src/gromacs/gmxlib/nonbonded/nb_kernel_avx_256_single/nb_kernel_ElecEwSw_VdwLJSw_GeomW4P1_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_ElecEwSw_VdwLJSw_GeomW4P1_VF_avx_256_single
  51  * Electrostatics interaction: Ewald
  52  * VdW interaction:            LennardJones
  53  * Geometry:                   Water4-Particle
  54  * Calculate force/pot:        PotentialAndForce
  55  */
  56 void
  57 nb_kernel_ElecEwSw_VdwLJSw_GeomW4P1_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     __m256           ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  87     real *           vdwioffsetptr1;
  88     __m256           ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
  89     real *           vdwioffsetptr2;
  90     __m256           ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
  91     real *           vdwioffsetptr3;
  92     __m256           ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
  93     int              vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D,vdwjidx0E,vdwjidx0F,vdwjidx0G,vdwjidx0H;
  94     __m256           jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  95     __m256           dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
  96     __m256           dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
  97     __m256           dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
  98     __m256           dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30;
  99     __m256           velec,felec,velecsum,facel,crf,krf,krf2;
 100     real             *charge;
 101     int              nvdwtype;
 102     __m256           rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
 103     int              *vdwtype;
 104     real             *vdwparam;
 105     __m256           one_sixth   = _mm256_set1_ps(1.0/6.0);
 106     __m256           one_twelfth = _mm256_set1_ps(1.0/12.0);
 107     __m256i          ewitab;
 108     __m128i          ewitab_lo,ewitab_hi;
 109     __m256           ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
 110     __m256           beta,beta2,beta3,zeta2,pmecorrF,pmecorrV,rinv3;
 111     real             *ewtab;
 112     __m256           rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
 113     real             rswitch_scalar,d_scalar;
 114     __m256           dummy_mask,cutoff_mask;
 115     __m256           signbit = _mm256_castsi256_ps( _mm256_set1_epi32(0x80000000) );
 116     __m256           one     = _mm256_set1_ps(1.0);
 117     __m256           two     = _mm256_set1_ps(2.0);
 118     x                = xx[0];
 119     f                = ff[0];
 120
 121     nri              = nlist->nri;
 122     iinr             = nlist->iinr;
 123     jindex           = nlist->jindex;
 124     jjnr             = nlist->jjnr;
 125     shiftidx         = nlist->shift;
 126     gid              = nlist->gid;
 127     shiftvec         = fr->shift_vec[0];
 128     fshift           = fr->fshift[0];
 129     facel            = _mm256_set1_ps(fr->ic->epsfac);
 130     charge           = mdatoms->chargeA;
 131     nvdwtype         = fr->ntype;
 132     vdwparam         = fr->nbfp;
 133     vdwtype          = mdatoms->typeA;
 134
 135     sh_ewald         = _mm256_set1_ps(fr->ic->sh_ewald);
 136     beta             = _mm256_set1_ps(fr->ic->ewaldcoeff_q);
 137     beta2            = _mm256_mul_ps(beta,beta);
 138     beta3            = _mm256_mul_ps(beta,beta2);
 139
 140     ewtab            = fr->ic->tabq_coul_FDV0;
 141     ewtabscale       = _mm256_set1_ps(fr->ic->tabq_scale);
 142     ewtabhalfspace   = _mm256_set1_ps(0.5/fr->ic->tabq_scale);
 143
 144     /* Setup water-specific parameters */
 145     inr              = nlist->iinr[0];
 146     iq1              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+1]));
 147     iq2              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+2]));
 148     iq3              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+3]));
 149     vdwioffsetptr0   = vdwparam+2*nvdwtype*vdwtype[inr+0];
 150
 151     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 152     rcutoff_scalar   = fr->ic->rcoulomb;
 153     rcutoff          = _mm256_set1_ps(rcutoff_scalar);
 154     rcutoff2         = _mm256_mul_ps(rcutoff,rcutoff);
 155
 156     rswitch_scalar   = fr->ic->rcoulomb_switch;
 157     rswitch          = _mm256_set1_ps(rswitch_scalar);
 158     /* Setup switch parameters */
 159     d_scalar         = rcutoff_scalar-rswitch_scalar;
 160     d                = _mm256_set1_ps(d_scalar);
 161     swV3             = _mm256_set1_ps(-10.0/(d_scalar*d_scalar*d_scalar));
 162     swV4             = _mm256_set1_ps( 15.0/(d_scalar*d_scalar*d_scalar*d_scalar));
 163     swV5             = _mm256_set1_ps( -6.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
 164     swF2             = _mm256_set1_ps(-30.0/(d_scalar*d_scalar*d_scalar));
 165     swF3             = _mm256_set1_ps( 60.0/(d_scalar*d_scalar*d_scalar*d_scalar));
 166     swF4             = _mm256_set1_ps(-30.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
 167
 168     /* Avoid stupid compiler warnings */
 169     jnrA = jnrB = jnrC = jnrD = jnrE = jnrF = jnrG = jnrH = 0;
 170     j_coord_offsetA = 0;
 171     j_coord_offsetB = 0;
 172     j_coord_offsetC = 0;
 173     j_coord_offsetD = 0;
 174     j_coord_offsetE = 0;
 175     j_coord_offsetF = 0;
 176     j_coord_offsetG = 0;
 177     j_coord_offsetH = 0;
 178
 179     outeriter        = 0;
 180     inneriter        = 0;
 181
 182     for(iidx=0;iidx<4*DIM;iidx++)
 183     {
 184         scratch[iidx] = 0.0;
 185     }
 186
 187     /* Start outer loop over neighborlists */
 188     for(iidx=0; iidx<nri; iidx++)
 189     {
 190         /* Load shift vector for this list */
 191         i_shift_offset   = DIM*shiftidx[iidx];
 192
 193         /* Load limits for loop over neighbors */
 194         j_index_start    = jindex[iidx];
 195         j_index_end      = jindex[iidx+1];
 196
 197         /* Get outer coordinate index */
 198         inr              = iinr[iidx];
 199         i_coord_offset   = DIM*inr;
 200
 201         /* Load i particle coords and add shift vector */
 202         gmx_mm256_load_shift_and_4rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,
 203                                                     &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
 204
 205         fix0             = _mm256_setzero_ps();
 206         fiy0             = _mm256_setzero_ps();
 207         fiz0             = _mm256_setzero_ps();
 208         fix1             = _mm256_setzero_ps();
 209         fiy1             = _mm256_setzero_ps();
 210         fiz1             = _mm256_setzero_ps();
 211         fix2             = _mm256_setzero_ps();
 212         fiy2             = _mm256_setzero_ps();
 213         fiz2             = _mm256_setzero_ps();
 214         fix3             = _mm256_setzero_ps();
 215         fiy3             = _mm256_setzero_ps();
 216         fiz3             = _mm256_setzero_ps();
 217
 218         /* Reset potential sums */
 219         velecsum         = _mm256_setzero_ps();
 220         vvdwsum          = _mm256_setzero_ps();
 221
 222         /* Start inner kernel loop */
 223         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+7]>=0; jidx+=8)
 224         {
 225
 226             /* Get j neighbor index, and coordinate index */
 227             jnrA             = jjnr[jidx];
 228             jnrB             = jjnr[jidx+1];
 229             jnrC             = jjnr[jidx+2];
 230             jnrD             = jjnr[jidx+3];
 231             jnrE             = jjnr[jidx+4];
 232             jnrF             = jjnr[jidx+5];
 233             jnrG             = jjnr[jidx+6];
 234             jnrH             = jjnr[jidx+7];
 235             j_coord_offsetA  = DIM*jnrA;
 236             j_coord_offsetB  = DIM*jnrB;
 237             j_coord_offsetC  = DIM*jnrC;
 238             j_coord_offsetD  = DIM*jnrD;
 239             j_coord_offsetE  = DIM*jnrE;
 240             j_coord_offsetF  = DIM*jnrF;
 241             j_coord_offsetG  = DIM*jnrG;
 242             j_coord_offsetH  = DIM*jnrH;
 243
 244             /* load j atom coordinates */
 245             gmx_mm256_load_1rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 246                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 247                                                  x+j_coord_offsetE,x+j_coord_offsetF,
 248                                                  x+j_coord_offsetG,x+j_coord_offsetH,
 249                                                  &jx0,&jy0,&jz0);
 250
 251             /* Calculate displacement vector */
 252             dx00             = _mm256_sub_ps(ix0,jx0);
 253             dy00             = _mm256_sub_ps(iy0,jy0);
 254             dz00             = _mm256_sub_ps(iz0,jz0);
 255             dx10             = _mm256_sub_ps(ix1,jx0);
 256             dy10             = _mm256_sub_ps(iy1,jy0);
 257             dz10             = _mm256_sub_ps(iz1,jz0);
 258             dx20             = _mm256_sub_ps(ix2,jx0);
 259             dy20             = _mm256_sub_ps(iy2,jy0);
 260             dz20             = _mm256_sub_ps(iz2,jz0);
 261             dx30             = _mm256_sub_ps(ix3,jx0);
 262             dy30             = _mm256_sub_ps(iy3,jy0);
 263             dz30             = _mm256_sub_ps(iz3,jz0);
 264
 265             /* Calculate squared distance and things based on it */
 266             rsq00            = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
 267             rsq10            = gmx_mm256_calc_rsq_ps(dx10,dy10,dz10);
 268             rsq20            = gmx_mm256_calc_rsq_ps(dx20,dy20,dz20);
 269             rsq30            = gmx_mm256_calc_rsq_ps(dx30,dy30,dz30);
 270
 271             rinv00           = avx256_invsqrt_f(rsq00);
 272             rinv10           = avx256_invsqrt_f(rsq10);
 273             rinv20           = avx256_invsqrt_f(rsq20);
 274             rinv30           = avx256_invsqrt_f(rsq30);
 275
 276             rinvsq00         = _mm256_mul_ps(rinv00,rinv00);
 277             rinvsq10         = _mm256_mul_ps(rinv10,rinv10);
 278             rinvsq20         = _mm256_mul_ps(rinv20,rinv20);
 279             rinvsq30         = _mm256_mul_ps(rinv30,rinv30);
 280
 281             /* Load parameters for j particles */
 282             jq0              = gmx_mm256_load_8real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
 283                                                                  charge+jnrC+0,charge+jnrD+0,
 284                                                                  charge+jnrE+0,charge+jnrF+0,
 285                                                                  charge+jnrG+0,charge+jnrH+0);
 286             vdwjidx0A        = 2*vdwtype[jnrA+0];
 287             vdwjidx0B        = 2*vdwtype[jnrB+0];
 288             vdwjidx0C        = 2*vdwtype[jnrC+0];
 289             vdwjidx0D        = 2*vdwtype[jnrD+0];
 290             vdwjidx0E        = 2*vdwtype[jnrE+0];
 291             vdwjidx0F        = 2*vdwtype[jnrF+0];
 292             vdwjidx0G        = 2*vdwtype[jnrG+0];
 293             vdwjidx0H        = 2*vdwtype[jnrH+0];
 294
 295             fjx0             = _mm256_setzero_ps();
 296             fjy0             = _mm256_setzero_ps();
 297             fjz0             = _mm256_setzero_ps();
 298
 299             /**************************
