src/gmxlib/nonbonded/nb_kernel_avx_256_single/nb_kernel_ElecEwSw_VdwNone_GeomW4W4_avx_256_single.c

   1 /*
   2  * Note: this file was generated by the Gromacs avx_256_single kernel generator.
   3  *
   4  *                This source code is part of
   5  *
   6  *                 G   R   O   M   A   C   S
   7  *
   8  * Copyright (c) 2001-2012, The GROMACS Development Team
   9  *
  10  * Gromacs is a library for molecular simulation and trajectory analysis,
  11  * written by Erik Lindahl, David van der Spoel, Berk Hess, and others - for
  12  * a full list of developers and information, check out http://www.gromacs.org
  13  *
  14  * This program is free software; you can redistribute it and/or modify it under
  15  * the terms of the GNU Lesser General Public License as published by the Free
  16  * Software Foundation; either version 2 of the License, or (at your option) any
  17  * later version.
  18  *
  19  * To help fund GROMACS development, we humbly ask that you cite
  20  * the papers people have written on it - you can find them on the website.
  21  */
  22 #ifdef HAVE_CONFIG_H
  23 #include <config.h>
  24 #endif
  25
  26 #include <math.h>
  27
  28 #include "../nb_kernel.h"
  29 #include "types/simple.h"
  30 #include "vec.h"
  31 #include "nrnb.h"
  32
  33 #include "gmx_math_x86_avx_256_single.h"
  34 #include "kernelutil_x86_avx_256_single.h"
  35
  36 /*
  37  * Gromacs nonbonded kernel:   nb_kernel_ElecEwSw_VdwNone_GeomW4W4_VF_avx_256_single
  38  * Electrostatics interaction: Ewald
  39  * VdW interaction:            None
  40  * Geometry:                   Water4-Water4
  41  * Calculate force/pot:        PotentialAndForce
  42  */
  43 void
  44 nb_kernel_ElecEwSw_VdwNone_GeomW4W4_VF_avx_256_single
  45                     (t_nblist * gmx_restrict                nlist,
  46                      rvec * gmx_restrict                    xx,
  47                      rvec * gmx_restrict                    ff,
  48                      t_forcerec * gmx_restrict              fr,
  49                      t_mdatoms * gmx_restrict               mdatoms,
  50                      nb_kernel_data_t * gmx_restrict        kernel_data,
  51                      t_nrnb * gmx_restrict                  nrnb)
  52 {
  53     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
  54      * just 0 for non-waters.
  55      * Suffixes A,B,C,D,E,F,G,H refer to j loop unrolling done with AVX, e.g. for the eight different
  56      * jnr indices corresponding to data put in the four positions in the SIMD register.
  57      */
  58     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
  59     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
  60     int              jnrA,jnrB,jnrC,jnrD;
  61     int              jnrE,jnrF,jnrG,jnrH;
  62     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
  63     int              jnrlistE,jnrlistF,jnrlistG,jnrlistH;
  64     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
  65     int              j_coord_offsetE,j_coord_offsetF,j_coord_offsetG,j_coord_offsetH;
  66     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  67     real             rcutoff_scalar;
  68     real             *shiftvec,*fshift,*x,*f;
  69     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD,*fjptrE,*fjptrF,*fjptrG,*fjptrH;
  70     real             scratch[4*DIM];
  71     __m256           tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
  72     real *           vdwioffsetptr1;
  73     __m256           ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
  74     real *           vdwioffsetptr2;
  75     __m256           ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
  76     real *           vdwioffsetptr3;
  77     __m256           ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
  78     int              vdwjidx1A,vdwjidx1B,vdwjidx1C,vdwjidx1D,vdwjidx1E,vdwjidx1F,vdwjidx1G,vdwjidx1H;
  79     __m256           jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
  80     int              vdwjidx2A,vdwjidx2B,vdwjidx2C,vdwjidx2D,vdwjidx2E,vdwjidx2F,vdwjidx2G,vdwjidx2H;
  81     __m256           jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
  82     int              vdwjidx3A,vdwjidx3B,vdwjidx3C,vdwjidx3D,vdwjidx3E,vdwjidx3F,vdwjidx3G,vdwjidx3H;
  83     __m256           jx3,jy3,jz3,fjx3,fjy3,fjz3,jq3,isaj3;
  84     __m256           dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11;
  85     __m256           dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12;
  86     __m256           dx13,dy13,dz13,rsq13,rinv13,rinvsq13,r13,qq13,c6_13,c12_13;
  87     __m256           dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21;
  88     __m256           dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22;
  89     __m256           dx23,dy23,dz23,rsq23,rinv23,rinvsq23,r23,qq23,c6_23,c12_23;
  90     __m256           dx31,dy31,dz31,rsq31,rinv31,rinvsq31,r31,qq31,c6_31,c12_31;
  91     __m256           dx32,dy32,dz32,rsq32,rinv32,rinvsq32,r32,qq32,c6_32,c12_32;
  92     __m256           dx33,dy33,dz33,rsq33,rinv33,rinvsq33,r33,qq33,c6_33,c12_33;
  93     __m256           velec,felec,velecsum,facel,crf,krf,krf2;
  94     real             *charge;
  95     __m256i          ewitab;
  96     __m128i          ewitab_lo,ewitab_hi;
  97     __m256           ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
  98     __m256           beta,beta2,beta3,zeta2,pmecorrF,pmecorrV,rinv3;
  99     real             *ewtab;
 100     __m256           rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
 101     real             rswitch_scalar,d_scalar;
 102     __m256           dummy_mask,cutoff_mask;
 103     __m256           signbit = _mm256_castsi256_ps( _mm256_set1_epi32(0x80000000) );
 104     __m256           one     = _mm256_set1_ps(1.0);
 105     __m256           two     = _mm256_set1_ps(2.0);
 106     x                = xx[0];
 107     f                = ff[0];
 108
 109     nri              = nlist->nri;
 110     iinr             = nlist->iinr;
 111     jindex           = nlist->jindex;
 112     jjnr             = nlist->jjnr;
 113     shiftidx         = nlist->shift;
 114     gid              = nlist->gid;
 115     shiftvec         = fr->shift_vec[0];
 116     fshift           = fr->fshift[0];
 117     facel            = _mm256_set1_ps(fr->epsfac);
 118     charge           = mdatoms->chargeA;
 119
 120     sh_ewald         = _mm256_set1_ps(fr->ic->sh_ewald);
 121     beta             = _mm256_set1_ps(fr->ic->ewaldcoeff);
 122     beta2            = _mm256_mul_ps(beta,beta);
 123     beta3            = _mm256_mul_ps(beta,beta2);
 124
 125     ewtab            = fr->ic->tabq_coul_FDV0;
 126     ewtabscale       = _mm256_set1_ps(fr->ic->tabq_scale);
 127     ewtabhalfspace   = _mm256_set1_ps(0.5/fr->ic->tabq_scale);
 128
 129     /* Setup water-specific parameters */
 130     inr              = nlist->iinr[0];
 131     iq1              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+1]));
 132     iq2              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+2]));
 133     iq3              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+3]));
 134
 135     jq1              = _mm256_set1_ps(charge[inr+1]);
 136     jq2              = _mm256_set1_ps(charge[inr+2]);
 137     jq3              = _mm256_set1_ps(charge[inr+3]);
 138     qq11             = _mm256_mul_ps(iq1,jq1);
 139     qq12             = _mm256_mul_ps(iq1,jq2);
 140     qq13             = _mm256_mul_ps(iq1,jq3);
 141     qq21             = _mm256_mul_ps(iq2,jq1);
 142     qq22             = _mm256_mul_ps(iq2,jq2);
 143     qq23             = _mm256_mul_ps(iq2,jq3);
 144     qq31             = _mm256_mul_ps(iq3,jq1);
 145     qq32             = _mm256_mul_ps(iq3,jq2);
 146     qq33             = _mm256_mul_ps(iq3,jq3);
 147
 148     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 149     rcutoff_scalar   = fr->rcoulomb;
 150     rcutoff          = _mm256_set1_ps(rcutoff_scalar);
 151     rcutoff2         = _mm256_mul_ps(rcutoff,rcutoff);
 152
 153     rswitch_scalar   = fr->rcoulomb_switch;
 154     rswitch          = _mm256_set1_ps(rswitch_scalar);
 155     /* Setup switch parameters */
 156     d_scalar         = rcutoff_scalar-rswitch_scalar;
 157     d                = _mm256_set1_ps(d_scalar);
 158     swV3             = _mm256_set1_ps(-10.0/(d_scalar*d_scalar*d_scalar));
 159     swV4             = _mm256_set1_ps( 15.0/(d_scalar*d_scalar*d_scalar*d_scalar));
 160     swV5             = _mm256_set1_ps( -6.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
 161     swF2             = _mm256_set1_ps(-30.0/(d_scalar*d_scalar*d_scalar));
 162     swF3             = _mm256_set1_ps( 60.0/(d_scalar*d_scalar*d_scalar*d_scalar));
 163     swF4             = _mm256_set1_ps(-30.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
 164
 165     /* Avoid stupid compiler warnings */
 166     jnrA = jnrB = jnrC = jnrD = jnrE = jnrF = jnrG = jnrH = 0;
 167     j_coord_offsetA = 0;
 168     j_coord_offsetB = 0;
 169     j_coord_offsetC = 0;
 170     j_coord_offsetD = 0;
 171     j_coord_offsetE = 0;
 172     j_coord_offsetF = 0;
 173     j_coord_offsetG = 0;
 174     j_coord_offsetH = 0;
 175
 176     outeriter        = 0;
 177     inneriter        = 0;
 178
 179     for(iidx=0;iidx<4*DIM;iidx++)
 180     {
 181         scratch[iidx] = 0.0;
 182     }
 183
 184     /* Start outer loop over neighborlists */
 185     for(iidx=0; iidx<nri; iidx++)
 186     {
 187         /* Load shift vector for this list */
 188         i_shift_offset   = DIM*shiftidx[iidx];
 189
 190         /* Load limits for loop over neighbors */
 191         j_index_start    = jindex[iidx];
 192         j_index_end      = jindex[iidx+1];
 193
 194         /* Get outer coordinate index */
 195         inr              = iinr[iidx];
 196         i_coord_offset   = DIM*inr;
 197
 198         /* Load i particle coords and add shift vector */
 199         gmx_mm256_load_shift_and_3rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset+DIM,
 200                                                     &ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
 201
 202         fix1             = _mm256_setzero_ps();
 203         fiy1             = _mm256_setzero_ps();
 204         fiz1             = _mm256_setzero_ps();
 205         fix2             = _mm256_setzero_ps();
 206         fiy2             = _mm256_setzero_ps();
 207         fiz2             = _mm256_setzero_ps();
 208         fix3             = _mm256_setzero_ps();
 209         fiy3             = _mm256_setzero_ps();
 210         fiz3             = _mm256_setzero_ps();
 211
 212         /* Reset potential sums */
 213         velecsum         = _mm256_setzero_ps();
 214
 215         /* Start inner kernel loop */
 216         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+7]>=0; jidx+=8)
 217         {
 218
 219             /* Get j neighbor index, and coordinate index */
 220             jnrA             = jjnr[jidx];
 221             jnrB             = jjnr[jidx+1];
 222             jnrC             = jjnr[jidx+2];
 223             jnrD             = jjnr[jidx+3];
 224             jnrE             = jjnr[jidx+4];
 225             jnrF             = jjnr[jidx+5];
 226             jnrG             = jjnr[jidx+6];
 227             jnrH             = jjnr[jidx+7];
 228             j_coord_offsetA  = DIM*jnrA;
 229             j_coord_offsetB  = DIM*jnrB;
 230             j_coord_offsetC  = DIM*jnrC;
 231             j_coord_offsetD  = DIM*jnrD;
 232             j_coord_offsetE  = DIM*jnrE;
 233             j_coord_offsetF  = DIM*jnrF;
 234             j_coord_offsetG  = DIM*jnrG;
 235             j_coord_offsetH  = DIM*jnrH;
 236
 237             /* load j atom coordinates */
 238             gmx_mm256_load_3rvec_8ptr_swizzle_ps(x+j_coord_offsetA+DIM,x+j_coord_offsetB+DIM,
 239                                                  x+j_coord_offsetC+DIM,x+j_coord_offsetD+DIM,
 240                                                  x+j_coord_offsetE+DIM,x+j_coord_offsetF+DIM,
 241                                                  x+j_coord_offsetG+DIM,x+j_coord_offsetH+DIM,
 242                                                  &jx1,&jy1,&jz1,&jx2,&jy2,&jz2,&jx3,&jy3,&jz3);
 243
 244             /* Calculate displacement vector */
 245             dx11             = _mm256_sub_ps(ix1,jx1);
 246             dy11             = _mm256_sub_ps(iy1,jy1);
 247             dz11             = _mm256_sub_ps(iz1,jz1);
 248             dx12             = _mm256_sub_ps(ix1,jx2);
 249             dy12             = _mm256_sub_ps(iy1,jy2);
 250             dz12             = _mm256_sub_ps(iz1,jz2);
 251             dx13             = _mm256_sub_ps(ix1,jx3);
 252             dy13             = _mm256_sub_ps(iy1,jy3);
 253             dz13             = _mm256_sub_ps(iz1,jz3);
 254             dx21             = _mm256_sub_ps(ix2,jx1);
 255             dy21             = _mm256_sub_ps(iy2,jy1);
 256             dz21             = _mm256_sub_ps(iz2,jz1);
 257             dx22             = _mm256_sub_ps(ix2,jx2);
 258             dy22             = _mm256_sub_ps(iy2,jy2);
 259             dz22             = _mm256_sub_ps(iz2,jz2);
 260             dx23             = _mm256_sub_ps(ix2,jx3);
 261             dy23             = _mm256_sub_ps(iy2,jy3);
 262             dz23             = _mm256_sub_ps(iz2,jz3);
 263             dx31             = _mm256_sub_ps(ix3,jx1);
 264             dy31             = _mm256_sub_ps(iy3,jy1);
 265             dz31             = _mm256_sub_ps(iz3,jz1);
 266             dx32             = _mm256_sub_ps(ix3,jx2);
 267             dy32             = _mm256_sub_ps(iy3,jy2);
 268             dz32             = _mm256_sub_ps(iz3,jz2);
 269             dx33             = _mm256_sub_ps(ix3,jx3);
 270             dy33             = _mm256_sub_ps(iy3,jy3);
 271             dz33             = _mm256_sub_ps(iz3,jz3);
 272
 273             /* Calculate squared distance and things based on it */
 274             rsq11            = gmx_mm256_calc_rsq_ps(dx11,dy11,dz11);
 275             rsq12            = gmx_mm256_calc_rsq_ps(dx12,dy12,dz12);
 276             rsq13            = gmx_mm256_calc_rsq_ps(dx13,dy13,dz13);
 277             rsq21            = gmx_mm256_calc_rsq_ps(dx21,dy21,dz21);
 278             rsq22            = gmx_mm256_calc_rsq_ps(dx22,dy22,dz22);
 279             rsq23            = gmx_mm256_calc_rsq_ps(dx23,dy23,dz23);
 280             rsq31            = gmx_mm256_calc_rsq_ps(dx31,dy31,dz31);
 281             rsq32            = gmx_mm256_calc_rsq_ps(dx32,dy32,dz32);
 282             rsq33            = gmx_mm256_calc_rsq_ps(dx33,dy33,dz33);
 283
 284             rinv11           = gmx_mm256_invsqrt_ps(rsq11);
 285             rinv12           = gmx_mm256_invsqrt_ps(rsq12);
 286             rinv13           = gmx_mm256_invsqrt_ps(rsq13);
 287             rinv21           = gmx_mm256_invsqrt_ps(rsq21);
 288             rinv22           = gmx_mm256_invsqrt_ps(rsq22);
 289             rinv23           = gmx_mm256_invsqrt_ps(rsq23);
 290             rinv31           = gmx_mm256_invsqrt_ps(rsq31);
 291             rinv32           = gmx_mm256_invsqrt_ps(rsq32);
 292             rinv33           = gmx_mm256_invsqrt_ps(rsq33);
 293
 294             rinvsq11         = _mm256_mul_ps(rinv11,rinv11);
 295             rinvsq12         = _mm256_mul_ps(rinv12,rinv12);
 296             rinvsq13         = _mm256_mul_ps(rinv13,rinv13);
 297             rinvsq21         = _mm256_mul_ps(rinv21,rinv21);
 298             rinvsq22         = _mm256_mul_ps(rinv22,rinv22);
 299             rinvsq23         = _mm256_mul_ps(rinv23,rinv23);
 300             rinvsq31         = _mm256_mul_ps(rinv31,rinv31);
 301             rinvsq32         = _mm256_mul_ps(rinv32,rinv32);
 302             rinvsq33         = _mm256_mul_ps(rinv33,rinv33);
 303
 304             fjx1             = _mm256_setzero_ps();
 305             fjy1             = _mm256_setzero_ps();
 306             fjz1             = _mm256_setzero_ps();
 307             fjx2             = _mm256_setzero_ps();
 308             fjy2             = _mm256_setzero_ps();
 309             fjz2             = _mm256_setzero_ps();
 310             fjx3             = _mm256_setzero_ps();
 311             fjy3             = _mm256_setzero_ps();
 312             fjz3             = _mm256_setzero_ps();
 313
 314             /**************************
 315              * CALCULATE INTERACTIONS *
 316              **************************/
 317
 318             if (gmx_mm256_any_lt(rsq11,rcutoff2))
 319             {
 320
 321             r11              = _mm256_mul_ps(rsq11,rinv11);
 322
 323             /* EWALD ELECTROSTATICS */
 324
 325             /* Analytical PME correction */
 326             zeta2            = _mm256_mul_ps(beta2,rsq11);
 327             rinv3            = _mm256_mul_ps(rinvsq11,rinv11);
 328             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
 329             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 330             felec            = _mm256_mul_ps(qq11,felec);
 331             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
 332             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 333             velec            = _mm256_sub_ps(rinv11,pmecorrV);
 334             velec            = _mm256_mul_ps(qq11,velec);
 335
 336             d                = _mm256_sub_ps(r11,rswitch);
 337             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 338             d2               = _mm256_mul_ps(d,d);
 339             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)))))));
 340
 341             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 342
 343             /* Evaluate switch function */
 344             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 345             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv11,_mm256_mul_ps(velec,dsw)) );
 346             velec            = _mm256_mul_ps(velec,sw);
 347             cutoff_mask      = _mm256_cmp_ps(rsq11,rcutoff2,_CMP_LT_OQ);
 348
 349             /* Update potential sum for this i atom from the interaction with this j atom. */
 350             velec            = _mm256_and_ps(velec,cutoff_mask);
 351             velecsum         = _mm256_add_ps(velecsum,velec);
 352
 353             fscal            = felec;
 354
 355             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 356
 357             /* Calculate temporary vectorial force */
 358             tx               = _mm256_mul_ps(fscal,dx11);
 359             ty               = _mm256_mul_ps(fscal,dy11);
 360             tz               = _mm256_mul_ps(fscal,dz11);
 361
 362             /* Update vectorial force */
 363             fix1             = _mm256_add_ps(fix1,tx);
 364             fiy1             = _mm256_add_ps(fiy1,ty);
 365             fiz1             = _mm256_add_ps(fiz1,tz);
 366
 367             fjx1             = _mm256_add_ps(fjx1,tx);
 368             fjy1             = _mm256_add_ps(fjy1,ty);
 369             fjz1             = _mm256_add_ps(fjz1,tz);
 370
 371             }
 372
 373             /**************************
 374              * CALCULATE INTERACTIONS *
 375              **************************/
 376
 377             if (gmx_mm256_any_lt(rsq12,rcutoff2))
 378             {
 379
 380             r12              = _mm256_mul_ps(rsq12,rinv12);
 381
 382             /* EWALD ELECTROSTATICS */
 383
 384             /* Analytical PME correction */
 385             zeta2            = _mm256_mul_ps(beta2,rsq12);
