src/gromacs/gmxlib/nonbonded/nb_kernel_avx_256_single/nb_kernel_ElecEwSw_VdwLJSw_GeomW4P1_avx_256_single.c

   1 /*
   2  * This file is part of the GROMACS molecular simulation package.
   3  *
   4  * Copyright (c) 2012,2013,2014, by the GROMACS development team, led by
   5  * Mark Abraham, David van der Spoel, Berk Hess, and Erik Lindahl,
   6  * and including many others, as listed in the AUTHORS file in the
   7  * top-level source directory and at http://www.gromacs.org.
   8  *
   9  * GROMACS is free software; you can redistribute it and/or
  10  * modify it under the terms of the GNU Lesser General Public License
  11  * as published by the Free Software Foundation; either version 2.1
  12  * of the License, or (at your option) any later version.
  13  *
  14  * GROMACS is distributed in the hope that it will be useful,
  15  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  16  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  17  * Lesser General Public License for more details.
  18  *
  19  * You should have received a copy of the GNU Lesser General Public
  20  * License along with GROMACS; if not, see
  21  * http://www.gnu.org/licenses, or write to the Free Software Foundation,
  22  * Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA.
  23  *
  24  * If you want to redistribute modifications to GROMACS, please
  25  * consider that scientific software is very special. Version
  26  * control is crucial - bugs must be traceable. We will be happy to
  27  * consider code for inclusion in the official distribution, but
  28  * derived work must not be called official GROMACS. Details are found
  29  * in the README & COPYING files - if they are missing, get the
  30  * official version at http://www.gromacs.org.
  31  *
  32  * To help us fund GROMACS development, we humbly ask that you cite
  33  * the research papers on the package. Check out http://www.gromacs.org.
  34  */
  35 /*
  36  * Note: this file was generated by the GROMACS avx_256_single kernel generator.
  37  */
  38 #include "config.h"
  39
  40 #include <math.h>
  41
  42 #include "../nb_kernel.h"
  43 #include "types/simple.h"
  44 #include "gromacs/math/vec.h"
  45 #include "nrnb.h"
  46
  47 #include "gromacs/simd/math_x86_avx_256_single.h"
  48 #include "kernelutil_x86_avx_256_single.h"
  49
  50 /*
  51  * Gromacs nonbonded kernel:   nb_kernel_ElecEwSw_VdwLJSw_GeomW4P1_VF_avx_256_single
  52  * Electrostatics interaction: Ewald
  53  * VdW interaction:            LennardJones
  54  * Geometry:                   Water4-Particle
  55  * Calculate force/pot:        PotentialAndForce
  56  */
  57 void
  58 nb_kernel_ElecEwSw_VdwLJSw_GeomW4P1_VF_avx_256_single
  59                     (t_nblist                    * gmx_restrict       nlist,
  60                      rvec                        * gmx_restrict          xx,
  61                      rvec                        * gmx_restrict          ff,
  62                      t_forcerec                  * gmx_restrict          fr,
  63                      t_mdatoms                   * gmx_restrict     mdatoms,
  64                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
  65                      t_nrnb                      * gmx_restrict        nrnb)
  66 {
  67     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
  68      * just 0 for non-waters.
  69      * Suffixes A,B,C,D,E,F,G,H refer to j loop unrolling done with AVX, e.g. for the eight different
  70      * jnr indices corresponding to data put in the four positions in the SIMD register.
  71      */
  72     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
  73     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
  74     int              jnrA,jnrB,jnrC,jnrD;
  75     int              jnrE,jnrF,jnrG,jnrH;
  76     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
  77     int              jnrlistE,jnrlistF,jnrlistG,jnrlistH;
  78     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
  79     int              j_coord_offsetE,j_coord_offsetF,j_coord_offsetG,j_coord_offsetH;
  80     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  81     real             rcutoff_scalar;
  82     real             *shiftvec,*fshift,*x,*f;
  83     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD,*fjptrE,*fjptrF,*fjptrG,*fjptrH;
  84     real             scratch[4*DIM];
  85     __m256           tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
  86     real *           vdwioffsetptr0;
  87     __m256           ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  88     real *           vdwioffsetptr1;
  89     __m256           ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
  90     real *           vdwioffsetptr2;
  91     __m256           ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
  92     real *           vdwioffsetptr3;
  93     __m256           ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
  94     int              vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D,vdwjidx0E,vdwjidx0F,vdwjidx0G,vdwjidx0H;
  95     __m256           jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  96     __m256           dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
  97     __m256           dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
  98     __m256           dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
  99     __m256           dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30;
 100     __m256           velec,felec,velecsum,facel,crf,krf,krf2;
 101     real             *charge;
 102     int              nvdwtype;
 103     __m256           rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
 104     int              *vdwtype;
 105     real             *vdwparam;
 106     __m256           one_sixth   = _mm256_set1_ps(1.0/6.0);
 107     __m256           one_twelfth = _mm256_set1_ps(1.0/12.0);
 108     __m256i          ewitab;
 109     __m128i          ewitab_lo,ewitab_hi;
 110     __m256           ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
 111     __m256           beta,beta2,beta3,zeta2,pmecorrF,pmecorrV,rinv3;
 112     real             *ewtab;
 113     __m256           rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
 114     real             rswitch_scalar,d_scalar;
 115     __m256           dummy_mask,cutoff_mask;
 116     __m256           signbit = _mm256_castsi256_ps( _mm256_set1_epi32(0x80000000) );
 117     __m256           one     = _mm256_set1_ps(1.0);
 118     __m256           two     = _mm256_set1_ps(2.0);
 119     x                = xx[0];
 120     f                = ff[0];
 121
 122     nri              = nlist->nri;
 123     iinr             = nlist->iinr;
 124     jindex           = nlist->jindex;
 125     jjnr             = nlist->jjnr;
 126     shiftidx         = nlist->shift;
 127     gid              = nlist->gid;
 128     shiftvec         = fr->shift_vec[0];
 129     fshift           = fr->fshift[0];
 130     facel            = _mm256_set1_ps(fr->epsfac);
 131     charge           = mdatoms->chargeA;
 132     nvdwtype         = fr->ntype;
 133     vdwparam         = fr->nbfp;
 134     vdwtype          = mdatoms->typeA;
 135
 136     sh_ewald         = _mm256_set1_ps(fr->ic->sh_ewald);
 137     beta             = _mm256_set1_ps(fr->ic->ewaldcoeff_q);
 138     beta2            = _mm256_mul_ps(beta,beta);
 139     beta3            = _mm256_mul_ps(beta,beta2);
 140
 141     ewtab            = fr->ic->tabq_coul_FDV0;
 142     ewtabscale       = _mm256_set1_ps(fr->ic->tabq_scale);
 143     ewtabhalfspace   = _mm256_set1_ps(0.5/fr->ic->tabq_scale);
 144
 145     /* Setup water-specific parameters */
 146     inr              = nlist->iinr[0];
 147     iq1              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+1]));
 148     iq2              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+2]));
 149     iq3              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+3]));
 150     vdwioffsetptr0   = vdwparam+2*nvdwtype*vdwtype[inr+0];
 151
 152     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 153     rcutoff_scalar   = fr->rcoulomb;
 154     rcutoff          = _mm256_set1_ps(rcutoff_scalar);
 155     rcutoff2         = _mm256_mul_ps(rcutoff,rcutoff);
 156
 157     rswitch_scalar   = fr->rcoulomb_switch;
 158     rswitch          = _mm256_set1_ps(rswitch_scalar);
 159     /* Setup switch parameters */
 160     d_scalar         = rcutoff_scalar-rswitch_scalar;
 161     d                = _mm256_set1_ps(d_scalar);
 162     swV3             = _mm256_set1_ps(-10.0/(d_scalar*d_scalar*d_scalar));
 163     swV4             = _mm256_set1_ps( 15.0/(d_scalar*d_scalar*d_scalar*d_scalar));
 164     swV5             = _mm256_set1_ps( -6.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
 165     swF2             = _mm256_set1_ps(-30.0/(d_scalar*d_scalar*d_scalar));
 166     swF3             = _mm256_set1_ps( 60.0/(d_scalar*d_scalar*d_scalar*d_scalar));
 167     swF4             = _mm256_set1_ps(-30.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
 168
 169     /* Avoid stupid compiler warnings */
 170     jnrA = jnrB = jnrC = jnrD = jnrE = jnrF = jnrG = jnrH = 0;
 171     j_coord_offsetA = 0;
 172     j_coord_offsetB = 0;
 173     j_coord_offsetC = 0;
 174     j_coord_offsetD = 0;
 175     j_coord_offsetE = 0;
 176     j_coord_offsetF = 0;
 177     j_coord_offsetG = 0;
 178     j_coord_offsetH = 0;
 179
 180     outeriter        = 0;
 181     inneriter        = 0;
 182
 183     for(iidx=0;iidx<4*DIM;iidx++)
 184     {
 185         scratch[iidx] = 0.0;
 186     }
 187
 188     /* Start outer loop over neighborlists */
 189     for(iidx=0; iidx<nri; iidx++)
 190     {
 191         /* Load shift vector for this list */
 192         i_shift_offset   = DIM*shiftidx[iidx];
 193
 194         /* Load limits for loop over neighbors */
 195         j_index_start    = jindex[iidx];
 196         j_index_end      = jindex[iidx+1];
 197
 198         /* Get outer coordinate index */
 199         inr              = iinr[iidx];
 200         i_coord_offset   = DIM*inr;
 201
 202         /* Load i particle coords and add shift vector */
 203         gmx_mm256_load_shift_and_4rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,
 204                                                     &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
 205
 206         fix0             = _mm256_setzero_ps();
 207         fiy0             = _mm256_setzero_ps();
 208         fiz0             = _mm256_setzero_ps();
 209         fix1             = _mm256_setzero_ps();
 210         fiy1             = _mm256_setzero_ps();
 211         fiz1             = _mm256_setzero_ps();
 212         fix2             = _mm256_setzero_ps();
 213         fiy2             = _mm256_setzero_ps();
 214         fiz2             = _mm256_setzero_ps();
 215         fix3             = _mm256_setzero_ps();
 216         fiy3             = _mm256_setzero_ps();
 217         fiz3             = _mm256_setzero_ps();
 218
 219         /* Reset potential sums */
 220         velecsum         = _mm256_setzero_ps();
 221         vvdwsum          = _mm256_setzero_ps();
 222
 223         /* Start inner kernel loop */
 224         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+7]>=0; jidx+=8)
 225         {
 226
