src/gromacs/gmxlib/nonbonded/nb_kernel_avx_256_single/nb_kernel_ElecEw_VdwLJEw_GeomW3W3_avx_256_single.cpp

   1 /*
   2  * This file is part of the GROMACS molecular simulation package.
   3  *
   4  * Copyright (c) 2012,2013,2014,2015,2017,2018, by the GROMACS development team, led by
   5  * Mark Abraham, David van der Spoel, Berk Hess, and Erik Lindahl,
   6  * and including many others, as listed in the AUTHORS file in the
   7  * top-level source directory and at http://www.gromacs.org.
   8  *
   9  * GROMACS is free software; you can redistribute it and/or
  10  * modify it under the terms of the GNU Lesser General Public License
  11  * as published by the Free Software Foundation; either version 2.1
  12  * of the License, or (at your option) any later version.
  13  *
  14  * GROMACS is distributed in the hope that it will be useful,
  15  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  16  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  17  * Lesser General Public License for more details.
  18  *
  19  * You should have received a copy of the GNU Lesser General Public
  20  * License along with GROMACS; if not, see
  21  * http://www.gnu.org/licenses, or write to the Free Software Foundation,
  22  * Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA.
  23  *
  24  * If you want to redistribute modifications to GROMACS, please
  25  * consider that scientific software is very special. Version
  26  * control is crucial - bugs must be traceable. We will be happy to
  27  * consider code for inclusion in the official distribution, but
  28  * derived work must not be called official GROMACS. Details are found
  29  * in the README & COPYING files - if they are missing, get the
  30  * official version at http://www.gromacs.org.
  31  *
  32  * To help us fund GROMACS development, we humbly ask that you cite
  33  * the research papers on the package. Check out http://www.gromacs.org.
  34  */
  35 /*
  36  * Note: this file was generated by the GROMACS avx_256_single kernel generator.
  37  */
  38 #include "gmxpre.h"
  39
  40 #include "config.h"
  41
  42 #include <math.h>
  43
  44 #include "../nb_kernel.h"
  45 #include "gromacs/gmxlib/nrnb.h"
  46
  47 #include "kernelutil_x86_avx_256_single.h"
  48
  49 /*
  50  * Gromacs nonbonded kernel:   nb_kernel_ElecEw_VdwLJEw_GeomW3W3_VF_avx_256_single
  51  * Electrostatics interaction: Ewald
  52  * VdW interaction:            LJEwald
  53  * Geometry:                   Water3-Water3
  54  * Calculate force/pot:        PotentialAndForce
  55  */
  56 void
  57 nb_kernel_ElecEw_VdwLJEw_GeomW3W3_VF_avx_256_single
  58                     (t_nblist                    * gmx_restrict       nlist,
  59                      rvec                        * gmx_restrict          xx,
  60                      rvec                        * gmx_restrict          ff,
  61                      struct t_forcerec           * gmx_restrict          fr,
  62                      t_mdatoms                   * gmx_restrict     mdatoms,
  63                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
  64                      t_nrnb                      * gmx_restrict        nrnb)
  65 {
  66     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
  67      * just 0 for non-waters.
  68      * Suffixes A,B,C,D,E,F,G,H refer to j loop unrolling done with AVX, e.g. for the eight different
  69      * jnr indices corresponding to data put in the four positions in the SIMD register.
  70      */
  71     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
  72     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
  73     int              jnrA,jnrB,jnrC,jnrD;
  74     int              jnrE,jnrF,jnrG,jnrH;
  75     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
  76     int              jnrlistE,jnrlistF,jnrlistG,jnrlistH;
  77     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
  78     int              j_coord_offsetE,j_coord_offsetF,j_coord_offsetG,j_coord_offsetH;
  79     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  80     real             rcutoff_scalar;
  81     real             *shiftvec,*fshift,*x,*f;
  82     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD,*fjptrE,*fjptrF,*fjptrG,*fjptrH;
  83     real             scratch[4*DIM];
  84     __m256           tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
  85     real *           vdwioffsetptr0;
  86     real *           vdwgridioffsetptr0;
  87     __m256           ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  88     real *           vdwioffsetptr1;
  89     real *           vdwgridioffsetptr1;
  90     __m256           ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
  91     real *           vdwioffsetptr2;
  92     real *           vdwgridioffsetptr2;
  93     __m256           ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
  94     int              vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D,vdwjidx0E,vdwjidx0F,vdwjidx0G,vdwjidx0H;
  95     __m256           jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  96     int              vdwjidx1A,vdwjidx1B,vdwjidx1C,vdwjidx1D,vdwjidx1E,vdwjidx1F,vdwjidx1G,vdwjidx1H;
  97     __m256           jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
  98     int              vdwjidx2A,vdwjidx2B,vdwjidx2C,vdwjidx2D,vdwjidx2E,vdwjidx2F,vdwjidx2G,vdwjidx2H;
  99     __m256           jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
 100     __m256           dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
 101     __m256           dx01,dy01,dz01,rsq01,rinv01,rinvsq01,r01,qq01,c6_01,c12_01;
 102     __m256           dx02,dy02,dz02,rsq02,rinv02,rinvsq02,r02,qq02,c6_02,c12_02;
 103     __m256           dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
 104     __m256           dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11;
 105     __m256           dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12;
 106     __m256           dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
 107     __m256           dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21;
 108     __m256           dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22;
 109     __m256           velec,felec,velecsum,facel,crf,krf,krf2;
 110     real             *charge;
 111     int              nvdwtype;
 112     __m256           rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
 113     int              *vdwtype;
 114     real             *vdwparam;
 115     __m256           one_sixth   = _mm256_set1_ps(1.0/6.0);
 116     __m256           one_twelfth = _mm256_set1_ps(1.0/12.0);
 117     __m256           c6grid_00;
 118     __m256           c6grid_01;
 119     __m256           c6grid_02;
 120     __m256           c6grid_10;
 121     __m256           c6grid_11;
 122     __m256           c6grid_12;
 123     __m256           c6grid_20;
 124     __m256           c6grid_21;
 125     __m256           c6grid_22;
 126     real             *vdwgridparam;
 127     __m256           ewclj,ewclj2,ewclj6,ewcljrsq,poly,exponent,f6A,f6B,sh_lj_ewald;
 128     __m256           one_half  = _mm256_set1_ps(0.5);
 129     __m256           minus_one = _mm256_set1_ps(-1.0);
 130     __m256i          ewitab;
 131     __m128i          ewitab_lo,ewitab_hi;
 132     __m256           ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
 133     __m256           beta,beta2,beta3,zeta2,pmecorrF,pmecorrV,rinv3;
 134     real             *ewtab;
 135     __m256           dummy_mask,cutoff_mask;
 136     __m256           signbit = _mm256_castsi256_ps( _mm256_set1_epi32(0x80000000) );
 137     __m256           one     = _mm256_set1_ps(1.0);
 138     __m256           two     = _mm256_set1_ps(2.0);
 139     x                = xx[0];
 140     f                = ff[0];
 141
 142     nri              = nlist->nri;
 143     iinr             = nlist->iinr;
 144     jindex           = nlist->jindex;
 145     jjnr             = nlist->jjnr;
 146     shiftidx         = nlist->shift;
 147     gid              = nlist->gid;
 148     shiftvec         = fr->shift_vec[0];
 149     fshift           = fr->fshift[0];
 150     facel            = _mm256_set1_ps(fr->ic->epsfac);
 151     charge           = mdatoms->chargeA;
 152     nvdwtype         = fr->ntype;
 153     vdwparam         = fr->nbfp;
 154     vdwtype          = mdatoms->typeA;
 155     vdwgridparam     = fr->ljpme_c6grid;
 156     sh_lj_ewald      = _mm256_set1_ps(fr->ic->sh_lj_ewald);
 157     ewclj            = _mm256_set1_ps(fr->ic->ewaldcoeff_lj);
 158     ewclj2           = _mm256_mul_ps(minus_one,_mm256_mul_ps(ewclj,ewclj));
 159
 160     sh_ewald         = _mm256_set1_ps(fr->ic->sh_ewald);
 161     beta             = _mm256_set1_ps(fr->ic->ewaldcoeff_q);
 162     beta2            = _mm256_mul_ps(beta,beta);
 163     beta3            = _mm256_mul_ps(beta,beta2);
 164
 165     ewtab            = fr->ic->tabq_coul_FDV0;
 166     ewtabscale       = _mm256_set1_ps(fr->ic->tabq_scale);
 167     ewtabhalfspace   = _mm256_set1_ps(0.5/fr->ic->tabq_scale);
 168
 169     /* Setup water-specific parameters */
 170     inr              = nlist->iinr[0];
 171     iq0              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+0]));
 172     iq1              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+1]));
 173     iq2              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+2]));
 174     vdwioffsetptr0   = vdwparam+2*nvdwtype*vdwtype[inr+0];
 175     vdwgridioffsetptr0 = vdwgridparam+2*nvdwtype*vdwtype[inr+0];
 176
 177     jq0              = _mm256_set1_ps(charge[inr+0]);
 178     jq1              = _mm256_set1_ps(charge[inr+1]);
 179     jq2              = _mm256_set1_ps(charge[inr+2]);
 180     vdwjidx0A        = 2*vdwtype[inr+0];
 181     qq00             = _mm256_mul_ps(iq0,jq0);
 182     c6_00            = _mm256_set1_ps(vdwioffsetptr0[vdwjidx0A]);
 183     c12_00           = _mm256_set1_ps(vdwioffsetptr0[vdwjidx0A+1]);
 184     c6grid_00        = _mm256_set1_ps(vdwgridioffsetptr0[vdwjidx0A]);
 185     qq01             = _mm256_mul_ps(iq0,jq1);
 186     qq02             = _mm256_mul_ps(iq0,jq2);
 187     qq10             = _mm256_mul_ps(iq1,jq0);
 188     qq11             = _mm256_mul_ps(iq1,jq1);
 189     qq12             = _mm256_mul_ps(iq1,jq2);
 190     qq20             = _mm256_mul_ps(iq2,jq0);
 191     qq21             = _mm256_mul_ps(iq2,jq1);
 192     qq22             = _mm256_mul_ps(iq2,jq2);
 193
 194     /* Avoid stupid compiler warnings */
 195     jnrA = jnrB = jnrC = jnrD = jnrE = jnrF = jnrG = jnrH = 0;
 196     j_coord_offsetA = 0;
 197     j_coord_offsetB = 0;
 198     j_coord_offsetC = 0;
 199     j_coord_offsetD = 0;
 200     j_coord_offsetE = 0;
 201     j_coord_offsetF = 0;
 202     j_coord_offsetG = 0;
 203     j_coord_offsetH = 0;
 204
 205     outeriter        = 0;
 206     inneriter        = 0;
 207
 208     for(iidx=0;iidx<4*DIM;iidx++)
 209     {
 210         scratch[iidx] = 0.0;
 211     }
 212
 213     /* Start outer loop over neighborlists */
 214     for(iidx=0; iidx<nri; iidx++)
 215     {
 216         /* Load shift vector for this list */
 217         i_shift_offset   = DIM*shiftidx[iidx];
 218
 219         /* Load limits for loop over neighbors */
 220         j_index_start    = jindex[iidx];
 221         j_index_end      = jindex[iidx+1];
 222
 223         /* Get outer coordinate index */
 224         inr              = iinr[iidx];
 225         i_coord_offset   = DIM*inr;
 226
 227         /* Load i particle coords and add shift vector */
 228         gmx_mm256_load_shift_and_3rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,
 229                                                     &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2);
 230
 231         fix0             = _mm256_setzero_ps();
 232         fiy0             = _mm256_setzero_ps();
 233         fiz0             = _mm256_setzero_ps();
 234         fix1             = _mm256_setzero_ps();
 235         fiy1             = _mm256_setzero_ps();
 236         fiz1             = _mm256_setzero_ps();
 237         fix2             = _mm256_setzero_ps();
 238         fiy2             = _mm256_setzero_ps();
 239         fiz2             = _mm256_setzero_ps();
 240
 241         /* Reset potential sums */
 242         velecsum         = _mm256_setzero_ps();
 243         vvdwsum          = _mm256_setzero_ps();
 244
 245         /* Start inner kernel loop */
 246         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+7]>=0; jidx+=8)
 247         {
 248
 249             /* Get j neighbor index, and coordinate index */
 250             jnrA             = jjnr[jidx];
 251             jnrB             = jjnr[jidx+1];
 252             jnrC             = jjnr[jidx+2];
 253             jnrD             = jjnr[jidx+3];
 254             jnrE             = jjnr[jidx+4];
 255             jnrF             = jjnr[jidx+5];
 256             jnrG             = jjnr[jidx+6];
 257             jnrH             = jjnr[jidx+7];
 258             j_coord_offsetA  = DIM*jnrA;
 259             j_coord_offsetB  = DIM*jnrB;
 260             j_coord_offsetC  = DIM*jnrC;
 261             j_coord_offsetD  = DIM*jnrD;
 262             j_coord_offsetE  = DIM*jnrE;
 263             j_coord_offsetF  = DIM*jnrF;
 264             j_coord_offsetG  = DIM*jnrG;
 265             j_coord_offsetH  = DIM*jnrH;
 266
 267             /* load j atom coordinates */
 268             gmx_mm256_load_3rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 269                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 270                                                  x+j_coord_offsetE,x+j_coord_offsetF,
 271                                                  x+j_coord_offsetG,x+j_coord_offsetH,
 272                                               &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,&jy2,&jz2);
 273
 274             /* Calculate displacement vector */
 275             dx00             = _mm256_sub_ps(ix0,jx0);
 276             dy00             = _mm256_sub_ps(iy0,jy0);
 277             dz00             = _mm256_sub_ps(iz0,jz0);
 278             dx01             = _mm256_sub_ps(ix0,jx1);
 279             dy01             = _mm256_sub_ps(iy0,jy1);
 280             dz01             = _mm256_sub_ps(iz0,jz1);
 281             dx02             = _mm256_sub_ps(ix0,jx2);
 282             dy02             = _mm256_sub_ps(iy0,jy2);
 283             dz02             = _mm256_sub_ps(iz0,jz2);
 284             dx10             = _mm256_sub_ps(ix1,jx0);