 386             rinv3            = _mm256_mul_ps(rinvsq12,rinv12);
 387             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
 388             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 389             felec            = _mm256_mul_ps(qq12,felec);
 390             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
 391             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 392             velec            = _mm256_sub_ps(rinv12,pmecorrV);
 393             velec            = _mm256_mul_ps(qq12,velec);
 394
 395             d                = _mm256_sub_ps(r12,rswitch);
 396             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 397             d2               = _mm256_mul_ps(d,d);
 398             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)))))));
 399
 400             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 401
 402             /* Evaluate switch function */
 403             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 404             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv12,_mm256_mul_ps(velec,dsw)) );
 405             velec            = _mm256_mul_ps(velec,sw);
 406             cutoff_mask      = _mm256_cmp_ps(rsq12,rcutoff2,_CMP_LT_OQ);
 407
 408             /* Update potential sum for this i atom from the interaction with this j atom. */
 409             velec            = _mm256_and_ps(velec,cutoff_mask);
 410             velecsum         = _mm256_add_ps(velecsum,velec);
 411
 412             fscal            = felec;
 413
 414             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 415
 416             /* Calculate temporary vectorial force */
 417             tx               = _mm256_mul_ps(fscal,dx12);
 418             ty               = _mm256_mul_ps(fscal,dy12);
 419             tz               = _mm256_mul_ps(fscal,dz12);
 420
 421             /* Update vectorial force */
 422             fix1             = _mm256_add_ps(fix1,tx);
 423             fiy1             = _mm256_add_ps(fiy1,ty);
 424             fiz1             = _mm256_add_ps(fiz1,tz);
 425
 426             fjx2             = _mm256_add_ps(fjx2,tx);
 427             fjy2             = _mm256_add_ps(fjy2,ty);
 428             fjz2             = _mm256_add_ps(fjz2,tz);
 429
 430             }
 431
 432             /**************************
 433              * CALCULATE INTERACTIONS *
 434              **************************/
 435
 436             if (gmx_mm256_any_lt(rsq13,rcutoff2))
 437             {
 438
 439             r13              = _mm256_mul_ps(rsq13,rinv13);
 440
 441             /* EWALD ELECTROSTATICS */
 442
 443             /* Analytical PME correction */
 444             zeta2            = _mm256_mul_ps(beta2,rsq13);
 445             rinv3            = _mm256_mul_ps(rinvsq13,rinv13);
 446             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
 447             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 448             felec            = _mm256_mul_ps(qq13,felec);
 449             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
 450             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 451             velec            = _mm256_sub_ps(rinv13,pmecorrV);
 452             velec            = _mm256_mul_ps(qq13,velec);
 453
 454             d                = _mm256_sub_ps(r13,rswitch);
 455             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 456             d2               = _mm256_mul_ps(d,d);
 457             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)))))));
 458
 459             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 460
 461             /* Evaluate switch function */
 462             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 463             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv13,_mm256_mul_ps(velec,dsw)) );
 464             velec            = _mm256_mul_ps(velec,sw);
 465             cutoff_mask      = _mm256_cmp_ps(rsq13,rcutoff2,_CMP_LT_OQ);
 466
 467             /* Update potential sum for this i atom from the interaction with this j atom. */
 468             velec            = _mm256_and_ps(velec,cutoff_mask);
 469             velecsum         = _mm256_add_ps(velecsum,velec);
 470
 471             fscal            = felec;
 472
 473             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 474
 475             /* Calculate temporary vectorial force */
 476             tx               = _mm256_mul_ps(fscal,dx13);
 477             ty               = _mm256_mul_ps(fscal,dy13);
 478             tz               = _mm256_mul_ps(fscal,dz13);
 479
 480             /* Update vectorial force */
 481             fix1             = _mm256_add_ps(fix1,tx);
 482             fiy1             = _mm256_add_ps(fiy1,ty);
 483             fiz1             = _mm256_add_ps(fiz1,tz);
 484
 485             fjx3             = _mm256_add_ps(fjx3,tx);
 486             fjy3             = _mm256_add_ps(fjy3,ty);
 487             fjz3             = _mm256_add_ps(fjz3,tz);
 488
 489             }
 490
 491             /**************************
 492              * CALCULATE INTERACTIONS *
 493              **************************/
 494
 495             if (gmx_mm256_any_lt(rsq21,rcutoff2))
 496             {
 497
 498             r21              = _mm256_mul_ps(rsq21,rinv21);
 499
 500             /* EWALD ELECTROSTATICS */
 501
 502             /* Analytical PME correction */
 503             zeta2            = _mm256_mul_ps(beta2,rsq21);
 504             rinv3            = _mm256_mul_ps(rinvsq21,rinv21);
 505             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
 506             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 507             felec            = _mm256_mul_ps(qq21,felec);
 508             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
 509             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 510             velec            = _mm256_sub_ps(rinv21,pmecorrV);
 511             velec            = _mm256_mul_ps(qq21,velec);
 512
 513             d                = _mm256_sub_ps(r21,rswitch);
 514             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 515             d2               = _mm256_mul_ps(d,d);
 516             sw               = _mm256_add_ps(one,_mm256_mul_ps(d2,_mm256_mul_ps(d,_mm256_add_ps(swV3,_mm256_mul_ps(d,_mm256_add_ps(swV4,_mm256_mul_ps(d,swV5)))))));
 517
 518             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 519
 520             /* Evaluate switch function */
 521             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 522             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv21,_mm256_mul_ps(velec,dsw)) );
 523             velec            = _mm256_mul_ps(velec,sw);
 524             cutoff_mask      = _mm256_cmp_ps(rsq21,rcutoff2,_CMP_LT_OQ);
 525
 526             /* Update potential sum for this i atom from the interaction with this j atom. */
 527             velec            = _mm256_and_ps(velec,cutoff_mask);
 528             velecsum         = _mm256_add_ps(velecsum,velec);
 529
 530             fscal            = felec;
 531
 532             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 533
 534             /* Calculate temporary vectorial force */
 535             tx               = _mm256_mul_ps(fscal,dx21);
 536             ty               = _mm256_mul_ps(fscal,dy21);
 537             tz               = _mm256_mul_ps(fscal,dz21);
 538
 539             /* Update vectorial force */
 540             fix2             = _mm256_add_ps(fix2,tx);
 541             fiy2             = _mm256_add_ps(fiy2,ty);
 542             fiz2             = _mm256_add_ps(fiz2,tz);
 543
 544             fjx1             = _mm256_add_ps(fjx1,tx);
 545             fjy1             = _mm256_add_ps(fjy1,ty);
 546             fjz1             = _mm256_add_ps(fjz1,tz);
 547
 548             }
 549
 550             /**************************
 551              * CALCULATE INTERACTIONS *
 552              **************************/
 553
 554             if (gmx_mm256_any_lt(rsq22,rcutoff2))
 555             {
 556
 557             r22              = _mm256_mul_ps(rsq22,rinv22);
 558
 559             /* EWALD ELECTROSTATICS */
 560
 561             /* Analytical PME correction */
 562             zeta2            = _mm256_mul_ps(beta2,rsq22);
 563             rinv3            = _mm256_mul_ps(rinvsq22,rinv22);
 564             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
 565             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 566             felec            = _mm256_mul_ps(qq22,felec);
 567             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
 568             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 569             velec            = _mm256_sub_ps(rinv22,pmecorrV);
 570             velec            = _mm256_mul_ps(qq22,velec);
 571
 572             d                = _mm256_sub_ps(r22,rswitch);
 573             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 574             d2               = _mm256_mul_ps(d,d);
 575             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)))))));
 576
 577             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 578
 579             /* Evaluate switch function */
 580             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 581             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv22,_mm256_mul_ps(velec,dsw)) );
 582             velec            = _mm256_mul_ps(velec,sw);
 583             cutoff_mask      = _mm256_cmp_ps(rsq22,rcutoff2,_CMP_LT_OQ);
 584
 585             /* Update potential sum for this i atom from the interaction with this j atom. */
 586             velec            = _mm256_and_ps(velec,cutoff_mask);
 587             velecsum         = _mm256_add_ps(velecsum,velec);
 588
 589             fscal            = felec;
 590
 591             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 592
 593             /* Calculate temporary vectorial force */
 594             tx               = _mm256_mul_ps(fscal,dx22);
 595             ty               = _mm256_mul_ps(fscal,dy22);
 596             tz               = _mm256_mul_ps(fscal,dz22);
 597
 598             /* Update vectorial force */
 599             fix2             = _mm256_add_ps(fix2,tx);
 600             fiy2             = _mm256_add_ps(fiy2,ty);
 601             fiz2             = _mm256_add_ps(fiz2,tz);
 602
 603             fjx2             = _mm256_add_ps(fjx2,tx);
 604             fjy2             = _mm256_add_ps(fjy2,ty);
 605             fjz2             = _mm256_add_ps(fjz2,tz);
 606
 607             }
 608
 609             /**************************
 610              * CALCULATE INTERACTIONS *
 611              **************************/
 612
 613             if (gmx_mm256_any_lt(rsq23,rcutoff2))
 614             {
 615
 616             r23              = _mm256_mul_ps(rsq23,rinv23);
 617
 618             /* EWALD ELECTROSTATICS */
 619
 620             /* Analytical PME correction */
 621             zeta2            = _mm256_mul_ps(beta2,rsq23);
 622             rinv3            = _mm256_mul_ps(rinvsq23,rinv23);
 623             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
 624             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 625             felec            = _mm256_mul_ps(qq23,felec);
 626             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
 627             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 628             velec            = _mm256_sub_ps(rinv23,pmecorrV);
 629             velec            = _mm256_mul_ps(qq23,velec);
 630
 631             d                = _mm256_sub_ps(r23,rswitch);
 632             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 633             d2               = _mm256_mul_ps(d,d);
 634             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)))))));
 635
 636             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 637
 638             /* Evaluate switch function */
 639             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 640             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv23,_mm256_mul_ps(velec,dsw)) );
 641             velec            = _mm256_mul_ps(velec,sw);
 642             cutoff_mask      = _mm256_cmp_ps(rsq23,rcutoff2,_CMP_LT_OQ);
 643
 644             /* Update potential sum for this i atom from the interaction with this j atom. */
 645             velec            = _mm256_and_ps(velec,cutoff_mask);
 646             velecsum         = _mm256_add_ps(velecsum,velec);
 647
 648             fscal            = felec;
 649
 650             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 651
 652             /* Calculate temporary vectorial force */
 653             tx               = _mm256_mul_ps(fscal,dx23);
 654             ty               = _mm256_mul_ps(fscal,dy23);
 655             tz               = _mm256_mul_ps(fscal,dz23);
 656
 657             /* Update vectorial force */
 658             fix2             = _mm256_add_ps(fix2,tx);
 659             fiy2             = _mm256_add_ps(fiy2,ty);
 660             fiz2             = _mm256_add_ps(fiz2,tz);
 661
 662             fjx3             = _mm256_add_ps(fjx3,tx);
 663             fjy3             = _mm256_add_ps(fjy3,ty);
 664             fjz3             = _mm256_add_ps(fjz3,tz);
 665
 666             }
 667
 668             /**************************
 669              * CALCULATE INTERACTIONS *
 670              **************************/
 671
 672             if (gmx_mm256_any_lt(rsq31,rcutoff2))
 673             {
 674
 675             r31              = _mm256_mul_ps(rsq31,rinv31);
 676
 677             /* EWALD ELECTROSTATICS */
 678
 679             /* Analytical PME correction */
 680             zeta2            = _mm256_mul_ps(beta2,rsq31);
 681             rinv3            = _mm256_mul_ps(rinvsq31,rinv31);
 682             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
 683             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 684             felec            = _mm256_mul_ps(qq31,felec);
 685             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
 686             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 687             velec            = _mm256_sub_ps(rinv31,pmecorrV);
 688             velec            = _mm256_mul_ps(qq31,velec);
 689
 690             d                = _mm256_sub_ps(r31,rswitch);
 691             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 692             d2               = _mm256_mul_ps(d,d);
 693             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)))))));
 694
 695             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 696
 697             /* Evaluate switch function */
 698             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 699             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv31,_mm256_mul_ps(velec,dsw)) );
 700             velec            = _mm256_mul_ps(velec,sw);
 701             cutoff_mask      = _mm256_cmp_ps(rsq31,rcutoff2,_CMP_LT_OQ);
 702
 703             /* Update potential sum for this i atom from the interaction with this j atom. */
 704             velec            = _mm256_and_ps(velec,cutoff_mask);
 705             velecsum         = _mm256_add_ps(velecsum,velec);
 706
 707             fscal            = felec;
 708
 709             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 710
 711             /* Calculate temporary vectorial force */
 712             tx               = _mm256_mul_ps(fscal,dx31);
 713             ty               = _mm256_mul_ps(fscal,dy31);
 714             tz               = _mm256_mul_ps(fscal,dz31);
 715
 716             /* Update vectorial force */
 717             fix3             = _mm256_add_ps(fix3,tx);
 718             fiy3             = _mm256_add_ps(fiy3,ty);
 719             fiz3             = _mm256_add_ps(fiz3,tz);
 720
 721             fjx1             = _mm256_add_ps(fjx1,tx);
 722             fjy1             = _mm256_add_ps(fjy1,ty);
 723             fjz1             = _mm256_add_ps(fjz1,tz);
 724
 725             }
 726
 727             /**************************
 728              * CALCULATE INTERACTIONS *
 729              **************************/
 730
 731             if (gmx_mm256_any_lt(rsq32,rcutoff2))
 732             {
 733
 734             r32              = _mm256_mul_ps(rsq32,rinv32);
 735
 736             /* EWALD ELECTROSTATICS */
 737
 738             /* Analytical PME correction */
 739             zeta2            = _mm256_mul_ps(beta2,rsq32);
 740             rinv3            = _mm256_mul_ps(rinvsq32,rinv32);
 741             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
 742             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 743             felec            = _mm256_mul_ps(qq32,felec);
 744             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
 745             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 746             velec            = _mm256_sub_ps(rinv32,pmecorrV);
 747             velec            = _mm256_mul_ps(qq32,velec);
 748
 749             d                = _mm256_sub_ps(r32,rswitch);
 750             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 751             d2               = _mm256_mul_ps(d,d);
 752             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)))))));
 753
 754             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 755
 756             /* Evaluate switch function */
 757             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 758             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv32,_mm256_mul_ps(velec,dsw)) );
 759             velec            = _mm256_mul_ps(velec,sw);
 760             cutoff_mask      = _mm256_cmp_ps(rsq32,rcutoff2,_CMP_LT_OQ);
 761
 762             /* Update potential sum for this i atom from the interaction with this j atom. */
 763             velec            = _mm256_and_ps(velec,cutoff_mask);
 764             velecsum         = _mm256_add_ps(velecsum,velec);
 765
 766             fscal            = felec;
 767
 768             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 769
 770             /* Calculate temporary vectorial force */
 771             tx               = _mm256_mul_ps(fscal,dx32);
 772             ty               = _mm256_mul_ps(fscal,dy32);
 773             tz               = _mm256_mul_ps(fscal,dz32);
 774
 775             /* Update vectorial force */
 776             fix3             = _mm256_add_ps(fix3,tx);
 777             fiy3             = _mm256_add_ps(fiy3,ty);
 778             fiz3             = _mm256_add_ps(fiz3,tz);
 779
 780             fjx2             = _mm256_add_ps(fjx2,tx);
 781             fjy2             = _mm256_add_ps(fjy2,ty);
 782             fjz2             = _mm256_add_ps(fjz2,tz);
 783
 784             }
 785
 786             /**************************
 787              * CALCULATE INTERACTIONS *
 788              **************************/
 789
 790             if (gmx_mm256_any_lt(rsq33,rcutoff2))
 791             {
 792
 793             r33              = _mm256_mul_ps(rsq33,rinv33);
 794
 795             /* EWALD ELECTROSTATICS */
 796
 797             /* Analytical PME correction */
 798             zeta2            = _mm256_mul_ps(beta2,rsq33);
 799             rinv3            = _mm256_mul_ps(rinvsq33,rinv33);
 800             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