 227             /* Get j neighbor index, and coordinate index */
 228             jnrA             = jjnr[jidx];
 229             jnrB             = jjnr[jidx+1];
 230             jnrC             = jjnr[jidx+2];
 231             jnrD             = jjnr[jidx+3];
 232             jnrE             = jjnr[jidx+4];
 233             jnrF             = jjnr[jidx+5];
 234             jnrG             = jjnr[jidx+6];
 235             jnrH             = jjnr[jidx+7];
 236             j_coord_offsetA  = DIM*jnrA;
 237             j_coord_offsetB  = DIM*jnrB;
 238             j_coord_offsetC  = DIM*jnrC;
 239             j_coord_offsetD  = DIM*jnrD;
 240             j_coord_offsetE  = DIM*jnrE;
 241             j_coord_offsetF  = DIM*jnrF;
 242             j_coord_offsetG  = DIM*jnrG;
 243             j_coord_offsetH  = DIM*jnrH;
 244
 245             /* load j atom coordinates */
 246             gmx_mm256_load_1rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 247                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 248                                                  x+j_coord_offsetE,x+j_coord_offsetF,
 249                                                  x+j_coord_offsetG,x+j_coord_offsetH,
 250                                                  &jx0,&jy0,&jz0);
 251
 252             /* Calculate displacement vector */
 253             dx00             = _mm256_sub_ps(ix0,jx0);
 254             dy00             = _mm256_sub_ps(iy0,jy0);
 255             dz00             = _mm256_sub_ps(iz0,jz0);
 256             dx10             = _mm256_sub_ps(ix1,jx0);
 257             dy10             = _mm256_sub_ps(iy1,jy0);
 258             dz10             = _mm256_sub_ps(iz1,jz0);
 259             dx20             = _mm256_sub_ps(ix2,jx0);
 260             dy20             = _mm256_sub_ps(iy2,jy0);
 261             dz20             = _mm256_sub_ps(iz2,jz0);
 262             dx30             = _mm256_sub_ps(ix3,jx0);
 263             dy30             = _mm256_sub_ps(iy3,jy0);
 264             dz30             = _mm256_sub_ps(iz3,jz0);
 265
 266             /* Calculate squared distance and things based on it */
 267             rsq00            = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
 268             rsq10            = gmx_mm256_calc_rsq_ps(dx10,dy10,dz10);
 269             rsq20            = gmx_mm256_calc_rsq_ps(dx20,dy20,dz20);
 270             rsq30            = gmx_mm256_calc_rsq_ps(dx30,dy30,dz30);
 271
 272             rinv00           = gmx_mm256_invsqrt_ps(rsq00);
 273             rinv10           = gmx_mm256_invsqrt_ps(rsq10);
 274             rinv20           = gmx_mm256_invsqrt_ps(rsq20);
 275             rinv30           = gmx_mm256_invsqrt_ps(rsq30);
 276
 277             rinvsq00         = _mm256_mul_ps(rinv00,rinv00);
 278             rinvsq10         = _mm256_mul_ps(rinv10,rinv10);
 279             rinvsq20         = _mm256_mul_ps(rinv20,rinv20);
 280             rinvsq30         = _mm256_mul_ps(rinv30,rinv30);
 281
 282             /* Load parameters for j particles */
 283             jq0              = gmx_mm256_load_8real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
 284                                                                  charge+jnrC+0,charge+jnrD+0,
 285                                                                  charge+jnrE+0,charge+jnrF+0,
 286                                                                  charge+jnrG+0,charge+jnrH+0);
 287             vdwjidx0A        = 2*vdwtype[jnrA+0];
 288             vdwjidx0B        = 2*vdwtype[jnrB+0];
 289             vdwjidx0C        = 2*vdwtype[jnrC+0];
 290             vdwjidx0D        = 2*vdwtype[jnrD+0];
 291             vdwjidx0E        = 2*vdwtype[jnrE+0];
 292             vdwjidx0F        = 2*vdwtype[jnrF+0];
 293             vdwjidx0G        = 2*vdwtype[jnrG+0];
 294             vdwjidx0H        = 2*vdwtype[jnrH+0];
 295
 296             fjx0             = _mm256_setzero_ps();
 297             fjy0             = _mm256_setzero_ps();
 298             fjz0             = _mm256_setzero_ps();
 299
 300             /**************************
 301              * CALCULATE INTERACTIONS *
 302              **************************/
 303
 304             if (gmx_mm256_any_lt(rsq00,rcutoff2))
 305             {
 306
 307             r00              = _mm256_mul_ps(rsq00,rinv00);
 308
 309             /* Compute parameters for interactions between i and j atoms */
 310             gmx_mm256_load_8pair_swizzle_ps(vdwioffsetptr0+vdwjidx0A,
 311                                             vdwioffsetptr0+vdwjidx0B,
 312                                             vdwioffsetptr0+vdwjidx0C,
 313                                             vdwioffsetptr0+vdwjidx0D,
 314                                             vdwioffsetptr0+vdwjidx0E,
 315                                             vdwioffsetptr0+vdwjidx0F,
 316                                             vdwioffsetptr0+vdwjidx0G,
 317                                             vdwioffsetptr0+vdwjidx0H,
 318                                             &c6_00,&c12_00);
 319
 320             /* LENNARD-JONES DISPERSION/REPULSION */
 321
 322             rinvsix          = _mm256_mul_ps(_mm256_mul_ps(rinvsq00,rinvsq00),rinvsq00);
 323             vvdw6            = _mm256_mul_ps(c6_00,rinvsix);
 324             vvdw12           = _mm256_mul_ps(c12_00,_mm256_mul_ps(rinvsix,rinvsix));
 325             vvdw             = _mm256_sub_ps( _mm256_mul_ps(vvdw12,one_twelfth) , _mm256_mul_ps(vvdw6,one_sixth) );
 326             fvdw             = _mm256_mul_ps(_mm256_sub_ps(vvdw12,vvdw6),rinvsq00);
 327
 328             d                = _mm256_sub_ps(r00,rswitch);
 329             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 330             d2               = _mm256_mul_ps(d,d);
 331             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)))))));
 332
 333             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 334
 335             /* Evaluate switch function */
 336             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 337             fvdw             = _mm256_sub_ps( _mm256_mul_ps(fvdw,sw) , _mm256_mul_ps(rinv00,_mm256_mul_ps(vvdw,dsw)) );
 338             vvdw             = _mm256_mul_ps(vvdw,sw);
 339             cutoff_mask      = _mm256_cmp_ps(rsq00,rcutoff2,_CMP_LT_OQ);
 340
 341             /* Update potential sum for this i atom from the interaction with this j atom. */
 342             vvdw             = _mm256_and_ps(vvdw,cutoff_mask);
 343             vvdwsum          = _mm256_add_ps(vvdwsum,vvdw);
 344
 345             fscal            = fvdw;
 346
 347             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 348
 349             /* Calculate temporary vectorial force */
 350             tx               = _mm256_mul_ps(fscal,dx00);
 351             ty               = _mm256_mul_ps(fscal,dy00);
 352             tz               = _mm256_mul_ps(fscal,dz00);
 353
 354             /* Update vectorial force */
 355             fix0             = _mm256_add_ps(fix0,tx);
 356             fiy0             = _mm256_add_ps(fiy0,ty);
 357             fiz0             = _mm256_add_ps(fiz0,tz);
 358
 359             fjx0             = _mm256_add_ps(fjx0,tx);
 360             fjy0             = _mm256_add_ps(fjy0,ty);
 361             fjz0             = _mm256_add_ps(fjz0,tz);
 362
 363             }
 364
 365             /**************************
 366              * CALCULATE INTERACTIONS *
 367              **************************/
 368
 369             if (gmx_mm256_any_lt(rsq10,rcutoff2))
 370             {
 371
 372             r10              = _mm256_mul_ps(rsq10,rinv10);
 373
 374             /* Compute parameters for interactions between i and j atoms */
 375             qq10             = _mm256_mul_ps(iq1,jq0);
 376
 377             /* EWALD ELECTROSTATICS */
 378
 379             /* Analytical PME correction */
 380             zeta2            = _mm256_mul_ps(beta2,rsq10);
 381             rinv3            = _mm256_mul_ps(rinvsq10,rinv10);
 382             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
 383             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 384             felec            = _mm256_mul_ps(qq10,felec);
 385             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
 386             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 387             velec            = _mm256_sub_ps(rinv10,pmecorrV);
 388             velec            = _mm256_mul_ps(qq10,velec);
 389
 390             d                = _mm256_sub_ps(r10,rswitch);
 391             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 392             d2               = _mm256_mul_ps(d,d);
 393             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)))))));
 394
 395             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 396
 397             /* Evaluate switch function */
 398             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 399             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv10,_mm256_mul_ps(velec,dsw)) );
 400             velec            = _mm256_mul_ps(velec,sw);
 401             cutoff_mask      = _mm256_cmp_ps(rsq10,rcutoff2,_CMP_LT_OQ);
 402
 403             /* Update potential sum for this i atom from the interaction with this j atom. */
 404             velec            = _mm256_and_ps(velec,cutoff_mask);
 405             velecsum         = _mm256_add_ps(velecsum,velec);
 406
 407             fscal            = felec;
 408
 409             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 410
 411             /* Calculate temporary vectorial force */
 412             tx               = _mm256_mul_ps(fscal,dx10);
 413             ty               = _mm256_mul_ps(fscal,dy10);
 414             tz               = _mm256_mul_ps(fscal,dz10);
 415
 416             /* Update vectorial force */
 417             fix1             = _mm256_add_ps(fix1,tx);
 418             fiy1             = _mm256_add_ps(fiy1,ty);
 419             fiz1             = _mm256_add_ps(fiz1,tz);
 420
 421             fjx0             = _mm256_add_ps(fjx0,tx);
 422             fjy0             = _mm256_add_ps(fjy0,ty);
 423             fjz0             = _mm256_add_ps(fjz0,tz);
 424
 425             }
 426
 427             /**************************
 428              * CALCULATE INTERACTIONS *
 429              **************************/
 430
 431             if (gmx_mm256_any_lt(rsq20,rcutoff2))
 432             {
 433
 434             r20              = _mm256_mul_ps(rsq20,rinv20);
 435
 436             /* Compute parameters for interactions between i and j atoms */
 437             qq20             = _mm256_mul_ps(iq2,jq0);
 438
 439             /* EWALD ELECTROSTATICS */
 440
 441             /* Analytical PME correction */
 442             zeta2            = _mm256_mul_ps(beta2,rsq20);
 443             rinv3            = _mm256_mul_ps(rinvsq20,rinv20);
 444             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
 445             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 446             felec            = _mm256_mul_ps(qq20,felec);
 447             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
