src/gromacs/gmxlib/nonbonded/nb_kernel_avx_256_double/nb_kernel_ElecEwSh_VdwLJSh_GeomW3W3_avx_256_double.c

   1 /*
   2  * Note: this file was generated by the Gromacs avx_256_double kernel generator.
   3  *
   4  *                This source code is part of
   5  *
   6  *                 G   R   O   M   A   C   S
   7  *
   8  * Copyright (c) 2001-2012, The GROMACS Development Team
   9  *
  10  * Gromacs is a library for molecular simulation and trajectory analysis,
  11  * written by Erik Lindahl, David van der Spoel, Berk Hess, and others - for
  12  * a full list of developers and information, check out http://www.gromacs.org
  13  *
  14  * This program is free software; you can redistribute it and/or modify it under
  15  * the terms of the GNU Lesser General Public License as published by the Free
  16  * Software Foundation; either version 2 of the License, or (at your option) any
  17  * later version.
  18  *
  19  * To help fund GROMACS development, we humbly ask that you cite
  20  * the papers people have written on it - you can find them on the website.
  21  */
  22 #ifdef HAVE_CONFIG_H
  23 #include <config.h>
  24 #endif
  25
  26 #include <math.h>
  27
  28 #include "../nb_kernel.h"
  29 #include "types/simple.h"
  30 #include "vec.h"
  31 #include "nrnb.h"
  32
  33 #include "gmx_math_x86_avx_256_double.h"
  34 #include "kernelutil_x86_avx_256_double.h"
  35
  36 /*
  37  * Gromacs nonbonded kernel:   nb_kernel_ElecEwSh_VdwLJSh_GeomW3W3_VF_avx_256_double
  38  * Electrostatics interaction: Ewald
  39  * VdW interaction:            LennardJones
  40  * Geometry:                   Water3-Water3
  41  * Calculate force/pot:        PotentialAndForce
  42  */
  43 void
  44 nb_kernel_ElecEwSh_VdwLJSh_GeomW3W3_VF_avx_256_double
  45                     (t_nblist * gmx_restrict                nlist,
  46                      rvec * gmx_restrict                    xx,
  47                      rvec * gmx_restrict                    ff,
  48                      t_forcerec * gmx_restrict              fr,
  49                      t_mdatoms * gmx_restrict               mdatoms,
  50                      nb_kernel_data_t * gmx_restrict        kernel_data,
  51                      t_nrnb * gmx_restrict                  nrnb)
  52 {
  53     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
  54      * just 0 for non-waters.
  55      * Suffixes A,B,C,D refer to j loop unrolling done with AVX, e.g. for the four different
  56      * jnr indices corresponding to data put in the four positions in the SIMD register.
  57      */
  58     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
  59     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
  60     int              jnrA,jnrB,jnrC,jnrD;
  61     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
  62     int              jnrlistE,jnrlistF,jnrlistG,jnrlistH;
  63     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
  64     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  65     real             rcutoff_scalar;
  66     real             *shiftvec,*fshift,*x,*f;
  67     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD;
  68     real             scratch[4*DIM];
  69     __m256d          tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
  70     real *           vdwioffsetptr0;
  71     __m256d          ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  72     real *           vdwioffsetptr1;
  73     __m256d          ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
  74     real *           vdwioffsetptr2;
  75     __m256d          ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
  76     int              vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D;
  77     __m256d          jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  78     int              vdwjidx1A,vdwjidx1B,vdwjidx1C,vdwjidx1D;
  79     __m256d          jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
  80     int              vdwjidx2A,vdwjidx2B,vdwjidx2C,vdwjidx2D;
  81     __m256d          jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
  82     __m256d          dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
  83     __m256d          dx01,dy01,dz01,rsq01,rinv01,rinvsq01,r01,qq01,c6_01,c12_01;
  84     __m256d          dx02,dy02,dz02,rsq02,rinv02,rinvsq02,r02,qq02,c6_02,c12_02;
  85     __m256d          dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
  86     __m256d          dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11;
  87     __m256d          dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12;
  88     __m256d          dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
  89     __m256d          dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21;
  90     __m256d          dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22;
  91     __m256d          velec,felec,velecsum,facel,crf,krf,krf2;
  92     real             *charge;
  93     int              nvdwtype;
  94     __m256d          rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
  95     int              *vdwtype;
  96     real             *vdwparam;
  97     __m256d          one_sixth   = _mm256_set1_pd(1.0/6.0);
  98     __m256d          one_twelfth = _mm256_set1_pd(1.0/12.0);
  99     __m128i          ewitab;
 100     __m256d          ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
 101     __m256d          beta,beta2,beta3,zeta2,pmecorrF,pmecorrV,rinv3;
 102     real             *ewtab;
 103     __m256d          dummy_mask,cutoff_mask;
 104     __m128           tmpmask0,tmpmask1;
 105     __m256d          signbit = _mm256_castsi256_pd( _mm256_set1_epi32(0x80000000) );
 106     __m256d          one     = _mm256_set1_pd(1.0);
 107     __m256d          two     = _mm256_set1_pd(2.0);
 108     x                = xx[0];
 109     f                = ff[0];
 110
 111     nri              = nlist->nri;
 112     iinr             = nlist->iinr;
 113     jindex           = nlist->jindex;
 114     jjnr             = nlist->jjnr;
 115     shiftidx         = nlist->shift;
 116     gid              = nlist->gid;
 117     shiftvec         = fr->shift_vec[0];
 118     fshift           = fr->fshift[0];
 119     facel            = _mm256_set1_pd(fr->epsfac);
 120     charge           = mdatoms->chargeA;
 121     nvdwtype         = fr->ntype;
 122     vdwparam         = fr->nbfp;
 123     vdwtype          = mdatoms->typeA;
 124
 125     sh_ewald         = _mm256_set1_pd(fr->ic->sh_ewald);
 126     beta             = _mm256_set1_pd(fr->ic->ewaldcoeff);
 127     beta2            = _mm256_mul_pd(beta,beta);
 128     beta3            = _mm256_mul_pd(beta,beta2);
 129
 130     ewtab            = fr->ic->tabq_coul_FDV0;
 131     ewtabscale       = _mm256_set1_pd(fr->ic->tabq_scale);
 132     ewtabhalfspace   = _mm256_set1_pd(0.5/fr->ic->tabq_scale);
 133
 134     /* Setup water-specific parameters */
 135     inr              = nlist->iinr[0];
 136     iq0              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+0]));
 137     iq1              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+1]));
 138     iq2              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+2]));
 139     vdwioffsetptr0   = vdwparam+2*nvdwtype*vdwtype[inr+0];
 140
 141     jq0              = _mm256_set1_pd(charge[inr+0]);
 142     jq1              = _mm256_set1_pd(charge[inr+1]);
 143     jq2              = _mm256_set1_pd(charge[inr+2]);
 144     vdwjidx0A        = 2*vdwtype[inr+0];
 145     qq00             = _mm256_mul_pd(iq0,jq0);
 146     c6_00            = _mm256_set1_pd(vdwioffsetptr0[vdwjidx0A]);
 147     c12_00           = _mm256_set1_pd(vdwioffsetptr0[vdwjidx0A+1]);
 148     qq01             = _mm256_mul_pd(iq0,jq1);
 149     qq02             = _mm256_mul_pd(iq0,jq2);
 150     qq10             = _mm256_mul_pd(iq1,jq0);
 151     qq11             = _mm256_mul_pd(iq1,jq1);
 152     qq12             = _mm256_mul_pd(iq1,jq2);
 153     qq20             = _mm256_mul_pd(iq2,jq0);
 154     qq21             = _mm256_mul_pd(iq2,jq1);
 155     qq22             = _mm256_mul_pd(iq2,jq2);
 156
 157     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 158     rcutoff_scalar   = fr->rcoulomb;
 159     rcutoff          = _mm256_set1_pd(rcutoff_scalar);
 160     rcutoff2         = _mm256_mul_pd(rcutoff,rcutoff);
 161
 162     sh_vdw_invrcut6  = _mm256_set1_pd(fr->ic->sh_invrc6);
 163     rvdw             = _mm256_set1_pd(fr->rvdw);
 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_3rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset,
 196                                                     &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2);
 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
 208         /* Reset potential sums */
 209         velecsum         = _mm256_setzero_pd();
 210         vvdwsum          = _mm256_setzero_pd();
 211
 212         /* Start inner kernel loop */
 213         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4)
 214         {
 215
 216             /* Get j neighbor index, and coordinate index */
 217             jnrA             = jjnr[jidx];
 218             jnrB             = jjnr[jidx+1];
 219             jnrC             = jjnr[jidx+2];
 220             jnrD             = jjnr[jidx+3];
 221             j_coord_offsetA  = DIM*jnrA;
 222             j_coord_offsetB  = DIM*jnrB;
 223             j_coord_offsetC  = DIM*jnrC;
 224             j_coord_offsetD  = DIM*jnrD;
 225
 226             /* load j atom coordinates */
 227             gmx_mm256_load_3rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
 228                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 229                                               &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,&jy2,&jz2);
 230
 231             /* Calculate displacement vector */
 232             dx00             = _mm256_sub_pd(ix0,jx0);
 233             dy00             = _mm256_sub_pd(iy0,jy0);
 234             dz00             = _mm256_sub_pd(iz0,jz0);
 235             dx01             = _mm256_sub_pd(ix0,jx1);
 236             dy01             = _mm256_sub_pd(iy0,jy1);
 237             dz01             = _mm256_sub_pd(iz0,jz1);
 238             dx02             = _mm256_sub_pd(ix0,jx2);
 239             dy02             = _mm256_sub_pd(iy0,jy2);
 240             dz02             = _mm256_sub_pd(iz0,jz2);
 241             dx10             = _mm256_sub_pd(ix1,jx0);
 242             dy10             = _mm256_sub_pd(iy1,jy0);
 243             dz10             = _mm256_sub_pd(iz1,jz0);
 244             dx11             = _mm256_sub_pd(ix1,jx1);
 245             dy11             = _mm256_sub_pd(iy1,jy1);
 246             dz11             = _mm256_sub_pd(iz1,jz1);
 247             dx12             = _mm256_sub_pd(ix1,jx2);
 248             dy12             = _mm256_sub_pd(iy1,jy2);
 249             dz12             = _mm256_sub_pd(iz1,jz2);
 250             dx20             = _mm256_sub_pd(ix2,jx0);
 251             dy20             = _mm256_sub_pd(iy2,jy0);
 252             dz20             = _mm256_sub_pd(iz2,jz0);
 253             dx21             = _mm256_sub_pd(ix2,jx1);
 254             dy21             = _mm256_sub_pd(iy2,jy1);
 255             dz21             = _mm256_sub_pd(iz2,jz1);
 256             dx22             = _mm256_sub_pd(ix2,jx2);
 257             dy22             = _mm256_sub_pd(iy2,jy2);
 258             dz22             = _mm256_sub_pd(iz2,jz2);
 259
 260             /* Calculate squared distance and things based on it */
 261             rsq00            = gmx_mm256_calc_rsq_pd(dx00,dy00,dz00);
 262             rsq01            = gmx_mm256_calc_rsq_pd(dx01,dy01,dz01);
 263             rsq02            = gmx_mm256_calc_rsq_pd(dx02,dy02,dz02);
 264             rsq10            = gmx_mm256_calc_rsq_pd(dx10,dy10,dz10);
 265             rsq11            = gmx_mm256_calc_rsq_pd(dx11,dy11,dz11);
 266             rsq12            = gmx_mm256_calc_rsq_pd(dx12,dy12,dz12);
 267             rsq20            = gmx_mm256_calc_rsq_pd(dx20,dy20,dz20);
 268             rsq21            = gmx_mm256_calc_rsq_pd(dx21,dy21,dz21);
 269             rsq22            = gmx_mm256_calc_rsq_pd(dx22,dy22,dz22);
 270
 271             rinv00           = gmx_mm256_invsqrt_pd(rsq00);
 272             rinv01           = gmx_mm256_invsqrt_pd(rsq01);
 273             rinv02           = gmx_mm256_invsqrt_pd(rsq02);
 274             rinv10           = gmx_mm256_invsqrt_pd(rsq10);
 275             rinv11           = gmx_mm256_invsqrt_pd(rsq11);
 276             rinv12           = gmx_mm256_invsqrt_pd(rsq12);
 277             rinv20           = gmx_mm256_invsqrt_pd(rsq20);
 278             rinv21           = gmx_mm256_invsqrt_pd(rsq21);
 279             rinv22           = gmx_mm256_invsqrt_pd(rsq22);
 280
 281             rinvsq00         = _mm256_mul_pd(rinv00,rinv00);
 282             rinvsq01         = _mm256_mul_pd(rinv01,rinv01);
 283             rinvsq02         = _mm256_mul_pd(rinv02,rinv02);
 284             rinvsq10         = _mm256_mul_pd(rinv10,rinv10);
 285             rinvsq11         = _mm256_mul_pd(rinv11,rinv11);
 286             rinvsq12         = _mm256_mul_pd(rinv12,rinv12);
 287             rinvsq20         = _mm256_mul_pd(rinv20,rinv20);
 288             rinvsq21         = _mm256_mul_pd(rinv21,rinv21);
 289             rinvsq22         = _mm256_mul_pd(rinv22,rinv22);
 290
 291             fjx0             = _mm256_setzero_pd();
 292             fjy0             = _mm256_setzero_pd();
 293             fjz0             = _mm256_setzero_pd();
 294             fjx1             = _mm256_setzero_pd();
 295             fjy1             = _mm256_setzero_pd();
 296             fjz1             = _mm256_setzero_pd();
 297             fjx2             = _mm256_setzero_pd();
 298             fjy2             = _mm256_setzero_pd();
 299             fjz2             = _mm256_setzero_pd();
 300
 301             /**************************
 302              * CALCULATE INTERACTIONS *
 303              **************************/
 304
 305             if (gmx_mm256_any_lt(rsq00,rcutoff2))
 306             {
 307
 308             r00              = _mm256_mul_pd(rsq00,rinv00);
 309
 310             /* EWALD ELECTROSTATICS */
 311
 312             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 313             ewrt             = _mm256_mul_pd(r00,ewtabscale);
 314             ewitab           = _mm256_cvttpd_epi32(ewrt);
 315             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
 316             ewitab           = _mm_slli_epi32(ewitab,2);
 317             ewtabF           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
 318             ewtabD           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
 319             ewtabV           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,2) );
 320             ewtabFn          = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,3) );
 321             GMX_MM256_FULLTRANSPOSE4_PD(ewtabF,ewtabD,ewtabV,ewtabFn);
 322             felec            = _mm256_add_pd(ewtabF,_mm256_mul_pd(eweps,ewtabD));
 323             velec            = _mm256_sub_pd(ewtabV,_mm256_mul_pd(_mm256_mul_pd(ewtabhalfspace,eweps),_mm256_add_pd(ewtabF,felec)));
 324             velec            = _mm256_mul_pd(qq00,_mm256_sub_pd(_mm256_sub_pd(rinv00,sh_ewald),velec));
 325             felec            = _mm256_mul_pd(_mm256_mul_pd(qq00,rinv00),_mm256_sub_pd(rinvsq00,felec));
 326
 327             /* LENNARD-JONES DISPERSION/REPULSION */
 328
 329             rinvsix          = _mm256_mul_pd(_mm256_mul_pd(rinvsq00,rinvsq00),rinvsq00);
 330             vvdw6            = _mm256_mul_pd(c6_00,rinvsix);
 331             vvdw12           = _mm256_mul_pd(c12_00,_mm256_mul_pd(rinvsix,rinvsix));
 332             vvdw             = _mm256_sub_pd(_mm256_mul_pd( _mm256_sub_pd(vvdw12 , _mm256_mul_pd(c12_00,_mm256_mul_pd(sh_vdw_invrcut6,sh_vdw_invrcut6))), one_twelfth) ,
 333                                           _mm256_mul_pd( _mm256_sub_pd(vvdw6,_mm256_mul_pd(c6_00,sh_vdw_invrcut6)),one_sixth));
 334             fvdw             = _mm256_mul_pd(_mm256_sub_pd(vvdw12,vvdw6),rinvsq00);
 335
 336             cutoff_mask      = _mm256_cmp_pd(rsq00,rcutoff2,_CMP_LT_OQ);
 337
 338             /* Update potential sum for this i atom from the interaction with this j atom. */
 339             velec            = _mm256_and_pd(velec,cutoff_mask);
 340             velecsum         = _mm256_add_pd(velecsum,velec);
 341             vvdw             = _mm256_and_pd(vvdw,cutoff_mask);
 342             vvdwsum          = _mm256_add_pd(vvdwsum,vvdw);
 343
 344             fscal            = _mm256_add_pd(felec,fvdw);
 345
 346             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 347