 285             dy10             = _mm256_sub_ps(iy1,jy0);
 286             dz10             = _mm256_sub_ps(iz1,jz0);
 287             dx11             = _mm256_sub_ps(ix1,jx1);
 288             dy11             = _mm256_sub_ps(iy1,jy1);
 289             dz11             = _mm256_sub_ps(iz1,jz1);
 290             dx12             = _mm256_sub_ps(ix1,jx2);
 291             dy12             = _mm256_sub_ps(iy1,jy2);
 292             dz12             = _mm256_sub_ps(iz1,jz2);
 293             dx20             = _mm256_sub_ps(ix2,jx0);
 294             dy20             = _mm256_sub_ps(iy2,jy0);
 295             dz20             = _mm256_sub_ps(iz2,jz0);
 296             dx21             = _mm256_sub_ps(ix2,jx1);
 297             dy21             = _mm256_sub_ps(iy2,jy1);
 298             dz21             = _mm256_sub_ps(iz2,jz1);
 299             dx22             = _mm256_sub_ps(ix2,jx2);
 300             dy22             = _mm256_sub_ps(iy2,jy2);
 301             dz22             = _mm256_sub_ps(iz2,jz2);
 302
 303             /* Calculate squared distance and things based on it */
 304             rsq00            = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
 305             rsq01            = gmx_mm256_calc_rsq_ps(dx01,dy01,dz01);
 306             rsq02            = gmx_mm256_calc_rsq_ps(dx02,dy02,dz02);
 307             rsq10            = gmx_mm256_calc_rsq_ps(dx10,dy10,dz10);
 308             rsq11            = gmx_mm256_calc_rsq_ps(dx11,dy11,dz11);
 309             rsq12            = gmx_mm256_calc_rsq_ps(dx12,dy12,dz12);
 310             rsq20            = gmx_mm256_calc_rsq_ps(dx20,dy20,dz20);
 311             rsq21            = gmx_mm256_calc_rsq_ps(dx21,dy21,dz21);
 312             rsq22            = gmx_mm256_calc_rsq_ps(dx22,dy22,dz22);
 313
 314             rinv00           = avx256_invsqrt_f(rsq00);
 315             rinv01           = avx256_invsqrt_f(rsq01);
 316             rinv02           = avx256_invsqrt_f(rsq02);
 317             rinv10           = avx256_invsqrt_f(rsq10);
 318             rinv11           = avx256_invsqrt_f(rsq11);
 319             rinv12           = avx256_invsqrt_f(rsq12);
 320             rinv20           = avx256_invsqrt_f(rsq20);
 321             rinv21           = avx256_invsqrt_f(rsq21);
 322             rinv22           = avx256_invsqrt_f(rsq22);
 323
 324             rinvsq00         = _mm256_mul_ps(rinv00,rinv00);
 325             rinvsq01         = _mm256_mul_ps(rinv01,rinv01);
 326             rinvsq02         = _mm256_mul_ps(rinv02,rinv02);
 327             rinvsq10         = _mm256_mul_ps(rinv10,rinv10);
 328             rinvsq11         = _mm256_mul_ps(rinv11,rinv11);
 329             rinvsq12         = _mm256_mul_ps(rinv12,rinv12);
 330             rinvsq20         = _mm256_mul_ps(rinv20,rinv20);
 331             rinvsq21         = _mm256_mul_ps(rinv21,rinv21);
 332             rinvsq22         = _mm256_mul_ps(rinv22,rinv22);
 333
 334             fjx0             = _mm256_setzero_ps();
 335             fjy0             = _mm256_setzero_ps();
 336             fjz0             = _mm256_setzero_ps();
 337             fjx1             = _mm256_setzero_ps();
 338             fjy1             = _mm256_setzero_ps();
 339             fjz1             = _mm256_setzero_ps();
 340             fjx2             = _mm256_setzero_ps();
 341             fjy2             = _mm256_setzero_ps();
 342             fjz2             = _mm256_setzero_ps();
 343
 344             /**************************
 345              * CALCULATE INTERACTIONS *
 346              **************************/
 347
 348             r00              = _mm256_mul_ps(rsq00,rinv00);
 349
 350             /* EWALD ELECTROSTATICS */
 351
 352             /* Analytical PME correction */
 353             zeta2            = _mm256_mul_ps(beta2,rsq00);
 354             rinv3            = _mm256_mul_ps(rinvsq00,rinv00);
 355             pmecorrF         = avx256_pmecorrF_f(zeta2);
 356             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 357             felec            = _mm256_mul_ps(qq00,felec);
 358             pmecorrV         = avx256_pmecorrV_f(zeta2);
 359             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 360             velec            = _mm256_sub_ps(rinv00,pmecorrV);
 361             velec            = _mm256_mul_ps(qq00,velec);
 362
 363             /* Analytical LJ-PME */
 364             rinvsix          = _mm256_mul_ps(_mm256_mul_ps(rinvsq00,rinvsq00),rinvsq00);
 365             ewcljrsq         = _mm256_mul_ps(ewclj2,rsq00);
 366             ewclj6           = _mm256_mul_ps(ewclj2,_mm256_mul_ps(ewclj2,ewclj2));
 367             exponent         = avx256_exp_f(ewcljrsq);
 368             /* poly = exp(-(beta*r)^2) * (1 + (beta*r)^2 + (beta*r)^4 /2) */
 369             poly             = _mm256_mul_ps(exponent,_mm256_add_ps(_mm256_sub_ps(one,ewcljrsq),_mm256_mul_ps(_mm256_mul_ps(ewcljrsq,ewcljrsq),one_half)));
 370             /* vvdw6 = [C6 - C6grid * (1-poly)]/r6 */
 371             vvdw6            = _mm256_mul_ps(_mm256_sub_ps(c6_00,_mm256_mul_ps(c6grid_00,_mm256_sub_ps(one,poly))),rinvsix);
 372             vvdw12           = _mm256_mul_ps(c12_00,_mm256_mul_ps(rinvsix,rinvsix));
 373             vvdw             = _mm256_sub_ps(_mm256_mul_ps(vvdw12,one_twelfth),_mm256_mul_ps(vvdw6,one_sixth));
 374             /* fvdw = vvdw12/r - (vvdw6/r + (C6grid * exponent * beta^6)/r) */
 375             fvdw             = _mm256_mul_ps(_mm256_sub_ps(vvdw12,_mm256_sub_ps(vvdw6,_mm256_mul_ps(_mm256_mul_ps(c6grid_00,one_sixth),_mm256_mul_ps(exponent,ewclj6)))),rinvsq00);
 376
 377             /* Update potential sum for this i atom from the interaction with this j atom. */
 378             velecsum         = _mm256_add_ps(velecsum,velec);
 379             vvdwsum          = _mm256_add_ps(vvdwsum,vvdw);
 380
 381             fscal            = _mm256_add_ps(felec,fvdw);
 382
 383             /* Calculate temporary vectorial force */
 384             tx               = _mm256_mul_ps(fscal,dx00);
 385             ty               = _mm256_mul_ps(fscal,dy00);
 386             tz               = _mm256_mul_ps(fscal,dz00);
 387
 388             /* Update vectorial force */
 389             fix0             = _mm256_add_ps(fix0,tx);
 390             fiy0             = _mm256_add_ps(fiy0,ty);
 391             fiz0             = _mm256_add_ps(fiz0,tz);
 392
 393             fjx0             = _mm256_add_ps(fjx0,tx);
 394             fjy0             = _mm256_add_ps(fjy0,ty);
 395             fjz0             = _mm256_add_ps(fjz0,tz);
 396
 397             /**************************
 398              * CALCULATE INTERACTIONS *
 399              **************************/
 400
 401             r01              = _mm256_mul_ps(rsq01,rinv01);
 402
 403             /* EWALD ELECTROSTATICS */
 404
 405             /* Analytical PME correction */
 406             zeta2            = _mm256_mul_ps(beta2,rsq01);
 407             rinv3            = _mm256_mul_ps(rinvsq01,rinv01);
 408             pmecorrF         = avx256_pmecorrF_f(zeta2);
 409             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 410             felec            = _mm256_mul_ps(qq01,felec);
 411             pmecorrV         = avx256_pmecorrV_f(zeta2);
 412             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 413             velec            = _mm256_sub_ps(rinv01,pmecorrV);
 414             velec            = _mm256_mul_ps(qq01,velec);
 415
 416             /* Update potential sum for this i atom from the interaction with this j atom. */
 417             velecsum         = _mm256_add_ps(velecsum,velec);
 418
 419             fscal            = felec;
 420
 421             /* Calculate temporary vectorial force */
 422             tx               = _mm256_mul_ps(fscal,dx01);
 423             ty               = _mm256_mul_ps(fscal,dy01);
 424             tz               = _mm256_mul_ps(fscal,dz01);
 425
 426             /* Update vectorial force */
 427             fix0             = _mm256_add_ps(fix0,tx);
 428             fiy0             = _mm256_add_ps(fiy0,ty);
 429             fiz0             = _mm256_add_ps(fiz0,tz);
 430
 431             fjx1             = _mm256_add_ps(fjx1,tx);
 432             fjy1             = _mm256_add_ps(fjy1,ty);
 433             fjz1             = _mm256_add_ps(fjz1,tz);
 434
 435             /**************************
 436              * CALCULATE INTERACTIONS *
 437              **************************/
 438
 439             r02              = _mm256_mul_ps(rsq02,rinv02);
 440
 441             /* EWALD ELECTROSTATICS */
 442
 443             /* Analytical PME correction */
 444             zeta2            = _mm256_mul_ps(beta2,rsq02);
 445             rinv3            = _mm256_mul_ps(rinvsq02,rinv02);
 446             pmecorrF         = avx256_pmecorrF_f(zeta2);
 447             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 448             felec            = _mm256_mul_ps(qq02,felec);
 449             pmecorrV         = avx256_pmecorrV_f(zeta2);
 450             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 451             velec            = _mm256_sub_ps(rinv02,pmecorrV);
 452             velec            = _mm256_mul_ps(qq02,velec);
 453
 454             /* Update potential sum for this i atom from the interaction with this j atom. */
 455             velecsum         = _mm256_add_ps(velecsum,velec);
 456
 457             fscal            = felec;
 458
 459             /* Calculate temporary vectorial force */
 460             tx               = _mm256_mul_ps(fscal,dx02);
 461             ty               = _mm256_mul_ps(fscal,dy02);
 462             tz               = _mm256_mul_ps(fscal,dz02);
 463
 464             /* Update vectorial force */
 465             fix0             = _mm256_add_ps(fix0,tx);
 466             fiy0             = _mm256_add_ps(fiy0,ty);
 467             fiz0             = _mm256_add_ps(fiz0,tz);
 468
 469             fjx2             = _mm256_add_ps(fjx2,tx);
 470             fjy2             = _mm256_add_ps(fjy2,ty);
 471             fjz2             = _mm256_add_ps(fjz2,tz);
 472
 473             /**************************
 474              * CALCULATE INTERACTIONS *
 475              **************************/
 476
 477             r10              = _mm256_mul_ps(rsq10,rinv10);
 478
 479             /* EWALD ELECTROSTATICS */
 480
 481             /* Analytical PME correction */
 482             zeta2            = _mm256_mul_ps(beta2,rsq10);
 483             rinv3            = _mm256_mul_ps(rinvsq10,rinv10);
 484             pmecorrF         = avx256_pmecorrF_f(zeta2);
 485             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 486             felec            = _mm256_mul_ps(qq10,felec);
 487             pmecorrV         = avx256_pmecorrV_f(zeta2);
 488             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 489             velec            = _mm256_sub_ps(rinv10,pmecorrV);
 490             velec            = _mm256_mul_ps(qq10,velec);
 491
 492             /* Update potential sum for this i atom from the interaction with this j atom. */
 493             velecsum         = _mm256_add_ps(velecsum,velec);
 494
 495             fscal            = felec;
 496
 497             /* Calculate temporary vectorial force */
 498             tx               = _mm256_mul_ps(fscal,dx10);
 499             ty               = _mm256_mul_ps(fscal,dy10);
 500             tz               = _mm256_mul_ps(fscal,dz10);
 501
 502             /* Update vectorial force */
 503             fix1             = _mm256_add_ps(fix1,tx);
 504             fiy1             = _mm256_add_ps(fiy1,ty);
 505             fiz1             = _mm256_add_ps(fiz1,tz);
 506
 507             fjx0             = _mm256_add_ps(fjx0,tx);
 508             fjy0             = _mm256_add_ps(fjy0,ty);
 509             fjz0             = _mm256_add_ps(fjz0,tz);
 510
 511             /**************************
 512              * CALCULATE INTERACTIONS *
 513              **************************/
 514
 515             r11              = _mm256_mul_ps(rsq11,rinv11);
 516
 517             /* EWALD ELECTROSTATICS */
 518
 519             /* Analytical PME correction */
 520             zeta2            = _mm256_mul_ps(beta2,rsq11);
 521             rinv3            = _mm256_mul_ps(rinvsq11,rinv11);
 522             pmecorrF         = avx256_pmecorrF_f(zeta2);
 523             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 524             felec            = _mm256_mul_ps(qq11,felec);
 525             pmecorrV         = avx256_pmecorrV_f(zeta2);
 526             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 527             velec            = _mm256_sub_ps(rinv11,pmecorrV);
 528             velec            = _mm256_mul_ps(qq11,velec);
 529
 530             /* Update potential sum for this i atom from the interaction with this j atom. */
 531             velecsum         = _mm256_add_ps(velecsum,velec);
 532
 533             fscal            = felec;
 534
 535             /* Calculate temporary vectorial force */
 536             tx               = _mm256_mul_ps(fscal,dx11);
 537             ty               = _mm256_mul_ps(fscal,dy11);
 538             tz               = _mm256_mul_ps(fscal,dz11);
 539
 540             /* Update vectorial force */
 541             fix1             = _mm256_add_ps(fix1,tx);
 542             fiy1             = _mm256_add_ps(fiy1,ty);
 543             fiz1             = _mm256_add_ps(fiz1,tz);
 544
 545             fjx1             = _mm256_add_ps(fjx1,tx);
 546             fjy1             = _mm256_add_ps(fjy1,ty);
 547             fjz1             = _mm256_add_ps(fjz1,tz);
 548
 549             /**************************
 550              * CALCULATE INTERACTIONS *
 551              **************************/
 552
 553             r12              = _mm256_mul_ps(rsq12,rinv12);
 554
 555             /* EWALD ELECTROSTATICS */
 556
 557             /* Analytical PME correction */
 558             zeta2            = _mm256_mul_ps(beta2,rsq12);
 559             rinv3            = _mm256_mul_ps(rinvsq12,rinv12);
 560             pmecorrF         = avx256_pmecorrF_f(zeta2);
 561             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 562             felec            = _mm256_mul_ps(qq12,felec);
 563             pmecorrV         = avx256_pmecorrV_f(zeta2);
 564             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 565             velec            = _mm256_sub_ps(rinv12,pmecorrV);