 448             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 449             velec            = _mm256_sub_ps(rinv20,pmecorrV);
 450             velec            = _mm256_mul_ps(qq20,velec);
 451
 452             d                = _mm256_sub_ps(r20,rswitch);
 453             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 454             d2               = _mm256_mul_ps(d,d);
 455             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)))))));
 456
 457             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 458
 459             /* Evaluate switch function */
 460             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 461             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv20,_mm256_mul_ps(velec,dsw)) );
 462             velec            = _mm256_mul_ps(velec,sw);
 463             cutoff_mask      = _mm256_cmp_ps(rsq20,rcutoff2,_CMP_LT_OQ);
 464
 465             /* Update potential sum for this i atom from the interaction with this j atom. */
 466             velec            = _mm256_and_ps(velec,cutoff_mask);
 467             velecsum         = _mm256_add_ps(velecsum,velec);
 468
 469             fscal            = felec;
 470
 471             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 472
 473             /* Calculate temporary vectorial force */
 474             tx               = _mm256_mul_ps(fscal,dx20);
 475             ty               = _mm256_mul_ps(fscal,dy20);
 476             tz               = _mm256_mul_ps(fscal,dz20);
 477
 478             /* Update vectorial force */
 479             fix2             = _mm256_add_ps(fix2,tx);
 480             fiy2             = _mm256_add_ps(fiy2,ty);
 481             fiz2             = _mm256_add_ps(fiz2,tz);
 482
 483             fjx0             = _mm256_add_ps(fjx0,tx);
 484             fjy0             = _mm256_add_ps(fjy0,ty);
 485             fjz0             = _mm256_add_ps(fjz0,tz);
 486
 487             }
 488
 489             /**************************
 490              * CALCULATE INTERACTIONS *
 491              **************************/
 492
 493             if (gmx_mm256_any_lt(rsq30,rcutoff2))
 494             {
 495
 496             r30              = _mm256_mul_ps(rsq30,rinv30);
 497
 498             /* Compute parameters for interactions between i and j atoms */
 499             qq30             = _mm256_mul_ps(iq3,jq0);
 500
 501             /* EWALD ELECTROSTATICS */
 502
 503             /* Analytical PME correction */
 504             zeta2            = _mm256_mul_ps(beta2,rsq30);
 505             rinv3            = _mm256_mul_ps(rinvsq30,rinv30);
 506             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
 507             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 508             felec            = _mm256_mul_ps(qq30,felec);
 509             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
 510             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 511             velec            = _mm256_sub_ps(rinv30,pmecorrV);
 512             velec            = _mm256_mul_ps(qq30,velec);
 513
 514             d                = _mm256_sub_ps(r30,rswitch);
 515             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 516             d2               = _mm256_mul_ps(d,d);
 517             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)))))));
 518
 519             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 520
 521             /* Evaluate switch function */
 522             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 523             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv30,_mm256_mul_ps(velec,dsw)) );
 524             velec            = _mm256_mul_ps(velec,sw);
 525             cutoff_mask      = _mm256_cmp_ps(rsq30,rcutoff2,_CMP_LT_OQ);
 526
 527             /* Update potential sum for this i atom from the interaction with this j atom. */
 528             velec            = _mm256_and_ps(velec,cutoff_mask);
 529             velecsum         = _mm256_add_ps(velecsum,velec);
 530
 531             fscal            = felec;
 532
 533             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 534
 535             /* Calculate temporary vectorial force */
 536             tx               = _mm256_mul_ps(fscal,dx30);
 537             ty               = _mm256_mul_ps(fscal,dy30);
 538             tz               = _mm256_mul_ps(fscal,dz30);
 539
 540             /* Update vectorial force */
 541             fix3             = _mm256_add_ps(fix3,tx);
 542             fiy3             = _mm256_add_ps(fiy3,ty);
 543             fiz3             = _mm256_add_ps(fiz3,tz);
 544
 545             fjx0             = _mm256_add_ps(fjx0,tx);
 546             fjy0             = _mm256_add_ps(fjy0,ty);
 547             fjz0             = _mm256_add_ps(fjz0,tz);
 548
 549             }
 550
 551             fjptrA             = f+j_coord_offsetA;
 552             fjptrB             = f+j_coord_offsetB;
 553             fjptrC             = f+j_coord_offsetC;
 554             fjptrD             = f+j_coord_offsetD;
 555             fjptrE             = f+j_coord_offsetE;
 556             fjptrF             = f+j_coord_offsetF;
 557             fjptrG             = f+j_coord_offsetG;
 558             fjptrH             = f+j_coord_offsetH;
 559
 560             gmx_mm256_decrement_1rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,fjx0,fjy0,fjz0);
 561
 562             /* Inner loop uses 386 flops */
 563         }
 564
 565         if(jidx<j_index_end)
 566         {
 567
 568             /* Get j neighbor index, and coordinate index */
 569             jnrlistA         = jjnr[jidx];
 570             jnrlistB         = jjnr[jidx+1];
 571             jnrlistC         = jjnr[jidx+2];
 572             jnrlistD         = jjnr[jidx+3];
 573             jnrlistE         = jjnr[jidx+4];
 574             jnrlistF         = jjnr[jidx+5];
 575             jnrlistG         = jjnr[jidx+6];
 576             jnrlistH         = jjnr[jidx+7];
 577             /* Sign of each element will be negative for non-real atoms.
 578              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
 579              * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
 580              */
 581             dummy_mask = gmx_mm256_set_m128(gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx+4)),_mm_setzero_si128())),
 582                                             gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128())));
 583
 584             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
 585             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
 586             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
 587             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
 588             jnrE       = (jnrlistE>=0) ? jnrlistE : 0;
 589             jnrF       = (jnrlistF>=0) ? jnrlistF : 0;
 590             jnrG       = (jnrlistG>=0) ? jnrlistG : 0;
 591             jnrH       = (jnrlistH>=0) ? jnrlistH : 0;
 592             j_coord_offsetA  = DIM*jnrA;
 593             j_coord_offsetB  = DIM*jnrB;
 594             j_coord_offsetC  = DIM*jnrC;
 595             j_coord_offsetD  = DIM*jnrD;
 596             j_coord_offsetE  = DIM*jnrE;
 597             j_coord_offsetF  = DIM*jnrF;
 598             j_coord_offsetG  = DIM*jnrG;
 599             j_coord_offsetH  = DIM*jnrH;
 600
 601             /* load j atom coordinates */
 602             gmx_mm256_load_1rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 603                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 604                                                  x+j_coord_offsetE,x+j_coord_offsetF,
 605                                                  x+j_coord_offsetG,x+j_coord_offsetH,
 606                                                  &jx0,&jy0,&jz0);
 607
 608             /* Calculate displacement vector */
 609             dx00             = _mm256_sub_ps(ix0,jx0);
 610             dy00             = _mm256_sub_ps(iy0,jy0);
 611             dz00             = _mm256_sub_ps(iz0,jz0);
 612             dx10             = _mm256_sub_ps(ix1,jx0);
 613             dy10             = _mm256_sub_ps(iy1,jy0);
 614             dz10             = _mm256_sub_ps(iz1,jz0);
 615             dx20             = _mm256_sub_ps(ix2,jx0);
 616             dy20             = _mm256_sub_ps(iy2,jy0);
 617             dz20             = _mm256_sub_ps(iz2,jz0);
 618             dx30             = _mm256_sub_ps(ix3,jx0);
 619             dy30             = _mm256_sub_ps(iy3,jy0);
 620             dz30             = _mm256_sub_ps(iz3,jz0);
 621
 622             /* Calculate squared distance and things based on it */
 623             rsq00            = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
 624             rsq10            = gmx_mm256_calc_rsq_ps(dx10,dy10,dz10);
 625             rsq20            = gmx_mm256_calc_rsq_ps(dx20,dy20,dz20);
 626             rsq30            = gmx_mm256_calc_rsq_ps(dx30,dy30,dz30);
 627
 628             rinv00           = gmx_mm256_invsqrt_ps(rsq00);
 629             rinv10           = gmx_mm256_invsqrt_ps(rsq10);
 630             rinv20           = gmx_mm256_invsqrt_ps(rsq20);
 631             rinv30           = gmx_mm256_invsqrt_ps(rsq30);
 632
 633             rinvsq00         = _mm256_mul_ps(rinv00,rinv00);
 634             rinvsq10         = _mm256_mul_ps(rinv10,rinv10);
 635             rinvsq20         = _mm256_mul_ps(rinv20,rinv20);
 636             rinvsq30         = _mm256_mul_ps(rinv30,rinv30);
 637
 638             /* Load parameters for j particles */
 639             jq0              = gmx_mm256_load_8real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
 640                                                                  charge+jnrC+0,charge+jnrD+0,
 641                                                                  charge+jnrE+0,charge+jnrF+0,
 642                                                                  charge+jnrG+0,charge+jnrH+0);
 643             vdwjidx0A        = 2*vdwtype[jnrA+0];
 644             vdwjidx0B        = 2*vdwtype[jnrB+0];
 645             vdwjidx0C        = 2*vdwtype[jnrC+0];
 646             vdwjidx0D        = 2*vdwtype[jnrD+0];
 647             vdwjidx0E        = 2*vdwtype[jnrE+0];
 648             vdwjidx0F        = 2*vdwtype[jnrF+0];
 649             vdwjidx0G        = 2*vdwtype[jnrG+0];
 650             vdwjidx0H        = 2*vdwtype[jnrH+0];
 651
 652             fjx0             = _mm256_setzero_ps();
 653             fjy0             = _mm256_setzero_ps();
 654             fjz0             = _mm256_setzero_ps();
 655
 656             /**************************
 657              * CALCULATE INTERACTIONS *
 658              **************************/
 659
 660             if (gmx_mm256_any_lt(rsq00,rcutoff2))
 661             {
 662
 663             r00              = _mm256_mul_ps(rsq00,rinv00);
 664             r00              = _mm256_andnot_ps(dummy_mask,r00);
 665
 666             /* Compute parameters for interactions between i and j atoms */
 667             gmx_mm256_load_8pair_swizzle_ps(vdwioffsetptr0+vdwjidx0A,
 668                                             vdwioffsetptr0+vdwjidx0B,
 669                                             vdwioffsetptr0+vdwjidx0C,