 801             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 802             felec            = _mm256_mul_ps(qq33,felec);
 803             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
 804             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 805             velec            = _mm256_sub_ps(rinv33,pmecorrV);
 806             velec            = _mm256_mul_ps(qq33,velec);
 807
 808             d                = _mm256_sub_ps(r33,rswitch);
 809             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 810             d2               = _mm256_mul_ps(d,d);
 811             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)))))));
 812
 813             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 814
 815             /* Evaluate switch function */
 816             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 817             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv33,_mm256_mul_ps(velec,dsw)) );
 818             velec            = _mm256_mul_ps(velec,sw);
 819             cutoff_mask      = _mm256_cmp_ps(rsq33,rcutoff2,_CMP_LT_OQ);
 820
 821             /* Update potential sum for this i atom from the interaction with this j atom. */
 822             velec            = _mm256_and_ps(velec,cutoff_mask);
 823             velecsum         = _mm256_add_ps(velecsum,velec);
 824
 825             fscal            = felec;
 826
 827             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 828
 829             /* Calculate temporary vectorial force */
 830             tx               = _mm256_mul_ps(fscal,dx33);
 831             ty               = _mm256_mul_ps(fscal,dy33);
 832             tz               = _mm256_mul_ps(fscal,dz33);
 833
 834             /* Update vectorial force */
 835             fix3             = _mm256_add_ps(fix3,tx);
 836             fiy3             = _mm256_add_ps(fiy3,ty);
 837             fiz3             = _mm256_add_ps(fiz3,tz);
 838
 839             fjx3             = _mm256_add_ps(fjx3,tx);
 840             fjy3             = _mm256_add_ps(fjy3,ty);
 841             fjz3             = _mm256_add_ps(fjz3,tz);
 842
 843             }
 844
 845             fjptrA             = f+j_coord_offsetA;
 846             fjptrB             = f+j_coord_offsetB;
 847             fjptrC             = f+j_coord_offsetC;
 848             fjptrD             = f+j_coord_offsetD;
 849             fjptrE             = f+j_coord_offsetE;
 850             fjptrF             = f+j_coord_offsetF;
 851             fjptrG             = f+j_coord_offsetG;
 852             fjptrH             = f+j_coord_offsetH;
 853
 854             gmx_mm256_decrement_3rvec_8ptr_swizzle_ps(fjptrA+DIM,fjptrB+DIM,fjptrC+DIM,fjptrD+DIM,
 855                                                       fjptrE+DIM,fjptrF+DIM,fjptrG+DIM,fjptrH+DIM,
 856                                                       fjx1,fjy1,fjz1,fjx2,fjy2,fjz2,fjx3,fjy3,fjz3);
 857
 858             /* Inner loop uses 972 flops */
 859         }
 860
 861         if(jidx<j_index_end)
 862         {
 863
 864             /* Get j neighbor index, and coordinate index */
 865             jnrlistA         = jjnr[jidx];
 866             jnrlistB         = jjnr[jidx+1];
 867             jnrlistC         = jjnr[jidx+2];
 868             jnrlistD         = jjnr[jidx+3];
 869             jnrlistE         = jjnr[jidx+4];
 870             jnrlistF         = jjnr[jidx+5];
 871             jnrlistG         = jjnr[jidx+6];
 872             jnrlistH         = jjnr[jidx+7];
 873             /* Sign of each element will be negative for non-real atoms.
 874              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
 875              * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
 876              */
 877             dummy_mask = gmx_mm256_set_m128(gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx+4)),_mm_setzero_si128())),
 878                                             gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128())));
 879
 880             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
 881             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
 882             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
 883             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
 884             jnrE       = (jnrlistE>=0) ? jnrlistE : 0;
 885             jnrF       = (jnrlistF>=0) ? jnrlistF : 0;
 886             jnrG       = (jnrlistG>=0) ? jnrlistG : 0;
 887             jnrH       = (jnrlistH>=0) ? jnrlistH : 0;
 888             j_coord_offsetA  = DIM*jnrA;
 889             j_coord_offsetB  = DIM*jnrB;
 890             j_coord_offsetC  = DIM*jnrC;
 891             j_coord_offsetD  = DIM*jnrD;
 892             j_coord_offsetE  = DIM*jnrE;
 893             j_coord_offsetF  = DIM*jnrF;
 894             j_coord_offsetG  = DIM*jnrG;
 895             j_coord_offsetH  = DIM*jnrH;
 896
 897             /* load j atom coordinates */
 898             gmx_mm256_load_3rvec_8ptr_swizzle_ps(x+j_coord_offsetA+DIM,x+j_coord_offsetB+DIM,
 899                                                  x+j_coord_offsetC+DIM,x+j_coord_offsetD+DIM,
 900                                                  x+j_coord_offsetE+DIM,x+j_coord_offsetF+DIM,
 901                                                  x+j_coord_offsetG+DIM,x+j_coord_offsetH+DIM,
 902                                                  &jx1,&jy1,&jz1,&jx2,&jy2,&jz2,&jx3,&jy3,&jz3);
 903
 904             /* Calculate displacement vector */
 905             dx11             = _mm256_sub_ps(ix1,jx1);
 906             dy11             = _mm256_sub_ps(iy1,jy1);
 907             dz11             = _mm256_sub_ps(iz1,jz1);
 908             dx12             = _mm256_sub_ps(ix1,jx2);
 909             dy12             = _mm256_sub_ps(iy1,jy2);
 910             dz12             = _mm256_sub_ps(iz1,jz2);
 911             dx13             = _mm256_sub_ps(ix1,jx3);
 912             dy13             = _mm256_sub_ps(iy1,jy3);
 913             dz13             = _mm256_sub_ps(iz1,jz3);
 914             dx21             = _mm256_sub_ps(ix2,jx1);
 915             dy21             = _mm256_sub_ps(iy2,jy1);
 916             dz21             = _mm256_sub_ps(iz2,jz1);
 917             dx22             = _mm256_sub_ps(ix2,jx2);
 918             dy22             = _mm256_sub_ps(iy2,jy2);
 919             dz22             = _mm256_sub_ps(iz2,jz2);
 920             dx23             = _mm256_sub_ps(ix2,jx3);
 921             dy23             = _mm256_sub_ps(iy2,jy3);
 922             dz23             = _mm256_sub_ps(iz2,jz3);
 923             dx31             = _mm256_sub_ps(ix3,jx1);
 924             dy31             = _mm256_sub_ps(iy3,jy1);
 925             dz31             = _mm256_sub_ps(iz3,jz1);
 926             dx32             = _mm256_sub_ps(ix3,jx2);
 927             dy32             = _mm256_sub_ps(iy3,jy2);
 928             dz32             = _mm256_sub_ps(iz3,jz2);
 929             dx33             = _mm256_sub_ps(ix3,jx3);
 930             dy33             = _mm256_sub_ps(iy3,jy3);
 931             dz33             = _mm256_sub_ps(iz3,jz3);
 932
 933             /* Calculate squared distance and things based on it */
 934             rsq11            = gmx_mm256_calc_rsq_ps(dx11,dy11,dz11);
 935             rsq12            = gmx_mm256_calc_rsq_ps(dx12,dy12,dz12);
 936             rsq13            = gmx_mm256_calc_rsq_ps(dx13,dy13,dz13);
 937             rsq21            = gmx_mm256_calc_rsq_ps(dx21,dy21,dz21);
 938             rsq22            = gmx_mm256_calc_rsq_ps(dx22,dy22,dz22);
 939             rsq23            = gmx_mm256_calc_rsq_ps(dx23,dy23,dz23);
 940             rsq31            = gmx_mm256_calc_rsq_ps(dx31,dy31,dz31);
 941             rsq32            = gmx_mm256_calc_rsq_ps(dx32,dy32,dz32);
 942             rsq33            = gmx_mm256_calc_rsq_ps(dx33,dy33,dz33);
 943
 944             rinv11           = gmx_mm256_invsqrt_ps(rsq11);
 945             rinv12           = gmx_mm256_invsqrt_ps(rsq12);
 946             rinv13           = gmx_mm256_invsqrt_ps(rsq13);
 947             rinv21           = gmx_mm256_invsqrt_ps(rsq21);
 948             rinv22           = gmx_mm256_invsqrt_ps(rsq22);
 949             rinv23           = gmx_mm256_invsqrt_ps(rsq23);
 950             rinv31           = gmx_mm256_invsqrt_ps(rsq31);
 951             rinv32           = gmx_mm256_invsqrt_ps(rsq32);
 952             rinv33           = gmx_mm256_invsqrt_ps(rsq33);
 953
 954             rinvsq11         = _mm256_mul_ps(rinv11,rinv11);
 955             rinvsq12         = _mm256_mul_ps(rinv12,rinv12);
 956             rinvsq13         = _mm256_mul_ps(rinv13,rinv13);
 957             rinvsq21         = _mm256_mul_ps(rinv21,rinv21);
 958             rinvsq22         = _mm256_mul_ps(rinv22,rinv22);
 959             rinvsq23         = _mm256_mul_ps(rinv23,rinv23);
 960             rinvsq31         = _mm256_mul_ps(rinv31,rinv31);
 961             rinvsq32         = _mm256_mul_ps(rinv32,rinv32);
 962             rinvsq33         = _mm256_mul_ps(rinv33,rinv33);
 963
 964             fjx1             = _mm256_setzero_ps();
 965             fjy1             = _mm256_setzero_ps();
 966             fjz1             = _mm256_setzero_ps();
 967             fjx2             = _mm256_setzero_ps();
 968             fjy2             = _mm256_setzero_ps();
 969             fjz2             = _mm256_setzero_ps();
 970             fjx3             = _mm256_setzero_ps();
 971             fjy3             = _mm256_setzero_ps();
 972             fjz3             = _mm256_setzero_ps();
 973
 974             /**************************
 975              * CALCULATE INTERACTIONS *
 976              **************************/
 977
 978             if (gmx_mm256_any_lt(rsq11,rcutoff2))
 979             {
 980
 981             r11              = _mm256_mul_ps(rsq11,rinv11);
 982             r11              = _mm256_andnot_ps(dummy_mask,r11);
 983
 984             /* EWALD ELECTROSTATICS */
 985
 986             /* Analytical PME correction */
 987             zeta2            = _mm256_mul_ps(beta2,rsq11);
 988             rinv3            = _mm256_mul_ps(rinvsq11,rinv11);
 989             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
 990             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 991             felec            = _mm256_mul_ps(qq11,felec);
 992             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
 993             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 994             velec            = _mm256_sub_ps(rinv11,pmecorrV);
 995             velec            = _mm256_mul_ps(qq11,velec);
 996
 997             d                = _mm256_sub_ps(r11,rswitch);
 998             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 999             d2               = _mm256_mul_ps(d,d);
1000             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)))))));
1001
1002             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1003
1004             /* Evaluate switch function */
1005             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1006             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv11,_mm256_mul_ps(velec,dsw)) );
1007             velec            = _mm256_mul_ps(velec,sw);
1008             cutoff_mask      = _mm256_cmp_ps(rsq11,rcutoff2,_CMP_LT_OQ);
1009
1010             /* Update potential sum for this i atom from the interaction with this j atom. */
1011             velec            = _mm256_and_ps(velec,cutoff_mask);
1012             velec            = _mm256_andnot_ps(dummy_mask,velec);
1013             velecsum         = _mm256_add_ps(velecsum,velec);
1014
1015             fscal            = felec;
1016
1017             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1018
1019             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1020
1021             /* Calculate temporary vectorial force */
1022             tx               = _mm256_mul_ps(fscal,dx11);
1023             ty               = _mm256_mul_ps(fscal,dy11);
1024             tz               = _mm256_mul_ps(fscal,dz11);
1025
1026             /* Update vectorial force */
1027             fix1             = _mm256_add_ps(fix1,tx);
1028             fiy1             = _mm256_add_ps(fiy1,ty);
1029             fiz1             = _mm256_add_ps(fiz1,tz);
1030
1031             fjx1             = _mm256_add_ps(fjx1,tx);
1032             fjy1             = _mm256_add_ps(fjy1,ty);
1033             fjz1             = _mm256_add_ps(fjz1,tz);
1034
1035             }
1036
1037             /**************************
1038              * CALCULATE INTERACTIONS *
1039              **************************/
1040
1041             if (gmx_mm256_any_lt(rsq12,rcutoff2))
1042             {
1043
1044             r12              = _mm256_mul_ps(rsq12,rinv12);
1045             r12              = _mm256_andnot_ps(dummy_mask,r12);
1046
1047             /* EWALD ELECTROSTATICS */
1048
1049             /* Analytical PME correction */
1050             zeta2            = _mm256_mul_ps(beta2,rsq12);
1051             rinv3            = _mm256_mul_ps(rinvsq12,rinv12);
1052             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
1053             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1054             felec            = _mm256_mul_ps(qq12,felec);
1055             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
1056             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1057             velec            = _mm256_sub_ps(rinv12,pmecorrV);
1058             velec            = _mm256_mul_ps(qq12,velec);
1059
1060             d                = _mm256_sub_ps(r12,rswitch);
1061             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1062             d2               = _mm256_mul_ps(d,d);
1063             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)))))));
1064
1065             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1066
1067             /* Evaluate switch function */
1068             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1069             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv12,_mm256_mul_ps(velec,dsw)) );
1070             velec            = _mm256_mul_ps(velec,sw);
1071             cutoff_mask      = _mm256_cmp_ps(rsq12,rcutoff2,_CMP_LT_OQ);
1072
1073             /* Update potential sum for this i atom from the interaction with this j atom. */
1074             velec            = _mm256_and_ps(velec,cutoff_mask);
1075             velec            = _mm256_andnot_ps(dummy_mask,velec);
1076             velecsum         = _mm256_add_ps(velecsum,velec);
1077
1078             fscal            = felec;
1079
1080             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1081
1082             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1083
1084             /* Calculate temporary vectorial force */
1085             tx               = _mm256_mul_ps(fscal,dx12);
1086             ty               = _mm256_mul_ps(fscal,dy12);
1087             tz               = _mm256_mul_ps(fscal,dz12);
1088
1089             /* Update vectorial force */
1090             fix1             = _mm256_add_ps(fix1,tx);
1091             fiy1             = _mm256_add_ps(fiy1,ty);
1092             fiz1             = _mm256_add_ps(fiz1,tz);
1093
1094             fjx2             = _mm256_add_ps(fjx2,tx);
1095             fjy2             = _mm256_add_ps(fjy2,ty);
1096             fjz2             = _mm256_add_ps(fjz2,tz);
1097
1098             }
1099
1100             /**************************
1101              * CALCULATE INTERACTIONS *
1102              **************************/
1103
1104             if (gmx_mm256_any_lt(rsq13,rcutoff2))
1105             {
1106
1107             r13              = _mm256_mul_ps(rsq13,rinv13);
1108             r13              = _mm256_andnot_ps(dummy_mask,r13);
1109
1110             /* EWALD ELECTROSTATICS */
1111
1112             /* Analytical PME correction */
1113             zeta2            = _mm256_mul_ps(beta2,rsq13);
1114             rinv3            = _mm256_mul_ps(rinvsq13,rinv13);
1115             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
1116             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1117             felec            = _mm256_mul_ps(qq13,felec);
1118             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
1119             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1120             velec            = _mm256_sub_ps(rinv13,pmecorrV);
1121             velec            = _mm256_mul_ps(qq13,velec);
1122
1123             d                = _mm256_sub_ps(r13,rswitch);
1124             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1125             d2               = _mm256_mul_ps(d,d);
1126             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)))))));
1127
1128             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1129
1130             /* Evaluate switch function */
1131             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1132             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv13,_mm256_mul_ps(velec,dsw)) );
1133             velec            = _mm256_mul_ps(velec,sw);
1134             cutoff_mask      = _mm256_cmp_ps(rsq13,rcutoff2,_CMP_LT_OQ);
1135
1136             /* Update potential sum for this i atom from the interaction with this j atom. */
1137             velec            = _mm256_and_ps(velec,cutoff_mask);
1138             velec            = _mm256_andnot_ps(dummy_mask,velec);
1139             velecsum         = _mm256_add_ps(velecsum,velec);
1140
1141             fscal            = felec;
1142
1143             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1144
1145             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1146
1147             /* Calculate temporary vectorial force */
1148             tx               = _mm256_mul_ps(fscal,dx13);
1149             ty               = _mm256_mul_ps(fscal,dy13);
1150             tz               = _mm256_mul_ps(fscal,dz13);
1151
1152             /* Update vectorial force */
1153             fix1             = _mm256_add_ps(fix1,tx);
1154             fiy1             = _mm256_add_ps(fiy1,ty);
1155             fiz1             = _mm256_add_ps(fiz1,tz);
1156
1157             fjx3             = _mm256_add_ps(fjx3,tx);
1158             fjy3             = _mm256_add_ps(fjy3,ty);
1159             fjz3             = _mm256_add_ps(fjz3,tz);
1160
1161             }
1162
1163             /**************************
1164              * CALCULATE INTERACTIONS *
1165              **************************/
1166
1167             if (gmx_mm256_any_lt(rsq21,rcutoff2))
1168             {
1169
1170             r21              = _mm256_mul_ps(rsq21,rinv21);
1171             r21              = _mm256_andnot_ps(dummy_mask,r21);
1172
1173             /* EWALD ELECTROSTATICS */
1174
1175             /* Analytical PME correction */
1176             zeta2            = _mm256_mul_ps(beta2,rsq21);
1177             rinv3            = _mm256_mul_ps(rinvsq21,rinv21);
1178             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
1179             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1180             felec            = _mm256_mul_ps(qq21,felec);
1181             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
1182             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1183             velec            = _mm256_sub_ps(rinv21,pmecorrV);
1184             velec            = _mm256_mul_ps(qq21,velec);
1185
1186             d                = _mm256_sub_ps(r21,rswitch);
