src/gromacs/gmxlib/nonbonded/nb_kernel_avx_256_single/nb_kernel_ElecEwSw_VdwLJSw_GeomW4W4_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 "gromacs/legacyheaders/types/simple.h"
  44 #include "gromacs/math/vec.h"
  45 #include "gromacs/legacyheaders/nrnb.h"
  46
  47 #include "gromacs/simd/math_x86_avx_256_single.h"
  48 #include "kernelutil_x86_avx_256_single.h"
  49
  50 /*
  51  * Gromacs nonbonded kernel:   nb_kernel_ElecEwSw_VdwLJSw_GeomW4W4_VF_avx_256_single
  52  * Electrostatics interaction: Ewald
  53  * VdW interaction:            LennardJones
  54  * Geometry:                   Water4-Water4
  55  * Calculate force/pot:        PotentialAndForce
  56  */
  57 void
  58 nb_kernel_ElecEwSw_VdwLJSw_GeomW4W4_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     int              vdwjidx1A,vdwjidx1B,vdwjidx1C,vdwjidx1D,vdwjidx1E,vdwjidx1F,vdwjidx1G,vdwjidx1H;
  97     __m256           jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
  98     int              vdwjidx2A,vdwjidx2B,vdwjidx2C,vdwjidx2D,vdwjidx2E,vdwjidx2F,vdwjidx2G,vdwjidx2H;
  99     __m256           jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
 100     int              vdwjidx3A,vdwjidx3B,vdwjidx3C,vdwjidx3D,vdwjidx3E,vdwjidx3F,vdwjidx3G,vdwjidx3H;
 101     __m256           jx3,jy3,jz3,fjx3,fjy3,fjz3,jq3,isaj3;
 102     __m256           dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
 103     __m256           dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11;
 104     __m256           dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12;
 105     __m256           dx13,dy13,dz13,rsq13,rinv13,rinvsq13,r13,qq13,c6_13,c12_13;
 106     __m256           dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21;
 107     __m256           dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22;
 108     __m256           dx23,dy23,dz23,rsq23,rinv23,rinvsq23,r23,qq23,c6_23,c12_23;
 109     __m256           dx31,dy31,dz31,rsq31,rinv31,rinvsq31,r31,qq31,c6_31,c12_31;
 110     __m256           dx32,dy32,dz32,rsq32,rinv32,rinvsq32,r32,qq32,c6_32,c12_32;
 111     __m256           dx33,dy33,dz33,rsq33,rinv33,rinvsq33,r33,qq33,c6_33,c12_33;
 112     __m256           velec,felec,velecsum,facel,crf,krf,krf2;
 113     real             *charge;
 114     int              nvdwtype;
 115     __m256           rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
 116     int              *vdwtype;
 117     real             *vdwparam;
 118     __m256           one_sixth   = _mm256_set1_ps(1.0/6.0);
 119     __m256           one_twelfth = _mm256_set1_ps(1.0/12.0);
 120     __m256i          ewitab;
 121     __m128i          ewitab_lo,ewitab_hi;
 122     __m256           ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
 123     __m256           beta,beta2,beta3,zeta2,pmecorrF,pmecorrV,rinv3;
 124     real             *ewtab;
 125     __m256           rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
 126     real             rswitch_scalar,d_scalar;
 127     __m256           dummy_mask,cutoff_mask;
 128     __m256           signbit = _mm256_castsi256_ps( _mm256_set1_epi32(0x80000000) );
 129     __m256           one     = _mm256_set1_ps(1.0);
 130     __m256           two     = _mm256_set1_ps(2.0);
 131     x                = xx[0];
 132     f                = ff[0];
 133
 134     nri              = nlist->nri;
 135     iinr             = nlist->iinr;
 136     jindex           = nlist->jindex;
 137     jjnr             = nlist->jjnr;
 138     shiftidx         = nlist->shift;
 139     gid              = nlist->gid;
 140     shiftvec         = fr->shift_vec[0];
 141     fshift           = fr->fshift[0];
 142     facel            = _mm256_set1_ps(fr->epsfac);
 143     charge           = mdatoms->chargeA;
 144     nvdwtype         = fr->ntype;
 145     vdwparam         = fr->nbfp;
 146     vdwtype          = mdatoms->typeA;
 147
 148     sh_ewald         = _mm256_set1_ps(fr->ic->sh_ewald);
 149     beta             = _mm256_set1_ps(fr->ic->ewaldcoeff_q);
 150     beta2            = _mm256_mul_ps(beta,beta);
 151     beta3            = _mm256_mul_ps(beta,beta2);
 152
 153     ewtab            = fr->ic->tabq_coul_FDV0;
 154     ewtabscale       = _mm256_set1_ps(fr->ic->tabq_scale);
 155     ewtabhalfspace   = _mm256_set1_ps(0.5/fr->ic->tabq_scale);
 156
 157     /* Setup water-specific parameters */
 158     inr              = nlist->iinr[0];
 159     iq1              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+1]));
 160     iq2              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+2]));
 161     iq3              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+3]));
 162     vdwioffsetptr0   = vdwparam+2*nvdwtype*vdwtype[inr+0];
 163
 164     jq1              = _mm256_set1_ps(charge[inr+1]);
 165     jq2              = _mm256_set1_ps(charge[inr+2]);
 166     jq3              = _mm256_set1_ps(charge[inr+3]);
 167     vdwjidx0A        = 2*vdwtype[inr+0];
 168     c6_00            = _mm256_set1_ps(vdwioffsetptr0[vdwjidx0A]);
 169     c12_00           = _mm256_set1_ps(vdwioffsetptr0[vdwjidx0A+1]);
 170     qq11             = _mm256_mul_ps(iq1,jq1);
 171     qq12             = _mm256_mul_ps(iq1,jq2);
 172     qq13             = _mm256_mul_ps(iq1,jq3);
 173     qq21             = _mm256_mul_ps(iq2,jq1);
 174     qq22             = _mm256_mul_ps(iq2,jq2);
 175     qq23             = _mm256_mul_ps(iq2,jq3);
 176     qq31             = _mm256_mul_ps(iq3,jq1);
 177     qq32             = _mm256_mul_ps(iq3,jq2);
 178     qq33             = _mm256_mul_ps(iq3,jq3);
 179
 180     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 181     rcutoff_scalar   = fr->rcoulomb;
 182     rcutoff          = _mm256_set1_ps(rcutoff_scalar);
 183     rcutoff2         = _mm256_mul_ps(rcutoff,rcutoff);
 184
 185     rswitch_scalar   = fr->rcoulomb_switch;
 186     rswitch          = _mm256_set1_ps(rswitch_scalar);
 187     /* Setup switch parameters */
 188     d_scalar         = rcutoff_scalar-rswitch_scalar;
 189     d                = _mm256_set1_ps(d_scalar);
 190     swV3             = _mm256_set1_ps(-10.0/(d_scalar*d_scalar*d_scalar));
 191     swV4             = _mm256_set1_ps( 15.0/(d_scalar*d_scalar*d_scalar*d_scalar));
 192     swV5             = _mm256_set1_ps( -6.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
 193     swF2             = _mm256_set1_ps(-30.0/(d_scalar*d_scalar*d_scalar));
 194     swF3             = _mm256_set1_ps( 60.0/(d_scalar*d_scalar*d_scalar*d_scalar));
 195     swF4             = _mm256_set1_ps(-30.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
 196
 197     /* Avoid stupid compiler warnings */
 198     jnrA = jnrB = jnrC = jnrD = jnrE = jnrF = jnrG = jnrH = 0;
 199     j_coord_offsetA = 0;
 200     j_coord_offsetB = 0;
 201     j_coord_offsetC = 0;
 202     j_coord_offsetD = 0;
 203     j_coord_offsetE = 0;
 204     j_coord_offsetF = 0;
 205     j_coord_offsetG = 0;
 206     j_coord_offsetH = 0;
 207
 208     outeriter        = 0;
 209     inneriter        = 0;
 210
 211     for(iidx=0;iidx<4*DIM;iidx++)
 212     {
 213         scratch[iidx] = 0.0;
 214     }
 215
 216     /* Start outer loop over neighborlists */
 217     for(iidx=0; iidx<nri; iidx++)
 218     {
 219         /* Load shift vector for this list */
 220         i_shift_offset   = DIM*shiftidx[iidx];
 221
 222         /* Load limits for loop over neighbors */
 223         j_index_start    = jindex[iidx];
 224         j_index_end      = jindex[iidx+1];
 225
 226         /* Get outer coordinate index */
 227         inr              = iinr[iidx];
 228         i_coord_offset   = DIM*inr;
 229
 230         /* Load i particle coords and add shift vector */
 231         gmx_mm256_load_shift_and_4rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,
 232                                                     &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
 233
 234         fix0             = _mm256_setzero_ps();
 235         fiy0             = _mm256_setzero_ps();
 236         fiz0             = _mm256_setzero_ps();
 237         fix1             = _mm256_setzero_ps();
 238         fiy1             = _mm256_setzero_ps();
 239         fiz1             = _mm256_setzero_ps();
 240         fix2             = _mm256_setzero_ps();
 241         fiy2             = _mm256_setzero_ps();
 242         fiz2             = _mm256_setzero_ps();
 243         fix3             = _mm256_setzero_ps();
 244         fiy3             = _mm256_setzero_ps();
 245         fiz3             = _mm256_setzero_ps();
 246
 247         /* Reset potential sums */
 248         velecsum         = _mm256_setzero_ps();
 249         vvdwsum          = _mm256_setzero_ps();
 250
 251         /* Start inner kernel loop */
 252         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+7]>=0; jidx+=8)
 253         {
 254
 255             /* Get j neighbor index, and coordinate index */
 256             jnrA             = jjnr[jidx];
 257             jnrB             = jjnr[jidx+1];
 258             jnrC             = jjnr[jidx+2];
 259             jnrD             = jjnr[jidx+3];
 260             jnrE             = jjnr[jidx+4];
 261             jnrF             = jjnr[jidx+5];
 262             jnrG             = jjnr[jidx+6];
 263             jnrH             = jjnr[jidx+7];
 264             j_coord_offsetA  = DIM*jnrA;
 265             j_coord_offsetB  = DIM*jnrB;
 266             j_coord_offsetC  = DIM*jnrC;
 267             j_coord_offsetD  = DIM*jnrD;
 268             j_coord_offsetE  = DIM*jnrE;
 269             j_coord_offsetF  = DIM*jnrF;
 270             j_coord_offsetG  = DIM*jnrG;
 271             j_coord_offsetH  = DIM*jnrH;
 272
 273             /* load j atom coordinates */
 274             gmx_mm256_load_4rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 275                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 276                                                  x+j_coord_offsetE,x+j_coord_offsetF,
 277                                                  x+j_coord_offsetG,x+j_coord_offsetH,
 278                                                  &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,
 279                                                  &jy2,&jz2,&jx3,&jy3,&jz3);
 280
 281             /* Calculate displacement vector */
 282             dx00             = _mm256_sub_ps(ix0,jx0);
 283             dy00             = _mm256_sub_ps(iy0,jy0);
 284             dz00             = _mm256_sub_ps(iz0,jz0);
 285             dx11             = _mm256_sub_ps(ix1,jx1);
 286             dy11             = _mm256_sub_ps(iy1,jy1);
 287             dz11             = _mm256_sub_ps(iz1,jz1);
 288             dx12             = _mm256_sub_ps(ix1,jx2);
 289             dy12             = _mm256_sub_ps(iy1,jy2);
 290             dz12             = _mm256_sub_ps(iz1,jz2);
 291             dx13             = _mm256_sub_ps(ix1,jx3);
 292             dy13             = _mm256_sub_ps(iy1,jy3);
 293             dz13             = _mm256_sub_ps(iz1,jz3);
 294             dx21             = _mm256_sub_ps(ix2,jx1);
 295             dy21             = _mm256_sub_ps(iy2,jy1);
 296             dz21             = _mm256_sub_ps(iz2,jz1);
 297             dx22             = _mm256_sub_ps(ix2,jx2);
 298             dy22             = _mm256_sub_ps(iy2,jy2);
 299             dz22             = _mm256_sub_ps(iz2,jz2);
 300             dx23             = _mm256_sub_ps(ix2,jx3);
 301             dy23             = _mm256_sub_ps(iy2,jy3);
 302             dz23             = _mm256_sub_ps(iz2,jz3);
 303             dx31             = _mm256_sub_ps(ix3,jx1);
 304             dy31             = _mm256_sub_ps(iy3,jy1);
 305             dz31             = _mm256_sub_ps(iz3,jz1);
 306             dx32             = _mm256_sub_ps(ix3,jx2);
 307             dy32             = _mm256_sub_ps(iy3,jy2);
 308             dz32             = _mm256_sub_ps(iz3,jz2);
 309             dx33             = _mm256_sub_ps(ix3,jx3);
 310             dy33             = _mm256_sub_ps(iy3,jy3);
 311             dz33             = _mm256_sub_ps(iz3,jz3);
 312
 313             /* Calculate squared distance and things based on it */
 314             rsq00            = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
 315             rsq11            = gmx_mm256_calc_rsq_ps(dx11,dy11,dz11);
 316             rsq12            = gmx_mm256_calc_rsq_ps(dx12,dy12,dz12);
 317             rsq13            = gmx_mm256_calc_rsq_ps(dx13,dy13,dz13);
 318             rsq21            = gmx_mm256_calc_rsq_ps(dx21,dy21,dz21);
 319             rsq22            = gmx_mm256_calc_rsq_ps(dx22,dy22,dz22);
 320             rsq23            = gmx_mm256_calc_rsq_ps(dx23,dy23,dz23);
 321             rsq31            = gmx_mm256_calc_rsq_ps(dx31,dy31,dz31);
 322             rsq32            = gmx_mm256_calc_rsq_ps(dx32,dy32,dz32);
 323             rsq33            = gmx_mm256_calc_rsq_ps(dx33,dy33,dz33);
 324
 325             rinv00           = gmx_mm256_invsqrt_ps(rsq00);
 326             rinv11           = gmx_mm256_invsqrt_ps(rsq11);
 327             rinv12           = gmx_mm256_invsqrt_ps(rsq12);
 328             rinv13           = gmx_mm256_invsqrt_ps(rsq13);
 329             rinv21           = gmx_mm256_invsqrt_ps(rsq21);
 330             rinv22           = gmx_mm256_invsqrt_ps(rsq22);
 331             rinv23           = gmx_mm256_invsqrt_ps(rsq23);
 332             rinv31           = gmx_mm256_invsqrt_ps(rsq31);
 333             rinv32           = gmx_mm256_invsqrt_ps(rsq32);
 334             rinv33           = gmx_mm256_invsqrt_ps(rsq33);
 335
 336             rinvsq00         = _mm256_mul_ps(rinv00,rinv00);
 337             rinvsq11         = _mm256_mul_ps(rinv11,rinv11);
 338             rinvsq12         = _mm256_mul_ps(rinv12,rinv12);
 339             rinvsq13         = _mm256_mul_ps(rinv13,rinv13);
 340             rinvsq21         = _mm256_mul_ps(rinv21,rinv21);
 341             rinvsq22         = _mm256_mul_ps(rinv22,rinv22);
 342             rinvsq23         = _mm256_mul_ps(rinv23,rinv23);
 343             rinvsq31         = _mm256_mul_ps(rinv31,rinv31);
 344             rinvsq32         = _mm256_mul_ps(rinv32,rinv32);
 345             rinvsq33         = _mm256_mul_ps(rinv33,rinv33);
 346
 347             fjx0             = _mm256_setzero_ps();
 348             fjy0             = _mm256_setzero_ps();
 349             fjz0             = _mm256_setzero_ps();
 350             fjx1             = _mm256_setzero_ps();
 351             fjy1             = _mm256_setzero_ps();
 352             fjz1             = _mm256_setzero_ps();
 353             fjx2             = _mm256_setzero_ps();
 354             fjy2             = _mm256_setzero_ps();
 355             fjz2             = _mm256_setzero_ps();
 356             fjx3             = _mm256_setzero_ps();
 357             fjy3             = _mm256_setzero_ps();
 358             fjz3             = _mm256_setzero_ps();
 359
 360             /**************************
 361              * CALCULATE INTERACTIONS *
 362              **************************/
 363
 364             if (gmx_mm256_any_lt(rsq00,rcutoff2))
 365             {
 366
 367             r00              = _mm256_mul_ps(rsq00,rinv00);
 368
 369             /* LENNARD-JONES DISPERSION/REPULSION */
 370
 371             rinvsix          = _mm256_mul_ps(_mm256_mul_ps(rinvsq00,rinvsq00),rinvsq00);
 372             vvdw6            = _mm256_mul_ps(c6_00,rinvsix);
 373             vvdw12           = _mm256_mul_ps(c12_00,_mm256_mul_ps(rinvsix,rinvsix));
 374             vvdw             = _mm256_sub_ps( _mm256_mul_ps(vvdw12,one_twelfth) , _mm256_mul_ps(vvdw6,one_sixth) );
 375             fvdw             = _mm256_mul_ps(_mm256_sub_ps(vvdw12,vvdw6),rinvsq00);
 376
 377             d                = _mm256_sub_ps(r00,rswitch);
 378             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 379             d2               = _mm256_mul_ps(d,d);
 380             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)))))));
 381
 382             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 383
 384             /* Evaluate switch function */
 385             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 386             fvdw             = _mm256_sub_ps( _mm256_mul_ps(fvdw,sw) , _mm256_mul_ps(rinv00,_mm256_mul_ps(vvdw,dsw)) );
 387             vvdw             = _mm256_mul_ps(vvdw,sw);
 388             cutoff_mask      = _mm256_cmp_ps(rsq00,rcutoff2,_CMP_LT_OQ);
 389
 390             /* Update potential sum for this i atom from the interaction with this j atom. */
 391             vvdw             = _mm256_and_ps(vvdw,cutoff_mask);
 392             vvdwsum          = _mm256_add_ps(vvdwsum,vvdw);
 393
 394             fscal            = fvdw;
 395
 396             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 397
 398             /* Calculate temporary vectorial force */
 399             tx               = _mm256_mul_ps(fscal,dx00);
 400             ty               = _mm256_mul_ps(fscal,dy00);
 401             tz               = _mm256_mul_ps(fscal,dz00);
 402
 403             /* Update vectorial force */
 404             fix0             = _mm256_add_ps(fix0,tx);
 405             fiy0             = _mm256_add_ps(fiy0,ty);
 406             fiz0             = _mm256_add_ps(fiz0,tz);
 407
 408             fjx0             = _mm256_add_ps(fjx0,tx);
 409             fjy0             = _mm256_add_ps(fjy0,ty);
 410             fjz0             = _mm256_add_ps(fjz0,tz);
 411
 412             }
 413
 414             /**************************
 415              * CALCULATE INTERACTIONS *
 416              **************************/
 417
 418             if (gmx_mm256_any_lt(rsq11,rcutoff2))
 419             {
 420
 421             r11              = _mm256_mul_ps(rsq11,rinv11);
 422
 423             /* EWALD ELECTROSTATICS */
 424
 425             /* Analytical PME correction */
 426             zeta2            = _mm256_mul_ps(beta2,rsq11);
 427             rinv3            = _mm256_mul_ps(rinvsq11,rinv11);
 428             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
 429             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 430             felec            = _mm256_mul_ps(qq11,felec);
 431             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
 432             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 433             velec            = _mm256_sub_ps(rinv11,pmecorrV);
 434             velec            = _mm256_mul_ps(qq11,velec);
 435
 436             d                = _mm256_sub_ps(r11,rswitch);
 437             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 438             d2               = _mm256_mul_ps(d,d);
 439             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)))))));
 440
 441             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 442
 443             /* Evaluate switch function */
 444             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 445             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv11,_mm256_mul_ps(velec,dsw)) );
 446             velec            = _mm256_mul_ps(velec,sw);
 447             cutoff_mask      = _mm256_cmp_ps(rsq11,rcutoff2,_CMP_LT_OQ);
 448
 449             /* Update potential sum for this i atom from the interaction with this j atom. */
 450             velec            = _mm256_and_ps(velec,cutoff_mask);
 451             velecsum         = _mm256_add_ps(velecsum,velec);
 452
 453             fscal            = felec;
 454
 455             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 456
 457             /* Calculate temporary vectorial force */
 458             tx               = _mm256_mul_ps(fscal,dx11);
 459             ty               = _mm256_mul_ps(fscal,dy11);
 460             tz               = _mm256_mul_ps(fscal,dz11);
 461
 462             /* Update vectorial force */
 463             fix1             = _mm256_add_ps(fix1,tx);
 464             fiy1             = _mm256_add_ps(fiy1,ty);
 465             fiz1             = _mm256_add_ps(fiz1,tz);
 466
 467             fjx1             = _mm256_add_ps(fjx1,tx);
 468             fjy1             = _mm256_add_ps(fjy1,ty);
 469             fjz1             = _mm256_add_ps(fjz1,tz);
 470
 471             }
 472
 473             /**************************
 474              * CALCULATE INTERACTIONS *
 475              **************************/
 476
 477             if (gmx_mm256_any_lt(rsq12,rcutoff2))
 478             {
 479
 480             r12              = _mm256_mul_ps(rsq12,rinv12);
 481
 482             /* EWALD ELECTROSTATICS */
 483
 484             /* Analytical PME correction */
 485             zeta2            = _mm256_mul_ps(beta2,rsq12);
 486             rinv3            = _mm256_mul_ps(rinvsq12,rinv12);
 487             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
 488             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 489             felec            = _mm256_mul_ps(qq12,felec);
 490             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
 491             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 492             velec            = _mm256_sub_ps(rinv12,pmecorrV);
 493             velec            = _mm256_mul_ps(qq12,velec);
 494
 495             d                = _mm256_sub_ps(r12,rswitch);
 496             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 497             d2               = _mm256_mul_ps(d,d);
 498             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)))))));
 499
 500             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 501
 502             /* Evaluate switch function */
 503             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 504             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv12,_mm256_mul_ps(velec,dsw)) );
 505             velec            = _mm256_mul_ps(velec,sw);