 300              * CALCULATE INTERACTIONS *
 301              **************************/
 302
 303             if (gmx_mm256_any_lt(rsq00,rcutoff2))
 304             {
 305
 306             r00              = _mm256_mul_ps(rsq00,rinv00);
 307
 308             /* Compute parameters for interactions between i and j atoms */
 309             gmx_mm256_load_8pair_swizzle_ps(vdwioffsetptr0+vdwjidx0A,
 310                                             vdwioffsetptr0+vdwjidx0B,
 311                                             vdwioffsetptr0+vdwjidx0C,
 312                                             vdwioffsetptr0+vdwjidx0D,
 313                                             vdwioffsetptr0+vdwjidx0E,
 314                                             vdwioffsetptr0+vdwjidx0F,
 315                                             vdwioffsetptr0+vdwjidx0G,
 316                                             vdwioffsetptr0+vdwjidx0H,
 317                                             &c6_00,&c12_00);
 318
 319             /* LENNARD-JONES DISPERSION/REPULSION */
 320
 321             rinvsix          = _mm256_mul_ps(_mm256_mul_ps(rinvsq00,rinvsq00),rinvsq00);
 322             vvdw6            = _mm256_mul_ps(c6_00,rinvsix);
 323             vvdw12           = _mm256_mul_ps(c12_00,_mm256_mul_ps(rinvsix,rinvsix));
 324             vvdw             = _mm256_sub_ps( _mm256_mul_ps(vvdw12,one_twelfth) , _mm256_mul_ps(vvdw6,one_sixth) );
 325             fvdw             = _mm256_mul_ps(_mm256_sub_ps(vvdw12,vvdw6),rinvsq00);
 326
 327             d                = _mm256_sub_ps(r00,rswitch);
 328             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 329             d2               = _mm256_mul_ps(d,d);
 330             sw               = _mm256_add_ps(one,_mm256_mul_ps(d2,_mm256_mul_ps(d,_mm256_add_ps(swV3,_mm256_mul_ps(d,_mm256_add_ps(swV4,_mm256_mul_ps(d,swV5)))))));
 331
 332             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 333
 334             /* Evaluate switch function */
 335             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 336             fvdw             = _mm256_sub_ps( _mm256_mul_ps(fvdw,sw) , _mm256_mul_ps(rinv00,_mm256_mul_ps(vvdw,dsw)) );
 337             vvdw             = _mm256_mul_ps(vvdw,sw);
 338             cutoff_mask      = _mm256_cmp_ps(rsq00,rcutoff2,_CMP_LT_OQ);
 339
 340             /* Update potential sum for this i atom from the interaction with this j atom. */
 341             vvdw             = _mm256_and_ps(vvdw,cutoff_mask);
 342             vvdwsum          = _mm256_add_ps(vvdwsum,vvdw);
 343
 344             fscal            = fvdw;
 345
 346             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 347
 348             /* Calculate temporary vectorial force */
 349             tx               = _mm256_mul_ps(fscal,dx00);
 350             ty               = _mm256_mul_ps(fscal,dy00);
 351             tz               = _mm256_mul_ps(fscal,dz00);
 352
 353             /* Update vectorial force */
 354             fix0             = _mm256_add_ps(fix0,tx);
 355             fiy0             = _mm256_add_ps(fiy0,ty);
 356             fiz0             = _mm256_add_ps(fiz0,tz);
 357
 358             fjx0             = _mm256_add_ps(fjx0,tx);
 359             fjy0             = _mm256_add_ps(fjy0,ty);
 360             fjz0             = _mm256_add_ps(fjz0,tz);
 361
 362             }
 363
 364             /**************************
 365              * CALCULATE INTERACTIONS *
 366              **************************/
 367
 368             if (gmx_mm256_any_lt(rsq10,rcutoff2))
 369             {
 370
 371             r10              = _mm256_mul_ps(rsq10,rinv10);
 372
 373             /* Compute parameters for interactions between i and j atoms */
 374             qq10             = _mm256_mul_ps(iq1,jq0);
 375
 376             /* EWALD ELECTROSTATICS */
 377
 378             /* Analytical PME correction */
 379             zeta2            = _mm256_mul_ps(beta2,rsq10);
 380             rinv3            = _mm256_mul_ps(rinvsq10,rinv10);
 381             pmecorrF         = avx256_pmecorrF_f(zeta2);
 382             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 383             felec            = _mm256_mul_ps(qq10,felec);
 384             pmecorrV         = avx256_pmecorrV_f(zeta2);
 385             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 386             velec            = _mm256_sub_ps(rinv10,pmecorrV);
 387             velec            = _mm256_mul_ps(qq10,velec);
 388
 389             d                = _mm256_sub_ps(r10,rswitch);
 390             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 391             d2               = _mm256_mul_ps(d,d);
 392             sw               = _mm256_add_ps(one,_mm256_mul_ps(d2,_mm256_mul_ps(d,_mm256_add_ps(swV3,_mm256_mul_ps(d,_mm256_add_ps(swV4,_mm256_mul_ps(d,swV5)))))));
 393
 394             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 395
 396             /* Evaluate switch function */
 397             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 398             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv10,_mm256_mul_ps(velec,dsw)) );
 399             velec            = _mm256_mul_ps(velec,sw);
 400             cutoff_mask      = _mm256_cmp_ps(rsq10,rcutoff2,_CMP_LT_OQ);
 401
 402             /* Update potential sum for this i atom from the interaction with this j atom. */
 403             velec            = _mm256_and_ps(velec,cutoff_mask);
 404             velecsum         = _mm256_add_ps(velecsum,velec);
 405
 406             fscal            = felec;
 407
 408             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 409
 410             /* Calculate temporary vectorial force */
 411             tx               = _mm256_mul_ps(fscal,dx10);
 412             ty               = _mm256_mul_ps(fscal,dy10);
 413             tz               = _mm256_mul_ps(fscal,dz10);
 414
 415             /* Update vectorial force */
 416             fix1             = _mm256_add_ps(fix1,tx);
 417             fiy1             = _mm256_add_ps(fiy1,ty);
 418             fiz1             = _mm256_add_ps(fiz1,tz);
 419
 420             fjx0             = _mm256_add_ps(fjx0,tx);
 421             fjy0             = _mm256_add_ps(fjy0,ty);
 422             fjz0             = _mm256_add_ps(fjz0,tz);
 423
 424             }
 425
 426             /**************************
 427              * CALCULATE INTERACTIONS *
 428              **************************/
 429
 430             if (gmx_mm256_any_lt(rsq20,rcutoff2))
 431             {
 432
 433             r20              = _mm256_mul_ps(rsq20,rinv20);
 434
 435             /* Compute parameters for interactions between i and j atoms */
 436             qq20             = _mm256_mul_ps(iq2,jq0);
 437
 438             /* EWALD ELECTROSTATICS */
 439
 440             /* Analytical PME correction */
 441             zeta2            = _mm256_mul_ps(beta2,rsq20);
 442             rinv3            = _mm256_mul_ps(rinvsq20,rinv20);
 443             pmecorrF         = avx256_pmecorrF_f(zeta2);
 444             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 445             felec            = _mm256_mul_ps(qq20,felec);
 446             pmecorrV         = avx256_pmecorrV_f(zeta2);
 447             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 448             velec            = _mm256_sub_ps(rinv20,pmecorrV);
 449             velec            = _mm256_mul_ps(qq20,velec);
 450
 451             d                = _mm256_sub_ps(r20,rswitch);
 452             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 453             d2               = _mm256_mul_ps(d,d);
 454             sw               = _mm256_add_ps(one,_mm256_mul_ps(d2,_mm256_mul_ps(d,_mm256_add_ps(swV3,_mm256_mul_ps(d,_mm256_add_ps(swV4,_mm256_mul_ps(d,swV5)))))));
 455
 456             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 457
 458             /* Evaluate switch function */
 459             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 460             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv20,_mm256_mul_ps(velec,dsw)) );
 461             velec            = _mm256_mul_ps(velec,sw);
 462             cutoff_mask      = _mm256_cmp_ps(rsq20,rcutoff2,_CMP_LT_OQ);
 463
 464             /* Update potential sum for this i atom from the interaction with this j atom. */
 465             velec            = _mm256_and_ps(velec,cutoff_mask);
 466             velecsum         = _mm256_add_ps(velecsum,velec);
 467
 468             fscal            = felec;
 469
 470             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 471
 472             /* Calculate temporary vectorial force */
 473             tx               = _mm256_mul_ps(fscal,dx20);
 474             ty               = _mm256_mul_ps(fscal,dy20);
 475             tz               = _mm256_mul_ps(fscal,dz20);
 476
 477             /* Update vectorial force */
 478             fix2             = _mm256_add_ps(fix2,tx);
 479             fiy2             = _mm256_add_ps(fiy2,ty);
 480             fiz2             = _mm256_add_ps(fiz2,tz);
 481
 482             fjx0             = _mm256_add_ps(fjx0,tx);
 483             fjy0             = _mm256_add_ps(fjy0,ty);
 484             fjz0             = _mm256_add_ps(fjz0,tz);
 485
 486             }
 487
 488             /**************************
 489              * CALCULATE INTERACTIONS *
 490              **************************/
 491
 492             if (gmx_mm256_any_lt(rsq30,rcutoff2))
 493             {
 494
 495             r30              = _mm256_mul_ps(rsq30,rinv30);
 496
 497             /* Compute parameters for interactions between i and j atoms */
 498             qq30             = _mm256_mul_ps(iq3,jq0);
 499
 500             /* EWALD ELECTROSTATICS */
 501
 502             /* Analytical PME correction */
 503             zeta2            = _mm256_mul_ps(beta2,rsq30);
 504             rinv3            = _mm256_mul_ps(rinvsq30,rinv30);
 505             pmecorrF         = avx256_pmecorrF_f(zeta2);
 506             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 507             felec            = _mm256_mul_ps(qq30,felec);
 508             pmecorrV         = avx256_pmecorrV_f(zeta2);
 509             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 510             velec            = _mm256_sub_ps(rinv30,pmecorrV);
 511             velec            = _mm256_mul_ps(qq30,velec);
 512
 513             d                = _mm256_sub_ps(r30,rswitch);
 514             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 515             d2               = _mm256_mul_ps(d,d);
 516             sw               = _mm256_add_ps(one,_mm256_mul_ps(d2,_mm256_mul_ps(d,_mm256_add_ps(swV3,_mm256_mul_ps(d,_mm256_add_ps(swV4,_mm256_mul_ps(d,swV5)))))));
 517
 518             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 519
 520             /* Evaluate switch function */
 521             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 522             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv30,_mm256_mul_ps(velec,dsw)) );
 523             velec            = _mm256_mul_ps(velec,sw);
 524             cutoff_mask      = _mm256_cmp_ps(rsq30,rcutoff2,_CMP_LT_OQ);
 525
 526             /* Update potential sum for this i atom from the interaction with this j atom. */