1187             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1188             d2               = _mm256_mul_ps(d,d);
1189             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)))))));
1190
1191             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1192
1193             /* Evaluate switch function */
1194             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1195             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv21,_mm256_mul_ps(velec,dsw)) );
1196             velec            = _mm256_mul_ps(velec,sw);
1197             cutoff_mask      = _mm256_cmp_ps(rsq21,rcutoff2,_CMP_LT_OQ);
1198
1199             /* Update potential sum for this i atom from the interaction with this j atom. */
1200             velec            = _mm256_and_ps(velec,cutoff_mask);
1201             velec            = _mm256_andnot_ps(dummy_mask,velec);
1202             velecsum         = _mm256_add_ps(velecsum,velec);
1203
1204             fscal            = felec;
1205
1206             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1207
1208             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1209
1210             /* Calculate temporary vectorial force */
1211             tx               = _mm256_mul_ps(fscal,dx21);
1212             ty               = _mm256_mul_ps(fscal,dy21);
1213             tz               = _mm256_mul_ps(fscal,dz21);
1214
1215             /* Update vectorial force */
1216             fix2             = _mm256_add_ps(fix2,tx);
1217             fiy2             = _mm256_add_ps(fiy2,ty);
1218             fiz2             = _mm256_add_ps(fiz2,tz);
1219
1220             fjx1             = _mm256_add_ps(fjx1,tx);
1221             fjy1             = _mm256_add_ps(fjy1,ty);
1222             fjz1             = _mm256_add_ps(fjz1,tz);
1223
1224             }
1225
1226             /**************************
1227              * CALCULATE INTERACTIONS *
1228              **************************/
1229
1230             if (gmx_mm256_any_lt(rsq22,rcutoff2))
1231             {
1232
1233             r22              = _mm256_mul_ps(rsq22,rinv22);
1234             r22              = _mm256_andnot_ps(dummy_mask,r22);
1235
1236             /* EWALD ELECTROSTATICS */
1237
1238             /* Analytical PME correction */
1239             zeta2            = _mm256_mul_ps(beta2,rsq22);
1240             rinv3            = _mm256_mul_ps(rinvsq22,rinv22);
1241             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
1242             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1243             felec            = _mm256_mul_ps(qq22,felec);
1244             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
1245             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1246             velec            = _mm256_sub_ps(rinv22,pmecorrV);
1247             velec            = _mm256_mul_ps(qq22,velec);
1248
1249             d                = _mm256_sub_ps(r22,rswitch);
1250             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1251             d2               = _mm256_mul_ps(d,d);
1252             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)))))));
1253
1254             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1255
1256             /* Evaluate switch function */
1257             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1258             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv22,_mm256_mul_ps(velec,dsw)) );
1259             velec            = _mm256_mul_ps(velec,sw);
1260             cutoff_mask      = _mm256_cmp_ps(rsq22,rcutoff2,_CMP_LT_OQ);
1261
1262             /* Update potential sum for this i atom from the interaction with this j atom. */
1263             velec            = _mm256_and_ps(velec,cutoff_mask);
1264             velec            = _mm256_andnot_ps(dummy_mask,velec);
1265             velecsum         = _mm256_add_ps(velecsum,velec);
1266
1267             fscal            = felec;
1268
1269             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1270
1271             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1272
1273             /* Calculate temporary vectorial force */
1274             tx               = _mm256_mul_ps(fscal,dx22);
1275             ty               = _mm256_mul_ps(fscal,dy22);
1276             tz               = _mm256_mul_ps(fscal,dz22);
1277
1278             /* Update vectorial force */
1279             fix2             = _mm256_add_ps(fix2,tx);
1280             fiy2             = _mm256_add_ps(fiy2,ty);
1281             fiz2             = _mm256_add_ps(fiz2,tz);
1282
1283             fjx2             = _mm256_add_ps(fjx2,tx);
1284             fjy2             = _mm256_add_ps(fjy2,ty);
1285             fjz2             = _mm256_add_ps(fjz2,tz);
1286
1287             }
1288
1289             /**************************
1290              * CALCULATE INTERACTIONS *
1291              **************************/
1292
1293             if (gmx_mm256_any_lt(rsq23,rcutoff2))
1294             {
1295
1296             r23              = _mm256_mul_ps(rsq23,rinv23);
1297             r23              = _mm256_andnot_ps(dummy_mask,r23);
1298
1299             /* EWALD ELECTROSTATICS */
1300
1301             /* Analytical PME correction */
1302             zeta2            = _mm256_mul_ps(beta2,rsq23);
1303             rinv3            = _mm256_mul_ps(rinvsq23,rinv23);
1304             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
1305             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1306             felec            = _mm256_mul_ps(qq23,felec);
1307             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
1308             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1309             velec            = _mm256_sub_ps(rinv23,pmecorrV);
1310             velec            = _mm256_mul_ps(qq23,velec);
1311
1312             d                = _mm256_sub_ps(r23,rswitch);
1313             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1314             d2               = _mm256_mul_ps(d,d);
1315             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)))))));
1316
1317             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1318
1319             /* Evaluate switch function */
1320             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1321             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv23,_mm256_mul_ps(velec,dsw)) );
1322             velec            = _mm256_mul_ps(velec,sw);
1323             cutoff_mask      = _mm256_cmp_ps(rsq23,rcutoff2,_CMP_LT_OQ);
1324
1325             /* Update potential sum for this i atom from the interaction with this j atom. */
1326             velec            = _mm256_and_ps(velec,cutoff_mask);
1327             velec            = _mm256_andnot_ps(dummy_mask,velec);
1328             velecsum         = _mm256_add_ps(velecsum,velec);
1329
1330             fscal            = felec;
1331
1332             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1333
1334             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1335
1336             /* Calculate temporary vectorial force */
1337             tx               = _mm256_mul_ps(fscal,dx23);
1338             ty               = _mm256_mul_ps(fscal,dy23);
1339             tz               = _mm256_mul_ps(fscal,dz23);
1340
1341             /* Update vectorial force */
1342             fix2             = _mm256_add_ps(fix2,tx);
1343             fiy2             = _mm256_add_ps(fiy2,ty);
1344             fiz2             = _mm256_add_ps(fiz2,tz);
1345
1346             fjx3             = _mm256_add_ps(fjx3,tx);
1347             fjy3             = _mm256_add_ps(fjy3,ty);
1348             fjz3             = _mm256_add_ps(fjz3,tz);
1349
1350             }
1351
1352             /**************************
1353              * CALCULATE INTERACTIONS *
1354              **************************/
1355
1356             if (gmx_mm256_any_lt(rsq31,rcutoff2))
1357             {
1358
1359             r31              = _mm256_mul_ps(rsq31,rinv31);
1360             r31              = _mm256_andnot_ps(dummy_mask,r31);
1361
1362             /* EWALD ELECTROSTATICS */
1363
1364             /* Analytical PME correction */
1365             zeta2            = _mm256_mul_ps(beta2,rsq31);
1366             rinv3            = _mm256_mul_ps(rinvsq31,rinv31);
1367             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
1368             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1369             felec            = _mm256_mul_ps(qq31,felec);
1370             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
1371             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1372             velec            = _mm256_sub_ps(rinv31,pmecorrV);
1373             velec            = _mm256_mul_ps(qq31,velec);
1374
1375             d                = _mm256_sub_ps(r31,rswitch);
1376             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1377             d2               = _mm256_mul_ps(d,d);
1378             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)))))));
1379
1380             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1381
1382             /* Evaluate switch function */
1383             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1384             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv31,_mm256_mul_ps(velec,dsw)) );
1385             velec            = _mm256_mul_ps(velec,sw);
1386             cutoff_mask      = _mm256_cmp_ps(rsq31,rcutoff2,_CMP_LT_OQ);
1387
1388             /* Update potential sum for this i atom from the interaction with this j atom. */
1389             velec            = _mm256_and_ps(velec,cutoff_mask);
1390             velec            = _mm256_andnot_ps(dummy_mask,velec);
1391             velecsum         = _mm256_add_ps(velecsum,velec);
1392
1393             fscal            = felec;
1394
1395             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1396
1397             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1398
1399             /* Calculate temporary vectorial force */
1400             tx               = _mm256_mul_ps(fscal,dx31);
1401             ty               = _mm256_mul_ps(fscal,dy31);
1402             tz               = _mm256_mul_ps(fscal,dz31);
1403
1404             /* Update vectorial force */
1405             fix3             = _mm256_add_ps(fix3,tx);
1406             fiy3             = _mm256_add_ps(fiy3,ty);
1407             fiz3             = _mm256_add_ps(fiz3,tz);
1408
1409             fjx1             = _mm256_add_ps(fjx1,tx);
1410             fjy1             = _mm256_add_ps(fjy1,ty);
1411             fjz1             = _mm256_add_ps(fjz1,tz);
1412
1413             }
1414
1415             /**************************
1416              * CALCULATE INTERACTIONS *
1417              **************************/
1418
1419             if (gmx_mm256_any_lt(rsq32,rcutoff2))
1420             {
1421
1422             r32              = _mm256_mul_ps(rsq32,rinv32);
1423             r32              = _mm256_andnot_ps(dummy_mask,r32);
1424
1425             /* EWALD ELECTROSTATICS */
1426
1427             /* Analytical PME correction */
1428             zeta2            = _mm256_mul_ps(beta2,rsq32);
1429             rinv3            = _mm256_mul_ps(rinvsq32,rinv32);
1430             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
1431             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1432             felec            = _mm256_mul_ps(qq32,felec);
1433             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
1434             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1435             velec            = _mm256_sub_ps(rinv32,pmecorrV);
1436             velec            = _mm256_mul_ps(qq32,velec);
1437
1438             d                = _mm256_sub_ps(r32,rswitch);
1439             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1440             d2               = _mm256_mul_ps(d,d);
1441             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)))))));
1442
1443             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1444
1445             /* Evaluate switch function */
1446             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1447             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv32,_mm256_mul_ps(velec,dsw)) );
1448             velec            = _mm256_mul_ps(velec,sw);
1449             cutoff_mask      = _mm256_cmp_ps(rsq32,rcutoff2,_CMP_LT_OQ);
1450
1451             /* Update potential sum for this i atom from the interaction with this j atom. */
1452             velec            = _mm256_and_ps(velec,cutoff_mask);
1453             velec            = _mm256_andnot_ps(dummy_mask,velec);
1454             velecsum         = _mm256_add_ps(velecsum,velec);
1455
1456             fscal            = felec;
1457
1458             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1459
1460             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1461
1462             /* Calculate temporary vectorial force */
1463             tx               = _mm256_mul_ps(fscal,dx32);
1464             ty               = _mm256_mul_ps(fscal,dy32);
1465             tz               = _mm256_mul_ps(fscal,dz32);
1466
1467             /* Update vectorial force */
1468             fix3             = _mm256_add_ps(fix3,tx);
1469             fiy3             = _mm256_add_ps(fiy3,ty);
1470             fiz3             = _mm256_add_ps(fiz3,tz);
1471
1472             fjx2             = _mm256_add_ps(fjx2,tx);
1473             fjy2             = _mm256_add_ps(fjy2,ty);
1474             fjz2             = _mm256_add_ps(fjz2,tz);
1475
1476             }
1477
1478             /**************************
1479              * CALCULATE INTERACTIONS *
1480              **************************/
1481
1482             if (gmx_mm256_any_lt(rsq33,rcutoff2))
1483             {
1484
1485             r33              = _mm256_mul_ps(rsq33,rinv33);
1486             r33              = _mm256_andnot_ps(dummy_mask,r33);
1487
1488             /* EWALD ELECTROSTATICS */
1489
1490             /* Analytical PME correction */
1491             zeta2            = _mm256_mul_ps(beta2,rsq33);
1492             rinv3            = _mm256_mul_ps(rinvsq33,rinv33);
1493             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
1494             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1495             felec            = _mm256_mul_ps(qq33,felec);
1496             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
1497             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1498             velec            = _mm256_sub_ps(rinv33,pmecorrV);
1499             velec            = _mm256_mul_ps(qq33,velec);
1500
1501             d                = _mm256_sub_ps(r33,rswitch);
1502             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1503             d2               = _mm256_mul_ps(d,d);
1504             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)))))));
1505
1506             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1507
1508             /* Evaluate switch function */
1509             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1510             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv33,_mm256_mul_ps(velec,dsw)) );
1511             velec            = _mm256_mul_ps(velec,sw);
1512             cutoff_mask      = _mm256_cmp_ps(rsq33,rcutoff2,_CMP_LT_OQ);
1513
1514             /* Update potential sum for this i atom from the interaction with this j atom. */
1515             velec            = _mm256_and_ps(velec,cutoff_mask);
1516             velec            = _mm256_andnot_ps(dummy_mask,velec);
1517             velecsum         = _mm256_add_ps(velecsum,velec);
1518
1519             fscal            = felec;
1520
1521             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1522
1523             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1524
1525             /* Calculate temporary vectorial force */
1526             tx               = _mm256_mul_ps(fscal,dx33);
1527             ty               = _mm256_mul_ps(fscal,dy33);
1528             tz               = _mm256_mul_ps(fscal,dz33);
1529
1530             /* Update vectorial force */
1531             fix3             = _mm256_add_ps(fix3,tx);
1532             fiy3             = _mm256_add_ps(fiy3,ty);
1533             fiz3             = _mm256_add_ps(fiz3,tz);
1534
1535             fjx3             = _mm256_add_ps(fjx3,tx);
1536             fjy3             = _mm256_add_ps(fjy3,ty);
1537             fjz3             = _mm256_add_ps(fjz3,tz);
1538
1539             }
1540
1541             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
1542             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
1543             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
1544             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
1545             fjptrE             = (jnrlistE>=0) ? f+j_coord_offsetE : scratch;
1546             fjptrF             = (jnrlistF>=0) ? f+j_coord_offsetF : scratch;
1547             fjptrG             = (jnrlistG>=0) ? f+j_coord_offsetG : scratch;
1548             fjptrH             = (jnrlistH>=0) ? f+j_coord_offsetH : scratch;
1549
1550             gmx_mm256_decrement_3rvec_8ptr_swizzle_ps(fjptrA+DIM,fjptrB+DIM,fjptrC+DIM,fjptrD+DIM,
1551                                                       fjptrE+DIM,fjptrF+DIM,fjptrG+DIM,fjptrH+DIM,
1552                                                       fjx1,fjy1,fjz1,fjx2,fjy2,fjz2,fjx3,fjy3,fjz3);
1553
1554             /* Inner loop uses 981 flops */
1555         }
1556
1557         /* End of innermost loop */
1558
1559         gmx_mm256_update_iforce_3atom_swizzle_ps(fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
1560                                                  f+i_coord_offset+DIM,fshift+i_shift_offset);
1561
1562         ggid                        = gid[iidx];
1563         /* Update potential energies */
1564         gmx_mm256_update_1pot_ps(velecsum,kernel_data->energygrp_elec+ggid);
1565
1566         /* Increment number of inner iterations */
1567         inneriter                  += j_index_end - j_index_start;
1568
1569         /* Outer loop uses 19 flops */
1570     }
1571
1572     /* Increment number of outer iterations */
1573     outeriter        += nri;
1574
1575     /* Update outer/inner flops */
1576
1577     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_W4W4_VF,outeriter*19 + inneriter*981);
1578 }
1579 /*
1580  * Gromacs nonbonded kernel:   nb_kernel_ElecEwSw_VdwNone_GeomW4W4_F_avx_256_single
1581  * Electrostatics interaction: Ewald
1582  * VdW interaction:            None
1583  * Geometry:                   Water4-Water4
1584  * Calculate force/pot:        Force
1585  */
1586 void
1587 nb_kernel_ElecEwSw_VdwNone_GeomW4W4_F_avx_256_single
1588                     (t_nblist * gmx_restrict                nlist,
1589                      rvec * gmx_restrict                    xx,
1590                      rvec * gmx_restrict                    ff,
1591                      t_forcerec * gmx_restrict              fr,
1592                      t_mdatoms * gmx_restrict               mdatoms,
1593                      nb_kernel_data_t * gmx_restrict        kernel_data,
1594                      t_nrnb * gmx_restrict                  nrnb)
1595 {
1596     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
1597      * just 0 for non-waters.