 566             velec            = _mm256_mul_ps(qq12,velec);
 567
 568             /* Update potential sum for this i atom from the interaction with this j atom. */
 569             velecsum         = _mm256_add_ps(velecsum,velec);
 570
 571             fscal            = felec;
 572
 573             /* Calculate temporary vectorial force */
 574             tx               = _mm256_mul_ps(fscal,dx12);
 575             ty               = _mm256_mul_ps(fscal,dy12);
 576             tz               = _mm256_mul_ps(fscal,dz12);
 577
 578             /* Update vectorial force */
 579             fix1             = _mm256_add_ps(fix1,tx);
 580             fiy1             = _mm256_add_ps(fiy1,ty);
 581             fiz1             = _mm256_add_ps(fiz1,tz);
 582
 583             fjx2             = _mm256_add_ps(fjx2,tx);
 584             fjy2             = _mm256_add_ps(fjy2,ty);
 585             fjz2             = _mm256_add_ps(fjz2,tz);
 586
 587             /**************************
 588              * CALCULATE INTERACTIONS *
 589              **************************/
 590
 591             r20              = _mm256_mul_ps(rsq20,rinv20);
 592
 593             /* EWALD ELECTROSTATICS */
 594
 595             /* Analytical PME correction */
 596             zeta2            = _mm256_mul_ps(beta2,rsq20);
 597             rinv3            = _mm256_mul_ps(rinvsq20,rinv20);
 598             pmecorrF         = avx256_pmecorrF_f(zeta2);
 599             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 600             felec            = _mm256_mul_ps(qq20,felec);
 601             pmecorrV         = avx256_pmecorrV_f(zeta2);
 602             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 603             velec            = _mm256_sub_ps(rinv20,pmecorrV);
 604             velec            = _mm256_mul_ps(qq20,velec);
 605
 606             /* Update potential sum for this i atom from the interaction with this j atom. */
 607             velecsum         = _mm256_add_ps(velecsum,velec);
 608
 609             fscal            = felec;
 610
 611             /* Calculate temporary vectorial force */
 612             tx               = _mm256_mul_ps(fscal,dx20);
 613             ty               = _mm256_mul_ps(fscal,dy20);
 614             tz               = _mm256_mul_ps(fscal,dz20);
 615
 616             /* Update vectorial force */
 617             fix2             = _mm256_add_ps(fix2,tx);
 618             fiy2             = _mm256_add_ps(fiy2,ty);
 619             fiz2             = _mm256_add_ps(fiz2,tz);
 620
 621             fjx0             = _mm256_add_ps(fjx0,tx);
 622             fjy0             = _mm256_add_ps(fjy0,ty);
 623             fjz0             = _mm256_add_ps(fjz0,tz);
 624
 625             /**************************
 626              * CALCULATE INTERACTIONS *
 627              **************************/
 628
 629             r21              = _mm256_mul_ps(rsq21,rinv21);
 630
 631             /* EWALD ELECTROSTATICS */
 632
 633             /* Analytical PME correction */
 634             zeta2            = _mm256_mul_ps(beta2,rsq21);
 635             rinv3            = _mm256_mul_ps(rinvsq21,rinv21);
 636             pmecorrF         = avx256_pmecorrF_f(zeta2);
 637             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 638             felec            = _mm256_mul_ps(qq21,felec);
 639             pmecorrV         = avx256_pmecorrV_f(zeta2);
 640             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 641             velec            = _mm256_sub_ps(rinv21,pmecorrV);
 642             velec            = _mm256_mul_ps(qq21,velec);
 643
 644             /* Update potential sum for this i atom from the interaction with this j atom. */
 645             velecsum         = _mm256_add_ps(velecsum,velec);
 646
 647             fscal            = felec;
 648
 649             /* Calculate temporary vectorial force */
 650             tx               = _mm256_mul_ps(fscal,dx21);
 651             ty               = _mm256_mul_ps(fscal,dy21);
 652             tz               = _mm256_mul_ps(fscal,dz21);
 653
 654             /* Update vectorial force */
 655             fix2             = _mm256_add_ps(fix2,tx);
 656             fiy2             = _mm256_add_ps(fiy2,ty);
 657             fiz2             = _mm256_add_ps(fiz2,tz);
 658
 659             fjx1             = _mm256_add_ps(fjx1,tx);
 660             fjy1             = _mm256_add_ps(fjy1,ty);
 661             fjz1             = _mm256_add_ps(fjz1,tz);
 662
 663             /**************************
 664              * CALCULATE INTERACTIONS *
 665              **************************/
 666
 667             r22              = _mm256_mul_ps(rsq22,rinv22);
 668
 669             /* EWALD ELECTROSTATICS */
 670
 671             /* Analytical PME correction */
 672             zeta2            = _mm256_mul_ps(beta2,rsq22);
 673             rinv3            = _mm256_mul_ps(rinvsq22,rinv22);
 674             pmecorrF         = avx256_pmecorrF_f(zeta2);
 675             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 676             felec            = _mm256_mul_ps(qq22,felec);
 677             pmecorrV         = avx256_pmecorrV_f(zeta2);
 678             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 679             velec            = _mm256_sub_ps(rinv22,pmecorrV);
 680             velec            = _mm256_mul_ps(qq22,velec);
 681
 682             /* Update potential sum for this i atom from the interaction with this j atom. */
 683             velecsum         = _mm256_add_ps(velecsum,velec);
 684
 685             fscal            = felec;
 686
 687             /* Calculate temporary vectorial force */
 688             tx               = _mm256_mul_ps(fscal,dx22);
 689             ty               = _mm256_mul_ps(fscal,dy22);
 690             tz               = _mm256_mul_ps(fscal,dz22);
 691
 692             /* Update vectorial force */
 693             fix2             = _mm256_add_ps(fix2,tx);
 694             fiy2             = _mm256_add_ps(fiy2,ty);
 695             fiz2             = _mm256_add_ps(fiz2,tz);
 696
 697             fjx2             = _mm256_add_ps(fjx2,tx);
 698             fjy2             = _mm256_add_ps(fjy2,ty);
 699             fjz2             = _mm256_add_ps(fjz2,tz);
 700
 701             fjptrA             = f+j_coord_offsetA;
 702             fjptrB             = f+j_coord_offsetB;
 703             fjptrC             = f+j_coord_offsetC;
 704             fjptrD             = f+j_coord_offsetD;
 705             fjptrE             = f+j_coord_offsetE;
 706             fjptrF             = f+j_coord_offsetF;
 707             fjptrG             = f+j_coord_offsetG;
 708             fjptrH             = f+j_coord_offsetH;
 709
 710             gmx_mm256_decrement_3rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,
 711                                                       fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2);
 712
 713             /* Inner loop uses 784 flops */
 714         }
 715
 716         if(jidx<j_index_end)
 717         {
 718
 719             /* Get j neighbor index, and coordinate index */
 720             jnrlistA         = jjnr[jidx];
 721             jnrlistB         = jjnr[jidx+1];
 722             jnrlistC         = jjnr[jidx+2];
 723             jnrlistD         = jjnr[jidx+3];
 724             jnrlistE         = jjnr[jidx+4];
 725             jnrlistF         = jjnr[jidx+5];
 726             jnrlistG         = jjnr[jidx+6];
 727             jnrlistH         = jjnr[jidx+7];
 728             /* Sign of each element will be negative for non-real atoms.
 729              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
 730              * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
 731              */
 732             dummy_mask = gmx_mm256_set_m128(gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx+4)),_mm_setzero_si128())),
 733                                             gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128())));
 734
 735             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
 736             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
 737             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
 738             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
 739             jnrE       = (jnrlistE>=0) ? jnrlistE : 0;
 740             jnrF       = (jnrlistF>=0) ? jnrlistF : 0;
 741             jnrG       = (jnrlistG>=0) ? jnrlistG : 0;
 742             jnrH       = (jnrlistH>=0) ? jnrlistH : 0;
 743             j_coord_offsetA  = DIM*jnrA;
 744             j_coord_offsetB  = DIM*jnrB;
 745             j_coord_offsetC  = DIM*jnrC;
 746             j_coord_offsetD  = DIM*jnrD;
 747             j_coord_offsetE  = DIM*jnrE;
 748             j_coord_offsetF  = DIM*jnrF;
 749             j_coord_offsetG  = DIM*jnrG;
 750             j_coord_offsetH  = DIM*jnrH;
 751
 752             /* load j atom coordinates */
 753             gmx_mm256_load_3rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 754                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 755                                                  x+j_coord_offsetE,x+j_coord_offsetF,
 756                                                  x+j_coord_offsetG,x+j_coord_offsetH,
 757                                               &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,&jy2,&jz2);
 758
 759             /* Calculate displacement vector */
 760             dx00             = _mm256_sub_ps(ix0,jx0);
 761             dy00             = _mm256_sub_ps(iy0,jy0);
 762             dz00             = _mm256_sub_ps(iz0,jz0);
 763             dx01             = _mm256_sub_ps(ix0,jx1);
 764             dy01             = _mm256_sub_ps(iy0,jy1);
 765             dz01             = _mm256_sub_ps(iz0,jz1);
 766             dx02             = _mm256_sub_ps(ix0,jx2);
 767             dy02             = _mm256_sub_ps(iy0,jy2);
 768             dz02             = _mm256_sub_ps(iz0,jz2);
 769             dx10             = _mm256_sub_ps(ix1,jx0);
 770             dy10             = _mm256_sub_ps(iy1,jy0);
 771             dz10             = _mm256_sub_ps(iz1,jz0);
 772             dx11             = _mm256_sub_ps(ix1,jx1);
 773             dy11             = _mm256_sub_ps(iy1,jy1);
 774             dz11             = _mm256_sub_ps(iz1,jz1);
 775             dx12             = _mm256_sub_ps(ix1,jx2);
 776             dy12             = _mm256_sub_ps(iy1,jy2);
 777             dz12             = _mm256_sub_ps(iz1,jz2);
 778             dx20             = _mm256_sub_ps(ix2,jx0);
 779             dy20             = _mm256_sub_ps(iy2,jy0);
 780             dz20             = _mm256_sub_ps(iz2,jz0);
 781             dx21             = _mm256_sub_ps(ix2,jx1);
 782             dy21             = _mm256_sub_ps(iy2,jy1);
 783             dz21             = _mm256_sub_ps(iz2,jz1);
 784             dx22             = _mm256_sub_ps(ix2,jx2);
 785             dy22             = _mm256_sub_ps(iy2,jy2);
 786             dz22             = _mm256_sub_ps(iz2,jz2);
 787
 788             /* Calculate squared distance and things based on it */
 789             rsq00            = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
 790             rsq01            = gmx_mm256_calc_rsq_ps(dx01,dy01,dz01);
 791             rsq02            = gmx_mm256_calc_rsq_ps(dx02,dy02,dz02);
 792             rsq10            = gmx_mm256_calc_rsq_ps(dx10,dy10,dz10);
 793             rsq11            = gmx_mm256_calc_rsq_ps(dx11,dy11,dz11);
 794             rsq12            = gmx_mm256_calc_rsq_ps(dx12,dy12,dz12);
 795             rsq20            = gmx_mm256_calc_rsq_ps(dx20,dy20,dz20);
 796             rsq21            = gmx_mm256_calc_rsq_ps(dx21,dy21,dz21);
 797             rsq22            = gmx_mm256_calc_rsq_ps(dx22,dy22,dz22);
 798
 799             rinv00           = avx256_invsqrt_f(rsq00);
 800             rinv01           = avx256_invsqrt_f(rsq01);
 801             rinv02           = avx256_invsqrt_f(rsq02);
 802             rinv10           = avx256_invsqrt_f(rsq10);
 803             rinv11           = avx256_invsqrt_f(rsq11);
 804             rinv12           = avx256_invsqrt_f(rsq12);
 805             rinv20           = avx256_invsqrt_f(rsq20);
 806             rinv21           = avx256_invsqrt_f(rsq21);
 807             rinv22           = avx256_invsqrt_f(rsq22);
 808
 809             rinvsq00         = _mm256_mul_ps(rinv00,rinv00);
 810             rinvsq01         = _mm256_mul_ps(rinv01,rinv01);
 811             rinvsq02         = _mm256_mul_ps(rinv02,rinv02);
 812             rinvsq10         = _mm256_mul_ps(rinv10,rinv10);
 813             rinvsq11         = _mm256_mul_ps(rinv11,rinv11);
 814             rinvsq12         = _mm256_mul_ps(rinv12,rinv12);
 815             rinvsq20         = _mm256_mul_ps(rinv20,rinv20);
 816             rinvsq21         = _mm256_mul_ps(rinv21,rinv21);
 817             rinvsq22         = _mm256_mul_ps(rinv22,rinv22);
 818
 819             fjx0             = _mm256_setzero_ps();
 820             fjy0             = _mm256_setzero_ps();
 821             fjz0             = _mm256_setzero_ps();
 822             fjx1             = _mm256_setzero_ps();
 823             fjy1             = _mm256_setzero_ps();
 824             fjz1             = _mm256_setzero_ps();
 825             fjx2             = _mm256_setzero_ps();
 826             fjy2             = _mm256_setzero_ps();
 827             fjz2             = _mm256_setzero_ps();
 828
 829             /**************************
 830              * CALCULATE INTERACTIONS *
 831              **************************/
 832
 833             r00              = _mm256_mul_ps(rsq00,rinv00);
 834             r00              = _mm256_andnot_ps(dummy_mask,r00);
 835
 836             /* EWALD ELECTROSTATICS */
 837
 838             /* Analytical PME correction */
 839             zeta2            = _mm256_mul_ps(beta2,rsq00);
 840             rinv3            = _mm256_mul_ps(rinvsq00,rinv00);
 841             pmecorrF         = avx256_pmecorrF_f(zeta2);
 842             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 843             felec            = _mm256_mul_ps(qq00,felec);
 844             pmecorrV         = avx256_pmecorrV_f(zeta2);
 845             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 846             velec            = _mm256_sub_ps(rinv00,pmecorrV);
 847             velec            = _mm256_mul_ps(qq00,velec);
 848
 849             /* Analytical LJ-PME */