 506             cutoff_mask      = _mm256_cmp_ps(rsq12,rcutoff2,_CMP_LT_OQ);
 507
 508             /* Update potential sum for this i atom from the interaction with this j atom. */
 509             velec            = _mm256_and_ps(velec,cutoff_mask);
 510             velecsum         = _mm256_add_ps(velecsum,velec);
 511
 512             fscal            = felec;
 513
 514             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 515
 516             /* Calculate temporary vectorial force */
 517             tx               = _mm256_mul_ps(fscal,dx12);
 518             ty               = _mm256_mul_ps(fscal,dy12);
 519             tz               = _mm256_mul_ps(fscal,dz12);
 520
 521             /* Update vectorial force */
 522             fix1             = _mm256_add_ps(fix1,tx);
 523             fiy1             = _mm256_add_ps(fiy1,ty);
 524             fiz1             = _mm256_add_ps(fiz1,tz);
 525
 526             fjx2             = _mm256_add_ps(fjx2,tx);
 527             fjy2             = _mm256_add_ps(fjy2,ty);
 528             fjz2             = _mm256_add_ps(fjz2,tz);
 529
 530             }
 531
 532             /**************************
 533              * CALCULATE INTERACTIONS *
 534              **************************/
 535
 536             if (gmx_mm256_any_lt(rsq13,rcutoff2))
 537             {
 538
 539             r13              = _mm256_mul_ps(rsq13,rinv13);
 540
 541             /* EWALD ELECTROSTATICS */
 542
 543             /* Analytical PME correction */
 544             zeta2            = _mm256_mul_ps(beta2,rsq13);
 545             rinv3            = _mm256_mul_ps(rinvsq13,rinv13);
 546             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
 547             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 548             felec            = _mm256_mul_ps(qq13,felec);
 549             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
 550             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 551             velec            = _mm256_sub_ps(rinv13,pmecorrV);
 552             velec            = _mm256_mul_ps(qq13,velec);
 553
 554             d                = _mm256_sub_ps(r13,rswitch);
 555             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 556             d2               = _mm256_mul_ps(d,d);
 557             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)))))));
 558
 559             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 560
 561             /* Evaluate switch function */
 562             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 563             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv13,_mm256_mul_ps(velec,dsw)) );
 564             velec            = _mm256_mul_ps(velec,sw);
 565             cutoff_mask      = _mm256_cmp_ps(rsq13,rcutoff2,_CMP_LT_OQ);
 566
 567             /* Update potential sum for this i atom from the interaction with this j atom. */
 568             velec            = _mm256_and_ps(velec,cutoff_mask);
 569             velecsum         = _mm256_add_ps(velecsum,velec);
 570
 571             fscal            = felec;
 572
 573             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 574
 575             /* Calculate temporary vectorial force */
 576             tx               = _mm256_mul_ps(fscal,dx13);
 577             ty               = _mm256_mul_ps(fscal,dy13);
 578             tz               = _mm256_mul_ps(fscal,dz13);
 579
 580             /* Update vectorial force */
 581             fix1             = _mm256_add_ps(fix1,tx);
 582             fiy1             = _mm256_add_ps(fiy1,ty);
 583             fiz1             = _mm256_add_ps(fiz1,tz);
 584
 585             fjx3             = _mm256_add_ps(fjx3,tx);
 586             fjy3             = _mm256_add_ps(fjy3,ty);
 587             fjz3             = _mm256_add_ps(fjz3,tz);
 588
 589             }
 590
 591             /**************************
 592              * CALCULATE INTERACTIONS *
 593              **************************/
 594
 595             if (gmx_mm256_any_lt(rsq21,rcutoff2))
 596             {
 597
 598             r21              = _mm256_mul_ps(rsq21,rinv21);
 599
 600             /* EWALD ELECTROSTATICS */
 601
 602             /* Analytical PME correction */
 603             zeta2            = _mm256_mul_ps(beta2,rsq21);
 604             rinv3            = _mm256_mul_ps(rinvsq21,rinv21);
 605             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
 606             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 607             felec            = _mm256_mul_ps(qq21,felec);
 608             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
 609             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 610             velec            = _mm256_sub_ps(rinv21,pmecorrV);
 611             velec            = _mm256_mul_ps(qq21,velec);
 612
 613             d                = _mm256_sub_ps(r21,rswitch);
 614             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 615             d2               = _mm256_mul_ps(d,d);
 616             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)))))));
 617
 618             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 619
 620             /* Evaluate switch function */
 621             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 622             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv21,_mm256_mul_ps(velec,dsw)) );
 623             velec            = _mm256_mul_ps(velec,sw);
 624             cutoff_mask      = _mm256_cmp_ps(rsq21,rcutoff2,_CMP_LT_OQ);
 625
 626             /* Update potential sum for this i atom from the interaction with this j atom. */
 627             velec            = _mm256_and_ps(velec,cutoff_mask);
 628             velecsum         = _mm256_add_ps(velecsum,velec);
 629
 630             fscal            = felec;
 631
 632             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 633
 634             /* Calculate temporary vectorial force */
 635             tx               = _mm256_mul_ps(fscal,dx21);
 636             ty               = _mm256_mul_ps(fscal,dy21);
 637             tz               = _mm256_mul_ps(fscal,dz21);
 638
 639             /* Update vectorial force */
 640             fix2             = _mm256_add_ps(fix2,tx);
 641             fiy2             = _mm256_add_ps(fiy2,ty);
 642             fiz2             = _mm256_add_ps(fiz2,tz);
 643
 644             fjx1             = _mm256_add_ps(fjx1,tx);
 645             fjy1             = _mm256_add_ps(fjy1,ty);
 646             fjz1             = _mm256_add_ps(fjz1,tz);
 647
 648             }
 649
 650             /**************************
 651              * CALCULATE INTERACTIONS *
 652              **************************/
 653
 654             if (gmx_mm256_any_lt(rsq22,rcutoff2))
 655             {
 656
 657             r22              = _mm256_mul_ps(rsq22,rinv22);
 658
 659             /* EWALD ELECTROSTATICS */
 660
 661             /* Analytical PME correction */
 662             zeta2            = _mm256_mul_ps(beta2,rsq22);
 663             rinv3            = _mm256_mul_ps(rinvsq22,rinv22);
 664             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
 665             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 666             felec            = _mm256_mul_ps(qq22,felec);
 667             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
 668             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 669             velec            = _mm256_sub_ps(rinv22,pmecorrV);
 670             velec            = _mm256_mul_ps(qq22,velec);
 671
 672             d                = _mm256_sub_ps(r22,rswitch);
 673             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 674             d2               = _mm256_mul_ps(d,d);
 675             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)))))));
 676
 677             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 678
 679             /* Evaluate switch function */
 680             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 681             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv22,_mm256_mul_ps(velec,dsw)) );
 682             velec            = _mm256_mul_ps(velec,sw);
 683             cutoff_mask      = _mm256_cmp_ps(rsq22,rcutoff2,_CMP_LT_OQ);
 684
 685             /* Update potential sum for this i atom from the interaction with this j atom. */
 686             velec            = _mm256_and_ps(velec,cutoff_mask);
 687             velecsum         = _mm256_add_ps(velecsum,velec);
 688
 689             fscal            = felec;
 690
 691             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 692
 693             /* Calculate temporary vectorial force */
 694             tx               = _mm256_mul_ps(fscal,dx22);
 695             ty               = _mm256_mul_ps(fscal,dy22);
 696             tz               = _mm256_mul_ps(fscal,dz22);
 697
 698             /* Update vectorial force */
 699             fix2             = _mm256_add_ps(fix2,tx);
 700             fiy2             = _mm256_add_ps(fiy2,ty);
 701             fiz2             = _mm256_add_ps(fiz2,tz);
 702
 703             fjx2             = _mm256_add_ps(fjx2,tx);
 704             fjy2             = _mm256_add_ps(fjy2,ty);
 705             fjz2             = _mm256_add_ps(fjz2,tz);
 706
 707             }
 708
 709             /**************************
 710              * CALCULATE INTERACTIONS *
 711              **************************/
 712
 713             if (gmx_mm256_any_lt(rsq23,rcutoff2))
 714             {
 715
 716             r23              = _mm256_mul_ps(rsq23,rinv23);
 717
 718             /* EWALD ELECTROSTATICS */
 719
 720             /* Analytical PME correction */
 721             zeta2            = _mm256_mul_ps(beta2,rsq23);
 722             rinv3            = _mm256_mul_ps(rinvsq23,rinv23);
 723             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
 724             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 725             felec            = _mm256_mul_ps(qq23,felec);
 726             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
 727             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 728             velec            = _mm256_sub_ps(rinv23,pmecorrV);
 729             velec            = _mm256_mul_ps(qq23,velec);
 730
 731             d                = _mm256_sub_ps(r23,rswitch);
 732             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 733             d2               = _mm256_mul_ps(d,d);
 734             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)))))));
 735
 736             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 737
 738             /* Evaluate switch function */
 739             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 740             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv23,_mm256_mul_ps(velec,dsw)) );
 741             velec            = _mm256_mul_ps(velec,sw);
 742             cutoff_mask      = _mm256_cmp_ps(rsq23,rcutoff2,_CMP_LT_OQ);
 743
 744             /* Update potential sum for this i atom from the interaction with this j atom. */
 745             velec            = _mm256_and_ps(velec,cutoff_mask);
 746             velecsum         = _mm256_add_ps(velecsum,velec);
 747
 748             fscal            = felec;
 749
 750             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 751
 752             /* Calculate temporary vectorial force */
 753             tx               = _mm256_mul_ps(fscal,dx23);
 754             ty               = _mm256_mul_ps(fscal,dy23);
 755             tz               = _mm256_mul_ps(fscal,dz23);
 756
 757             /* Update vectorial force */
 758             fix2             = _mm256_add_ps(fix2,tx);
 759             fiy2             = _mm256_add_ps(fiy2,ty);
 760             fiz2             = _mm256_add_ps(fiz2,tz);
 761
 762             fjx3             = _mm256_add_ps(fjx3,tx);
 763             fjy3             = _mm256_add_ps(fjy3,ty);
 764             fjz3             = _mm256_add_ps(fjz3,tz);
 765
 766             }
 767
 768             /**************************
 769              * CALCULATE INTERACTIONS *
 770              **************************/
 771
 772             if (gmx_mm256_any_lt(rsq31,rcutoff2))
 773             {
 774
 775             r31              = _mm256_mul_ps(rsq31,rinv31);
 776
 777             /* EWALD ELECTROSTATICS */
 778
 779             /* Analytical PME correction */
 780             zeta2            = _mm256_mul_ps(beta2,rsq31);
 781             rinv3            = _mm256_mul_ps(rinvsq31,rinv31);
 782             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
 783             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 784             felec            = _mm256_mul_ps(qq31,felec);
 785             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
 786             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 787             velec            = _mm256_sub_ps(rinv31,pmecorrV);
 788             velec            = _mm256_mul_ps(qq31,velec);
 789
 790             d                = _mm256_sub_ps(r31,rswitch);
 791             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 792             d2               = _mm256_mul_ps(d,d);
 793             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)))))));
 794
 795             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 796
 797             /* Evaluate switch function */
 798             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 799             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv31,_mm256_mul_ps(velec,dsw)) );
 800             velec            = _mm256_mul_ps(velec,sw);
 801             cutoff_mask      = _mm256_cmp_ps(rsq31,rcutoff2,_CMP_LT_OQ);
 802
 803             /* Update potential sum for this i atom from the interaction with this j atom. */
 804             velec            = _mm256_and_ps(velec,cutoff_mask);
 805             velecsum         = _mm256_add_ps(velecsum,velec);
 806
 807             fscal            = felec;
 808
 809             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 810
 811             /* Calculate temporary vectorial force */
 812             tx               = _mm256_mul_ps(fscal,dx31);
 813             ty               = _mm256_mul_ps(fscal,dy31);
 814             tz               = _mm256_mul_ps(fscal,dz31);
 815
 816             /* Update vectorial force */
 817             fix3             = _mm256_add_ps(fix3,tx);
 818             fiy3             = _mm256_add_ps(fiy3,ty);
 819             fiz3             = _mm256_add_ps(fiz3,tz);
 820
 821             fjx1             = _mm256_add_ps(fjx1,tx);
 822             fjy1             = _mm256_add_ps(fjy1,ty);
 823             fjz1             = _mm256_add_ps(fjz1,tz);
 824
 825             }
 826
 827             /**************************
 828              * CALCULATE INTERACTIONS *
 829              **************************/
 830
 831             if (gmx_mm256_any_lt(rsq32,rcutoff2))
 832             {
 833
 834             r32              = _mm256_mul_ps(rsq32,rinv32);
 835
 836             /* EWALD ELECTROSTATICS */
 837
 838             /* Analytical PME correction */
 839             zeta2            = _mm256_mul_ps(beta2,rsq32);
 840             rinv3            = _mm256_mul_ps(rinvsq32,rinv32);
 841             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
 842             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 843             felec            = _mm256_mul_ps(qq32,felec);
 844             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
 845             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 846             velec            = _mm256_sub_ps(rinv32,pmecorrV);
 847             velec            = _mm256_mul_ps(qq32,velec);
 848
 849             d                = _mm256_sub_ps(r32,rswitch);
 850             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 851             d2               = _mm256_mul_ps(d,d);
 852             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)))))));
 853
 854             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 855
 856             /* Evaluate switch function */
 857             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 858             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv32,_mm256_mul_ps(velec,dsw)) );
 859             velec            = _mm256_mul_ps(velec,sw);
 860             cutoff_mask      = _mm256_cmp_ps(rsq32,rcutoff2,_CMP_LT_OQ);
 861
 862             /* Update potential sum for this i atom from the interaction with this j atom. */
 863             velec            = _mm256_and_ps(velec,cutoff_mask);
 864             velecsum         = _mm256_add_ps(velecsum,velec);
 865
 866             fscal            = felec;
 867
 868             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 869
 870             /* Calculate temporary vectorial force */
 871             tx               = _mm256_mul_ps(fscal,dx32);
 872             ty               = _mm256_mul_ps(fscal,dy32);
 873             tz               = _mm256_mul_ps(fscal,dz32);
 874
 875             /* Update vectorial force */
 876             fix3             = _mm256_add_ps(fix3,tx);
 877             fiy3             = _mm256_add_ps(fiy3,ty);
 878             fiz3             = _mm256_add_ps(fiz3,tz);
 879
 880             fjx2             = _mm256_add_ps(fjx2,tx);
 881             fjy2             = _mm256_add_ps(fjy2,ty);
 882             fjz2             = _mm256_add_ps(fjz2,tz);
 883
 884             }
 885
 886             /**************************
 887              * CALCULATE INTERACTIONS *
 888              **************************/
 889
 890             if (gmx_mm256_any_lt(rsq33,rcutoff2))
 891             {
 892
 893             r33              = _mm256_mul_ps(rsq33,rinv33);
 894
 895             /* EWALD ELECTROSTATICS */
 896
 897             /* Analytical PME correction */
 898             zeta2            = _mm256_mul_ps(beta2,rsq33);
 899             rinv3            = _mm256_mul_ps(rinvsq33,rinv33);
 900             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
 901             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 902             felec            = _mm256_mul_ps(qq33,felec);
 903             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
 904             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 905             velec            = _mm256_sub_ps(rinv33,pmecorrV);
 906             velec            = _mm256_mul_ps(qq33,velec);
 907
 908             d                = _mm256_sub_ps(r33,rswitch);
 909             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 910             d2               = _mm256_mul_ps(d,d);
 911             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)))))));
 912
 913             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 914
 915             /* Evaluate switch function */
 916             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 917             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv33,_mm256_mul_ps(velec,dsw)) );
 918             velec            = _mm256_mul_ps(velec,sw);
 919             cutoff_mask      = _mm256_cmp_ps(rsq33,rcutoff2,_CMP_LT_OQ);
 920
 921             /* Update potential sum for this i atom from the interaction with this j atom. */
 922             velec            = _mm256_and_ps(velec,cutoff_mask);
 923             velecsum         = _mm256_add_ps(velecsum,velec);
 924
 925             fscal            = felec;
 926
 927             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 928
 929             /* Calculate temporary vectorial force */
 930             tx               = _mm256_mul_ps(fscal,dx33);
 931             ty               = _mm256_mul_ps(fscal,dy33);
 932             tz               = _mm256_mul_ps(fscal,dz33);
 933
 934             /* Update vectorial force */
 935             fix3             = _mm256_add_ps(fix3,tx);
 936             fiy3             = _mm256_add_ps(fiy3,ty);
 937             fiz3             = _mm256_add_ps(fiz3,tz);
 938
 939             fjx3             = _mm256_add_ps(fjx3,tx);
 940             fjy3             = _mm256_add_ps(fjy3,ty);
 941             fjz3             = _mm256_add_ps(fjz3,tz);
 942
 943             }
 944
 945             fjptrA             = f+j_coord_offsetA;
 946             fjptrB             = f+j_coord_offsetB;
 947             fjptrC             = f+j_coord_offsetC;
 948             fjptrD             = f+j_coord_offsetD;
 949             fjptrE             = f+j_coord_offsetE;
 950             fjptrF             = f+j_coord_offsetF;
 951             fjptrG             = f+j_coord_offsetG;
 952             fjptrH             = f+j_coord_offsetH;
 953
 954             gmx_mm256_decrement_4rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,
 955                                                       fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,
 956                                                       fjx2,fjy2,fjz2,fjx3,fjy3,fjz3);
 957
 958             /* Inner loop uses 1034 flops */
 959         }
 960
 961         if(jidx<j_index_end)
 962         {
 963
 964             /* Get j neighbor index, and coordinate index */
 965             jnrlistA         = jjnr[jidx];
 966             jnrlistB         = jjnr[jidx+1];
 967             jnrlistC         = jjnr[jidx+2];
 968             jnrlistD         = jjnr[jidx+3];
 969             jnrlistE         = jjnr[jidx+4];
 970             jnrlistF         = jjnr[jidx+5];
 971             jnrlistG         = jjnr[jidx+6];
 972             jnrlistH         = jjnr[jidx+7];
 973             /* Sign of each element will be negative for non-real atoms.