1598      * Suffixes A,B,C,D,E,F,G,H refer to j loop unrolling done with AVX, e.g. for the eight different
1599      * jnr indices corresponding to data put in the four positions in the SIMD register.
1600      */
1601     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
1602     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
1603     int              jnrA,jnrB,jnrC,jnrD;
1604     int              jnrE,jnrF,jnrG,jnrH;
1605     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
1606     int              jnrlistE,jnrlistF,jnrlistG,jnrlistH;
1607     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
1608     int              j_coord_offsetE,j_coord_offsetF,j_coord_offsetG,j_coord_offsetH;
1609     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
1610     real             rcutoff_scalar;
1611     real             *shiftvec,*fshift,*x,*f;
1612     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD,*fjptrE,*fjptrF,*fjptrG,*fjptrH;
1613     real             scratch[4*DIM];
1614     __m256           tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
1615     real *           vdwioffsetptr1;
1616     __m256           ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
1617     real *           vdwioffsetptr2;
1618     __m256           ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
1619     real *           vdwioffsetptr3;
1620     __m256           ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
1621     int              vdwjidx1A,vdwjidx1B,vdwjidx1C,vdwjidx1D,vdwjidx1E,vdwjidx1F,vdwjidx1G,vdwjidx1H;
1622     __m256           jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
1623     int              vdwjidx2A,vdwjidx2B,vdwjidx2C,vdwjidx2D,vdwjidx2E,vdwjidx2F,vdwjidx2G,vdwjidx2H;
1624     __m256           jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
1625     int              vdwjidx3A,vdwjidx3B,vdwjidx3C,vdwjidx3D,vdwjidx3E,vdwjidx3F,vdwjidx3G,vdwjidx3H;
1626     __m256           jx3,jy3,jz3,fjx3,fjy3,fjz3,jq3,isaj3;
1627     __m256           dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11;
1628     __m256           dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12;
1629     __m256           dx13,dy13,dz13,rsq13,rinv13,rinvsq13,r13,qq13,c6_13,c12_13;
1630     __m256           dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21;
1631     __m256           dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22;
1632     __m256           dx23,dy23,dz23,rsq23,rinv23,rinvsq23,r23,qq23,c6_23,c12_23;
1633     __m256           dx31,dy31,dz31,rsq31,rinv31,rinvsq31,r31,qq31,c6_31,c12_31;
1634     __m256           dx32,dy32,dz32,rsq32,rinv32,rinvsq32,r32,qq32,c6_32,c12_32;
1635     __m256           dx33,dy33,dz33,rsq33,rinv33,rinvsq33,r33,qq33,c6_33,c12_33;
1636     __m256           velec,felec,velecsum,facel,crf,krf,krf2;
1637     real             *charge;
1638     __m256i          ewitab;
1639     __m128i          ewitab_lo,ewitab_hi;
1640     __m256           ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
1641     __m256           beta,beta2,beta3,zeta2,pmecorrF,pmecorrV,rinv3;
1642     real             *ewtab;
1643     __m256           rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
1644     real             rswitch_scalar,d_scalar;
1645     __m256           dummy_mask,cutoff_mask;
1646     __m256           signbit = _mm256_castsi256_ps( _mm256_set1_epi32(0x80000000) );
1647     __m256           one     = _mm256_set1_ps(1.0);
1648     __m256           two     = _mm256_set1_ps(2.0);
1649     x                = xx[0];
1650     f                = ff[0];
1651
1652     nri              = nlist->nri;
1653     iinr             = nlist->iinr;
1654     jindex           = nlist->jindex;
1655     jjnr             = nlist->jjnr;
1656     shiftidx         = nlist->shift;
1657     gid              = nlist->gid;
1658     shiftvec         = fr->shift_vec[0];
1659     fshift           = fr->fshift[0];
1660     facel            = _mm256_set1_ps(fr->epsfac);
1661     charge           = mdatoms->chargeA;
1662
1663     sh_ewald         = _mm256_set1_ps(fr->ic->sh_ewald);
1664     beta             = _mm256_set1_ps(fr->ic->ewaldcoeff);
1665     beta2            = _mm256_mul_ps(beta,beta);
1666     beta3            = _mm256_mul_ps(beta,beta2);
1667
1668     ewtab            = fr->ic->tabq_coul_FDV0;
1669     ewtabscale       = _mm256_set1_ps(fr->ic->tabq_scale);
1670     ewtabhalfspace   = _mm256_set1_ps(0.5/fr->ic->tabq_scale);
1671
1672     /* Setup water-specific parameters */
1673     inr              = nlist->iinr[0];
1674     iq1              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+1]));
1675     iq2              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+2]));
1676     iq3              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+3]));
1677
1678     jq1              = _mm256_set1_ps(charge[inr+1]);
1679     jq2              = _mm256_set1_ps(charge[inr+2]);
1680     jq3              = _mm256_set1_ps(charge[inr+3]);
1681     qq11             = _mm256_mul_ps(iq1,jq1);
1682     qq12             = _mm256_mul_ps(iq1,jq2);
1683     qq13             = _mm256_mul_ps(iq1,jq3);
1684     qq21             = _mm256_mul_ps(iq2,jq1);
1685     qq22             = _mm256_mul_ps(iq2,jq2);
1686     qq23             = _mm256_mul_ps(iq2,jq3);
1687     qq31             = _mm256_mul_ps(iq3,jq1);
1688     qq32             = _mm256_mul_ps(iq3,jq2);
1689     qq33             = _mm256_mul_ps(iq3,jq3);
1690
1691     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
1692     rcutoff_scalar   = fr->rcoulomb;
1693     rcutoff          = _mm256_set1_ps(rcutoff_scalar);
1694     rcutoff2         = _mm256_mul_ps(rcutoff,rcutoff);
1695
1696     rswitch_scalar   = fr->rcoulomb_switch;
1697     rswitch          = _mm256_set1_ps(rswitch_scalar);
1698     /* Setup switch parameters */
1699     d_scalar         = rcutoff_scalar-rswitch_scalar;
1700     d                = _mm256_set1_ps(d_scalar);
1701     swV3             = _mm256_set1_ps(-10.0/(d_scalar*d_scalar*d_scalar));
1702     swV4             = _mm256_set1_ps( 15.0/(d_scalar*d_scalar*d_scalar*d_scalar));
1703     swV5             = _mm256_set1_ps( -6.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
1704     swF2             = _mm256_set1_ps(-30.0/(d_scalar*d_scalar*d_scalar));
1705     swF3             = _mm256_set1_ps( 60.0/(d_scalar*d_scalar*d_scalar*d_scalar));
1706     swF4             = _mm256_set1_ps(-30.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
1707
1708     /* Avoid stupid compiler warnings */
1709     jnrA = jnrB = jnrC = jnrD = jnrE = jnrF = jnrG = jnrH = 0;
1710     j_coord_offsetA = 0;
1711     j_coord_offsetB = 0;
1712     j_coord_offsetC = 0;
1713     j_coord_offsetD = 0;
1714     j_coord_offsetE = 0;
1715     j_coord_offsetF = 0;
1716     j_coord_offsetG = 0;
1717     j_coord_offsetH = 0;
1718
1719     outeriter        = 0;
1720     inneriter        = 0;
1721
1722     for(iidx=0;iidx<4*DIM;iidx++)
1723     {
1724         scratch[iidx] = 0.0;
1725     }
1726
1727     /* Start outer loop over neighborlists */
1728     for(iidx=0; iidx<nri; iidx++)
1729     {
1730         /* Load shift vector for this list */
1731         i_shift_offset   = DIM*shiftidx[iidx];
1732
1733         /* Load limits for loop over neighbors */
1734         j_index_start    = jindex[iidx];
1735         j_index_end      = jindex[iidx+1];
1736
1737         /* Get outer coordinate index */
1738         inr              = iinr[iidx];
1739         i_coord_offset   = DIM*inr;
1740
1741         /* Load i particle coords and add shift vector */
1742         gmx_mm256_load_shift_and_3rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset+DIM,
1743                                                     &ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
1744
1745         fix1             = _mm256_setzero_ps();
1746         fiy1             = _mm256_setzero_ps();
1747         fiz1             = _mm256_setzero_ps();
1748         fix2             = _mm256_setzero_ps();
1749         fiy2             = _mm256_setzero_ps();
1750         fiz2             = _mm256_setzero_ps();
1751         fix3             = _mm256_setzero_ps();
1752         fiy3             = _mm256_setzero_ps();
1753         fiz3             = _mm256_setzero_ps();
1754
1755         /* Start inner kernel loop */
1756         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+7]>=0; jidx+=8)
1757         {
1758
1759             /* Get j neighbor index, and coordinate index */
1760             jnrA             = jjnr[jidx];
1761             jnrB             = jjnr[jidx+1];
1762             jnrC             = jjnr[jidx+2];
1763             jnrD             = jjnr[jidx+3];
1764             jnrE             = jjnr[jidx+4];
1765             jnrF             = jjnr[jidx+5];
1766             jnrG             = jjnr[jidx+6];
1767             jnrH             = jjnr[jidx+7];
1768             j_coord_offsetA  = DIM*jnrA;
1769             j_coord_offsetB  = DIM*jnrB;
1770             j_coord_offsetC  = DIM*jnrC;
1771             j_coord_offsetD  = DIM*jnrD;
1772             j_coord_offsetE  = DIM*jnrE;
1773             j_coord_offsetF  = DIM*jnrF;
1774             j_coord_offsetG  = DIM*jnrG;
1775             j_coord_offsetH  = DIM*jnrH;
1776
1777             /* load j atom coordinates */
1778             gmx_mm256_load_3rvec_8ptr_swizzle_ps(x+j_coord_offsetA+DIM,x+j_coord_offsetB+DIM,
1779                                                  x+j_coord_offsetC+DIM,x+j_coord_offsetD+DIM,
1780                                                  x+j_coord_offsetE+DIM,x+j_coord_offsetF+DIM,
1781                                                  x+j_coord_offsetG+DIM,x+j_coord_offsetH+DIM,
1782                                                  &jx1,&jy1,&jz1,&jx2,&jy2,&jz2,&jx3,&jy3,&jz3);
1783
1784             /* Calculate displacement vector */
1785             dx11             = _mm256_sub_ps(ix1,jx1);
1786             dy11             = _mm256_sub_ps(iy1,jy1);
1787             dz11             = _mm256_sub_ps(iz1,jz1);
1788             dx12             = _mm256_sub_ps(ix1,jx2);
1789             dy12             = _mm256_sub_ps(iy1,jy2);
1790             dz12             = _mm256_sub_ps(iz1,jz2);
1791             dx13             = _mm256_sub_ps(ix1,jx3);
1792             dy13             = _mm256_sub_ps(iy1,jy3);
1793             dz13             = _mm256_sub_ps(iz1,jz3);
1794             dx21             = _mm256_sub_ps(ix2,jx1);
1795             dy21             = _mm256_sub_ps(iy2,jy1);
1796             dz21             = _mm256_sub_ps(iz2,jz1);
1797             dx22             = _mm256_sub_ps(ix2,jx2);
1798             dy22             = _mm256_sub_ps(iy2,jy2);
1799             dz22             = _mm256_sub_ps(iz2,jz2);
1800             dx23             = _mm256_sub_ps(ix2,jx3);
1801             dy23             = _mm256_sub_ps(iy2,jy3);
1802             dz23             = _mm256_sub_ps(iz2,jz3);
1803             dx31             = _mm256_sub_ps(ix3,jx1);
1804             dy31             = _mm256_sub_ps(iy3,jy1);
1805             dz31             = _mm256_sub_ps(iz3,jz1);
1806             dx32             = _mm256_sub_ps(ix3,jx2);
1807             dy32             = _mm256_sub_ps(iy3,jy2);
1808             dz32             = _mm256_sub_ps(iz3,jz2);
1809             dx33             = _mm256_sub_ps(ix3,jx3);
1810             dy33             = _mm256_sub_ps(iy3,jy3);
1811             dz33             = _mm256_sub_ps(iz3,jz3);
1812
1813             /* Calculate squared distance and things based on it */
1814             rsq11            = gmx_mm256_calc_rsq_ps(dx11,dy11,dz11);
1815             rsq12            = gmx_mm256_calc_rsq_ps(dx12,dy12,dz12);
1816             rsq13            = gmx_mm256_calc_rsq_ps(dx13,dy13,dz13);
1817             rsq21            = gmx_mm256_calc_rsq_ps(dx21,dy21,dz21);
1818             rsq22            = gmx_mm256_calc_rsq_ps(dx22,dy22,dz22);
1819             rsq23            = gmx_mm256_calc_rsq_ps(dx23,dy23,dz23);
1820             rsq31            = gmx_mm256_calc_rsq_ps(dx31,dy31,dz31);
1821             rsq32            = gmx_mm256_calc_rsq_ps(dx32,dy32,dz32);
1822             rsq33            = gmx_mm256_calc_rsq_ps(dx33,dy33,dz33);
1823
1824             rinv11           = gmx_mm256_invsqrt_ps(rsq11);
1825             rinv12           = gmx_mm256_invsqrt_ps(rsq12);
1826             rinv13           = gmx_mm256_invsqrt_ps(rsq13);
1827             rinv21           = gmx_mm256_invsqrt_ps(rsq21);
1828             rinv22           = gmx_mm256_invsqrt_ps(rsq22);
1829             rinv23           = gmx_mm256_invsqrt_ps(rsq23);
1830             rinv31           = gmx_mm256_invsqrt_ps(rsq31);
1831             rinv32           = gmx_mm256_invsqrt_ps(rsq32);
1832             rinv33           = gmx_mm256_invsqrt_ps(rsq33);
1833
1834             rinvsq11         = _mm256_mul_ps(rinv11,rinv11);
1835             rinvsq12         = _mm256_mul_ps(rinv12,rinv12);
1836             rinvsq13         = _mm256_mul_ps(rinv13,rinv13);
1837             rinvsq21         = _mm256_mul_ps(rinv21,rinv21);
1838             rinvsq22         = _mm256_mul_ps(rinv22,rinv22);
1839             rinvsq23         = _mm256_mul_ps(rinv23,rinv23);
1840             rinvsq31         = _mm256_mul_ps(rinv31,rinv31);
1841             rinvsq32         = _mm256_mul_ps(rinv32,rinv32);
1842             rinvsq33         = _mm256_mul_ps(rinv33,rinv33);
1843
1844             fjx1             = _mm256_setzero_ps();
1845             fjy1             = _mm256_setzero_ps();
1846             fjz1             = _mm256_setzero_ps();
1847             fjx2             = _mm256_setzero_ps();
1848             fjy2             = _mm256_setzero_ps();
1849             fjz2             = _mm256_setzero_ps();
1850             fjx3             = _mm256_setzero_ps();
1851             fjy3             = _mm256_setzero_ps();
1852             fjz3             = _mm256_setzero_ps();
1853
1854             /**************************
1855              * CALCULATE INTERACTIONS *
1856              **************************/
1857
1858             if (gmx_mm256_any_lt(rsq11,rcutoff2))
1859             {
1860
1861             r11              = _mm256_mul_ps(rsq11,rinv11);
1862
1863             /* EWALD ELECTROSTATICS */
1864
1865             /* Analytical PME correction */
1866             zeta2            = _mm256_mul_ps(beta2,rsq11);
1867             rinv3            = _mm256_mul_ps(rinvsq11,rinv11);
1868             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
1869             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1870             felec            = _mm256_mul_ps(qq11,felec);
1871             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
1872             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1873             velec            = _mm256_sub_ps(rinv11,pmecorrV);
1874             velec            = _mm256_mul_ps(qq11,velec);
1875
1876             d                = _mm256_sub_ps(r11,rswitch);
1877             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1878             d2               = _mm256_mul_ps(d,d);
1879             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)))))));
1880
1881             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1882
1883             /* Evaluate switch function */
1884             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1885             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv11,_mm256_mul_ps(velec,dsw)) );
1886             cutoff_mask      = _mm256_cmp_ps(rsq11,rcutoff2,_CMP_LT_OQ);
1887
1888             fscal            = felec;
1889
1890             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1891
1892             /* Calculate temporary vectorial force */
1893             tx               = _mm256_mul_ps(fscal,dx11);
1894             ty               = _mm256_mul_ps(fscal,dy11);
1895             tz               = _mm256_mul_ps(fscal,dz11);
1896
1897             /* Update vectorial force */
1898             fix1             = _mm256_add_ps(fix1,tx);
1899             fiy1             = _mm256_add_ps(fiy1,ty);
1900             fiz1             = _mm256_add_ps(fiz1,tz);
1901
1902             fjx1             = _mm256_add_ps(fjx1,tx);
1903             fjy1             = _mm256_add_ps(fjy1,ty);
1904             fjz1             = _mm256_add_ps(fjz1,tz);
1905
1906             }
1907
1908             /**************************
1909              * CALCULATE INTERACTIONS *
1910              **************************/
1911
1912             if (gmx_mm256_any_lt(rsq12,rcutoff2))
1913             {
1914
1915             r12              = _mm256_mul_ps(rsq12,rinv12);
1916
1917             /* EWALD ELECTROSTATICS */
1918
1919             /* Analytical PME correction */
1920             zeta2            = _mm256_mul_ps(beta2,rsq12);
1921             rinv3            = _mm256_mul_ps(rinvsq12,rinv12);
1922             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
1923             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1924             felec            = _mm256_mul_ps(qq12,felec);
1925             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