 850             rinvsix          = _mm256_mul_ps(_mm256_mul_ps(rinvsq00,rinvsq00),rinvsq00);
 851             ewcljrsq         = _mm256_mul_ps(ewclj2,rsq00);
 852             ewclj6           = _mm256_mul_ps(ewclj2,_mm256_mul_ps(ewclj2,ewclj2));
 853             exponent         = avx256_exp_f(ewcljrsq);
 854             /* poly = exp(-(beta*r)^2) * (1 + (beta*r)^2 + (beta*r)^4 /2) */
 855             poly             = _mm256_mul_ps(exponent,_mm256_add_ps(_mm256_sub_ps(one,ewcljrsq),_mm256_mul_ps(_mm256_mul_ps(ewcljrsq,ewcljrsq),one_half)));
 856             /* vvdw6 = [C6 - C6grid * (1-poly)]/r6 */
 857             vvdw6            = _mm256_mul_ps(_mm256_sub_ps(c6_00,_mm256_mul_ps(c6grid_00,_mm256_sub_ps(one,poly))),rinvsix);
 858             vvdw12           = _mm256_mul_ps(c12_00,_mm256_mul_ps(rinvsix,rinvsix));
 859             vvdw             = _mm256_sub_ps(_mm256_mul_ps(vvdw12,one_twelfth),_mm256_mul_ps(vvdw6,one_sixth));
 860             /* fvdw = vvdw12/r - (vvdw6/r + (C6grid * exponent * beta^6)/r) */
 861             fvdw             = _mm256_mul_ps(_mm256_sub_ps(vvdw12,_mm256_sub_ps(vvdw6,_mm256_mul_ps(_mm256_mul_ps(c6grid_00,one_sixth),_mm256_mul_ps(exponent,ewclj6)))),rinvsq00);
 862
 863             /* Update potential sum for this i atom from the interaction with this j atom. */
 864             velec            = _mm256_andnot_ps(dummy_mask,velec);
 865             velecsum         = _mm256_add_ps(velecsum,velec);
 866             vvdw             = _mm256_andnot_ps(dummy_mask,vvdw);
 867             vvdwsum          = _mm256_add_ps(vvdwsum,vvdw);
 868
 869             fscal            = _mm256_add_ps(felec,fvdw);
 870
 871             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
 872
 873             /* Calculate temporary vectorial force */
 874             tx               = _mm256_mul_ps(fscal,dx00);
 875             ty               = _mm256_mul_ps(fscal,dy00);
 876             tz               = _mm256_mul_ps(fscal,dz00);
 877
 878             /* Update vectorial force */
 879             fix0             = _mm256_add_ps(fix0,tx);
 880             fiy0             = _mm256_add_ps(fiy0,ty);
 881             fiz0             = _mm256_add_ps(fiz0,tz);
 882
 883             fjx0             = _mm256_add_ps(fjx0,tx);
 884             fjy0             = _mm256_add_ps(fjy0,ty);
 885             fjz0             = _mm256_add_ps(fjz0,tz);
 886
 887             /**************************
 888              * CALCULATE INTERACTIONS *
 889              **************************/
 890
 891             r01              = _mm256_mul_ps(rsq01,rinv01);
 892             r01              = _mm256_andnot_ps(dummy_mask,r01);
 893
 894             /* EWALD ELECTROSTATICS */
 895
 896             /* Analytical PME correction */
 897             zeta2            = _mm256_mul_ps(beta2,rsq01);
 898             rinv3            = _mm256_mul_ps(rinvsq01,rinv01);
 899             pmecorrF         = avx256_pmecorrF_f(zeta2);
 900             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 901             felec            = _mm256_mul_ps(qq01,felec);
 902             pmecorrV         = avx256_pmecorrV_f(zeta2);
 903             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 904             velec            = _mm256_sub_ps(rinv01,pmecorrV);
 905             velec            = _mm256_mul_ps(qq01,velec);
 906
 907             /* Update potential sum for this i atom from the interaction with this j atom. */
 908             velec            = _mm256_andnot_ps(dummy_mask,velec);
 909             velecsum         = _mm256_add_ps(velecsum,velec);
 910
 911             fscal            = felec;
 912
 913             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
 914
 915             /* Calculate temporary vectorial force */
 916             tx               = _mm256_mul_ps(fscal,dx01);
 917             ty               = _mm256_mul_ps(fscal,dy01);
 918             tz               = _mm256_mul_ps(fscal,dz01);
 919
 920             /* Update vectorial force */
 921             fix0             = _mm256_add_ps(fix0,tx);
 922             fiy0             = _mm256_add_ps(fiy0,ty);
 923             fiz0             = _mm256_add_ps(fiz0,tz);
 924
 925             fjx1             = _mm256_add_ps(fjx1,tx);
 926             fjy1             = _mm256_add_ps(fjy1,ty);
 927             fjz1             = _mm256_add_ps(fjz1,tz);
 928
 929             /**************************
 930              * CALCULATE INTERACTIONS *
 931              **************************/
 932
 933             r02              = _mm256_mul_ps(rsq02,rinv02);
 934             r02              = _mm256_andnot_ps(dummy_mask,r02);
 935
 936             /* EWALD ELECTROSTATICS */
 937
 938             /* Analytical PME correction */
 939             zeta2            = _mm256_mul_ps(beta2,rsq02);
 940             rinv3            = _mm256_mul_ps(rinvsq02,rinv02);
 941             pmecorrF         = avx256_pmecorrF_f(zeta2);
 942             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 943             felec            = _mm256_mul_ps(qq02,felec);
 944             pmecorrV         = avx256_pmecorrV_f(zeta2);
 945             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 946             velec            = _mm256_sub_ps(rinv02,pmecorrV);
 947             velec            = _mm256_mul_ps(qq02,velec);
 948
 949             /* Update potential sum for this i atom from the interaction with this j atom. */
 950             velec            = _mm256_andnot_ps(dummy_mask,velec);
 951             velecsum         = _mm256_add_ps(velecsum,velec);
 952
 953             fscal            = felec;
 954
 955             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
 956
 957             /* Calculate temporary vectorial force */
 958             tx               = _mm256_mul_ps(fscal,dx02);
 959             ty               = _mm256_mul_ps(fscal,dy02);
 960             tz               = _mm256_mul_ps(fscal,dz02);
 961
 962             /* Update vectorial force */
 963             fix0             = _mm256_add_ps(fix0,tx);
 964             fiy0             = _mm256_add_ps(fiy0,ty);
 965             fiz0             = _mm256_add_ps(fiz0,tz);
 966
 967             fjx2             = _mm256_add_ps(fjx2,tx);
 968             fjy2             = _mm256_add_ps(fjy2,ty);
 969             fjz2             = _mm256_add_ps(fjz2,tz);
 970
 971             /**************************
 972              * CALCULATE INTERACTIONS *
 973              **************************/
 974
 975             r10              = _mm256_mul_ps(rsq10,rinv10);
 976             r10              = _mm256_andnot_ps(dummy_mask,r10);
 977
 978             /* EWALD ELECTROSTATICS */
 979
 980             /* Analytical PME correction */
 981             zeta2            = _mm256_mul_ps(beta2,rsq10);
 982             rinv3            = _mm256_mul_ps(rinvsq10,rinv10);
 983             pmecorrF         = avx256_pmecorrF_f(zeta2);
 984             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 985             felec            = _mm256_mul_ps(qq10,felec);
 986             pmecorrV         = avx256_pmecorrV_f(zeta2);
 987             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 988             velec            = _mm256_sub_ps(rinv10,pmecorrV);
 989             velec            = _mm256_mul_ps(qq10,velec);
 990
 991             /* Update potential sum for this i atom from the interaction with this j atom. */
 992             velec            = _mm256_andnot_ps(dummy_mask,velec);
 993             velecsum         = _mm256_add_ps(velecsum,velec);
 994
 995             fscal            = felec;
 996
 997             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
 998
 999             /* Calculate temporary vectorial force */
1000             tx               = _mm256_mul_ps(fscal,dx10);
1001             ty               = _mm256_mul_ps(fscal,dy10);
1002             tz               = _mm256_mul_ps(fscal,dz10);
1003
1004             /* Update vectorial force */
1005             fix1             = _mm256_add_ps(fix1,tx);
1006             fiy1             = _mm256_add_ps(fiy1,ty);
1007             fiz1             = _mm256_add_ps(fiz1,tz);
1008
1009             fjx0             = _mm256_add_ps(fjx0,tx);
1010             fjy0             = _mm256_add_ps(fjy0,ty);
1011             fjz0             = _mm256_add_ps(fjz0,tz);
1012
1013             /**************************
1014              * CALCULATE INTERACTIONS *
1015              **************************/
1016
1017             r11              = _mm256_mul_ps(rsq11,rinv11);
1018             r11              = _mm256_andnot_ps(dummy_mask,r11);
1019
1020             /* EWALD ELECTROSTATICS */
1021
1022             /* Analytical PME correction */
1023             zeta2            = _mm256_mul_ps(beta2,rsq11);
1024             rinv3            = _mm256_mul_ps(rinvsq11,rinv11);
1025             pmecorrF         = avx256_pmecorrF_f(zeta2);
1026             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1027             felec            = _mm256_mul_ps(qq11,felec);
1028             pmecorrV         = avx256_pmecorrV_f(zeta2);
1029             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1030             velec            = _mm256_sub_ps(rinv11,pmecorrV);
1031             velec            = _mm256_mul_ps(qq11,velec);
1032
1033             /* Update potential sum for this i atom from the interaction with this j atom. */
1034             velec            = _mm256_andnot_ps(dummy_mask,velec);
1035             velecsum         = _mm256_add_ps(velecsum,velec);
1036
1037             fscal            = felec;
1038
1039             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1040
1041             /* Calculate temporary vectorial force */
1042             tx               = _mm256_mul_ps(fscal,dx11);
1043             ty               = _mm256_mul_ps(fscal,dy11);
1044             tz               = _mm256_mul_ps(fscal,dz11);
1045
1046             /* Update vectorial force */
1047             fix1             = _mm256_add_ps(fix1,tx);
1048             fiy1             = _mm256_add_ps(fiy1,ty);
1049             fiz1             = _mm256_add_ps(fiz1,tz);
1050
1051             fjx1             = _mm256_add_ps(fjx1,tx);
1052             fjy1             = _mm256_add_ps(fjy1,ty);
1053             fjz1             = _mm256_add_ps(fjz1,tz);
1054
1055             /**************************
1056              * CALCULATE INTERACTIONS *
1057              **************************/
1058
1059             r12              = _mm256_mul_ps(rsq12,rinv12);
1060             r12              = _mm256_andnot_ps(dummy_mask,r12);
1061
1062             /* EWALD ELECTROSTATICS */
1063
1064             /* Analytical PME correction */
1065             zeta2            = _mm256_mul_ps(beta2,rsq12);
1066             rinv3            = _mm256_mul_ps(rinvsq12,rinv12);
1067             pmecorrF         = avx256_pmecorrF_f(zeta2);
1068             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1069             felec            = _mm256_mul_ps(qq12,felec);
1070             pmecorrV         = avx256_pmecorrV_f(zeta2);
1071             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1072             velec            = _mm256_sub_ps(rinv12,pmecorrV);
1073             velec            = _mm256_mul_ps(qq12,velec);
1074
1075             /* Update potential sum for this i atom from the interaction with this j atom. */
1076             velec            = _mm256_andnot_ps(dummy_mask,velec);
1077             velecsum         = _mm256_add_ps(velecsum,velec);
1078
1079             fscal            = felec;
1080
1081             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1082
1083             /* Calculate temporary vectorial force */
1084             tx               = _mm256_mul_ps(fscal,dx12);
1085             ty               = _mm256_mul_ps(fscal,dy12);
1086             tz               = _mm256_mul_ps(fscal,dz12);
1087
1088             /* Update vectorial force */
1089             fix1             = _mm256_add_ps(fix1,tx);
1090             fiy1             = _mm256_add_ps(fiy1,ty);
1091             fiz1             = _mm256_add_ps(fiz1,tz);
1092
1093             fjx2             = _mm256_add_ps(fjx2,tx);
1094             fjy2             = _mm256_add_ps(fjy2,ty);
1095             fjz2             = _mm256_add_ps(fjz2,tz);
1096
1097             /**************************
1098              * CALCULATE INTERACTIONS *
1099              **************************/
1100
1101             r20              = _mm256_mul_ps(rsq20,rinv20);
1102             r20              = _mm256_andnot_ps(dummy_mask,r20);
1103
1104             /* EWALD ELECTROSTATICS */
1105
1106             /* Analytical PME correction */
1107             zeta2            = _mm256_mul_ps(beta2,rsq20);
1108             rinv3            = _mm256_mul_ps(rinvsq20,rinv20);
1109             pmecorrF         = avx256_pmecorrF_f(zeta2);
1110             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1111             felec            = _mm256_mul_ps(qq20,felec);
1112             pmecorrV         = avx256_pmecorrV_f(zeta2);
1113             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1114             velec            = _mm256_sub_ps(rinv20,pmecorrV);
1115             velec            = _mm256_mul_ps(qq20,velec);
1116
1117             /* Update potential sum for this i atom from the interaction with this j atom. */
1118             velec            = _mm256_andnot_ps(dummy_mask,velec);
1119             velecsum         = _mm256_add_ps(velecsum,velec);
1120
1121             fscal            = felec;
1122
1123             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1124
1125             /* Calculate temporary vectorial force */
1126             tx               = _mm256_mul_ps(fscal,dx20);
1127             ty               = _mm256_mul_ps(fscal,dy20);
1128             tz               = _mm256_mul_ps(fscal,dz20);
1129
1130             /* Update vectorial force */
1131             fix2             = _mm256_add_ps(fix2,tx);
1132             fiy2             = _mm256_add_ps(fiy2,ty);
1133             fiz2             = _mm256_add_ps(fiz2,tz);
1134
1135             fjx0             = _mm256_add_ps(fjx0,tx);
1136             fjy0             = _mm256_add_ps(fjy0,ty);
1137             fjz0             = _mm256_add_ps(fjz0,tz);
1138
1139             /**************************
1140              * CALCULATE INTERACTIONS *