 974              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
 975              * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
 976              */
 977             dummy_mask = gmx_mm256_set_m128(gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx+4)),_mm_setzero_si128())),
 978                                             gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128())));
 979
 980             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
 981             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
 982             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
 983             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
 984             jnrE       = (jnrlistE>=0) ? jnrlistE : 0;
 985             jnrF       = (jnrlistF>=0) ? jnrlistF : 0;
 986             jnrG       = (jnrlistG>=0) ? jnrlistG : 0;
 987             jnrH       = (jnrlistH>=0) ? jnrlistH : 0;
 988             j_coord_offsetA  = DIM*jnrA;
 989             j_coord_offsetB  = DIM*jnrB;
 990             j_coord_offsetC  = DIM*jnrC;
 991             j_coord_offsetD  = DIM*jnrD;
 992             j_coord_offsetE  = DIM*jnrE;
 993             j_coord_offsetF  = DIM*jnrF;
 994             j_coord_offsetG  = DIM*jnrG;
 995             j_coord_offsetH  = DIM*jnrH;
 996
 997             /* load j atom coordinates */
 998             gmx_mm256_load_4rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 999                                                  x+j_coord_offsetC,x+j_coord_offsetD,
1000                                                  x+j_coord_offsetE,x+j_coord_offsetF,
1001                                                  x+j_coord_offsetG,x+j_coord_offsetH,
1002                                                  &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,
1003                                                  &jy2,&jz2,&jx3,&jy3,&jz3);
1004
1005             /* Calculate displacement vector */
1006             dx00             = _mm256_sub_ps(ix0,jx0);
1007             dy00             = _mm256_sub_ps(iy0,jy0);
1008             dz00             = _mm256_sub_ps(iz0,jz0);
1009             dx11             = _mm256_sub_ps(ix1,jx1);
1010             dy11             = _mm256_sub_ps(iy1,jy1);
1011             dz11             = _mm256_sub_ps(iz1,jz1);
1012             dx12             = _mm256_sub_ps(ix1,jx2);
1013             dy12             = _mm256_sub_ps(iy1,jy2);
1014             dz12             = _mm256_sub_ps(iz1,jz2);
1015             dx13             = _mm256_sub_ps(ix1,jx3);
1016             dy13             = _mm256_sub_ps(iy1,jy3);
1017             dz13             = _mm256_sub_ps(iz1,jz3);
1018             dx21             = _mm256_sub_ps(ix2,jx1);
1019             dy21             = _mm256_sub_ps(iy2,jy1);
1020             dz21             = _mm256_sub_ps(iz2,jz1);
1021             dx22             = _mm256_sub_ps(ix2,jx2);
1022             dy22             = _mm256_sub_ps(iy2,jy2);
1023             dz22             = _mm256_sub_ps(iz2,jz2);
1024             dx23             = _mm256_sub_ps(ix2,jx3);
1025             dy23             = _mm256_sub_ps(iy2,jy3);
1026             dz23             = _mm256_sub_ps(iz2,jz3);
1027             dx31             = _mm256_sub_ps(ix3,jx1);
1028             dy31             = _mm256_sub_ps(iy3,jy1);
1029             dz31             = _mm256_sub_ps(iz3,jz1);
1030             dx32             = _mm256_sub_ps(ix3,jx2);
1031             dy32             = _mm256_sub_ps(iy3,jy2);
1032             dz32             = _mm256_sub_ps(iz3,jz2);
1033             dx33             = _mm256_sub_ps(ix3,jx3);
1034             dy33             = _mm256_sub_ps(iy3,jy3);
1035             dz33             = _mm256_sub_ps(iz3,jz3);
1036
1037             /* Calculate squared distance and things based on it */
1038             rsq00            = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
1039             rsq11            = gmx_mm256_calc_rsq_ps(dx11,dy11,dz11);
1040             rsq12            = gmx_mm256_calc_rsq_ps(dx12,dy12,dz12);
1041             rsq13            = gmx_mm256_calc_rsq_ps(dx13,dy13,dz13);
1042             rsq21            = gmx_mm256_calc_rsq_ps(dx21,dy21,dz21);
1043             rsq22            = gmx_mm256_calc_rsq_ps(dx22,dy22,dz22);
1044             rsq23            = gmx_mm256_calc_rsq_ps(dx23,dy23,dz23);
1045             rsq31            = gmx_mm256_calc_rsq_ps(dx31,dy31,dz31);
1046             rsq32            = gmx_mm256_calc_rsq_ps(dx32,dy32,dz32);
1047             rsq33            = gmx_mm256_calc_rsq_ps(dx33,dy33,dz33);
1048
1049             rinv00           = gmx_mm256_invsqrt_ps(rsq00);
1050             rinv11           = gmx_mm256_invsqrt_ps(rsq11);
1051             rinv12           = gmx_mm256_invsqrt_ps(rsq12);
1052             rinv13           = gmx_mm256_invsqrt_ps(rsq13);
1053             rinv21           = gmx_mm256_invsqrt_ps(rsq21);
1054             rinv22           = gmx_mm256_invsqrt_ps(rsq22);
1055             rinv23           = gmx_mm256_invsqrt_ps(rsq23);
1056             rinv31           = gmx_mm256_invsqrt_ps(rsq31);
1057             rinv32           = gmx_mm256_invsqrt_ps(rsq32);
1058             rinv33           = gmx_mm256_invsqrt_ps(rsq33);
1059
1060             rinvsq00         = _mm256_mul_ps(rinv00,rinv00);
1061             rinvsq11         = _mm256_mul_ps(rinv11,rinv11);
1062             rinvsq12         = _mm256_mul_ps(rinv12,rinv12);
1063             rinvsq13         = _mm256_mul_ps(rinv13,rinv13);
1064             rinvsq21         = _mm256_mul_ps(rinv21,rinv21);
1065             rinvsq22         = _mm256_mul_ps(rinv22,rinv22);
1066             rinvsq23         = _mm256_mul_ps(rinv23,rinv23);
1067             rinvsq31         = _mm256_mul_ps(rinv31,rinv31);
1068             rinvsq32         = _mm256_mul_ps(rinv32,rinv32);
1069             rinvsq33         = _mm256_mul_ps(rinv33,rinv33);
1070
1071             fjx0             = _mm256_setzero_ps();
1072             fjy0             = _mm256_setzero_ps();
1073             fjz0             = _mm256_setzero_ps();
1074             fjx1             = _mm256_setzero_ps();
1075             fjy1             = _mm256_setzero_ps();
1076             fjz1             = _mm256_setzero_ps();
1077             fjx2             = _mm256_setzero_ps();
1078             fjy2             = _mm256_setzero_ps();
1079             fjz2             = _mm256_setzero_ps();
1080             fjx3             = _mm256_setzero_ps();
1081             fjy3             = _mm256_setzero_ps();
1082             fjz3             = _mm256_setzero_ps();
1083
1084             /**************************
1085              * CALCULATE INTERACTIONS *
1086              **************************/
1087
1088             if (gmx_mm256_any_lt(rsq00,rcutoff2))
1089             {
1090
1091             r00              = _mm256_mul_ps(rsq00,rinv00);
1092             r00              = _mm256_andnot_ps(dummy_mask,r00);
1093
1094             /* LENNARD-JONES DISPERSION/REPULSION */
1095
1096             rinvsix          = _mm256_mul_ps(_mm256_mul_ps(rinvsq00,rinvsq00),rinvsq00);
1097             vvdw6            = _mm256_mul_ps(c6_00,rinvsix);
1098             vvdw12           = _mm256_mul_ps(c12_00,_mm256_mul_ps(rinvsix,rinvsix));
1099             vvdw             = _mm256_sub_ps( _mm256_mul_ps(vvdw12,one_twelfth) , _mm256_mul_ps(vvdw6,one_sixth) );
1100             fvdw             = _mm256_mul_ps(_mm256_sub_ps(vvdw12,vvdw6),rinvsq00);
1101
1102             d                = _mm256_sub_ps(r00,rswitch);
1103             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1104             d2               = _mm256_mul_ps(d,d);
1105             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)))))));
1106
1107             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1108
1109             /* Evaluate switch function */
1110             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1111             fvdw             = _mm256_sub_ps( _mm256_mul_ps(fvdw,sw) , _mm256_mul_ps(rinv00,_mm256_mul_ps(vvdw,dsw)) );
1112             vvdw             = _mm256_mul_ps(vvdw,sw);
1113             cutoff_mask      = _mm256_cmp_ps(rsq00,rcutoff2,_CMP_LT_OQ);
1114
1115             /* Update potential sum for this i atom from the interaction with this j atom. */
1116             vvdw             = _mm256_and_ps(vvdw,cutoff_mask);
1117             vvdw             = _mm256_andnot_ps(dummy_mask,vvdw);
1118             vvdwsum          = _mm256_add_ps(vvdwsum,vvdw);
1119
1120             fscal            = fvdw;
1121
1122             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1123
1124             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1125
1126             /* Calculate temporary vectorial force */
1127             tx               = _mm256_mul_ps(fscal,dx00);
1128             ty               = _mm256_mul_ps(fscal,dy00);
1129             tz               = _mm256_mul_ps(fscal,dz00);
1130
1131             /* Update vectorial force */
1132             fix0             = _mm256_add_ps(fix0,tx);
1133             fiy0             = _mm256_add_ps(fiy0,ty);
1134             fiz0             = _mm256_add_ps(fiz0,tz);
1135
1136             fjx0             = _mm256_add_ps(fjx0,tx);
1137             fjy0             = _mm256_add_ps(fjy0,ty);
1138             fjz0             = _mm256_add_ps(fjz0,tz);
1139
1140             }
1141
1142             /**************************
1143              * CALCULATE INTERACTIONS *
1144              **************************/
1145
1146             if (gmx_mm256_any_lt(rsq11,rcutoff2))
1147             {
1148
1149             r11              = _mm256_mul_ps(rsq11,rinv11);
1150             r11              = _mm256_andnot_ps(dummy_mask,r11);
1151
1152             /* EWALD ELECTROSTATICS */
1153
1154             /* Analytical PME correction */
1155             zeta2            = _mm256_mul_ps(beta2,rsq11);
1156             rinv3            = _mm256_mul_ps(rinvsq11,rinv11);
1157             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
1158             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1159             felec            = _mm256_mul_ps(qq11,felec);
1160             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
1161             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1162             velec            = _mm256_sub_ps(rinv11,pmecorrV);
1163             velec            = _mm256_mul_ps(qq11,velec);
1164
1165             d                = _mm256_sub_ps(r11,rswitch);
1166             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1167             d2               = _mm256_mul_ps(d,d);
1168             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)))))));
1169
1170             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1171
1172             /* Evaluate switch function */
1173             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1174             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv11,_mm256_mul_ps(velec,dsw)) );
1175             velec            = _mm256_mul_ps(velec,sw);
1176             cutoff_mask      = _mm256_cmp_ps(rsq11,rcutoff2,_CMP_LT_OQ);
1177
1178             /* Update potential sum for this i atom from the interaction with this j atom. */
1179             velec            = _mm256_and_ps(velec,cutoff_mask);
1180             velec            = _mm256_andnot_ps(dummy_mask,velec);
1181             velecsum         = _mm256_add_ps(velecsum,velec);
1182
1183             fscal            = felec;
1184
1185             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1186
1187             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1188
1189             /* Calculate temporary vectorial force */
1190             tx               = _mm256_mul_ps(fscal,dx11);
1191             ty               = _mm256_mul_ps(fscal,dy11);
1192             tz               = _mm256_mul_ps(fscal,dz11);
1193
1194             /* Update vectorial force */
1195             fix1             = _mm256_add_ps(fix1,tx);
1196             fiy1             = _mm256_add_ps(fiy1,ty);
1197             fiz1             = _mm256_add_ps(fiz1,tz);
1198
1199             fjx1             = _mm256_add_ps(fjx1,tx);
1200             fjy1             = _mm256_add_ps(fjy1,ty);
1201             fjz1             = _mm256_add_ps(fjz1,tz);
1202
1203             }
1204
1205             /**************************
1206              * CALCULATE INTERACTIONS *
1207              **************************/
1208
1209             if (gmx_mm256_any_lt(rsq12,rcutoff2))
1210             {
1211
1212             r12              = _mm256_mul_ps(rsq12,rinv12);
1213             r12              = _mm256_andnot_ps(dummy_mask,r12);
1214
1215             /* EWALD ELECTROSTATICS */
1216
1217             /* Analytical PME correction */
1218             zeta2            = _mm256_mul_ps(beta2,rsq12);
1219             rinv3            = _mm256_mul_ps(rinvsq12,rinv12);
1220             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
1221             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1222             felec            = _mm256_mul_ps(qq12,felec);
1223             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
1224             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1225             velec            = _mm256_sub_ps(rinv12,pmecorrV);
1226             velec            = _mm256_mul_ps(qq12,velec);
1227
1228             d                = _mm256_sub_ps(r12,rswitch);
1229             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1230             d2               = _mm256_mul_ps(d,d);
1231             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)))))));
1232
1233             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1234
1235             /* Evaluate switch function */
1236             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1237             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv12,_mm256_mul_ps(velec,dsw)) );
1238             velec            = _mm256_mul_ps(velec,sw);
1239             cutoff_mask      = _mm256_cmp_ps(rsq12,rcutoff2,_CMP_LT_OQ);
1240
1241             /* Update potential sum for this i atom from the interaction with this j atom. */
1242             velec            = _mm256_and_ps(velec,cutoff_mask);
1243             velec            = _mm256_andnot_ps(dummy_mask,velec);
1244             velecsum         = _mm256_add_ps(velecsum,velec);
1245
1246             fscal            = felec;
1247
1248             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1249
1250             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1251
1252             /* Calculate temporary vectorial force */
1253             tx               = _mm256_mul_ps(fscal,dx12);
1254             ty               = _mm256_mul_ps(fscal,dy12);
1255             tz               = _mm256_mul_ps(fscal,dz12);
1256
1257             /* Update vectorial force */
1258             fix1             = _mm256_add_ps(fix1,tx);
1259             fiy1             = _mm256_add_ps(fiy1,ty);
1260             fiz1             = _mm256_add_ps(fiz1,tz);
1261
1262             fjx2             = _mm256_add_ps(fjx2,tx);
1263             fjy2             = _mm256_add_ps(fjy2,ty);
1264             fjz2             = _mm256_add_ps(fjz2,tz);
1265
1266             }
1267
1268             /**************************
1269              * CALCULATE INTERACTIONS *
1270              **************************/
1271
1272             if (gmx_mm256_any_lt(rsq13,rcutoff2))
1273             {
1274
1275             r13              = _mm256_mul_ps(rsq13,rinv13);
1276             r13              = _mm256_andnot_ps(dummy_mask,r13);
1277
1278             /* EWALD ELECTROSTATICS */
1279
1280             /* Analytical PME correction */
1281             zeta2            = _mm256_mul_ps(beta2,rsq13);
1282             rinv3            = _mm256_mul_ps(rinvsq13,rinv13);
1283             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
1284             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1285             felec            = _mm256_mul_ps(qq13,felec);
1286             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
1287             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1288             velec            = _mm256_sub_ps(rinv13,pmecorrV);
1289             velec            = _mm256_mul_ps(qq13,velec);
1290
1291             d                = _mm256_sub_ps(r13,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(rinv13,_mm256_mul_ps(velec,dsw)) );
1301             velec            = _mm256_mul_ps(velec,sw);
1302             cutoff_mask      = _mm256_cmp_ps(rsq13,rcutoff2,_CMP_LT_OQ);
1303
1304             /* Update potential sum for this i atom from the interaction with this j atom. */
1305             velec            = _mm256_and_ps(velec,cutoff_mask);
1306             velec            = _mm256_andnot_ps(dummy_mask,velec);
1307             velecsum         = _mm256_add_ps(velecsum,velec);
1308
1309             fscal            = felec;
1310
1311             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1312
1313             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1314
1315             /* Calculate temporary vectorial force */
1316             tx               = _mm256_mul_ps(fscal,dx13);
1317             ty               = _mm256_mul_ps(fscal,dy13);
1318             tz               = _mm256_mul_ps(fscal,dz13);
1319
1320             /* Update vectorial force */
1321             fix1             = _mm256_add_ps(fix1,tx);
1322             fiy1             = _mm256_add_ps(fiy1,ty);
1323             fiz1             = _mm256_add_ps(fiz1,tz);
1324
1325             fjx3             = _mm256_add_ps(fjx3,tx);
1326             fjy3             = _mm256_add_ps(fjy3,ty);
1327             fjz3             = _mm256_add_ps(fjz3,tz);
1328
1329             }
1330
1331             /**************************
1332              * CALCULATE INTERACTIONS *
1333              **************************/
1334
1335             if (gmx_mm256_any_lt(rsq21,rcutoff2))
1336             {
1337
1338             r21              = _mm256_mul_ps(rsq21,rinv21);
1339             r21              = _mm256_andnot_ps(dummy_mask,r21);
1340
1341             /* EWALD ELECTROSTATICS */
1342
1343             /* Analytical PME correction */
1344             zeta2            = _mm256_mul_ps(beta2,rsq21);
1345             rinv3            = _mm256_mul_ps(rinvsq21,rinv21);
1346             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
1347             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1348             felec            = _mm256_mul_ps(qq21,felec);
1349             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
1350             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1351             velec            = _mm256_sub_ps(rinv21,pmecorrV);
1352             velec            = _mm256_mul_ps(qq21,velec);
1353
1354             d                = _mm256_sub_ps(r21,rswitch);
1355             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1356             d2               = _mm256_mul_ps(d,d);
1357             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)))))));
1358
1359             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1360
1361             /* Evaluate switch function */
