero_pd();
 214
 215         /* Start inner kernel loop */
 216         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4)
 217         {
 218
 219             /* Get j neighbor index, and coordinate index */
 220             jnrA             = jjnr[jidx];
 221             jnrB             = jjnr[jidx+1];
 222             jnrC             = jjnr[jidx+2];
 223             jnrD             = jjnr[jidx+3];
 224             j_coord_offsetA  = DIM*jnrA;
 225             j_coord_offsetB  = DIM*jnrB;
 226             j_coord_offsetC  = DIM*jnrC;
 227             j_coord_offsetD  = DIM*jnrD;
 228
 229             /* load j atom coordinates */
 230             gmx_mm256_load_1rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
 231                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 232                                                  &jx0,&jy0,&jz0);
 233
 234             /* Calculate displacement vector */
 235             dx00             = _mm256_sub_pd(ix0,jx0);
 236             dy00             = _mm256_sub_pd(iy0,jy0);
 237             dz00             = _mm256_sub_pd(iz0,jz0);
 238             dx10             = _mm256_sub_pd(ix1,jx0);
 239             dy10             = _mm256_sub_pd(iy1,jy0);
 240             dz10             = _mm256_sub_pd(iz1,jz0);
 241             dx20             = _mm256_sub_pd(ix2,jx0);
 242             dy20             = _mm256_sub_pd(iy2,jy0);
 243             dz20             = _mm256_sub_pd(iz2,jz0);
 244             dx30             = _mm256_sub_pd(ix3,jx0);
 245             dy30             = _mm256_sub_pd(iy3,jy0);
 246             dz30             = _mm256_sub_pd(iz3,jz0);
 247
 248             /* Calculate squared distance and things based on it */
 249             rsq00            = gmx_mm256_calc_rsq_pd(dx00,dy00,dz00);
 250             rsq10            = gmx_mm256_calc_rsq_pd(dx10,dy10,dz10);
 251             rsq20            = gmx_mm256_calc_rsq_pd(dx20,dy20,dz20);
 252             rsq30            = gmx_mm256_calc_rsq_pd(dx30,dy30,dz30);
 253
 254             rinv00           = gmx_mm256_invsqrt_pd(rsq00);
 255             rinv10           = gmx_mm256_invsqrt_pd(rsq10);
 256             rinv20           = gmx_mm256_invsqrt_pd(rsq20);
 257             rinv30           = gmx_mm256_invsqrt_pd(rsq30);
 258
 259             rinvsq00         = _mm256_mul_pd(rinv00,rinv00);
 260             rinvsq10         = _mm256_mul_pd(rinv10,rinv10);
 261             rinvsq20         = _mm256_mul_pd(rinv20,rinv20);
 262             rinvsq30         = _mm256_mul_pd(rinv30,rinv30);
 263
 264             /* Load parameters for j particles */
 265             jq0              = gmx_mm256_load_4real_swizzle_pd(charge+jnrA+0,charge+jnrB+0,
 266                                                                  charge+jnrC+0,charge+jnrD+0);
 267             vdwjidx0A        = 2*vdwtype[jnrA+0];
 268             vdwjidx0B        = 2*vdwtype[jnrB+0];
 269             vdwjidx0C        = 2*vdwtype[jnrC+0];
 270             vdwjidx0D        = 2*vdwtype[jnrD+0];
 271
 272             fjx0             = _mm256_setzero_pd();
 273             fjy0             = _mm256_setzero_pd();
 274             fjz0             = _mm256_setzero_pd();
 275
 276             /**************************
 277              * CALCULATE INTERACTIONS *
 278              **************************/
 279
 280             r00              = _mm256_mul_pd(rsq00,rinv00);
 281
 282             /* Compute parameters for interactions between i and j atoms */
 283             gmx_mm256_load_4pair_swizzle_pd(vdwioffsetptr0+vdwjidx0A,
 284                                             vdwioffsetptr0+vdwjidx0B,
 285                                             vdwioffsetptr0+vdwjidx0C,
 286                                             vdwioffsetptr0+vdwjidx0D,
 287                                             &c6_00,&c12_00);
 288
 289             c6grid_00       = gmx_mm256_load_4real_swizzle_pd(vdwgridioffsetptr0+vdwjidx0A,
 290                                                                   vdwgridioffsetptr0+vdwjidx0B,
 291                                                                   vdwgridioffsetptr0+vdwjidx0C,
 292                                                                   vdwgridioffsetptr0+vdwjidx0D);
 293
 294             /* Analytical LJ-PME */
 295             rinvsix          = _mm256_mul_pd(_mm256_mul_pd(rinvsq00,rinvsq00),rinvsq00);
 296             ewcljrsq         = _mm256_mul_pd(ewclj2,rsq00);
 297             ewclj6           = _mm256_mul_pd(ewclj2,_mm256_mul_pd(ewclj2,ewclj2));
 298             exponent         = gmx_simd_exp_d(ewcljrsq);
 299             /* poly = exp(-(beta*r)^2) * (1 + (beta*r)^2 + (beta*r)^4 /2) */
 300             poly             = _mm256_mul_pd(exponent,_mm256_add_pd(_mm256_sub_pd(one,ewcljrsq),_mm256_mul_pd(_mm256_mul_pd(ewcljrsq,ewcljrsq),one_half)));
 301             /* vvdw6 = [C6 - C6grid * (1-poly)]/r6 */
 302             vvdw6            = _mm256_mul_pd(_mm256_sub_pd(c6_00,_mm256_mul_pd(c6grid_00,_mm256_sub_pd(one,poly))),rinvsix);
 303             vvdw12           = _mm256_mul_pd(c12_00,_mm256_mul_pd(rinvsix,rinvsix));
 304             vvdw             = _mm256_sub_pd(_mm256_mul_pd(vvdw12,one_twelfth),_mm256_mul_pd(vvdw6,one_sixth));
 305             /* fvdw = vvdw12/r - (vvdw6/r + (C6grid * exponent * beta^6)/r) */
 306             fvdw             = _mm256_mul_pd(_mm256_sub_pd(vvdw12,_mm256_sub_pd(vvdw6,_mm256_mul_pd(_mm256_mul_pd(c6grid_00,one_sixth),_mm256_mul_pd(exponent,ewclj6)))),rinvsq00);
 307
 308             /* Update potential sum for this i atom from the interaction with this j atom. */
 309             vvdwsum          = _mm256_add_pd(vvdwsum,vvdw);
 310
 311             fscal            = fvdw;
 312
 313             /* Calculate temporary vectorial force */
 314             tx               = _mm256_mul_pd(fscal,dx00);
 315             ty               = _mm256_mul_pd(fscal,dy00);
 316             tz               = _mm256_mul_pd(fscal,dz00);
 317
 318             /* Update vectorial force */
 319             fix0             = _mm256_add_pd(fix0,tx);
 320             fiy0             = _mm256_add_pd(fiy0,ty);
 321             fiz0             = _mm256_add_pd(fiz0,tz);
 322
 323             fjx0             = _mm256_add_pd(fjx0,tx);
 324             fjy0             = _mm256_add_pd(fjy0,ty);
 325             fjz0             = _mm256_add_pd(fjz0,tz);
 326
 327             /**************************
 328              * CALCULATE INTERACTIONS *
 329              **************************/
 330
 331             r10              = _mm256_mul_pd(rsq10,rinv10);
 332
 333             /* Compute parameters for interactions between i and j atoms */
 334             qq10             = _mm256_mul_pd(iq1,jq0);
 335
 336             /* EWALD ELECTROSTATICS */
 337
 338             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 339             ewrt             = _mm256_mul_pd(r10,ewtabscale);
 340             ewitab           = _mm256_cvttpd_epi32(ewrt);
 341             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
 342             ewitab           = _mm_slli_epi32(ewitab,2);
 343             ewtabF           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
 344             ewtabD           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
 345             ewtabV           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,2) );