1141              **************************/
1142
1143             r21              = _mm256_mul_ps(rsq21,rinv21);
1144             r21              = _mm256_andnot_ps(dummy_mask,r21);
1145
1146             /* EWALD ELECTROSTATICS */
1147
1148             /* Analytical PME correction */
1149             zeta2            = _mm256_mul_ps(beta2,rsq21);
1150             rinv3            = _mm256_mul_ps(rinvsq21,rinv21);
1151             pmecorrF         = avx256_pmecorrF_f(zeta2);
1152             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1153             felec            = _mm256_mul_ps(qq21,felec);
1154             pmecorrV         = avx256_pmecorrV_f(zeta2);
1155             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1156             velec            = _mm256_sub_ps(rinv21,pmecorrV);
1157             velec            = _mm256_mul_ps(qq21,velec);
1158
1159             /* Update potential sum for this i atom from the interaction with this j atom. */
1160             velec            = _mm256_andnot_ps(dummy_mask,velec);
1161             velecsum         = _mm256_add_ps(velecsum,velec);
1162
1163             fscal            = felec;
1164
1165             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1166
1167             /* Calculate temporary vectorial force */
1168             tx               = _mm256_mul_ps(fscal,dx21);
1169             ty               = _mm256_mul_ps(fscal,dy21);
1170             tz               = _mm256_mul_ps(fscal,dz21);
1171
1172             /* Update vectorial force */
1173             fix2             = _mm256_add_ps(fix2,tx);
1174             fiy2             = _mm256_add_ps(fiy2,ty);
1175             fiz2             = _mm256_add_ps(fiz2,tz);
1176
1177             fjx1             = _mm256_add_ps(fjx1,tx);
1178             fjy1             = _mm256_add_ps(fjy1,ty);
1179             fjz1             = _mm256_add_ps(fjz1,tz);
1180
1181             /**************************
1182              * CALCULATE INTERACTIONS *
1183              **************************/
1184
1185             r22              = _mm256_mul_ps(rsq22,rinv22);
1186             r22              = _mm256_andnot_ps(dummy_mask,r22);
1187
1188             /* EWALD ELECTROSTATICS */
1189
1190             /* Analytical PME correction */
1191             zeta2            = _mm256_mul_ps(beta2,rsq22);
1192             rinv3            = _mm256_mul_ps(rinvsq22,rinv22);
1193             pmecorrF         = avx256_pmecorrF_f(zeta2);
1194             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1195             felec            = _mm256_mul_ps(qq22,felec);
1196             pmecorrV         = avx256_pmecorrV_f(zeta2);
1197             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1198             velec            = _mm256_sub_ps(rinv22,pmecorrV);
1199             velec            = _mm256_mul_ps(qq22,velec);
1200
1201             /* Update potential sum for this i atom from the interaction with this j atom. */
1202             velec            = _mm256_andnot_ps(dummy_mask,velec);
1203             velecsum         = _mm256_add_ps(velecsum,velec);
1204
1205             fscal            = felec;
1206
1207             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1208
1209             /* Calculate temporary vectorial force */
1210             tx               = _mm256_mul_ps(fscal,dx22);
1211             ty               = _mm256_mul_ps(fscal,dy22);
1212             tz               = _mm256_mul_ps(fscal,dz22);
1213
1214             /* Update vectorial force */
1215             fix2             = _mm256_add_ps(fix2,tx);
1216             fiy2             = _mm256_add_ps(fiy2,ty);
1217             fiz2             = _mm256_add_ps(fiz2,tz);
1218
1219             fjx2             = _mm256_add_ps(fjx2,tx);
1220             fjy2             = _mm256_add_ps(fjy2,ty);
1221             fjz2             = _mm256_add_ps(fjz2,tz);
1222
1223             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
1224             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
1225             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
1226             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
1227             fjptrE             = (jnrlistE>=0) ? f+j_coord_offsetE : scratch;
1228             fjptrF             = (jnrlistF>=0) ? f+j_coord_offsetF : scratch;
1229             fjptrG             = (jnrlistG>=0) ? f+j_coord_offsetG : scratch;
1230             fjptrH             = (jnrlistH>=0) ? f+j_coord_offsetH : scratch;
1231
1232             gmx_mm256_decrement_3rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,
1233                                                       fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2);
1234
1235             /* Inner loop uses 793 flops */
1236         }
1237
1238         /* End of innermost loop */
1239
1240         gmx_mm256_update_iforce_3atom_swizzle_ps(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,
1241                                                  f+i_coord_offset,fshift+i_shift_offset);
1242
1243         ggid                        = gid[iidx];
1244         /* Update potential energies */
1245         gmx_mm256_update_1pot_ps(velecsum,kernel_data->energygrp_elec+ggid);
1246         gmx_mm256_update_1pot_ps(vvdwsum,kernel_data->energygrp_vdw+ggid);
1247
1248         /* Increment number of inner iterations */
1249         inneriter                  += j_index_end - j_index_start;
1250
1251         /* Outer loop uses 20 flops */
1252     }
1253
1254     /* Increment number of outer iterations */
1255     outeriter        += nri;
1256
1257     /* Update outer/inner flops */
1258
1259     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3W3_VF,outeriter*20 + inneriter*793);
1260 }
1261 /*
1262  * Gromacs nonbonded kernel:   nb_kernel_ElecEw_VdwLJEw_GeomW3W3_F_avx_256_single
1263  * Electrostatics interaction: Ewald
1264  * VdW interaction:            LJEwald
1265  * Geometry:                   Water3-Water3
1266  * Calculate force/pot:        Force
1267  */
1268 void
1269 nb_kernel_ElecEw_VdwLJEw_GeomW3W3_F_avx_256_single
1270                     (t_nblist                    * gmx_restrict       nlist,
1271                      rvec                        * gmx_restrict          xx,
1272                      rvec                        * gmx_restrict          ff,
1273                      struct t_forcerec           * gmx_restrict          fr,
1274                      t_mdatoms                   * gmx_restrict     mdatoms,
1275                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
1276                      t_nrnb                      * gmx_restrict        nrnb)
1277 {
1278     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
1279      * just 0 for non-waters.
1280      * Suffixes A,B,C,D,E,F,G,H refer to j loop unrolling done with AVX, e.g. for the eight different
1281      * jnr indices corresponding to data put in the four positions in the SIMD register.
1282      */
1283     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
1284     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
1285     int              jnrA,jnrB,jnrC,jnrD;
1286     int              jnrE,jnrF,jnrG,jnrH;
1287     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
1288     int              jnrlistE,jnrlistF,jnrlistG,jnrlistH;
1289     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
1290     int              j_coord_offsetE,j_coord_offsetF,j_coord_offsetG,j_coord_offsetH;
1291     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
1292     real             rcutoff_scalar;
1293     real             *shiftvec,*fshift,*x,*f;
1294     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD,*fjptrE,*fjptrF,*fjptrG,*fjptrH;
1295     real             scratch[4*DIM];
1296     __m256           tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
1297     real *           vdwioffsetptr0;
1298     real *           vdwgridioffsetptr0;
1299     __m256           ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
1300     real *           vdwioffsetptr1;
1301     real *           vdwgridioffsetptr1;
1302     __m256           ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
1303     real *           vdwioffsetptr2;
1304     real *           vdwgridioffsetptr2;
1305     __m256           ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
1306     int              vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D,vdwjidx0E,vdwjidx0F,vdwjidx0G,vdwjidx0H;
1307     __m256           jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
1308     int              vdwjidx1A,vdwjidx1B,vdwjidx1C,vdwjidx1D,vdwjidx1E,vdwjidx1F,vdwjidx1G,vdwjidx1H;
1309     __m256           jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
1310     int              vdwjidx2A,vdwjidx2B,vdwjidx2C,vdwjidx2D,vdwjidx2E,vdwjidx2F,vdwjidx2G,vdwjidx2H;
1311     __m256           jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
1312     __m256           dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
1313     __m256           dx01,dy01,dz01,rsq01,rinv01,rinvsq01,r01,qq01,c6_01,c12_01;
1314     __m256           dx02,dy02,dz02,rsq02,rinv02,rinvsq02,r02,qq02,c6_02,c12_02;
1315     __m256           dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
1316     __m256           dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11;
1317     __m256           dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12;
1318     __m256           dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
1319     __m256           dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21;
1320     __m256           dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22;
1321     __m256           velec,felec,velecsum,facel,crf,krf,krf2;
1322     real             *charge;
1323     int              nvdwtype;
1324     __m256           rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
1325     int              *vdwtype;
1326     real             *vdwparam;
1327     __m256           one_sixth   = _mm256_set1_ps(1.0/6.0);
1328     __m256           one_twelfth = _mm256_set1_ps(1.0/12.0);
1329     __m256           c6grid_00;
1330     __m256           c6grid_01;
1331     __m256           c6grid_02;
1332     __m256           c6grid_10;
1333     __m256           c6grid_11;
1334     __m256           c6grid_12;
1335     __m256           c6grid_20;
1336     __m256           c6grid_21;
1337     __m256           c6grid_22;
1338     real             *vdwgridparam;
1339     __m256           ewclj,ewclj2,ewclj6,ewcljrsq,poly,exponent,f6A,f6B,sh_lj_ewald;
1340     __m256           one_half  = _mm256_set1_ps(0.5);
1341     __m256           minus_one = _mm256_set1_ps(-1.0);
1342     __m256i          ewitab;
1343     __m128i          ewitab_lo,ewitab_hi;
1344     __m256           ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
1345     __m256           beta,beta2,beta3,zeta2,pmecorrF,pmecorrV,rinv3;
1346     real             *ewtab;
1347     __m256           dummy_mask,cutoff_mask;
1348     __m256           signbit = _mm256_castsi256_ps( _mm256_set1_epi32(0x80000000) );
1349     __m256           one     = _mm256_set1_ps(1.0);
1350     __m256           two     = _mm256_set1_ps(2.0);
1351     x                = xx[0];
1352     f                = ff[0];
1353
1354     nri              = nlist->nri;
1355     iinr             = nlist->iinr;
1356     jindex           = nlist->jindex;
1357     jjnr             = nlist->jjnr;
1358     shiftidx         = nlist->shift;
1359     gid              = nlist->gid;
1360     shiftvec         = fr->shift_vec[0];
1361     fshift           = fr->fshift[0];
1362     facel            = _mm256_set1_ps(fr->ic->epsfac);
1363     charge           = mdatoms->chargeA;
1364     nvdwtype         = fr->ntype;
1365     vdwparam         = fr->nbfp;
1366     vdwtype          = mdatoms->typeA;
1367     vdwgridparam     = fr->ljpme_c6grid;
1368     sh_lj_ewald      = _mm256_set1_ps(fr->ic->sh_lj_ewald);
1369     ewclj            = _mm256_set1_ps(fr->ic->ewaldcoeff_lj);
1370     ewclj2           = _mm256_mul_ps(minus_one,_mm256_mul_ps(ewclj,ewclj));
1371
1372     sh_ewald         = _mm256_set1_ps(fr->ic->sh_ewald);
1373     beta             = _mm256_set1_ps(fr->ic->ewaldcoeff_q);
1374     beta2            = _mm256_mul_ps(beta,beta);
1375     beta3            = _mm256_mul_ps(beta,beta2);
1376
1377     ewtab            = fr->ic->tabq_coul_F;
1378     ewtabscale       = _mm256_set1_ps(fr->ic->tabq_scale);
1379     ewtabhalfspace   = _mm256_set1_ps(0.5/fr->ic->tabq_scale);
1380
1381     /* Setup water-specific parameters */
1382     inr              = nlist->iinr[0];
1383     iq0              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+0]));
1384     iq1              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+1]));
1385     iq2              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+2]));
1386     vdwioffsetptr0   = vdwparam+2*nvdwtype*vdwtype[inr+0];
1387     vdwgridioffsetptr0 = vdwgridparam+2*nvdwtype*vdwtype[inr+0];
1388
1389     jq0              = _mm256_set1_ps(charge[inr+0]);
1390     jq1              = _mm256_set1_ps(charge[inr+1]);
1391     jq2              = _mm256_set1_ps(charge[inr+2]);
1392     vdwjidx0A        = 2*vdwtype[inr+0];
1393     qq00             = _mm256_mul_ps(iq0,jq0);
1394     c6_00            = _mm256_set1_ps(vdwioffsetptr0[vdwjidx0A]);
1395     c12_00           = _mm256_set1_ps(vdwioffsetptr0[vdwjidx0A+1]);
1396     c6grid_00        = _mm256_set1_ps(vdwgridioffsetptr0[vdwjidx0A]);
1397     qq01             = _mm256_mul_ps(iq0,jq1);
1398     qq02             = _mm256_mul_ps(iq0,jq2);
1399     qq10             = _mm256_mul_ps(iq1,jq0);
1400     qq11             = _mm256_mul_ps(iq1,jq1);
1401     qq12             = _mm256_mul_ps(iq1,jq2);
1402     qq20             = _mm256_mul_ps(iq2,jq0);
1403     qq21             = _mm256_mul_ps(iq2,jq1);
1404     qq22             = _mm256_mul_ps(iq2,jq2);
1405
1406     /* Avoid stupid compiler warnings */
1407     jnrA = jnrB = jnrC = jnrD = jnrE = jnrF = jnrG = jnrH = 0;
1408     j_coord_offsetA = 0;
1409     j_coord_offsetB = 0;
1410     j_coord_offsetC = 0;
1411     j_coord_offsetD = 0;
1412     j_coord_offsetE = 0;
1413     j_coord_offsetF = 0;
1414     j_coord_offsetG = 0;
1415     j_coord_offsetH = 0;
1416
1417     outeriter        = 0;
1418     inneriter        = 0;
1419
1420     for(iidx=0;iidx<4*DIM;iidx++)
1421     {
1422         scratch[iidx] = 0.0;
1423     }
1424
1425     /* Start outer loop over neighborlists */
1426     for(iidx=0; iidx<nri; iidx++)