1362             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1363             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv21,_mm256_mul_ps(velec,dsw)) );
1364             velec            = _mm256_mul_ps(velec,sw);
1365             cutoff_mask      = _mm256_cmp_ps(rsq21,rcutoff2,_CMP_LT_OQ);
1366
1367             /* Update potential sum for this i atom from the interaction with this j atom. */
1368             velec            = _mm256_and_ps(velec,cutoff_mask);
1369             velec            = _mm256_andnot_ps(dummy_mask,velec);
1370             velecsum         = _mm256_add_ps(velecsum,velec);
1371
1372             fscal            = felec;
1373
1374             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1375
1376             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1377
1378             /* Calculate temporary vectorial force */
1379             tx               = _mm256_mul_ps(fscal,dx21);
1380             ty               = _mm256_mul_ps(fscal,dy21);
1381             tz               = _mm256_mul_ps(fscal,dz21);
1382
1383             /* Update vectorial force */
1384             fix2             = _mm256_add_ps(fix2,tx);
1385             fiy2             = _mm256_add_ps(fiy2,ty);
1386             fiz2             = _mm256_add_ps(fiz2,tz);
1387
1388             fjx1             = _mm256_add_ps(fjx1,tx);
1389             fjy1             = _mm256_add_ps(fjy1,ty);
1390             fjz1             = _mm256_add_ps(fjz1,tz);
1391
1392             }
1393
1394             /**************************
1395              * CALCULATE INTERACTIONS *
1396              **************************/
1397
1398             if (gmx_mm256_any_lt(rsq22,rcutoff2))
1399             {
1400
1401             r22              = _mm256_mul_ps(rsq22,rinv22);
1402             r22              = _mm256_andnot_ps(dummy_mask,r22);
1403
1404             /* EWALD ELECTROSTATICS */
1405
1406             /* Analytical PME correction */
1407             zeta2            = _mm256_mul_ps(beta2,rsq22);
1408             rinv3            = _mm256_mul_ps(rinvsq22,rinv22);
1409             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
1410             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1411             felec            = _mm256_mul_ps(qq22,felec);
1412             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
1413             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1414             velec            = _mm256_sub_ps(rinv22,pmecorrV);
1415             velec            = _mm256_mul_ps(qq22,velec);
1416
1417             d                = _mm256_sub_ps(r22,rswitch);
1418             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1419             d2               = _mm256_mul_ps(d,d);
1420             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)))))));
1421
1422             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1423
1424             /* Evaluate switch function */
1425             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1426             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv22,_mm256_mul_ps(velec,dsw)) );
1427             velec            = _mm256_mul_ps(velec,sw);
1428             cutoff_mask      = _mm256_cmp_ps(rsq22,rcutoff2,_CMP_LT_OQ);
1429
1430             /* Update potential sum for this i atom from the interaction with this j atom. */
1431             velec            = _mm256_and_ps(velec,cutoff_mask);
1432             velec            = _mm256_andnot_ps(dummy_mask,velec);
1433             velecsum         = _mm256_add_ps(velecsum,velec);
1434
1435             fscal            = felec;
1436
1437             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1438
1439             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1440
1441             /* Calculate temporary vectorial force */
1442             tx               = _mm256_mul_ps(fscal,dx22);
1443             ty               = _mm256_mul_ps(fscal,dy22);
1444             tz               = _mm256_mul_ps(fscal,dz22);
1445
1446             /* Update vectorial force */
1447             fix2             = _mm256_add_ps(fix2,tx);
1448             fiy2             = _mm256_add_ps(fiy2,ty);
1449             fiz2             = _mm256_add_ps(fiz2,tz);
1450
1451             fjx2             = _mm256_add_ps(fjx2,tx);
1452             fjy2             = _mm256_add_ps(fjy2,ty);
1453             fjz2             = _mm256_add_ps(fjz2,tz);
1454
1455             }
1456
1457             /**************************
1458              * CALCULATE INTERACTIONS *
1459              **************************/
1460
1461             if (gmx_mm256_any_lt(rsq23,rcutoff2))
1462             {
1463
1464             r23              = _mm256_mul_ps(rsq23,rinv23);
1465             r23              = _mm256_andnot_ps(dummy_mask,r23);
1466
1467             /* EWALD ELECTROSTATICS */
1468
1469             /* Analytical PME correction */
1470             zeta2            = _mm256_mul_ps(beta2,rsq23);
1471             rinv3            = _mm256_mul_ps(rinvsq23,rinv23);
1472             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
1473             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1474             felec            = _mm256_mul_ps(qq23,felec);
1475             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
1476             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1477             velec            = _mm256_sub_ps(rinv23,pmecorrV);
1478             velec            = _mm256_mul_ps(qq23,velec);
1479
1480             d                = _mm256_sub_ps(r23,rswitch);
1481             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1482             d2               = _mm256_mul_ps(d,d);
1483             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)))))));
1484
1485             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1486
1487             /* Evaluate switch function */
1488             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1489             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv23,_mm256_mul_ps(velec,dsw)) );
1490             velec            = _mm256_mul_ps(velec,sw);
1491             cutoff_mask      = _mm256_cmp_ps(rsq23,rcutoff2,_CMP_LT_OQ);
1492
1493             /* Update potential sum for this i atom from the interaction with this j atom. */
1494             velec            = _mm256_and_ps(velec,cutoff_mask);
1495             velec            = _mm256_andnot_ps(dummy_mask,velec);
1496             velecsum         = _mm256_add_ps(velecsum,velec);
1497
1498             fscal            = felec;
1499
1500             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1501
1502             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1503
1504             /* Calculate temporary vectorial force */
1505             tx               = _mm256_mul_ps(fscal,dx23);
1506             ty               = _mm256_mul_ps(fscal,dy23);
1507             tz               = _mm256_mul_ps(fscal,dz23);
1508
1509             /* Update vectorial force */
1510             fix2             = _mm256_add_ps(fix2,tx);
1511             fiy2             = _mm256_add_ps(fiy2,ty);
1512             fiz2             = _mm256_add_ps(fiz2,tz);
1513
1514             fjx3             = _mm256_add_ps(fjx3,tx);
1515             fjy3             = _mm256_add_ps(fjy3,ty);
1516             fjz3             = _mm256_add_ps(fjz3,tz);
1517
1518             }
1519
1520             /**************************
1521              * CALCULATE INTERACTIONS *
1522              **************************/
1523
1524             if (gmx_mm256_any_lt(rsq31,rcutoff2))
1525             {
1526
1527             r31              = _mm256_mul_ps(rsq31,rinv31);
1528             r31              = _mm256_andnot_ps(dummy_mask,r31);
1529
1530             /* EWALD ELECTROSTATICS */
1531
1532             /* Analytical PME correction */
1533             zeta2            = _mm256_mul_ps(beta2,rsq31);
1534             rinv3            = _mm256_mul_ps(rinvsq31,rinv31);
1535             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
1536             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1537             felec            = _mm256_mul_ps(qq31,felec);
1538             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
1539             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1540             velec            = _mm256_sub_ps(rinv31,pmecorrV);
1541             velec            = _mm256_mul_ps(qq31,velec);
1542
1543             d                = _mm256_sub_ps(r31,rswitch);
1544             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1545             d2               = _mm256_mul_ps(d,d);
1546             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)))))));
1547
1548             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1549
1550             /* Evaluate switch function */
1551             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1552             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv31,_mm256_mul_ps(velec,dsw)) );
1553             velec            = _mm256_mul_ps(velec,sw);
1554             cutoff_mask      = _mm256_cmp_ps(rsq31,rcutoff2,_CMP_LT_OQ);
1555
1556             /* Update potential sum for this i atom from the interaction with this j atom. */
1557             velec            = _mm256_and_ps(velec,cutoff_mask);
1558             velec            = _mm256_andnot_ps(dummy_mask,velec);
1559             velecsum         = _mm256_add_ps(velecsum,velec);
1560
1561             fscal            = felec;
1562
1563             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1564
1565             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1566
1567             /* Calculate temporary vectorial force */
1568             tx               = _mm256_mul_ps(fscal,dx31);
1569             ty               = _mm256_mul_ps(fscal,dy31);
1570             tz               = _mm256_mul_ps(fscal,dz31);
1571
1572             /* Update vectorial force */
1573             fix3             = _mm256_add_ps(fix3,tx);
1574             fiy3             = _mm256_add_ps(fiy3,ty);
1575             fiz3             = _mm256_add_ps(fiz3,tz);
1576
1577             fjx1             = _mm256_add_ps(fjx1,tx);
1578             fjy1             = _mm256_add_ps(fjy1,ty);
1579             fjz1             = _mm256_add_ps(fjz1,tz);
1580
1581             }
1582
1583             /**************************
1584              * CALCULATE INTERACTIONS *
1585              **************************/
1586
1587             if (gmx_mm256_any_lt(rsq32,rcutoff2))
1588             {
1589
1590             r32              = _mm256_mul_ps(rsq32,rinv32);
1591             r32              = _mm256_andnot_ps(dummy_mask,r32);
1592
1593             /* EWALD ELECTROSTATICS */
1594
1595             /* Analytical PME correction */
1596             zeta2            = _mm256_mul_ps(beta2,rsq32);
1597             rinv3            = _mm256_mul_ps(rinvsq32,rinv32);
1598             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
1599             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1600             felec            = _mm256_mul_ps(qq32,felec);
1601             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
1602             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1603             velec            = _mm256_sub_ps(rinv32,pmecorrV);
1604             velec            = _mm256_mul_ps(qq32,velec);
1605
1606             d                = _mm256_sub_ps(r32,rswitch);
1607             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1608             d2               = _mm256_mul_ps(d,d);
1609             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)))))));
1610
1611             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1612
1613             /* Evaluate switch function */
1614             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1615             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv32,_mm256_mul_ps(velec,dsw)) );
1616             velec            = _mm256_mul_ps(velec,sw);
1617             cutoff_mask      = _mm256_cmp_ps(rsq32,rcutoff2,_CMP_LT_OQ);
1618
1619             /* Update potential sum for this i atom from the interaction with this j atom. */
1620             velec            = _mm256_and_ps(velec,cutoff_mask);
1621             velec            = _mm256_andnot_ps(dummy_mask,velec);
1622             velecsum         = _mm256_add_ps(velecsum,velec);
1623
1624             fscal            = felec;
1625
1626             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1627
1628             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1629
1630             /* Calculate temporary vectorial force */
1631             tx               = _mm256_mul_ps(fscal,dx32);
1632             ty               = _mm256_mul_ps(fscal,dy32);
1633             tz               = _mm256_mul_ps(fscal,dz32);
1634
1635             /* Update vectorial force */
1636             fix3             = _mm256_add_ps(fix3,tx);
1637             fiy3             = _mm256_add_ps(fiy3,ty);
1638             fiz3             = _mm256_add_ps(fiz3,tz);
1639
1640             fjx2             = _mm256_add_ps(fjx2,tx);
1641             fjy2             = _mm256_add_ps(fjy2,ty);
1642             fjz2             = _mm256_add_ps(fjz2,tz);
1643
1644             }
1645
1646             /**************************
1647              * CALCULATE INTERACTIONS *
1648              **************************/
1649
1650             if (gmx_mm256_any_lt(rsq33,rcutoff2))
1651             {
1652
1653             r33              = _mm256_mul_ps(rsq33,rinv33);
1654             r33              = _mm256_andnot_ps(dummy_mask,r33);
1655
1656             /* EWALD ELECTROSTATICS */
1657
1658             /* Analytical PME correction */
1659             zeta2            = _mm256_mul_ps(beta2,rsq33);
1660             rinv3            = _mm256_mul_ps(rinvsq33,rinv33);
1661             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
1662             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1663             felec            = _mm256_mul_ps(qq33,felec);
1664             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
1665             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1666             velec            = _mm256_sub_ps(rinv33,pmecorrV);
1667             velec            = _mm256_mul_ps(qq33,velec);
1668
1669             d                = _mm256_sub_ps(r33,rswitch);
1670             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1671             d2               = _mm256_mul_ps(d,d);
1672             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)))))));
1673
1674             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1675
1676             /* Evaluate switch function */
1677             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1678             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv33,_mm256_mul_ps(velec,dsw)) );
1679             velec            = _mm256_mul_ps(velec,sw);
1680             cutoff_mask      = _mm256_cmp_ps(rsq33,rcutoff2,_CMP_LT_OQ);
1681
1682             /* Update potential sum for this i atom from the interaction with this j atom. */
1683             velec            = _mm256_and_ps(velec,cutoff_mask);
1684             velec            = _mm256_andnot_ps(dummy_mask,velec);
1685             velecsum         = _mm256_add_ps(velecsum,velec);
1686
1687             fscal            = felec;
1688
1689             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1690
1691             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1692
1693             /* Calculate temporary vectorial force */
1694             tx               = _mm256_mul_ps(fscal,dx33);
1695             ty               = _mm256_mul_ps(fscal,dy33);
1696             tz               = _mm256_mul_ps(fscal,dz33);
1697
1698             /* Update vectorial force */
1699             fix3             = _mm256_add_ps(fix3,tx);
1700             fiy3             = _mm256_add_ps(fiy3,ty);
1701             fiz3             = _mm256_add_ps(fiz3,tz);
1702
1703             fjx3             = _mm256_add_ps(fjx3,tx);
1704             fjy3             = _mm256_add_ps(fjy3,ty);
1705             fjz3             = _mm256_add_ps(fjz3,tz);
1706
1707             }
1708
1709             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
1710             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
1711             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
1712             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
1713             fjptrE             = (jnrlistE>=0) ? f+j_coord_offsetE : scratch;
1714             fjptrF             = (jnrlistF>=0) ? f+j_coord_offsetF : scratch;
1715             fjptrG             = (jnrlistG>=0) ? f+j_coord_offsetG : scratch;
1716             fjptrH             = (jnrlistH>=0) ? f+j_coord_offsetH : scratch;
1717
1718             gmx_mm256_decrement_4rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,
1719                                                       fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,
1720                                                       fjx2,fjy2,fjz2,fjx3,fjy3,fjz3);
1721
1722             /* Inner loop uses 1044 flops */
1723         }
1724
1725         /* End of innermost loop */
1726
1727         gmx_mm256_update_iforce_4atom_swizzle_ps(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
1728                                                  f+i_coord_offset,fshift+i_shift_offset);
1729
1730         ggid                        = gid[iidx];
1731         /* Update potential energies */
1732         gmx_mm256_update_1pot_ps(velecsum,kernel_data->energygrp_elec+ggid);
1733         gmx_mm256_update_1pot_ps(vvdwsum,kernel_data->energygrp_vdw+ggid);
1734
1735         /* Increment number of inner iterations */
1736         inneriter                  += j_index_end - j_index_start;
1737
1738         /* Outer loop uses 26 flops */
1739     }
1740
1741     /* Increment number of outer iterations */
1742     outeriter        += nri;
1743
1744     /* Update outer/inner flops */
1745
1746     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4W4_VF,outeriter*26 + inneriter*1044);
1747 }
1748 /*
1749  * Gromacs nonbonded kernel:   nb_kernel_ElecEwSw_VdwLJSw_GeomW4W4_F_avx_256_single
1750  * Electrostatics interaction: Ewald
1751  * VdW interaction:            LennardJones
1752  * Geometry:                   Water4-Water4
1753  * Calculate force/pot:        Force
1754  */
1755 void
1756 nb_kernel_ElecEwSw_VdwLJSw_GeomW4W4_F_avx_256_single
1757                     (t_nblist                    * gmx_restrict       nlist,
1758                      rvec                        * gmx_restrict          xx,
1759                      rvec                        * gmx_restrict          ff,
1760                      t_forcerec                  * gmx_restrict          fr,
1761                      t_mdatoms                   * gmx_restrict     mdatoms,
1762                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
1763                      t_nrnb                      * gmx_restrict        nrnb)
1764 {
1765     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
1766      * just 0 for non-waters.