 346             ewtabFn          = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,3) );
 347             GMX_MM256_FULLTRANSPOSE4_PD(ewtabF,ewtabD,ewtabV,ewtabFn);
 348             felec            = _mm256_add_pd(ewtabF,_mm256_mul_pd(eweps,ewtabD));
 349             velec            = _mm256_sub_pd(ewtabV,_mm256_mul_pd(_mm256_mul_pd(ewtabhalfspace,eweps),_mm256_add_pd(ewtabF,felec)));
 350             velec            = _mm256_mul_pd(qq10,_mm256_sub_pd(rinv10,velec));
 351             felec            = _mm256_mul_pd(_mm256_mul_pd(qq10,rinv10),_mm256_sub_pd(rinvsq10,felec));
 352
 353             /* Update potential sum for this i atom from the interaction with this j atom. */
 354             velecsum         = _mm256_add_pd(velecsum,velec);
 355
 356             fscal            = felec;
 357
 358             /* Calculate temporary vectorial force */
 359             tx               = _mm256_mul_pd(fscal,dx10);
 360             ty               = _mm256_mul_pd(fscal,dy10);
 361             tz               = _mm256_mul_pd(fscal,dz10);
 362
 363             /* Update vectorial force */
 364             fix1             = _mm256_add_pd(fix1,tx);
 365             fiy1             = _mm256_add_pd(fiy1,ty);
 366             fiz1             = _mm256_add_pd(fiz1,tz);
 367
 368             fjx0             = _mm256_add_pd(fjx0,tx);
 369             fjy0             = _mm256_add_pd(fjy0,ty);
 370             fjz0             = _mm256_add_pd(fjz0,tz);
 371
 372             /**************************
 373              * CALCULATE INTERACTIONS *
 374              **************************/
 375
 376             r20              = _mm256_mul_pd(rsq20,rinv20);
 377
 378             /* Compute parameters for interactions between i and j atoms */
 379             qq20             = _mm256_mul_pd(iq2,jq0);
 380
 381             /* EWALD ELECTROSTATICS */
 382
 383             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 384             ewrt             = _mm256_mul_pd(r20,ewtabscale);
 385             ewitab           = _mm256_cvttpd_epi32(ewrt);
 386             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
 387             ewitab           = _mm_slli_epi32(ewitab,2);
 388             ewtabF           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
 389             ewtabD           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
 390             ewtabV           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,2) );
 391             ewtabFn          = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,3) );
 392             GMX_MM256_FULLTRANSPOSE4_PD(ewtabF,ewtabD,ewtabV,ewtabFn);
 393             felec            = _mm256_add_pd(ewtabF,_mm256_mul_pd(eweps,ewtabD));
 394             velec            = _mm256_sub_pd(ewtabV,_mm256_mul_pd(_mm256_mul_pd(ewtabhalfspace,eweps),_mm256_add_pd(ewtabF,felec)));
 395             velec            = _mm256_mul_pd(qq20,_mm256_sub_pd(rinv20,velec));
 396             felec            = _mm256_mul_pd(_mm256_mul_pd(qq20,rinv20),_mm256_sub_pd(rinvsq20,felec));
 397
 398             /* Update potential sum for this i atom from the interaction with this j atom. */
 399             velecsum         = _mm256_add_pd(velecsum,velec);
 400
 401             fscal            = felec;
 402
 403             /* Calculate temporary vectorial force */
 404             tx               = _mm256_mul_pd(fscal,dx20);
 405             ty               = _mm256_mul_pd(fscal,dy20);
 406             tz               = _mm256_mul_pd(fscal,dz20);
 407
 408             /* Update vectorial force */
 409             fix2             = _mm256_add_pd(fix2,tx);
 410             fiy2             = _mm256_add_pd(fiy2,ty);
 411             fiz2             = _mm256_add_pd(fiz2,tz);
 412
 413             fjx0             = _mm256_add_pd(fjx0,tx);
 414             fjy0             = _mm256_add_pd(fjy0,ty);
 415             fjz0             = _mm256_add_pd(fjz0,tz);
 416
 417             /**************************
 418              * CALCULATE INTERACTIONS *
 419              **************************/
 420
 421             r30              = _mm256_mul_pd(rsq30,rinv30);
 422
 423             /* Compute parameters for interactions between i and j atoms */
 424             qq30             = _mm256_mul_pd(iq3,jq0);
 425
 426             /* EWALD ELECTROSTATICS */
 427
 428             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 429             ewrt             = _mm256_mul_pd(r30,ewtabscale);
 430             ewitab           = _mm256_cvttpd_epi32(ewrt);
 431             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
 432             ewitab           = _mm_slli_epi32(ewitab,2);
 433             ewtabF           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
 434             ewtabD           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
 435             ewtabV           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,2) );
 436             ewtabFn          = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,3) );
 437             GMX_MM256_FULLTRANSPOSE4_PD(ewtabF,ewtabD,ewtabV,ewtabFn);
 438             felec            = _mm256_add_pd(ewtabF,_mm256_mul_pd(eweps,ewtabD));
 439             velec            = _mm256_sub_pd(ewtabV,_mm256_mul_pd(_mm256_mul_pd(ewtabhalfspace,eweps),_mm256_add_pd(ewtabF,felec)));
 440             velec            = _mm256_mul_pd(qq30,_mm256_sub_pd(rinv30,velec));
 441             felec            = _mm256_mul_pd(_mm256_mul_pd(qq30,rinv30),_mm256_sub_pd(rinvsq30,felec));
 442
 443             /* Update potential sum for this i atom from the interaction with this j atom. */
 444             velecsum         = _mm256_add_pd(velecsum,velec);
 445
 446             fscal            = felec;
 447
 448             /* Calculate temporary vectorial force */
 449             tx               = _mm256_mul_pd(fscal,dx30);
 450             ty               = _mm256_mul_pd(fscal,dy30);
 451             tz               = _mm256_mul_pd(fscal,dz30);
 452
 453             /* Update vectorial force */
 454             fix3             = _mm256_add_pd(fix3,tx);
 455             fiy3             = _mm256_add_pd(fiy3,ty);
 456             fiz3             = _mm256_add_pd(fiz3,tz);
 457
 458             fjx0             = _mm256_add_pd(fjx0,tx);
 459             fjy0             = _mm256_add_pd(fjy0,ty);
 460             fjz0             = _mm256_add_pd(fjz0,tz);
 461
 462             fjptrA             = f+j_coord_offsetA;
 463             fjptrB             = f+j_coord_offsetB;
 464             fjptrC             = f+j_coord_offsetC;
 465             fjptrD             = f+j_coord_offsetD;
 466
 467             gmx_mm256_decrement_1rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0);
 468
 469             /* Inner loop uses 180 flops */
 470         }
 471
 472         if(jidx<j_index_end)
 473         {
 474
 475             /* Get j neighbor index, and coordinate index */
 476             jnrlistA         = jjnr[jidx];
 477             jnrlistB         = jjnr[jidx+1];
 478             jnrlistC         = jjnr[jidx+2];
 479             jnrlistD         = jjnr[jidx+3];
 480             /* Sign of each element will be negative for non-real atoms.
 481              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
 482              * so use it as val = _mm_andnot_pd(mask,val) to clear dummy entries.