1926             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1927             velec            = _mm256_sub_ps(rinv12,pmecorrV);
1928             velec            = _mm256_mul_ps(qq12,velec);
1929
1930             d                = _mm256_sub_ps(r12,rswitch);
1931             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1932             d2               = _mm256_mul_ps(d,d);
1933             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)))))));
1934
1935             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1936
1937             /* Evaluate switch function */
1938             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1939             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv12,_mm256_mul_ps(velec,dsw)) );
1940             cutoff_mask      = _mm256_cmp_ps(rsq12,rcutoff2,_CMP_LT_OQ);
1941
1942             fscal            = felec;
1943
1944             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1945
1946             /* Calculate temporary vectorial force */
1947             tx               = _mm256_mul_ps(fscal,dx12);
1948             ty               = _mm256_mul_ps(fscal,dy12);
1949             tz               = _mm256_mul_ps(fscal,dz12);
1950
1951             /* Update vectorial force */
1952             fix1             = _mm256_add_ps(fix1,tx);
1953             fiy1             = _mm256_add_ps(fiy1,ty);
1954             fiz1             = _mm256_add_ps(fiz1,tz);
1955
1956             fjx2             = _mm256_add_ps(fjx2,tx);
1957             fjy2             = _mm256_add_ps(fjy2,ty);
1958             fjz2             = _mm256_add_ps(fjz2,tz);
1959
1960             }
1961
1962             /**************************
1963              * CALCULATE INTERACTIONS *
1964              **************************/
1965
1966             if (gmx_mm256_any_lt(rsq13,rcutoff2))
1967             {
1968
1969             r13              = _mm256_mul_ps(rsq13,rinv13);
1970
1971             /* EWALD ELECTROSTATICS */
1972
1973             /* Analytical PME correction */
1974             zeta2            = _mm256_mul_ps(beta2,rsq13);
1975             rinv3            = _mm256_mul_ps(rinvsq13,rinv13);
1976             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
1977             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1978             felec            = _mm256_mul_ps(qq13,felec);
1979             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
1980             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1981             velec            = _mm256_sub_ps(rinv13,pmecorrV);
1982             velec            = _mm256_mul_ps(qq13,velec);
1983
1984             d                = _mm256_sub_ps(r13,rswitch);
1985             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1986             d2               = _mm256_mul_ps(d,d);
1987             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)))))));
1988
1989             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1990
1991             /* Evaluate switch function */
1992             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1993             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv13,_mm256_mul_ps(velec,dsw)) );
1994             cutoff_mask      = _mm256_cmp_ps(rsq13,rcutoff2,_CMP_LT_OQ);
1995
1996             fscal            = felec;
1997
1998             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1999
2000             /* Calculate temporary vectorial force */
2001             tx               = _mm256_mul_ps(fscal,dx13);
2002             ty               = _mm256_mul_ps(fscal,dy13);
2003             tz               = _mm256_mul_ps(fscal,dz13);
2004
2005             /* Update vectorial force */
2006             fix1             = _mm256_add_ps(fix1,tx);
2007             fiy1             = _mm256_add_ps(fiy1,ty);
2008             fiz1             = _mm256_add_ps(fiz1,tz);
2009
2010             fjx3             = _mm256_add_ps(fjx3,tx);
2011             fjy3             = _mm256_add_ps(fjy3,ty);
2012             fjz3             = _mm256_add_ps(fjz3,tz);
2013
2014             }
2015
2016             /**************************
2017              * CALCULATE INTERACTIONS *
2018              **************************/
2019
2020             if (gmx_mm256_any_lt(rsq21,rcutoff2))
2021             {
2022
2023             r21              = _mm256_mul_ps(rsq21,rinv21);
2024
2025             /* EWALD ELECTROSTATICS */
2026
2027             /* Analytical PME correction */
2028             zeta2            = _mm256_mul_ps(beta2,rsq21);
2029             rinv3            = _mm256_mul_ps(rinvsq21,rinv21);
2030             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
2031             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2032             felec            = _mm256_mul_ps(qq21,felec);
2033             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
2034             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2035             velec            = _mm256_sub_ps(rinv21,pmecorrV);
2036             velec            = _mm256_mul_ps(qq21,velec);
2037
2038             d                = _mm256_sub_ps(r21,rswitch);
2039             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2040             d2               = _mm256_mul_ps(d,d);
2041             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)))))));
2042
2043             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2044
2045             /* Evaluate switch function */
2046             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2047             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv21,_mm256_mul_ps(velec,dsw)) );
2048             cutoff_mask      = _mm256_cmp_ps(rsq21,rcutoff2,_CMP_LT_OQ);
2049
2050             fscal            = felec;
2051
2052             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2053
2054             /* Calculate temporary vectorial force */
2055             tx               = _mm256_mul_ps(fscal,dx21);
2056             ty               = _mm256_mul_ps(fscal,dy21);
2057             tz               = _mm256_mul_ps(fscal,dz21);
2058
2059             /* Update vectorial force */
2060             fix2             = _mm256_add_ps(fix2,tx);
2061             fiy2             = _mm256_add_ps(fiy2,ty);
2062             fiz2             = _mm256_add_ps(fiz2,tz);
2063
2064             fjx1             = _mm256_add_ps(fjx1,tx);
2065             fjy1             = _mm256_add_ps(fjy1,ty);
2066             fjz1             = _mm256_add_ps(fjz1,tz);
2067
2068             }
2069
2070             /**************************
2071              * CALCULATE INTERACTIONS *
2072              **************************/
2073
2074             if (gmx_mm256_any_lt(rsq22,rcutoff2))
2075             {
2076
2077             r22              = _mm256_mul_ps(rsq22,rinv22);
2078
2079             /* EWALD ELECTROSTATICS */
2080
2081             /* Analytical PME correction */
2082             zeta2            = _mm256_mul_ps(beta2,rsq22);
2083             rinv3            = _mm256_mul_ps(rinvsq22,rinv22);
2084             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
2085             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2086             felec            = _mm256_mul_ps(qq22,felec);
2087             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
2088             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2089             velec            = _mm256_sub_ps(rinv22,pmecorrV);
2090             velec            = _mm256_mul_ps(qq22,velec);
2091
2092             d                = _mm256_sub_ps(r22,rswitch);
2093             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2094             d2               = _mm256_mul_ps(d,d);
2095             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)))))));
2096
2097             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2098
2099             /* Evaluate switch function */
2100             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2101             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv22,_mm256_mul_ps(velec,dsw)) );
2102             cutoff_mask      = _mm256_cmp_ps(rsq22,rcutoff2,_CMP_LT_OQ);
2103
2104             fscal            = felec;
2105
2106             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2107
2108             /* Calculate temporary vectorial force */
2109             tx               = _mm256_mul_ps(fscal,dx22);
2110             ty               = _mm256_mul_ps(fscal,dy22);
2111             tz               = _mm256_mul_ps(fscal,dz22);
2112
2113             /* Update vectorial force */
2114             fix2             = _mm256_add_ps(fix2,tx);
2115             fiy2             = _mm256_add_ps(fiy2,ty);
2116             fiz2             = _mm256_add_ps(fiz2,tz);
2117
2118             fjx2             = _mm256_add_ps(fjx2,tx);
2119             fjy2             = _mm256_add_ps(fjy2,ty);
2120             fjz2             = _mm256_add_ps(fjz2,tz);
2121
2122             }
2123
2124             /**************************
2125              * CALCULATE INTERACTIONS *
2126              **************************/
2127
2128             if (gmx_mm256_any_lt(rsq23,rcutoff2))
2129             {
2130
2131             r23              = _mm256_mul_ps(rsq23,rinv23);
2132
2133             /* EWALD ELECTROSTATICS */
2134
2135             /* Analytical PME correction */
2136             zeta2            = _mm256_mul_ps(beta2,rsq23);
2137             rinv3            = _mm256_mul_ps(rinvsq23,rinv23);
2138             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
2139             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2140             felec            = _mm256_mul_ps(qq23,felec);
2141             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
2142             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2143             velec            = _mm256_sub_ps(rinv23,pmecorrV);
2144             velec            = _mm256_mul_ps(qq23,velec);
2145
2146             d                = _mm256_sub_ps(r23,rswitch);
2147             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2148             d2               = _mm256_mul_ps(d,d);
2149             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)))))));
2150
2151             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2152
2153             /* Evaluate switch function */
2154             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2155             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv23,_mm256_mul_ps(velec,dsw)) );
2156             cutoff_mask      = _mm256_cmp_ps(rsq23,rcutoff2,_CMP_LT_OQ);
2157
2158             fscal            = felec;
2159
2160             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2161
2162             /* Calculate temporary vectorial force */
2163             tx               = _mm256_mul_ps(fscal,dx23);
2164             ty               = _mm256_mul_ps(fscal,dy23);
2165             tz               = _mm256_mul_ps(fscal,dz23);
2166
2167             /* Update vectorial force */
2168             fix2             = _mm256_add_ps(fix2,tx);
2169             fiy2             = _mm256_add_ps(fiy2,ty);
2170             fiz2             = _mm256_add_ps(fiz2,tz);
2171
2172             fjx3             = _mm256_add_ps(fjx3,tx);
2173             fjy3             = _mm256_add_ps(fjy3,ty);
2174             fjz3             = _mm256_add_ps(fjz3,tz);
2175
2176             }
2177
2178             /**************************
2179              * CALCULATE INTERACTIONS *
2180              **************************/
2181
2182             if (gmx_mm256_any_lt(rsq31,rcutoff2))
2183             {
2184
2185             r31              = _mm256_mul_ps(rsq31,rinv31);
2186
2187             /* EWALD ELECTROSTATICS */
2188
2189             /* Analytical PME correction */
2190             zeta2            = _mm256_mul_ps(beta2,rsq31);
2191             rinv3            = _mm256_mul_ps(rinvsq31,rinv31);
2192             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
2193             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2194             felec            = _mm256_mul_ps(qq31,felec);
2195             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
2196             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2197             velec            = _mm256_sub_ps(rinv31,pmecorrV);
2198             velec            = _mm256_mul_ps(qq31,velec);
2199
2200             d                = _mm256_sub_ps(r31,rswitch);
2201             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2202             d2               = _mm256_mul_ps(d,d);
2203             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)))))));
2204
2205             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2206
2207             /* Evaluate switch function */
2208             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2209             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv31,_mm256_mul_ps(velec,dsw)) );
2210             cutoff_mask      = _mm256_cmp_ps(rsq31,rcutoff2,_CMP_LT_OQ);
2211
2212             fscal            = felec;
2213
2214             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2215
2216             /* Calculate temporary vectorial force */
2217             tx               = _mm256_mul_ps(fscal,dx31);
2218             ty               = _mm256_mul_ps(fscal,dy31);
2219             tz               = _mm256_mul_ps(fscal,dz31);
2220
2221             /* Update vectorial force */
2222             fix3             = _mm256_add_ps(fix3,tx);
2223             fiy3             = _mm256_add_ps(fiy3,ty);
2224             fiz3             = _mm256_add_ps(fiz3,tz);
2225
2226             fjx1             = _mm256_add_ps(fjx1,tx);
2227             fjy1             = _mm256_add_ps(fjy1,ty);
2228             fjz1             = _mm256_add_ps(fjz1,tz);
2229
2230             }
2231
2232             /**************************
2233              * CALCULATE INTERACTIONS *
2234              **************************/
2235
2236             if (gmx_mm256_any_lt(rsq32,rcutoff2))
2237             {
2238
2239             r32              = _mm256_mul_ps(rsq32,rinv32);
2240
2241             /* EWALD ELECTROSTATICS */
2242
2243             /* Analytical PME correction */
2244             zeta2            = _mm256_mul_ps(beta2,rsq32);
2245             rinv3            = _mm256_mul_ps(rinvsq32,rinv32);
2246             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
2247             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2248             felec            = _mm256_mul_ps(qq32,felec);
2249             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
2250             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2251             velec            = _mm256_sub_ps(rinv32,pmecorrV);
2252             velec            = _mm256_mul_ps(qq32,velec);
2253
2254             d                = _mm256_sub_ps(r32,rswitch);
2255             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2256             d2               = _mm256_mul_ps(d,d);
2257             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)))))));
2258
2259             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2260
2261             /* Evaluate switch function */
2262             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2263             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv32,_mm256_mul_ps(velec,dsw)) );
2264             cutoff_mask      = _mm256_cmp_ps(rsq32,rcutoff2,_CMP_LT_OQ);
2265
2266             fscal            = felec;
2267
2268             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2269
2270             /* Calculate temporary vectorial force */
2271             tx               = _mm256_mul_ps(fscal,dx32);
2272             ty               = _mm256_mul_ps(fscal,dy32);
2273             tz               = _mm256_mul_ps(fscal,dz32);
2274
2275             /* Update vectorial force */
2276             fix3             = _mm256_add_ps(fix3,tx);
2277             fiy3             = _mm256_add_ps(fiy3,ty);
2278             fiz3             = _mm256_add_ps(fiz3,tz);
2279
2280             fjx2             = _mm256_add_ps(fjx2,tx);
2281             fjy2             = _mm256_add_ps(fjy2,ty);
2282             fjz2             = _mm256_add_ps(fjz2,tz);
2283
2284             }
2285
2286             /**************************
2287              * CALCULATE INTERACTIONS *
2288              **************************/
2289
2290             if (gmx_mm256_any_lt(rsq33,rcutoff2))
2291             {
2292
2293             r33              = _mm256_mul_ps(rsq33,rinv33);
2294
2295             /* EWALD ELECTROSTATICS */
2296
2297             /* Analytical PME correction */
2298             zeta2            = _mm256_mul_ps(beta2,rsq33);
2299             rinv3            = _mm256_mul_ps(rinvsq33,rinv33);
2300             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
2301             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2302             felec            = _mm256_mul_ps(qq33,felec);