 348             /* Calculate temporary vectorial force */
 349             tx               = _mm256_mul_pd(fscal,dx00);
 350             ty               = _mm256_mul_pd(fscal,dy00);
 351             tz               = _mm256_mul_pd(fscal,dz00);
 352
 353             /* Update vectorial force */
 354             fix0             = _mm256_add_pd(fix0,tx);
 355             fiy0             = _mm256_add_pd(fiy0,ty);
 356             fiz0             = _mm256_add_pd(fiz0,tz);
 357
 358             fjx0             = _mm256_add_pd(fjx0,tx);
 359             fjy0             = _mm256_add_pd(fjy0,ty);
 360             fjz0             = _mm256_add_pd(fjz0,tz);
 361
 362             }
 363
 364             /**************************
 365              * CALCULATE INTERACTIONS *
 366              **************************/
 367
 368             if (gmx_mm256_any_lt(rsq01,rcutoff2))
 369             {
 370
 371             r01              = _mm256_mul_pd(rsq01,rinv01);
 372
 373             /* EWALD ELECTROSTATICS */
 374
 375             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 376             ewrt             = _mm256_mul_pd(r01,ewtabscale);
 377             ewitab           = _mm256_cvttpd_epi32(ewrt);
 378             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
 379             ewitab           = _mm_slli_epi32(ewitab,2);
 380             ewtabF           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
 381             ewtabD           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
 382             ewtabV           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,2) );
 383             ewtabFn          = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,3) );
 384             GMX_MM256_FULLTRANSPOSE4_PD(ewtabF,ewtabD,ewtabV,ewtabFn);
 385             felec            = _mm256_add_pd(ewtabF,_mm256_mul_pd(eweps,ewtabD));
 386             velec            = _mm256_sub_pd(ewtabV,_mm256_mul_pd(_mm256_mul_pd(ewtabhalfspace,eweps),_mm256_add_pd(ewtabF,felec)));
 387             velec            = _mm256_mul_pd(qq01,_mm256_sub_pd(_mm256_sub_pd(rinv01,sh_ewald),velec));
 388             felec            = _mm256_mul_pd(_mm256_mul_pd(qq01,rinv01),_mm256_sub_pd(rinvsq01,felec));
 389
 390             cutoff_mask      = _mm256_cmp_pd(rsq01,rcutoff2,_CMP_LT_OQ);
 391
 392             /* Update potential sum for this i atom from the interaction with this j atom. */
 393             velec            = _mm256_and_pd(velec,cutoff_mask);
 394             velecsum         = _mm256_add_pd(velecsum,velec);
 395
 396             fscal            = felec;
 397
 398             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 399
 400             /* Calculate temporary vectorial force */
 401             tx               = _mm256_mul_pd(fscal,dx01);
 402             ty               = _mm256_mul_pd(fscal,dy01);
 403             tz               = _mm256_mul_pd(fscal,dz01);
 404
 405             /* Update vectorial force */
 406             fix0             = _mm256_add_pd(fix0,tx);
 407             fiy0             = _mm256_add_pd(fiy0,ty);
 408             fiz0             = _mm256_add_pd(fiz0,tz);
 409
 410             fjx1             = _mm256_add_pd(fjx1,tx);
 411             fjy1             = _mm256_add_pd(fjy1,ty);
 412             fjz1             = _mm256_add_pd(fjz1,tz);
 413
 414             }
 415
 416             /**************************
 417              * CALCULATE INTERACTIONS *
 418              **************************/
 419
 420             if (gmx_mm256_any_lt(rsq02,rcutoff2))
 421             {
 422
 423             r02              = _mm256_mul_pd(rsq02,rinv02);
 424
 425             /* EWALD ELECTROSTATICS */
 426
 427             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 428             ewrt             = _mm256_mul_pd(r02,ewtabscale);
 429             ewitab           = _mm256_cvttpd_epi32(ewrt);
 430             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
 431             ewitab           = _mm_slli_epi32(ewitab,2);
 432             ewtabF           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
 433             ewtabD           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
 434             ewtabV           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,2) );
 435             ewtabFn          = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,3) );
 436             GMX_MM256_FULLTRANSPOSE4_PD(ewtabF,ewtabD,ewtabV,ewtabFn);
 437             felec            = _mm256_add_pd(ewtabF,_mm256_mul_pd(eweps,ewtabD));
 438             velec            = _mm256_sub_pd(ewtabV,_mm256_mul_pd(_mm256_mul_pd(ewtabhalfspace,eweps),_mm256_add_pd(ewtabF,felec)));
 439             velec            = _mm256_mul_pd(qq02,_mm256_sub_pd(_mm256_sub_pd(rinv02,sh_ewald),velec));
 440             felec            = _mm256_mul_pd(_mm256_mul_pd(qq02,rinv02),_mm256_sub_pd(rinvsq02,felec));
 441
 442             cutoff_mask      = _mm256_cmp_pd(rsq02,rcutoff2,_CMP_LT_OQ);
 443
 444             /* Update potential sum for this i atom from the interaction with this j atom. */
 445             velec            = _mm256_and_pd(velec,cutoff_mask);
 446             velecsum         = _mm256_add_pd(velecsum,velec);
 447
 448             fscal            = felec;
 449
 450             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 451
 452             /* Calculate temporary vectorial force */
 453             tx               = _mm256_mul_pd(fscal,dx02);
 454             ty               = _mm256_mul_pd(fscal,dy02);
 455             tz               = _mm256_mul_pd(fscal,dz02);
 456
 457             /* Update vectorial force */
 458             fix0             = _mm256_add_pd(fix0,tx);
 459             fiy0             = _mm256_add_pd(fiy0,ty);
 460             fiz0             = _mm256_add_pd(fiz0,tz);
 461
 462             fjx2             = _mm256_add_pd(fjx2,tx);
 463             fjy2             = _mm256_add_pd(fjy2,ty);
 464             fjz2             = _mm256_add_pd(fjz2,tz);
 465
 466             }
 467
 468             /**************************
 469              * CALCULATE INTERACTIONS *
 470              **************************/
 471
 472             if (gmx_mm256_any_lt(rsq10,rcutoff2))
 473             {
 474
 475             r10              = _mm256_mul_pd(rsq10,rinv10);
 476
 477             /* EWALD ELECTROSTATICS */
 478
 479             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 480             ewrt             = _mm256_mul_pd(r10,ewtabscale);
 481             ewitab           = _mm256_cvttpd_epi32(ewrt);
 482             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
 483             ewitab           = _mm_slli_epi32(ewitab,2);
 484             ewtabF           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
 485             ewtabD           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
 486             ewtabV           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,2) );
 487             ewtabFn          = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,3) );
 488             GMX_MM256_FULLTRANSPOSE4_PD(ewtabF,ewtabD,ewtabV,ewtabFn);
 489             felec            = _mm256_add_pd(ewtabF,_mm256_mul_pd(eweps,ewtabD));
 490             velec            = _mm256_sub_pd(ewtabV,_mm256_mul_pd(_mm256_mul_pd(ewtabhalfspace,eweps),_mm256_add_pd(ewtabF,felec)));
 491             velec            = _mm256_mul_pd(qq10,_mm256_sub_pd(_mm256_sub_pd(rinv10,sh_ewald),velec));
 492             felec            = _mm256_mul_pd(_mm256_mul_pd(qq10,rinv10),_mm256_sub_pd(rinvsq10,felec));
 493
 494             cutoff_mask      = _mm256_cmp_pd(rsq10,rcutoff2,_CMP_LT_OQ);
 495
 496             /* Update potential sum for this i atom from the interaction with this j atom. */
 497             velec            = _mm256_and_pd(velec,cutoff_mask);
 498             velecsum         = _mm256_add_pd(velecsum,velec);
 499
 500             fscal            = felec;
 501
 502             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 503
 504             /* Calculate temporary vectorial force */
 505             tx               = _mm256_mul_pd(fscal,dx10);
 506             ty               = _mm256_mul_pd(fscal,dy10);
 507             tz               = _mm256_mul_pd(fscal,dz10);
 508
 509             /* Update vectorial force */
 510             fix1             = _mm256_add_pd(fix1,tx);
 511             fiy1             = _mm256_add_pd(fiy1,ty);
 512             fiz1             = _mm256_add_pd(fiz1,tz);
 513
 514             fjx0             = _mm256_add_pd(fjx0,tx);
 515             fjy0             = _mm256_add_pd(fjy0,ty);
 516             fjz0             = _mm256_add_pd(fjz0,tz);
 517
 518             }
 519
 520             /**************************
 521              * CALCULATE INTERACTIONS *
 522              **************************/
 523
 524             if (gmx_mm256_any_lt(rsq11,rcutoff2))
 525             {
 526
 527             r11              = _mm256_mul_pd(rsq11,rinv11);
 528
 529             /* EWALD ELECTROSTATICS */
 530
 531             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 532             ewrt             = _mm256_mul_pd(r11,ewtabscale);
 533             ewitab           = _mm256_cvttpd_epi32(ewrt);
 534             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
 535             ewitab           = _mm_slli_epi32(ewitab,2);
 536             ewtabF           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
 537             ewtabD           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
 538             ewtabV           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,2) );
 539             ewtabFn          = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,3) );
 540             GMX_MM256_FULLTRANSPOSE4_PD(ewtabF,ewtabD,ewtabV,ewtabFn);
 541             felec            = _mm256_add_pd(ewtabF,_mm256_mul_pd(eweps,ewtabD));
 542             velec            = _mm256_sub_pd(ewtabV,_mm256_mul_pd(_mm256_mul_pd(ewtabhalfspace,eweps),_mm256_add_pd(ewtabF,felec)));
 543             velec            = _mm256_mul_pd(qq11,_mm256_sub_pd(_mm256_sub_pd(rinv11,sh_ewald),velec));
 544             felec            = _mm256_mul_pd(_mm256_mul_pd(qq11,rinv11),_mm256_sub_pd(rinvsq11,felec));
 545
 546             cutoff_mask      = _mm256_cmp_pd(rsq11,rcutoff2,_CMP_LT_OQ);
 547
 548             /* Update potential sum for this i atom from the interaction with this j atom. */
 549             velec            = _mm256_and_pd(velec,cutoff_mask);
 550             velecsum         = _mm256_add_pd(velecsum,velec);
 551
 552             fscal            = felec;
 553
 554             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 555
 556             /* Calculate temporary vectorial force */
 557             tx               = _mm256_mul_pd(fscal,dx11);
 558             ty               = _mm256_mul_pd(fscal,dy11);
 559             tz               = _mm256_mul_pd(fscal,dz11);
 560
 561             /* Update vectorial force */
 562             fix1             = _mm256_add_pd(fix1,tx);
 563             fiy1             = _mm256_add_pd(fiy1,ty);
 564             fiz1             = _mm256_add_pd(fiz1,tz);
 565
 566             fjx1             = _mm256_add_pd(fjx1,tx);
 567             fjy1             = _mm256_add_pd(fjy1,ty);
 568             fjz1             = _mm256_add_pd(fjz1,tz);
 569
 570             }
 571
 572             /**************************
 573              * CALCULATE INTERACTIONS *
 574              **************************/
 575
 576             if (gmx_mm256_any_lt(rsq12,rcutoff2))
 577             {
 578
 579             r12              = _mm256_mul_pd(rsq12,rinv12);
 580
 581             /* EWALD ELECTROSTATICS */
 582
 583             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 584             ewrt             = _mm256_mul_pd(r12,ewtabscale);
 585             ewitab           = _mm256_cvttpd_epi32(ewrt);
 586             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
 587             ewitab           = _mm_slli_epi32(ewitab,2);
 588             ewtabF           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
 589             ewtabD           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
 590             ewtabV           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,2) );
 591             ewtabFn          = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,3) );
 592             GMX_MM256_FULLTRANSPOSE4_PD(ewtabF,ewtabD,ewtabV,ewtabFn);
 593             felec            = _mm256_add_pd(ewtabF,_mm256_mul_pd(eweps,ewtabD));
 594             velec            = _mm256_sub_pd(ewtabV,_mm256_mul_pd(_mm256_mul_pd(ewtabhalfspace,eweps),_mm256_add_pd(ewtabF,felec)));
 595             velec            = _mm256_mul_pd(qq12,_mm256_sub_pd(_mm256_sub_pd(rinv12,sh_ewald),velec));
 596             felec            = _mm256_mul_pd(_mm256_mul_pd(qq12,rinv12),_mm256_sub_pd(rinvsq12,felec));
 597
 598             cutoff_mask      = _mm256_cmp_pd(rsq12,rcutoff2,_CMP_LT_OQ);
 599
 600             /* Update potential sum for this i atom from the interaction with this j atom. */
 601             velec            = _mm256_and_pd(velec,cutoff_mask);
 602             velecsum         = _mm256_add_pd(velecsum,velec);
 603
 604             fscal            = felec;
 605
 606             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 607
 608             /* Calculate temporary vectorial force */
 609             tx               = _mm256_mul_pd(fscal,dx12);
 610             ty               = _mm256_mul_pd(fscal,dy12);
 611             tz               = _mm256_mul_pd(fscal,dz12);
 612
 613             /* Update vectorial force */
 614             fix1             = _mm256_add_pd(fix1,tx);
 615             fiy1             = _mm256_add_pd(fiy1,ty);
 616             fiz1             = _mm256_add_pd(fiz1,tz);
 617
 618             fjx2             = _mm256_add_pd(fjx2,tx);
 619             fjy2             = _mm256_add_pd(fjy2,ty);
 620             fjz2             = _mm256_add_pd(fjz2,tz);
 621
 622             }
 623
 624             /**************************
 625              * CALCULATE INTERACTIONS *
 626              **************************/
 627
 628             if (gmx_mm256_any_lt(rsq20,rcutoff2))
 629             {
 630
 631             r20              = _mm256_mul_pd(rsq20,rinv20);
 632
 633             /* EWALD ELECTROSTATICS */
 634
 635             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 636             ewrt             = _mm256_mul_pd(r20,ewtabscale);
 637             ewitab           = _mm256_cvttpd_epi32(ewrt);
 638             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
 639             ewitab           = _mm_slli_epi32(ewitab,2);
 640             ewtabF           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
 641             ewtabD           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
 642             ewtabV           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,2) );
 643             ewtabFn          = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,3) );
 644             GMX_MM256_FULLTRANSPOSE4_PD(ewtabF,ewtabD,ewtabV,ewtabFn);
 645             felec            = _mm256_add_pd(ewtabF,_mm256_mul_pd(eweps,ewtabD));
 646             velec            = _mm256_sub_pd(ewtabV,_mm256_mul_pd(_mm256_mul_pd(ewtabhalfspace,eweps),_mm256_add_pd(ewtabF,felec)));
 647             velec            = _mm256_mul_pd(qq20,_mm256_sub_pd(_mm256_sub_pd(rinv20,sh_ewald),velec));
 648             felec            = _mm256_mul_pd(_mm256_mul_pd(qq20,rinv20),_mm256_sub_pd(rinvsq20,felec));
 649
 650             cutoff_mask      = _mm256_cmp_pd(rsq20,rcutoff2,_CMP_LT_OQ);
 651
 652             /* Update potential sum for this i atom from the interaction with this j atom. */
 653             velec            = _mm256_and_pd(velec,cutoff_mask);
 654             velecsum         = _mm256_add_pd(velecsum,velec);
 655
 656             fscal            = felec;
 657
 658             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 659
 660             /* Calculate temporary vectorial force */
 661             tx               = _mm256_mul_pd(fscal,dx20);
 662             ty               = _mm256_mul_pd(fscal,dy20);
 663             tz               = _mm256_mul_pd(fscal,dz20);
 664
 665             /* Update vectorial force */
 666             fix2             = _mm256_add_pd(fix2,tx);
 667             fiy2             = _mm256_add_pd(fiy2,ty);
 668             fiz2             = _mm256_add_pd(fiz2,tz);
 669
 670             fjx0             = _mm256_add_pd(fjx0,tx);
 671             fjy0             = _mm256_add_pd(fjy0,ty);
 672             fjz0             = _mm256_add_pd(fjz0,tz);
 673
 674             }