 670                                             vdwioffsetptr0+vdwjidx0D,
 671                                             vdwioffsetptr0+vdwjidx0E,
 672                                             vdwioffsetptr0+vdwjidx0F,
 673                                             vdwioffsetptr0+vdwjidx0G,
 674                                             vdwioffsetptr0+vdwjidx0H,
 675                                             &c6_00,&c12_00);
 676
 677             /* LENNARD-JONES DISPERSION/REPULSION */
 678
 679             rinvsix          = _mm256_mul_ps(_mm256_mul_ps(rinvsq00,rinvsq00),rinvsq00);
 680             vvdw6            = _mm256_mul_ps(c6_00,rinvsix);
 681             vvdw12           = _mm256_mul_ps(c12_00,_mm256_mul_ps(rinvsix,rinvsix));
 682             vvdw             = _mm256_sub_ps( _mm256_mul_ps(vvdw12,one_twelfth) , _mm256_mul_ps(vvdw6,one_sixth) );
 683             fvdw             = _mm256_mul_ps(_mm256_sub_ps(vvdw12,vvdw6),rinvsq00);
 684
 685             d                = _mm256_sub_ps(r00,rswitch);
 686             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 687             d2               = _mm256_mul_ps(d,d);
 688             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)))))));
 689
 690             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 691
 692             /* Evaluate switch function */
 693             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 694             fvdw             = _mm256_sub_ps( _mm256_mul_ps(fvdw,sw) , _mm256_mul_ps(rinv00,_mm256_mul_ps(vvdw,dsw)) );
 695             vvdw             = _mm256_mul_ps(vvdw,sw);
 696             cutoff_mask      = _mm256_cmp_ps(rsq00,rcutoff2,_CMP_LT_OQ);
 697
 698             /* Update potential sum for this i atom from the interaction with this j atom. */
 699             vvdw             = _mm256_and_ps(vvdw,cutoff_mask);
 700             vvdw             = _mm256_andnot_ps(dummy_mask,vvdw);
 701             vvdwsum          = _mm256_add_ps(vvdwsum,vvdw);
 702
 703             fscal            = fvdw;
 704
 705             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 706
 707             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
 708
 709             /* Calculate temporary vectorial force */
 710             tx               = _mm256_mul_ps(fscal,dx00);
 711             ty               = _mm256_mul_ps(fscal,dy00);
 712             tz               = _mm256_mul_ps(fscal,dz00);
 713
 714             /* Update vectorial force */
 715             fix0             = _mm256_add_ps(fix0,tx);
 716             fiy0             = _mm256_add_ps(fiy0,ty);
 717             fiz0             = _mm256_add_ps(fiz0,tz);
 718
 719             fjx0             = _mm256_add_ps(fjx0,tx);
 720             fjy0             = _mm256_add_ps(fjy0,ty);
 721             fjz0             = _mm256_add_ps(fjz0,tz);
 722
 723             }
 724
 725             /**************************
 726              * CALCULATE INTERACTIONS *
 727              **************************/
 728
 729             if (gmx_mm256_any_lt(rsq10,rcutoff2))
 730             {
 731
 732             r10              = _mm256_mul_ps(rsq10,rinv10);
 733             r10              = _mm256_andnot_ps(dummy_mask,r10);
 734
 735             /* Compute parameters for interactions between i and j atoms */
 736             qq10             = _mm256_mul_ps(iq1,jq0);
 737
 738             /* EWALD ELECTROSTATICS */
 739
 740             /* Analytical PME correction */
 741             zeta2            = _mm256_mul_ps(beta2,rsq10);
 742             rinv3            = _mm256_mul_ps(rinvsq10,rinv10);
 743             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
 744             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 745             felec            = _mm256_mul_ps(qq10,felec);
 746             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
 747             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 748             velec            = _mm256_sub_ps(rinv10,pmecorrV);
 749             velec            = _mm256_mul_ps(qq10,velec);
 750
 751             d                = _mm256_sub_ps(r10,rswitch);
 752             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 753             d2               = _mm256_mul_ps(d,d);
 754             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)))))));
 755
 756             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 757
 758             /* Evaluate switch function */
 759             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 760             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv10,_mm256_mul_ps(velec,dsw)) );
 761             velec            = _mm256_mul_ps(velec,sw);
 762             cutoff_mask      = _mm256_cmp_ps(rsq10,rcutoff2,_CMP_LT_OQ);
 763
 764             /* Update potential sum for this i atom from the interaction with this j atom. */
 765             velec            = _mm256_and_ps(velec,cutoff_mask);
 766             velec            = _mm256_andnot_ps(dummy_mask,velec);
 767             velecsum         = _mm256_add_ps(velecsum,velec);
 768
 769             fscal            = felec;
 770
 771             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 772
 773             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
 774
 775             /* Calculate temporary vectorial force */
 776             tx               = _mm256_mul_ps(fscal,dx10);
 777             ty               = _mm256_mul_ps(fscal,dy10);
 778             tz               = _mm256_mul_ps(fscal,dz10);
 779
 780             /* Update vectorial force */
 781             fix1             = _mm256_add_ps(fix1,tx);
 782             fiy1             = _mm256_add_ps(fiy1,ty);
 783             fiz1             = _mm256_add_ps(fiz1,tz);
 784
 785             fjx0             = _mm256_add_ps(fjx0,tx);
 786             fjy0             = _mm256_add_ps(fjy0,ty);
 787             fjz0             = _mm256_add_ps(fjz0,tz);
 788
 789             }
 790
 791             /**************************
 792              * CALCULATE INTERACTIONS *
 793              **************************/
 794
 795             if (gmx_mm256_any_lt(rsq20,rcutoff2))
 796             {
 797
 798             r20              = _mm256_mul_ps(rsq20,rinv20);
 799             r20              = _mm256_andnot_ps(dummy_mask,r20);
 800
 801             /* Compute parameters for interactions between i and j atoms */
 802             qq20             = _mm256_mul_ps(iq2,jq0);
 803
 804             /* EWALD ELECTROSTATICS */
 805
 806             /* Analytical PME correction */
 807             zeta2            = _mm256_mul_ps(beta2,rsq20);
 808             rinv3            = _mm256_mul_ps(rinvsq20,rinv20);
 809             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
 810             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 811             felec            = _mm256_mul_ps(qq20,felec);
 812             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
 813             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 814             velec            = _mm256_sub_ps(rinv20,pmecorrV);
 815             velec            = _mm256_mul_ps(qq20,velec);
 816
 817             d                = _mm256_sub_ps(r20,rswitch);
 818             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 819             d2               = _mm256_mul_ps(d,d);
 820             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)))))));
 821
 822             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 823
 824             /* Evaluate switch function */
 825             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 826             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv20,_mm256_mul_ps(velec,dsw)) );
 827             velec            = _mm256_mul_ps(velec,sw);
 828             cutoff_mask      = _mm256_cmp_ps(rsq20,rcutoff2,_CMP_LT_OQ);
 829
 830             /* Update potential sum for this i atom from the interaction with this j atom. */
 831             velec            = _mm256_and_ps(velec,cutoff_mask);
 832             velec            = _mm256_andnot_ps(dummy_mask,velec);
 833             velecsum         = _mm256_add_ps(velecsum,velec);
 834
 835             fscal            = felec;
 836
 837             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 838
 839             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
 840
 841             /* Calculate temporary vectorial force */
 842             tx               = _mm256_mul_ps(fscal,dx20);
 843             ty               = _mm256_mul_ps(fscal,dy20);
 844             tz               = _mm256_mul_ps(fscal,dz20);
 845
 846             /* Update vectorial force */
 847             fix2             = _mm256_add_ps(fix2,tx);
 848             fiy2             = _mm256_add_ps(fiy2,ty);
 849             fiz2             = _mm256_add_ps(fiz2,tz);
 850
 851             fjx0             = _mm256_add_ps(fjx0,tx);
 852             fjy0             = _mm256_add_ps(fjy0,ty);
 853             fjz0             = _mm256_add_ps(fjz0,tz);
 854
 855             }
 856
 857             /**************************
 858              * CALCULATE INTERACTIONS *
 859              **************************/
 860
 861             if (gmx_mm256_any_lt(rsq30,rcutoff2))
 862             {
 863
 864             r30              = _mm256_mul_ps(rsq30,rinv30);
 865             r30              = _mm256_andnot_ps(dummy_mask,r30);
 866
 867             /* Compute parameters for interactions between i and j atoms */
 868             qq30             = _mm256_mul_ps(iq3,jq0);
 869
 870             /* EWALD ELECTROSTATICS */
 871
 872             /* Analytical PME correction */
 873             zeta2            = _mm256_mul_ps(beta2,rsq30);
 874             rinv3            = _mm256_mul_ps(rinvsq30,rinv30);
 875             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
 876             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 877             felec            = _mm256_mul_ps(qq30,felec);
 878             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
 879             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 880             velec            = _mm256_sub_ps(rinv30,pmecorrV);
 881             velec            = _mm256_mul_ps(qq30,velec);
 882
 883             d                = _mm256_sub_ps(r30,rswitch);
 884             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 885             d2               = _mm256_mul_ps(d,d);
 886             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)))))));
 887
 888             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 889
 890             /* Evaluate switch function */
 891             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 892             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv30,_mm256_mul_ps(velec,dsw)) );
 893             velec            = _mm256_mul_ps(velec,sw);
 894             cutoff_mask      = _mm256_cmp_ps(rsq30,rcutoff2,_CMP_LT_OQ);
 895
 896             /* Update potential sum for this i atom from the interaction with this j atom. */
 897             velec            = _mm256_and_ps(velec,cutoff_mask);
 898             velec            = _mm256_andnot_ps(dummy_mask,velec);