 527             velec            = _mm256_and_ps(velec,cutoff_mask);
 528             velecsum         = _mm256_add_ps(velecsum,velec);
 529
 530             fscal            = felec;
 531
 532             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 533
 534             /* Calculate temporary vectorial force */
 535             tx               = _mm256_mul_ps(fscal,dx30);
 536             ty               = _mm256_mul_ps(fscal,dy30);
 537             tz               = _mm256_mul_ps(fscal,dz30);
 538
 539             /* Update vectorial force */
 540             fix3             = _mm256_add_ps(fix3,tx);
 541             fiy3             = _mm256_add_ps(fiy3,ty);
 542             fiz3             = _mm256_add_ps(fiz3,tz);
 543
 544             fjx0             = _mm256_add_ps(fjx0,tx);
 545             fjy0             = _mm256_add_ps(fjy0,ty);
 546             fjz0             = _mm256_add_ps(fjz0,tz);
 547
 548             }
 549
 550             fjptrA             = f+j_coord_offsetA;
 551             fjptrB             = f+j_coord_offsetB;
 552             fjptrC             = f+j_coord_offsetC;
 553             fjptrD             = f+j_coord_offsetD;
 554             fjptrE             = f+j_coord_offsetE;
 555             fjptrF             = f+j_coord_offsetF;
 556             fjptrG             = f+j_coord_offsetG;
 557             fjptrH             = f+j_coord_offsetH;
 558
 559             gmx_mm256_decrement_1rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,fjx0,fjy0,fjz0);
 560
 561             /* Inner loop uses 386 flops */
 562         }
 563
 564         if(jidx<j_index_end)
 565         {
 566
 567             /* Get j neighbor index, and coordinate index */
 568             jnrlistA         = jjnr[jidx];
 569             jnrlistB         = jjnr[jidx+1];
 570             jnrlistC         = jjnr[jidx+2];
 571             jnrlistD         = jjnr[jidx+3];
 572             jnrlistE         = jjnr[jidx+4];
 573             jnrlistF         = jjnr[jidx+5];
 574             jnrlistG         = jjnr[jidx+6];
 575             jnrlistH         = jjnr[jidx+7];
 576             /* Sign of each element will be negative for non-real atoms.
 577              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
 578              * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
 579              */
 580             dummy_mask = gmx_mm256_set_m128(gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx+4)),_mm_setzero_si128())),
 581                                             gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128())));
 582
 583             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
 584             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
 585             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
 586             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
 587             jnrE       = (jnrlistE>=0) ? jnrlistE : 0;
 588             jnrF       = (jnrlistF>=0) ? jnrlistF : 0;
 589             jnrG       = (jnrlistG>=0) ? jnrlistG : 0;
 590             jnrH       = (jnrlistH>=0) ? jnrlistH : 0;
 591             j_coord_offsetA  = DIM*jnrA;
 592             j_coord_offsetB  = DIM*jnrB;
 593             j_coord_offsetC  = DIM*jnrC;
 594             j_coord_offsetD  = DIM*jnrD;
 595             j_coord_offsetE  = DIM*jnrE;
 596             j_coord_offsetF  = DIM*jnrF;
 597             j_coord_offsetG  = DIM*jnrG;
 598             j_coord_offsetH  = DIM*jnrH;
 599
 600             /* load j atom coordinates */
 601             gmx_mm256_load_1rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 602                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 603                                                  x+j_coord_offsetE,x+j_coord_offsetF,
 604                                                  x+j_coord_offsetG,x+j_coord_offsetH,
 605                                                  &jx0,&jy0,&jz0);
 606
 607             /* Calculate displacement vector */
 608             dx00             = _mm256_sub_ps(ix0,jx0);
 609             dy00             = _mm256_sub_ps(iy0,jy0);
 610             dz00             = _mm256_sub_ps(iz0,jz0);
 611             dx10             = _mm256_sub_ps(ix1,jx0);
 612             dy10             = _mm256_sub_ps(iy1,jy0);
 613             dz10             = _mm256_sub_ps(iz1,jz0);
 614             dx20             = _mm256_sub_ps(ix2,jx0);
 615             dy20             = _mm256_sub_ps(iy2,jy0);
 616             dz20             = _mm256_sub_ps(iz2,jz0);
 617             dx30             = _mm256_sub_ps(ix3,jx0);
 618             dy30             = _mm256_sub_ps(iy3,jy0);
 619             dz30             = _mm256_sub_ps(iz3,jz0);
 620
 621             /* Calculate squared distance and things based on it */
 622             rsq00            = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
 623             rsq10            = gmx_mm256_calc_rsq_ps(dx10,dy10,dz10);
 624             rsq20            = gmx_mm256_calc_rsq_ps(dx20,dy20,dz20);
 625             rsq30            = gmx_mm256_calc_rsq_ps(dx30,dy30,dz30);
 626
 627             rinv00           = avx256_invsqrt_f(rsq00);
 628             rinv10           = avx256_invsqrt_f(rsq10);
 629             rinv20           = avx256_invsqrt_f(rsq20);
 630             rinv30           = avx256_invsqrt_f(rsq30);
 631
 632             rinvsq00         = _mm256_mul_ps(rinv00,rinv00);
 633             rinvsq10         = _mm256_mul_ps(rinv10,rinv10);
 634             rinvsq20         = _mm256_mul_ps(rinv20,rinv20);
 635             rinvsq30         = _mm256_mul_ps(rinv30,rinv30);
 636
 637             /* Load parameters for j particles */
 638             jq0              = gmx_mm256_load_8real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
 639                                                                  charge+jnrC+0,charge+jnrD+0,
 640                                                                  charge+jnrE+0,charge+jnrF+0,
 641                                                                  charge+jnrG+0,charge+jnrH+0);
 642             vdwjidx0A        = 2*vdwtype[jnrA+0];
 643             vdwjidx0B        = 2*vdwtype[jnrB+0];
 644             vdwjidx0C        = 2*vdwtype[jnrC+0];
 645             vdwjidx0D        = 2*vdwtype[jnrD+0];
 646             vdwjidx0E        = 2*vdwtype[jnrE+0];
 647             vdwjidx0F        = 2*vdwtype[jnrF+0];
 648             vdwjidx0G        = 2*vdwtype[jnrG+0];
 649             vdwjidx0H        = 2*vdwtype[jnrH+0];
 650
 651             fjx0             = _mm256_setzero_ps();
 652             fjy0             = _mm256_setzero_ps();
 653             fjz0             = _mm256_setzero_ps();
 654
 655             /**************************
 656              * CALCULATE INTERACTIONS *
 657              **************************/
 658
 659             if (gmx_mm256_any_lt(rsq00,rcutoff2))
 660             {
 661
 662             r00              = _mm256_mul_ps(rsq00,rinv00);
 663             r00              = _mm256_andnot_ps(dummy_mask,r00);
 664
 665             /* Compute parameters for interactions between i and j atoms */
 666             gmx_mm256_load_8pair_swizzle_ps(vdwioffsetptr0+vdwjidx0A,
 667                                             vdwioffsetptr0+vdwjidx0B,
 668                                             vdwioffsetptr0+vdwjidx0C,
 669                                             vdwioffsetptr0+vdwjidx0D,
 670                                             vdwioffsetptr0+vdwjidx0E,
 671                                             vdwioffsetptr0+vdwjidx0F,
 672                                             vdwioffsetptr0+vdwjidx0G,
 673                                             vdwioffsetptr0+vdwjidx0H,
 674                                             &c6_00,&c12_00);
 675
 676             /* LENNARD-JONES DISPERSION/REPULSION */
 677
 678             rinvsix          = _mm256_mul_ps(_mm256_mul_ps(rinvsq00,rinvsq00),rinvsq00);
 679             vvdw6            = _mm256_mul_ps(c6_00,rinvsix);
 680             vvdw12           = _mm256_mul_ps(c12_00,_mm256_mul_ps(rinvsix,rinvsix));
 681             vvdw             = _mm256_sub_ps( _mm256_mul_ps(vvdw12,one_twelfth) , _mm256_mul_ps(vvdw6,one_sixth) );
 682             fvdw             = _mm256_mul_ps(_mm256_sub_ps(vvdw12,vvdw6),rinvsq00);
 683
 684             d                = _mm256_sub_ps(r00,rswitch);
 685             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 686             d2               = _mm256_mul_ps(d,d);
 687             sw               = _mm256_add_ps(one,_mm256_mul_ps(d2,_mm256_mul_ps(d,_mm256_add_ps(swV3,_mm256_mul_ps(d,_mm256_add_ps(swV4,_mm256_mul_ps(d,swV5)))))));
 688
 689             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 690
 691             /* Evaluate switch function */
 692             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 693             fvdw             = _mm256_sub_ps( _mm256_mul_ps(fvdw,sw) , _mm256_mul_ps(rinv00,_mm256_mul_ps(vvdw,dsw)) );
 694             vvdw             = _mm256_mul_ps(vvdw,sw);
 695             cutoff_mask      = _mm256_cmp_ps(rsq00,rcutoff2,_CMP_LT_OQ);
 696
 697             /* Update potential sum for this i atom from the interaction with this j atom. */
 698             vvdw             = _mm256_and_ps(vvdw,cutoff_mask);
 699             vvdw             = _mm256_andnot_ps(dummy_mask,vvdw);
 700             vvdwsum          = _mm256_add_ps(vvdwsum,vvdw);
 701
 702             fscal            = fvdw;
 703
 704             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 705
 706             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
 707
 708             /* Calculate temporary vectorial force */
 709             tx               = _mm256_mul_ps(fscal,dx00);
 710             ty               = _mm256_mul_ps(fscal,dy00);
 711             tz               = _mm256_mul_ps(fscal,dz00);
 712
 713             /* Update vectorial force */
 714             fix0             = _mm256_add_ps(fix0,tx);
 715             fiy0             = _mm256_add_ps(fiy0,ty);
 716             fiz0             = _mm256_add_ps(fiz0,tz);
 717
 718             fjx0             = _mm256_add_ps(fjx0,tx);
 719             fjy0             = _mm256_add_ps(fjy0,ty);
 720             fjz0             = _mm256_add_ps(fjz0,tz);
 721
 722             }
 723
 724             /**************************
 725              * CALCULATE INTERACTIONS *
 726              **************************/
 727
 728             if (gmx_mm256_any_lt(rsq10,rcutoff2))
 729             {
 730
 731             r10              = _mm256_mul_ps(rsq10,rinv10);
 732             r10              = _mm256_andnot_ps(dummy_mask,r10);
 733
 734             /* Compute parameters for interactions between i and j atoms */
 735             qq10             = _mm256_mul_ps(iq1,jq0);
 736
 737             /* EWALD ELECTROSTATICS */
 738
 739             /* Analytical PME correction */
 740             zeta2            = _mm256_mul_ps(beta2,rsq10);
 741             rinv3            = _mm256_mul_ps(rinvsq10,rinv10);
 742             pmecorrF         = avx256_pmecorrF_f(zeta2);
 743             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 744             felec            = _mm256_mul_ps(qq10,felec);
 745             pmecorrV         = avx256_pmecorrV_f(zeta2);