 675
 676             /**************************
 677              * CALCULATE INTERACTIONS *
 678              **************************/
 679
 680             if (gmx_mm256_any_lt(rsq21,rcutoff2))
 681             {
 682
 683             r21              = _mm256_mul_pd(rsq21,rinv21);
 684
 685             /* EWALD ELECTROSTATICS */
 686
 687             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 688             ewrt             = _mm256_mul_pd(r21,ewtabscale);
 689             ewitab           = _mm256_cvttpd_epi32(ewrt);
 690             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
 691             ewitab           = _mm_slli_epi32(ewitab,2);
 692             ewtabF           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
 693             ewtabD           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
 694             ewtabV           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,2) );
 695             ewtabFn          = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,3) );
 696             GMX_MM256_FULLTRANSPOSE4_PD(ewtabF,ewtabD,ewtabV,ewtabFn);
 697             felec            = _mm256_add_pd(ewtabF,_mm256_mul_pd(eweps,ewtabD));
 698             velec            = _mm256_sub_pd(ewtabV,_mm256_mul_pd(_mm256_mul_pd(ewtabhalfspace,eweps),_mm256_add_pd(ewtabF,felec)));
 699             velec            = _mm256_mul_pd(qq21,_mm256_sub_pd(_mm256_sub_pd(rinv21,sh_ewald),velec));
 700             felec            = _mm256_mul_pd(_mm256_mul_pd(qq21,rinv21),_mm256_sub_pd(rinvsq21,felec));
 701
 702             cutoff_mask      = _mm256_cmp_pd(rsq21,rcutoff2,_CMP_LT_OQ);
 703
 704             /* Update potential sum for this i atom from the interaction with this j atom. */
 705             velec            = _mm256_and_pd(velec,cutoff_mask);
 706             velecsum         = _mm256_add_pd(velecsum,velec);
 707
 708             fscal            = felec;
 709
 710             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 711
 712             /* Calculate temporary vectorial force */
 713             tx               = _mm256_mul_pd(fscal,dx21);
 714             ty               = _mm256_mul_pd(fscal,dy21);
 715             tz               = _mm256_mul_pd(fscal,dz21);
 716
 717             /* Update vectorial force */
 718             fix2             = _mm256_add_pd(fix2,tx);
 719             fiy2             = _mm256_add_pd(fiy2,ty);
 720             fiz2             = _mm256_add_pd(fiz2,tz);
 721
 722             fjx1             = _mm256_add_pd(fjx1,tx);
 723             fjy1             = _mm256_add_pd(fjy1,ty);
 724             fjz1             = _mm256_add_pd(fjz1,tz);
 725
 726             }
 727
 728             /**************************
 729              * CALCULATE INTERACTIONS *
 730              **************************/
 731
 732             if (gmx_mm256_any_lt(rsq22,rcutoff2))
 733             {
 734
 735             r22              = _mm256_mul_pd(rsq22,rinv22);
 736
 737             /* EWALD ELECTROSTATICS */
 738
 739             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 740             ewrt             = _mm256_mul_pd(r22,ewtabscale);
 741             ewitab           = _mm256_cvttpd_epi32(ewrt);
 742             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
 743             ewitab           = _mm_slli_epi32(ewitab,2);
 744             ewtabF           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
 745             ewtabD           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
 746             ewtabV           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,2) );
 747             ewtabFn          = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,3) );
 748             GMX_MM256_FULLTRANSPOSE4_PD(ewtabF,ewtabD,ewtabV,ewtabFn);
 749             felec            = _mm256_add_pd(ewtabF,_mm256_mul_pd(eweps,ewtabD));
 750             velec            = _mm256_sub_pd(ewtabV,_mm256_mul_pd(_mm256_mul_pd(ewtabhalfspace,eweps),_mm256_add_pd(ewtabF,felec)));
 751             velec            = _mm256_mul_pd(qq22,_mm256_sub_pd(_mm256_sub_pd(rinv22,sh_ewald),velec));
 752             felec            = _mm256_mul_pd(_mm256_mul_pd(qq22,rinv22),_mm256_sub_pd(rinvsq22,felec));
 753
 754             cutoff_mask      = _mm256_cmp_pd(rsq22,rcutoff2,_CMP_LT_OQ);
 755
 756             /* Update potential sum for this i atom from the interaction with this j atom. */
 757             velec            = _mm256_and_pd(velec,cutoff_mask);
 758             velecsum         = _mm256_add_pd(velecsum,velec);
 759
 760             fscal            = felec;
 761
 762             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 763
 764             /* Calculate temporary vectorial force */
 765             tx               = _mm256_mul_pd(fscal,dx22);
 766             ty               = _mm256_mul_pd(fscal,dy22);
 767             tz               = _mm256_mul_pd(fscal,dz22);
 768
 769             /* Update vectorial force */
 770             fix2             = _mm256_add_pd(fix2,tx);
 771             fiy2             = _mm256_add_pd(fiy2,ty);
 772             fiz2             = _mm256_add_pd(fiz2,tz);
 773
 774             fjx2             = _mm256_add_pd(fjx2,tx);
 775             fjy2             = _mm256_add_pd(fjy2,ty);
 776             fjz2             = _mm256_add_pd(fjz2,tz);
 777
 778             }
 779
 780             fjptrA             = f+j_coord_offsetA;
 781             fjptrB             = f+j_coord_offsetB;
 782             fjptrC             = f+j_coord_offsetC;
 783             fjptrD             = f+j_coord_offsetD;
 784
 785             gmx_mm256_decrement_3rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,
 786                                                       fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2);
 787
 788             /* Inner loop uses 432 flops */
 789         }
 790
 791         if(jidx<j_index_end)
 792         {
 793
 794             /* Get j neighbor index, and coordinate index */
 795             jnrlistA         = jjnr[jidx];
 796             jnrlistB         = jjnr[jidx+1];
 797             jnrlistC         = jjnr[jidx+2];
 798             jnrlistD         = jjnr[jidx+3];
 799             /* Sign of each element will be negative for non-real atoms.
 800              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
 801              * so use it as val = _mm_andnot_pd(mask,val) to clear dummy entries.
 802              */
 803             tmpmask0 = gmx_mm_castsi128_pd(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128()));
 804
 805             tmpmask1 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(3,3,2,2));
 806             tmpmask0 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(1,1,0,0));
 807             dummy_mask = _mm256_castps_pd(gmx_mm256_set_m128(tmpmask1,tmpmask0));
 808
 809             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
 810             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
 811             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
 812             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
 813             j_coord_offsetA  = DIM*jnrA;
 814             j_coord_offsetB  = DIM*jnrB;
 815             j_coord_offsetC  = DIM*jnrC;
 816             j_coord_offsetD  = DIM*jnrD;
 817
 818             /* load j atom coordinates */
 819             gmx_mm256_load_3rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
 820                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 821                                               &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,&jy2,&jz2);
 822
 823             /* Calculate displacement vector */
 824             dx00             = _mm256_sub_pd(ix0,jx0);
 825             dy00             = _mm256_sub_pd(iy0,jy0);
 826             dz00             = _mm256_sub_pd(iz0,jz0);
 827             dx01             = _mm256_sub_pd(ix0,jx1);
 828             dy01             = _mm256_sub_pd(iy0,jy1);
 829             dz01             = _mm256_sub_pd(iz0,jz1);
 830             dx02             = _mm256_sub_pd(ix0,jx2);
 831             dy02             = _mm256_sub_pd(iy0,jy2);
 832             dz02             = _mm256_sub_pd(iz0,jz2);
 833             dx10             = _mm256_sub_pd(ix1,jx0);
 834             dy10             = _mm256_sub_pd(iy1,jy0);
 835             dz10             = _mm256_sub_pd(iz1,jz0);
 836             dx11             = _mm256_sub_pd(ix1,jx1);
 837             dy11             = _mm256_sub_pd(iy1,jy1);
 838             dz11             = _mm256_sub_pd(iz1,jz1);
 839             dx12             = _mm256_sub_pd(ix1,jx2);
 840             dy12             = _mm256_sub_pd(iy1,jy2);
 841             dz12             = _mm256_sub_pd(iz1,jz2);
 842             dx20             = _mm256_sub_pd(ix2,jx0);
 843             dy20             = _mm256_sub_pd(iy2,jy0);
 844             dz20             = _mm256_sub_pd(iz2,jz0);
 845             dx21             = _mm256_sub_pd(ix2,jx1);
 846             dy21             = _mm256_sub_pd(iy2,jy1);
 847             dz21             = _mm256_sub_pd(iz2,jz1);
 848             dx22             = _mm256_sub_pd(ix2,jx2);
 849             dy22             = _mm256_sub_pd(iy2,jy2);
 850             dz22             = _mm256_sub_pd(iz2,jz2);
 851
 852             /* Calculate squared distance and things based on it */
 853             rsq00            = gmx_mm256_calc_rsq_pd(dx00,dy00,dz00);
 854             rsq01            = gmx_mm256_calc_rsq_pd(dx01,dy01,dz01);
 855             rsq02            = gmx_mm256_calc_rsq_pd(dx02,dy02,dz02);
 856             rsq10            = gmx_mm256_calc_rsq_pd(dx10,dy10,dz10);
 857             rsq11            = gmx_mm256_calc_rsq_pd(dx11,dy11,dz11);
 858             rsq12            = gmx_mm256_calc_rsq_pd(dx12,dy12,dz12);
 859             rsq20            = gmx_mm256_calc_rsq_pd(dx20,dy20,dz20);
 860             rsq21            = gmx_mm256_calc_rsq_pd(dx21,dy21,dz21);
 861             rsq22            = gmx_mm256_calc_rsq_pd(dx22,dy22,dz22);
 862
 863             rinv00           = gmx_mm256_invsqrt_pd(rsq00);
 864             rinv01           = gmx_mm256_invsqrt_pd(rsq01);
 865             rinv02           = gmx_mm256_invsqrt_pd(rsq02);
 866             rinv10           = gmx_mm256_invsqrt_pd(rsq10);
 867             rinv11           = gmx_mm256_invsqrt_pd(rsq11);
 868             rinv12           = gmx_mm256_invsqrt_pd(rsq12);
 869             rinv20           = gmx_mm256_invsqrt_pd(rsq20);
 870             rinv21           = gmx_mm256_invsqrt_pd(rsq21);
 871             rinv22           = gmx_mm256_invsqrt_pd(rsq22);
 872
 873             rinvsq00         = _mm256_mul_pd(rinv00,rinv00);
 874             rinvsq01         = _mm256_mul_pd(rinv01,rinv01);
 875             rinvsq02         = _mm256_mul_pd(rinv02,rinv02);
 876             rinvsq10         = _mm256_mul_pd(rinv10,rinv10);
 877             rinvsq11         = _mm256_mul_pd(rinv11,rinv11);
 878             rinvsq12         = _mm256_mul_pd(rinv12,rinv12);
 879             rinvsq20         = _mm256_mul_pd(rinv20,rinv20);
 880             rinvsq21         = _mm256_mul_pd(rinv21,rinv21);
 881             rinvsq22         = _mm256_mul_pd(rinv22,rinv22);
 882
 883             fjx0             = _mm256_setzero_pd();
 884             fjy0             = _mm256_setzero_pd();
 885             fjz0             = _mm256_setzero_pd();
 886             fjx1             = _mm256_setzero_pd();
 887             fjy1             = _mm256_setzero_pd();
 888             fjz1             = _mm256_setzero_pd();
 889             fjx2             = _mm256_setzero_pd();
 890             fjy2             = _mm256_setzero_pd();
 891             fjz2             = _mm256_setzero_pd();
 892
 893             /**************************
 894              * CALCULATE INTERACTIONS *
 895              **************************/
 896
 897             if (gmx_mm256_any_lt(rsq00,rcutoff2))
 898             {
 899
 900             r00              = _mm256_mul_pd(rsq00,rinv00);
 901             r00              = _mm256_andnot_pd(dummy_mask,r00);
 902
 903             /* EWALD ELECTROSTATICS */
 904
 905             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 906             ewrt             = _mm256_mul_pd(r00,ewtabscale);
 907             ewitab           = _mm256_cvttpd_epi32(ewrt);
 908             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
 909             ewitab           = _mm_slli_epi32(ewitab,2);
 910             ewtabF           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
 911             ewtabD           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
 912             ewtabV           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,2) );
 913             ewtabFn          = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,3) );
 914             GMX_MM256_FULLTRANSPOSE4_PD(ewtabF,ewtabD,ewtabV,ewtabFn);
 915             felec            = _mm256_add_pd(ewtabF,_mm256_mul_pd(eweps,ewtabD));
 916             velec            = _mm256_sub_pd(ewtabV,_mm256_mul_pd(_mm256_mul_pd(ewtabhalfspace,eweps),_mm256_add_pd(ewtabF,felec)));
 917             velec            = _mm256_mul_pd(qq00,_mm256_sub_pd(_mm256_sub_pd(rinv00,sh_ewald),velec));
 918             felec            = _mm256_mul_pd(_mm256_mul_pd(qq00,rinv00),_mm256_sub_pd(rinvsq00,felec));
 919
 920             /* LENNARD-JONES DISPERSION/REPULSION */
 921
 922             rinvsix          = _mm256_mul_pd(_mm256_mul_pd(rinvsq00,rinvsq00),rinvsq00);
 923             vvdw6            = _mm256_mul_pd(c6_00,rinvsix);
 924             vvdw12           = _mm256_mul_pd(c12_00,_mm256_mul_pd(rinvsix,rinvsix));
 925             vvdw             = _mm256_sub_pd(_mm256_mul_pd( _mm256_sub_pd(vvdw12 , _mm256_mul_pd(c12_00,_mm256_mul_pd(sh_vdw_invrcut6,sh_vdw_invrcut6))), one_twelfth) ,
 926                                           _mm256_mul_pd( _mm256_sub_pd(vvdw6,_mm256_mul_pd(c6_00,sh_vdw_invrcut6)),one_sixth));
 927             fvdw             = _mm256_mul_pd(_mm256_sub_pd(vvdw12,vvdw6),rinvsq00);
 928
 929             cutoff_mask      = _mm256_cmp_pd(rsq00,rcutoff2,_CMP_LT_OQ);
 930
 931             /* Update potential sum for this i atom from the interaction with this j atom. */
 932             velec            = _mm256_and_pd(velec,cutoff_mask);
 933             velec            = _mm256_andnot_pd(dummy_mask,velec);
 934             velecsum         = _mm256_add_pd(velecsum,velec);
 935             vvdw             = _mm256_and_pd(vvdw,cutoff_mask);
 936             vvdw             = _mm256_andnot_pd(dummy_mask,vvdw);
 937             vvdwsum          = _mm256_add_pd(vvdwsum,vvdw);
 938
 939             fscal            = _mm256_add_pd(felec,fvdw);
 940
 941             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 942
 943             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
 944
 945             /* Calculate temporary vectorial force */
 946             tx               = _mm256_mul_pd(fscal,dx00);
 947             ty               = _mm256_mul_pd(fscal,dy00);
 948             tz               = _mm256_mul_pd(fscal,dz00);
 949
 950             /* Update vectorial force */
 951             fix0             = _mm256_add_pd(fix0,tx);
 952             fiy0             = _mm256_add_pd(fiy0,ty);
 953             fiz0             = _mm256_add_pd(fiz0,tz);
 954
 955             fjx0             = _mm256_add_pd(fjx0,tx);
 956             fjy0             = _mm256_add_pd(fjy0,ty);
 957             fjz0             = _mm256_add_pd(fjz0,tz);
 958
 959             }
 960
 961             /**************************
 962              * CALCULATE INTERACTIONS *
 963              **************************/
 964
 965             if (gmx_mm256_any_lt(rsq01,rcutoff2))
 966             {
 967
 968             r01              = _mm256_mul_pd(rsq01,rinv01);
 969             r01              = _mm256_andnot_pd(dummy_mask,r01);
 970
 971             /* EWALD ELECTROSTATICS */
 972
 973             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 974             ewrt             = _mm256_mul_pd(r01,ewtabscale);
 975             ewitab           = _mm256_cvttpd_epi32(ewrt);
 976             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
 977             ewitab           = _mm_slli_epi32(ewitab,2);
 978             ewtabF           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
 979             ewtabD           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
 980             ewtabV           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,2) );
 981             ewtabFn          = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,3) );
 982             GMX_MM256_FULLTRANSPOSE4_PD(ewtabF,ewtabD,ewtabV,ewtabFn);