 483              */
 484             tmpmask0 = gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128()));
 485
 486             tmpmask1 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(3,3,2,2));
 487             tmpmask0 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(1,1,0,0));
 488             dummy_mask = _mm256_castps_pd(gmx_mm256_set_m128(tmpmask1,tmpmask0));
 489
 490             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
 491             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
 492             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
 493             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
 494             j_coord_offsetA  = DIM*jnrA;
 495             j_coord_offsetB  = DIM*jnrB;
 496             j_coord_offsetC  = DIM*jnrC;
 497             j_coord_offsetD  = DIM*jnrD;
 498
 499             /* load j atom coordinates */
 500             gmx_mm256_load_1rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
 501                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 502                                                  &jx0,&jy0,&jz0);
 503
 504             /* Calculate displacement vector */
 505             dx00             = _mm256_sub_pd(ix0,jx0);
 506             dy00             = _mm256_sub_pd(iy0,jy0);
 507             dz00             = _mm256_sub_pd(iz0,jz0);
 508             dx10             = _mm256_sub_pd(ix1,jx0);
 509             dy10             = _mm256_sub_pd(iy1,jy0);
 510             dz10             = _mm256_sub_pd(iz1,jz0);
 511             dx20             = _mm256_sub_pd(ix2,jx0);
 512             dy20             = _mm256_sub_pd(iy2,jy0);
 513             dz20             = _mm256_sub_pd(iz2,jz0);
 514             dx30             = _mm256_sub_pd(ix3,jx0);
 515             dy30             = _mm256_sub_pd(iy3,jy0);
 516             dz30             = _mm256_sub_pd(iz3,jz0);
 517
 518             /* Calculate squared distance and things based on it */
 519             rsq00            = gmx_mm256_calc_rsq_pd(dx00,dy00,dz00);
 520             rsq10            = gmx_mm256_calc_rsq_pd(dx10,dy10,dz10);
 521             rsq20            = gmx_mm256_calc_rsq_pd(dx20,dy20,dz20);
 522             rsq30            = gmx_mm256_calc_rsq_pd(dx30,dy30,dz30);
 523
 524             rinv00           = gmx_mm256_invsqrt_pd(rsq00);
 525             rinv10           = gmx_mm256_invsqrt_pd(rsq10);
 526             rinv20           = gmx_mm256_invsqrt_pd(rsq20);
 527             rinv30           = gmx_mm256_invsqrt_pd(rsq30);
 528
 529             rinvsq00         = _mm256_mul_pd(rinv00,rinv00);
 530             rinvsq10         = _mm256_mul_pd(rinv10,rinv10);
 531             rinvsq20         = _mm256_mul_pd(rinv20,rinv20);
 532             rinvsq30         = _mm256_mul_pd(rinv30,rinv30);
 533
 534             /* Load parameters for j particles */
 535             jq0              = gmx_mm256_load_4real_swizzle_pd(charge+jnrA+0,charge+jnrB+0,
 536                                                                  charge+jnrC+0,charge+jnrD+0);
 537             vdwjidx0A        = 2*vdwtype[jnrA+0];
 538             vdwjidx0B        = 2*vdwtype[jnrB+0];
 539             vdwjidx0C        = 2*vdwtype[jnrC+0];
 540             vdwjidx0D        = 2*vdwtype[jnrD+0];
 541
 542             fjx0             = _mm256_setzero_pd();
 543             fjy0             = _mm256_setzero_pd();
 544             fjz0             = _mm256_setzero_pd();
 545
 546             /**************************
 547              * CALCULATE INTERACTIONS *
 548              **************************/
 549
 550             r00              = _mm256_mul_pd(rsq00,rinv00);
 551             r00              = _mm256_andnot_pd(dummy_mask,r00);
 552
 553             /* Compute parameters for interactions between i and j atoms */
 554             gmx_mm256_load_4pair_swizzle_pd(vdwioffsetptr0+vdwjidx0A,
 555                                             vdwioffsetptr0+vdwjidx0B,
 556                                             vdwioffsetptr0+vdwjidx0C,
 557                                             vdwioffsetptr0+vdwjidx0D,
 558                                             &c6_00,&c12_00);
 559
 560             c6grid_00       = gmx_mm256_load_4real_swizzle_pd(vdwgridioffsetptr0+vdwjidx0A,
 561                                                                   vdwgridioffsetptr0+vdwjidx0B,
 562                                                                   vdwgridioffsetptr0+vdwjidx0C,
 563                                                                   vdwgridioffsetptr0+vdwjidx0D);
 564
 565             /* Analytical LJ-PME */
 566             rinvsix          = _mm256_mul_pd(_mm256_mul_pd(rinvsq00,rinvsq00),rinvsq00);
 567             ewcljrsq         = _mm256_mul_pd(ewclj2,rsq00);
 568             ewclj6           = _mm256_mul_pd(ewclj2,_mm256_mul_pd(ewclj2,ewclj2));
 569             exponent         = gmx_simd_exp_d(ewcljrsq);
 570             /* poly = exp(-(beta*r)^2) * (1 + (beta*r)^2 + (beta*r)^4 /2) */
 571             poly             = _mm256_mul_pd(exponent,_mm256_add_pd(_mm256_sub_pd(one,ewcljrsq),_mm256_mul_pd(_mm256_mul_pd(ewcljrsq,ewcljrsq),one_half)));
 572             /* vvdw6 = [C6 - C6grid * (1-poly)]/r6 */
 573             vvdw6            = _mm256_mul_pd(_mm256_sub_pd(c6_00,_mm256_mul_pd(c6grid_00,_mm256_sub_pd(one,poly))),rinvsix);
 574             vvdw12           = _mm256_mul_pd(c12_00,_mm256_mul_pd(rinvsix,rinvsix));
 575             vvdw             = _mm256_sub_pd(_mm256_mul_pd(vvdw12,one_twelfth),_mm256_mul_pd(vvdw6,one_sixth));
 576             /* fvdw = vvdw12/r - (vvdw6/r + (C6grid * exponent * beta^6)/r) */
 577             fvdw             = _mm256_mul_pd(_mm256_sub_pd(vvdw12,_mm256_sub_pd(vvdw6,_mm256_mul_pd(_mm256_mul_pd(c6grid_00,one_sixth),_mm256_mul_pd(exponent,ewclj6)))),rinvsq00);
 578
 579             /* Update potential sum for this i atom from the interaction with this j atom. */
 580             vvdw             = _mm256_andnot_pd(dummy_mask,vvdw);
 581             vvdwsum          = _mm256_add_pd(vvdwsum,vvdw);
 582
 583             fscal            = fvdw;
 584
 585             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
 586
 587             /* Calculate temporary vectorial force */
 588             tx               = _mm256_mul_pd(fscal,dx00);
 589             ty               = _mm256_mul_pd(fscal,dy00);
 590             tz               = _mm256_mul_pd(fscal,dz00);
 591
 592             /* Update vectorial force */
 593             fix0             = _mm256_add_pd(fix0,tx);
 594             fiy0             = _mm256_add_pd(fiy0,ty);
 595             fiz0             = _mm256_add_pd(fiz0,tz);
 596
 597             fjx0             = _mm256_add_pd(fjx0,tx);
 598             fjy0             = _mm256_add_pd(fjy0,ty);
 599             fjz0             = _mm256_add_pd(fjz0,tz);
 600
 601             /**************************
 602              * CALCULATE INTERACTIONS *
 603              **************************/
 604
 605             r10              = _mm256_mul_pd(rsq10,rinv10);
 606             r10              = _mm256_andnot_pd(dummy_mask,r10);
 607
 608             /* Compute parameters for interactions between i and j atoms */
 609             qq10             = _mm256_mul_pd(iq1,jq0);
 610
 611             /* EWALD ELECTROSTATICS */
 612
 613             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 614             ewrt             = _mm256_mul_pd(r10,ewtabscale);
 615             ewitab           = _mm256_cvttpd_epi32(ewrt);
 616             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
 617             ewitab           = _mm_slli_epi32(ewitab,2);
 618             ewtabF           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
 619             ewtabD           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
 620             ewtabV           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,2) );
 621             ewtabFn          = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,3) );
 622             GMX_MM256_FULLTRANSPOSE4_PD(ewtabF,ewtabD,ewtabV,ewtabFn);
 623             felec            = _mm256_add_pd(ewtabF,_mm256_mul_pd(eweps,ewtabD));
 624             velec            = _mm256_sub_pd(ewtabV,_mm256_mul_pd(_mm256_mul_pd(ewtabhalfspace,eweps),_mm256_add_pd(ewtabF,felec)));
 625             velec            = _mm256_mul_pd(qq10,_mm256_sub_pd(rinv10,velec));
 626             felec            = _mm256_mul_pd(_mm256_mul_pd(qq10,rinv10),_mm256_sub_pd(rinvsq10,felec));
 627
 628             /* Update potential sum for this i atom from the interaction with this j atom. */
 629             velec            = _mm256_andnot_pd(dummy_mask,velec);
 630             velecsum         = _mm256_add_pd(velecsum,velec);