1767      * Suffixes A,B,C,D,E,F,G,H refer to j loop unrolling done with AVX, e.g. for the eight different
1768      * jnr indices corresponding to data put in the four positions in the SIMD register.
1769      */
1770     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
1771     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
1772     int              jnrA,jnrB,jnrC,jnrD;
1773     int              jnrE,jnrF,jnrG,jnrH;
1774     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
1775     int              jnrlistE,jnrlistF,jnrlistG,jnrlistH;
1776     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
1777     int              j_coord_offsetE,j_coord_offsetF,j_coord_offsetG,j_coord_offsetH;
1778     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
1779     real             rcutoff_scalar;
1780     real             *shiftvec,*fshift,*x,*f;
1781     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD,*fjptrE,*fjptrF,*fjptrG,*fjptrH;
1782     real             scratch[4*DIM];
1783     __m256           tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
1784     real *           vdwioffsetptr0;
1785     __m256           ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
1786     real *           vdwioffsetptr1;
1787     __m256           ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
1788     real *           vdwioffsetptr2;
1789     __m256           ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
1790     real *           vdwioffsetptr3;
1791     __m256           ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
1792     int              vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D,vdwjidx0E,vdwjidx0F,vdwjidx0G,vdwjidx0H;
1793     __m256           jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
1794     int              vdwjidx1A,vdwjidx1B,vdwjidx1C,vdwjidx1D,vdwjidx1E,vdwjidx1F,vdwjidx1G,vdwjidx1H;
1795     __m256           jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
1796     int              vdwjidx2A,vdwjidx2B,vdwjidx2C,vdwjidx2D,vdwjidx2E,vdwjidx2F,vdwjidx2G,vdwjidx2H;
1797     __m256           jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
1798     int              vdwjidx3A,vdwjidx3B,vdwjidx3C,vdwjidx3D,vdwjidx3E,vdwjidx3F,vdwjidx3G,vdwjidx3H;
1799     __m256           jx3,jy3,jz3,fjx3,fjy3,fjz3,jq3,isaj3;
1800     __m256           dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
1801     __m256           dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11;
1802     __m256           dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12;
1803     __m256           dx13,dy13,dz13,rsq13,rinv13,rinvsq13,r13,qq13,c6_13,c12_13;
1804     __m256           dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21;
1805     __m256           dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22;
1806     __m256           dx23,dy23,dz23,rsq23,rinv23,rinvsq23,r23,qq23,c6_23,c12_23;
1807     __m256           dx31,dy31,dz31,rsq31,rinv31,rinvsq31,r31,qq31,c6_31,c12_31;
1808     __m256           dx32,dy32,dz32,rsq32,rinv32,rinvsq32,r32,qq32,c6_32,c12_32;
1809     __m256           dx33,dy33,dz33,rsq33,rinv33,rinvsq33,r33,qq33,c6_33,c12_33;
1810     __m256           velec,felec,velecsum,facel,crf,krf,krf2;
1811     real             *charge;
1812     int              nvdwtype;
1813     __m256           rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
1814     int              *vdwtype;
1815     real             *vdwparam;
1816     __m256           one_sixth   = _mm256_set1_ps(1.0/6.0);
1817     __m256           one_twelfth = _mm256_set1_ps(1.0/12.0);
1818     __m256i          ewitab;
1819     __m128i          ewitab_lo,ewitab_hi;
1820     __m256           ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
1821     __m256           beta,beta2,beta3,zeta2,pmecorrF,pmecorrV,rinv3;
1822     real             *ewtab;
1823     __m256           rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
1824     real             rswitch_scalar,d_scalar;
1825     __m256           dummy_mask,cutoff_mask;
1826     __m256           signbit = _mm256_castsi256_ps( _mm256_set1_epi32(0x80000000) );
1827     __m256           one     = _mm256_set1_ps(1.0);
1828     __m256           two     = _mm256_set1_ps(2.0);
1829     x                = xx[0];
1830     f                = ff[0];
1831
1832     nri              = nlist->nri;
1833     iinr             = nlist->iinr;
1834     jindex           = nlist->jindex;
1835     jjnr             = nlist->jjnr;
1836     shiftidx         = nlist->shift;
1837     gid              = nlist->gid;
1838     shiftvec         = fr->shift_vec[0];
1839     fshift           = fr->fshift[0];
1840     facel            = _mm256_set1_ps(fr->epsfac);
1841     charge           = mdatoms->chargeA;
1842     nvdwtype         = fr->ntype;
1843     vdwparam         = fr->nbfp;
1844     vdwtype          = mdatoms->typeA;
1845
1846     sh_ewald         = _mm256_set1_ps(fr->ic->sh_ewald);
1847     beta             = _mm256_set1_ps(fr->ic->ewaldcoeff_q);
1848     beta2            = _mm256_mul_ps(beta,beta);
1849     beta3            = _mm256_mul_ps(beta,beta2);
1850
1851     ewtab            = fr->ic->tabq_coul_FDV0;
1852     ewtabscale       = _mm256_set1_ps(fr->ic->tabq_scale);
1853     ewtabhalfspace   = _mm256_set1_ps(0.5/fr->ic->tabq_scale);
1854
1855     /* Setup water-specific parameters */
1856     inr              = nlist->iinr[0];
1857     iq1              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+1]));
1858     iq2              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+2]));
1859     iq3              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+3]));
1860     vdwioffsetptr0   = vdwparam+2*nvdwtype*vdwtype[inr+0];
1861
1862     jq1              = _mm256_set1_ps(charge[inr+1]);
1863     jq2              = _mm256_set1_ps(charge[inr+2]);
1864     jq3              = _mm256_set1_ps(charge[inr+3]);
1865     vdwjidx0A        = 2*vdwtype[inr+0];
1866     c6_00            = _mm256_set1_ps(vdwioffsetptr0[vdwjidx0A]);
1867     c12_00           = _mm256_set1_ps(vdwioffsetptr0[vdwjidx0A+1]);
1868     qq11             = _mm256_mul_ps(iq1,jq1);
1869     qq12             = _mm256_mul_ps(iq1,jq2);
1870     qq13             = _mm256_mul_ps(iq1,jq3);
1871     qq21             = _mm256_mul_ps(iq2,jq1);
1872     qq22             = _mm256_mul_ps(iq2,jq2);
1873     qq23             = _mm256_mul_ps(iq2,jq3);
1874     qq31             = _mm256_mul_ps(iq3,jq1);
1875     qq32             = _mm256_mul_ps(iq3,jq2);
1876     qq33             = _mm256_mul_ps(iq3,jq3);
1877
1878     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
1879     rcutoff_scalar   = fr->rcoulomb;
1880     rcutoff          = _mm256_set1_ps(rcutoff_scalar);
1881     rcutoff2         = _mm256_mul_ps(rcutoff,rcutoff);
1882
1883     rswitch_scalar   = fr->rcoulomb_switch;
1884     rswitch          = _mm256_set1_ps(rswitch_scalar);
1885     /* Setup switch parameters */
1886     d_scalar         = rcutoff_scalar-rswitch_scalar;
1887     d                = _mm256_set1_ps(d_scalar);
1888     swV3             = _mm256_set1_ps(-10.0/(d_scalar*d_scalar*d_scalar));
1889     swV4             = _mm256_set1_ps( 15.0/(d_scalar*d_scalar*d_scalar*d_scalar));
1890     swV5             = _mm256_set1_ps( -6.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
1891     swF2             = _mm256_set1_ps(-30.0/(d_scalar*d_scalar*d_scalar));
1892     swF3             = _mm256_set1_ps( 60.0/(d_scalar*d_scalar*d_scalar*d_scalar));
1893     swF4             = _mm256_set1_ps(-30.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
1894
1895     /* Avoid stupid compiler warnings */
1896     jnrA = jnrB = jnrC = jnrD = jnrE = jnrF = jnrG = jnrH = 0;
1897     j_coord_offsetA = 0;
1898     j_coord_offsetB = 0;
1899     j_coord_offsetC = 0;
1900     j_coord_offsetD = 0;
1901     j_coord_offsetE = 0;
1902     j_coord_offsetF = 0;
1903     j_coord_offsetG = 0;
1904     j_coord_offsetH = 0;
1905
1906     outeriter        = 0;
1907     inneriter        = 0;
1908
1909     for(iidx=0;iidx<4*DIM;iidx++)
1910     {
1911         scratch[iidx] = 0.0;
1912     }
1913
1914     /* Start outer loop over neighborlists */
1915     for(iidx=0; iidx<nri; iidx++)
1916     {
1917         /* Load shift vector for this list */
1918         i_shift_offset   = DIM*shiftidx[iidx];
1919
1920         /* Load limits for loop over neighbors */
1921         j_index_start    = jindex[iidx];
1922         j_index_end      = jindex[iidx+1];
1923
1924         /* Get outer coordinate index */
1925         inr              = iinr[iidx];
1926         i_coord_offset   = DIM*inr;
1927
1928         /* Load i particle coords and add shift vector */
1929         gmx_mm256_load_shift_and_4rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,
1930                                                     &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
1931
1932         fix0             = _mm256_setzero_ps();
1933         fiy0             = _mm256_setzero_ps();
1934         fiz0             = _mm256_setzero_ps();
1935         fix1             = _mm256_setzero_ps();
1936         fiy1             = _mm256_setzero_ps();
1937         fiz1             = _mm256_setzero_ps();
1938         fix2             = _mm256_setzero_ps();
1939         fiy2             = _mm256_setzero_ps();
1940         fiz2             = _mm256_setzero_ps();
1941         fix3             = _mm256_setzero_ps();
1942         fiy3             = _mm256_setzero_ps();
1943         fiz3             = _mm256_setzero_ps();
1944
1945         /* Start inner kernel loop */
1946         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+7]>=0; jidx+=8)
1947         {
1948
1949             /* Get j neighbor index, and coordinate index */
1950             jnrA             = jjnr[jidx];
1951             jnrB             = jjnr[jidx+1];
1952             jnrC             = jjnr[jidx+2];
1953             jnrD             = jjnr[jidx+3];
1954             jnrE             = jjnr[jidx+4];
1955             jnrF             = jjnr[jidx+5];
1956             jnrG             = jjnr[jidx+6];
1957             jnrH             = jjnr[jidx+7];
1958             j_coord_offsetA  = DIM*jnrA;
1959             j_coord_offsetB  = DIM*jnrB;
1960             j_coord_offsetC  = DIM*jnrC;
1961             j_coord_offsetD  = DIM*jnrD;
1962             j_coord_offsetE  = DIM*jnrE;
1963             j_coord_offsetF  = DIM*jnrF;
1964             j_coord_offsetG  = DIM*jnrG;
1965             j_coord_offsetH  = DIM*jnrH;
1966
1967             /* load j atom coordinates */
1968             gmx_mm256_load_4rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
1969                                                  x+j_coord_offsetC,x+j_coord_offsetD,
1970                                                  x+j_coord_offsetE,x+j_coord_offsetF,
1971                                                  x+j_coord_offsetG,x+j_coord_offsetH,
1972                                                  &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,
1973                                                  &jy2,&jz2,&jx3,&jy3,&jz3);
1974
1975             /* Calculate displacement vector */
1976             dx00             = _mm256_sub_ps(ix0,jx0);
1977             dy00             = _mm256_sub_ps(iy0,jy0);
1978             dz00             = _mm256_sub_ps(iz0,jz0);
1979             dx11             = _mm256_sub_ps(ix1,jx1);
1980             dy11             = _mm256_sub_ps(iy1,jy1);
1981             dz11             = _mm256_sub_ps(iz1,jz1);
1982             dx12             = _mm256_sub_ps(ix1,jx2);
1983             dy12             = _mm256_sub_ps(iy1,jy2);
1984             dz12             = _mm256_sub_ps(iz1,jz2);
1985             dx13             = _mm256_sub_ps(ix1,jx3);
1986             dy13             = _mm256_sub_ps(iy1,jy3);
1987             dz13             = _mm256_sub_ps(iz1,jz3);
1988             dx21             = _mm256_sub_ps(ix2,jx1);
1989             dy21             = _mm256_sub_ps(iy2,jy1);
1990             dz21             = _mm256_sub_ps(iz2,jz1);
1991             dx22             = _mm256_sub_ps(ix2,jx2);
1992             dy22             = _mm256_sub_ps(iy2,jy2);
1993             dz22             = _mm256_sub_ps(iz2,jz2);
1994             dx23             = _mm256_sub_ps(ix2,jx3);
1995             dy23             = _mm256_sub_ps(iy2,jy3);
1996             dz23             = _mm256_sub_ps(iz2,jz3);
1997             dx31             = _mm256_sub_ps(ix3,jx1);
1998             dy31             = _mm256_sub_ps(iy3,jy1);
1999             dz31             = _mm256_sub_ps(iz3,jz1);
2000             dx32             = _mm256_sub_ps(ix3,jx2);
2001             dy32             = _mm256_sub_ps(iy3,jy2);
2002             dz32             = _mm256_sub_ps(iz3,jz2);
2003             dx33             = _mm256_sub_ps(ix3,jx3);
2004             dy33             = _mm256_sub_ps(iy3,jy3);
2005             dz33             = _mm256_sub_ps(iz3,jz3);
2006
2007             /* Calculate squared distance and things based on it */
2008             rsq00            = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
2009             rsq11            = gmx_mm256_calc_rsq_ps(dx11,dy11,dz11);
2010             rsq12            = gmx_mm256_calc_rsq_ps(dx12,dy12,dz12);
2011             rsq13            = gmx_mm256_calc_rsq_ps(dx13,dy13,dz13);
2012             rsq21            = gmx_mm256_calc_rsq_ps(dx21,dy21,dz21);
2013             rsq22            = gmx_mm256_calc_rsq_ps(dx22,dy22,dz22);
2014             rsq23            = gmx_mm256_calc_rsq_ps(dx23,dy23,dz23);
2015             rsq31            = gmx_mm256_calc_rsq_ps(dx31,dy31,dz31);
2016             rsq32            = gmx_mm256_calc_rsq_ps(dx32,dy32,dz32);
2017             rsq33            = gmx_mm256_calc_rsq_ps(dx33,dy33,dz33);
2018
2019             rinv00           = gmx_mm256_invsqrt_ps(rsq00);
2020             rinv11           = gmx_mm256_invsqrt_ps(rsq11);
2021             rinv12           = gmx_mm256_invsqrt_ps(rsq12);
2022             rinv13           = gmx_mm256_invsqrt_ps(rsq13);
2023             rinv21           = gmx_mm256_invsqrt_ps(rsq21);
2024             rinv22           = gmx_mm256_invsqrt_ps(rsq22);
2025             rinv23           = gmx_mm256_invsqrt_ps(rsq23);
2026             rinv31           = gmx_mm256_invsqrt_ps(rsq31);
2027             rinv32           = gmx_mm256_invsqrt_ps(rsq32);
2028             rinv33           = gmx_mm256_invsqrt_ps(rsq33);
2029
2030             rinvsq00         = _mm256_mul_ps(rinv00,rinv00);
2031             rinvsq11         = _mm256_mul_ps(rinv11,rinv11);
2032             rinvsq12         = _mm256_mul_ps(rinv12,rinv12);
2033             rinvsq13         = _mm256_mul_ps(rinv13,rinv13);
2034             rinvsq21         = _mm256_mul_ps(rinv21,rinv21);
2035             rinvsq22         = _mm256_mul_ps(rinv22,rinv22);
2036             rinvsq23         = _mm256_mul_ps(rinv23,rinv23);
2037             rinvsq31         = _mm256_mul_ps(rinv31,rinv31);
2038             rinvsq32         = _mm256_mul_ps(rinv32,rinv32);
2039             rinvsq33         = _mm256_mul_ps(rinv33,rinv33);
2040
2041             fjx0             = _mm256_setzero_ps();
2042             fjy0             = _mm256_setzero_ps();
2043             fjz0             = _mm256_setzero_ps();
2044             fjx1             = _mm256_setzero_ps();
2045             fjy1             = _mm256_setzero_ps();
2046             fjz1             = _mm256_setzero_ps();
2047             fjx2             = _mm256_setzero_ps();
2048             fjy2             = _mm256_setzero_ps();
2049             fjz2             = _mm256_setzero_ps();
2050             fjx3             = _mm256_setzero_ps();
2051             fjy3             = _mm256_setzero_ps();
2052             fjz3             = _mm256_setzero_ps();
2053
2054             /**************************
2055              * CALCULATE INTERACTIONS *
2056              **************************/
2057
2058             if (gmx_mm256_any_lt(rsq00,rcutoff2))
2059             {
2060
2061             r00              = _mm256_mul_ps(rsq00,rinv00);
2062
2063             /* LENNARD-JONES DISPERSION/REPULSION */
2064
2065             rinvsix          = _mm256_mul_ps(_mm256_mul_ps(rinvsq00,rinvsq00),rinvsq00);
2066             vvdw6            = _mm256_mul_ps(c6_00,rinvsix);
2067             vvdw12           = _mm256_mul_ps(c12_00,_mm256_mul_ps(rinvsix,rinvsix));
2068             vvdw             = _mm256_sub_ps( _mm256_mul_ps(vvdw12,one_twelfth) , _mm256_mul_ps(vvdw6,one_sixth) );
2069             fvdw             = _mm256_mul_ps(_mm256_sub_ps(vvdw12,vvdw6),rinvsq00);
2070
2071             d                = _mm256_sub_ps(r00,rswitch);
2072             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2073             d2               = _mm256_mul_ps(d,d);
2074             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)))))));
2075
2076             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2077
2078             /* Evaluate switch function */
2079             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2080             fvdw             = _mm256_sub_ps( _mm256_mul_ps(fvdw,sw) , _mm256_mul_ps(rinv00,_mm256_mul_ps(vvdw,dsw)) );
2081             cutoff_mask      = _mm256_cmp_ps(rsq00,rcutoff2,_CMP_LT_OQ);
2082
2083             fscal            = fvdw;
2084
2085             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2086
2087             /* Calculate temporary vectorial force */
2088             tx               = _mm256_mul_ps(fscal,dx00);
2089             ty               = _mm256_mul_ps(fscal,dy00);
2090             tz               = _mm256_mul_ps(fscal,dz00);
2091
2092             /* Update vectorial force */
2093             fix0             = _mm256_add_ps(fix0,tx);
2094             fiy0             = _mm256_add_ps(fiy0,ty);
2095             fiz0             = _mm256_add_ps(fiz0,tz);
2096
2097             fjx0             = _mm256_add_ps(fjx0,tx);
2098             fjy0             = _mm256_add_ps(fjy0,ty);
2099             fjz0             = _mm256_add_ps(fjz0,tz);
2100
2101             }
2102
2103             /**************************
2104              * CALCULATE INTERACTIONS *
2105              **************************/
2106
2107             if (gmx_mm256_any_lt(rsq11,rcutoff2))
2108             {
2109
2110             r11              = _mm256_mul_ps(rsq11,rinv11);
2111
2112             /* EWALD ELECTROSTATICS */
2113
2114             /* Analytical PME correction */
2115             zeta2            = _mm256_mul_ps(beta2,rsq11);
2116             rinv3            = _mm256_mul_ps(rinvsq11,rinv11);
2117             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
2118             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2119             felec            = _mm256_mul_ps(qq11,felec);
2120             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
2121             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2122             velec            = _mm256_sub_ps(rinv11,pmecorrV);