 983             felec            = _mm256_add_pd(ewtabF,_mm256_mul_pd(eweps,ewtabD));
 984             velec            = _mm256_sub_pd(ewtabV,_mm256_mul_pd(_mm256_mul_pd(ewtabhalfspace,eweps),_mm256_add_pd(ewtabF,felec)));
 985             velec            = _mm256_mul_pd(qq01,_mm256_sub_pd(_mm256_sub_pd(rinv01,sh_ewald),velec));
 986             felec            = _mm256_mul_pd(_mm256_mul_pd(qq01,rinv01),_mm256_sub_pd(rinvsq01,felec));
 987
 988             cutoff_mask      = _mm256_cmp_pd(rsq01,rcutoff2,_CMP_LT_OQ);
 989
 990             /* Update potential sum for this i atom from the interaction with this j atom. */
 991             velec            = _mm256_and_pd(velec,cutoff_mask);
 992             velec            = _mm256_andnot_pd(dummy_mask,velec);
 993             velecsum         = _mm256_add_pd(velecsum,velec);
 994
 995             fscal            = felec;
 996
 997             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 998
 999             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
1000
1001             /* Calculate temporary vectorial force */
1002             tx               = _mm256_mul_pd(fscal,dx01);
1003             ty               = _mm256_mul_pd(fscal,dy01);
1004             tz               = _mm256_mul_pd(fscal,dz01);
1005
1006             /* Update vectorial force */
1007             fix0             = _mm256_add_pd(fix0,tx);
1008             fiy0             = _mm256_add_pd(fiy0,ty);
1009             fiz0             = _mm256_add_pd(fiz0,tz);
1010
1011             fjx1             = _mm256_add_pd(fjx1,tx);
1012             fjy1             = _mm256_add_pd(fjy1,ty);
1013             fjz1             = _mm256_add_pd(fjz1,tz);
1014
1015             }
1016
1017             /**************************
1018              * CALCULATE INTERACTIONS *
1019              **************************/
1020
1021             if (gmx_mm256_any_lt(rsq02,rcutoff2))
1022             {
1023
1024             r02              = _mm256_mul_pd(rsq02,rinv02);
1025             r02              = _mm256_andnot_pd(dummy_mask,r02);
1026
1027             /* EWALD ELECTROSTATICS */
1028
1029             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1030             ewrt             = _mm256_mul_pd(r02,ewtabscale);
1031             ewitab           = _mm256_cvttpd_epi32(ewrt);
1032             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
1033             ewitab           = _mm_slli_epi32(ewitab,2);
1034             ewtabF           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
1035             ewtabD           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
1036             ewtabV           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,2) );
1037             ewtabFn          = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,3) );
1038             GMX_MM256_FULLTRANSPOSE4_PD(ewtabF,ewtabD,ewtabV,ewtabFn);
1039             felec            = _mm256_add_pd(ewtabF,_mm256_mul_pd(eweps,ewtabD));
1040             velec            = _mm256_sub_pd(ewtabV,_mm256_mul_pd(_mm256_mul_pd(ewtabhalfspace,eweps),_mm256_add_pd(ewtabF,felec)));
1041             velec            = _mm256_mul_pd(qq02,_mm256_sub_pd(_mm256_sub_pd(rinv02,sh_ewald),velec));
1042             felec            = _mm256_mul_pd(_mm256_mul_pd(qq02,rinv02),_mm256_sub_pd(rinvsq02,felec));
1043
1044             cutoff_mask      = _mm256_cmp_pd(rsq02,rcutoff2,_CMP_LT_OQ);
1045
1046             /* Update potential sum for this i atom from the interaction with this j atom. */
1047             velec            = _mm256_and_pd(velec,cutoff_mask);
1048             velec            = _mm256_andnot_pd(dummy_mask,velec);
1049             velecsum         = _mm256_add_pd(velecsum,velec);
1050
1051             fscal            = felec;
1052
1053             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1054
1055             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
1056
1057             /* Calculate temporary vectorial force */
1058             tx               = _mm256_mul_pd(fscal,dx02);
1059             ty               = _mm256_mul_pd(fscal,dy02);
1060             tz               = _mm256_mul_pd(fscal,dz02);
1061
1062             /* Update vectorial force */
1063             fix0             = _mm256_add_pd(fix0,tx);
1064             fiy0             = _mm256_add_pd(fiy0,ty);
1065             fiz0             = _mm256_add_pd(fiz0,tz);
1066
1067             fjx2             = _mm256_add_pd(fjx2,tx);
1068             fjy2             = _mm256_add_pd(fjy2,ty);
1069             fjz2             = _mm256_add_pd(fjz2,tz);
1070
1071             }
1072
1073             /**************************
1074              * CALCULATE INTERACTIONS *
1075              **************************/
1076
1077             if (gmx_mm256_any_lt(rsq10,rcutoff2))
1078             {
1079
1080             r10              = _mm256_mul_pd(rsq10,rinv10);
1081             r10              = _mm256_andnot_pd(dummy_mask,r10);
1082
1083             /* EWALD ELECTROSTATICS */
1084
1085             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1086             ewrt             = _mm256_mul_pd(r10,ewtabscale);
1087             ewitab           = _mm256_cvttpd_epi32(ewrt);
1088             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
1089             ewitab           = _mm_slli_epi32(ewitab,2);
1090             ewtabF           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
1091             ewtabD           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
1092             ewtabV           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,2) );
1093             ewtabFn          = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,3) );
1094             GMX_MM256_FULLTRANSPOSE4_PD(ewtabF,ewtabD,ewtabV,ewtabFn);
1095             felec            = _mm256_add_pd(ewtabF,_mm256_mul_pd(eweps,ewtabD));
1096             velec            = _mm256_sub_pd(ewtabV,_mm256_mul_pd(_mm256_mul_pd(ewtabhalfspace,eweps),_mm256_add_pd(ewtabF,felec)));
1097             velec            = _mm256_mul_pd(qq10,_mm256_sub_pd(_mm256_sub_pd(rinv10,sh_ewald),velec));
1098             felec            = _mm256_mul_pd(_mm256_mul_pd(qq10,rinv10),_mm256_sub_pd(rinvsq10,felec));
1099
1100             cutoff_mask      = _mm256_cmp_pd(rsq10,rcutoff2,_CMP_LT_OQ);
1101
1102             /* Update potential sum for this i atom from the interaction with this j atom. */
1103             velec            = _mm256_and_pd(velec,cutoff_mask);
1104             velec            = _mm256_andnot_pd(dummy_mask,velec);
1105             velecsum         = _mm256_add_pd(velecsum,velec);
1106
1107             fscal            = felec;
1108
1109             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1110
1111             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
1112
1113             /* Calculate temporary vectorial force */
1114             tx               = _mm256_mul_pd(fscal,dx10);
1115             ty               = _mm256_mul_pd(fscal,dy10);
1116             tz               = _mm256_mul_pd(fscal,dz10);
1117
1118             /* Update vectorial force */
1119             fix1             = _mm256_add_pd(fix1,tx);
1120             fiy1             = _mm256_add_pd(fiy1,ty);
1121             fiz1             = _mm256_add_pd(fiz1,tz);
1122
1123             fjx0             = _mm256_add_pd(fjx0,tx);
1124             fjy0             = _mm256_add_pd(fjy0,ty);
1125             fjz0             = _mm256_add_pd(fjz0,tz);
1126
1127             }
1128
1129             /**************************
1130              * CALCULATE INTERACTIONS *
1131              **************************/
1132
1133             if (gmx_mm256_any_lt(rsq11,rcutoff2))
1134             {
1135
1136             r11              = _mm256_mul_pd(rsq11,rinv11);
1137             r11              = _mm256_andnot_pd(dummy_mask,r11);
1138
1139             /* EWALD ELECTROSTATICS */
1140
1141             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1142             ewrt             = _mm256_mul_pd(r11,ewtabscale);
1143             ewitab           = _mm256_cvttpd_epi32(ewrt);
1144             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
1145             ewitab           = _mm_slli_epi32(ewitab,2);
1146             ewtabF           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
1147             ewtabD           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
1148             ewtabV           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,2) );
1149             ewtabFn          = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,3) );
1150             GMX_MM256_FULLTRANSPOSE4_PD(ewtabF,ewtabD,ewtabV,ewtabFn);
1151             felec            = _mm256_add_pd(ewtabF,_mm256_mul_pd(eweps,ewtabD));
1152             velec            = _mm256_sub_pd(ewtabV,_mm256_mul_pd(_mm256_mul_pd(ewtabhalfspace,eweps),_mm256_add_pd(ewtabF,felec)));
1153             velec            = _mm256_mul_pd(qq11,_mm256_sub_pd(_mm256_sub_pd(rinv11,sh_ewald),velec));
1154             felec            = _mm256_mul_pd(_mm256_mul_pd(qq11,rinv11),_mm256_sub_pd(rinvsq11,felec));
1155
1156             cutoff_mask      = _mm256_cmp_pd(rsq11,rcutoff2,_CMP_LT_OQ);
1157
1158             /* Update potential sum for this i atom from the interaction with this j atom. */
1159             velec            = _mm256_and_pd(velec,cutoff_mask);
1160             velec            = _mm256_andnot_pd(dummy_mask,velec);
1161             velecsum         = _mm256_add_pd(velecsum,velec);
1162
1163             fscal            = felec;
1164
1165             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1166
1167             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
1168
1169             /* Calculate temporary vectorial force */
1170             tx               = _mm256_mul_pd(fscal,dx11);
1171             ty               = _mm256_mul_pd(fscal,dy11);
1172             tz               = _mm256_mul_pd(fscal,dz11);
1173
1174             /* Update vectorial force */
1175             fix1             = _mm256_add_pd(fix1,tx);
1176             fiy1             = _mm256_add_pd(fiy1,ty);
1177             fiz1             = _mm256_add_pd(fiz1,tz);
1178
1179             fjx1             = _mm256_add_pd(fjx1,tx);
1180             fjy1             = _mm256_add_pd(fjy1,ty);
1181             fjz1             = _mm256_add_pd(fjz1,tz);
1182
1183             }
1184
1185             /**************************
1186              * CALCULATE INTERACTIONS *
1187              **************************/
1188
1189             if (gmx_mm256_any_lt(rsq12,rcutoff2))
1190             {
1191
1192             r12              = _mm256_mul_pd(rsq12,rinv12);
1193             r12              = _mm256_andnot_pd(dummy_mask,r12);
1194
1195             /* EWALD ELECTROSTATICS */
1196
1197             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1198             ewrt             = _mm256_mul_pd(r12,ewtabscale);
1199             ewitab           = _mm256_cvttpd_epi32(ewrt);
1200             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
1201             ewitab           = _mm_slli_epi32(ewitab,2);
1202             ewtabF           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
1203             ewtabD           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
1204             ewtabV           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,2) );
1205             ewtabFn          = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,3) );
1206             GMX_MM256_FULLTRANSPOSE4_PD(ewtabF,ewtabD,ewtabV,ewtabFn);
1207             felec            = _mm256_add_pd(ewtabF,_mm256_mul_pd(eweps,ewtabD));
1208             velec            = _mm256_sub_pd(ewtabV,_mm256_mul_pd(_mm256_mul_pd(ewtabhalfspace,eweps),_mm256_add_pd(ewtabF,felec)));
1209             velec            = _mm256_mul_pd(qq12,_mm256_sub_pd(_mm256_sub_pd(rinv12,sh_ewald),velec));
1210             felec            = _mm256_mul_pd(_mm256_mul_pd(qq12,rinv12),_mm256_sub_pd(rinvsq12,felec));
1211
1212             cutoff_mask      = _mm256_cmp_pd(rsq12,rcutoff2,_CMP_LT_OQ);
1213
1214             /* Update potential sum for this i atom from the interaction with this j atom. */
1215             velec            = _mm256_and_pd(velec,cutoff_mask);
1216             velec            = _mm256_andnot_pd(dummy_mask,velec);
1217             velecsum         = _mm256_add_pd(velecsum,velec);
1218
1219             fscal            = felec;
1220
1221             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1222
1223             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
1224
1225             /* Calculate temporary vectorial force */
1226             tx               = _mm256_mul_pd(fscal,dx12);
1227             ty               = _mm256_mul_pd(fscal,dy12);
1228             tz               = _mm256_mul_pd(fscal,dz12);
1229
1230             /* Update vectorial force */
1231             fix1             = _mm256_add_pd(fix1,tx);
1232             fiy1             = _mm256_add_pd(fiy1,ty);
1233             fiz1             = _mm256_add_pd(fiz1,tz);
1234
1235             fjx2             = _mm256_add_pd(fjx2,tx);
1236             fjy2             = _mm256_add_pd(fjy2,ty);
1237             fjz2             = _mm256_add_pd(fjz2,tz);
1238
1239             }
1240
1241             /**************************
1242              * CALCULATE INTERACTIONS *
1243              **************************/
1244
1245             if (gmx_mm256_any_lt(rsq20,rcutoff2))
1246             {
1247
1248             r20              = _mm256_mul_pd(rsq20,rinv20);
1249             r20              = _mm256_andnot_pd(dummy_mask,r20);
1250
1251             /* EWALD ELECTROSTATICS */
1252
1253             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1254             ewrt             = _mm256_mul_pd(r20,ewtabscale);
1255             ewitab           = _mm256_cvttpd_epi32(ewrt);
1256             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
1257             ewitab           = _mm_slli_epi32(ewitab,2);
1258             ewtabF           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
1259             ewtabD           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
1260             ewtabV           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,2) );
1261             ewtabFn          = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,3) );
1262             GMX_MM256_FULLTRANSPOSE4_PD(ewtabF,ewtabD,ewtabV,ewtabFn);
1263             felec            = _mm256_add_pd(ewtabF,_mm256_mul_pd(eweps,ewtabD));
1264             velec            = _mm256_sub_pd(ewtabV,_mm256_mul_pd(_mm256_mul_pd(ewtabhalfspace,eweps),_mm256_add_pd(ewtabF,felec)));
1265             velec            = _mm256_mul_pd(qq20,_mm256_sub_pd(_mm256_sub_pd(rinv20,sh_ewald),velec));
1266             felec            = _mm256_mul_pd(_mm256_mul_pd(qq20,rinv20),_mm256_sub_pd(rinvsq20,felec));
1267
1268             cutoff_mask      = _mm256_cmp_pd(rsq20,rcutoff2,_CMP_LT_OQ);
1269
1270             /* Update potential sum for this i atom from the interaction with this j atom. */
1271             velec            = _mm256_and_pd(velec,cutoff_mask);
1272             velec            = _mm256_andnot_pd(dummy_mask,velec);
1273             velecsum         = _mm256_add_pd(velecsum,velec);
1274
1275             fscal            = felec;
1276
1277             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1278
1279             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
1280
1281             /* Calculate temporary vectorial force */
1282             tx               = _mm256_mul_pd(fscal,dx20);
1283             ty               = _mm256_mul_pd(fscal,dy20);
1284             tz               = _mm256_mul_pd(fscal,dz20);
1285
1286             /* Update vectorial force */
1287             fix2             = _mm256_add_pd(fix2,tx);
1288             fiy2             = _mm256_add_pd(fiy2,ty);
1289             fiz2             = _mm256_add_pd(fiz2,tz);
1290
1291             fjx0             = _mm256_add_pd(fjx0,tx);
1292             fjy0             = _mm256_add_pd(fjy0,ty);
1293             fjz0             = _mm256_add_pd(fjz0,tz);
1294
1295             }
1296
1297             /**************************
1298              * CALCULATE INTERACTIONS *
1299              **************************/
1300
1301             if (gmx_mm256_any_lt(rsq21,rcutoff2))
1302             {
1303
1304             r21              = _mm256_mul_pd(rsq21,rinv21);
1305             r21              = _mm256_andnot_pd(dummy_mask,r21);
1306
1307             /* EWALD ELECTROSTATICS */
1308
1309             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1310             ewrt             = _mm256_mul_pd(r21,ewtabscale);