 631
 632             fscal            = felec;
 633
 634             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
 635
 636             /* Calculate temporary vectorial force */
 637             tx               = _mm256_mul_pd(fscal,dx10);
 638             ty               = _mm256_mul_pd(fscal,dy10);
 639             tz               = _mm256_mul_pd(fscal,dz10);
 640
 641             /* Update vectorial force */
 642             fix1             = _mm256_add_pd(fix1,tx);
 643             fiy1             = _mm256_add_pd(fiy1,ty);
 644             fiz1             = _mm256_add_pd(fiz1,tz);
 645
 646             fjx0             = _mm256_add_pd(fjx0,tx);
 647             fjy0             = _mm256_add_pd(fjy0,ty);
 648             fjz0             = _mm256_add_pd(fjz0,tz);
 649
 650             /**************************
 651              * CALCULATE INTERACTIONS *
 652              **************************/
 653
 654             r20              = _mm256_mul_pd(rsq20,rinv20);
 655             r20              = _mm256_andnot_pd(dummy_mask,r20);
 656
 657             /* Compute parameters for interactions between i and j atoms */
 658             qq20             = _mm256_mul_pd(iq2,jq0);
 659
 660             /* EWALD ELECTROSTATICS */
 661
 662             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 663             ewrt             = _mm256_mul_pd(r20,ewtabscale);
 664             ewitab           = _mm256_cvttpd_epi32(ewrt);
 665             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
 666             ewitab           = _mm_slli_epi32(ewitab,2);
 667             ewtabF           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
 668             ewtabD           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
 669             ewtabV           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,2) );
 670             ewtabFn          = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,3) );
 671             GMX_MM256_FULLTRANSPOSE4_PD(ewtabF,ewtabD,ewtabV,ewtabFn);
 672             felec            = _mm256_add_pd(ewtabF,_mm256_mul_pd(eweps,ewtabD));
 673             velec            = _mm256_sub_pd(ewtabV,_mm256_mul_pd(_mm256_mul_pd(ewtabhalfspace,eweps),_mm256_add_pd(ewtabF,felec)));
 674             velec            = _mm256_mul_pd(qq20,_mm256_sub_pd(rinv20,velec));
 675             felec            = _mm256_mul_pd(_mm256_mul_pd(qq20,rinv20),_mm256_sub_pd(rinvsq20,felec));
 676
 677             /* Update potential sum for this i atom from the interaction with this j atom. */
 678             velec            = _mm256_andnot_pd(dummy_mask,velec);
 679             velecsum         = _mm256_add_pd(velecsum,velec);
 680
 681             fscal            = felec;
 682
 683             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
 684
 685             /* Calculate temporary vectorial force */
 686             tx               = _mm256_mul_pd(fscal,dx20);
 687             ty               = _mm256_mul_pd(fscal,dy20);
 688             tz               = _mm256_mul_pd(fscal,dz20);
 689
 690             /* Update vectorial force */
 691             fix2             = _mm256_add_pd(fix2,tx);
 692             fiy2             = _mm256_add_pd(fiy2,ty);
 693             fiz2             = _mm256_add_pd(fiz2,tz);
 694
 695             fjx0             = _mm256_add_pd(fjx0,tx);
 696             fjy0             = _mm256_add_pd(fjy0,ty);
 697             fjz0             = _mm256_add_pd(fjz0,tz);
 698
 699             /**************************
 700              * CALCULATE INTERACTIONS *
 701              **************************/
 702
 703             r30              = _mm256_mul_pd(rsq30,rinv30);
 704             r30              = _mm256_andnot_pd(dummy_mask,r30);
 705
 706             /* Compute parameters for interactions between i and j atoms */
 707             qq30             = _mm256_mul_pd(iq3,jq0);
 708
 709             /* EWALD ELECTROSTATICS */
 710
 711             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 712             ewrt             = _mm256_mul_pd(r30,ewtabscale);
 713             ewitab           = _mm256_cvttpd_epi32(ewrt);
 714             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
 715             ewitab           = _mm_slli_epi32(ewitab,2);
 716             ewtabF           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
 717             ewtabD           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
 718             ewtabV           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,2) );
 719             ewtabFn          = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,3) );
 720             GMX_MM256_FULLTRANSPOSE4_PD(ewtabF,ewtabD,ewtabV,ewtabFn);
 721             felec            = _mm256_add_pd(ewtabF,_mm256_mul_pd(eweps,ewtabD));
 722             velec            = _mm256_sub_pd(ewtabV,_mm256_mul_pd(_mm256_mul_pd(ewtabhalfspace,eweps),_mm256_add_pd(ewtabF,felec)));
 723             velec            = _mm256_mul_pd(qq30,_mm256_sub_pd(rinv30,velec));
 724             felec            = _mm256_mul_pd(_mm256_mul_pd(qq30,rinv30),_mm256_sub_pd(rinvsq30,felec));
 725
 726             /* Update potential sum for this i atom from the interaction with this j atom. */
 727             velec            = _mm256_andnot_pd(dummy_mask,velec);
 728             velecsum         = _mm256_add_pd(velecsum,velec);
 729
 730             fscal            = felec;
 731
 732             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
 733
 734             /* Calculate temporary vectorial force */
 735             tx               = _mm256_mul_pd(fscal,dx30);
 736             ty               = _mm256_mul_pd(fscal,dy30);
 737             tz               = _mm256_mul_pd(fscal,dz30);
 738
 739             /* Update vectorial force */
 740             fix3             = _mm256_add_pd(fix3,tx);
 741             fiy3             = _mm256_add_pd(fiy3,ty);
 742             fiz3             = _mm256_add_pd(fiz3,tz);
 743
 744             fjx0             = _mm256_add_pd(fjx0,tx);
 745             fjy0             = _mm256_add_pd(fjy0,ty);
 746             fjz0             = _mm256_add_pd(fjz0,tz);
 747
 748             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
 749             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
 750             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
 751             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
 752
 753             gmx_mm256_decrement_1rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0);
 754
 755             /* Inner loop uses 184 flops */
 756         }
 757
 758         /* End of innermost loop */
 759
 760         gmx_mm256_update_iforce_4atom_swizzle_pd(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
 761                                                  f+i_coord_offset,fshift+i_shift_offset);
 762
 763         ggid                        = gid[iidx];
 764         /* Update potential energies */
 765         gmx_mm256_update_1pot_pd(velecsum,kernel_data->energygrp_elec+ggid);
 766         gmx_mm256_update_1pot_pd(vvdwsum,kernel_data->energygrp_vdw+ggid);
 767
 768         /* Increment number of inner iterations */
 769         inneriter                  += j_index_end - j_index_start;
 770
 771         /* Outer loop uses 26 flops */
 772     }
 773
 774     /* Increment number of outer iterations */
 775     outeriter        += nri;
 776
 777     /* Update outer/inner flops */
 778
 779     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_VF,outeriter*26 + inneriter*184);
 780 }
 781 /*
 782  * Gromacs nonbonded kernel:   nb_kernel_ElecEw_VdwLJEw_GeomW4P1_F_avx_256_double
 783  * Electrostatics interaction: Ewald
 784  * VdW interaction:            LJEwald
 785  * Geometry:                   Water4-Particle
 786  * Calculate force/pot:        Force
 787  */
 788 void
 789 nb_kernel_ElecEw_VdwLJEw_GeomW4P1_F_avx_256_double
 790                     (t_nblist                    * gmx_restrict       nlist,
 791                      rvec                        * gmx_restrict          xx,
 792                      rvec                        * gmx_restrict          ff,
 793                      t_forcerec                  * gmx_restrict          fr,
 794                      t_mdatoms                   * gmx_restrict     mdatoms,
 795                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 796                      t_nrnb                      * gmx_restrict        nrnb)
 797 {
 798     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
 799      * just 0 for non-waters.
 800      * Suffixes A,B,C,D refer to j loop unrolling done with AVX, e.g. for the four different
 801      * jnr indices corresponding to data put in the four positions in the SIMD register.