 899             velecsum         = _mm256_add_ps(velecsum,velec);
 900
 901             fscal            = felec;
 902
 903             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 904
 905             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
 906
 907             /* Calculate temporary vectorial force */
 908             tx               = _mm256_mul_ps(fscal,dx30);
 909             ty               = _mm256_mul_ps(fscal,dy30);
 910             tz               = _mm256_mul_ps(fscal,dz30);
 911
 912             /* Update vectorial force */
 913             fix3             = _mm256_add_ps(fix3,tx);
 914             fiy3             = _mm256_add_ps(fiy3,ty);
 915             fiz3             = _mm256_add_ps(fiz3,tz);
 916
 917             fjx0             = _mm256_add_ps(fjx0,tx);
 918             fjy0             = _mm256_add_ps(fjy0,ty);
 919             fjz0             = _mm256_add_ps(fjz0,tz);
 920
 921             }
 922
 923             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
 924             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
 925             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
 926             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
 927             fjptrE             = (jnrlistE>=0) ? f+j_coord_offsetE : scratch;
 928             fjptrF             = (jnrlistF>=0) ? f+j_coord_offsetF : scratch;
 929             fjptrG             = (jnrlistG>=0) ? f+j_coord_offsetG : scratch;
 930             fjptrH             = (jnrlistH>=0) ? f+j_coord_offsetH : scratch;
 931
 932             gmx_mm256_decrement_1rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,fjx0,fjy0,fjz0);
 933
 934             /* Inner loop uses 390 flops */
 935         }
 936
 937         /* End of innermost loop */
 938
 939         gmx_mm256_update_iforce_4atom_swizzle_ps(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
 940                                                  f+i_coord_offset,fshift+i_shift_offset);
 941
 942         ggid                        = gid[iidx];
 943         /* Update potential energies */
 944         gmx_mm256_update_1pot_ps(velecsum,kernel_data->energygrp_elec+ggid);
 945         gmx_mm256_update_1pot_ps(vvdwsum,kernel_data->energygrp_vdw+ggid);
 946
 947         /* Increment number of inner iterations */
 948         inneriter                  += j_index_end - j_index_start;
 949
 950         /* Outer loop uses 26 flops */
 951     }
 952
 953     /* Increment number of outer iterations */
 954     outeriter        += nri;
 955
 956     /* Update outer/inner flops */
 957
 958     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_VF,outeriter*26 + inneriter*390);
 959 }
 960 /*
 961  * Gromacs nonbonded kernel:   nb_kernel_ElecEwSw_VdwLJSw_GeomW4P1_F_avx_256_single
 962  * Electrostatics interaction: Ewald
 963  * VdW interaction:            LennardJones
 964  * Geometry:                   Water4-Particle
 965  * Calculate force/pot:        Force
 966  */
 967 void
 968 nb_kernel_ElecEwSw_VdwLJSw_GeomW4P1_F_avx_256_single
 969                     (t_nblist                    * gmx_restrict       nlist,
 970                      rvec                        * gmx_restrict          xx,
 971                      rvec                        * gmx_restrict          ff,
 972                      t_forcerec                  * gmx_restrict          fr,
 973                      t_mdatoms                   * gmx_restrict     mdatoms,
 974                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 975                      t_nrnb                      * gmx_restrict        nrnb)
 976 {
 977     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
 978      * just 0 for non-waters.
 979      * Suffixes A,B,C,D,E,F,G,H refer to j loop unrolling done with AVX, e.g. for the eight different
 980      * jnr indices corresponding to data put in the four positions in the SIMD register.
 981      */
 982     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
 983     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
 984     int              jnrA,jnrB,jnrC,jnrD;
 985     int              jnrE,jnrF,jnrG,jnrH;
 986     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
 987     int              jnrlistE,jnrlistF,jnrlistG,jnrlistH;
 988     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
 989     int              j_coord_offsetE,j_coord_offsetF,j_coord_offsetG,j_coord_offsetH;
 990     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 991     real             rcutoff_scalar;
 992     real             *shiftvec,*fshift,*x,*f;
 993     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD,*fjptrE,*fjptrF,*fjptrG,*fjptrH;
 994     real             scratch[4*DIM];
 995     __m256           tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
 996     real *           vdwioffsetptr0;
 997     __m256           ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 998     real *           vdwioffsetptr1;
 999     __m256           ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
1000     real *           vdwioffsetptr2;
1001     __m256           ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
1002     real *           vdwioffsetptr3;
1003     __m256           ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
1004     int              vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D,vdwjidx0E,vdwjidx0F,vdwjidx0G,vdwjidx0H;
1005     __m256           jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
1006     __m256           dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
1007     __m256           dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
1008     __m256           dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
1009     __m256           dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30;
1010     __m256           velec,felec,velecsum,facel,crf,krf,krf2;
1011     real             *charge;
1012     int              nvdwtype;
1013     __m256           rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
1014     int              *vdwtype;
1015     real             *vdwparam;
1016     __m256           one_sixth   = _mm256_set1_ps(1.0/6.0);
1017     __m256           one_twelfth = _mm256_set1_ps(1.0/12.0);
1018     __m256i          ewitab;
1019     __m128i          ewitab_lo,ewitab_hi;
1020     __m256           ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
1021     __m256           beta,beta2,beta3,zeta2,pmecorrF,pmecorrV,rinv3;
1022     real             *ewtab;
1023     __m256           rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
1024     real             rswitch_scalar,d_scalar;
1025     __m256           dummy_mask,cutoff_mask;
1026     __m256           signbit = _mm256_castsi256_ps( _mm256_set1_epi32(0x80000000) );
1027     __m256           one     = _mm256_set1_ps(1.0);
1028     __m256           two     = _mm256_set1_ps(2.0);
1029     x                = xx[0];
1030     f                = ff[0];
1031
1032     nri              = nlist->nri;
1033     iinr             = nlist->iinr;
1034     jindex           = nlist->jindex;
1035     jjnr             = nlist->jjnr;
1036     shiftidx         = nlist->shift;
1037     gid              = nlist->gid;
1038     shiftvec         = fr->shift_vec[0];
1039     fshift           = fr->fshift[0];
1040     facel            = _mm256_set1_ps(fr->epsfac);
1041     charge           = mdatoms->chargeA;
1042     nvdwtype         = fr->ntype;
1043     vdwparam         = fr->nbfp;
1044     vdwtype          = mdatoms->typeA;
1045
1046     sh_ewald         = _mm256_set1_ps(fr->ic->sh_ewald);
1047     beta             = _mm256_set1_ps(fr->ic->ewaldcoeff_q);
1048     beta2            = _mm256_mul_ps(beta,beta);
1049     beta3            = _mm256_mul_ps(beta,beta2);
1050
1051     ewtab            = fr->ic->tabq_coul_FDV0;
1052     ewtabscale       = _mm256_set1_ps(fr->ic->tabq_scale);
1053     ewtabhalfspace   = _mm256_set1_ps(0.5/fr->ic->tabq_scale);
1054
1055     /* Setup water-specific parameters */
1056     inr              = nlist->iinr[0];
1057     iq1              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+1]));
1058     iq2              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+2]));
1059     iq3              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+3]));
1060     vdwioffsetptr0   = vdwparam+2*nvdwtype*vdwtype[inr+0];
1061
1062     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
1063     rcutoff_scalar   = fr->rcoulomb;
1064     rcutoff          = _mm256_set1_ps(rcutoff_scalar);
1065     rcutoff2         = _mm256_mul_ps(rcutoff,rcutoff);
1066
1067     rswitch_scalar   = fr->rcoulomb_switch;
1068     rswitch          = _mm256_set1_ps(rswitch_scalar);
1069     /* Setup switch parameters */
1070     d_scalar         = rcutoff_scalar-rswitch_scalar;
1071     d                = _mm256_set1_ps(d_scalar);
1072     swV3             = _mm256_set1_ps(-10.0/(d_scalar*d_scalar*d_scalar));
1073     swV4             = _mm256_set1_ps( 15.0/(d_scalar*d_scalar*d_scalar*d_scalar));
1074     swV5             = _mm256_set1_ps( -6.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
1075     swF2             = _mm256_set1_ps(-30.0/(d_scalar*d_scalar*d_scalar));
1076     swF3             = _mm256_set1_ps( 60.0/(d_scalar*d_scalar*d_scalar*d_scalar));
1077     swF4             = _mm256_set1_ps(-30.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
1078
1079     /* Avoid stupid compiler warnings */
1080     jnrA = jnrB = jnrC = jnrD = jnrE = jnrF = jnrG = jnrH = 0;
1081     j_coord_offsetA = 0;
1082     j_coord_offsetB = 0;
1083     j_coord_offsetC = 0;
1084     j_coord_offsetD = 0;
1085     j_coord_offsetE = 0;
1086     j_coord_offsetF = 0;
1087     j_coord_offsetG = 0;
1088     j_coord_offsetH = 0;
1089
1090     outeriter        = 0;
1091     inneriter        = 0;
1092
1093     for(iidx=0;iidx<4*DIM;iidx++)
1094     {
1095         scratch[iidx] = 0.0;
1096     }
1097
1098     /* Start outer loop over neighborlists */
1099     for(iidx=0; iidx<nri; iidx++)
1100     {
1101         /* Load shift vector for this list */
1102         i_shift_offset   = DIM*shiftidx[iidx];
1103
1104         /* Load limits for loop over neighbors */
1105         j_index_start    = jindex[iidx];
1106         j_index_end      = jindex[iidx+1];
1107
1108         /* Get outer coordinate index */
1109         inr              = iinr[iidx];
1110         i_coord_offset   = DIM*inr;
1111
1112         /* Load i particle coords and add shift vector */
1113         gmx_mm256_load_shift_and_4rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,
1114                                                     &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
1115
1116         fix0             = _mm256_setzero_ps();
1117         fiy0             = _mm256_setzero_ps();
1118         fiz0             = _mm256_setzero_ps();
1119         fix1             = _mm256_setzero_ps();
1120         fiy1             = _mm256_setzero_ps();
1121         fiz1             = _mm256_setzero_ps();
1122         fix2             = _mm256_setzero_ps();
1123         fiy2             = _mm256_setzero_ps();
1124         fiz2             = _mm256_setzero_ps();
1125         fix3             = _mm256_setzero_ps();
1126         fiy3             = _mm256_setzero_ps();