1311             ewitab           = _mm256_cvttpd_epi32(ewrt);
1312             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
1313             ewitab           = _mm_slli_epi32(ewitab,2);
1314             ewtabF           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
1315             ewtabD           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
1316             ewtabV           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,2) );
1317             ewtabFn          = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,3) );
1318             GMX_MM256_FULLTRANSPOSE4_PD(ewtabF,ewtabD,ewtabV,ewtabFn);
1319             felec            = _mm256_add_pd(ewtabF,_mm256_mul_pd(eweps,ewtabD));
1320             velec            = _mm256_sub_pd(ewtabV,_mm256_mul_pd(_mm256_mul_pd(ewtabhalfspace,eweps),_mm256_add_pd(ewtabF,felec)));
1321             velec            = _mm256_mul_pd(qq21,_mm256_sub_pd(_mm256_sub_pd(rinv21,sh_ewald),velec));
1322             felec            = _mm256_mul_pd(_mm256_mul_pd(qq21,rinv21),_mm256_sub_pd(rinvsq21,felec));
1323
1324             cutoff_mask      = _mm256_cmp_pd(rsq21,rcutoff2,_CMP_LT_OQ);
1325
1326             /* Update potential sum for this i atom from the interaction with this j atom. */
1327             velec            = _mm256_and_pd(velec,cutoff_mask);
1328             velec            = _mm256_andnot_pd(dummy_mask,velec);
1329             velecsum         = _mm256_add_pd(velecsum,velec);
1330
1331             fscal            = felec;
1332
1333             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1334
1335             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
1336
1337             /* Calculate temporary vectorial force */
1338             tx               = _mm256_mul_pd(fscal,dx21);
1339             ty               = _mm256_mul_pd(fscal,dy21);
1340             tz               = _mm256_mul_pd(fscal,dz21);
1341
1342             /* Update vectorial force */
1343             fix2             = _mm256_add_pd(fix2,tx);
1344             fiy2             = _mm256_add_pd(fiy2,ty);
1345             fiz2             = _mm256_add_pd(fiz2,tz);
1346
1347             fjx1             = _mm256_add_pd(fjx1,tx);
1348             fjy1             = _mm256_add_pd(fjy1,ty);
1349             fjz1             = _mm256_add_pd(fjz1,tz);
1350
1351             }
1352
1353             /**************************
1354              * CALCULATE INTERACTIONS *
1355              **************************/
1356
1357             if (gmx_mm256_any_lt(rsq22,rcutoff2))
1358             {
1359
1360             r22              = _mm256_mul_pd(rsq22,rinv22);
1361             r22              = _mm256_andnot_pd(dummy_mask,r22);
1362
1363             /* EWALD ELECTROSTATICS */
1364
1365             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1366             ewrt             = _mm256_mul_pd(r22,ewtabscale);
1367             ewitab           = _mm256_cvttpd_epi32(ewrt);
1368             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
1369             ewitab           = _mm_slli_epi32(ewitab,2);
1370             ewtabF           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
1371             ewtabD           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
1372             ewtabV           = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,2) );
1373             ewtabFn          = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,3) );
1374             GMX_MM256_FULLTRANSPOSE4_PD(ewtabF,ewtabD,ewtabV,ewtabFn);
1375             felec            = _mm256_add_pd(ewtabF,_mm256_mul_pd(eweps,ewtabD));
1376             velec            = _mm256_sub_pd(ewtabV,_mm256_mul_pd(_mm256_mul_pd(ewtabhalfspace,eweps),_mm256_add_pd(ewtabF,felec)));
1377             velec            = _mm256_mul_pd(qq22,_mm256_sub_pd(_mm256_sub_pd(rinv22,sh_ewald),velec));
1378             felec            = _mm256_mul_pd(_mm256_mul_pd(qq22,rinv22),_mm256_sub_pd(rinvsq22,felec));
1379
1380             cutoff_mask      = _mm256_cmp_pd(rsq22,rcutoff2,_CMP_LT_OQ);
1381
1382             /* Update potential sum for this i atom from the interaction with this j atom. */
1383             velec            = _mm256_and_pd(velec,cutoff_mask);
1384             velec            = _mm256_andnot_pd(dummy_mask,velec);
1385             velecsum         = _mm256_add_pd(velecsum,velec);
1386
1387             fscal            = felec;
1388
1389             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1390
1391             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
1392
1393             /* Calculate temporary vectorial force */
1394             tx               = _mm256_mul_pd(fscal,dx22);
1395             ty               = _mm256_mul_pd(fscal,dy22);
1396             tz               = _mm256_mul_pd(fscal,dz22);
1397
1398             /* Update vectorial force */
1399             fix2             = _mm256_add_pd(fix2,tx);
1400             fiy2             = _mm256_add_pd(fiy2,ty);
1401             fiz2             = _mm256_add_pd(fiz2,tz);
1402
1403             fjx2             = _mm256_add_pd(fjx2,tx);
1404             fjy2             = _mm256_add_pd(fjy2,ty);
1405             fjz2             = _mm256_add_pd(fjz2,tz);
1406
1407             }
1408
1409             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
1410             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
1411             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
1412             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
1413
1414             gmx_mm256_decrement_3rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,
1415                                                       fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2);
1416
1417             /* Inner loop uses 441 flops */
1418         }
1419
1420         /* End of innermost loop */
1421
1422         gmx_mm256_update_iforce_3atom_swizzle_pd(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,
1423                                                  f+i_coord_offset,fshift+i_shift_offset);
1424
1425         ggid                        = gid[iidx];
1426         /* Update potential energies */
1427         gmx_mm256_update_1pot_pd(velecsum,kernel_data->energygrp_elec+ggid);
1428         gmx_mm256_update_1pot_pd(vvdwsum,kernel_data->energygrp_vdw+ggid);
1429
1430         /* Increment number of inner iterations */
1431         inneriter                  += j_index_end - j_index_start;
1432
1433         /* Outer loop uses 20 flops */
1434     }
1435
1436     /* Increment number of outer iterations */
1437     outeriter        += nri;
1438
1439     /* Update outer/inner flops */
1440
1441     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3W3_VF,outeriter*20 + inneriter*441);
1442 }
1443 /*
1444  * Gromacs nonbonded kernel:   nb_kernel_ElecEwSh_VdwLJSh_GeomW3W3_F_avx_256_double
1445  * Electrostatics interaction: Ewald
1446  * VdW interaction:            LennardJones
1447  * Geometry:                   Water3-Water3
1448  * Calculate force/pot:        Force
1449  */
1450 void
1451 nb_kernel_ElecEwSh_VdwLJSh_GeomW3W3_F_avx_256_double
1452                     (t_nblist * gmx_restrict                nlist,
1453                      rvec * gmx_restrict                    xx,
1454                      rvec * gmx_restrict                    ff,
1455                      t_forcerec * gmx_restrict              fr,
1456                      t_mdatoms * gmx_restrict               mdatoms,
1457                      nb_kernel_data_t * gmx_restrict        kernel_data,
1458                      t_nrnb * gmx_restrict                  nrnb)
1459 {
1460     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
1461      * just 0 for non-waters.
1462      * Suffixes A,B,C,D refer to j loop unrolling done with AVX, e.g. for the four different
1463      * jnr indices corresponding to data put in the four positions in the SIMD register.
1464      */
1465     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
1466     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
1467     int              jnrA,jnrB,jnrC,jnrD;
1468     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
1469     int              jnrlistE,jnrlistF,jnrlistG,jnrlistH;
1470     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
1471     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
1472     real             rcutoff_scalar;
1473     real             *shiftvec,*fshift,*x,*f;
1474     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD;
1475     real             scratch[4*DIM];
1476     __m256d          tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
1477     real *           vdwioffsetptr0;
1478     __m256d          ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
1479     real *           vdwioffsetptr1;
1480     __m256d          ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
1481     real *           vdwioffsetptr2;
1482     __m256d          ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
1483     int              vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D;
1484     __m256d          jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
1485     int              vdwjidx1A,vdwjidx1B,vdwjidx1C,vdwjidx1D;
1486     __m256d          jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
1487     int              vdwjidx2A,vdwjidx2B,vdwjidx2C,vdwjidx2D;
1488     __m256d          jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
1489     __m256d          dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
1490     __m256d          dx01,dy01,dz01,rsq01,rinv01,rinvsq01,r01,qq01,c6_01,c12_01;
1491     __m256d          dx02,dy02,dz02,rsq02,rinv02,rinvsq02,r02,qq02,c6_02,c12_02;
1492     __m256d          dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
1493     __m256d          dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11;
1494     __m256d          dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12;
1495     __m256d          dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
1496     __m256d          dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21;
1497     __m256d          dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22;
1498     __m256d          velec,felec,velecsum,facel,crf,krf,krf2;
1499     real             *charge;
1500     int              nvdwtype;
1501     __m256d          rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
1502     int              *vdwtype;
1503     real             *vdwparam;
1504     __m256d          one_sixth   = _mm256_set1_pd(1.0/6.0);
1505     __m256d          one_twelfth = _mm256_set1_pd(1.0/12.0);
1506     __m128i          ewitab;
1507     __m256d          ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
1508     __m256d          beta,beta2,beta3,zeta2,pmecorrF,pmecorrV,rinv3;
1509     real             *ewtab;
1510     __m256d          dummy_mask,cutoff_mask;
1511     __m128           tmpmask0,tmpmask1;
1512     __m256d          signbit = _mm256_castsi256_pd( _mm256_set1_epi32(0x80000000) );
1513     __m256d          one     = _mm256_set1_pd(1.0);
1514     __m256d          two     = _mm256_set1_pd(2.0);
1515     x                = xx[0];
1516     f                = ff[0];
1517
1518     nri              = nlist->nri;
1519     iinr             = nlist->iinr;
1520     jindex           = nlist->jindex;
1521     jjnr             = nlist->jjnr;
1522     shiftidx         = nlist->shift;
1523     gid              = nlist->gid;
1524     shiftvec         = fr->shift_vec[0];
1525     fshift           = fr->fshift[0];
1526     facel            = _mm256_set1_pd(fr->epsfac);
1527     charge           = mdatoms->chargeA;
1528     nvdwtype         = fr->ntype;
1529     vdwparam         = fr->nbfp;
1530     vdwtype          = mdatoms->typeA;
1531
1532     sh_ewald         = _mm256_set1_pd(fr->ic->sh_ewald);
1533     beta             = _mm256_set1_pd(fr->ic->ewaldcoeff);
1534     beta2            = _mm256_mul_pd(beta,beta);
1535     beta3            = _mm256_mul_pd(beta,beta2);
1536
1537     ewtab            = fr->ic->tabq_coul_F;
1538     ewtabscale       = _mm256_set1_pd(fr->ic->tabq_scale);
1539     ewtabhalfspace   = _mm256_set1_pd(0.5/fr->ic->tabq_scale);
1540
1541     /* Setup water-specific parameters */
1542     inr              = nlist->iinr[0];
1543     iq0              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+0]));
1544     iq1              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+1]));
1545     iq2              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+2]));
1546     vdwioffsetptr0   = vdwparam+2*nvdwtype*vdwtype[inr+0];
1547
1548     jq0              = _mm256_set1_pd(charge[inr+0]);
1549     jq1              = _mm256_set1_pd(charge[inr+1]);
1550     jq2              = _mm256_set1_pd(charge[inr+2]);
1551     vdwjidx0A        = 2*vdwtype[inr+0];
1552     qq00             = _mm256_mul_pd(iq0,jq0);
1553     c6_00            = _mm256_set1_pd(vdwioffsetptr0[vdwjidx0A]);
1554     c12_00           = _mm256_set1_pd(vdwioffsetptr0[vdwjidx0A+1]);
1555     qq01             = _mm256_mul_pd(iq0,jq1);
1556     qq02             = _mm256_mul_pd(iq0,jq2);
1557     qq10             = _mm256_mul_pd(iq1,jq0);
1558     qq11             = _mm256_mul_pd(iq1,jq1);
1559     qq12             = _mm256_mul_pd(iq1,jq2);
1560     qq20             = _mm256_mul_pd(iq2,jq0);
1561     qq21             = _mm256_mul_pd(iq2,jq1);
1562     qq22             = _mm256_mul_pd(iq2,jq2);
1563
1564     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
1565     rcutoff_scalar   = fr->rcoulomb;
1566     rcutoff          = _mm256_set1_pd(rcutoff_scalar);
1567     rcutoff2         = _mm256_mul_pd(rcutoff,rcutoff);
1568
1569     sh_vdw_invrcut6  = _mm256_set1_pd(fr->ic->sh_invrc6);
1570     rvdw             = _mm256_set1_pd(fr->rvdw);
1571
1572     /* Avoid stupid compiler warnings */
1573     jnrA = jnrB = jnrC = jnrD = 0;
1574     j_coord_offsetA = 0;
1575     j_coord_offsetB = 0;
1576     j_coord_offsetC = 0;
1577     j_coord_offsetD = 0;
1578
1579     outeriter        = 0;
1580     inneriter        = 0;
1581
1582     for(iidx=0;iidx<4*DIM;iidx++)
1583     {
1584         scratch[iidx] = 0.0;
1585     }
1586
1587     /* Start outer loop over neighborlists */
1588     for(iidx=0; iidx<nri; iidx++)
1589     {
1590         /* Load shift vector for this list */
1591         i_shift_offset   = DIM*shiftidx[iidx];
1592
1593         /* Load limits for loop over neighbors */
1594         j_index_start    = jindex[iidx];
1595         j_index_end      = jindex[iidx+1];
1596
1597         /* Get outer coordinate index */
1598         inr              = iinr[iidx];
1599         i_coord_offset   = DIM*inr;
1600
1601         /* Load i particle coords and add shift vector */
1602         gmx_mm256_load_shift_and_3rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset,
1603                                                     &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2);
1604
1605         fix0             = _mm256_setzero_pd();
1606         fiy0             = _mm256_setzero_pd();
1607         fiz0             = _mm256_setzero_pd();
1608         fix1             = _mm256_setzero_pd();
1609         fiy1             = _mm256_setzero_pd();
1610         fiz1             = _mm256_setzero_pd();
1611         fix2             = _mm256_setzero_pd();
1612         fiy2             = _mm256_setzero_pd();
1613         fiz2             = _mm256_setzero_pd();
1614
1615         /* Start inner kernel loop */
1616         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4)
1617         {
1618
1619             /* Get j neighbor index, and coordinate index */
1620             jnrA             = jjnr[jidx];
1621             jnrB             = jjnr[jidx+1];
1622             jnrC             = jjnr[jidx+2];
1623             jnrD             = jjnr[jidx+3];
1624             j_coord_offsetA  = DIM*jnrA;
1625             j_coord_offsetB  = DIM*jnrB;
1626             j_coord_offsetC  = DIM*jnrC;
1627             j_coord_offsetD  = DIM*jnrD;
1628
1629             /* load j atom coordinates */
1630             gmx_mm256_load_3rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
1631                                                  x+j_coord_offsetC,x+j_coord_offsetD,
1632                                               &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,&jy2,&jz2);
1633
1634             /* Calculate displacement vector */
1635             dx00             = _mm256_sub_pd(ix0,jx0);
1636             dy00             = _mm256_sub_pd(iy0,jy0);
1637             dz00             = _mm256_sub_pd(iz0,jz0);
1638             dx01             = _mm256_sub_pd(ix0,jx1);
1639             dy01             = _mm256_sub_pd(iy0,jy1);