 802      */
 803     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
 804     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
 805     int              jnrA,jnrB,jnrC,jnrD;
 806     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
 807     int              jnrlistE,jnrlistF,jnrlistG,jnrlistH;
 808     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
 809     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 810     real             rcutoff_scalar;
 811     real             *shiftvec,*fshift,*x,*f;
 812     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD;
 813     real             scratch[4*DIM];
 814     __m256d          tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
 815     real *           vdwioffsetptr0;
 816     real *           vdwgridioffsetptr0;
 817     __m256d          ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 818     real *           vdwioffsetptr1;
 819     real *           vdwgridioffsetptr1;
 820     __m256d          ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
 821     real *           vdwioffsetptr2;
 822     real *           vdwgridioffsetptr2;
 823     __m256d          ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
 824     real *           vdwioffsetptr3;
 825     real *           vdwgridioffsetptr3;
 826     __m256d          ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
 827     int              vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D;
 828     __m256d          jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 829     __m256d          dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
 830     __m256d          dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
 831     __m256d          dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
 832     __m256d          dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30;
 833     __m256d          velec,felec,velecsum,facel,crf,krf,krf2;
 834     real             *charge;
 835     int              nvdwtype;
 836     __m256d          rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
 837     int              *vdwtype;
 838     real             *vdwparam;
 839     __m256d          one_sixth   = _mm256_set1_pd(1.0/6.0);
 840     __m256d          one_twelfth = _mm256_set1_pd(1.0/12.0);
 841     __m256d           c6grid_00;
 842     __m256d           c6grid_10;
 843     __m256d           c6grid_20;
 844     __m256d           c6grid_30;
 845     real             *vdwgridparam;
 846     __m256d           ewclj,ewclj2,ewclj6,ewcljrsq,poly,exponent,f6A,f6B,sh_lj_ewald;
 847     __m256d           one_half  = _mm256_set1_pd(0.5);
 848     __m256d           minus_one = _mm256_set1_pd(-1.0);
 849     __m128i          ewitab;
 850     __m256d          ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
 851     __m256d          beta,beta2,beta3,zeta2,pmecorrF,pmecorrV,rinv3;
 852     real             *ewtab;
 853     __m256d          dummy_mask,cutoff_mask;
 854     __m128           tmpmask0,tmpmask1;
 855     __m256d          signbit = _mm256_castsi256_pd( _mm256_set1_epi32(0x80000000) );
 856     __m256d          one     = _mm256_set1_pd(1.0);
 857     __m256d          two     = _mm256_set1_pd(2.0);
 858     x                = xx[0];
 859     f                = ff[0];
 860
 861     nri              = nlist->nri;
 862     iinr             = nlist->iinr;
 863     jindex           = nlist->jindex;
 864     jjnr             = nlist->jjnr;
 865     shiftidx         = nlist->shift;
 866     gid              = nlist->gid;
 867     shiftvec         = fr->shift_vec[0];
 868     fshift           = fr->fshift[0];
 869     facel            = _mm256_set1_pd(fr->epsfac);
 870     charge           = mdatoms->chargeA;
 871     nvdwtype         = fr->ntype;
 872     vdwparam         = fr->nbfp;
 873     vdwtype          = mdatoms->typeA;
 874     vdwgridparam     = fr->ljpme_c6grid;
 875     sh_lj_ewald      = _mm256_set1_pd(fr->ic->sh_lj_ewald);
 876     ewclj            = _mm256_set1_pd(fr->ewaldcoeff_lj);
 877     ewclj2           = _mm256_mul_pd(minus_one,_mm256_mul_pd(ewclj,ewclj));
 878
 879     sh_ewald         = _mm256_set1_pd(fr->ic->sh_ewald);
 880     beta             = _mm256_set1_pd(fr->ic->ewaldcoeff_q);
 881     beta2            = _mm256_mul_pd(beta,beta);
 882     beta3            = _mm256_mul_pd(beta,beta2);
 883
 884     ewtab            = fr->ic->tabq_coul_F;
 885     ewtabscale       = _mm256_set1_pd(fr->ic->tabq_scale);
 886     ewtabhalfspace   = _mm256_set1_pd(0.5/fr->ic->tabq_scale);
 887
 888     /* Setup water-specific parameters */
 889     inr              = nlist->iinr[0];
 890     iq1              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+1]));
 891     iq2              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+2]));
 892     iq3              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+3]));
 893     vdwioffsetptr0   = vdwparam+2*nvdwtype*vdwtype[inr+0];
 894     vdwgridioffsetptr0 = vdwgridparam+2*nvdwtype*vdwtype[inr+0];
 895
 896     /* Avoid stupid compiler warnings */
 897     jnrA = jnrB = jnrC = jnrD = 0;
 898     j_coord_offsetA = 0;
 899     j_coord_offsetB = 0;
 900     j_coord_offsetC = 0;
 901     j_coord_offsetD = 0;
 902
 903     outeriter        = 0;
 904     inneriter        = 0;
 905
 906     for(iidx=0;iidx<4*DIM;iidx++)
 907     {
 908         scratch[iidx] = 0.0;
 909     }
 910
 911     /* Start outer loop over neighborlists */
 912     for(iidx=0; iidx<nri; iidx++)
 913     {
 914         /* Load shift vector for this list */
 915         i_shift_offset   = DIM*shiftidx[iidx];
 916
 917         /* Load limits for loop over neighbors */
 918         j_index_start    = jindex[iidx];
 919         j_index_end      = jindex[iidx+1];
 920
 921         /* Get outer coordinate index */
 922         inr              = iinr[iidx];
 923         i_coord_offset   = DIM*inr;
 924
 925         /* Load i particle coords and add shift vector */
 926         gmx_mm256_load_shift_and_4rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset,
 927                                                     &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
 928
 929         fix0             = _mm256_setzero_pd();
 930         fiy0             = _mm256_setzero_pd();
 931         fiz0             = _mm256_setzero_pd();
 932         fix1             = _mm256_setzero_pd();
 933         fiy1             = _mm256_setzero_pd();
 934         fiz1             = _mm256_setzero_pd();
 935         fix2             = _mm256_setzero_pd();
 936         fiy2             = _mm256_setzero_pd();
 937         fiz2             = _mm256_setzero_pd();
 938         fix3             = _mm256_setzero_pd();
 939         fiy3             = _mm256_setzero_pd();
 940         fiz3             = _mm256_setzero_pd();
 941
 942         /* Start inner kernel loop */
 943         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4)
 944         {
 945
 946             /* Get j neighbor index, and coordinate index */
 947             jnrA             = jjnr[jidx];
 948             jnrB             = jjnr[jidx+1];
 949             jnrC             = jjnr[jidx+2];
 950             jnrD             = jjnr[jidx+3];
 951             j_coord_offsetA  = DIM*jnrA;
 952             j_coord_offsetB  = DIM*jnrB;
 953             j_coord_offsetC  = DIM*jnrC;
 954             j_coord_offsetD  = DIM*jnrD;
 955
 956             /* load j atom coordinates */
 957             gmx_mm256_load_1rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
 958                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 959                                                  &jx0,&jy0,&jz0);
 960
 961             /* Calculate displacement vector */
 962             dx00             = _mm256_sub_pd(ix0,jx0);
 963             dy00             = _mm256_sub_pd(iy0,jy0);
 964             dz00             = _mm256_sub_pd(iz0,jz0);
 965             dx10             = _mm256_sub_pd(ix1,jx0);
 966             dy10             = _mm256_sub_pd(iy1,jy0);
 967             dz10             = _mm256_sub_pd(iz1,jz0);
 968             dx20             = _mm256_sub_pd(ix2,jx0);
 969             dy20             = _mm256_sub_pd(iy2,jy0);
 970             dz20             = _mm256_sub_pd(iz2,jz0);
 971             dx30             = _mm256_sub_pd(ix3,jx0);
 972             dy30             = _mm256_sub_pd(iy3,jy0);
 973             dz30             = _mm256_sub_pd(iz3,jz0);
 974
 975             /* Calculate squared distance and things based on it */
 976             rsq00            = gmx_mm256_calc_rsq_pd(dx00,dy00,dz00);
 977             rsq10            = gmx_mm256_calc_rsq_pd(dx10,dy10,dz10);
 978             rsq20            = gmx_mm256_calc_rsq_pd(dx20,dy20,dz20);
 979             rsq30            = gmx_mm256_calc_rsq_pd(dx30,dy30,dz30);
 980
 981             rinv00           = gmx_mm256_invsqrt_pd(rsq00);
 982             rinv10           = gmx_mm256_invsqrt_pd(rsq10);
 983             rinv20           = gmx_mm256_invsqrt_pd(rsq20);
 984             rinv30           = gmx_mm256_invsqrt_pd(rsq30);
 985
 986             rinvsq00         = _mm256_mul_pd(rinv00,rinv00);
 987             rinvsq10         = _mm256_mul_pd(rinv10,rinv10);
 988             rinvsq20         = _mm256_mul_pd(rinv20,rinv20);
 989             rinvsq30         = _mm256_mul_pd(rinv30,rinv30);
 990
 991             /* Load parameters for j particles */
 992             jq0              = gmx_mm256_load_4real_swizzle_pd(charge+jnrA+0,charge+jnrB+0,
 993                                                                  charge+jnrC+0,charge+jnrD+0);
 994             vdwjidx0A        = 2*vdwtype[jnrA+0];
 995             vdwjidx0B        = 2*vdwtype[jnrB+0];
 996             vdwjidx0C        = 2*vdwtype[jnrC+0];