 746             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 747             velec            = _mm256_sub_ps(rinv10,pmecorrV);
 748             velec            = _mm256_mul_ps(qq10,velec);
 749
 750             d                = _mm256_sub_ps(r10,rswitch);
 751             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 752             d2               = _mm256_mul_ps(d,d);
 753             sw               = _mm256_add_ps(one,_mm256_mul_ps(d2,_mm256_mul_ps(d,_mm256_add_ps(swV3,_mm256_mul_ps(d,_mm256_add_ps(swV4,_mm256_mul_ps(d,swV5)))))));
 754
 755             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 756
 757             /* Evaluate switch function */
 758             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 759             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv10,_mm256_mul_ps(velec,dsw)) );
 760             velec            = _mm256_mul_ps(velec,sw);
 761             cutoff_mask      = _mm256_cmp_ps(rsq10,rcutoff2,_CMP_LT_OQ);
 762
 763             /* Update potential sum for this i atom from the interaction with this j atom. */
 764             velec            = _mm256_and_ps(velec,cutoff_mask);
 765             velec            = _mm256_andnot_ps(dummy_mask,velec);
 766             velecsum         = _mm256_add_ps(velecsum,velec);
 767
 768             fscal            = felec;
 769
 770             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 771
 772             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
 773
 774             /* Calculate temporary vectorial force */
 775             tx               = _mm256_mul_ps(fscal,dx10);
 776             ty               = _mm256_mul_ps(fscal,dy10);
 777             tz               = _mm256_mul_ps(fscal,dz10);
 778
 779             /* Update vectorial force */
 780             fix1             = _mm256_add_ps(fix1,tx);
 781             fiy1             = _mm256_add_ps(fiy1,ty);
 782             fiz1             = _mm256_add_ps(fiz1,tz);
 783
 784             fjx0             = _mm256_add_ps(fjx0,tx);
 785             fjy0             = _mm256_add_ps(fjy0,ty);
 786             fjz0             = _mm256_add_ps(fjz0,tz);
 787
 788             }
 789
 790             /**************************
 791              * CALCULATE INTERACTIONS *
 792              **************************/
 793
 794             if (gmx_mm256_any_lt(rsq20,rcutoff2))
 795             {
 796
 797             r20              = _mm256_mul_ps(rsq20,rinv20);
 798             r20              = _mm256_andnot_ps(dummy_mask,r20);
 799
 800             /* Compute parameters for interactions between i and j atoms */
 801             qq20             = _mm256_mul_ps(iq2,jq0);
 802
 803             /* EWALD ELECTROSTATICS */
 804
 805             /* Analytical PME correction */
 806             zeta2            = _mm256_mul_ps(beta2,rsq20);
 807             rinv3            = _mm256_mul_ps(rinvsq20,rinv20);
 808             pmecorrF         = avx256_pmecorrF_f(zeta2);
 809             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 810             felec            = _mm256_mul_ps(qq20,felec);
 811             pmecorrV         = avx256_pmecorrV_f(zeta2);
 812             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 813             velec            = _mm256_sub_ps(rinv20,pmecorrV);
 814             velec            = _mm256_mul_ps(qq20,velec);
 815
 816             d                = _mm256_sub_ps(r20,rswitch);
 817             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 818             d2               = _mm256_mul_ps(d,d);
 819             sw               = _mm256_add_ps(one,_mm256_mul_ps(d2,_mm256_mul_ps(d,_mm256_add_ps(swV3,_mm256_mul_ps(d,_mm256_add_ps(swV4,_mm256_mul_ps(d,swV5)))))));
 820
 821             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 822
 823             /* Evaluate switch function */
 824             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 825             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv20,_mm256_mul_ps(velec,dsw)) );
 826             velec            = _mm256_mul_ps(velec,sw);
 827             cutoff_mask      = _mm256_cmp_ps(rsq20,rcutoff2,_CMP_LT_OQ);
 828
 829             /* Update potential sum for this i atom from the interaction with this j atom. */
 830             velec            = _mm256_and_ps(velec,cutoff_mask);
 831             velec            = _mm256_andnot_ps(dummy_mask,velec);
 832             velecsum         = _mm256_add_ps(velecsum,velec);
 833
 834             fscal            = felec;
 835
 836             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 837
 838             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
 839
 840             /* Calculate temporary vectorial force */
 841             tx               = _mm256_mul_ps(fscal,dx20);
 842             ty               = _mm256_mul_ps(fscal,dy20);
 843             tz               = _mm256_mul_ps(fscal,dz20);
 844
 845             /* Update vectorial force */
 846             fix2             = _mm256_add_ps(fix2,tx);
 847             fiy2             = _mm256_add_ps(fiy2,ty);
 848             fiz2             = _mm256_add_ps(fiz2,tz);
 849
 850             fjx0             = _mm256_add_ps(fjx0,tx);
 851             fjy0             = _mm256_add_ps(fjy0,ty);
 852             fjz0             = _mm256_add_ps(fjz0,tz);
 853
 854             }
 855
 856             /**************************
 857              * CALCULATE INTERACTIONS *
 858              **************************/
 859
 860             if (gmx_mm256_any_lt(rsq30,rcutoff2))
 861             {
 862
 863             r30              = _mm256_mul_ps(rsq30,rinv30);
 864             r30              = _mm256_andnot_ps(dummy_mask,r30);
 865
 866             /* Compute parameters for interactions between i and j atoms */
 867             qq30             = _mm256_mul_ps(iq3,jq0);
 868
 869             /* EWALD ELECTROSTATICS */
 870
 871             /* Analytical PME correction */
 872             zeta2            = _mm256_mul_ps(beta2,rsq30);
 873             rinv3            = _mm256_mul_ps(rinvsq30,rinv30);
 874             pmecorrF         = avx256_pmecorrF_f(zeta2);
 875             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 876             felec            = _mm256_mul_ps(qq30,felec);
 877             pmecorrV         = avx256_pmecorrV_f(zeta2);
 878             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 879             velec            = _mm256_sub_ps(rinv30,pmecorrV);
 880             velec            = _mm256_mul_ps(qq30,velec);
 881
 882             d                = _mm256_sub_ps(r30,rswitch);
 883             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 884             d2               = _mm256_mul_ps(d,d);
 885             sw               = _mm256_add_ps(one,_mm256_mul_ps(d2,_mm256_mul_ps(d,_mm256_add_ps(swV3,_mm256_mul_ps(d,_mm256_add_ps(swV4,_mm256_mul_ps(d,swV5)))))));
 886
 887             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 888
 889             /* Evaluate switch function */
 890             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 891             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv30,_mm256_mul_ps(velec,dsw)) );
 892             velec            = _mm256_mul_ps(velec,sw);
 893             cutoff_mask      = _mm256_cmp_ps(rsq30,rcutoff2,_CMP_LT_OQ);
 894
 895             /* Update potential sum for this i atom from the interaction with this j atom. */
 896             velec            = _mm256_and_ps(velec,cutoff_mask);
 897             velec            = _mm256_andnot_ps(dummy_mask,velec);
 898             velecsum         = _mm256_add_ps(velecsum,velec);
 899
 900             fscal            = felec;
 901
 902             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 903
 904             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
 905
 906             /* Calculate temporary vectorial force */
 907             tx               = _mm256_mul_ps(fscal,dx30);
 908             ty               = _mm256_mul_ps(fscal,dy30);
 909             tz               = _mm256_mul_ps(fscal,dz30);
 910
 911             /* Update vectorial force */
 912             fix3             = _mm256_add_ps(fix3,tx);
 913             fiy3             = _mm256_add_ps(fiy3,ty);
 914             fiz3             = _mm256_add_ps(fiz3,tz);
 915
 916             fjx0             = _mm256_add_ps(fjx0,tx);
 917             fjy0             = _mm256_add_ps(fjy0,ty);
 918             fjz0             = _mm256_add_ps(fjz0,tz);
 919
 920             }
 921
 922             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
 923             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
 924             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
 925             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
 926             fjptrE             = (jnrlistE>=0) ? f+j_coord_offsetE : scratch;
 927             fjptrF             = (jnrlistF>=0) ? f+j_coord_offsetF : scratch;
 928             fjptrG             = (jnrlistG>=0) ? f+j_coord_offsetG : scratch;
 929             fjptrH             = (jnrlistH>=0) ? f+j_coord_offsetH : scratch;
 930
 931             gmx_mm256_decrement_1rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,fjx0,fjy0,fjz0);
 932
 933             /* Inner loop uses 390 flops */
 934         }
 935
 936         /* End of innermost loop */
 937
 938         gmx_mm256_update_iforce_4atom_swizzle_ps(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
 939                                                  f+i_coord_offset,fshift+i_shift_offset);
 940
 941         ggid                        = gid[iidx];
 942         /* Update potential energies */
 943         gmx_mm256_update_1pot_ps(velecsum,kernel_data->energygrp_elec+ggid);
 944         gmx_mm256_update_1pot_ps(vvdwsum,kernel_data->energygrp_vdw+ggid);
 945
 946         /* Increment number of inner iterations */
 947         inneriter                  += j_index_end - j_index_start;
 948
 949         /* Outer loop uses 26 flops */
 950     }
 951
 952     /* Increment number of outer iterations */
 953     outeriter        += nri;
 954
 955     /* Update outer/inner flops */
 956
 957     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_VF,outeriter*26 + inneriter*390);
 958 }
 959 /*
 960  * Gromacs nonbonded kernel:   nb_kernel_ElecEwSw_VdwLJSw_GeomW4P1_F_avx_256_single
 961  * Electrostatics interaction: Ewald
 962  * VdW interaction:            LennardJones
 963  * Geometry:                   Water4-Particle
 964  * Calculate force/pot:        Force
 965  */
 966 void
 967 nb_kernel_ElecEwSw_VdwLJSw_GeomW4P1_F_avx_256_single
 968                     (t_nblist                    * gmx_restrict       nlist,
 969                      rvec                        * gmx_restrict          xx,
 970                      rvec                        * gmx_restrict          ff,
 971                      struct t_forcerec           * gmx_restrict          fr,
 972                      t_mdatoms                   * gmx_restrict     mdatoms,
 973                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 974                      t_nrnb                      * gmx_restrict        nrnb)
 975 {
 976     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
 977      * just 0 for non-waters.