1640             dz01             = _mm256_sub_pd(iz0,jz1);
1641             dx02             = _mm256_sub_pd(ix0,jx2);
1642             dy02             = _mm256_sub_pd(iy0,jy2);
1643             dz02             = _mm256_sub_pd(iz0,jz2);
1644             dx10             = _mm256_sub_pd(ix1,jx0);
1645             dy10             = _mm256_sub_pd(iy1,jy0);
1646             dz10             = _mm256_sub_pd(iz1,jz0);
1647             dx11             = _mm256_sub_pd(ix1,jx1);
1648             dy11             = _mm256_sub_pd(iy1,jy1);
1649             dz11             = _mm256_sub_pd(iz1,jz1);
1650             dx12             = _mm256_sub_pd(ix1,jx2);
1651             dy12             = _mm256_sub_pd(iy1,jy2);
1652             dz12             = _mm256_sub_pd(iz1,jz2);
1653             dx20             = _mm256_sub_pd(ix2,jx0);
1654             dy20             = _mm256_sub_pd(iy2,jy0);
1655             dz20             = _mm256_sub_pd(iz2,jz0);
1656             dx21             = _mm256_sub_pd(ix2,jx1);
1657             dy21             = _mm256_sub_pd(iy2,jy1);
1658             dz21             = _mm256_sub_pd(iz2,jz1);
1659             dx22             = _mm256_sub_pd(ix2,jx2);
1660             dy22             = _mm256_sub_pd(iy2,jy2);
1661             dz22             = _mm256_sub_pd(iz2,jz2);
1662
1663             /* Calculate squared distance and things based on it */
1664             rsq00            = gmx_mm256_calc_rsq_pd(dx00,dy00,dz00);
1665             rsq01            = gmx_mm256_calc_rsq_pd(dx01,dy01,dz01);
1666             rsq02            = gmx_mm256_calc_rsq_pd(dx02,dy02,dz02);
1667             rsq10            = gmx_mm256_calc_rsq_pd(dx10,dy10,dz10);
1668             rsq11            = gmx_mm256_calc_rsq_pd(dx11,dy11,dz11);
1669             rsq12            = gmx_mm256_calc_rsq_pd(dx12,dy12,dz12);
1670             rsq20            = gmx_mm256_calc_rsq_pd(dx20,dy20,dz20);
1671             rsq21            = gmx_mm256_calc_rsq_pd(dx21,dy21,dz21);
1672             rsq22            = gmx_mm256_calc_rsq_pd(dx22,dy22,dz22);
1673
1674             rinv00           = gmx_mm256_invsqrt_pd(rsq00);
1675             rinv01           = gmx_mm256_invsqrt_pd(rsq01);
1676             rinv02           = gmx_mm256_invsqrt_pd(rsq02);
1677             rinv10           = gmx_mm256_invsqrt_pd(rsq10);
1678             rinv11           = gmx_mm256_invsqrt_pd(rsq11);
1679             rinv12           = gmx_mm256_invsqrt_pd(rsq12);
1680             rinv20           = gmx_mm256_invsqrt_pd(rsq20);
1681             rinv21           = gmx_mm256_invsqrt_pd(rsq21);
1682             rinv22           = gmx_mm256_invsqrt_pd(rsq22);
1683
1684             rinvsq00         = _mm256_mul_pd(rinv00,rinv00);
1685             rinvsq01         = _mm256_mul_pd(rinv01,rinv01);
1686             rinvsq02         = _mm256_mul_pd(rinv02,rinv02);
1687             rinvsq10         = _mm256_mul_pd(rinv10,rinv10);
1688             rinvsq11         = _mm256_mul_pd(rinv11,rinv11);
1689             rinvsq12         = _mm256_mul_pd(rinv12,rinv12);
1690             rinvsq20         = _mm256_mul_pd(rinv20,rinv20);
1691             rinvsq21         = _mm256_mul_pd(rinv21,rinv21);
1692             rinvsq22         = _mm256_mul_pd(rinv22,rinv22);
1693
1694             fjx0             = _mm256_setzero_pd();
1695             fjy0             = _mm256_setzero_pd();
1696             fjz0             = _mm256_setzero_pd();
1697             fjx1             = _mm256_setzero_pd();
1698             fjy1             = _mm256_setzero_pd();
1699             fjz1             = _mm256_setzero_pd();
1700             fjx2             = _mm256_setzero_pd();
1701             fjy2             = _mm256_setzero_pd();
1702             fjz2             = _mm256_setzero_pd();
1703
1704             /**************************
1705              * CALCULATE INTERACTIONS *
1706              **************************/
1707
1708             if (gmx_mm256_any_lt(rsq00,rcutoff2))
1709             {
1710
1711             r00              = _mm256_mul_pd(rsq00,rinv00);
1712
1713             /* EWALD ELECTROSTATICS */
1714
1715             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1716             ewrt             = _mm256_mul_pd(r00,ewtabscale);
1717             ewitab           = _mm256_cvttpd_epi32(ewrt);
1718             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
1719             gmx_mm256_load_4pair_swizzle_pd(ewtab + _mm_extract_epi32(ewitab,0),ewtab + _mm_extract_epi32(ewitab,1),
1720                                             ewtab + _mm_extract_epi32(ewitab,2),ewtab + _mm_extract_epi32(ewitab,3),
1721                                             &ewtabF,&ewtabFn);
1722             felec            = _mm256_add_pd(_mm256_mul_pd( _mm256_sub_pd(one,eweps),ewtabF),_mm256_mul_pd(eweps,ewtabFn));
1723             felec            = _mm256_mul_pd(_mm256_mul_pd(qq00,rinv00),_mm256_sub_pd(rinvsq00,felec));
1724
1725             /* LENNARD-JONES DISPERSION/REPULSION */
1726
1727             rinvsix          = _mm256_mul_pd(_mm256_mul_pd(rinvsq00,rinvsq00),rinvsq00);
1728             fvdw             = _mm256_mul_pd(_mm256_sub_pd(_mm256_mul_pd(c12_00,rinvsix),c6_00),_mm256_mul_pd(rinvsix,rinvsq00));
1729
1730             cutoff_mask      = _mm256_cmp_pd(rsq00,rcutoff2,_CMP_LT_OQ);
1731
1732             fscal            = _mm256_add_pd(felec,fvdw);
1733
1734             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1735
1736             /* Calculate temporary vectorial force */
1737             tx               = _mm256_mul_pd(fscal,dx00);
1738             ty               = _mm256_mul_pd(fscal,dy00);
1739             tz               = _mm256_mul_pd(fscal,dz00);
1740
1741             /* Update vectorial force */
1742             fix0             = _mm256_add_pd(fix0,tx);
1743             fiy0             = _mm256_add_pd(fiy0,ty);
1744             fiz0             = _mm256_add_pd(fiz0,tz);
1745
1746             fjx0             = _mm256_add_pd(fjx0,tx);
1747             fjy0             = _mm256_add_pd(fjy0,ty);
1748             fjz0             = _mm256_add_pd(fjz0,tz);
1749
1750             }
1751
1752             /**************************
1753              * CALCULATE INTERACTIONS *
1754              **************************/
1755
1756             if (gmx_mm256_any_lt(rsq01,rcutoff2))
1757             {
1758
1759             r01              = _mm256_mul_pd(rsq01,rinv01);
1760
1761             /* EWALD ELECTROSTATICS */
1762
1763             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1764             ewrt             = _mm256_mul_pd(r01,ewtabscale);
1765             ewitab           = _mm256_cvttpd_epi32(ewrt);
1766             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
1767             gmx_mm256_load_4pair_swizzle_pd(ewtab + _mm_extract_epi32(ewitab,0),ewtab + _mm_extract_epi32(ewitab,1),
1768                                             ewtab + _mm_extract_epi32(ewitab,2),ewtab + _mm_extract_epi32(ewitab,3),
1769                                             &ewtabF,&ewtabFn);
1770             felec            = _mm256_add_pd(_mm256_mul_pd( _mm256_sub_pd(one,eweps),ewtabF),_mm256_mul_pd(eweps,ewtabFn));
1771             felec            = _mm256_mul_pd(_mm256_mul_pd(qq01,rinv01),_mm256_sub_pd(rinvsq01,felec));
1772
1773             cutoff_mask      = _mm256_cmp_pd(rsq01,rcutoff2,_CMP_LT_OQ);
1774
1775             fscal            = felec;
1776
1777             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1778
1779             /* Calculate temporary vectorial force */
1780             tx               = _mm256_mul_pd(fscal,dx01);
1781             ty               = _mm256_mul_pd(fscal,dy01);
1782             tz               = _mm256_mul_pd(fscal,dz01);
1783
1784             /* Update vectorial force */
1785             fix0             = _mm256_add_pd(fix0,tx);
1786             fiy0             = _mm256_add_pd(fiy0,ty);
1787             fiz0             = _mm256_add_pd(fiz0,tz);
1788
1789             fjx1             = _mm256_add_pd(fjx1,tx);
1790             fjy1             = _mm256_add_pd(fjy1,ty);
1791             fjz1             = _mm256_add_pd(fjz1,tz);
1792
1793             }
1794
1795             /**************************
1796              * CALCULATE INTERACTIONS *
1797              **************************/
1798
1799             if (gmx_mm256_any_lt(rsq02,rcutoff2))
1800             {
1801
1802             r02              = _mm256_mul_pd(rsq02,rinv02);
1803
1804             /* EWALD ELECTROSTATICS */
1805
1806             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1807             ewrt             = _mm256_mul_pd(r02,ewtabscale);
1808             ewitab           = _mm256_cvttpd_epi32(ewrt);
1809             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
1810             gmx_mm256_load_4pair_swizzle_pd(ewtab + _mm_extract_epi32(ewitab,0),ewtab + _mm_extract_epi32(ewitab,1),
1811                                             ewtab + _mm_extract_epi32(ewitab,2),ewtab + _mm_extract_epi32(ewitab,3),
1812                                             &ewtabF,&ewtabFn);
1813             felec            = _mm256_add_pd(_mm256_mul_pd( _mm256_sub_pd(one,eweps),ewtabF),_mm256_mul_pd(eweps,ewtabFn));
1814             felec            = _mm256_mul_pd(_mm256_mul_pd(qq02,rinv02),_mm256_sub_pd(rinvsq02,felec));
1815
1816             cutoff_mask      = _mm256_cmp_pd(rsq02,rcutoff2,_CMP_LT_OQ);
1817
1818             fscal            = felec;
1819
1820             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1821
1822             /* Calculate temporary vectorial force */
1823             tx               = _mm256_mul_pd(fscal,dx02);
1824             ty               = _mm256_mul_pd(fscal,dy02);
1825             tz               = _mm256_mul_pd(fscal,dz02);
1826
1827             /* Update vectorial force */
1828             fix0             = _mm256_add_pd(fix0,tx);
1829             fiy0             = _mm256_add_pd(fiy0,ty);
1830             fiz0             = _mm256_add_pd(fiz0,tz);
1831
1832             fjx2             = _mm256_add_pd(fjx2,tx);
1833             fjy2             = _mm256_add_pd(fjy2,ty);
1834             fjz2             = _mm256_add_pd(fjz2,tz);
1835
1836             }
1837
1838             /**************************
1839              * CALCULATE INTERACTIONS *
1840              **************************/
1841
1842             if (gmx_mm256_any_lt(rsq10,rcutoff2))
1843             {
1844
1845             r10              = _mm256_mul_pd(rsq10,rinv10);
1846
1847             /* EWALD ELECTROSTATICS */
1848
1849             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1850             ewrt             = _mm256_mul_pd(r10,ewtabscale);
1851             ewitab           = _mm256_cvttpd_epi32(ewrt);
1852             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
1853             gmx_mm256_load_4pair_swizzle_pd(ewtab + _mm_extract_epi32(ewitab,0),ewtab + _mm_extract_epi32(ewitab,1),
1854                                             ewtab + _mm_extract_epi32(ewitab,2),ewtab + _mm_extract_epi32(ewitab,3),
1855                                             &ewtabF,&ewtabFn);
1856             felec            = _mm256_add_pd(_mm256_mul_pd( _mm256_sub_pd(one,eweps),ewtabF),_mm256_mul_pd(eweps,ewtabFn));
1857             felec            = _mm256_mul_pd(_mm256_mul_pd(qq10,rinv10),_mm256_sub_pd(rinvsq10,felec));
1858
1859             cutoff_mask      = _mm256_cmp_pd(rsq10,rcutoff2,_CMP_LT_OQ);
1860
1861             fscal            = felec;
1862
1863             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1864
1865             /* Calculate temporary vectorial force */
1866             tx               = _mm256_mul_pd(fscal,dx10);
1867             ty               = _mm256_mul_pd(fscal,dy10);
1868             tz               = _mm256_mul_pd(fscal,dz10);
1869
1870             /* Update vectorial force */
1871             fix1             = _mm256_add_pd(fix1,tx);
1872             fiy1             = _mm256_add_pd(fiy1,ty);
1873             fiz1             = _mm256_add_pd(fiz1,tz);
1874
1875             fjx0             = _mm256_add_pd(fjx0,tx);
1876             fjy0             = _mm256_add_pd(fjy0,ty);
1877             fjz0             = _mm256_add_pd(fjz0,tz);
1878
1879             }
1880
1881             /**************************
1882              * CALCULATE INTERACTIONS *
1883              **************************/
1884
1885             if (gmx_mm256_any_lt(rsq11,rcutoff2))
1886             {
1887
1888             r11              = _mm256_mul_pd(rsq11,rinv11);
1889
1890             /* EWALD ELECTROSTATICS */
1891
1892             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1893             ewrt             = _mm256_mul_pd(r11,ewtabscale);
1894             ewitab           = _mm256_cvttpd_epi32(ewrt);
1895             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
1896             gmx_mm256_load_4pair_swizzle_pd(ewtab + _mm_extract_epi32(ewitab,0),ewtab + _mm_extract_epi32(ewitab,1),
1897                                             ewtab + _mm_extract_epi32(ewitab,2),ewtab + _mm_extract_epi32(ewitab,3),
1898                                             &ewtabF,&ewtabFn);
1899             felec            = _mm256_add_pd(_mm256_mul_pd( _mm256_sub_pd(one,eweps),ewtabF),_mm256_mul_pd(eweps,ewtabFn));
1900             felec            = _mm256_mul_pd(_mm256_mul_pd(qq11,rinv11),_mm256_sub_pd(rinvsq11,felec));
1901
1902             cutoff_mask      = _mm256_cmp_pd(rsq11,rcutoff2,_CMP_LT_OQ);
1903
1904             fscal            = felec;
1905
1906             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1907
1908             /* Calculate temporary vectorial force */
1909             tx               = _mm256_mul_pd(fscal,dx11);
1910             ty               = _mm256_mul_pd(fscal,dy11);
1911             tz               = _mm256_mul_pd(fscal,dz11);
1912
1913             /* Update vectorial force */
1914             fix1             = _mm256_add_pd(fix1,tx);
1915             fiy1             = _mm256_add_pd(fiy1,ty);
1916             fiz1             = _mm256_add_pd(fiz1,tz);
1917
1918             fjx1             = _mm256_add_pd(fjx1,tx);
1919             fjy1             = _mm256_add_pd(fjy1,ty);
1920             fjz1             = _mm256_add_pd(fjz1,tz);
1921
1922             }
1923
1924             /**************************
1925              * CALCULATE INTERACTIONS *
1926              **************************/
1927
1928             if (gmx_mm256_any_lt(rsq12,rcutoff2))
1929             {
1930
1931             r12              = _mm256_mul_pd(rsq12,rinv12);
1932
1933             /* EWALD ELECTROSTATICS */
1934
1935             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1936             ewrt             = _mm256_mul_pd(r12,ewtabscale);
1937             ewitab           = _mm256_cvttpd_epi32(ewrt);
1938             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
1939             gmx_mm256_load_4pair_swizzle_pd(ewtab + _mm_extract_epi32(ewitab,0),ewtab + _mm_extract_epi32(ewitab,1),
1940                                             ewtab + _mm_extract_epi32(ewitab,2),ewtab + _mm_extract_epi32(ewitab,3),
1941                                             &ewtabF,&ewtabFn);
1942             felec            = _mm256_add_pd(_mm256_mul_pd( _mm256_sub_pd(one,eweps),ewtabF),_mm256_mul_pd(eweps,ewtabFn));
1943             felec            = _mm256_mul_pd(_mm256_mul_pd(qq12,rinv12),_mm256_sub_pd(rinvsq12,felec));
1944
1945             cutoff_mask      = _mm256_cmp_pd(rsq12,rcutoff2,_CMP_LT_OQ);
1946
1947             fscal            = felec;
1948
1949             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1950
1951             /* Calculate temporary vectorial force */
1952             tx               = _mm256_mul_pd(fscal,dx12);
1953             ty               = _mm256_mul_pd(fscal,dy12);
1954             tz               = _mm256_mul_pd(fscal,dz12);
1955
1956             /* Update vectorial force */
1957             fix1             = _mm256_add_pd(fix1,tx);
1958             fiy1             = _mm256_add_pd(fiy1,ty);
1959             fiz1             = _mm256_add_pd(fiz1,tz);
1960
1961             fjx2             = _mm256_add_pd(fjx2,tx);
1962             fjy2             = _mm256_add_pd(fjy2,ty);
1963             fjz2             = _mm256_add_pd(fjz2,tz);
1964
1965             }
1966
1967             /**************************
1968              * CALCULATE INTERACTIONS *
1969              **************************/
1970
1971             if (gmx_mm256_any_lt(rsq20,rcutoff2))
1972             {
1973
1974             r20              = _mm256_mul_pd(rsq20,rinv20);
1975
1976             /* EWALD ELECTROSTATICS */
1977