1127         fiz3             = _mm256_setzero_ps();
1128
1129         /* Start inner kernel loop */
1130         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+7]>=0; jidx+=8)
1131         {
1132
1133             /* Get j neighbor index, and coordinate index */
1134             jnrA             = jjnr[jidx];
1135             jnrB             = jjnr[jidx+1];
1136             jnrC             = jjnr[jidx+2];
1137             jnrD             = jjnr[jidx+3];
1138             jnrE             = jjnr[jidx+4];
1139             jnrF             = jjnr[jidx+5];
1140             jnrG             = jjnr[jidx+6];
1141             jnrH             = jjnr[jidx+7];
1142             j_coord_offsetA  = DIM*jnrA;
1143             j_coord_offsetB  = DIM*jnrB;
1144             j_coord_offsetC  = DIM*jnrC;
1145             j_coord_offsetD  = DIM*jnrD;
1146             j_coord_offsetE  = DIM*jnrE;
1147             j_coord_offsetF  = DIM*jnrF;
1148             j_coord_offsetG  = DIM*jnrG;
1149             j_coord_offsetH  = DIM*jnrH;
1150
1151             /* load j atom coordinates */
1152             gmx_mm256_load_1rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
1153                                                  x+j_coord_offsetC,x+j_coord_offsetD,
1154                                                  x+j_coord_offsetE,x+j_coord_offsetF,
1155                                                  x+j_coord_offsetG,x+j_coord_offsetH,
1156                                                  &jx0,&jy0,&jz0);
1157
1158             /* Calculate displacement vector */
1159             dx00             = _mm256_sub_ps(ix0,jx0);
1160             dy00             = _mm256_sub_ps(iy0,jy0);
1161             dz00             = _mm256_sub_ps(iz0,jz0);
1162             dx10             = _mm256_sub_ps(ix1,jx0);
1163             dy10             = _mm256_sub_ps(iy1,jy0);
1164             dz10             = _mm256_sub_ps(iz1,jz0);
1165             dx20             = _mm256_sub_ps(ix2,jx0);
1166             dy20             = _mm256_sub_ps(iy2,jy0);
1167             dz20             = _mm256_sub_ps(iz2,jz0);
1168             dx30             = _mm256_sub_ps(ix3,jx0);
1169             dy30             = _mm256_sub_ps(iy3,jy0);
1170             dz30             = _mm256_sub_ps(iz3,jz0);
1171
1172             /* Calculate squared distance and things based on it */
1173             rsq00            = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
1174             rsq10            = gmx_mm256_calc_rsq_ps(dx10,dy10,dz10);
1175             rsq20            = gmx_mm256_calc_rsq_ps(dx20,dy20,dz20);
1176             rsq30            = gmx_mm256_calc_rsq_ps(dx30,dy30,dz30);
1177
1178             rinv00           = gmx_mm256_invsqrt_ps(rsq00);
1179             rinv10           = gmx_mm256_invsqrt_ps(rsq10);
1180             rinv20           = gmx_mm256_invsqrt_ps(rsq20);
1181             rinv30           = gmx_mm256_invsqrt_ps(rsq30);
1182
1183             rinvsq00         = _mm256_mul_ps(rinv00,rinv00);
1184             rinvsq10         = _mm256_mul_ps(rinv10,rinv10);
1185             rinvsq20         = _mm256_mul_ps(rinv20,rinv20);
1186             rinvsq30         = _mm256_mul_ps(rinv30,rinv30);
1187
1188             /* Load parameters for j particles */
1189             jq0              = gmx_mm256_load_8real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
1190                                                                  charge+jnrC+0,charge+jnrD+0,
1191                                                                  charge+jnrE+0,charge+jnrF+0,
1192                                                                  charge+jnrG+0,charge+jnrH+0);
1193             vdwjidx0A        = 2*vdwtype[jnrA+0];
1194             vdwjidx0B        = 2*vdwtype[jnrB+0];
1195             vdwjidx0C        = 2*vdwtype[jnrC+0];
1196             vdwjidx0D        = 2*vdwtype[jnrD+0];
1197             vdwjidx0E        = 2*vdwtype[jnrE+0];
1198             vdwjidx0F        = 2*vdwtype[jnrF+0];
1199             vdwjidx0G        = 2*vdwtype[jnrG+0];
1200             vdwjidx0H        = 2*vdwtype[jnrH+0];
1201
1202             fjx0             = _mm256_setzero_ps();
1203             fjy0             = _mm256_setzero_ps();
1204             fjz0             = _mm256_setzero_ps();
1205
1206             /**************************
1207              * CALCULATE INTERACTIONS *
1208              **************************/
1209
1210             if (gmx_mm256_any_lt(rsq00,rcutoff2))
1211             {
1212
1213             r00              = _mm256_mul_ps(rsq00,rinv00);
1214
1215             /* Compute parameters for interactions between i and j atoms */
1216             gmx_mm256_load_8pair_swizzle_ps(vdwioffsetptr0+vdwjidx0A,
1217                                             vdwioffsetptr0+vdwjidx0B,
1218                                             vdwioffsetptr0+vdwjidx0C,
1219                                             vdwioffsetptr0+vdwjidx0D,
1220                                             vdwioffsetptr0+vdwjidx0E,
1221                                             vdwioffsetptr0+vdwjidx0F,
1222                                             vdwioffsetptr0+vdwjidx0G,
1223                                             vdwioffsetptr0+vdwjidx0H,
1224                                             &c6_00,&c12_00);
1225
1226             /* LENNARD-JONES DISPERSION/REPULSION */
1227
1228             rinvsix          = _mm256_mul_ps(_mm256_mul_ps(rinvsq00,rinvsq00),rinvsq00);
1229             vvdw6            = _mm256_mul_ps(c6_00,rinvsix);
1230             vvdw12           = _mm256_mul_ps(c12_00,_mm256_mul_ps(rinvsix,rinvsix));
1231             vvdw             = _mm256_sub_ps( _mm256_mul_ps(vvdw12,one_twelfth) , _mm256_mul_ps(vvdw6,one_sixth) );
1232             fvdw             = _mm256_mul_ps(_mm256_sub_ps(vvdw12,vvdw6),rinvsq00);
1233
1234             d                = _mm256_sub_ps(r00,rswitch);
1235             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1236             d2               = _mm256_mul_ps(d,d);
1237             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)))))));
1238
1239             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1240
1241             /* Evaluate switch function */
1242             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1243             fvdw             = _mm256_sub_ps( _mm256_mul_ps(fvdw,sw) , _mm256_mul_ps(rinv00,_mm256_mul_ps(vvdw,dsw)) );
1244             cutoff_mask      = _mm256_cmp_ps(rsq00,rcutoff2,_CMP_LT_OQ);
1245
1246             fscal            = fvdw;
1247
1248             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1249
1250             /* Calculate temporary vectorial force */
1251             tx               = _mm256_mul_ps(fscal,dx00);
1252             ty               = _mm256_mul_ps(fscal,dy00);
1253             tz               = _mm256_mul_ps(fscal,dz00);
1254
1255             /* Update vectorial force */
1256             fix0             = _mm256_add_ps(fix0,tx);
1257             fiy0             = _mm256_add_ps(fiy0,ty);
1258             fiz0             = _mm256_add_ps(fiz0,tz);
1259
1260             fjx0             = _mm256_add_ps(fjx0,tx);
1261             fjy0             = _mm256_add_ps(fjy0,ty);
1262             fjz0             = _mm256_add_ps(fjz0,tz);
1263
1264             }
1265
1266             /**************************
1267              * CALCULATE INTERACTIONS *
1268              **************************/
1269
1270             if (gmx_mm256_any_lt(rsq10,rcutoff2))
1271             {
1272
1273             r10              = _mm256_mul_ps(rsq10,rinv10);
1274
1275             /* Compute parameters for interactions between i and j atoms */
1276             qq10             = _mm256_mul_ps(iq1,jq0);
1277
1278             /* EWALD ELECTROSTATICS */
1279
1280             /* Analytical PME correction */
1281             zeta2            = _mm256_mul_ps(beta2,rsq10);
1282             rinv3            = _mm256_mul_ps(rinvsq10,rinv10);
1283             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
1284             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1285             felec            = _mm256_mul_ps(qq10,felec);
1286             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
1287             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1288             velec            = _mm256_sub_ps(rinv10,pmecorrV);
1289             velec            = _mm256_mul_ps(qq10,velec);
1290
1291             d                = _mm256_sub_ps(r10,rswitch);
1292             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1293             d2               = _mm256_mul_ps(d,d);
1294             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)))))));
1295
1296             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1297
1298             /* Evaluate switch function */
1299             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1300             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv10,_mm256_mul_ps(velec,dsw)) );
1301             cutoff_mask      = _mm256_cmp_ps(rsq10,rcutoff2,_CMP_LT_OQ);
1302
1303             fscal            = felec;
1304
1305             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1306
1307             /* Calculate temporary vectorial force */
1308             tx               = _mm256_mul_ps(fscal,dx10);
1309             ty               = _mm256_mul_ps(fscal,dy10);
1310             tz               = _mm256_mul_ps(fscal,dz10);
1311
1312             /* Update vectorial force */
1313             fix1             = _mm256_add_ps(fix1,tx);
1314             fiy1             = _mm256_add_ps(fiy1,ty);
1315             fiz1             = _mm256_add_ps(fiz1,tz);
1316
1317             fjx0             = _mm256_add_ps(fjx0,tx);
1318             fjy0             = _mm256_add_ps(fjy0,ty);
1319             fjz0             = _mm256_add_ps(fjz0,tz);
1320
1321             }
1322
1323             /**************************
1324              * CALCULATE INTERACTIONS *
1325              **************************/
1326
1327             if (gmx_mm256_any_lt(rsq20,rcutoff2))
1328             {
1329
1330             r20              = _mm256_mul_ps(rsq20,rinv20);
1331
1332             /* Compute parameters for interactions between i and j atoms */
1333             qq20             = _mm256_mul_ps(iq2,jq0);
1334
1335             /* EWALD ELECTROSTATICS */
1336
1337             /* Analytical PME correction */
1338             zeta2            = _mm256_mul_ps(beta2,rsq20);
1339             rinv3            = _mm256_mul_ps(rinvsq20,rinv20);
1340             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
1341             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1342             felec            = _mm256_mul_ps(qq20,felec);
1343             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
1344             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1345             velec            = _mm256_sub_ps(rinv20,pmecorrV);
1346             velec            = _mm256_mul_ps(qq20,velec);
1347
1348             d                = _mm256_sub_ps(r20,rswitch);
1349             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1350             d2               = _mm256_mul_ps(d,d);