 978      * Suffixes A,B,C,D,E,F,G,H refer to j loop unrolling done with AVX, e.g. for the eight different
 979      * jnr indices corresponding to data put in the four positions in the SIMD register.
 980      */
 981     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
 982     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
 983     int              jnrA,jnrB,jnrC,jnrD;
 984     int              jnrE,jnrF,jnrG,jnrH;
 985     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
 986     int              jnrlistE,jnrlistF,jnrlistG,jnrlistH;
 987     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
 988     int              j_coord_offsetE,j_coord_offsetF,j_coord_offsetG,j_coord_offsetH;
 989     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 990     real             rcutoff_scalar;
 991     real             *shiftvec,*fshift,*x,*f;
 992     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD,*fjptrE,*fjptrF,*fjptrG,*fjptrH;
 993     real             scratch[4*DIM];
 994     __m256           tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
 995     real *           vdwioffsetptr0;
 996     __m256           ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 997     real *           vdwioffsetptr1;
 998     __m256           ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
 999     real *           vdwioffsetptr2;
1000     __m256           ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
1001     real *           vdwioffsetptr3;
1002     __m256           ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
1003     int              vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D,vdwjidx0E,vdwjidx0F,vdwjidx0G,vdwjidx0H;
1004     __m256           jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
1005     __m256           dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
1006     __m256           dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
1007     __m256           dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
1008     __m256           dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30;
1009     __m256           velec,felec,velecsum,facel,crf,krf,krf2;
1010     real             *charge;
1011     int              nvdwtype;
1012     __m256           rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
1013     int              *vdwtype;
1014     real             *vdwparam;
1015     __m256           one_sixth   = _mm256_set1_ps(1.0/6.0);
1016     __m256           one_twelfth = _mm256_set1_ps(1.0/12.0);
1017     __m256i          ewitab;
1018     __m128i          ewitab_lo,ewitab_hi;
1019     __m256           ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
1020     __m256           beta,beta2,beta3,zeta2,pmecorrF,pmecorrV,rinv3;
1021     real             *ewtab;
1022     __m256           rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
1023     real             rswitch_scalar,d_scalar;
1024     __m256           dummy_mask,cutoff_mask;
1025     __m256           signbit = _mm256_castsi256_ps( _mm256_set1_epi32(0x80000000) );
1026     __m256           one     = _mm256_set1_ps(1.0);
1027     __m256           two     = _mm256_set1_ps(2.0);
1028     x                = xx[0];
1029     f                = ff[0];
1030
1031     nri              = nlist->nri;
1032     iinr             = nlist->iinr;
1033     jindex           = nlist->jindex;
1034     jjnr             = nlist->jjnr;
1035     shiftidx         = nlist->shift;
1036     gid              = nlist->gid;
1037     shiftvec         = fr->shift_vec[0];
1038     fshift           = fr->fshift[0];
1039     facel            = _mm256_set1_ps(fr->ic->epsfac);
1040     charge           = mdatoms->chargeA;
1041     nvdwtype         = fr->ntype;
1042     vdwparam         = fr->nbfp;
1043     vdwtype          = mdatoms->typeA;
1044
1045     sh_ewald         = _mm256_set1_ps(fr->ic->sh_ewald);
1046     beta             = _mm256_set1_ps(fr->ic->ewaldcoeff_q);
1047     beta2            = _mm256_mul_ps(beta,beta);
1048     beta3            = _mm256_mul_ps(beta,beta2);
1049
1050     ewtab            = fr->ic->tabq_coul_FDV0;
1051     ewtabscale       = _mm256_set1_ps(fr->ic->tabq_scale);
1052     ewtabhalfspace   = _mm256_set1_ps(0.5/fr->ic->tabq_scale);
1053
1054     /* Setup water-specific parameters */
1055     inr              = nlist->iinr[0];
1056     iq1              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+1]));
1057     iq2              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+2]));
1058     iq3              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+3]));
1059     vdwioffsetptr0   = vdwparam+2*nvdwtype*vdwtype[inr+0];
1060
1061     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
1062     rcutoff_scalar   = fr->ic->rcoulomb;
1063     rcutoff          = _mm256_set1_ps(rcutoff_scalar);
1064     rcutoff2         = _mm256_mul_ps(rcutoff,rcutoff);
1065
1066     rswitch_scalar   = fr->ic->rcoulomb_switch;
1067     rswitch          = _mm256_set1_ps(rswitch_scalar);
1068     /* Setup switch parameters */
1069     d_scalar         = rcutoff_scalar-rswitch_scalar;
1070     d                = _mm256_set1_ps(d_scalar);
1071     swV3             = _mm256_set1_ps(-10.0/(d_scalar*d_scalar*d_scalar));
1072     swV4             = _mm256_set1_ps( 15.0/(d_scalar*d_scalar*d_scalar*d_scalar));
1073     swV5             = _mm256_set1_ps( -6.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
1074     swF2             = _mm256_set1_ps(-30.0/(d_scalar*d_scalar*d_scalar));
1075     swF3             = _mm256_set1_ps( 60.0/(d_scalar*d_scalar*d_scalar*d_scalar));
1076     swF4             = _mm256_set1_ps(-30.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
1077
1078     /* Avoid stupid compiler warnings */
1079     jnrA = jnrB = jnrC = jnrD = jnrE = jnrF = jnrG = jnrH = 0;
1080     j_coord_offsetA = 0;
1081     j_coord_offsetB = 0;
1082     j_coord_offsetC = 0;
1083     j_coord_offsetD = 0;
1084     j_coord_offsetE = 0;
1085     j_coord_offsetF = 0;
1086     j_coord_offsetG = 0;
1087     j_coord_offsetH = 0;
1088
1089     outeriter        = 0;
1090     inneriter        = 0;
1091
1092     for(iidx=0;iidx<4*DIM;iidx++)
1093     {
1094         scratch[iidx] = 0.0;
1095     }
1096
1097     /* Start outer loop over neighborlists */
1098     for(iidx=0; iidx<nri; iidx++)
1099     {
1100         /* Load shift vector for this list */
1101         i_shift_offset   = DIM*shiftidx[iidx];
1102
1103         /* Load limits for loop over neighbors */
1104         j_index_start    = jindex[iidx];
1105         j_index_end      = jindex[iidx+1];
1106
1107         /* Get outer coordinate index */
1108         inr              = iinr[iidx];
1109         i_coord_offset   = DIM*inr;
1110
1111         /* Load i particle coords and add shift vector */
1112         gmx_mm256_load_shift_and_4rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,
1113                                                     &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
1114
1115         fix0             = _mm256_setzero_ps();
1116         fiy0             = _mm256_setzero_ps();
1117         fiz0             = _mm256_setzero_ps();
1118         fix1             = _mm256_setzero_ps();
1119         fiy1             = _mm256_setzero_ps();
1120         fiz1             = _mm256_setzero_ps();
1121         fix2             = _mm256_setzero_ps();
1122         fiy2             = _mm256_setzero_ps();
1123         fiz2             = _mm256_setzero_ps();
1124         fix3             = _mm256_setzero_ps();
1125         fiy3             = _mm256_setzero_ps();
1126         fiz3             = _mm256_setzero_ps();
1127
1128         /* Start inner kernel loop */
1129         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+7]>=0; jidx+=8)
1130         {
1131
1132             /* Get j neighbor index, and coordinate index */
1133             jnrA             = jjnr[jidx];
1134             jnrB             = jjnr[jidx+1];
1135             jnrC             = jjnr[jidx+2];
1136             jnrD             = jjnr[jidx+3];
1137             jnrE             = jjnr[jidx+4];
1138             jnrF             = jjnr[jidx+5];
1139             jnrG             = jjnr[jidx+6];
1140             jnrH             = jjnr[jidx+7];
1141             j_coord_offsetA  = DIM*jnrA;
1142             j_coord_offsetB  = DIM*jnrB;
1143             j_coord_offsetC  = DIM*jnrC;
1144             j_coord_offsetD  = DIM*jnrD;
1145             j_coord_offsetE  = DIM*jnrE;
1146             j_coord_offsetF  = DIM*jnrF;
1147             j_coord_offsetG  = DIM*jnrG;
1148             j_coord_offsetH  = DIM*jnrH;
1149
1150             /* load j atom coordinates */
1151             gmx_mm256_load_1rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
1152                                                  x+j_coord_offsetC,x+j_coord_offsetD,
1153                                                  x+j_coord_offsetE,x+j_coord_offsetF,
1154                                                  x+j_coord_offsetG,x+j_coord_offsetH,
1155                                                  &jx0,&jy0,&jz0);
1156
1157             /* Calculate displacement vector */
1158             dx00             = _mm256_sub_ps(ix0,jx0);
1159             dy00             = _mm256_sub_ps(iy0,jy0);
1160             dz00             = _mm256_sub_ps(iz0,jz0);
1161             dx10             = _mm256_sub_ps(ix1,jx0);
1162             dy10             = _mm256_sub_ps(iy1,jy0);
1163             dz10             = _mm256_sub_ps(iz1,jz0);
1164             dx20             = _mm256_sub_ps(ix2,jx0);
1165             dy20             = _mm256_sub_ps(iy2,jy0);
1166             dz20             = _mm256_sub_ps(iz2,jz0);
1167             dx30             = _mm256_sub_ps(ix3,jx0);
1168             dy30             = _mm256_sub_ps(iy3,jy0);
1169             dz30             = _mm256_sub_ps(iz3,jz0);
1170
1171             /* Calculate squared distance and things based on it */
1172             rsq00            = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
1173             rsq10            = gmx_mm256_calc_rsq_ps(dx10,dy10,dz10);
1174             rsq20            = gmx_mm256_calc_rsq_ps(dx20,dy20,dz20);
1175             rsq30            = gmx_mm256_calc_rsq_ps(dx30,dy30,dz30);
1176
1177             rinv00           = avx256_invsqrt_f(rsq00);
1178             rinv10           = avx256_invsqrt_f(rsq10);
1179             rinv20           = avx256_invsqrt_f(rsq20);
1180             rinv30           = avx256_invsqrt_f(rsq30);
1181