1978             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1979             ewrt             = _mm256_mul_pd(r20,ewtabscale);
1980             ewitab           = _mm256_cvttpd_epi32(ewrt);
1981             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
1982             gmx_mm256_load_4pair_swizzle_pd(ewtab + _mm_extract_epi32(ewitab,0),ewtab + _mm_extract_epi32(ewitab,1),
1983                                             ewtab + _mm_extract_epi32(ewitab,2),ewtab + _mm_extract_epi32(ewitab,3),
1984                                             &ewtabF,&ewtabFn);
1985             felec            = _mm256_add_pd(_mm256_mul_pd( _mm256_sub_pd(one,eweps),ewtabF),_mm256_mul_pd(eweps,ewtabFn));
1986             felec            = _mm256_mul_pd(_mm256_mul_pd(qq20,rinv20),_mm256_sub_pd(rinvsq20,felec));
1987
1988             cutoff_mask      = _mm256_cmp_pd(rsq20,rcutoff2,_CMP_LT_OQ);
1989
1990             fscal            = felec;
1991
1992             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1993
1994             /* Calculate temporary vectorial force */
1995             tx               = _mm256_mul_pd(fscal,dx20);
1996             ty               = _mm256_mul_pd(fscal,dy20);
1997             tz               = _mm256_mul_pd(fscal,dz20);
1998
1999             /* Update vectorial force */
2000             fix2             = _mm256_add_pd(fix2,tx);
2001             fiy2             = _mm256_add_pd(fiy2,ty);
2002             fiz2             = _mm256_add_pd(fiz2,tz);
2003
2004             fjx0             = _mm256_add_pd(fjx0,tx);
2005             fjy0             = _mm256_add_pd(fjy0,ty);
2006             fjz0             = _mm256_add_pd(fjz0,tz);
2007
2008             }
2009
2010             /**************************
2011              * CALCULATE INTERACTIONS *
2012              **************************/
2013
2014             if (gmx_mm256_any_lt(rsq21,rcutoff2))
2015             {
2016
2017             r21              = _mm256_mul_pd(rsq21,rinv21);
2018
2019             /* EWALD ELECTROSTATICS */
2020
2021             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
2022             ewrt             = _mm256_mul_pd(r21,ewtabscale);
2023             ewitab           = _mm256_cvttpd_epi32(ewrt);
2024             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
2025             gmx_mm256_load_4pair_swizzle_pd(ewtab + _mm_extract_epi32(ewitab,0),ewtab + _mm_extract_epi32(ewitab,1),
2026                                             ewtab + _mm_extract_epi32(ewitab,2),ewtab + _mm_extract_epi32(ewitab,3),
2027                                             &ewtabF,&ewtabFn);
2028             felec            = _mm256_add_pd(_mm256_mul_pd( _mm256_sub_pd(one,eweps),ewtabF),_mm256_mul_pd(eweps,ewtabFn));
2029             felec            = _mm256_mul_pd(_mm256_mul_pd(qq21,rinv21),_mm256_sub_pd(rinvsq21,felec));
2030
2031             cutoff_mask      = _mm256_cmp_pd(rsq21,rcutoff2,_CMP_LT_OQ);
2032
2033             fscal            = felec;
2034
2035             fscal            = _mm256_and_pd(fscal,cutoff_mask);
2036
2037             /* Calculate temporary vectorial force */
2038             tx               = _mm256_mul_pd(fscal,dx21);
2039             ty               = _mm256_mul_pd(fscal,dy21);
2040             tz               = _mm256_mul_pd(fscal,dz21);
2041
2042             /* Update vectorial force */
2043             fix2             = _mm256_add_pd(fix2,tx);
2044             fiy2             = _mm256_add_pd(fiy2,ty);
2045             fiz2             = _mm256_add_pd(fiz2,tz);
2046
2047             fjx1             = _mm256_add_pd(fjx1,tx);
2048             fjy1             = _mm256_add_pd(fjy1,ty);
2049             fjz1             = _mm256_add_pd(fjz1,tz);
2050
2051             }
2052
2053             /**************************
2054              * CALCULATE INTERACTIONS *
2055              **************************/
2056
2057             if (gmx_mm256_any_lt(rsq22,rcutoff2))
2058             {
2059
2060             r22              = _mm256_mul_pd(rsq22,rinv22);
2061
2062             /* EWALD ELECTROSTATICS */
2063
2064             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
2065             ewrt             = _mm256_mul_pd(r22,ewtabscale);
2066             ewitab           = _mm256_cvttpd_epi32(ewrt);
2067             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
2068             gmx_mm256_load_4pair_swizzle_pd(ewtab + _mm_extract_epi32(ewitab,0),ewtab + _mm_extract_epi32(ewitab,1),
2069                                             ewtab + _mm_extract_epi32(ewitab,2),ewtab + _mm_extract_epi32(ewitab,3),
2070                                             &ewtabF,&ewtabFn);
2071             felec            = _mm256_add_pd(_mm256_mul_pd( _mm256_sub_pd(one,eweps),ewtabF),_mm256_mul_pd(eweps,ewtabFn));
2072             felec            = _mm256_mul_pd(_mm256_mul_pd(qq22,rinv22),_mm256_sub_pd(rinvsq22,felec));
2073
2074             cutoff_mask      = _mm256_cmp_pd(rsq22,rcutoff2,_CMP_LT_OQ);
2075
2076             fscal            = felec;
2077
2078             fscal            = _mm256_and_pd(fscal,cutoff_mask);
2079
2080             /* Calculate temporary vectorial force */
2081             tx               = _mm256_mul_pd(fscal,dx22);
2082             ty               = _mm256_mul_pd(fscal,dy22);
2083             tz               = _mm256_mul_pd(fscal,dz22);
2084
2085             /* Update vectorial force */
2086             fix2             = _mm256_add_pd(fix2,tx);
2087             fiy2             = _mm256_add_pd(fiy2,ty);
2088             fiz2             = _mm256_add_pd(fiz2,tz);
2089
2090             fjx2             = _mm256_add_pd(fjx2,tx);
2091             fjy2             = _mm256_add_pd(fjy2,ty);
2092             fjz2             = _mm256_add_pd(fjz2,tz);
2093
2094             }
2095
2096             fjptrA             = f+j_coord_offsetA;
2097             fjptrB             = f+j_coord_offsetB;
2098             fjptrC             = f+j_coord_offsetC;
2099             fjptrD             = f+j_coord_offsetD;
2100
2101             gmx_mm256_decrement_3rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,
2102                                                       fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2);
2103
2104             /* Inner loop uses 358 flops */
2105         }
2106
2107         if(jidx<j_index_end)
2108         {
2109
2110             /* Get j neighbor index, and coordinate index */
2111             jnrlistA         = jjnr[jidx];
2112             jnrlistB         = jjnr[jidx+1];
2113             jnrlistC         = jjnr[jidx+2];
2114             jnrlistD         = jjnr[jidx+3];
2115             /* Sign of each element will be negative for non-real atoms.
2116              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
2117              * so use it as val = _mm_andnot_pd(mask,val) to clear dummy entries.
2118              */
2119             tmpmask0 = gmx_mm_castsi128_pd(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128()));
2120
2121             tmpmask1 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(3,3,2,2));
2122             tmpmask0 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(1,1,0,0));
2123             dummy_mask = _mm256_castps_pd(gmx_mm256_set_m128(tmpmask1,tmpmask0));
2124
2125             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
2126             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
2127             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
2128             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
2129             j_coord_offsetA  = DIM*jnrA;
2130             j_coord_offsetB  = DIM*jnrB;
2131             j_coord_offsetC  = DIM*jnrC;
2132             j_coord_offsetD  = DIM*jnrD;
2133
2134             /* load j atom coordinates */
2135             gmx_mm256_load_3rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
2136                                                  x+j_coord_offsetC,x+j_coord_offsetD,
2137                                               &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,&jy2,&jz2);
2138
2139             /* Calculate displacement vector */
2140             dx00             = _mm256_sub_pd(ix0,jx0);
2141             dy00             = _mm256_sub_pd(iy0,jy0);
2142             dz00             = _mm256_sub_pd(iz0,jz0);
2143             dx01             = _mm256_sub_pd(ix0,jx1);
2144             dy01             = _mm256_sub_pd(iy0,jy1);
2145             dz01             = _mm256_sub_pd(iz0,jz1);
2146             dx02             = _mm256_sub_pd(ix0,jx2);
2147             dy02             = _mm256_sub_pd(iy0,jy2);
2148             dz02             = _mm256_sub_pd(iz0,jz2);
2149             dx10             = _mm256_sub_pd(ix1,jx0);
2150             dy10             = _mm256_sub_pd(iy1,jy0);
2151             dz10             = _mm256_sub_pd(iz1,jz0);
2152             dx11             = _mm256_sub_pd(ix1,jx1);
2153             dy11             = _mm256_sub_pd(iy1,jy1);
2154             dz11             = _mm256_sub_pd(iz1,jz1);
2155             dx12             = _mm256_sub_pd(ix1,jx2);
2156             dy12             = _mm256_sub_pd(iy1,jy2);
2157             dz12             = _mm256_sub_pd(iz1,jz2);
2158             dx20             = _mm256_sub_pd(ix2,jx0);
2159             dy20             = _mm256_sub_pd(iy2,jy0);
2160             dz20             = _mm256_sub_pd(iz2,jz0);
2161             dx21             = _mm256_sub_pd(ix2,jx1);
2162             dy21             = _mm256_sub_pd(iy2,jy1);
2163             dz21             = _mm256_sub_pd(iz2,jz1);
2164             dx22             = _mm256_sub_pd(ix2,jx2);
2165             dy22             = _mm256_sub_pd(iy2,jy2);
2166             dz22             = _mm256_sub_pd(iz2,jz2);
2167
2168             /* Calculate squared distance and things based on it */
2169             rsq00            = gmx_mm256_calc_rsq_pd(dx00,dy00,dz00);
2170             rsq01            = gmx_mm256_calc_rsq_pd(dx01,dy01,dz01);
2171             rsq02            = gmx_mm256_calc_rsq_pd(dx02,dy02,dz02);
2172             rsq10            = gmx_mm256_calc_rsq_pd(dx10,dy10,dz10);
2173             rsq11            = gmx_mm256_calc_rsq_pd(dx11,dy11,dz11);
2174             rsq12            = gmx_mm256_calc_rsq_pd(dx12,dy12,dz12);
2175             rsq20            = gmx_mm256_calc_rsq_pd(dx20,dy20,dz20);
2176             rsq21            = gmx_mm256_calc_rsq_pd(dx21,dy21,dz21);
2177             rsq22            = gmx_mm256_calc_rsq_pd(dx22,dy22,dz22);
2178
2179             rinv00           = gmx_mm256_invsqrt_pd(rsq00);
2180             rinv01           = gmx_mm256_invsqrt_pd(rsq01);
2181             rinv02           = gmx_mm256_invsqrt_pd(rsq02);
2182             rinv10           = gmx_mm256_invsqrt_pd(rsq10);
2183             rinv11           = gmx_mm256_invsqrt_pd(rsq11);
2184             rinv12           = gmx_mm256_invsqrt_pd(rsq12);
2185             rinv20           = gmx_mm256_invsqrt_pd(rsq20);
2186             rinv21           = gmx_mm256_invsqrt_pd(rsq21);
2187             rinv22           = gmx_mm256_invsqrt_pd(rsq22);
2188
2189             rinvsq00         = _mm256_mul_pd(rinv00,rinv00);
2190             rinvsq01         = _mm256_mul_pd(rinv01,rinv01);
2191             rinvsq02         = _mm256_mul_pd(rinv02,rinv02);
2192             rinvsq10         = _mm256_mul_pd(rinv10,rinv10);
2193             rinvsq11         = _mm256_mul_pd(rinv11,rinv11);
2194             rinvsq12         = _mm256_mul_pd(rinv12,rinv12);
2195             rinvsq20         = _mm256_mul_pd(rinv20,rinv20);
2196             rinvsq21         = _mm256_mul_pd(rinv21,rinv21);
2197             rinvsq22         = _mm256_mul_pd(rinv22,rinv22);
2198
2199             fjx0             = _mm256_setzero_pd();
2200             fjy0             = _mm256_setzero_pd();
2201             fjz0             = _mm256_setzero_pd();
2202             fjx1             = _mm256_setzero_pd();
2203             fjy1             = _mm256_setzero_pd();
2204             fjz1             = _mm256_setzero_pd();
2205             fjx2             = _mm256_setzero_pd();
2206             fjy2             = _mm256_setzero_pd();
2207             fjz2             = _mm256_setzero_pd();
2208
2209             /**************************
2210              * CALCULATE INTERACTIONS *
2211              **************************/
2212
2213             if (gmx_mm256_any_lt(rsq00,rcutoff2))
2214             {
2215
2216             r00              = _mm256_mul_pd(rsq00,rinv00);
2217             r00              = _mm256_andnot_pd(dummy_mask,r00);
2218
2219             /* EWALD ELECTROSTATICS */
2220
2221             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
2222             ewrt             = _mm256_mul_pd(r00,ewtabscale);
2223             ewitab           = _mm256_cvttpd_epi32(ewrt);
2224             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
2225             gmx_mm256_load_4pair_swizzle_pd(ewtab + _mm_extract_epi32(ewitab,0),ewtab + _mm_extract_epi32(ewitab,1),
2226                                             ewtab + _mm_extract_epi32(ewitab,2),ewtab + _mm_extract_epi32(ewitab,3),
2227                                             &ewtabF,&ewtabFn);
2228             felec            = _mm256_add_pd(_mm256_mul_pd( _mm256_sub_pd(one,eweps),ewtabF),_mm256_mul_pd(eweps,ewtabFn));
2229             felec            = _mm256_mul_pd(_mm256_mul_pd(qq00,rinv00),_mm256_sub_pd(rinvsq00,felec));
2230
2231             /* LENNARD-JONES DISPERSION/REPULSION */
2232
2233             rinvsix          = _mm256_mul_pd(_mm256_mul_pd(rinvsq00,rinvsq00),rinvsq00);
2234             fvdw             = _mm256_mul_pd(_mm256_sub_pd(_mm256_mul_pd(c12_00,rinvsix),c6_00),_mm256_mul_pd(rinvsix,rinvsq00));
2235
2236             cutoff_mask      = _mm256_cmp_pd(rsq00,rcutoff2,_CMP_LT_OQ);
2237
2238             fscal            = _mm256_add_pd(felec,fvdw);
2239
2240             fscal            = _mm256_and_pd(fscal,cutoff_mask);
2241
2242             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
2243
2244             /* Calculate temporary vectorial force */
2245             tx               = _mm256_mul_pd(fscal,dx00);
2246             ty               = _mm256_mul_pd(fscal,dy00);
2247             tz               = _mm256_mul_pd(fscal,dz00);
2248
2249             /* Update vectorial force */
2250             fix0             = _mm256_add_pd(fix0,tx);
2251             fiy0             = _mm256_add_pd(fiy0,ty);
2252             fiz0             = _mm256_add_pd(fiz0,tz);
2253
2254             fjx0             = _mm256_add_pd(fjx0,tx);
2255             fjy0             = _mm256_add_pd(fjy0,ty);
2256             fjz0             = _mm256_add_pd(fjz0,tz);
2257
2258             }
2259
2260             /**************************
2261              * CALCULATE INTERACTIONS *
2262              **************************/
2263
2264             if (gmx_mm256_any_lt(rsq01,rcutoff2))
2265             {
2266
2267             r01              = _mm256_mul_pd(rsq01,rinv01);
2268             r01              = _mm256_andnot_pd(dummy_mask,r01);
2269
2270             /* EWALD ELECTROSTATICS */
2271
2272             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
2273             ewrt             = _mm256_mul_pd(r01,ewtabscale);
2274             ewitab           = _mm256_cvttpd_epi32(ewrt);
2275             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
2276             gmx_mm256_load_4pair_swizzle_pd(ewtab + _mm_extract_epi32(ewitab,0),ewtab + _mm_extract_epi32(ewitab,1),
2277                                             ewtab + _mm_extract_epi32(ewitab,2),ewtab + _mm_extract_epi32(ewitab,3),
2278                                             &ewtabF,&ewtabFn);
2279             felec            = _mm256_add_pd(_mm256_mul_pd( _mm256_sub_pd(one,eweps),ewtabF),_mm256_mul_pd(eweps,ewtabFn));
2280             felec            = _mm256_mul_pd(_mm256_mul_pd(qq01,rinv01),_mm256_sub_pd(rinvsq01,felec));
2281
2282             cutoff_mask      = _mm256_cmp_pd(rsq01,rcutoff2,_CMP_LT_OQ);
2283
2284             fscal            = felec;
2285
2286             fscal            = _mm256_and_pd(fscal,cutoff_mask);
2287
2288             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
2289
2290             /* Calculate temporary vectorial force */
2291             tx               = _mm256_mul_pd(fscal,dx01);
2292             ty               = _mm256_mul_pd(fscal,dy01);
2293             tz               = _mm256_mul_pd(fscal,dz01);
2294
2295             /* Update vectorial force */
2296             fix0             = _mm256_add_pd(fix0,tx);
2297             fiy0             = _mm256_add_pd(fiy0,ty);
2298             fiz0             = _mm256_add_pd(fiz0,tz);
2299
2300             fjx1             = _mm256_add_pd(fjx1,tx);
2301             fjy1             = _mm256_add_pd(fjy1,ty);