1427     {
1428         /* Load shift vector for this list */
1429         i_shift_offset   = DIM*shiftidx[iidx];
1430
1431         /* Load limits for loop over neighbors */
1432         j_index_start    = jindex[iidx];
1433         j_index_end      = jindex[iidx+1];
1434
1435         /* Get outer coordinate index */
1436         inr              = iinr[iidx];
1437         i_coord_offset   = DIM*inr;
1438
1439         /* Load i particle coords and add shift vector */
1440         gmx_mm256_load_shift_and_3rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,
1441                                                     &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2);
1442
1443         fix0             = _mm256_setzero_ps();
1444         fiy0             = _mm256_setzero_ps();
1445         fiz0             = _mm256_setzero_ps();
1446         fix1             = _mm256_setzero_ps();
1447         fiy1             = _mm256_setzero_ps();
1448         fiz1             = _mm256_setzero_ps();
1449         fix2             = _mm256_setzero_ps();
1450         fiy2             = _mm256_setzero_ps();
1451         fiz2             = _mm256_setzero_ps();
1452
1453         /* Start inner kernel loop */
1454         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+7]>=0; jidx+=8)
1455         {
1456
1457             /* Get j neighbor index, and coordinate index */
1458             jnrA             = jjnr[jidx];
1459             jnrB             = jjnr[jidx+1];
1460             jnrC             = jjnr[jidx+2];
1461             jnrD             = jjnr[jidx+3];
1462             jnrE             = jjnr[jidx+4];
1463             jnrF             = jjnr[jidx+5];
1464             jnrG             = jjnr[jidx+6];
1465             jnrH             = jjnr[jidx+7];
1466             j_coord_offsetA  = DIM*jnrA;
1467             j_coord_offsetB  = DIM*jnrB;
1468             j_coord_offsetC  = DIM*jnrC;
1469             j_coord_offsetD  = DIM*jnrD;
1470             j_coord_offsetE  = DIM*jnrE;
1471             j_coord_offsetF  = DIM*jnrF;
1472             j_coord_offsetG  = DIM*jnrG;
1473             j_coord_offsetH  = DIM*jnrH;
1474
1475             /* load j atom coordinates */
1476             gmx_mm256_load_3rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
1477                                                  x+j_coord_offsetC,x+j_coord_offsetD,
1478                                                  x+j_coord_offsetE,x+j_coord_offsetF,
1479                                                  x+j_coord_offsetG,x+j_coord_offsetH,
1480                                               &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,&jy2,&jz2);
1481
1482             /* Calculate displacement vector */
1483             dx00             = _mm256_sub_ps(ix0,jx0);
1484             dy00             = _mm256_sub_ps(iy0,jy0);
1485             dz00             = _mm256_sub_ps(iz0,jz0);
1486             dx01             = _mm256_sub_ps(ix0,jx1);
1487             dy01             = _mm256_sub_ps(iy0,jy1);
1488             dz01             = _mm256_sub_ps(iz0,jz1);
1489             dx02             = _mm256_sub_ps(ix0,jx2);
1490             dy02             = _mm256_sub_ps(iy0,jy2);
1491             dz02             = _mm256_sub_ps(iz0,jz2);
1492             dx10             = _mm256_sub_ps(ix1,jx0);
1493             dy10             = _mm256_sub_ps(iy1,jy0);
1494             dz10             = _mm256_sub_ps(iz1,jz0);
1495             dx11             = _mm256_sub_ps(ix1,jx1);
1496             dy11             = _mm256_sub_ps(iy1,jy1);
1497             dz11             = _mm256_sub_ps(iz1,jz1);
1498             dx12             = _mm256_sub_ps(ix1,jx2);
1499             dy12             = _mm256_sub_ps(iy1,jy2);
1500             dz12             = _mm256_sub_ps(iz1,jz2);
1501             dx20             = _mm256_sub_ps(ix2,jx0);
1502             dy20             = _mm256_sub_ps(iy2,jy0);
1503             dz20             = _mm256_sub_ps(iz2,jz0);
1504             dx21             = _mm256_sub_ps(ix2,jx1);
1505             dy21             = _mm256_sub_ps(iy2,jy1);
1506             dz21             = _mm256_sub_ps(iz2,jz1);
1507             dx22             = _mm256_sub_ps(ix2,jx2);
1508             dy22             = _mm256_sub_ps(iy2,jy2);
1509             dz22             = _mm256_sub_ps(iz2,jz2);
1510
1511             /* Calculate squared distance and things based on it */
1512             rsq00            = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
1513             rsq01            = gmx_mm256_calc_rsq_ps(dx01,dy01,dz01);
1514             rsq02            = gmx_mm256_calc_rsq_ps(dx02,dy02,dz02);
1515             rsq10            = gmx_mm256_calc_rsq_ps(dx10,dy10,dz10);
1516             rsq11            = gmx_mm256_calc_rsq_ps(dx11,dy11,dz11);
1517             rsq12            = gmx_mm256_calc_rsq_ps(dx12,dy12,dz12);
1518             rsq20            = gmx_mm256_calc_rsq_ps(dx20,dy20,dz20);
1519             rsq21            = gmx_mm256_calc_rsq_ps(dx21,dy21,dz21);
1520             rsq22            = gmx_mm256_calc_rsq_ps(dx22,dy22,dz22);
1521
1522             rinv00           = avx256_invsqrt_f(rsq00);
1523             rinv01           = avx256_invsqrt_f(rsq01);
1524             rinv02           = avx256_invsqrt_f(rsq02);
1525             rinv10           = avx256_invsqrt_f(rsq10);
1526             rinv11           = avx256_invsqrt_f(rsq11);
1527             rinv12           = avx256_invsqrt_f(rsq12);
1528             rinv20           = avx256_invsqrt_f(rsq20);
1529             rinv21           = avx256_invsqrt_f(rsq21);
1530             rinv22           = avx256_invsqrt_f(rsq22);
1531
1532             rinvsq00         = _mm256_mul_ps(rinv00,rinv00);
1533             rinvsq01         = _mm256_mul_ps(rinv01,rinv01);
1534             rinvsq02         = _mm256_mul_ps(rinv02,rinv02);
1535             rinvsq10         = _mm256_mul_ps(rinv10,rinv10);
1536             rinvsq11         = _mm256_mul_ps(rinv11,rinv11);
1537             rinvsq12         = _mm256_mul_ps(rinv12,rinv12);
1538             rinvsq20         = _mm256_mul_ps(rinv20,rinv20);
1539             rinvsq21         = _mm256_mul_ps(rinv21,rinv21);
1540             rinvsq22         = _mm256_mul_ps(rinv22,rinv22);
1541
1542             fjx0             = _mm256_setzero_ps();
1543             fjy0             = _mm256_setzero_ps();
1544             fjz0             = _mm256_setzero_ps();
1545             fjx1             = _mm256_setzero_ps();
1546             fjy1             = _mm256_setzero_ps();
1547             fjz1             = _mm256_setzero_ps();
1548             fjx2             = _mm256_setzero_ps();
1549             fjy2             = _mm256_setzero_ps();
1550             fjz2             = _mm256_setzero_ps();
1551
1552             /**************************
1553              * CALCULATE INTERACTIONS *
1554              **************************/
1555
1556             r00              = _mm256_mul_ps(rsq00,rinv00);
1557
1558             /* EWALD ELECTROSTATICS */
1559
1560             /* Analytical PME correction */
1561             zeta2            = _mm256_mul_ps(beta2,rsq00);
1562             rinv3            = _mm256_mul_ps(rinvsq00,rinv00);
1563             pmecorrF         = avx256_pmecorrF_f(zeta2);
1564             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1565             felec            = _mm256_mul_ps(qq00,felec);
1566
1567             /* Analytical LJ-PME */
1568             rinvsix          = _mm256_mul_ps(_mm256_mul_ps(rinvsq00,rinvsq00),rinvsq00);
1569             ewcljrsq         = _mm256_mul_ps(ewclj2,rsq00);
1570             ewclj6           = _mm256_mul_ps(ewclj2,_mm256_mul_ps(ewclj2,ewclj2));
1571             exponent         = avx256_exp_f(ewcljrsq);
1572             /* poly = exp(-(beta*r)^2) * (1 + (beta*r)^2 + (beta*r)^4 /2) */
1573             poly             = _mm256_mul_ps(exponent,_mm256_add_ps(_mm256_sub_ps(one,ewcljrsq),_mm256_mul_ps(_mm256_mul_ps(ewcljrsq,ewcljrsq),one_half)));
1574             /* f6A = 6 * C6grid * (1 - poly) */
1575             f6A              = _mm256_mul_ps(c6grid_00,_mm256_sub_ps(one,poly));
1576             /* f6B = C6grid * exponent * beta^6 */
1577             f6B              = _mm256_mul_ps(_mm256_mul_ps(c6grid_00,one_sixth),_mm256_mul_ps(exponent,ewclj6));
1578             /* fvdw = 12*C12/r13 - ((6*C6 - f6A)/r6 + f6B)/r */
1579             fvdw              = _mm256_mul_ps(_mm256_add_ps(_mm256_mul_ps(_mm256_sub_ps(_mm256_mul_ps(c12_00,rinvsix),_mm256_sub_ps(c6_00,f6A)),rinvsix),f6B),rinvsq00);
1580
1581             fscal            = _mm256_add_ps(felec,fvdw);
1582
1583             /* Calculate temporary vectorial force */
1584             tx               = _mm256_mul_ps(fscal,dx00);
1585             ty               = _mm256_mul_ps(fscal,dy00);
1586             tz               = _mm256_mul_ps(fscal,dz00);
1587
1588             /* Update vectorial force */
1589             fix0             = _mm256_add_ps(fix0,tx);
1590             fiy0             = _mm256_add_ps(fiy0,ty);
1591             fiz0             = _mm256_add_ps(fiz0,tz);
1592
1593             fjx0             = _mm256_add_ps(fjx0,tx);
1594             fjy0             = _mm256_add_ps(fjy0,ty);
1595             fjz0             = _mm256_add_ps(fjz0,tz);
1596
1597             /**************************
1598              * CALCULATE INTERACTIONS *
1599              **************************/
1600
1601             r01              = _mm256_mul_ps(rsq01,rinv01);
1602
1603             /* EWALD ELECTROSTATICS */
1604
1605             /* Analytical PME correction */
1606             zeta2            = _mm256_mul_ps(beta2,rsq01);
1607             rinv3            = _mm256_mul_ps(rinvsq01,rinv01);
1608             pmecorrF         = avx256_pmecorrF_f(zeta2);
1609             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1610             felec            = _mm256_mul_ps(qq01,felec);
1611
1612             fscal            = felec;
1613
1614             /* Calculate temporary vectorial force */
1615             tx               = _mm256_mul_ps(fscal,dx01);
1616             ty               = _mm256_mul_ps(fscal,dy01);
1617             tz               = _mm256_mul_ps(fscal,dz01);
1618
1619             /* Update vectorial force */
1620             fix0             = _mm256_add_ps(fix0,tx);
1621             fiy0             = _mm256_add_ps(fiy0,ty);
1622             fiz0             = _mm256_add_ps(fiz0,tz);
1623
1624             fjx1             = _mm256_add_ps(fjx1,tx);
1625             fjy1             = _mm256_add_ps(fjy1,ty);
1626             fjz1             = _mm256_add_ps(fjz1,tz);
1627
1628             /**************************
1629              * CALCULATE INTERACTIONS *
1630              **************************/
1631
1632             r02              = _mm256_mul_ps(rsq02,rinv02);
1633
1634             /* EWALD ELECTROSTATICS */
1635
1636             /* Analytical PME correction */
1637             zeta2            = _mm256_mul_ps(beta2,rsq02);
1638             rinv3            = _mm256_mul_ps(rinvsq02,rinv02);
1639             pmecorrF         = avx256_pmecorrF_f(zeta2);
1640             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1641             felec            = _mm256_mul_ps(qq02,felec);
1642
1643             fscal            = felec;
1644
1645             /* Calculate temporary vectorial force */
1646             tx               = _mm256_mul_ps(fscal,dx02);
1647             ty               = _mm256_mul_ps(fscal,dy02);
1648             tz               = _mm256_mul_ps(fscal,dz02);
1649
1650             /* Update vectorial force */
1651             fix0             = _mm256_add_ps(fix0,tx);
1652             fiy0             = _mm256_add_ps(fiy0,ty);
1653             fiz0             = _mm256_add_ps(fiz0,tz);
1654
1655             fjx2             = _mm256_add_ps(fjx2,tx);
1656             fjy2             = _mm256_add_ps(fjy2,ty);
1657             fjz2             = _mm256_add_ps(fjz2,tz);
1658
1659             /**************************
1660              * CALCULATE INTERACTIONS *
1661              **************************/
1662
1663             r10              = _mm256_mul_ps(rsq10,rinv10);
1664
1665             /* EWALD ELECTROSTATICS */
1666
1667             /* Analytical PME correction */
1668             zeta2            = _mm256_mul_ps(beta2,rsq10);
1669             rinv3            = _mm256_mul_ps(rinvsq10,rinv10);
1670             pmecorrF         = avx256_pmecorrF_f(zeta2);
1671             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1672             felec            = _mm256_mul_ps(qq10,felec);
1673
1674             fscal            = felec;
1675
1676             /* Calculate temporary vectorial force */
1677             tx               = _mm256_mul_ps(fscal,dx10);
1678             ty               = _mm256_mul_ps(fscal,dy10);
1679             tz               = _mm256_mul_ps(fscal,dz10);
1680
1681             /* Update vectorial force */
1682             fix1             = _mm256_add_ps(fix1,tx);
1683             fiy1             = _mm256_add_ps(fiy1,ty);
1684             fiz1             = _mm256_add_ps(fiz1,tz);
1685
1686             fjx0             = _mm256_add_ps(fjx0,tx);
1687             fjy0             = _mm256_add_ps(fjy0,ty);
1688             fjz0             = _mm256_add_ps(fjz0,tz);
1689
1690             /**************************
1691              * CALCULATE INTERACTIONS *
1692              **************************/
1693
1694             r11              = _mm256_mul_ps(rsq11,rinv11);
1695
1696             /* EWALD ELECTROSTATICS */
1697
1698             /* Analytical PME correction */
1699             zeta2            = _mm256_mul_ps(beta2,rsq11);
1700             rinv3            = _mm256_mul_ps(rinvsq11,rinv11);
1701             pmecorrF         = avx256_pmecorrF_f(zeta2);
1702             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1703             felec            = _mm256_mul_ps(qq11,felec);
1704
1705             fscal            = felec;
1706
1707             /* Calculate temporary vectorial force */
1708             tx               = _mm256_mul_ps(fscal,dx11);
1709             ty               = _mm256_mul_ps(fscal,dy11);
1710             tz               = _mm256_mul_ps(fscal,dz11);
1711
1712             /* Update vectorial force */
1713             fix1             = _mm256_add_ps(fix1,tx);
1714             fiy1             = _mm256_add_ps(fiy1,ty);
1715             fiz1             = _mm256_add_ps(fiz1,tz);
1716