1182             rinvsq00         = _mm256_mul_ps(rinv00,rinv00);
1183             rinvsq10         = _mm256_mul_ps(rinv10,rinv10);
1184             rinvsq20         = _mm256_mul_ps(rinv20,rinv20);
1185             rinvsq30         = _mm256_mul_ps(rinv30,rinv30);
1186
1187             /* Load parameters for j particles */
1188             jq0              = gmx_mm256_load_8real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
1189                                                                  charge+jnrC+0,charge+jnrD+0,
1190                                                                  charge+jnrE+0,charge+jnrF+0,
1191                                                                  charge+jnrG+0,charge+jnrH+0);
1192             vdwjidx0A        = 2*vdwtype[jnrA+0];
1193             vdwjidx0B        = 2*vdwtype[jnrB+0];
1194             vdwjidx0C        = 2*vdwtype[jnrC+0];
1195             vdwjidx0D        = 2*vdwtype[jnrD+0];
1196             vdwjidx0E        = 2*vdwtype[jnrE+0];
1197             vdwjidx0F        = 2*vdwtype[jnrF+0];
1198             vdwjidx0G        = 2*vdwtype[jnrG+0];
1199             vdwjidx0H        = 2*vdwtype[jnrH+0];
1200
1201             fjx0             = _mm256_setzero_ps();
1202             fjy0             = _mm256_setzero_ps();
1203             fjz0             = _mm256_setzero_ps();
1204
1205             /**************************
1206              * CALCULATE INTERACTIONS *
1207              **************************/
1208
1209             if (gmx_mm256_any_lt(rsq00,rcutoff2))
1210             {
1211
1212             r00              = _mm256_mul_ps(rsq00,rinv00);
1213
1214             /* Compute parameters for interactions between i and j atoms */
1215             gmx_mm256_load_8pair_swizzle_ps(vdwioffsetptr0+vdwjidx0A,
1216                                             vdwioffsetptr0+vdwjidx0B,
1217                                             vdwioffsetptr0+vdwjidx0C,
1218                                             vdwioffsetptr0+vdwjidx0D,
1219                                             vdwioffsetptr0+vdwjidx0E,
1220                                             vdwioffsetptr0+vdwjidx0F,
1221                                             vdwioffsetptr0+vdwjidx0G,
1222                                             vdwioffsetptr0+vdwjidx0H,
1223                                             &c6_00,&c12_00);
1224
1225             /* LENNARD-JONES DISPERSION/REPULSION */
1226
1227             rinvsix          = _mm256_mul_ps(_mm256_mul_ps(rinvsq00,rinvsq00),rinvsq00);
1228             vvdw6            = _mm256_mul_ps(c6_00,rinvsix);
1229             vvdw12           = _mm256_mul_ps(c12_00,_mm256_mul_ps(rinvsix,rinvsix));
1230             vvdw             = _mm256_sub_ps( _mm256_mul_ps(vvdw12,one_twelfth) , _mm256_mul_ps(vvdw6,one_sixth) );
1231             fvdw             = _mm256_mul_ps(_mm256_sub_ps(vvdw12,vvdw6),rinvsq00);
1232
1233             d                = _mm256_sub_ps(r00,rswitch);
1234             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1235             d2               = _mm256_mul_ps(d,d);
1236             sw               = _mm256_add_ps(one,_mm256_mul_ps(d2,_mm256_mul_ps(d,_mm256_add_ps(swV3,_mm256_mul_ps(d,_mm256_add_ps(swV4,_mm256_mul_ps(d,swV5)))))));
1237
1238             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1239
1240             /* Evaluate switch function */
1241             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1242             fvdw             = _mm256_sub_ps( _mm256_mul_ps(fvdw,sw) , _mm256_mul_ps(rinv00,_mm256_mul_ps(vvdw,dsw)) );
1243             cutoff_mask      = _mm256_cmp_ps(rsq00,rcutoff2,_CMP_LT_OQ);
1244
1245             fscal            = fvdw;
1246
1247             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1248
1249             /* Calculate temporary vectorial force */
1250             tx               = _mm256_mul_ps(fscal,dx00);
1251             ty               = _mm256_mul_ps(fscal,dy00);
1252             tz               = _mm256_mul_ps(fscal,dz00);
1253
1254             /* Update vectorial force */
1255             fix0             = _mm256_add_ps(fix0,tx);
1256             fiy0             = _mm256_add_ps(fiy0,ty);
1257             fiz0             = _mm256_add_ps(fiz0,tz);
1258
1259             fjx0             = _mm256_add_ps(fjx0,tx);
1260             fjy0             = _mm256_add_ps(fjy0,ty);
1261             fjz0             = _mm256_add_ps(fjz0,tz);
1262
1263             }
1264
1265             /**************************
1266              * CALCULATE INTERACTIONS *
1267              **************************/
1268
1269             if (gmx_mm256_any_lt(rsq10,rcutoff2))
1270             {
1271
1272             r10              = _mm256_mul_ps(rsq10,rinv10);
1273
1274             /* Compute parameters for interactions between i and j atoms */
1275             qq10             = _mm256_mul_ps(iq1,jq0);
1276
1277             /* EWALD ELECTROSTATICS */
1278
1279             /* Analytical PME correction */
1280             zeta2            = _mm256_mul_ps(beta2,rsq10);
1281             rinv3            = _mm256_mul_ps(rinvsq10,rinv10);
1282             pmecorrF         = avx256_pmecorrF_f(zeta2);
1283             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1284             felec            = _mm256_mul_ps(qq10,felec);
1285             pmecorrV         = avx256_pmecorrV_f(zeta2);
1286             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1287             velec            = _mm256_sub_ps(rinv10,pmecorrV);
1288             velec            = _mm256_mul_ps(qq10,velec);
1289
1290             d                = _mm256_sub_ps(r10,rswitch);
1291             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1292             d2               = _mm256_mul_ps(d,d);
1293             sw               = _mm256_add_ps(one,_mm256_mul_ps(d2,_mm256_mul_ps(d,_mm256_add_ps(swV3,_mm256_mul_ps(d,_mm256_add_ps(swV4,_mm256_mul_ps(d,swV5)))))));
1294
1295             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1296
1297             /* Evaluate switch function */
1298             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1299             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv10,_mm256_mul_ps(velec,dsw)) );
1300             cutoff_mask      = _mm256_cmp_ps(rsq10,rcutoff2,_CMP_LT_OQ);
1301
1302             fscal            = felec;
1303
1304             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1305
1306             /* Calculate temporary vectorial force */
1307             tx               = _mm256_mul_ps(fscal,dx10);
1308             ty               = _mm256_mul_ps(fscal,dy10);
1309             tz               = _mm256_mul_ps(fscal,dz10);
1310
1311             /* Update vectorial force */
1312             fix1             = _mm256_add_ps(fix1,tx);
1313             fiy1             = _mm256_add_ps(fiy1,ty);
1314             fiz1             = _mm256_add_ps(fiz1,tz);
1315
1316             fjx0             = _mm256_add_ps(fjx0,tx);
1317             fjy0             = _mm256_add_ps(fjy0,ty);
1318             fjz0             = _mm256_add_ps(fjz0,tz);
1319
1320             }
1321
1322             /**************************
1323              * CALCULATE INTERACTIONS *
1324              **************************/
1325
1326             if (gmx_mm256_any_lt(rsq20,rcutoff2))
1327             {
1328
1329             r20              = _mm256_mul_ps(rsq20,rinv20);
1330
1331             /* Compute parameters for interactions between i and j atoms */
1332             qq20             = _mm256_mul_ps(iq2,jq0);
1333
1334             /* EWALD ELECTROSTATICS */
1335
1336             /* Analytical PME correction */
1337             zeta2            = _mm256_mul_ps(beta2,rsq20);
1338             rinv3            = _mm256_mul_ps(rinvsq20,rinv20);
1339             pmecorrF         = avx256_pmecorrF_f(zeta2);
1340             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1341             felec            = _mm256_mul_ps(qq20,felec);
1342             pmecorrV         = avx256_pmecorrV_f(zeta2);
1343             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1344             velec            = _mm256_sub_ps(rinv20,pmecorrV);
1345             velec            = _mm256_mul_ps(qq20,velec);
1346
1347             d                = _mm256_sub_ps(r20,rswitch);
1348             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1349             d2               = _mm256_mul_ps(d,d);
1350             sw               = _mm256_add_ps(one,_mm256_mul_ps(d2,_mm256_mul_ps(d,_mm256_add_ps(swV3,_mm256_mul_ps(d,_mm256_add_ps(swV4,_mm256_mul_ps(d,swV5)))))));
1351
1352             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1353
1354             /* Evaluate switch function */
1355             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1356             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv20,_mm256_mul_ps(velec,dsw)) );
1357             cutoff_mask      = _mm256_cmp_ps(rsq20,rcutoff2,_CMP_LT_OQ);
1358
1359             fscal            = felec;
1360
1361             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1362
1363             /* Calculate temporary vectorial force */
1364             tx               = _mm256_mul_ps(fscal,dx20);
1365             ty               = _mm256_mul_ps(fscal,dy20);
1366             tz               = _mm256_mul_ps(fscal,dz20);
1367
1368             /* Update vectorial force */
1369             fix2             = _mm256_add_ps(fix2,tx);
1370             fiy2             = _mm256_add_ps(fiy2,ty);
1371             fiz2             = _mm256_add_ps(fiz2,tz);
1372
1373             fjx0             = _mm256_add_ps(fjx0,tx);
1374             fjy0             = _mm256_add_ps(fjy0,ty);
1375             fjz0             = _mm256_add_ps(fjz0,tz);
1376
1377             }
1378
1379             /**************************
1380              * CALCULATE INTERACTIONS *
1381              **************************/
1382
1383             if (gmx_mm256_any_lt(rsq30,rcutoff2))
1384             {
1385
1386             r30              = _mm256_mul_ps(rsq30,rinv30);
1387
1388             /* Compute parameters for interactions between i and j atoms */
1389             qq30             = _mm256_mul_ps(iq3,jq0);
1390
1391             /* EWALD ELECTROSTATICS */
1392
1393             /* Analytical PME correction */
1394             zeta2            = _mm256_mul_ps(beta2,rsq30);
1395             rinv3            = _mm256_mul_ps(rinvsq30,rinv30);