1351             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)))))));
1352
1353             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1354
1355             /* Evaluate switch function */
1356             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1357             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv20,_mm256_mul_ps(velec,dsw)) );
1358             cutoff_mask      = _mm256_cmp_ps(rsq20,rcutoff2,_CMP_LT_OQ);
1359
1360             fscal            = felec;
1361
1362             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1363
1364             /* Calculate temporary vectorial force */
1365             tx               = _mm256_mul_ps(fscal,dx20);
1366             ty               = _mm256_mul_ps(fscal,dy20);
1367             tz               = _mm256_mul_ps(fscal,dz20);
1368
1369             /* Update vectorial force */
1370             fix2             = _mm256_add_ps(fix2,tx);
1371             fiy2             = _mm256_add_ps(fiy2,ty);
1372             fiz2             = _mm256_add_ps(fiz2,tz);
1373
1374             fjx0             = _mm256_add_ps(fjx0,tx);
1375             fjy0             = _mm256_add_ps(fjy0,ty);
1376             fjz0             = _mm256_add_ps(fjz0,tz);
1377
1378             }
1379
1380             /**************************
1381              * CALCULATE INTERACTIONS *
1382              **************************/
1383
1384             if (gmx_mm256_any_lt(rsq30,rcutoff2))
1385             {
1386
1387             r30              = _mm256_mul_ps(rsq30,rinv30);
1388
1389             /* Compute parameters for interactions between i and j atoms */
1390             qq30             = _mm256_mul_ps(iq3,jq0);
1391
1392             /* EWALD ELECTROSTATICS */
1393
1394             /* Analytical PME correction */
1395             zeta2            = _mm256_mul_ps(beta2,rsq30);
1396             rinv3            = _mm256_mul_ps(rinvsq30,rinv30);
1397             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
1398             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1399             felec            = _mm256_mul_ps(qq30,felec);
1400             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
1401             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1402             velec            = _mm256_sub_ps(rinv30,pmecorrV);
1403             velec            = _mm256_mul_ps(qq30,velec);
1404
1405             d                = _mm256_sub_ps(r30,rswitch);
1406             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1407             d2               = _mm256_mul_ps(d,d);
1408             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)))))));
1409
1410             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1411
1412             /* Evaluate switch function */
1413             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1414             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv30,_mm256_mul_ps(velec,dsw)) );
1415             cutoff_mask      = _mm256_cmp_ps(rsq30,rcutoff2,_CMP_LT_OQ);
1416
1417             fscal            = felec;
1418
1419             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1420
1421             /* Calculate temporary vectorial force */
1422             tx               = _mm256_mul_ps(fscal,dx30);
1423             ty               = _mm256_mul_ps(fscal,dy30);
1424             tz               = _mm256_mul_ps(fscal,dz30);
1425
1426             /* Update vectorial force */
1427             fix3             = _mm256_add_ps(fix3,tx);
1428             fiy3             = _mm256_add_ps(fiy3,ty);
1429             fiz3             = _mm256_add_ps(fiz3,tz);
1430
1431             fjx0             = _mm256_add_ps(fjx0,tx);
1432             fjy0             = _mm256_add_ps(fjy0,ty);
1433             fjz0             = _mm256_add_ps(fjz0,tz);
1434
1435             }
1436
1437             fjptrA             = f+j_coord_offsetA;
1438             fjptrB             = f+j_coord_offsetB;
1439             fjptrC             = f+j_coord_offsetC;
1440             fjptrD             = f+j_coord_offsetD;
1441             fjptrE             = f+j_coord_offsetE;
1442             fjptrF             = f+j_coord_offsetF;
1443             fjptrG             = f+j_coord_offsetG;
1444             fjptrH             = f+j_coord_offsetH;
1445
1446             gmx_mm256_decrement_1rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,fjx0,fjy0,fjz0);
1447
1448             /* Inner loop uses 374 flops */
1449         }
1450
1451         if(jidx<j_index_end)
1452         {
1453
1454             /* Get j neighbor index, and coordinate index */
1455             jnrlistA         = jjnr[jidx];
1456             jnrlistB         = jjnr[jidx+1];
1457             jnrlistC         = jjnr[jidx+2];
1458             jnrlistD         = jjnr[jidx+3];
1459             jnrlistE         = jjnr[jidx+4];
1460             jnrlistF         = jjnr[jidx+5];
1461             jnrlistG         = jjnr[jidx+6];
1462             jnrlistH         = jjnr[jidx+7];
1463             /* Sign of each element will be negative for non-real atoms.
1464              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
1465              * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
1466              */
1467             dummy_mask = gmx_mm256_set_m128(gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx+4)),_mm_setzero_si128())),
1468                                             gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128())));
1469
1470             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
1471             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
1472             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
1473             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
1474             jnrE       = (jnrlistE>=0) ? jnrlistE : 0;
1475             jnrF       = (jnrlistF>=0) ? jnrlistF : 0;
1476             jnrG       = (jnrlistG>=0) ? jnrlistG : 0;
1477             jnrH       = (jnrlistH>=0) ? jnrlistH : 0;
1478             j_coord_offsetA  = DIM*jnrA;
1479             j_coord_offsetB  = DIM*jnrB;
1480             j_coord_offsetC  = DIM*jnrC;
1481             j_coord_offsetD  = DIM*jnrD;
1482             j_coord_offsetE  = DIM*jnrE;
1483             j_coord_offsetF  = DIM*jnrF;
1484             j_coord_offsetG  = DIM*jnrG;
1485             j_coord_offsetH  = DIM*jnrH;
1486
1487             /* load j atom coordinates */
1488             gmx_mm256_load_1rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
1489                                                  x+j_coord_offsetC,x+j_coord_offsetD,
1490                                                  x+j_coord_offsetE,x+j_coord_offsetF,
1491                                                  x+j_coord_offsetG,x+j_coord_offsetH,
1492                                                  &jx0,&jy0,&jz0);
1493
1494             /* Calculate displacement vector */
1495             dx00             = _mm256_sub_ps(ix0,jx0);
1496             dy00             = _mm256_sub_ps(iy0,jy0);
1497             dz00             = _mm256_sub_ps(iz0,jz0);
1498             dx10             = _mm256_sub_ps(ix1,jx0);
1499             dy10             = _mm256_sub_ps(iy1,jy0);
1500             dz10             = _mm256_sub_ps(iz1,jz0);
1501             dx20             = _mm256_sub_ps(ix2,jx0);
1502             dy20             = _mm256_sub_ps(iy2,jy0);
1503             dz20             = _mm256_sub_ps(iz2,jz0);
1504             dx30             = _mm256_sub_ps(ix3,jx0);
1505             dy30             = _mm256_sub_ps(iy3,jy0);
1506             dz30             = _mm256_sub_ps(iz3,jz0);
1507
1508             /* Calculate squared distance and things based on it */
1509             rsq00            = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
1510             rsq10            = gmx_mm256_calc_rsq_ps(dx10,dy10,dz10);
1511             rsq20            = gmx_mm256_calc_rsq_ps(dx20,dy20,dz20);
1512             rsq30            = gmx_mm256_calc_rsq_ps(dx30,dy30,dz30);
1513
1514             rinv00           = gmx_mm256_invsqrt_ps(rsq00);
1515             rinv10           = gmx_mm256_invsqrt_ps(rsq10);
1516             rinv20           = gmx_mm256_invsqrt_ps(rsq20);
1517             rinv30           = gmx_mm256_invsqrt_ps(rsq30);
1518
1519             rinvsq00         = _mm256_mul_ps(rinv00,rinv00);
1520             rinvsq10         = _mm256_mul_ps(rinv10,rinv10);
1521             rinvsq20         = _mm256_mul_ps(rinv20,rinv20);
1522             rinvsq30         = _mm256_mul_ps(rinv30,rinv30);
1523
1524             /* Load parameters for j particles */
1525             jq0              = gmx_mm256_load_8real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
1526                                                                  charge+jnrC+0,charge+jnrD+0,
1527                                                                  charge+jnrE+0,charge+jnrF+0,
1528                                                                  charge+jnrG+0,charge+jnrH+0);
1529             vdwjidx0A        = 2*vdwtype[jnrA+0];
1530             vdwjidx0B        = 2*vdwtype[jnrB+0];
1531             vdwjidx0C        = 2*vdwtype[jnrC+0];
1532             vdwjidx0D        = 2*vdwtype[jnrD+0];
1533             vdwjidx0E        = 2*vdwtype[jnrE+0];
1534             vdwjidx0F        = 2*vdwtype[jnrF+0];
1535             vdwjidx0G        = 2*vdwtype[jnrG+0];
1536             vdwjidx0H        = 2*vdwtype[jnrH+0];
1537
1538             fjx0             = _mm256_setzero_ps();
1539             fjy0             = _mm256_setzero_ps();
1540             fjz0             = _mm256_setzero_ps();
1541
1542             /**************************
1543              * CALCULATE INTERACTIONS *
1544              **************************/
1545
1546             if (gmx_mm256_any_lt(rsq00,rcutoff2))
1547             {
1548
1549             r00              = _mm256_mul_ps(rsq00,rinv00);
1550             r00              = _mm256_andnot_ps(dummy_mask,r00);
1551
1552             /* Compute parameters for interactions between i and j atoms */
1553             gmx_mm256_load_8pair_swizzle_ps(vdwioffsetptr0+vdwjidx0A,
1554                                             vdwioffsetptr0+vdwjidx0B,
1555                                             vdwioffsetptr0+vdwjidx0C,
1556                                             vdwioffsetptr0+vdwjidx0D,
1557                                             vdwioffsetptr0+vdwjidx0E,
1558                                             vdwioffsetptr0+vdwjidx0F,
1559                                             vdwioffsetptr0+vdwjidx0G,
1560                                             vdwioffsetptr0+vdwjidx0H,
1561                                             &c6_00,&c12_00);
1562
1563             /* LENNARD-JONES DISPERSION/REPULSION */
1564
1565             rinvsix          = _mm256_mul_ps(_mm256_mul_ps(rinvsq00,rinvsq00),rinvsq00);
1566             vvdw6            = _mm256_mul_ps(c6_00,rinvsix);
1567             vvdw12           = _mm256_mul_ps(c12_00,_mm256_mul_ps(rinvsix,rinvsix));
1568             vvdw             = _mm256_sub_ps( _mm256_mul_ps(vvdw12,one_twelfth) , _mm256_mul_ps(vvdw6,one_sixth) );
1569             fvdw             = _mm256_mul_ps(_mm256_sub_ps(vvdw12,vvdw6),rinvsq00);
1570