2123             velec            = _mm256_mul_ps(qq11,velec);
2124
2125             d                = _mm256_sub_ps(r11,rswitch);
2126             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2127             d2               = _mm256_mul_ps(d,d);
2128             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)))))));
2129
2130             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2131
2132             /* Evaluate switch function */
2133             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2134             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv11,_mm256_mul_ps(velec,dsw)) );
2135             cutoff_mask      = _mm256_cmp_ps(rsq11,rcutoff2,_CMP_LT_OQ);
2136
2137             fscal            = felec;
2138
2139             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2140
2141             /* Calculate temporary vectorial force */
2142             tx               = _mm256_mul_ps(fscal,dx11);
2143             ty               = _mm256_mul_ps(fscal,dy11);
2144             tz               = _mm256_mul_ps(fscal,dz11);
2145
2146             /* Update vectorial force */
2147             fix1             = _mm256_add_ps(fix1,tx);
2148             fiy1             = _mm256_add_ps(fiy1,ty);
2149             fiz1             = _mm256_add_ps(fiz1,tz);
2150
2151             fjx1             = _mm256_add_ps(fjx1,tx);
2152             fjy1             = _mm256_add_ps(fjy1,ty);
2153             fjz1             = _mm256_add_ps(fjz1,tz);
2154
2155             }
2156
2157             /**************************
2158              * CALCULATE INTERACTIONS *
2159              **************************/
2160
2161             if (gmx_mm256_any_lt(rsq12,rcutoff2))
2162             {
2163
2164             r12              = _mm256_mul_ps(rsq12,rinv12);
2165
2166             /* EWALD ELECTROSTATICS */
2167
2168             /* Analytical PME correction */
2169             zeta2            = _mm256_mul_ps(beta2,rsq12);
2170             rinv3            = _mm256_mul_ps(rinvsq12,rinv12);
2171             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
2172             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2173             felec            = _mm256_mul_ps(qq12,felec);
2174             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
2175             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2176             velec            = _mm256_sub_ps(rinv12,pmecorrV);
2177             velec            = _mm256_mul_ps(qq12,velec);
2178
2179             d                = _mm256_sub_ps(r12,rswitch);
2180             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2181             d2               = _mm256_mul_ps(d,d);
2182             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)))))));
2183
2184             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2185
2186             /* Evaluate switch function */
2187             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2188             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv12,_mm256_mul_ps(velec,dsw)) );
2189             cutoff_mask      = _mm256_cmp_ps(rsq12,rcutoff2,_CMP_LT_OQ);
2190
2191             fscal            = felec;
2192
2193             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2194
2195             /* Calculate temporary vectorial force */
2196             tx               = _mm256_mul_ps(fscal,dx12);
2197             ty               = _mm256_mul_ps(fscal,dy12);
2198             tz               = _mm256_mul_ps(fscal,dz12);
2199
2200             /* Update vectorial force */
2201             fix1             = _mm256_add_ps(fix1,tx);
2202             fiy1             = _mm256_add_ps(fiy1,ty);
2203             fiz1             = _mm256_add_ps(fiz1,tz);
2204
2205             fjx2             = _mm256_add_ps(fjx2,tx);
2206             fjy2             = _mm256_add_ps(fjy2,ty);
2207             fjz2             = _mm256_add_ps(fjz2,tz);
2208
2209             }
2210
2211             /**************************
2212              * CALCULATE INTERACTIONS *
2213              **************************/
2214
2215             if (gmx_mm256_any_lt(rsq13,rcutoff2))
2216             {
2217
2218             r13              = _mm256_mul_ps(rsq13,rinv13);
2219
2220             /* EWALD ELECTROSTATICS */
2221
2222             /* Analytical PME correction */
2223             zeta2            = _mm256_mul_ps(beta2,rsq13);
2224             rinv3            = _mm256_mul_ps(rinvsq13,rinv13);
2225             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
2226             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2227             felec            = _mm256_mul_ps(qq13,felec);
2228             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
2229             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2230             velec            = _mm256_sub_ps(rinv13,pmecorrV);
2231             velec            = _mm256_mul_ps(qq13,velec);
2232
2233             d                = _mm256_sub_ps(r13,rswitch);
2234             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2235             d2               = _mm256_mul_ps(d,d);
2236             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)))))));
2237
2238             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2239
2240             /* Evaluate switch function */
2241             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2242             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv13,_mm256_mul_ps(velec,dsw)) );
2243             cutoff_mask      = _mm256_cmp_ps(rsq13,rcutoff2,_CMP_LT_OQ);
2244
2245             fscal            = felec;
2246
2247             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2248
2249             /* Calculate temporary vectorial force */
2250             tx               = _mm256_mul_ps(fscal,dx13);
2251             ty               = _mm256_mul_ps(fscal,dy13);
2252             tz               = _mm256_mul_ps(fscal,dz13);
2253
2254             /* Update vectorial force */
2255             fix1             = _mm256_add_ps(fix1,tx);
2256             fiy1             = _mm256_add_ps(fiy1,ty);
2257             fiz1             = _mm256_add_ps(fiz1,tz);
2258
2259             fjx3             = _mm256_add_ps(fjx3,tx);
2260             fjy3             = _mm256_add_ps(fjy3,ty);
2261             fjz3             = _mm256_add_ps(fjz3,tz);
2262
2263             }
2264
2265             /**************************
2266              * CALCULATE INTERACTIONS *
2267              **************************/
2268
2269             if (gmx_mm256_any_lt(rsq21,rcutoff2))
2270             {
2271
2272             r21              = _mm256_mul_ps(rsq21,rinv21);
2273
2274             /* EWALD ELECTROSTATICS */
2275
2276             /* Analytical PME correction */
2277             zeta2            = _mm256_mul_ps(beta2,rsq21);
2278             rinv3            = _mm256_mul_ps(rinvsq21,rinv21);
2279             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
2280             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2281             felec            = _mm256_mul_ps(qq21,felec);
2282             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
2283             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2284             velec            = _mm256_sub_ps(rinv21,pmecorrV);
2285             velec            = _mm256_mul_ps(qq21,velec);
2286
2287             d                = _mm256_sub_ps(r21,rswitch);
2288             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2289             d2               = _mm256_mul_ps(d,d);
2290             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)))))));
2291
2292             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2293
2294             /* Evaluate switch function */
2295             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2296             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv21,_mm256_mul_ps(velec,dsw)) );
2297             cutoff_mask      = _mm256_cmp_ps(rsq21,rcutoff2,_CMP_LT_OQ);
2298
2299             fscal            = felec;
2300
2301             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2302
2303             /* Calculate temporary vectorial force */
2304             tx               = _mm256_mul_ps(fscal,dx21);
2305             ty               = _mm256_mul_ps(fscal,dy21);
2306             tz               = _mm256_mul_ps(fscal,dz21);
2307
2308             /* Update vectorial force */
2309             fix2             = _mm256_add_ps(fix2,tx);
2310             fiy2             = _mm256_add_ps(fiy2,ty);
2311             fiz2             = _mm256_add_ps(fiz2,tz);
2312
2313             fjx1             = _mm256_add_ps(fjx1,tx);
2314             fjy1             = _mm256_add_ps(fjy1,ty);
2315             fjz1             = _mm256_add_ps(fjz1,tz);
2316
2317             }
2318
2319             /**************************
2320              * CALCULATE INTERACTIONS *
2321              **************************/
2322
2323             if (gmx_mm256_any_lt(rsq22,rcutoff2))
2324             {
2325
2326             r22              = _mm256_mul_ps(rsq22,rinv22);
2327
2328             /* EWALD ELECTROSTATICS */
2329
2330             /* Analytical PME correction */
2331             zeta2            = _mm256_mul_ps(beta2,rsq22);
2332             rinv3            = _mm256_mul_ps(rinvsq22,rinv22);
2333             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
2334             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2335             felec            = _mm256_mul_ps(qq22,felec);
2336             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
2337             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2338             velec            = _mm256_sub_ps(rinv22,pmecorrV);
2339             velec            = _mm256_mul_ps(qq22,velec);
2340
2341             d                = _mm256_sub_ps(r22,rswitch);
2342             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2343             d2               = _mm256_mul_ps(d,d);
2344             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)))))));
2345
2346             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2347
2348             /* Evaluate switch function */
2349             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2350             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv22,_mm256_mul_ps(velec,dsw)) );
2351             cutoff_mask      = _mm256_cmp_ps(rsq22,rcutoff2,_CMP_LT_OQ);
2352
2353             fscal            = felec;
2354
2355             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2356
2357             /* Calculate temporary vectorial force */
2358             tx               = _mm256_mul_ps(fscal,dx22);
2359             ty               = _mm256_mul_ps(fscal,dy22);
2360             tz               = _mm256_mul_ps(fscal,dz22);
2361
2362             /* Update vectorial force */
2363             fix2             = _mm256_add_ps(fix2,tx);
2364             fiy2             = _mm256_add_ps(fiy2,ty);
2365             fiz2             = _mm256_add_ps(fiz2,tz);
2366
2367             fjx2             = _mm256_add_ps(fjx2,tx);
2368             fjy2             = _mm256_add_ps(fjy2,ty);
2369             fjz2             = _mm256_add_ps(fjz2,tz);
2370
2371             }
2372
2373             /**************************
2374              * CALCULATE INTERACTIONS *
2375              **************************/
2376
2377             if (gmx_mm256_any_lt(rsq23,rcutoff2))
2378             {
2379
2380             r23              = _mm256_mul_ps(rsq23,rinv23);
2381
2382             /* EWALD ELECTROSTATICS */
2383
2384             /* Analytical PME correction */
2385             zeta2            = _mm256_mul_ps(beta2,rsq23);
2386             rinv3            = _mm256_mul_ps(rinvsq23,rinv23);
2387             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
2388             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2389             felec            = _mm256_mul_ps(qq23,felec);
2390             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
2391             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2392             velec            = _mm256_sub_ps(rinv23,pmecorrV);
2393             velec            = _mm256_mul_ps(qq23,velec);
2394
2395             d                = _mm256_sub_ps(r23,rswitch);
2396             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2397             d2               = _mm256_mul_ps(d,d);
2398             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)))))));
2399
2400             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2401
2402             /* Evaluate switch function */
2403             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2404             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv23,_mm256_mul_ps(velec,dsw)) );
2405             cutoff_mask      = _mm256_cmp_ps(rsq23,rcutoff2,_CMP_LT_OQ);
2406
2407             fscal            = felec;
2408
2409             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2410
2411             /* Calculate temporary vectorial force */
2412             tx               = _mm256_mul_ps(fscal,dx23);
2413             ty               = _mm256_mul_ps(fscal,dy23);
2414             tz               = _mm256_mul_ps(fscal,dz23);
2415
2416             /* Update vectorial force */
2417             fix2             = _mm256_add_ps(fix2,tx);
2418             fiy2             = _mm256_add_ps(fiy2,ty);
2419             fiz2             = _mm256_add_ps(fiz2,tz);
2420
2421             fjx3             = _mm256_add_ps(fjx3,tx);
2422             fjy3             = _mm256_add_ps(fjy3,ty);
2423             fjz3             = _mm256_add_ps(fjz3,tz);
2424
2425             }
2426
2427             /**************************
2428              * CALCULATE INTERACTIONS *
2429              **************************/
2430
2431             if (gmx_mm256_any_lt(rsq31,rcutoff2))
2432             {
2433
2434             r31              = _mm256_mul_ps(rsq31,rinv31);
2435
2436             /* EWALD ELECTROSTATICS */
2437
2438             /* Analytical PME correction */
2439             zeta2            = _mm256_mul_ps(beta2,rsq31);
2440             rinv3            = _mm256_mul_ps(rinvsq31,rinv31);
2441             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
2442             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2443             felec            = _mm256_mul_ps(qq31,felec);
2444             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
2445             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2446             velec            = _mm256_sub_ps(rinv31,pmecorrV);
2447             velec            = _mm256_mul_ps(qq31,velec);
2448
2449             d                = _mm256_sub_ps(r31,rswitch);
2450             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2451             d2               = _mm256_mul_ps(d,d);
2452             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)))))));
2453
2454             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2455
2456             /* Evaluate switch function */
2457             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2458             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv31,_mm256_mul_ps(velec,dsw)) );
2459             cutoff_mask      = _mm256_cmp_ps(rsq31,rcutoff2,_CMP_LT_OQ);
2460
2461             fscal            = felec;
2462
2463             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2464
2465             /* Calculate temporary vectorial force */
2466             tx               = _mm256_mul_ps(fscal,dx31);
2467             ty               = _mm256_mul_ps(fscal,dy31);
2468             tz               = _mm256_mul_ps(fscal,dz31);
2469
2470             /* Update vectorial force */
2471             fix3             = _mm256_add_ps(fix3,tx);
2472             fiy3             = _mm256_add_ps(fiy3,ty);
2473             fiz3             = _mm256_add_ps(fiz3,tz);
2474
2475             fjx1             = _mm256_add_ps(fjx1,tx);
2476             fjy1             = _mm256_add_ps(fjy1,ty);
2477             fjz1             = _mm256_add_ps(fjz1,tz);
2478
2479             }
2480
2481             /**************************
2482              * CALCULATE INTERACTIONS *
2483              **************************/
2484
2485             if (gmx_mm256_any_lt(rsq32,rcutoff2))
2486             {
2487
2488             r32              = _mm256_mul_ps(rsq32,rinv32);
2489
2490             /* EWALD ELECTROSTATICS */
2491
2492             /* Analytical PME correction */
2493             zeta2            = _mm256_mul_ps(beta2,rsq32);
2494             rinv3            = _mm256_mul_ps(rinvsq32,rinv32);
2495             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
2496             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2497             felec            = _mm256_mul_ps(qq32,felec);
2498             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
2499             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2500             velec            = _mm256_sub_ps(rinv32,pmecorrV);
2501             velec            = _mm256_mul_ps(qq32,velec);
2502
2503             d                = _mm256_sub_ps(r32,rswitch);
2504             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2505             d2               = _mm256_mul_ps(d,d);
2506             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)))))));
2507
2508             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2509
2510             /* Evaluate switch function */
2511             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2512             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv32,_mm256_mul_ps(velec,dsw)) );
2513             cutoff_mask      = _mm256_cmp_ps(rsq32,rcutoff2,_CMP_LT_OQ);
2514
2515             fscal            = felec;
2516
2517             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2518
2519             /* Calculate temporary vectorial force */
2520             tx               = _mm256_mul_ps(fscal,dx32);
2521             ty               = _mm256_mul_ps(fscal,dy32);
2522             tz               = _mm256_mul_ps(fscal,dz32);
2523
2524             /* Update vectorial force */
2525             fix3             = _mm256_add_ps(fix3,tx);
2526             fiy3             = _mm256_add_ps(fiy3,ty);
2527             fiz3             = _mm256_add_ps(fiz3,tz);
2528
2529             fjx2             = _mm256_add_ps(fjx2,tx);
2530             fjy2             = _mm256_add_ps(fjy2,ty);
2531             fjz2             = _mm256_add_ps(fjz2,tz);
2532
2533             }
2534
2535             /**************************
2536              * CALCULATE INTERACTIONS *
2537              **************************/
2538