2303             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
2304             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2305             velec            = _mm256_sub_ps(rinv33,pmecorrV);
2306             velec            = _mm256_mul_ps(qq33,velec);
2307
2308             d                = _mm256_sub_ps(r33,rswitch);
2309             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2310             d2               = _mm256_mul_ps(d,d);
2311             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)))))));
2312
2313             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2314
2315             /* Evaluate switch function */
2316             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2317             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv33,_mm256_mul_ps(velec,dsw)) );
2318             cutoff_mask      = _mm256_cmp_ps(rsq33,rcutoff2,_CMP_LT_OQ);
2319
2320             fscal            = felec;
2321
2322             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2323
2324             /* Calculate temporary vectorial force */
2325             tx               = _mm256_mul_ps(fscal,dx33);
2326             ty               = _mm256_mul_ps(fscal,dy33);
2327             tz               = _mm256_mul_ps(fscal,dz33);
2328
2329             /* Update vectorial force */
2330             fix3             = _mm256_add_ps(fix3,tx);
2331             fiy3             = _mm256_add_ps(fiy3,ty);
2332             fiz3             = _mm256_add_ps(fiz3,tz);
2333
2334             fjx3             = _mm256_add_ps(fjx3,tx);
2335             fjy3             = _mm256_add_ps(fjy3,ty);
2336             fjz3             = _mm256_add_ps(fjz3,tz);
2337
2338             }
2339
2340             fjptrA             = f+j_coord_offsetA;
2341             fjptrB             = f+j_coord_offsetB;
2342             fjptrC             = f+j_coord_offsetC;
2343             fjptrD             = f+j_coord_offsetD;
2344             fjptrE             = f+j_coord_offsetE;
2345             fjptrF             = f+j_coord_offsetF;
2346             fjptrG             = f+j_coord_offsetG;
2347             fjptrH             = f+j_coord_offsetH;
2348
2349             gmx_mm256_decrement_3rvec_8ptr_swizzle_ps(fjptrA+DIM,fjptrB+DIM,fjptrC+DIM,fjptrD+DIM,
2350                                                       fjptrE+DIM,fjptrF+DIM,fjptrG+DIM,fjptrH+DIM,
2351                                                       fjx1,fjy1,fjz1,fjx2,fjy2,fjz2,fjx3,fjy3,fjz3);
2352
2353             /* Inner loop uses 945 flops */
2354         }
2355
2356         if(jidx<j_index_end)
2357         {
2358
2359             /* Get j neighbor index, and coordinate index */
2360             jnrlistA         = jjnr[jidx];
2361             jnrlistB         = jjnr[jidx+1];
2362             jnrlistC         = jjnr[jidx+2];
2363             jnrlistD         = jjnr[jidx+3];
2364             jnrlistE         = jjnr[jidx+4];
2365             jnrlistF         = jjnr[jidx+5];
2366             jnrlistG         = jjnr[jidx+6];
2367             jnrlistH         = jjnr[jidx+7];
2368             /* Sign of each element will be negative for non-real atoms.
2369              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
2370              * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
2371              */
2372             dummy_mask = gmx_mm256_set_m128(gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx+4)),_mm_setzero_si128())),
2373                                             gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128())));
2374
2375             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
2376             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
2377             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
2378             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
2379             jnrE       = (jnrlistE>=0) ? jnrlistE : 0;
2380             jnrF       = (jnrlistF>=0) ? jnrlistF : 0;
2381             jnrG       = (jnrlistG>=0) ? jnrlistG : 0;
2382             jnrH       = (jnrlistH>=0) ? jnrlistH : 0;
2383             j_coord_offsetA  = DIM*jnrA;
2384             j_coord_offsetB  = DIM*jnrB;
2385             j_coord_offsetC  = DIM*jnrC;
2386             j_coord_offsetD  = DIM*jnrD;
2387             j_coord_offsetE  = DIM*jnrE;
2388             j_coord_offsetF  = DIM*jnrF;
2389             j_coord_offsetG  = DIM*jnrG;
2390             j_coord_offsetH  = DIM*jnrH;
2391
2392             /* load j atom coordinates */
2393             gmx_mm256_load_3rvec_8ptr_swizzle_ps(x+j_coord_offsetA+DIM,x+j_coord_offsetB+DIM,
2394                                                  x+j_coord_offsetC+DIM,x+j_coord_offsetD+DIM,
2395                                                  x+j_coord_offsetE+DIM,x+j_coord_offsetF+DIM,
2396                                                  x+j_coord_offsetG+DIM,x+j_coord_offsetH+DIM,
2397                                                  &jx1,&jy1,&jz1,&jx2,&jy2,&jz2,&jx3,&jy3,&jz3);
2398
2399             /* Calculate displacement vector */
2400             dx11             = _mm256_sub_ps(ix1,jx1);
2401             dy11             = _mm256_sub_ps(iy1,jy1);
2402             dz11             = _mm256_sub_ps(iz1,jz1);
2403             dx12             = _mm256_sub_ps(ix1,jx2);
2404             dy12             = _mm256_sub_ps(iy1,jy2);
2405             dz12             = _mm256_sub_ps(iz1,jz2);
2406             dx13             = _mm256_sub_ps(ix1,jx3);
2407             dy13             = _mm256_sub_ps(iy1,jy3);
2408             dz13             = _mm256_sub_ps(iz1,jz3);
2409             dx21             = _mm256_sub_ps(ix2,jx1);
2410             dy21             = _mm256_sub_ps(iy2,jy1);
2411             dz21             = _mm256_sub_ps(iz2,jz1);
2412             dx22             = _mm256_sub_ps(ix2,jx2);
2413             dy22             = _mm256_sub_ps(iy2,jy2);
2414             dz22             = _mm256_sub_ps(iz2,jz2);
2415             dx23             = _mm256_sub_ps(ix2,jx3);
2416             dy23             = _mm256_sub_ps(iy2,jy3);
2417             dz23             = _mm256_sub_ps(iz2,jz3);
2418             dx31             = _mm256_sub_ps(ix3,jx1);
2419             dy31             = _mm256_sub_ps(iy3,jy1);
2420             dz31             = _mm256_sub_ps(iz3,jz1);
2421             dx32             = _mm256_sub_ps(ix3,jx2);
2422             dy32             = _mm256_sub_ps(iy3,jy2);
2423             dz32             = _mm256_sub_ps(iz3,jz2);
2424             dx33             = _mm256_sub_ps(ix3,jx3);
2425             dy33             = _mm256_sub_ps(iy3,jy3);
2426             dz33             = _mm256_sub_ps(iz3,jz3);
2427
2428             /* Calculate squared distance and things based on it */
2429             rsq11            = gmx_mm256_calc_rsq_ps(dx11,dy11,dz11);
2430             rsq12            = gmx_mm256_calc_rsq_ps(dx12,dy12,dz12);
2431             rsq13            = gmx_mm256_calc_rsq_ps(dx13,dy13,dz13);
2432             rsq21            = gmx_mm256_calc_rsq_ps(dx21,dy21,dz21);
2433             rsq22            = gmx_mm256_calc_rsq_ps(dx22,dy22,dz22);
2434             rsq23            = gmx_mm256_calc_rsq_ps(dx23,dy23,dz23);
2435             rsq31            = gmx_mm256_calc_rsq_ps(dx31,dy31,dz31);
2436             rsq32            = gmx_mm256_calc_rsq_ps(dx32,dy32,dz32);
2437             rsq33            = gmx_mm256_calc_rsq_ps(dx33,dy33,dz33);
2438
2439             rinv11           = gmx_mm256_invsqrt_ps(rsq11);
2440             rinv12           = gmx_mm256_invsqrt_ps(rsq12);
2441             rinv13           = gmx_mm256_invsqrt_ps(rsq13);
2442             rinv21           = gmx_mm256_invsqrt_ps(rsq21);
2443             rinv22           = gmx_mm256_invsqrt_ps(rsq22);
2444             rinv23           = gmx_mm256_invsqrt_ps(rsq23);
2445             rinv31           = gmx_mm256_invsqrt_ps(rsq31);
2446             rinv32           = gmx_mm256_invsqrt_ps(rsq32);
2447             rinv33           = gmx_mm256_invsqrt_ps(rsq33);
2448
2449             rinvsq11         = _mm256_mul_ps(rinv11,rinv11);
2450             rinvsq12         = _mm256_mul_ps(rinv12,rinv12);
2451             rinvsq13         = _mm256_mul_ps(rinv13,rinv13);
2452             rinvsq21         = _mm256_mul_ps(rinv21,rinv21);
2453             rinvsq22         = _mm256_mul_ps(rinv22,rinv22);
2454             rinvsq23         = _mm256_mul_ps(rinv23,rinv23);
2455             rinvsq31         = _mm256_mul_ps(rinv31,rinv31);
2456             rinvsq32         = _mm256_mul_ps(rinv32,rinv32);
2457             rinvsq33         = _mm256_mul_ps(rinv33,rinv33);
2458
2459             fjx1             = _mm256_setzero_ps();
2460             fjy1             = _mm256_setzero_ps();
2461             fjz1             = _mm256_setzero_ps();
2462             fjx2             = _mm256_setzero_ps();
2463             fjy2             = _mm256_setzero_ps();
2464             fjz2             = _mm256_setzero_ps();
2465             fjx3             = _mm256_setzero_ps();
2466             fjy3             = _mm256_setzero_ps();
2467             fjz3             = _mm256_setzero_ps();
2468
2469             /**************************
2470              * CALCULATE INTERACTIONS *
2471              **************************/
2472
2473             if (gmx_mm256_any_lt(rsq11,rcutoff2))
2474             {
2475
2476             r11              = _mm256_mul_ps(rsq11,rinv11);
2477             r11              = _mm256_andnot_ps(dummy_mask,r11);
2478
2479             /* EWALD ELECTROSTATICS */
2480
2481             /* Analytical PME correction */
2482             zeta2            = _mm256_mul_ps(beta2,rsq11);
2483             rinv3            = _mm256_mul_ps(rinvsq11,rinv11);
2484             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
2485             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2486             felec            = _mm256_mul_ps(qq11,felec);
2487             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
2488             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2489             velec            = _mm256_sub_ps(rinv11,pmecorrV);
2490             velec            = _mm256_mul_ps(qq11,velec);
2491
2492             d                = _mm256_sub_ps(r11,rswitch);
2493             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2494             d2               = _mm256_mul_ps(d,d);
2495             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)))))));
2496
2497             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2498
2499             /* Evaluate switch function */
2500             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2501             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv11,_mm256_mul_ps(velec,dsw)) );
2502             cutoff_mask      = _mm256_cmp_ps(rsq11,rcutoff2,_CMP_LT_OQ);
2503
2504             fscal            = felec;
2505
2506             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2507
2508             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
2509
2510             /* Calculate temporary vectorial force */
2511             tx               = _mm256_mul_ps(fscal,dx11);
2512             ty               = _mm256_mul_ps(fscal,dy11);
2513             tz               = _mm256_mul_ps(fscal,dz11);
2514
2515             /* Update vectorial force */
2516             fix1             = _mm256_add_ps(fix1,tx);
2517             fiy1             = _mm256_add_ps(fiy1,ty);
2518             fiz1             = _mm256_add_ps(fiz1,tz);
2519
2520             fjx1             = _mm256_add_ps(fjx1,tx);
2521             fjy1             = _mm256_add_ps(fjy1,ty);
2522             fjz1             = _mm256_add_ps(fjz1,tz);
2523
2524             }
2525
2526             /**************************
2527              * CALCULATE INTERACTIONS *
2528              **************************/
2529
2530             if (gmx_mm256_any_lt(rsq12,rcutoff2))
2531             {
2532
2533             r12              = _mm256_mul_ps(rsq12,rinv12);
2534             r12              = _mm256_andnot_ps(dummy_mask,r12);
2535
2536             /* EWALD ELECTROSTATICS */
2537
2538             /* Analytical PME correction */
2539             zeta2            = _mm256_mul_ps(beta2,rsq12);
2540             rinv3            = _mm256_mul_ps(rinvsq12,rinv12);
2541             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
2542             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2543             felec            = _mm256_mul_ps(qq12,felec);
2544             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
2545             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2546             velec            = _mm256_sub_ps(rinv12,pmecorrV);
2547             velec            = _mm256_mul_ps(qq12,velec);
2548
2549             d                = _mm256_sub_ps(r12,rswitch);
2550             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2551             d2               = _mm256_mul_ps(d,d);
2552             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)))))));
2553
2554             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2555
2556             /* Evaluate switch function */
2557             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2558             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv12,_mm256_mul_ps(velec,dsw)) );
2559             cutoff_mask      = _mm256_cmp_ps(rsq12,rcutoff2,_CMP_LT_OQ);
2560
2561             fscal            = felec;
2562
2563             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2564
2565             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
2566
2567             /* Calculate temporary vectorial force */
2568             tx               = _mm256_mul_ps(fscal,dx12);
2569             ty               = _mm256_mul_ps(fscal,dy12);
2570             tz               = _mm256_mul_ps(fscal,dz12);
2571
2572             /* Update vectorial force */
2573             fix1             = _mm256_add_ps(fix1,tx);
2574             fiy1             = _mm256_add_ps(fiy1,ty);
2575             fiz1             = _mm256_add_ps(fiz1,tz);
2576
2577             fjx2             = _mm256_add_ps(fjx2,tx);
2578             fjy2             = _mm256_add_ps(fjy2,ty);
2579             fjz2             = _mm256_add_ps(fjz2,tz);
2580
2581             }
2582
2583             /**************************
2584              * CALCULATE INTERACTIONS *
2585              **************************/
2586
2587             if (gmx_mm256_any_lt(rsq13,rcutoff2))
2588             {
2589
2590             r13              = _mm256_mul_ps(rsq13,rinv13);
2591             r13              = _mm256_andnot_ps(dummy_mask,r13);
2592
2593             /* EWALD ELECTROSTATICS */
2594
2595             /* Analytical PME correction */
2596             zeta2            = _mm256_mul_ps(beta2,rsq13);
2597             rinv3            = _mm256_mul_ps(rinvsq13,rinv13);
2598             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
2599             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2600             felec            = _mm256_mul_ps(qq13,felec);
2601             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
2602             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2603             velec            = _mm256_sub_ps(rinv13,pmecorrV);
2604             velec            = _mm256_mul_ps(qq13,velec);
2605
2606             d                = _mm256_sub_ps(r13,rswitch);
2607             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2608             d2               = _mm256_mul_ps(d,d);
2609             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)))))));
2610
2611             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2612
2613             /* Evaluate switch function */
2614             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2615             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv13,_mm256_mul_ps(velec,dsw)) );
2616             cutoff_mask      = _mm256_cmp_ps(rsq13,rcutoff2,_CMP_LT_OQ);
2617
2618             fscal            = felec;
2619
2620             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2621
2622             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
2623
2624             /* Calculate temporary vectorial force */
2625             tx               = _mm256_mul_ps(fscal,dx13);
2626             ty               = _mm256_mul_ps(fscal,dy13);
2627             tz               = _mm256_mul_ps(fscal,dz13);
2628
2629             /* Update vectorial force */
2630             fix1             = _mm256_add_ps(fix1,tx);
2631             fiy1             = _mm256_add_ps(fiy1,ty);
2632             fiz1             = _mm256_add_ps(fiz1,tz);
2633
2634             fjx3             = _mm256_add_ps(fjx3,tx);
2635             fjy3             = _mm256_add_ps(fjy3,ty);
2636             fjz3             = _mm256_add_ps(fjz3,tz);
2637
2638             }
2639
2640             /**************************
2641              * CALCULATE INTERACTIONS *
2642              **************************/
2643
2644             if (gmx_mm256_any_lt(rsq21,rcutoff2))