1571             d                = _mm256_sub_ps(r00,rswitch);
1572             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1573             d2               = _mm256_mul_ps(d,d);
1574             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)))))));
1575
1576             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1577
1578             /* Evaluate switch function */
1579             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1580             fvdw             = _mm256_sub_ps( _mm256_mul_ps(fvdw,sw) , _mm256_mul_ps(rinv00,_mm256_mul_ps(vvdw,dsw)) );
1581             cutoff_mask      = _mm256_cmp_ps(rsq00,rcutoff2,_CMP_LT_OQ);
1582
1583             fscal            = fvdw;
1584
1585             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1586
1587             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1588
1589             /* Calculate temporary vectorial force */
1590             tx               = _mm256_mul_ps(fscal,dx00);
1591             ty               = _mm256_mul_ps(fscal,dy00);
1592             tz               = _mm256_mul_ps(fscal,dz00);
1593
1594             /* Update vectorial force */
1595             fix0             = _mm256_add_ps(fix0,tx);
1596             fiy0             = _mm256_add_ps(fiy0,ty);
1597             fiz0             = _mm256_add_ps(fiz0,tz);
1598
1599             fjx0             = _mm256_add_ps(fjx0,tx);
1600             fjy0             = _mm256_add_ps(fjy0,ty);
1601             fjz0             = _mm256_add_ps(fjz0,tz);
1602
1603             }
1604
1605             /**************************
1606              * CALCULATE INTERACTIONS *
1607              **************************/
1608
1609             if (gmx_mm256_any_lt(rsq10,rcutoff2))
1610             {
1611
1612             r10              = _mm256_mul_ps(rsq10,rinv10);
1613             r10              = _mm256_andnot_ps(dummy_mask,r10);
1614
1615             /* Compute parameters for interactions between i and j atoms */
1616             qq10             = _mm256_mul_ps(iq1,jq0);
1617
1618             /* EWALD ELECTROSTATICS */
1619
1620             /* Analytical PME correction */
1621             zeta2            = _mm256_mul_ps(beta2,rsq10);
1622             rinv3            = _mm256_mul_ps(rinvsq10,rinv10);
1623             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
1624             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1625             felec            = _mm256_mul_ps(qq10,felec);
1626             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
1627             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1628             velec            = _mm256_sub_ps(rinv10,pmecorrV);
1629             velec            = _mm256_mul_ps(qq10,velec);
1630
1631             d                = _mm256_sub_ps(r10,rswitch);
1632             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1633             d2               = _mm256_mul_ps(d,d);
1634             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)))))));
1635
1636             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1637
1638             /* Evaluate switch function */
1639             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1640             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv10,_mm256_mul_ps(velec,dsw)) );
1641             cutoff_mask      = _mm256_cmp_ps(rsq10,rcutoff2,_CMP_LT_OQ);
1642
1643             fscal            = felec;
1644
1645             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1646
1647             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1648
1649             /* Calculate temporary vectorial force */
1650             tx               = _mm256_mul_ps(fscal,dx10);
1651             ty               = _mm256_mul_ps(fscal,dy10);
1652             tz               = _mm256_mul_ps(fscal,dz10);
1653
1654             /* Update vectorial force */
1655             fix1             = _mm256_add_ps(fix1,tx);
1656             fiy1             = _mm256_add_ps(fiy1,ty);
1657             fiz1             = _mm256_add_ps(fiz1,tz);
1658
1659             fjx0             = _mm256_add_ps(fjx0,tx);
1660             fjy0             = _mm256_add_ps(fjy0,ty);
1661             fjz0             = _mm256_add_ps(fjz0,tz);
1662
1663             }
1664
1665             /**************************
1666              * CALCULATE INTERACTIONS *
1667              **************************/
1668
1669             if (gmx_mm256_any_lt(rsq20,rcutoff2))
1670             {
1671
1672             r20              = _mm256_mul_ps(rsq20,rinv20);
1673             r20              = _mm256_andnot_ps(dummy_mask,r20);
1674
1675             /* Compute parameters for interactions between i and j atoms */
1676             qq20             = _mm256_mul_ps(iq2,jq0);
1677
1678             /* EWALD ELECTROSTATICS */
1679
1680             /* Analytical PME correction */
1681             zeta2            = _mm256_mul_ps(beta2,rsq20);
1682             rinv3            = _mm256_mul_ps(rinvsq20,rinv20);
1683             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
1684             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1685             felec            = _mm256_mul_ps(qq20,felec);
1686             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
1687             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1688             velec            = _mm256_sub_ps(rinv20,pmecorrV);
1689             velec            = _mm256_mul_ps(qq20,velec);
1690
1691             d                = _mm256_sub_ps(r20,rswitch);
1692             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1693             d2               = _mm256_mul_ps(d,d);
1694             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)))))));
1695
1696             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1697
1698             /* Evaluate switch function */
1699             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1700             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv20,_mm256_mul_ps(velec,dsw)) );
1701             cutoff_mask      = _mm256_cmp_ps(rsq20,rcutoff2,_CMP_LT_OQ);
1702
1703             fscal            = felec;
1704
1705             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1706
1707             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1708
1709             /* Calculate temporary vectorial force */
1710             tx               = _mm256_mul_ps(fscal,dx20);
1711             ty               = _mm256_mul_ps(fscal,dy20);
1712             tz               = _mm256_mul_ps(fscal,dz20);
1713
1714             /* Update vectorial force */
1715             fix2             = _mm256_add_ps(fix2,tx);
1716             fiy2             = _mm256_add_ps(fiy2,ty);
1717             fiz2             = _mm256_add_ps(fiz2,tz);
1718
1719             fjx0             = _mm256_add_ps(fjx0,tx);
1720             fjy0             = _mm256_add_ps(fjy0,ty);
1721             fjz0             = _mm256_add_ps(fjz0,tz);
1722
1723             }
1724
1725             /**************************
1726              * CALCULATE INTERACTIONS *
1727              **************************/
1728
1729             if (gmx_mm256_any_lt(rsq30,rcutoff2))
1730             {
1731
1732             r30              = _mm256_mul_ps(rsq30,rinv30);
1733             r30              = _mm256_andnot_ps(dummy_mask,r30);
1734
1735             /* Compute parameters for interactions between i and j atoms */
1736             qq30             = _mm256_mul_ps(iq3,jq0);
1737
1738             /* EWALD ELECTROSTATICS */
1739
1740             /* Analytical PME correction */
1741             zeta2            = _mm256_mul_ps(beta2,rsq30);
1742             rinv3            = _mm256_mul_ps(rinvsq30,rinv30);
1743             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
1744             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1745             felec            = _mm256_mul_ps(qq30,felec);
1746             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
1747             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1748             velec            = _mm256_sub_ps(rinv30,pmecorrV);
1749             velec            = _mm256_mul_ps(qq30,velec);
1750
1751             d                = _mm256_sub_ps(r30,rswitch);
1752             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1753             d2               = _mm256_mul_ps(d,d);
1754             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)))))));
1755
1756             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1757
1758             /* Evaluate switch function */
1759             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1760             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv30,_mm256_mul_ps(velec,dsw)) );
1761             cutoff_mask      = _mm256_cmp_ps(rsq30,rcutoff2,_CMP_LT_OQ);
1762
1763             fscal            = felec;
1764
1765             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1766
1767             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1768
1769             /* Calculate temporary vectorial force */
1770             tx               = _mm256_mul_ps(fscal,dx30);
1771             ty               = _mm256_mul_ps(fscal,dy30);
1772             tz               = _mm256_mul_ps(fscal,dz30);
1773
1774             /* Update vectorial force */
1775             fix3             = _mm256_add_ps(fix3,tx);
1776             fiy3             = _mm256_add_ps(fiy3,ty);
1777             fiz3             = _mm256_add_ps(fiz3,tz);
1778
1779             fjx0             = _mm256_add_ps(fjx0,tx);
1780             fjy0             = _mm256_add_ps(fjy0,ty);
1781             fjz0             = _mm256_add_ps(fjz0,tz);
1782
1783             }
1784
1785             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
1786             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
1787             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
1788             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
1789             fjptrE             = (jnrlistE>=0) ? f+j_coord_offsetE : scratch;
1790             fjptrF             = (jnrlistF>=0) ? f+j_coord_offsetF : scratch;
1791             fjptrG             = (jnrlistG>=0) ? f+j_coord_offsetG : scratch;
1792             fjptrH             = (jnrlistH>=0) ? f+j_coord_offsetH : scratch;
1793
1794             gmx_mm256_decrement_1rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,fjx0,fjy0,fjz0);
1795
1796             /* Inner loop uses 378 flops */
1797         }
1798
1799         /* End of innermost loop */
1800
1801         gmx_mm256_update_iforce_4atom_swizzle_ps(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
1802                                                  f+i_coord_offset,fshift+i_shift_offset);
1803
1804         /* Increment number of inner iterations */
1805         inneriter                  += j_index_end - j_index_start;
1806
1807         /* Outer loop uses 24 flops */
1808     }
1809
1810     /* Increment number of outer iterations */
1811     outeriter        += nri;
1812
1813     /* Update outer/inner flops */
1814
1815     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_F,outeriter*24 + inneriter*378);
1816 }