2539             if (gmx_mm256_any_lt(rsq33,rcutoff2))
2540             {
2541
2542             r33              = _mm256_mul_ps(rsq33,rinv33);
2543
2544             /* EWALD ELECTROSTATICS */
2545
2546             /* Analytical PME correction */
2547             zeta2            = _mm256_mul_ps(beta2,rsq33);
2548             rinv3            = _mm256_mul_ps(rinvsq33,rinv33);
2549             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
2550             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2551             felec            = _mm256_mul_ps(qq33,felec);
2552             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
2553             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2554             velec            = _mm256_sub_ps(rinv33,pmecorrV);
2555             velec            = _mm256_mul_ps(qq33,velec);
2556
2557             d                = _mm256_sub_ps(r33,rswitch);
2558             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2559             d2               = _mm256_mul_ps(d,d);
2560             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)))))));
2561
2562             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2563
2564             /* Evaluate switch function */
2565             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2566             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv33,_mm256_mul_ps(velec,dsw)) );
2567             cutoff_mask      = _mm256_cmp_ps(rsq33,rcutoff2,_CMP_LT_OQ);
2568
2569             fscal            = felec;
2570
2571             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2572
2573             /* Calculate temporary vectorial force */
2574             tx               = _mm256_mul_ps(fscal,dx33);
2575             ty               = _mm256_mul_ps(fscal,dy33);
2576             tz               = _mm256_mul_ps(fscal,dz33);
2577
2578             /* Update vectorial force */
2579             fix3             = _mm256_add_ps(fix3,tx);
2580             fiy3             = _mm256_add_ps(fiy3,ty);
2581             fiz3             = _mm256_add_ps(fiz3,tz);
2582
2583             fjx3             = _mm256_add_ps(fjx3,tx);
2584             fjy3             = _mm256_add_ps(fjy3,ty);
2585             fjz3             = _mm256_add_ps(fjz3,tz);
2586
2587             }
2588
2589             fjptrA             = f+j_coord_offsetA;
2590             fjptrB             = f+j_coord_offsetB;
2591             fjptrC             = f+j_coord_offsetC;
2592             fjptrD             = f+j_coord_offsetD;
2593             fjptrE             = f+j_coord_offsetE;
2594             fjptrF             = f+j_coord_offsetF;
2595             fjptrG             = f+j_coord_offsetG;
2596             fjptrH             = f+j_coord_offsetH;
2597
2598             gmx_mm256_decrement_4rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,
2599                                                       fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,
2600                                                       fjx2,fjy2,fjz2,fjx3,fjy3,fjz3);
2601
2602             /* Inner loop uses 1004 flops */
2603         }
2604
2605         if(jidx<j_index_end)
2606         {
2607
2608             /* Get j neighbor index, and coordinate index */
2609             jnrlistA         = jjnr[jidx];
2610             jnrlistB         = jjnr[jidx+1];
2611             jnrlistC         = jjnr[jidx+2];
2612             jnrlistD         = jjnr[jidx+3];
2613             jnrlistE         = jjnr[jidx+4];
2614             jnrlistF         = jjnr[jidx+5];
2615             jnrlistG         = jjnr[jidx+6];
2616             jnrlistH         = jjnr[jidx+7];
2617             /* Sign of each element will be negative for non-real atoms.
2618              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
2619              * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
2620              */
2621             dummy_mask = gmx_mm256_set_m128(gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx+4)),_mm_setzero_si128())),
2622                                             gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128())));
2623
2624             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
2625             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
2626             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
2627             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
2628             jnrE       = (jnrlistE>=0) ? jnrlistE : 0;
2629             jnrF       = (jnrlistF>=0) ? jnrlistF : 0;
2630             jnrG       = (jnrlistG>=0) ? jnrlistG : 0;
2631             jnrH       = (jnrlistH>=0) ? jnrlistH : 0;
2632             j_coord_offsetA  = DIM*jnrA;
2633             j_coord_offsetB  = DIM*jnrB;
2634             j_coord_offsetC  = DIM*jnrC;
2635             j_coord_offsetD  = DIM*jnrD;
2636             j_coord_offsetE  = DIM*jnrE;
2637             j_coord_offsetF  = DIM*jnrF;
2638             j_coord_offsetG  = DIM*jnrG;
2639             j_coord_offsetH  = DIM*jnrH;
2640
2641             /* load j atom coordinates */
2642             gmx_mm256_load_4rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
2643                                                  x+j_coord_offsetC,x+j_coord_offsetD,
2644                                                  x+j_coord_offsetE,x+j_coord_offsetF,
2645                                                  x+j_coord_offsetG,x+j_coord_offsetH,
2646                                                  &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,
2647                                                  &jy2,&jz2,&jx3,&jy3,&jz3);
2648
2649             /* Calculate displacement vector */
2650             dx00             = _mm256_sub_ps(ix0,jx0);
2651             dy00             = _mm256_sub_ps(iy0,jy0);
2652             dz00             = _mm256_sub_ps(iz0,jz0);
2653             dx11             = _mm256_sub_ps(ix1,jx1);
2654             dy11             = _mm256_sub_ps(iy1,jy1);
2655             dz11             = _mm256_sub_ps(iz1,jz1);
2656             dx12             = _mm256_sub_ps(ix1,jx2);
2657             dy12             = _mm256_sub_ps(iy1,jy2);
2658             dz12             = _mm256_sub_ps(iz1,jz2);
2659             dx13             = _mm256_sub_ps(ix1,jx3);
2660             dy13             = _mm256_sub_ps(iy1,jy3);
2661             dz13             = _mm256_sub_ps(iz1,jz3);
2662             dx21             = _mm256_sub_ps(ix2,jx1);
2663             dy21             = _mm256_sub_ps(iy2,jy1);
2664             dz21             = _mm256_sub_ps(iz2,jz1);
2665             dx22             = _mm256_sub_ps(ix2,jx2);
2666             dy22             = _mm256_sub_ps(iy2,jy2);
2667             dz22             = _mm256_sub_ps(iz2,jz2);
2668             dx23             = _mm256_sub_ps(ix2,jx3);
2669             dy23             = _mm256_sub_ps(iy2,jy3);
2670             dz23             = _mm256_sub_ps(iz2,jz3);
2671             dx31             = _mm256_sub_ps(ix3,jx1);
2672             dy31             = _mm256_sub_ps(iy3,jy1);
2673             dz31             = _mm256_sub_ps(iz3,jz1);
2674             dx32             = _mm256_sub_ps(ix3,jx2);
2675             dy32             = _mm256_sub_ps(iy3,jy2);
2676             dz32             = _mm256_sub_ps(iz3,jz2);
2677             dx33             = _mm256_sub_ps(ix3,jx3);
2678             dy33             = _mm256_sub_ps(iy3,jy3);
2679             dz33             = _mm256_sub_ps(iz3,jz3);
2680
2681             /* Calculate squared distance and things based on it */
2682             rsq00            = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
2683             rsq11            = gmx_mm256_calc_rsq_ps(dx11,dy11,dz11);
2684             rsq12            = gmx_mm256_calc_rsq_ps(dx12,dy12,dz12);
2685             rsq13            = gmx_mm256_calc_rsq_ps(dx13,dy13,dz13);
2686             rsq21            = gmx_mm256_calc_rsq_ps(dx21,dy21,dz21);
2687             rsq22            = gmx_mm256_calc_rsq_ps(dx22,dy22,dz22);
2688             rsq23            = gmx_mm256_calc_rsq_ps(dx23,dy23,dz23);
2689             rsq31            = gmx_mm256_calc_rsq_ps(dx31,dy31,dz31);
2690             rsq32            = gmx_mm256_calc_rsq_ps(dx32,dy32,dz32);
2691             rsq33            = gmx_mm256_calc_rsq_ps(dx33,dy33,dz33);
2692
2693             rinv00           = gmx_mm256_invsqrt_ps(rsq00);
2694             rinv11           = gmx_mm256_invsqrt_ps(rsq11);
2695             rinv12           = gmx_mm256_invsqrt_ps(rsq12);
2696             rinv13           = gmx_mm256_invsqrt_ps(rsq13);
2697             rinv21           = gmx_mm256_invsqrt_ps(rsq21);
2698             rinv22           = gmx_mm256_invsqrt_ps(rsq22);
2699             rinv23           = gmx_mm256_invsqrt_ps(rsq23);
2700             rinv31           = gmx_mm256_invsqrt_ps(rsq31);
2701             rinv32           = gmx_mm256_invsqrt_ps(rsq32);
2702             rinv33           = gmx_mm256_invsqrt_ps(rsq33);
2703
2704             rinvsq00         = _mm256_mul_ps(rinv00,rinv00);
2705             rinvsq11         = _mm256_mul_ps(rinv11,rinv11);
2706             rinvsq12         = _mm256_mul_ps(rinv12,rinv12);
2707             rinvsq13         = _mm256_mul_ps(rinv13,rinv13);
2708             rinvsq21         = _mm256_mul_ps(rinv21,rinv21);
2709             rinvsq22         = _mm256_mul_ps(rinv22,rinv22);
2710             rinvsq23         = _mm256_mul_ps(rinv23,rinv23);
2711             rinvsq31         = _mm256_mul_ps(rinv31,rinv31);
2712             rinvsq32         = _mm256_mul_ps(rinv32,rinv32);
2713             rinvsq33         = _mm256_mul_ps(rinv33,rinv33);
2714
2715             fjx0             = _mm256_setzero_ps();
2716             fjy0             = _mm256_setzero_ps();
2717             fjz0             = _mm256_setzero_ps();
2718             fjx1             = _mm256_setzero_ps();
2719             fjy1             = _mm256_setzero_ps();
2720             fjz1             = _mm256_setzero_ps();
2721             fjx2             = _mm256_setzero_ps();
2722             fjy2             = _mm256_setzero_ps();
2723             fjz2             = _mm256_setzero_ps();
2724             fjx3             = _mm256_setzero_ps();
2725             fjy3             = _mm256_setzero_ps();
2726             fjz3             = _mm256_setzero_ps();
2727
2728             /**************************
2729              * CALCULATE INTERACTIONS *
2730              **************************/
2731
2732             if (gmx_mm256_any_lt(rsq00,rcutoff2))
2733             {
2734
2735             r00              = _mm256_mul_ps(rsq00,rinv00);
2736             r00              = _mm256_andnot_ps(dummy_mask,r00);
2737
2738             /* LENNARD-JONES DISPERSION/REPULSION */
2739
2740             rinvsix          = _mm256_mul_ps(_mm256_mul_ps(rinvsq00,rinvsq00),rinvsq00);
2741             vvdw6            = _mm256_mul_ps(c6_00,rinvsix);
2742             vvdw12           = _mm256_mul_ps(c12_00,_mm256_mul_ps(rinvsix,rinvsix));
2743             vvdw             = _mm256_sub_ps( _mm256_mul_ps(vvdw12,one_twelfth) , _mm256_mul_ps(vvdw6,one_sixth) );
2744             fvdw             = _mm256_mul_ps(_mm256_sub_ps(vvdw12,vvdw6),rinvsq00);
2745
2746             d                = _mm256_sub_ps(r00,rswitch);
2747             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2748             d2               = _mm256_mul_ps(d,d);
2749             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)))))));
2750
2751             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2752
2753             /* Evaluate switch function */
2754             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2755             fvdw             = _mm256_sub_ps( _mm256_mul_ps(fvdw,sw) , _mm256_mul_ps(rinv00,_mm256_mul_ps(vvdw,dsw)) );
2756             cutoff_mask      = _mm256_cmp_ps(rsq00,rcutoff2,_CMP_LT_OQ);
2757
2758             fscal            = fvdw;
2759
2760             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2761
2762             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
2763
2764             /* Calculate temporary vectorial force */
2765             tx               = _mm256_mul_ps(fscal,dx00);
2766             ty               = _mm256_mul_ps(fscal,dy00);
2767             tz               = _mm256_mul_ps(fscal,dz00);
2768
2769             /* Update vectorial force */
2770             fix0             = _mm256_add_ps(fix0,tx);
2771             fiy0             = _mm256_add_ps(fiy0,ty);
2772             fiz0             = _mm256_add_ps(fiz0,tz);
2773
2774             fjx0             = _mm256_add_ps(fjx0,tx);
2775             fjy0             = _mm256_add_ps(fjy0,ty);
2776             fjz0             = _mm256_add_ps(fjz0,tz);
2777
2778             }
2779
2780             /**************************
2781              * CALCULATE INTERACTIONS *
2782              **************************/
2783
2784             if (gmx_mm256_any_lt(rsq11,rcutoff2))
2785             {
2786
2787             r11              = _mm256_mul_ps(rsq11,rinv11);
2788             r11              = _mm256_andnot_ps(dummy_mask,r11);
2789
2790             /* EWALD ELECTROSTATICS */
2791
2792             /* Analytical PME correction */
2793             zeta2            = _mm256_mul_ps(beta2,rsq11);
2794             rinv3            = _mm256_mul_ps(rinvsq11,rinv11);
2795             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
2796             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2797             felec            = _mm256_mul_ps(qq11,felec);
2798             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
2799             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2800             velec            = _mm256_sub_ps(rinv11,pmecorrV);
2801             velec            = _mm256_mul_ps(qq11,velec);
2802
2803             d                = _mm256_sub_ps(r11,rswitch);
2804             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2805             d2               = _mm256_mul_ps(d,d);
2806             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)))))));
2807
2808             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2809
2810             /* Evaluate switch function */
2811             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2812             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv11,_mm256_mul_ps(velec,dsw)) );
2813             cutoff_mask      = _mm256_cmp_ps(rsq11,rcutoff2,_CMP_LT_OQ);
2814
2815             fscal            = felec;
2816
2817             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2818
2819             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
2820
2821             /* Calculate temporary vectorial force */
2822             tx               = _mm256_mul_ps(fscal,dx11);
2823             ty               = _mm256_mul_ps(fscal,dy11);
2824             tz               = _mm256_mul_ps(fscal,dz11);
2825
2826             /* Update vectorial force */
2827             fix1             = _mm256_add_ps(fix1,tx);
2828             fiy1             = _mm256_add_ps(fiy1,ty);
2829             fiz1             = _mm256_add_ps(fiz1,tz);
2830
2831             fjx1             = _mm256_add_ps(fjx1,tx);
2832             fjy1             = _mm256_add_ps(fjy1,ty);
2833             fjz1             = _mm256_add_ps(fjz1,tz);
2834
2835             }
2836
2837             /**************************
2838              * CALCULATE INTERACTIONS *
2839              **************************/
2840
2841             if (gmx_mm256_any_lt(rsq12,rcutoff2))
2842             {
2843
2844             r12              = _mm256_mul_ps(rsq12,rinv12);
2845             r12              = _mm256_andnot_ps(dummy_mask,r12);
2846
2847             /* EWALD ELECTROSTATICS */
2848
2849             /* Analytical PME correction */
2850             zeta2            = _mm256_mul_ps(beta2,rsq12);
2851             rinv3            = _mm256_mul_ps(rinvsq12,rinv12);
2852             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
2853             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2854             felec            = _mm256_mul_ps(qq12,felec);
2855             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
2856             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2857             velec            = _mm256_sub_ps(rinv12,pmecorrV);
2858             velec            = _mm256_mul_ps(qq12,velec);
2859
2860             d                = _mm256_sub_ps(r12,rswitch);
2861             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2862             d2               = _mm256_mul_ps(d,d);
2863             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)))))));
2864
2865             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2866
2867             /* Evaluate switch function */
2868             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2869             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv12,_mm256_mul_ps(velec,dsw)) );
2870             cutoff_mask      = _mm256_cmp_ps(rsq12,rcutoff2,_CMP_LT_OQ);
2871
2872             fscal            = felec;
2873
2874             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2875
2876             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
2877
2878             /* Calculate temporary vectorial force */
2879             tx               = _mm256_mul_ps(fscal,dx12);
2880             ty               = _mm256_mul_ps(fscal,dy12);
2881             tz               = _mm256_mul_ps(fscal,dz12);
2882
2883             /* Update vectorial force */
2884             fix1             = _mm256_add_ps(fix1,tx);
2885             fiy1             = _mm256_add_ps(fiy1,ty);
2886             fiz1             = _mm256_add_ps(fiz1,tz);
2887
2888             fjx2             = _mm256_add_ps(fjx2,tx);
2889             fjy2             = _mm256_add_ps(fjy2,ty);
2890             fjz2             = _mm256_add_ps(fjz2,tz);
2891
2892             }
2893
2894             /**************************
2895              * CALCULATE INTERACTIONS *
2896              **************************/
2897
2898             if (gmx_mm256_any_lt(rsq13,rcutoff2))
2899             {
2900
2901             r13              = _mm256_mul_ps(rsq13,rinv13);
2902             r13              = _mm256_andnot_ps(dummy_mask,r13);
2903
2904             /* EWALD ELECTROSTATICS */
2905
2906             /* Analytical PME correction */
2907             zeta2            = _mm256_mul_ps(beta2,rsq13);
2908             rinv3            = _mm256_mul_ps(rinvsq13,rinv13);
2909             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
2910             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2911             felec            = _mm256_mul_ps(qq13,felec);
2912             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
2913             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2914             velec            = _mm256_sub_ps(rinv13,pmecorrV);