2645             {
2646
2647             r21              = _mm256_mul_ps(rsq21,rinv21);
2648             r21              = _mm256_andnot_ps(dummy_mask,r21);
2649
2650             /* EWALD ELECTROSTATICS */
2651
2652             /* Analytical PME correction */
2653             zeta2            = _mm256_mul_ps(beta2,rsq21);
2654             rinv3            = _mm256_mul_ps(rinvsq21,rinv21);
2655             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
2656             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2657             felec            = _mm256_mul_ps(qq21,felec);
2658             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
2659             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2660             velec            = _mm256_sub_ps(rinv21,pmecorrV);
2661             velec            = _mm256_mul_ps(qq21,velec);
2662
2663             d                = _mm256_sub_ps(r21,rswitch);
2664             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2665             d2               = _mm256_mul_ps(d,d);
2666             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)))))));
2667
2668             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2669
2670             /* Evaluate switch function */
2671             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2672             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv21,_mm256_mul_ps(velec,dsw)) );
2673             cutoff_mask      = _mm256_cmp_ps(rsq21,rcutoff2,_CMP_LT_OQ);
2674
2675             fscal            = felec;
2676
2677             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2678
2679             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
2680
2681             /* Calculate temporary vectorial force */
2682             tx               = _mm256_mul_ps(fscal,dx21);
2683             ty               = _mm256_mul_ps(fscal,dy21);
2684             tz               = _mm256_mul_ps(fscal,dz21);
2685
2686             /* Update vectorial force */
2687             fix2             = _mm256_add_ps(fix2,tx);
2688             fiy2             = _mm256_add_ps(fiy2,ty);
2689             fiz2             = _mm256_add_ps(fiz2,tz);
2690
2691             fjx1             = _mm256_add_ps(fjx1,tx);
2692             fjy1             = _mm256_add_ps(fjy1,ty);
2693             fjz1             = _mm256_add_ps(fjz1,tz);
2694
2695             }
2696
2697             /**************************
2698              * CALCULATE INTERACTIONS *
2699              **************************/
2700
2701             if (gmx_mm256_any_lt(rsq22,rcutoff2))
2702             {
2703
2704             r22              = _mm256_mul_ps(rsq22,rinv22);
2705             r22              = _mm256_andnot_ps(dummy_mask,r22);
2706
2707             /* EWALD ELECTROSTATICS */
2708
2709             /* Analytical PME correction */
2710             zeta2            = _mm256_mul_ps(beta2,rsq22);
2711             rinv3            = _mm256_mul_ps(rinvsq22,rinv22);
2712             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
2713             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2714             felec            = _mm256_mul_ps(qq22,felec);
2715             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
2716             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2717             velec            = _mm256_sub_ps(rinv22,pmecorrV);
2718             velec            = _mm256_mul_ps(qq22,velec);
2719
2720             d                = _mm256_sub_ps(r22,rswitch);
2721             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2722             d2               = _mm256_mul_ps(d,d);
2723             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)))))));
2724
2725             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2726
2727             /* Evaluate switch function */
2728             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2729             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv22,_mm256_mul_ps(velec,dsw)) );
2730             cutoff_mask      = _mm256_cmp_ps(rsq22,rcutoff2,_CMP_LT_OQ);
2731
2732             fscal            = felec;
2733
2734             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2735
2736             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
2737
2738             /* Calculate temporary vectorial force */
2739             tx               = _mm256_mul_ps(fscal,dx22);
2740             ty               = _mm256_mul_ps(fscal,dy22);
2741             tz               = _mm256_mul_ps(fscal,dz22);
2742
2743             /* Update vectorial force */
2744             fix2             = _mm256_add_ps(fix2,tx);
2745             fiy2             = _mm256_add_ps(fiy2,ty);
2746             fiz2             = _mm256_add_ps(fiz2,tz);
2747
2748             fjx2             = _mm256_add_ps(fjx2,tx);
2749             fjy2             = _mm256_add_ps(fjy2,ty);
2750             fjz2             = _mm256_add_ps(fjz2,tz);
2751
2752             }
2753
2754             /**************************
2755              * CALCULATE INTERACTIONS *
2756              **************************/
2757
2758             if (gmx_mm256_any_lt(rsq23,rcutoff2))
2759             {
2760
2761             r23              = _mm256_mul_ps(rsq23,rinv23);
2762             r23              = _mm256_andnot_ps(dummy_mask,r23);
2763
2764             /* EWALD ELECTROSTATICS */
2765
2766             /* Analytical PME correction */
2767             zeta2            = _mm256_mul_ps(beta2,rsq23);
2768             rinv3            = _mm256_mul_ps(rinvsq23,rinv23);
2769             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
2770             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2771             felec            = _mm256_mul_ps(qq23,felec);
2772             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
2773             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2774             velec            = _mm256_sub_ps(rinv23,pmecorrV);
2775             velec            = _mm256_mul_ps(qq23,velec);
2776
2777             d                = _mm256_sub_ps(r23,rswitch);
2778             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2779             d2               = _mm256_mul_ps(d,d);
2780             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)))))));
2781
2782             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2783
2784             /* Evaluate switch function */
2785             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2786             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv23,_mm256_mul_ps(velec,dsw)) );
2787             cutoff_mask      = _mm256_cmp_ps(rsq23,rcutoff2,_CMP_LT_OQ);
2788
2789             fscal            = felec;
2790
2791             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2792
2793             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
2794
2795             /* Calculate temporary vectorial force */
2796             tx               = _mm256_mul_ps(fscal,dx23);
2797             ty               = _mm256_mul_ps(fscal,dy23);
2798             tz               = _mm256_mul_ps(fscal,dz23);
2799
2800             /* Update vectorial force */
2801             fix2             = _mm256_add_ps(fix2,tx);
2802             fiy2             = _mm256_add_ps(fiy2,ty);
2803             fiz2             = _mm256_add_ps(fiz2,tz);
2804
2805             fjx3             = _mm256_add_ps(fjx3,tx);
2806             fjy3             = _mm256_add_ps(fjy3,ty);
2807             fjz3             = _mm256_add_ps(fjz3,tz);
2808
2809             }
2810
2811             /**************************
2812              * CALCULATE INTERACTIONS *
2813              **************************/
2814
2815             if (gmx_mm256_any_lt(rsq31,rcutoff2))
2816             {
2817
2818             r31              = _mm256_mul_ps(rsq31,rinv31);
2819             r31              = _mm256_andnot_ps(dummy_mask,r31);
2820
2821             /* EWALD ELECTROSTATICS */
2822
2823             /* Analytical PME correction */
2824             zeta2            = _mm256_mul_ps(beta2,rsq31);
2825             rinv3            = _mm256_mul_ps(rinvsq31,rinv31);
2826             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
2827             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2828             felec            = _mm256_mul_ps(qq31,felec);
2829             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
2830             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2831             velec            = _mm256_sub_ps(rinv31,pmecorrV);
2832             velec            = _mm256_mul_ps(qq31,velec);
2833
2834             d                = _mm256_sub_ps(r31,rswitch);
2835             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2836             d2               = _mm256_mul_ps(d,d);
2837             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)))))));
2838
2839             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2840
2841             /* Evaluate switch function */
2842             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2843             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv31,_mm256_mul_ps(velec,dsw)) );
2844             cutoff_mask      = _mm256_cmp_ps(rsq31,rcutoff2,_CMP_LT_OQ);
2845
2846             fscal            = felec;
2847
2848             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2849
2850             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
2851
2852             /* Calculate temporary vectorial force */
2853             tx               = _mm256_mul_ps(fscal,dx31);
2854             ty               = _mm256_mul_ps(fscal,dy31);
2855             tz               = _mm256_mul_ps(fscal,dz31);
2856
2857             /* Update vectorial force */
2858             fix3             = _mm256_add_ps(fix3,tx);
2859             fiy3             = _mm256_add_ps(fiy3,ty);
2860             fiz3             = _mm256_add_ps(fiz3,tz);
2861
2862             fjx1             = _mm256_add_ps(fjx1,tx);
2863             fjy1             = _mm256_add_ps(fjy1,ty);
2864             fjz1             = _mm256_add_ps(fjz1,tz);
2865
2866             }
2867
2868             /**************************
2869              * CALCULATE INTERACTIONS *
2870              **************************/
2871
2872             if (gmx_mm256_any_lt(rsq32,rcutoff2))
2873             {
2874
2875             r32              = _mm256_mul_ps(rsq32,rinv32);
2876             r32              = _mm256_andnot_ps(dummy_mask,r32);
2877
2878             /* EWALD ELECTROSTATICS */
2879
2880             /* Analytical PME correction */
2881             zeta2            = _mm256_mul_ps(beta2,rsq32);
2882             rinv3            = _mm256_mul_ps(rinvsq32,rinv32);
2883             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
2884             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2885             felec            = _mm256_mul_ps(qq32,felec);
2886             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
2887             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2888             velec            = _mm256_sub_ps(rinv32,pmecorrV);
2889             velec            = _mm256_mul_ps(qq32,velec);
2890
2891             d                = _mm256_sub_ps(r32,rswitch);
2892             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2893             d2               = _mm256_mul_ps(d,d);
2894             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)))))));
2895
2896             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2897
2898             /* Evaluate switch function */
2899             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2900             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv32,_mm256_mul_ps(velec,dsw)) );
2901             cutoff_mask      = _mm256_cmp_ps(rsq32,rcutoff2,_CMP_LT_OQ);
2902
2903             fscal            = felec;
2904
2905             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2906
2907             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
2908
2909             /* Calculate temporary vectorial force */
2910             tx               = _mm256_mul_ps(fscal,dx32);
2911             ty               = _mm256_mul_ps(fscal,dy32);
2912             tz               = _mm256_mul_ps(fscal,dz32);
2913
2914             /* Update vectorial force */
2915             fix3             = _mm256_add_ps(fix3,tx);
2916             fiy3             = _mm256_add_ps(fiy3,ty);
2917             fiz3             = _mm256_add_ps(fiz3,tz);
2918
2919             fjx2             = _mm256_add_ps(fjx2,tx);
2920             fjy2             = _mm256_add_ps(fjy2,ty);
2921             fjz2             = _mm256_add_ps(fjz2,tz);
2922
2923             }
2924
2925             /**************************
2926              * CALCULATE INTERACTIONS *
2927              **************************/
2928
2929             if (gmx_mm256_any_lt(rsq33,rcutoff2))
2930             {
2931
2932             r33              = _mm256_mul_ps(rsq33,rinv33);
2933             r33              = _mm256_andnot_ps(dummy_mask,r33);
2934
2935             /* EWALD ELECTROSTATICS */
2936
2937             /* Analytical PME correction */
2938             zeta2            = _mm256_mul_ps(beta2,rsq33);
2939             rinv3            = _mm256_mul_ps(rinvsq33,rinv33);
2940             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
2941             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2942             felec            = _mm256_mul_ps(qq33,felec);
2943             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
2944             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2945             velec            = _mm256_sub_ps(rinv33,pmecorrV);
2946             velec            = _mm256_mul_ps(qq33,velec);
2947
2948             d                = _mm256_sub_ps(r33,rswitch);
2949             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2950             d2               = _mm256_mul_ps(d,d);
2951             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)))))));
2952
2953             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2954
2955             /* Evaluate switch function */
2956             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2957             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv33,_mm256_mul_ps(velec,dsw)) );
2958             cutoff_mask      = _mm256_cmp_ps(rsq33,rcutoff2,_CMP_LT_OQ);
2959
2960             fscal            = felec;
2961
2962             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2963
2964             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
2965
2966             /* Calculate temporary vectorial force */
2967             tx               = _mm256_mul_ps(fscal,dx33);
2968             ty               = _mm256_mul_ps(fscal,dy33);
2969             tz               = _mm256_mul_ps(fscal,dz33);
2970
2971             /* Update vectorial force */
2972             fix3             = _mm256_add_ps(fix3,tx);
2973             fiy3             = _mm256_add_ps(fiy3,ty);
2974             fiz3             = _mm256_add_ps(fiz3,tz);
2975
2976             fjx3             = _mm256_add_ps(fjx3,tx);
2977             fjy3             = _mm256_add_ps(fjy3,ty);
2978             fjz3             = _mm256_add_ps(fjz3,tz);
2979
2980             }
2981
2982             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
2983             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
2984             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
2985             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
2986             fjptrE             = (jnrlistE>=0) ? f+j_coord_offsetE : scratch;
2987             fjptrF             = (jnrlistF>=0) ? f+j_coord_offsetF : scratch;
2988             fjptrG             = (jnrlistG>=0) ? f+j_coord_offsetG : scratch;
2989             fjptrH             = (jnrlistH>=0) ? f+j_coord_offsetH : scratch;
2990
2991             gmx_mm256_decrement_3rvec_8ptr_swizzle_ps(fjptrA+DIM,fjptrB+DIM,fjptrC+DIM,fjptrD+DIM,
2992                                                       fjptrE+DIM,fjptrF+DIM,fjptrG+DIM,fjptrH+DIM,
2993                                                       fjx1,fjy1,fjz1,fjx2,fjy2,fjz2,fjx3,fjy3,fjz3);
2994
2995             /* Inner loop uses 954 flops */
2996         }
2997
2998         /* End of innermost loop */
2999
3000         gmx_mm256_update_iforce_3atom_swizzle_ps(fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
3001                                                  f+i_coord_offset+DIM,fshift+i_shift_offset);
3002
3003         /* Increment number of inner iterations */
3004         inneriter                  += j_index_end - j_index_start;
3005
3006         /* Outer loop uses 18 flops */
3007     }
3008
3009     /* Increment number of outer iterations */
3010     outeriter        += nri;
3011
3012     /* Update outer/inner flops */
3013
3014     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_W4W4_F,outeriter*18 + inneriter*954);
3015 }