 997             vdwjidx0D        = 2*vdwtype[jnrD+0];
 998
 999             fjx0             = _mm256_setzero_pd();
1000             fjy0             = _mm256_setzero_pd();
1001             fjz0             = _mm256_setzero_pd();
1002
1003             /**************************
1004              * CALCULATE INTERACTIONS *
1005              **************************/
1006
1007             r00              = _mm256_mul_pd(rsq00,rinv00);
1008
1009             /* Compute parameters for interactions between i and j atoms */
1010             gmx_mm256_load_4pair_swizzle_pd(vdwioffsetptr0+vdwjidx0A,
1011                                             vdwioffsetptr0+vdwjidx0B,
1012                                             vdwioffsetptr0+vdwjidx0C,
1013                                             vdwioffsetptr0+vdwjidx0D,
1014                                             &c6_00,&c12_00);
1015
1016             c6grid_00       = gmx_mm256_load_4real_swizzle_pd(vdwgridioffsetptr0+vdwjidx0A,
1017                                                                   vdwgridioffsetptr0+vdwjidx0B,
1018                                                                   vdwgridioffsetptr0+vdwjidx0C,
1019                                                                   vdwgridioffsetptr0+vdwjidx0D);
1020
1021             /* Analytical LJ-PME */
1022             rinvsix          = _mm256_mul_pd(_mm256_mul_pd(rinvsq00,rinvsq00),rinvsq00);
1023             ewcljrsq         = _mm256_mul_pd(ewclj2,rsq00);
1024             ewclj6           = _mm256_mul_pd(ewclj2,_mm256_mul_pd(ewclj2,ewclj2));
1025             exponent         = gmx_simd_exp_d(ewcljrsq);
1026             /* poly = exp(-(beta*r)^2) * (1 + (beta*r)^2 + (beta*r)^4 /2) */
1027             poly             = _mm256_mul_pd(exponent,_mm256_add_pd(_mm256_sub_pd(one,ewcljrsq),_mm256_mul_pd(_mm256_mul_pd(ewcljrsq,ewcljrsq),one_half)));
1028             /* f6A = 6 * C6grid * (1 - poly) */
1029             f6A              = _mm256_mul_pd(c6grid_00,_mm256_sub_pd(one,poly));
1030             /* f6B = C6grid * exponent * beta^6 */
1031             f6B              = _mm256_mul_pd(_mm256_mul_pd(c6grid_00,one_sixth),_mm256_mul_pd(exponent,ewclj6));
1032             /* fvdw = 12*C12/r13 - ((6*C6 - f6A)/r6 + f6B)/r */
1033             fvdw              = _mm256_mul_pd(_mm256_add_pd(_mm256_mul_pd(_mm256_sub_pd(_mm256_mul_pd(c12_00,rinvsix),_mm256_sub_pd(c6_00,f6A)),rinvsix),f6B),rinvsq00);
1034
1035             fscal            = fvdw;
1036
1037             /* Calculate temporary vectorial force */
1038             tx               = _mm256_mul_pd(fscal,dx00);
1039             ty               = _mm256_mul_pd(fscal,dy00);
1040             tz               = _mm256_mul_pd(fscal,dz00);
1041
1042             /* Update vectorial force */
1043             fix0             = _mm256_add_pd(fix0,tx);
1044             fiy0             = _mm256_add_pd(fiy0,ty);
1045             fiz0             = _mm256_add_pd(fiz0,tz);
1046
1047             fjx0             = _mm256_add_pd(fjx0,tx);
1048             fjy0             = _mm256_add_pd(fjy0,ty);
1049             fjz0             = _mm256_add_pd(fjz0,tz);
1050
1051             /**************************
1052              * CALCULATE INTERACTIONS *
1053              **************************/
1054
1055             r10              = _mm256_mul_pd(rsq10,rinv10);
1056
1057             /* Compute parameters for interactions between i and j atoms */
1058             qq10             = _mm256_mul_pd(iq1,jq0);
1059
1060             /* EWALD ELECTROSTATICS */
1061
1062             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1063             ewrt             = _mm256_mul_pd(r10,ewtabscale);
1064             ewitab           = _mm256_cvttpd_epi32(ewrt);
1065             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
1066             gmx_mm256_load_4pair_swizzle_pd(ewtab + _mm_extract_epi32(ewitab,0),ewtab + _mm_extract_epi32(ewitab,1),
1067                                             ewtab + _mm_extract_epi32(ewitab,2),ewtab + _mm_extract_epi32(ewitab,3),
1068                                             &ewtabF,&ewtabFn);
1069             felec            = _mm256_add_pd(_mm256_mul_pd( _mm256_sub_pd(one,eweps),ewtabF),_mm256_mul_pd(eweps,ewtabFn));
1070             felec            = _mm256_mul_pd(_mm256_mul_pd(qq10,rinv10),_mm256_sub_pd(rinvsq10,felec));
1071
1072             fscal            = felec;
1073
1074             /* Calculate temporary vectorial force */
1075             tx               = _mm256_mul_pd(fscal,dx10);
1076             ty               = _mm256_mul_pd(fscal,dy10);
1077             tz               = _mm256_mul_pd(fscal,dz10);
1078
1079             /* Update vectorial force */
1080             fix1             = _mm256_add_pd(fix1,tx);
1081             fiy1             = _mm256_add_pd(fiy1,ty);
1082             fiz1             = _mm256_add_pd(fiz1,tz);
1083
1084             fjx0             = _mm256_add_pd(fjx0,tx);
1085             fjy0             = _mm256_add_pd(fjy0,ty);
1086             fjz0             = _mm256_add_pd(fjz0,tz);
1087
1088             /**************************
1089              * CALCULATE INTERACTIONS *
1090              **************************/
1091
1092             r20              = _mm256_mul_pd(rsq20,rinv20);
1093
1094             /* Compute parameters for interactions between i and j atoms */
1095             qq20             = _mm256_mul_pd(iq2,jq0);
1096
1097             /* EWALD ELECTROSTATICS */
1098
1099             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1100             ewrt             = _mm256_mul_pd(r20,ewtabscale);
1101             ewitab           = _mm256_cvttpd_epi32(ewrt);
1102             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
1103             gmx_mm256_load_4pair_swizzle_pd(ewtab + _mm_extract_epi32(ewitab,0),ewtab + _mm_extract_epi32(ewitab,1),
1104                                             ewtab + _mm_extract_epi32(ewitab,2),ewtab + _mm_extract_epi32(ewitab,3),
1105                                             &ewtabF,&ewtabFn);
1106             felec            = _mm256_add_pd(_mm256_mul_pd( _mm256_sub_pd(one,eweps),ewtabF),_mm256_mul_pd(eweps,ewtabFn));
1107             felec            = _mm256_mul_pd(_mm256_mul_pd(qq20,rinv20),_mm256_sub_pd(rinvsq20,felec));
1108
1109             fscal            = felec;
1110
1111             /* Calculate temporary vectorial force */
1112             tx               = _mm256_mul_pd(fscal,dx20);
1113             ty               = _mm256_mul_pd(fscal,dy20);
1114             tz               = _mm256_mul_pd(fscal,dz20);
1115
1116             /* Update vectorial force */
1117             fix2             = _mm256_add_pd(fix2,tx);
1118             fiy2             = _mm256_add_pd(fiy2,ty);
1119             fiz2             = _mm256_add_pd(fiz2,tz);
1120
1121             fjx0             = _mm256_add_pd(fjx0,tx);
1122             fjy0             = _mm256_add_pd(fjy0,ty);
1123             fjz0             = _mm256_add_pd(fjz0,tz);
1124
1125             /**************************
1126              * CALCULATE INTERACTIONS *
1127              **************************/
1128
1129             r30              = _mm256_mul_pd(rsq30,rinv30);
1130
1131             /* Compute parameters for interactions between i and j atoms */
1132             qq30             = _mm256_mul_pd(iq3,jq0);
1133
1134             /* EWALD ELECTROSTATICS */
1135
1136             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1137             ewrt             = _mm256_mul_pd(r30,ewtabscale);
1138             ewitab           = _mm256_cvttpd_epi32(ewrt);
1139             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
1140             gmx_mm256_load_4pair_swizzle_pd(ewtab + _mm_extract_epi32(ewitab,0),ewtab + _mm_extract_epi32(ewitab,1),
1141                                             ewtab + _mm_extract_epi32(ewitab,2),ewtab + _mm_extract_epi32(ewitab,3),
1142                                             &ewtabF,&ewtabFn);
1143             felec            = _mm256_add_pd(_mm256_mul_pd( _mm256_sub_pd(one,eweps),ewtabF),_mm256_mul_pd(eweps,ewtabFn));
1144             felec            = _mm256_mul_pd(_mm256_mul_pd(qq30,rinv30),_mm256_sub_pd(rinvsq30,felec));
1145
1146             fscal            = felec;
1147
1148             /* Calculate temporary vectorial force */
1149             tx               = _mm256_mul_pd(fscal,dx30);
1150             ty               = _mm256_mul_pd(fscal,dy30);
1151             tz               = _mm256_mul_pd(fscal,dz30);
1152
1153             /* Update vectorial force */
1154             fix3             = _mm256_add_pd(fix3,tx);
1155             fiy3             = _mm256_add_pd(fiy3,ty);
1156             fiz3             = _mm256_add_pd(fiz3,tz);
1157
1158             fjx0             = _mm256_add_pd(fjx0,tx);
1159             fjy0             = _mm256_add_pd(fjy0,ty);
1160             fjz0             = _mm256_add_pd(fjz0,tz);
1161
1162             fjptrA             = f+j_coord_offsetA;
1163             fjptrB             = f+j_coord_offsetB;
1164             fjptrC             = f+j_coord_offsetC;
1165             fjptrD             = f+j_coord_offsetD;
1166
1167             gmx_mm256_decrement_1rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0);
1168
1169             /* Inner loop uses 157 flops */
1170         }
1171
1172         if(jidx<j_index_end)
1173         {
1174
1175             /* Get j neighbor index, and coordinate index */
1176             jnrlistA         = jjnr[jidx];
1177             jnrlistB         = jjnr[jidx+1];
1178             jnrlistC         = jjnr[jidx+2];
1179             jnrlistD         = jjnr[jidx+3];
1180             /* Sign of each element will be negative for non-real atoms.