2302             fjz1             = _mm256_add_pd(fjz1,tz);
2303
2304             }
2305
2306             /**************************
2307              * CALCULATE INTERACTIONS *
2308              **************************/
2309
2310             if (gmx_mm256_any_lt(rsq02,rcutoff2))
2311             {
2312
2313             r02              = _mm256_mul_pd(rsq02,rinv02);
2314             r02              = _mm256_andnot_pd(dummy_mask,r02);
2315
2316             /* EWALD ELECTROSTATICS */
2317
2318             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
2319             ewrt             = _mm256_mul_pd(r02,ewtabscale);
2320             ewitab           = _mm256_cvttpd_epi32(ewrt);
2321             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
2322             gmx_mm256_load_4pair_swizzle_pd(ewtab + _mm_extract_epi32(ewitab,0),ewtab + _mm_extract_epi32(ewitab,1),
2323                                             ewtab + _mm_extract_epi32(ewitab,2),ewtab + _mm_extract_epi32(ewitab,3),
2324                                             &ewtabF,&ewtabFn);
2325             felec            = _mm256_add_pd(_mm256_mul_pd( _mm256_sub_pd(one,eweps),ewtabF),_mm256_mul_pd(eweps,ewtabFn));
2326             felec            = _mm256_mul_pd(_mm256_mul_pd(qq02,rinv02),_mm256_sub_pd(rinvsq02,felec));
2327
2328             cutoff_mask      = _mm256_cmp_pd(rsq02,rcutoff2,_CMP_LT_OQ);
2329
2330             fscal            = felec;
2331
2332             fscal            = _mm256_and_pd(fscal,cutoff_mask);
2333
2334             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
2335
2336             /* Calculate temporary vectorial force */
2337             tx               = _mm256_mul_pd(fscal,dx02);
2338             ty               = _mm256_mul_pd(fscal,dy02);
2339             tz               = _mm256_mul_pd(fscal,dz02);
2340
2341             /* Update vectorial force */
2342             fix0             = _mm256_add_pd(fix0,tx);
2343             fiy0             = _mm256_add_pd(fiy0,ty);
2344             fiz0             = _mm256_add_pd(fiz0,tz);
2345
2346             fjx2             = _mm256_add_pd(fjx2,tx);
2347             fjy2             = _mm256_add_pd(fjy2,ty);
2348             fjz2             = _mm256_add_pd(fjz2,tz);
2349
2350             }
2351
2352             /**************************
2353              * CALCULATE INTERACTIONS *
2354              **************************/
2355
2356             if (gmx_mm256_any_lt(rsq10,rcutoff2))
2357             {
2358
2359             r10              = _mm256_mul_pd(rsq10,rinv10);
2360             r10              = _mm256_andnot_pd(dummy_mask,r10);
2361
2362             /* EWALD ELECTROSTATICS */
2363
2364             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
2365             ewrt             = _mm256_mul_pd(r10,ewtabscale);
2366             ewitab           = _mm256_cvttpd_epi32(ewrt);
2367             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
2368             gmx_mm256_load_4pair_swizzle_pd(ewtab + _mm_extract_epi32(ewitab,0),ewtab + _mm_extract_epi32(ewitab,1),
2369                                             ewtab + _mm_extract_epi32(ewitab,2),ewtab + _mm_extract_epi32(ewitab,3),
2370                                             &ewtabF,&ewtabFn);
2371             felec            = _mm256_add_pd(_mm256_mul_pd( _mm256_sub_pd(one,eweps),ewtabF),_mm256_mul_pd(eweps,ewtabFn));
2372             felec            = _mm256_mul_pd(_mm256_mul_pd(qq10,rinv10),_mm256_sub_pd(rinvsq10,felec));
2373
2374             cutoff_mask      = _mm256_cmp_pd(rsq10,rcutoff2,_CMP_LT_OQ);
2375
2376             fscal            = felec;
2377
2378             fscal            = _mm256_and_pd(fscal,cutoff_mask);
2379
2380             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
2381
2382             /* Calculate temporary vectorial force */
2383             tx               = _mm256_mul_pd(fscal,dx10);
2384             ty               = _mm256_mul_pd(fscal,dy10);
2385             tz               = _mm256_mul_pd(fscal,dz10);
2386
2387             /* Update vectorial force */
2388             fix1             = _mm256_add_pd(fix1,tx);
2389             fiy1             = _mm256_add_pd(fiy1,ty);
2390             fiz1             = _mm256_add_pd(fiz1,tz);
2391
2392             fjx0             = _mm256_add_pd(fjx0,tx);
2393             fjy0             = _mm256_add_pd(fjy0,ty);
2394             fjz0             = _mm256_add_pd(fjz0,tz);
2395
2396             }
2397
2398             /**************************
2399              * CALCULATE INTERACTIONS *
2400              **************************/
2401
2402             if (gmx_mm256_any_lt(rsq11,rcutoff2))
2403             {
2404
2405             r11              = _mm256_mul_pd(rsq11,rinv11);
2406             r11              = _mm256_andnot_pd(dummy_mask,r11);
2407
2408             /* EWALD ELECTROSTATICS */
2409
2410             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
2411             ewrt             = _mm256_mul_pd(r11,ewtabscale);
2412             ewitab           = _mm256_cvttpd_epi32(ewrt);
2413             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
2414             gmx_mm256_load_4pair_swizzle_pd(ewtab + _mm_extract_epi32(ewitab,0),ewtab + _mm_extract_epi32(ewitab,1),
2415                                             ewtab + _mm_extract_epi32(ewitab,2),ewtab + _mm_extract_epi32(ewitab,3),
2416                                             &ewtabF,&ewtabFn);
2417             felec            = _mm256_add_pd(_mm256_mul_pd( _mm256_sub_pd(one,eweps),ewtabF),_mm256_mul_pd(eweps,ewtabFn));
2418             felec            = _mm256_mul_pd(_mm256_mul_pd(qq11,rinv11),_mm256_sub_pd(rinvsq11,felec));
2419
2420             cutoff_mask      = _mm256_cmp_pd(rsq11,rcutoff2,_CMP_LT_OQ);
2421
2422             fscal            = felec;
2423
2424             fscal            = _mm256_and_pd(fscal,cutoff_mask);
2425
2426             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
2427
2428             /* Calculate temporary vectorial force */
2429             tx               = _mm256_mul_pd(fscal,dx11);
2430             ty               = _mm256_mul_pd(fscal,dy11);
2431             tz               = _mm256_mul_pd(fscal,dz11);
2432
2433             /* Update vectorial force */
2434             fix1             = _mm256_add_pd(fix1,tx);
2435             fiy1             = _mm256_add_pd(fiy1,ty);
2436             fiz1             = _mm256_add_pd(fiz1,tz);
2437
2438             fjx1             = _mm256_add_pd(fjx1,tx);
2439             fjy1             = _mm256_add_pd(fjy1,ty);
2440             fjz1             = _mm256_add_pd(fjz1,tz);
2441
2442             }
2443
2444             /**************************
2445              * CALCULATE INTERACTIONS *
2446              **************************/
2447
2448             if (gmx_mm256_any_lt(rsq12,rcutoff2))
2449             {
2450
2451             r12              = _mm256_mul_pd(rsq12,rinv12);
2452             r12              = _mm256_andnot_pd(dummy_mask,r12);
2453
2454             /* EWALD ELECTROSTATICS */
2455
2456             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
2457             ewrt             = _mm256_mul_pd(r12,ewtabscale);
2458             ewitab           = _mm256_cvttpd_epi32(ewrt);
2459             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
2460             gmx_mm256_load_4pair_swizzle_pd(ewtab + _mm_extract_epi32(ewitab,0),ewtab + _mm_extract_epi32(ewitab,1),
2461                                             ewtab + _mm_extract_epi32(ewitab,2),ewtab + _mm_extract_epi32(ewitab,3),
2462                                             &ewtabF,&ewtabFn);
2463             felec            = _mm256_add_pd(_mm256_mul_pd( _mm256_sub_pd(one,eweps),ewtabF),_mm256_mul_pd(eweps,ewtabFn));
2464             felec            = _mm256_mul_pd(_mm256_mul_pd(qq12,rinv12),_mm256_sub_pd(rinvsq12,felec));
2465
2466             cutoff_mask      = _mm256_cmp_pd(rsq12,rcutoff2,_CMP_LT_OQ);
2467
2468             fscal            = felec;
2469
2470             fscal            = _mm256_and_pd(fscal,cutoff_mask);
2471
2472             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
2473
2474             /* Calculate temporary vectorial force */
2475             tx               = _mm256_mul_pd(fscal,dx12);
2476             ty               = _mm256_mul_pd(fscal,dy12);
2477             tz               = _mm256_mul_pd(fscal,dz12);
2478
2479             /* Update vectorial force */
2480             fix1             = _mm256_add_pd(fix1,tx);
2481             fiy1             = _mm256_add_pd(fiy1,ty);
2482             fiz1             = _mm256_add_pd(fiz1,tz);
2483
2484             fjx2             = _mm256_add_pd(fjx2,tx);
2485             fjy2             = _mm256_add_pd(fjy2,ty);
2486             fjz2             = _mm256_add_pd(fjz2,tz);
2487
2488             }
2489
2490             /**************************
2491              * CALCULATE INTERACTIONS *
2492              **************************/
2493
2494             if (gmx_mm256_any_lt(rsq20,rcutoff2))
2495             {
2496
2497             r20              = _mm256_mul_pd(rsq20,rinv20);
2498             r20              = _mm256_andnot_pd(dummy_mask,r20);
2499
2500             /* EWALD ELECTROSTATICS */
2501
2502             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
2503             ewrt             = _mm256_mul_pd(r20,ewtabscale);
2504             ewitab           = _mm256_cvttpd_epi32(ewrt);
2505             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
2506             gmx_mm256_load_4pair_swizzle_pd(ewtab + _mm_extract_epi32(ewitab,0),ewtab + _mm_extract_epi32(ewitab,1),
2507                                             ewtab + _mm_extract_epi32(ewitab,2),ewtab + _mm_extract_epi32(ewitab,3),
2508                                             &ewtabF,&ewtabFn);
2509             felec            = _mm256_add_pd(_mm256_mul_pd( _mm256_sub_pd(one,eweps),ewtabF),_mm256_mul_pd(eweps,ewtabFn));
2510             felec            = _mm256_mul_pd(_mm256_mul_pd(qq20,rinv20),_mm256_sub_pd(rinvsq20,felec));
2511
2512             cutoff_mask      = _mm256_cmp_pd(rsq20,rcutoff2,_CMP_LT_OQ);
2513
2514             fscal            = felec;
2515
2516             fscal            = _mm256_and_pd(fscal,cutoff_mask);
2517
2518             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
2519
2520             /* Calculate temporary vectorial force */
2521             tx               = _mm256_mul_pd(fscal,dx20);
2522             ty               = _mm256_mul_pd(fscal,dy20);
2523             tz               = _mm256_mul_pd(fscal,dz20);
2524
2525             /* Update vectorial force */
2526             fix2             = _mm256_add_pd(fix2,tx);
2527             fiy2             = _mm256_add_pd(fiy2,ty);
2528             fiz2             = _mm256_add_pd(fiz2,tz);
2529
2530             fjx0             = _mm256_add_pd(fjx0,tx);
2531             fjy0             = _mm256_add_pd(fjy0,ty);
2532             fjz0             = _mm256_add_pd(fjz0,tz);
2533
2534             }
2535
2536             /**************************
2537              * CALCULATE INTERACTIONS *
2538              **************************/
2539
2540             if (gmx_mm256_any_lt(rsq21,rcutoff2))
2541             {
2542
2543             r21              = _mm256_mul_pd(rsq21,rinv21);
2544             r21              = _mm256_andnot_pd(dummy_mask,r21);
2545
2546             /* EWALD ELECTROSTATICS */
2547
2548             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
2549             ewrt             = _mm256_mul_pd(r21,ewtabscale);
2550             ewitab           = _mm256_cvttpd_epi32(ewrt);
2551             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
2552             gmx_mm256_load_4pair_swizzle_pd(ewtab + _mm_extract_epi32(ewitab,0),ewtab + _mm_extract_epi32(ewitab,1),
2553                                             ewtab + _mm_extract_epi32(ewitab,2),ewtab + _mm_extract_epi32(ewitab,3),
2554                                             &ewtabF,&ewtabFn);
2555             felec            = _mm256_add_pd(_mm256_mul_pd( _mm256_sub_pd(one,eweps),ewtabF),_mm256_mul_pd(eweps,ewtabFn));
2556             felec            = _mm256_mul_pd(_mm256_mul_pd(qq21,rinv21),_mm256_sub_pd(rinvsq21,felec));
2557
2558             cutoff_mask      = _mm256_cmp_pd(rsq21,rcutoff2,_CMP_LT_OQ);
2559
2560             fscal            = felec;
2561
2562             fscal            = _mm256_and_pd(fscal,cutoff_mask);
2563
2564             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
2565
2566             /* Calculate temporary vectorial force */
2567             tx               = _mm256_mul_pd(fscal,dx21);
2568             ty               = _mm256_mul_pd(fscal,dy21);
2569             tz               = _mm256_mul_pd(fscal,dz21);
2570
2571             /* Update vectorial force */
2572             fix2             = _mm256_add_pd(fix2,tx);
2573             fiy2             = _mm256_add_pd(fiy2,ty);
2574             fiz2             = _mm256_add_pd(fiz2,tz);
2575
2576             fjx1             = _mm256_add_pd(fjx1,tx);
2577             fjy1             = _mm256_add_pd(fjy1,ty);
2578             fjz1             = _mm256_add_pd(fjz1,tz);
2579
2580             }
2581
2582             /**************************
2583              * CALCULATE INTERACTIONS *
2584              **************************/
2585
2586             if (gmx_mm256_any_lt(rsq22,rcutoff2))
2587             {
2588
2589             r22              = _mm256_mul_pd(rsq22,rinv22);
2590             r22              = _mm256_andnot_pd(dummy_mask,r22);
2591
2592             /* EWALD ELECTROSTATICS */
2593
2594             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
2595             ewrt             = _mm256_mul_pd(r22,ewtabscale);
2596             ewitab           = _mm256_cvttpd_epi32(ewrt);
2597             eweps            = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
2598             gmx_mm256_load_4pair_swizzle_pd(ewtab + _mm_extract_epi32(ewitab,0),ewtab + _mm_extract_epi32(ewitab,1),
2599                                             ewtab + _mm_extract_epi32(ewitab,2),ewtab + _mm_extract_epi32(ewitab,3),
2600                                             &ewtabF,&ewtabFn);
2601             felec            = _mm256_add_pd(_mm256_mul_pd( _mm256_sub_pd(one,eweps),ewtabF),_mm256_mul_pd(eweps,ewtabFn));
2602             felec            = _mm256_mul_pd(_mm256_mul_pd(qq22,rinv22),_mm256_sub_pd(rinvsq22,felec));
2603
2604             cutoff_mask      = _mm256_cmp_pd(rsq22,rcutoff2,_CMP_LT_OQ);
2605
2606             fscal            = felec;
2607
2608             fscal            = _mm256_and_pd(fscal,cutoff_mask);
2609
2610             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
2611
2612             /* Calculate temporary vectorial force */
2613             tx               = _mm256_mul_pd(fscal,dx22);
2614             ty               = _mm256_mul_pd(fscal,dy22);
2615             tz               = _mm256_mul_pd(fscal,dz22);
2616
2617             /* Update vectorial force */
2618             fix2             = _mm256_add_pd(fix2,tx);
2619             fiy2             = _mm256_add_pd(fiy2,ty);
2620             fiz2             = _mm256_add_pd(fiz2,tz);
2621
2622             fjx2             = _mm256_add_pd(fjx2,tx);
2623             fjy2             = _mm256_add_pd(fjy2,ty);
2624             fjz2             = _mm256_add_pd(fjz2,tz);
2625
2626             }
2627
2628             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
2629             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
2630             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
2631             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
2632
2633             gmx_mm256_decrement_3rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,
2634                                                       fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2);
2635
2636             /* Inner loop uses 367 flops */
2637         }
2638
2639         /* End of innermost loop */
2640
2641         gmx_mm256_update_iforce_3atom_swizzle_pd(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,
2642                                                  f+i_coord_offset,fshift+i_shift_offset);
2643
2644         /* Increment number of inner iterations */
2645         inneriter                  += j_index_end - j_index_start;
2646
2647         /* Outer loop uses 18 flops */
2648     }
2649
2650     /* Increment number of outer iterations */
2651     outeriter        += nri;
2652
2653     /* Update outer/inner flops */
2654
2655     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3W3_F,outeriter*18 + inneriter*367);
2656 }