1396             pmecorrF         = avx256_pmecorrF_f(zeta2);
1397             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1398             felec            = _mm256_mul_ps(qq30,felec);
1399             pmecorrV         = avx256_pmecorrV_f(zeta2);
1400             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1401             velec            = _mm256_sub_ps(rinv30,pmecorrV);
1402             velec            = _mm256_mul_ps(qq30,velec);
1403
1404             d                = _mm256_sub_ps(r30,rswitch);
1405             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1406             d2               = _mm256_mul_ps(d,d);
1407             sw               = _mm256_add_ps(one,_mm256_mul_ps(d2,_mm256_mul_ps(d,_mm256_add_ps(swV3,_mm256_mul_ps(d,_mm256_add_ps(swV4,_mm256_mul_ps(d,swV5)))))));
1408
1409             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1410
1411             /* Evaluate switch function */
1412             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1413             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv30,_mm256_mul_ps(velec,dsw)) );
1414             cutoff_mask      = _mm256_cmp_ps(rsq30,rcutoff2,_CMP_LT_OQ);
1415
1416             fscal            = felec;
1417
1418             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1419
1420             /* Calculate temporary vectorial force */
1421             tx               = _mm256_mul_ps(fscal,dx30);
1422             ty               = _mm256_mul_ps(fscal,dy30);
1423             tz               = _mm256_mul_ps(fscal,dz30);
1424
1425             /* Update vectorial force */
1426             fix3             = _mm256_add_ps(fix3,tx);
1427             fiy3             = _mm256_add_ps(fiy3,ty);
1428             fiz3             = _mm256_add_ps(fiz3,tz);
1429
1430             fjx0             = _mm256_add_ps(fjx0,tx);
1431             fjy0             = _mm256_add_ps(fjy0,ty);
1432             fjz0             = _mm256_add_ps(fjz0,tz);
1433
1434             }
1435
1436             fjptrA             = f+j_coord_offsetA;
1437             fjptrB             = f+j_coord_offsetB;
1438             fjptrC             = f+j_coord_offsetC;
1439             fjptrD             = f+j_coord_offsetD;
1440             fjptrE             = f+j_coord_offsetE;
1441             fjptrF             = f+j_coord_offsetF;
1442             fjptrG             = f+j_coord_offsetG;
1443             fjptrH             = f+j_coord_offsetH;
1444
1445             gmx_mm256_decrement_1rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,fjx0,fjy0,fjz0);
1446
1447             /* Inner loop uses 374 flops */
1448         }
1449
1450         if(jidx<j_index_end)
1451         {
1452
1453             /* Get j neighbor index, and coordinate index */
1454             jnrlistA         = jjnr[jidx];
1455             jnrlistB         = jjnr[jidx+1];
1456             jnrlistC         = jjnr[jidx+2];
1457             jnrlistD         = jjnr[jidx+3];
1458             jnrlistE         = jjnr[jidx+4];
1459             jnrlistF         = jjnr[jidx+5];
1460             jnrlistG         = jjnr[jidx+6];
1461             jnrlistH         = jjnr[jidx+7];
1462             /* Sign of each element will be negative for non-real atoms.
1463              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
1464              * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
1465              */
1466             dummy_mask = gmx_mm256_set_m128(gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx+4)),_mm_setzero_si128())),
1467                                             gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128())));
1468
1469             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
1470             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
1471             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
1472             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
1473             jnrE       = (jnrlistE>=0) ? jnrlistE : 0;
1474             jnrF       = (jnrlistF>=0) ? jnrlistF : 0;
1475             jnrG       = (jnrlistG>=0) ? jnrlistG : 0;
1476             jnrH       = (jnrlistH>=0) ? jnrlistH : 0;
1477             j_coord_offsetA  = DIM*jnrA;
1478             j_coord_offsetB  = DIM*jnrB;
1479             j_coord_offsetC  = DIM*jnrC;
1480             j_coord_offsetD  = DIM*jnrD;
1481             j_coord_offsetE  = DIM*jnrE;
1482             j_coord_offsetF  = DIM*jnrF;
1483             j_coord_offsetG  = DIM*jnrG;
1484             j_coord_offsetH  = DIM*jnrH;
1485
1486             /* load j atom coordinates */
1487             gmx_mm256_load_1rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
1488                                                  x+j_coord_offsetC,x+j_coord_offsetD,
1489                                                  x+j_coord_offsetE,x+j_coord_offsetF,
1490                                                  x+j_coord_offsetG,x+j_coord_offsetH,
1491                                                  &jx0,&jy0,&jz0);
1492
1493             /* Calculate displacement vector */
1494             dx00             = _mm256_sub_ps(ix0,jx0);
1495             dy00             = _mm256_sub_ps(iy0,jy0);
1496             dz00             = _mm256_sub_ps(iz0,jz0);
1497             dx10             = _mm256_sub_ps(ix1,jx0);
1498             dy10             = _mm256_sub_ps(iy1,jy0);
1499             dz10             = _mm256_sub_ps(iz1,jz0);
1500             dx20             = _mm256_sub_ps(ix2,jx0);
1501             dy20             = _mm256_sub_ps(iy2,jy0);
1502             dz20             = _mm256_sub_ps(iz2,jz0);
1503             dx30             = _mm256_sub_ps(ix3,jx0);
1504             dy30             = _mm256_sub_ps(iy3,jy0);
1505             dz30             = _mm256_sub_ps(iz3,jz0);
1506
1507             /* Calculate squared distance and things based on it */
1508             rsq00            = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
1509             rsq10            = gmx_mm256_calc_rsq_ps(dx10,dy10,dz10);
1510             rsq20            = gmx_mm256_calc_rsq_ps(dx20,dy20,dz20);
1511             rsq30            = gmx_mm256_calc_rsq_ps(dx30,dy30,dz30);
1512
1513             rinv00           = avx256_invsqrt_f(rsq00);
1514             rinv10           = avx256_invsqrt_f(rsq10);
1515             rinv20           = avx256_invsqrt_f(rsq20);
1516             rinv30           = avx256_invsqrt_f(rsq30);
1517
1518             rinvsq00         = _mm256_mul_ps(rinv00,rinv00);
1519             rinvsq10         = _mm256_mul_ps(rinv10,rinv10);
1520             rinvsq20         = _mm256_mul_ps(rinv20,rinv20);
1521             rinvsq30         = _mm256_mul_ps(rinv30,rinv30);
1522
1523             /* Load parameters for j particles */
1524             jq0              = gmx_mm256_load_8real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
1525                                                                  charge+jnrC+0,charge+jnrD+0,
1526                                                                  charge+jnrE+0,charge+jnrF+0,
1527                                                                  charge+jnrG+0,charge+jnrH+0);
1528             vdwjidx0A        = 2*vdwtype[jnrA+0];
1529             vdwjidx0B        = 2*vdwtype[jnrB+0];
1530             vdwjidx0C        = 2*vdwtype[jnrC+0];
1531             vdwjidx0D        = 2*vdwtype[jnrD+0];
1532             vdwjidx0E        = 2*vdwtype[jnrE+0];
1533             vdwjidx0F        = 2*vdwtype[jnrF+0];
1534             vdwjidx0G        = 2*vdwtype[jnrG+0];
1535             vdwjidx0H        = 2*vdwtype[jnrH+0];
1536
1537             fjx0             = _mm256_setzero_ps();
1538             fjy0             = _mm256_setzero_ps();
1539             fjz0             = _mm256_setzero_ps();
1540
1541             /**************************
1542              * CALCULATE INTERACTIONS *
1543              **************************/
1544
1545             if (gmx_mm256_any_lt(rsq00,rcutoff2))
1546             {
1547
1548             r00              = _mm256_mul_ps(rsq00,rinv00);
1549             r00              = _mm256_andnot_ps(dummy_mask,r00);
1550
1551             /* Compute parameters for interactions between i and j atoms */
1552             gmx_mm256_load_8pair_swizzle_ps(vdwioffsetptr0+vdwjidx0A,
1553                                             vdwioffsetptr0+vdwjidx0B,
1554                                             vdwioffsetptr0+vdwjidx0C,
1555                                             vdwioffsetptr0+vdwjidx0D,
1556                                             vdwioffsetptr0+vdwjidx0E,
1557                                             vdwioffsetptr0+vdwjidx0F,
1558                                             vdwioffsetptr0+vdwjidx0G,
1559                                             vdwioffsetptr0+vdwjidx0H,
1560                                             &c6_00,&c12_00);
1561
1562             /* LENNARD-JONES DISPERSION/REPULSION */
1563
1564             rinvsix          = _mm256_mul_ps(_mm256_mul_ps(rinvsq00,rinvsq00),rinvsq00);
1565             vvdw6            = _mm256_mul_ps(c6_00,rinvsix);
1566             vvdw12           = _mm256_mul_ps(c12_00,_mm256_mul_ps(rinvsix,rinvsix));
1567             vvdw             = _mm256_sub_ps( _mm256_mul_ps(vvdw12,one_twelfth) , _mm256_mul_ps(vvdw6,one_sixth) );
1568             fvdw             = _mm256_mul_ps(_mm256_sub_ps(vvdw12,vvdw6),rinvsq00);
1569
1570             d                = _mm256_sub_ps(r00,rswitch);
1571             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1572             d2               = _mm256_mul_ps(d,d);
1573             sw               = _mm256_add_ps(one,_mm256_mul_ps(d2,_mm256_mul_ps(d,_mm256_add_ps(swV3,_mm256_mul_ps(d,_mm256_add_ps(swV4,_mm256_mul_ps(d,swV5)))))));
1574
1575             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1576
1577             /* Evaluate switch function */
1578             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1579             fvdw             = _mm256_sub_ps( _mm256_mul_ps(fvdw,sw) , _mm256_mul_ps(rinv00,_mm256_mul_ps(vvdw,dsw)) );
1580             cutoff_mask      = _mm256_cmp_ps(rsq00,rcutoff2,_CMP_LT_OQ);
1581
1582             fscal            = fvdw;
1583
1584             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1585
1586             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1587
1588             /* Calculate temporary vectorial force */
1589             tx               = _mm256_mul_ps(fscal,dx00);