1181              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
1182              * so use it as val = _mm_andnot_pd(mask,val) to clear dummy entries.
1183              */
1184             tmpmask0 = gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128()));
1185
1186             tmpmask1 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(3,3,2,2));
1187             tmpmask0 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(1,1,0,0));
1188             dummy_mask = _mm256_castps_pd(gmx_mm256_set_m128(tmpmask1,tmpmask0));
1189
1190             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
1191             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
1192             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
1193             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
1194             j_coord_offsetA  = DIM*jnrA;
1195             j_coord_offsetB  = DIM*jnrB;
1196             j_coord_offsetC  = DIM*jnrC;
1197             j_coord_offsetD  = DIM*jnrD;
1198
1199             /* load j atom coordinates */
1200             gmx_mm256_load_1rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
1201                                                  x+j_coord_offsetC,x+j_coord_offsetD,
1202                                                  &jx0,&jy0,&jz0);
1203
1204             /* Calculate displacement vector */
1205             dx00             = _mm256_sub_pd(ix0,jx0);
1206             dy00             = _mm256_sub_pd(iy0,jy0);
1207             dz00             = _mm256_sub_pd(iz0,jz0);
1208             dx10             = _mm256_sub_pd(ix1,jx0);
1209             dy10             = _mm256_sub_pd(iy1,jy0);
1210             dz10             = _mm256_sub_pd(iz1,jz0);
1211             dx20             = _mm256_sub_pd(ix2,jx0);
1212             dy20             = _mm256_sub_pd(iy2,jy0);
1213             dz20             = _mm256_sub_pd(iz2,jz0);
1214             dx30             = _mm256_sub_pd(ix3,jx0);
1215             dy30             = _mm256_sub_pd(iy3,jy0);
1216             dz30             = _mm256_sub_pd(iz3,jz0);
1217
1218             /* Calculate squared distance and things based on it */
1219             rsq00            = gmx_mm256_calc_rsq_pd(dx00,dy00,dz00);
1220             rsq10            = gmx_mm256_calc_rsq_pd(dx10,dy10,dz10);
1221             rsq20            = gmx_mm256_calc_rsq_pd(dx20,dy20,dz20);
1222             rsq30            = gmx_mm256_calc_rsq_pd(dx30,dy30,dz30);
1223
1224             rinv00           = gmx_mm256_invsqrt_pd(rsq00);
1225             rinv10           = gmx_mm256_invsqrt_pd(rsq10);
1226             rinv20           = gmx_mm256_invsqrt_pd(rsq20);
1227             rinv30           = gmx_mm256_invsqrt_pd(rsq30);
1228
1229             rinvsq00         = _mm256_mul_pd(rinv00,rinv00);
1230             rinvsq10         = _mm256_mul_pd(rinv10,rinv10);
1231             rinvsq20         = _mm256_mul_pd(rinv20,rinv20);
1232             rinvsq30         = _mm256_mul_pd(rinv30,rinv30);
1233
1234             /* Load parameters for j particles */
1235             jq0              = gmx_mm256_load_4real_swizzle_pd(charge+jnrA+0,charge+jnrB+0,
1236                                                                  charge+jnrC+0,charge+jnrD+0);
1237             vdwjidx0A        = 2*vdwtype[jnrA+0];
1238             vdwjidx0B        = 2*vdwtype[jnrB+0];
1239             vdwjidx0C        = 2*vdwtype[jnrC+0];
1240             vdwjidx0D        = 2*vdwtype[jnrD+0];
1241
1242             fjx0             = _mm256_setzero_pd();
1243             fjy0             = _mm256_setzero_pd();
1244             fjz0             = _mm256_setzero_pd();
1245
1246             /**************************
1247              * CALCULATE INTERACTIONS *
1248              **************************/
1249
1250             r00              = _mm256_mul_pd(rsq00,rinv00);
1251             r00              = _mm256_andnot_pd(dummy_mask,r00);
1252
1253             /* Compute parameters for interactions between i and j atoms */
1254             gmx_mm256_load_4pair_swizzle_pd(vdwioffsetptr0+vdwjidx0A,
1255                                             vdwioffsetptr0+vdwjidx0B,
1256                                             vdwioffsetptr0+vdwjidx0C,
1257                                             vdwioffsetptr0+vdwjidx0D,
1258                                             &c6_00,&c12_00);
1259
1260             c6grid_00       = gmx_mm256_load_4real_swizzle_pd(vdwgridioffsetptr0+vdwjidx0A,
1261                                                                   vdwgridioffsetptr0+vdwjidx0B,
1262                                                                   vdwgridioffsetptr0+vdwjidx0C,
1263                                                                   vdwgridioffsetptr0+vdwjidx0D);
1264
1265             /* Analytical LJ-PME */
1266             rinvsix          = _mm256_mul_pd(_mm256_mul_pd(rinvsq00,rinvsq00),rinvsq00);
1267             ewcljrsq         = _mm256_mul_pd(ewclj2,rsq00);
1268             ewclj6           = _mm256_mul_pd(ewclj2,_mm256_mul_pd(ewclj2,ewclj2));
1269             exponent         = gmx_simd_exp_d(ewcljrsq);
1270             /* poly = exp(-(beta*r)^2) * (1 + (beta*r)^2 + (beta*r)^4 /2) */
1271             poly             = _mm256_mul_pd(exponent,_mm256_add_pd(_mm256_sub_pd(one,ewcljrsq),_mm256_mul_pd(_mm256_mul_pd(ewcljrsq,ewcljrsq),one_half)));
1272             /* f6A = 6 * C6grid * (1 - poly) */
1273             f6A              = _mm256_mul_pd(c6grid_00,_mm256_sub_pd(one,poly));
1274             /* f6B = C6grid * exponent * beta^6 */
1275             f6B              = _mm256_mul_pd(_mm256_mul_pd(c6grid_00,one_sixth),_mm256_mul_pd(exponent,ewclj6));
1276             /* fvdw = 12*C12/r13 - ((6*C6 - f6A)/r6 + f6B)/r */
1277             fvdw              = _mm256_mul_pd(_mm256_add_pd(_mm256_mul_pd(_mm256_sub_pd(_mm256_mul_pd(c12_00,rinvsix),_mm256_sub_pd(c6_00,f6A)),rinvsix),f6B),rinvsq00);
1278
1279             fscal            = fvdw;
1280
1281             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
1282
1283             /* Calculate temporary vectorial force */
1284             tx               = _mm256_mul_pd(fscal,dx00);
1285             ty               = _mm256_mul_pd(fscal,dy00);
1286             tz               = _mm256_mul_pd(fscal,dz00);
1287
1288             /* Update vectorial force */
1289             fix0             = _mm256_add_pd(fix0,tx);
1290             fiy0             = _mm256_add_pd(fiy0,ty);
1291             fiz0             = _mm256_add_pd(fiz0,tz);
1292
1293             fjx0             = _mm256_add_pd(fjx0,tx);
1294             fjy0             = _mm256_add_pd(fjy0,ty);
1295             fjz0             = _mm256_add_pd(fjz0,tz);
1296
1297             /**************************
1298              * CALCULATE INTERACTIONS *
1299              **************************/
1300
1301             r10              = _mm256_mul_pd(rsq10,rinv10);
1302             r10              = _mm256_andnot_pd(dummy_mask,r10);
1303
1304             /* Compute parameters for interactions between i and j atoms */