2915             velec            = _mm256_mul_ps(qq13,velec);
2916
2917             d                = _mm256_sub_ps(r13,rswitch);
2918             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2919             d2               = _mm256_mul_ps(d,d);
2920             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)))))));
2921
2922             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2923
2924             /* Evaluate switch function */
2925             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2926             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv13,_mm256_mul_ps(velec,dsw)) );
2927             cutoff_mask      = _mm256_cmp_ps(rsq13,rcutoff2,_CMP_LT_OQ);
2928
2929             fscal            = felec;
2930
2931             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2932
2933             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
2934
2935             /* Calculate temporary vectorial force */
2936             tx               = _mm256_mul_ps(fscal,dx13);
2937             ty               = _mm256_mul_ps(fscal,dy13);
2938             tz               = _mm256_mul_ps(fscal,dz13);
2939
2940             /* Update vectorial force */
2941             fix1             = _mm256_add_ps(fix1,tx);
2942             fiy1             = _mm256_add_ps(fiy1,ty);
2943             fiz1             = _mm256_add_ps(fiz1,tz);
2944
2945             fjx3             = _mm256_add_ps(fjx3,tx);
2946             fjy3             = _mm256_add_ps(fjy3,ty);
2947             fjz3             = _mm256_add_ps(fjz3,tz);
2948
2949             }
2950
2951             /**************************
2952              * CALCULATE INTERACTIONS *
2953              **************************/
2954
2955             if (gmx_mm256_any_lt(rsq21,rcutoff2))
2956             {
2957
2958             r21              = _mm256_mul_ps(rsq21,rinv21);
2959             r21              = _mm256_andnot_ps(dummy_mask,r21);
2960
2961             /* EWALD ELECTROSTATICS */
2962
2963             /* Analytical PME correction */
2964             zeta2            = _mm256_mul_ps(beta2,rsq21);
2965             rinv3            = _mm256_mul_ps(rinvsq21,rinv21);
2966             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
2967             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2968             felec            = _mm256_mul_ps(qq21,felec);
2969             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
2970             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2971             velec            = _mm256_sub_ps(rinv21,pmecorrV);
2972             velec            = _mm256_mul_ps(qq21,velec);
2973
2974             d                = _mm256_sub_ps(r21,rswitch);
2975             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2976             d2               = _mm256_mul_ps(d,d);
2977             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)))))));
2978
2979             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2980
2981             /* Evaluate switch function */
2982             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2983             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv21,_mm256_mul_ps(velec,dsw)) );
2984             cutoff_mask      = _mm256_cmp_ps(rsq21,rcutoff2,_CMP_LT_OQ);
2985
2986             fscal            = felec;
2987
2988             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2989
2990             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
2991
2992             /* Calculate temporary vectorial force */
2993             tx               = _mm256_mul_ps(fscal,dx21);
2994             ty               = _mm256_mul_ps(fscal,dy21);
2995             tz               = _mm256_mul_ps(fscal,dz21);
2996
2997             /* Update vectorial force */
2998             fix2             = _mm256_add_ps(fix2,tx);
2999             fiy2             = _mm256_add_ps(fiy2,ty);
3000             fiz2             = _mm256_add_ps(fiz2,tz);
3001
3002             fjx1             = _mm256_add_ps(fjx1,tx);
3003             fjy1             = _mm256_add_ps(fjy1,ty);
3004             fjz1             = _mm256_add_ps(fjz1,tz);
3005
3006             }
3007
3008             /**************************
3009              * CALCULATE INTERACTIONS *
3010              **************************/
3011
3012             if (gmx_mm256_any_lt(rsq22,rcutoff2))
3013             {
3014
3015             r22              = _mm256_mul_ps(rsq22,rinv22);
3016             r22              = _mm256_andnot_ps(dummy_mask,r22);
3017
3018             /* EWALD ELECTROSTATICS */
3019
3020             /* Analytical PME correction */
3021             zeta2            = _mm256_mul_ps(beta2,rsq22);
3022             rinv3            = _mm256_mul_ps(rinvsq22,rinv22);
3023             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
3024             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
3025             felec            = _mm256_mul_ps(qq22,felec);
3026             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
3027             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
3028             velec            = _mm256_sub_ps(rinv22,pmecorrV);
3029             velec            = _mm256_mul_ps(qq22,velec);
3030
3031             d                = _mm256_sub_ps(r22,rswitch);
3032             d                = _mm256_max_ps(d,_mm256_setzero_ps());
3033             d2               = _mm256_mul_ps(d,d);
3034             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)))))));
3035
3036             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
3037
3038             /* Evaluate switch function */
3039             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
3040             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv22,_mm256_mul_ps(velec,dsw)) );
3041             cutoff_mask      = _mm256_cmp_ps(rsq22,rcutoff2,_CMP_LT_OQ);
3042
3043             fscal            = felec;
3044
3045             fscal            = _mm256_and_ps(fscal,cutoff_mask);
3046
3047             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
3048
3049             /* Calculate temporary vectorial force */
3050             tx               = _mm256_mul_ps(fscal,dx22);
3051             ty               = _mm256_mul_ps(fscal,dy22);
3052             tz               = _mm256_mul_ps(fscal,dz22);
3053
3054             /* Update vectorial force */
3055             fix2             = _mm256_add_ps(fix2,tx);
3056             fiy2             = _mm256_add_ps(fiy2,ty);
3057             fiz2             = _mm256_add_ps(fiz2,tz);
3058
3059             fjx2             = _mm256_add_ps(fjx2,tx);
3060             fjy2             = _mm256_add_ps(fjy2,ty);
3061             fjz2             = _mm256_add_ps(fjz2,tz);
3062
3063             }
3064
3065             /**************************
3066              * CALCULATE INTERACTIONS *
3067              **************************/
3068
3069             if (gmx_mm256_any_lt(rsq23,rcutoff2))
3070             {
3071
3072             r23              = _mm256_mul_ps(rsq23,rinv23);
3073             r23              = _mm256_andnot_ps(dummy_mask,r23);
3074
3075             /* EWALD ELECTROSTATICS */
3076
3077             /* Analytical PME correction */
3078             zeta2            = _mm256_mul_ps(beta2,rsq23);
3079             rinv3            = _mm256_mul_ps(rinvsq23,rinv23);
3080             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
3081             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
3082             felec            = _mm256_mul_ps(qq23,felec);
3083             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
3084             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
3085             velec            = _mm256_sub_ps(rinv23,pmecorrV);
3086             velec            = _mm256_mul_ps(qq23,velec);
3087
3088             d                = _mm256_sub_ps(r23,rswitch);
3089             d                = _mm256_max_ps(d,_mm256_setzero_ps());
3090             d2               = _mm256_mul_ps(d,d);
3091             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)))))));
3092
3093             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
3094
3095             /* Evaluate switch function */
3096             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
3097             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv23,_mm256_mul_ps(velec,dsw)) );
3098             cutoff_mask      = _mm256_cmp_ps(rsq23,rcutoff2,_CMP_LT_OQ);
3099
3100             fscal            = felec;
3101
3102             fscal            = _mm256_and_ps(fscal,cutoff_mask);
3103
3104             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
3105
3106             /* Calculate temporary vectorial force */
3107             tx               = _mm256_mul_ps(fscal,dx23);
3108             ty               = _mm256_mul_ps(fscal,dy23);
3109             tz               = _mm256_mul_ps(fscal,dz23);
3110
3111             /* Update vectorial force */
3112             fix2             = _mm256_add_ps(fix2,tx);
3113             fiy2             = _mm256_add_ps(fiy2,ty);
3114             fiz2             = _mm256_add_ps(fiz2,tz);
3115
3116             fjx3             = _mm256_add_ps(fjx3,tx);
3117             fjy3             = _mm256_add_ps(fjy3,ty);
3118             fjz3             = _mm256_add_ps(fjz3,tz);
3119
3120             }
3121
3122             /**************************
3123              * CALCULATE INTERACTIONS *
3124              **************************/
3125
3126             if (gmx_mm256_any_lt(rsq31,rcutoff2))
3127             {
3128
3129             r31              = _mm256_mul_ps(rsq31,rinv31);
3130             r31              = _mm256_andnot_ps(dummy_mask,r31);
3131
3132             /* EWALD ELECTROSTATICS */
3133
3134             /* Analytical PME correction */
3135             zeta2            = _mm256_mul_ps(beta2,rsq31);
3136             rinv3            = _mm256_mul_ps(rinvsq31,rinv31);
3137             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
3138             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
3139             felec            = _mm256_mul_ps(qq31,felec);
3140             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
3141             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
3142             velec            = _mm256_sub_ps(rinv31,pmecorrV);
3143             velec            = _mm256_mul_ps(qq31,velec);
3144
3145             d                = _mm256_sub_ps(r31,rswitch);
3146             d                = _mm256_max_ps(d,_mm256_setzero_ps());
3147             d2               = _mm256_mul_ps(d,d);
3148             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)))))));
3149
3150             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
3151
3152             /* Evaluate switch function */
3153             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
3154             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv31,_mm256_mul_ps(velec,dsw)) );
3155             cutoff_mask      = _mm256_cmp_ps(rsq31,rcutoff2,_CMP_LT_OQ);
3156
3157             fscal            = felec;
3158
3159             fscal            = _mm256_and_ps(fscal,cutoff_mask);
3160
3161             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
3162
3163             /* Calculate temporary vectorial force */
3164             tx               = _mm256_mul_ps(fscal,dx31);
3165             ty               = _mm256_mul_ps(fscal,dy31);
3166             tz               = _mm256_mul_ps(fscal,dz31);
3167
3168             /* Update vectorial force */
3169             fix3             = _mm256_add_ps(fix3,tx);
3170             fiy3             = _mm256_add_ps(fiy3,ty);
3171             fiz3             = _mm256_add_ps(fiz3,tz);
3172
3173             fjx1             = _mm256_add_ps(fjx1,tx);
3174             fjy1             = _mm256_add_ps(fjy1,ty);
3175             fjz1             = _mm256_add_ps(fjz1,tz);
3176
3177             }
3178
3179             /**************************
3180              * CALCULATE INTERACTIONS *
3181              **************************/
3182
3183             if (gmx_mm256_any_lt(rsq32,rcutoff2))
3184             {
3185
3186             r32              = _mm256_mul_ps(rsq32,rinv32);
3187             r32              = _mm256_andnot_ps(dummy_mask,r32);
3188
3189             /* EWALD ELECTROSTATICS */
3190
3191             /* Analytical PME correction */
3192             zeta2            = _mm256_mul_ps(beta2,rsq32);
3193             rinv3            = _mm256_mul_ps(rinvsq32,rinv32);
3194             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
3195             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
3196             felec            = _mm256_mul_ps(qq32,felec);
3197             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
3198             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
3199             velec            = _mm256_sub_ps(rinv32,pmecorrV);
3200             velec            = _mm256_mul_ps(qq32,velec);
3201
3202             d                = _mm256_sub_ps(r32,rswitch);
3203             d                = _mm256_max_ps(d,_mm256_setzero_ps());
3204             d2               = _mm256_mul_ps(d,d);
3205             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)))))));
3206
3207             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
3208
3209             /* Evaluate switch function */
3210             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
3211             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv32,_mm256_mul_ps(velec,dsw)) );
3212             cutoff_mask      = _mm256_cmp_ps(rsq32,rcutoff2,_CMP_LT_OQ);
3213
3214             fscal            = felec;
3215
3216             fscal            = _mm256_and_ps(fscal,cutoff_mask);
3217
3218             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
3219
3220             /* Calculate temporary vectorial force */
3221             tx               = _mm256_mul_ps(fscal,dx32);
3222             ty               = _mm256_mul_ps(fscal,dy32);
3223             tz               = _mm256_mul_ps(fscal,dz32);
3224
3225             /* Update vectorial force */
3226             fix3             = _mm256_add_ps(fix3,tx);
3227             fiy3             = _mm256_add_ps(fiy3,ty);
3228             fiz3             = _mm256_add_ps(fiz3,tz);
3229
3230             fjx2             = _mm256_add_ps(fjx2,tx);
3231             fjy2             = _mm256_add_ps(fjy2,ty);
3232             fjz2             = _mm256_add_ps(fjz2,tz);
3233
3234             }
3235
3236             /**************************
3237              * CALCULATE INTERACTIONS *
3238              **************************/
3239
3240             if (gmx_mm256_any_lt(rsq33,rcutoff2))
3241             {
3242
3243             r33              = _mm256_mul_ps(rsq33,rinv33);
3244             r33              = _mm256_andnot_ps(dummy_mask,r33);
3245
3246             /* EWALD ELECTROSTATICS */
3247
3248             /* Analytical PME correction */
3249             zeta2            = _mm256_mul_ps(beta2,rsq33);
3250             rinv3            = _mm256_mul_ps(rinvsq33,rinv33);
3251             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
3252             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
3253             felec            = _mm256_mul_ps(qq33,felec);
3254             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
3255             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
3256             velec            = _mm256_sub_ps(rinv33,pmecorrV);
3257             velec            = _mm256_mul_ps(qq33,velec);
3258
3259             d                = _mm256_sub_ps(r33,rswitch);
3260             d                = _mm256_max_ps(d,_mm256_setzero_ps());
3261             d2               = _mm256_mul_ps(d,d);
3262             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)))))));
3263
3264             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
3265
3266             /* Evaluate switch function */
3267             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
3268             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv33,_mm256_mul_ps(velec,dsw)) );
3269             cutoff_mask      = _mm256_cmp_ps(rsq33,rcutoff2,_CMP_LT_OQ);
3270
3271             fscal            = felec;
3272
3273             fscal            = _mm256_and_ps(fscal,cutoff_mask);
3274
3275             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
3276
3277             /* Calculate temporary vectorial force */
3278             tx               = _mm256_mul_ps(fscal,dx33);
3279             ty               = _mm256_mul_ps(fscal,dy33);
3280             tz               = _mm256_mul_ps(fscal,dz33);
3281
3282             /* Update vectorial force */
3283             fix3             = _mm256_add_ps(fix3,tx);
3284             fiy3             = _mm256_add_ps(fiy3,ty);
3285             fiz3             = _mm256_add_ps(fiz3,tz);
3286
3287             fjx3             = _mm256_add_ps(fjx3,tx);
3288             fjy3             = _mm256_add_ps(fjy3,ty);
3289             fjz3             = _mm256_add_ps(fjz3,tz);
3290
3291             }
3292
3293             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
3294             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
3295             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
3296             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
3297             fjptrE             = (jnrlistE>=0) ? f+j_coord_offsetE : scratch;
3298             fjptrF             = (jnrlistF>=0) ? f+j_coord_offsetF : scratch;
3299             fjptrG             = (jnrlistG>=0) ? f+j_coord_offsetG : scratch;
3300             fjptrH             = (jnrlistH>=0) ? f+j_coord_offsetH : scratch;
3301
3302             gmx_mm256_decrement_4rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,
3303                                                       fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,
3304                                                       fjx2,fjy2,fjz2,fjx3,fjy3,fjz3);
3305
3306             /* Inner loop uses 1014 flops */
3307         }
3308
3309         /* End of innermost loop */
3310
3311         gmx_mm256_update_iforce_4atom_swizzle_ps(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
3312                                                  f+i_coord_offset,fshift+i_shift_offset);
3313
3314         /* Increment number of inner iterations */
3315         inneriter                  += j_index_end - j_index_start;
3316
3317         /* Outer loop uses 24 flops */
3318     }
3319
3320     /* Increment number of outer iterations */
3321     outeriter        += nri;
3322
3323     /* Update outer/inner flops */
3324
3325     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4W4_F,outeriter*24 + inneriter*1014);
3326 }