1717             fjx1             = _mm256_add_ps(fjx1,tx);
1718             fjy1             = _mm256_add_ps(fjy1,ty);
1719             fjz1             = _mm256_add_ps(fjz1,tz);
1720
1721             /**************************
1722              * CALCULATE INTERACTIONS *
1723              **************************/
1724
1725             r12              = _mm256_mul_ps(rsq12,rinv12);
1726
1727             /* EWALD ELECTROSTATICS */
1728
1729             /* Analytical PME correction */
1730             zeta2            = _mm256_mul_ps(beta2,rsq12);
1731             rinv3            = _mm256_mul_ps(rinvsq12,rinv12);
1732             pmecorrF         = avx256_pmecorrF_f(zeta2);
1733             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1734             felec            = _mm256_mul_ps(qq12,felec);
1735
1736             fscal            = felec;
1737
1738             /* Calculate temporary vectorial force */
1739             tx               = _mm256_mul_ps(fscal,dx12);
1740             ty               = _mm256_mul_ps(fscal,dy12);
1741             tz               = _mm256_mul_ps(fscal,dz12);
1742
1743             /* Update vectorial force */
1744             fix1             = _mm256_add_ps(fix1,tx);
1745             fiy1             = _mm256_add_ps(fiy1,ty);
1746             fiz1             = _mm256_add_ps(fiz1,tz);
1747
1748             fjx2             = _mm256_add_ps(fjx2,tx);
1749             fjy2             = _mm256_add_ps(fjy2,ty);
1750             fjz2             = _mm256_add_ps(fjz2,tz);
1751
1752             /**************************
1753              * CALCULATE INTERACTIONS *
1754              **************************/
1755
1756             r20              = _mm256_mul_ps(rsq20,rinv20);
1757
1758             /* EWALD ELECTROSTATICS */
1759
1760             /* Analytical PME correction */
1761             zeta2            = _mm256_mul_ps(beta2,rsq20);
1762             rinv3            = _mm256_mul_ps(rinvsq20,rinv20);
1763             pmecorrF         = avx256_pmecorrF_f(zeta2);
1764             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1765             felec            = _mm256_mul_ps(qq20,felec);
1766
1767             fscal            = felec;
1768
1769             /* Calculate temporary vectorial force */
1770             tx               = _mm256_mul_ps(fscal,dx20);
1771             ty               = _mm256_mul_ps(fscal,dy20);
1772             tz               = _mm256_mul_ps(fscal,dz20);
1773
1774             /* Update vectorial force */
1775             fix2             = _mm256_add_ps(fix2,tx);
1776             fiy2             = _mm256_add_ps(fiy2,ty);
1777             fiz2             = _mm256_add_ps(fiz2,tz);
1778
1779             fjx0             = _mm256_add_ps(fjx0,tx);
1780             fjy0             = _mm256_add_ps(fjy0,ty);
1781             fjz0             = _mm256_add_ps(fjz0,tz);
1782
1783             /**************************
1784              * CALCULATE INTERACTIONS *
1785              **************************/
1786
1787             r21              = _mm256_mul_ps(rsq21,rinv21);
1788
1789             /* EWALD ELECTROSTATICS */
1790
1791             /* Analytical PME correction */
1792             zeta2            = _mm256_mul_ps(beta2,rsq21);
1793             rinv3            = _mm256_mul_ps(rinvsq21,rinv21);
1794             pmecorrF         = avx256_pmecorrF_f(zeta2);
1795             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1796             felec            = _mm256_mul_ps(qq21,felec);
1797
1798             fscal            = felec;
1799
1800             /* Calculate temporary vectorial force */
1801             tx               = _mm256_mul_ps(fscal,dx21);
1802             ty               = _mm256_mul_ps(fscal,dy21);
1803             tz               = _mm256_mul_ps(fscal,dz21);
1804
1805             /* Update vectorial force */
1806             fix2             = _mm256_add_ps(fix2,tx);
1807             fiy2             = _mm256_add_ps(fiy2,ty);
1808             fiz2             = _mm256_add_ps(fiz2,tz);
1809
1810             fjx1             = _mm256_add_ps(fjx1,tx);
1811             fjy1             = _mm256_add_ps(fjy1,ty);
1812             fjz1             = _mm256_add_ps(fjz1,tz);
1813
1814             /**************************
1815              * CALCULATE INTERACTIONS *
1816              **************************/
1817
1818             r22              = _mm256_mul_ps(rsq22,rinv22);
1819
1820             /* EWALD ELECTROSTATICS */
1821
1822             /* Analytical PME correction */
1823             zeta2            = _mm256_mul_ps(beta2,rsq22);
1824             rinv3            = _mm256_mul_ps(rinvsq22,rinv22);
1825             pmecorrF         = avx256_pmecorrF_f(zeta2);
1826             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1827             felec            = _mm256_mul_ps(qq22,felec);
1828
1829             fscal            = felec;
1830
1831             /* Calculate temporary vectorial force */
1832             tx               = _mm256_mul_ps(fscal,dx22);
1833             ty               = _mm256_mul_ps(fscal,dy22);
1834             tz               = _mm256_mul_ps(fscal,dz22);
1835
1836             /* Update vectorial force */
1837             fix2             = _mm256_add_ps(fix2,tx);
1838             fiy2             = _mm256_add_ps(fiy2,ty);
1839             fiz2             = _mm256_add_ps(fiz2,tz);
1840
1841             fjx2             = _mm256_add_ps(fjx2,tx);
1842             fjy2             = _mm256_add_ps(fjy2,ty);
1843             fjz2             = _mm256_add_ps(fjz2,tz);
1844
1845             fjptrA             = f+j_coord_offsetA;
1846             fjptrB             = f+j_coord_offsetB;
1847             fjptrC             = f+j_coord_offsetC;
1848             fjptrD             = f+j_coord_offsetD;
1849             fjptrE             = f+j_coord_offsetE;
1850             fjptrF             = f+j_coord_offsetF;
1851             fjptrG             = f+j_coord_offsetG;
1852             fjptrH             = f+j_coord_offsetH;
1853
1854             gmx_mm256_decrement_3rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,
1855                                                       fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2);
1856
1857             /* Inner loop uses 527 flops */
1858         }
1859
1860         if(jidx<j_index_end)
1861         {
1862
1863             /* Get j neighbor index, and coordinate index */
1864             jnrlistA         = jjnr[jidx];
1865             jnrlistB         = jjnr[jidx+1];
1866             jnrlistC         = jjnr[jidx+2];
1867             jnrlistD         = jjnr[jidx+3];
1868             jnrlistE         = jjnr[jidx+4];
1869             jnrlistF         = jjnr[jidx+5];
1870             jnrlistG         = jjnr[jidx+6];
1871             jnrlistH         = jjnr[jidx+7];
1872             /* Sign of each element will be negative for non-real atoms.
1873              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
1874              * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
1875              */
1876             dummy_mask = gmx_mm256_set_m128(gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx+4)),_mm_setzero_si128())),
1877                                             gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128())));
1878
1879             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
1880             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
1881             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
1882             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
1883             jnrE       = (jnrlistE>=0) ? jnrlistE : 0;
1884             jnrF       = (jnrlistF>=0) ? jnrlistF : 0;
1885             jnrG       = (jnrlistG>=0) ? jnrlistG : 0;
1886             jnrH       = (jnrlistH>=0) ? jnrlistH : 0;
1887             j_coord_offsetA  = DIM*jnrA;
1888             j_coord_offsetB  = DIM*jnrB;
1889             j_coord_offsetC  = DIM*jnrC;
1890             j_coord_offsetD  = DIM*jnrD;
1891             j_coord_offsetE  = DIM*jnrE;
1892             j_coord_offsetF  = DIM*jnrF;
1893             j_coord_offsetG  = DIM*jnrG;
1894             j_coord_offsetH  = DIM*jnrH;
1895
1896             /* load j atom coordinates */
1897             gmx_mm256_load_3rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
1898                                                  x+j_coord_offsetC,x+j_coord_offsetD,
1899                                                  x+j_coord_offsetE,x+j_coord_offsetF,
1900                                                  x+j_coord_offsetG,x+j_coord_offsetH,
1901                                               &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,&jy2,&jz2);
1902
1903             /* Calculate displacement vector */
1904             dx00             = _mm256_sub_ps(ix0,jx0);
1905             dy00             = _mm256_sub_ps(iy0,jy0);
1906             dz00             = _mm256_sub_ps(iz0,jz0);
1907             dx01             = _mm256_sub_ps(ix0,jx1);
1908             dy01             = _mm256_sub_ps(iy0,jy1);
1909             dz01             = _mm256_sub_ps(iz0,jz1);
1910             dx02             = _mm256_sub_ps(ix0,jx2);
1911             dy02             = _mm256_sub_ps(iy0,jy2);
1912             dz02             = _mm256_sub_ps(iz0,jz2);
1913             dx10             = _mm256_sub_ps(ix1,jx0);
1914             dy10             = _mm256_sub_ps(iy1,jy0);
1915             dz10             = _mm256_sub_ps(iz1,jz0);
1916             dx11             = _mm256_sub_ps(ix1,jx1);
1917             dy11             = _mm256_sub_ps(iy1,jy1);
1918             dz11             = _mm256_sub_ps(iz1,jz1);
1919             dx12             = _mm256_sub_ps(ix1,jx2);
1920             dy12             = _mm256_sub_ps(iy1,jy2);
1921             dz12             = _mm256_sub_ps(iz1,jz2);
1922             dx20             = _mm256_sub_ps(ix2,jx0);
1923             dy20             = _mm256_sub_ps(iy2,jy0);
1924             dz20             = _mm256_sub_ps(iz2,jz0);
1925             dx21             = _mm256_sub_ps(ix2,jx1);
1926             dy21             = _mm256_sub_ps(iy2,jy1);
1927             dz21             = _mm256_sub_ps(iz2,jz1);
1928             dx22             = _mm256_sub_ps(ix2,jx2);
1929             dy22             = _mm256_sub_ps(iy2,jy2);
1930             dz22             = _mm256_sub_ps(iz2,jz2);
1931
1932             /* Calculate squared distance and things based on it */
1933             rsq00            = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
1934             rsq01            = gmx_mm256_calc_rsq_ps(dx01,dy01,dz01);
1935             rsq02            = gmx_mm256_calc_rsq_ps(dx02,dy02,dz02);
1936             rsq10            = gmx_mm256_calc_rsq_ps(dx10,dy10,dz10);
1937             rsq11            = gmx_mm256_calc_rsq_ps(dx11,dy11,dz11);
1938             rsq12            = gmx_mm256_calc_rsq_ps(dx12,dy12,dz12);
1939             rsq20            = gmx_mm256_calc_rsq_ps(dx20,dy20,dz20);
1940             rsq21            = gmx_mm256_calc_rsq_ps(dx21,dy21,dz21);
1941             rsq22            = gmx_mm256_calc_rsq_ps(dx22,dy22,dz22);
1942
1943             rinv00           = avx256_invsqrt_f(rsq00);
1944             rinv01           = avx256_invsqrt_f(rsq01);
1945             rinv02           = avx256_invsqrt_f(rsq02);
1946             rinv10           = avx256_invsqrt_f(rsq10);
1947             rinv11           = avx256_invsqrt_f(rsq11);
1948             rinv12           = avx256_invsqrt_f(rsq12);
1949             rinv20           = avx256_invsqrt_f(rsq20);
1950             rinv21           = avx256_invsqrt_f(rsq21);
1951             rinv22           = avx256_invsqrt_f(rsq22);
1952
1953             rinvsq00         = _mm256_mul_ps(rinv00,rinv00);
1954             rinvsq01         = _mm256_mul_ps(rinv01,rinv01);
1955             rinvsq02         = _mm256_mul_ps(rinv02,rinv02);
1956             rinvsq10         = _mm256_mul_ps(rinv10,rinv10);
1957             rinvsq11         = _mm256_mul_ps(rinv11,rinv11);
1958             rinvsq12         = _mm256_mul_ps(rinv12,rinv12);
1959             rinvsq20         = _mm256_mul_ps(rinv20,rinv20);
1960             rinvsq21         = _mm256_mul_ps(rinv21,rinv21);
1961             rinvsq22         = _mm256_mul_ps(rinv22,rinv22);
1962
1963             fjx0             = _mm256_setzero_ps();
1964             fjy0             = _mm256_setzero_ps();
1965             fjz0             = _mm256_setzero_ps();
1966             fjx1             = _mm256_setzero_ps();
1967             fjy1             = _mm256_setzero_ps();
1968             fjz1             = _mm256_setzero_ps();
1969             fjx2             = _mm256_setzero_ps();
1970             fjy2             = _mm256_setzero_ps();
1971             fjz2             = _mm256_setzero_ps();
1972
1973             /**************************
1974              * CALCULATE INTERACTIONS *
1975              **************************/
1976
1977             r00              = _mm256_mul_ps(rsq00,rinv00);
1978             r00              = _mm256_andnot_ps(dummy_mask,r00);
1979
1980             /* EWALD ELECTROSTATICS */
1981
1982             /* Analytical PME correction */
1983             zeta2            = _mm256_mul_ps(beta2,rsq00);
1984             rinv3            = _mm256_mul_ps(rinvsq00,rinv00);
1985             pmecorrF         = avx256_pmecorrF_f(zeta2);
1986             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1987             felec            = _mm256_mul_ps(qq00,felec);
1988
1989             /* Analytical LJ-PME */
1990             rinvsix          = _mm256_mul_ps(_mm256_mul_ps(rinvsq00,rinvsq00),rinvsq00);
1991             ewcljrsq         = _mm256_mul_ps(ewclj2,rsq00);
1992             ewclj6           = _mm256_mul_ps(ewclj2,_mm256_mul_ps(ewclj2,ewclj2));
1993             exponent         = avx256_exp_f(ewcljrsq);
1994             /* poly = exp(-(beta*r)^2) * (1 + (beta*r)^2 + (beta*r)^4 /2) */
1995             poly             = _mm256_mul_ps(exponent,_mm256_add_ps(_mm256_sub_ps(one,ewcljrsq),_mm256_mul_ps(_mm256_mul_ps(ewcljrsq,ewcljrsq),one_half)));
1996             /* f6A = 6 * C6grid * (1 - poly) */
1997             f6A              = _mm256_mul_ps(c6grid_00,_mm256_sub_ps(one,poly));
1998             /* f6B = C6grid * exponent * beta^6 */
1999             f6B              = _mm256_mul_ps(_mm256_mul_ps(c6grid_00,one_sixth),_mm256_mul_ps(exponent,ewclj6));
2000             /* fvdw = 12*C12/r13 - ((6*C6 - f6A)/r6 + f6B)/r */