1305             qq10             = _mm256_mul_pd(iq1,jq0);
1306
1307             /* EWALD ELECTROSTATICS */
1308
1309             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1310             ewrt             = _mm256_mul_pd(r10,ewtabscale);
1311             ewitab           = _mm256_cvttpd_epi32(ewrt);
1312             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
1313             gmx_mm256_load_4pair_swizzle_pd(ewtab + _mm_extract_epi32(ewitab,0),ewtab + _mm_extract_epi32(ewitab,1),
1314                                             ewtab + _mm_extract_epi32(ewitab,2),ewtab + _mm_extract_epi32(ewitab,3),
1315                                             &ewtabF,&ewtabFn);
1316             felec            = _mm256_add_pd(_mm256_mul_pd( _mm256_sub_pd(one,eweps),ewtabF),_mm256_mul_pd(eweps,ewtabFn));
1317             felec            = _mm256_mul_pd(_mm256_mul_pd(qq10,rinv10),_mm256_sub_pd(rinvsq10,felec));
1318
1319             fscal            = felec;
1320
1321             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
1322
1323             /* Calculate temporary vectorial force */
1324             tx               = _mm256_mul_pd(fscal,dx10);
1325             ty               = _mm256_mul_pd(fscal,dy10);
1326             tz               = _mm256_mul_pd(fscal,dz10);
1327
1328             /* Update vectorial force */
1329             fix1             = _mm256_add_pd(fix1,tx);
1330             fiy1             = _mm256_add_pd(fiy1,ty);
1331             fiz1             = _mm256_add_pd(fiz1,tz);
1332
1333             fjx0             = _mm256_add_pd(fjx0,tx);
1334             fjy0             = _mm256_add_pd(fjy0,ty);
1335             fjz0             = _mm256_add_pd(fjz0,tz);
1336
1337             /**************************
1338              * CALCULATE INTERACTIONS *
1339              **************************/
1340
1341             r20              = _mm256_mul_pd(rsq20,rinv20);
1342             r20              = _mm256_andnot_pd(dummy_mask,r20);
1343
1344             /* Compute parameters for interactions between i and j atoms */
1345             qq20             = _mm256_mul_pd(iq2,jq0);
1346
1347             /* EWALD ELECTROSTATICS */
1348
1349             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1350             ewrt             = _mm256_mul_pd(r20,ewtabscale);
1351             ewitab           = _mm256_cvttpd_epi32(ewrt);
1352             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
1353             gmx_mm256_load_4pair_swizzle_pd(ewtab + _mm_extract_epi32(ewitab,0),ewtab + _mm_extract_epi32(ewitab,1),
1354                                             ewtab + _mm_extract_epi32(ewitab,2),ewtab + _mm_extract_epi32(ewitab,3),
1355                                             &ewtabF,&ewtabFn);
1356             felec            = _mm256_add_pd(_mm256_mul_pd( _mm256_sub_pd(one,eweps),ewtabF),_mm256_mul_pd(eweps,ewtabFn));
1357             felec            = _mm256_mul_pd(_mm256_mul_pd(qq20,rinv20),_mm256_sub_pd(rinvsq20,felec));
1358
1359             fscal            = felec;
1360
1361             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
1362
1363             /* Calculate temporary vectorial force */
1364             tx               = _mm256_mul_pd(fscal,dx20);
1365             ty               = _mm256_mul_pd(fscal,dy20);
1366             tz               = _mm256_mul_pd(fscal,dz20);
1367
1368             /* Update vectorial force */
1369             fix2             = _mm256_add_pd(fix2,tx);
1370             fiy2             = _mm256_add_pd(fiy2,ty);
1371             fiz2             = _mm256_add_pd(fiz2,tz);
1372
1373             fjx0             = _mm256_add_pd(fjx0,tx);
1374             fjy0             = _mm256_add_pd(fjy0,ty);
1375             fjz0             = _mm256_add_pd(fjz0,tz);
1376
1377             /**************************
1378              * CALCULATE INTERACTIONS *
1379              **************************/
1380
1381             r30              = _mm256_mul_pd(rsq30,rinv30);
1382             r30              = _mm256_andnot_pd(dummy_mask,r30);
1383
1384             /* Compute parameters for interactions between i and j atoms */
1385             qq30             = _mm256_mul_pd(iq3,jq0);
1386
1387             /* EWALD ELECTROSTATICS */
1388
1389             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1390             ewrt             = _mm256_mul_pd(r30,ewtabscale);
1391             ewitab           = _mm256_cvttpd_epi32(ewrt);
1392             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
1393             gmx_mm256_load_4pair_swizzle_pd(ewtab + _mm_extract_epi32(ewitab,0),ewtab + _mm_extract_epi32(ewitab,1),
1394                                             ewtab + _mm_extract_epi32(ewitab,2),ewtab + _mm_extract_epi32(ewitab,3),
1395                                             &ewtabF,&ewtabFn);
1396             felec            = _mm256_add_pd(_mm256_mul_pd( _mm256_sub_pd(one,eweps),ewtabF),_mm256_mul_pd(eweps,ewtabFn));
1397             felec            = _mm256_mul_pd(_mm256_mul_pd(qq30,rinv30),_mm256_sub_pd(rinvsq30,felec));
1398
1399             fscal            = felec;
1400
1401             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
1402
1403             /* Calculate temporary vectorial force */
1404             tx               = _mm256_mul_pd(fscal,dx30);
1405             ty               = _mm256_mul_pd(fscal,dy30);
1406             tz               = _mm256_mul_pd(fscal,dz30);
1407
1408             /* Update vectorial force */
1409             fix3             = _mm256_add_pd(fix3,tx);
1410             fiy3             = _mm256_add_pd(fiy3,ty);
1411             fiz3             = _mm256_add_pd(fiz3,tz);
1412
1413             fjx0             = _mm256_add_pd(fjx0,tx);
1414             fjy0             = _mm256_add_pd(fjy0,ty);
1415             fjz0             = _mm256_add_pd(fjz0,tz);
1416
1417             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
1418             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
1419             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
1420             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
1421
1422             gmx_mm256_decrement_1rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0);
1423
1424             /* Inner loop uses 161 flops */
1425         }
1426
1427         /* End of innermost loop */
1428
1429         gmx_mm256_update_iforce_4atom_swizzle_pd(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
1430                                                  f+i_coord_offset,fshift+i_shift_offset);
1431
1432         /* Increment number of inner iterations */
1433         inneriter                  += j_index_end - j_index_start;
1434
1435         /* Outer loop uses 24 flops */
1436     }
1437
1438     /* Increment number of outer iterations */
1439     outeriter        += nri;
1440
1441     /* Update outer/inner flops */
1442
1443     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_F,outeriter*24 + inneriter*161);
1444 }