1590             ty               = _mm256_mul_ps(fscal,dy00);
1591             tz               = _mm256_mul_ps(fscal,dz00);
1592
1593             /* Update vectorial force */
1594             fix0             = _mm256_add_ps(fix0,tx);
1595             fiy0             = _mm256_add_ps(fiy0,ty);
1596             fiz0             = _mm256_add_ps(fiz0,tz);
1597
1598             fjx0             = _mm256_add_ps(fjx0,tx);
1599             fjy0             = _mm256_add_ps(fjy0,ty);
1600             fjz0             = _mm256_add_ps(fjz0,tz);
1601
1602             }
1603
1604             /**************************
1605              * CALCULATE INTERACTIONS *
1606              **************************/
1607
1608             if (gmx_mm256_any_lt(rsq10,rcutoff2))
1609             {
1610
1611             r10              = _mm256_mul_ps(rsq10,rinv10);
1612             r10              = _mm256_andnot_ps(dummy_mask,r10);
1613
1614             /* Compute parameters for interactions between i and j atoms */
1615             qq10             = _mm256_mul_ps(iq1,jq0);
1616
1617             /* EWALD ELECTROSTATICS */
1618
1619             /* Analytical PME correction */
1620             zeta2            = _mm256_mul_ps(beta2,rsq10);
1621             rinv3            = _mm256_mul_ps(rinvsq10,rinv10);
1622             pmecorrF         = avx256_pmecorrF_f(zeta2);
1623             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1624             felec            = _mm256_mul_ps(qq10,felec);
1625             pmecorrV         = avx256_pmecorrV_f(zeta2);
1626             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1627             velec            = _mm256_sub_ps(rinv10,pmecorrV);
1628             velec            = _mm256_mul_ps(qq10,velec);
1629
1630             d                = _mm256_sub_ps(r10,rswitch);
1631             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1632             d2               = _mm256_mul_ps(d,d);
1633             sw               = _mm256_add_ps(one,_mm256_mul_ps(d2,_mm256_mul_ps(d,_mm256_add_ps(swV3,_mm256_mul_ps(d,_mm256_add_ps(swV4,_mm256_mul_ps(d,swV5)))))));
1634
1635             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1636
1637             /* Evaluate switch function */
1638             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1639             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv10,_mm256_mul_ps(velec,dsw)) );
1640             cutoff_mask      = _mm256_cmp_ps(rsq10,rcutoff2,_CMP_LT_OQ);
1641
1642             fscal            = felec;
1643
1644             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1645
1646             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1647
1648             /* Calculate temporary vectorial force */
1649             tx               = _mm256_mul_ps(fscal,dx10);
1650             ty               = _mm256_mul_ps(fscal,dy10);
1651             tz               = _mm256_mul_ps(fscal,dz10);
1652
1653             /* Update vectorial force */
1654             fix1             = _mm256_add_ps(fix1,tx);
1655             fiy1             = _mm256_add_ps(fiy1,ty);
1656             fiz1             = _mm256_add_ps(fiz1,tz);
1657
1658             fjx0             = _mm256_add_ps(fjx0,tx);
1659             fjy0             = _mm256_add_ps(fjy0,ty);
1660             fjz0             = _mm256_add_ps(fjz0,tz);
1661
1662             }
1663
1664             /**************************
1665              * CALCULATE INTERACTIONS *
1666              **************************/
1667
1668             if (gmx_mm256_any_lt(rsq20,rcutoff2))
1669             {
1670
1671             r20              = _mm256_mul_ps(rsq20,rinv20);
1672             r20              = _mm256_andnot_ps(dummy_mask,r20);
1673
1674             /* Compute parameters for interactions between i and j atoms */
1675             qq20             = _mm256_mul_ps(iq2,jq0);
1676
1677             /* EWALD ELECTROSTATICS */
1678
1679             /* Analytical PME correction */
1680             zeta2            = _mm256_mul_ps(beta2,rsq20);
1681             rinv3            = _mm256_mul_ps(rinvsq20,rinv20);
1682             pmecorrF         = avx256_pmecorrF_f(zeta2);
1683             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1684             felec            = _mm256_mul_ps(qq20,felec);
1685             pmecorrV         = avx256_pmecorrV_f(zeta2);
1686             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1687             velec            = _mm256_sub_ps(rinv20,pmecorrV);
1688             velec            = _mm256_mul_ps(qq20,velec);
1689
1690             d                = _mm256_sub_ps(r20,rswitch);
1691             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1692             d2               = _mm256_mul_ps(d,d);
1693             sw               = _mm256_add_ps(one,_mm256_mul_ps(d2,_mm256_mul_ps(d,_mm256_add_ps(swV3,_mm256_mul_ps(d,_mm256_add_ps(swV4,_mm256_mul_ps(d,swV5)))))));
1694
1695             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1696
1697             /* Evaluate switch function */
1698             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1699             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv20,_mm256_mul_ps(velec,dsw)) );
1700             cutoff_mask      = _mm256_cmp_ps(rsq20,rcutoff2,_CMP_LT_OQ);
1701
1702             fscal            = felec;
1703
1704             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1705
1706             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1707
1708             /* Calculate temporary vectorial force */
1709             tx               = _mm256_mul_ps(fscal,dx20);
1710             ty               = _mm256_mul_ps(fscal,dy20);
1711             tz               = _mm256_mul_ps(fscal,dz20);
1712
1713             /* Update vectorial force */
1714             fix2             = _mm256_add_ps(fix2,tx);
1715             fiy2             = _mm256_add_ps(fiy2,ty);
1716             fiz2             = _mm256_add_ps(fiz2,tz);
1717
1718             fjx0             = _mm256_add_ps(fjx0,tx);
1719             fjy0             = _mm256_add_ps(fjy0,ty);
1720             fjz0             = _mm256_add_ps(fjz0,tz);
1721
1722             }
1723
1724             /**************************
1725              * CALCULATE INTERACTIONS *
1726              **************************/
1727
1728             if (gmx_mm256_any_lt(rsq30,rcutoff2))
1729             {
1730
1731             r30              = _mm256_mul_ps(rsq30,rinv30);
1732             r30              = _mm256_andnot_ps(dummy_mask,r30);
1733
1734             /* Compute parameters for interactions between i and j atoms */
1735             qq30             = _mm256_mul_ps(iq3,jq0);
1736
1737             /* EWALD ELECTROSTATICS */
1738
1739             /* Analytical PME correction */
1740             zeta2            = _mm256_mul_ps(beta2,rsq30);
1741             rinv3            = _mm256_mul_ps(rinvsq30,rinv30);
1742             pmecorrF         = avx256_pmecorrF_f(zeta2);
1743             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1744             felec            = _mm256_mul_ps(qq30,felec);
1745             pmecorrV         = avx256_pmecorrV_f(zeta2);
1746             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1747             velec            = _mm256_sub_ps(rinv30,pmecorrV);
1748             velec            = _mm256_mul_ps(qq30,velec);
1749
1750             d                = _mm256_sub_ps(r30,rswitch);
1751             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1752             d2               = _mm256_mul_ps(d,d);
1753             sw               = _mm256_add_ps(one,_mm256_mul_ps(d2,_mm256_mul_ps(d,_mm256_add_ps(swV3,_mm256_mul_ps(d,_mm256_add_ps(swV4,_mm256_mul_ps(d,swV5)))))));
1754
1755             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1756
1757             /* Evaluate switch function */
1758             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1759             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv30,_mm256_mul_ps(velec,dsw)) );
1760             cutoff_mask      = _mm256_cmp_ps(rsq30,rcutoff2,_CMP_LT_OQ);
1761
1762             fscal            = felec;
1763
1764             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1765
1766             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1767
1768             /* Calculate temporary vectorial force */
1769             tx               = _mm256_mul_ps(fscal,dx30);
1770             ty               = _mm256_mul_ps(fscal,dy30);
1771             tz               = _mm256_mul_ps(fscal,dz30);
1772
1773             /* Update vectorial force */
1774             fix3             = _mm256_add_ps(fix3,tx);
1775             fiy3             = _mm256_add_ps(fiy3,ty);
1776             fiz3             = _mm256_add_ps(fiz3,tz);
1777
1778             fjx0             = _mm256_add_ps(fjx0,tx);
1779             fjy0             = _mm256_add_ps(fjy0,ty);
1780             fjz0             = _mm256_add_ps(fjz0,tz);
1781
1782             }
1783
1784             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
1785             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
1786             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
1787             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
1788             fjptrE             = (jnrlistE>=0) ? f+j_coord_offsetE : scratch;
1789             fjptrF             = (jnrlistF>=0) ? f+j_coord_offsetF : scratch;
1790             fjptrG             = (jnrlistG>=0) ? f+j_coord_offsetG : scratch;
1791             fjptrH             = (jnrlistH>=0) ? f+j_coord_offsetH : scratch;
1792
1793             gmx_mm256_decrement_1rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,fjx0,fjy0,fjz0);
1794
1795             /* Inner loop uses 378 flops */
1796         }
1797
1798         /* End of innermost loop */
1799
1800         gmx_mm256_update_iforce_4atom_swizzle_ps(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
1801                                                  f+i_coord_offset,fshift+i_shift_offset);
1802
1803         /* Increment number of inner iterations */
1804         inneriter                  += j_index_end - j_index_start;
1805
1806         /* Outer loop uses 24 flops */
1807     }
1808
1809     /* Increment number of outer iterations */
1810     outeriter        += nri;
1811
1812     /* Update outer/inner flops */
1813
1814     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_F,outeriter*24 + inneriter*378);
1815 }