2001             fvdw              = _mm256_mul_ps(_mm256_add_ps(_mm256_mul_ps(_mm256_sub_ps(_mm256_mul_ps(c12_00,rinvsix),_mm256_sub_ps(c6_00,f6A)),rinvsix),f6B),rinvsq00);
2002
2003             fscal            = _mm256_add_ps(felec,fvdw);
2004
2005             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
2006
2007             /* Calculate temporary vectorial force */
2008             tx               = _mm256_mul_ps(fscal,dx00);
2009             ty               = _mm256_mul_ps(fscal,dy00);
2010             tz               = _mm256_mul_ps(fscal,dz00);
2011
2012             /* Update vectorial force */
2013             fix0             = _mm256_add_ps(fix0,tx);
2014             fiy0             = _mm256_add_ps(fiy0,ty);
2015             fiz0             = _mm256_add_ps(fiz0,tz);
2016
2017             fjx0             = _mm256_add_ps(fjx0,tx);
2018             fjy0             = _mm256_add_ps(fjy0,ty);
2019             fjz0             = _mm256_add_ps(fjz0,tz);
2020
2021             /**************************
2022              * CALCULATE INTERACTIONS *
2023              **************************/
2024
2025             r01              = _mm256_mul_ps(rsq01,rinv01);
2026             r01              = _mm256_andnot_ps(dummy_mask,r01);
2027
2028             /* EWALD ELECTROSTATICS */
2029
2030             /* Analytical PME correction */
2031             zeta2            = _mm256_mul_ps(beta2,rsq01);
2032             rinv3            = _mm256_mul_ps(rinvsq01,rinv01);
2033             pmecorrF         = avx256_pmecorrF_f(zeta2);
2034             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2035             felec            = _mm256_mul_ps(qq01,felec);
2036
2037             fscal            = felec;
2038
2039             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
2040
2041             /* Calculate temporary vectorial force */
2042             tx               = _mm256_mul_ps(fscal,dx01);
2043             ty               = _mm256_mul_ps(fscal,dy01);
2044             tz               = _mm256_mul_ps(fscal,dz01);
2045
2046             /* Update vectorial force */
2047             fix0             = _mm256_add_ps(fix0,tx);
2048             fiy0             = _mm256_add_ps(fiy0,ty);
2049             fiz0             = _mm256_add_ps(fiz0,tz);
2050
2051             fjx1             = _mm256_add_ps(fjx1,tx);
2052             fjy1             = _mm256_add_ps(fjy1,ty);
2053             fjz1             = _mm256_add_ps(fjz1,tz);
2054
2055             /**************************
2056              * CALCULATE INTERACTIONS *
2057              **************************/
2058
2059             r02              = _mm256_mul_ps(rsq02,rinv02);
2060             r02              = _mm256_andnot_ps(dummy_mask,r02);
2061
2062             /* EWALD ELECTROSTATICS */
2063
2064             /* Analytical PME correction */
2065             zeta2            = _mm256_mul_ps(beta2,rsq02);
2066             rinv3            = _mm256_mul_ps(rinvsq02,rinv02);
2067             pmecorrF         = avx256_pmecorrF_f(zeta2);
2068             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2069             felec            = _mm256_mul_ps(qq02,felec);
2070
2071             fscal            = felec;
2072
2073             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
2074
2075             /* Calculate temporary vectorial force */
2076             tx               = _mm256_mul_ps(fscal,dx02);
2077             ty               = _mm256_mul_ps(fscal,dy02);
2078             tz               = _mm256_mul_ps(fscal,dz02);
2079
2080             /* Update vectorial force */
2081             fix0             = _mm256_add_ps(fix0,tx);
2082             fiy0             = _mm256_add_ps(fiy0,ty);
2083             fiz0             = _mm256_add_ps(fiz0,tz);
2084
2085             fjx2             = _mm256_add_ps(fjx2,tx);
2086             fjy2             = _mm256_add_ps(fjy2,ty);
2087             fjz2             = _mm256_add_ps(fjz2,tz);
2088
2089             /**************************
2090              * CALCULATE INTERACTIONS *
2091              **************************/
2092
2093             r10              = _mm256_mul_ps(rsq10,rinv10);
2094             r10              = _mm256_andnot_ps(dummy_mask,r10);
2095
2096             /* EWALD ELECTROSTATICS */
2097
2098             /* Analytical PME correction */
2099             zeta2            = _mm256_mul_ps(beta2,rsq10);
2100             rinv3            = _mm256_mul_ps(rinvsq10,rinv10);
2101             pmecorrF         = avx256_pmecorrF_f(zeta2);
2102             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2103             felec            = _mm256_mul_ps(qq10,felec);
2104
2105             fscal            = felec;
2106
2107             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
2108
2109             /* Calculate temporary vectorial force */
2110             tx               = _mm256_mul_ps(fscal,dx10);
2111             ty               = _mm256_mul_ps(fscal,dy10);
2112             tz               = _mm256_mul_ps(fscal,dz10);
2113
2114             /* Update vectorial force */
2115             fix1             = _mm256_add_ps(fix1,tx);
2116             fiy1             = _mm256_add_ps(fiy1,ty);
2117             fiz1             = _mm256_add_ps(fiz1,tz);
2118
2119             fjx0             = _mm256_add_ps(fjx0,tx);
2120             fjy0             = _mm256_add_ps(fjy0,ty);
2121             fjz0             = _mm256_add_ps(fjz0,tz);
2122
2123             /**************************
2124              * CALCULATE INTERACTIONS *
2125              **************************/
2126
2127             r11              = _mm256_mul_ps(rsq11,rinv11);
2128             r11              = _mm256_andnot_ps(dummy_mask,r11);
2129
2130             /* EWALD ELECTROSTATICS */
2131
2132             /* Analytical PME correction */
2133             zeta2            = _mm256_mul_ps(beta2,rsq11);
2134             rinv3            = _mm256_mul_ps(rinvsq11,rinv11);
2135             pmecorrF         = avx256_pmecorrF_f(zeta2);
2136             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2137             felec            = _mm256_mul_ps(qq11,felec);
2138
2139             fscal            = felec;
2140
2141             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
2142
2143             /* Calculate temporary vectorial force */
2144             tx               = _mm256_mul_ps(fscal,dx11);
2145             ty               = _mm256_mul_ps(fscal,dy11);
2146             tz               = _mm256_mul_ps(fscal,dz11);
2147
2148             /* Update vectorial force */
2149             fix1             = _mm256_add_ps(fix1,tx);
2150             fiy1             = _mm256_add_ps(fiy1,ty);
2151             fiz1             = _mm256_add_ps(fiz1,tz);
2152
2153             fjx1             = _mm256_add_ps(fjx1,tx);
2154             fjy1             = _mm256_add_ps(fjy1,ty);
2155             fjz1             = _mm256_add_ps(fjz1,tz);
2156
2157             /**************************
2158              * CALCULATE INTERACTIONS *
2159              **************************/
2160
2161             r12              = _mm256_mul_ps(rsq12,rinv12);
2162             r12              = _mm256_andnot_ps(dummy_mask,r12);
2163
2164             /* EWALD ELECTROSTATICS */
2165
2166             /* Analytical PME correction */
2167             zeta2            = _mm256_mul_ps(beta2,rsq12);
2168             rinv3            = _mm256_mul_ps(rinvsq12,rinv12);
2169             pmecorrF         = avx256_pmecorrF_f(zeta2);
2170             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2171             felec            = _mm256_mul_ps(qq12,felec);
2172
2173             fscal            = felec;
2174
2175             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
2176
2177             /* Calculate temporary vectorial force */
2178             tx               = _mm256_mul_ps(fscal,dx12);
2179             ty               = _mm256_mul_ps(fscal,dy12);
2180             tz               = _mm256_mul_ps(fscal,dz12);
2181
2182             /* Update vectorial force */
2183             fix1             = _mm256_add_ps(fix1,tx);
2184             fiy1             = _mm256_add_ps(fiy1,ty);
2185             fiz1             = _mm256_add_ps(fiz1,tz);
2186
2187             fjx2             = _mm256_add_ps(fjx2,tx);
2188             fjy2             = _mm256_add_ps(fjy2,ty);
2189             fjz2             = _mm256_add_ps(fjz2,tz);
2190
2191             /**************************
2192              * CALCULATE INTERACTIONS *
2193              **************************/
2194
2195             r20              = _mm256_mul_ps(rsq20,rinv20);
2196             r20              = _mm256_andnot_ps(dummy_mask,r20);
2197
2198             /* EWALD ELECTROSTATICS */
2199
2200             /* Analytical PME correction */
2201             zeta2            = _mm256_mul_ps(beta2,rsq20);
2202             rinv3            = _mm256_mul_ps(rinvsq20,rinv20);
2203             pmecorrF         = avx256_pmecorrF_f(zeta2);
2204             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2205             felec            = _mm256_mul_ps(qq20,felec);
2206
2207             fscal            = felec;
2208
2209             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
2210
2211             /* Calculate temporary vectorial force */
2212             tx               = _mm256_mul_ps(fscal,dx20);
2213             ty               = _mm256_mul_ps(fscal,dy20);
2214             tz               = _mm256_mul_ps(fscal,dz20);
2215
2216             /* Update vectorial force */
2217             fix2             = _mm256_add_ps(fix2,tx);
2218             fiy2             = _mm256_add_ps(fiy2,ty);
2219             fiz2             = _mm256_add_ps(fiz2,tz);
2220
2221             fjx0             = _mm256_add_ps(fjx0,tx);
2222             fjy0             = _mm256_add_ps(fjy0,ty);
2223             fjz0             = _mm256_add_ps(fjz0,tz);
2224
2225             /**************************
2226              * CALCULATE INTERACTIONS *
2227              **************************/
2228
2229             r21              = _mm256_mul_ps(rsq21,rinv21);
2230             r21              = _mm256_andnot_ps(dummy_mask,r21);
2231
2232             /* EWALD ELECTROSTATICS */
2233
2234             /* Analytical PME correction */
2235             zeta2            = _mm256_mul_ps(beta2,rsq21);
2236             rinv3            = _mm256_mul_ps(rinvsq21,rinv21);
2237             pmecorrF         = avx256_pmecorrF_f(zeta2);
2238             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2239             felec            = _mm256_mul_ps(qq21,felec);
2240
2241             fscal            = felec;
2242
2243             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
2244
2245             /* Calculate temporary vectorial force */
2246             tx               = _mm256_mul_ps(fscal,dx21);
2247             ty               = _mm256_mul_ps(fscal,dy21);
2248             tz               = _mm256_mul_ps(fscal,dz21);
2249
2250             /* Update vectorial force */
2251             fix2             = _mm256_add_ps(fix2,tx);
2252             fiy2             = _mm256_add_ps(fiy2,ty);
2253             fiz2             = _mm256_add_ps(fiz2,tz);
2254
2255             fjx1             = _mm256_add_ps(fjx1,tx);
2256             fjy1             = _mm256_add_ps(fjy1,ty);
2257             fjz1             = _mm256_add_ps(fjz1,tz);
2258
2259             /**************************
2260              * CALCULATE INTERACTIONS *
2261              **************************/
2262
2263             r22              = _mm256_mul_ps(rsq22,rinv22);
2264             r22              = _mm256_andnot_ps(dummy_mask,r22);
2265
2266             /* EWALD ELECTROSTATICS */
2267
2268             /* Analytical PME correction */
2269             zeta2            = _mm256_mul_ps(beta2,rsq22);
2270             rinv3            = _mm256_mul_ps(rinvsq22,rinv22);
2271             pmecorrF         = avx256_pmecorrF_f(zeta2);
2272             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2273             felec            = _mm256_mul_ps(qq22,felec);
2274
2275             fscal            = felec;
2276
2277             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
2278
2279             /* Calculate temporary vectorial force */
2280             tx               = _mm256_mul_ps(fscal,dx22);
2281             ty               = _mm256_mul_ps(fscal,dy22);
2282             tz               = _mm256_mul_ps(fscal,dz22);
2283
2284             /* Update vectorial force */
2285             fix2             = _mm256_add_ps(fix2,tx);
2286             fiy2             = _mm256_add_ps(fiy2,ty);
2287             fiz2             = _mm256_add_ps(fiz2,tz);
2288
2289             fjx2             = _mm256_add_ps(fjx2,tx);
2290             fjy2             = _mm256_add_ps(fjy2,ty);
2291             fjz2             = _mm256_add_ps(fjz2,tz);
2292
2293             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
2294             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
2295             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
2296             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
2297             fjptrE             = (jnrlistE>=0) ? f+j_coord_offsetE : scratch;
2298             fjptrF             = (jnrlistF>=0) ? f+j_coord_offsetF : scratch;
2299             fjptrG             = (jnrlistG>=0) ? f+j_coord_offsetG : scratch;
2300             fjptrH             = (jnrlistH>=0) ? f+j_coord_offsetH : scratch;
2301
2302             gmx_mm256_decrement_3rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,
2303                                                       fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2);
2304
2305             /* Inner loop uses 536 flops */
2306         }
2307
2308         /* End of innermost loop */
2309
2310         gmx_mm256_update_iforce_3atom_swizzle_ps(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,
2311                                                  f+i_coord_offset,fshift+i_shift_offset);
2312
2313         /* Increment number of inner iterations */
2314         inneriter                  += j_index_end - j_index_start;
2315
2316         /* Outer loop uses 18 flops */
2317     }
2318
2319     /* Increment number of outer iterations */
2320     outeriter        += nri;
2321
2322     /* Update outer/inner flops */
2323
2324     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3W3_F,outeriter*18 + inneriter*536);
2325 }