src/gromacs/gmxlib/nonbonded/nb_kernel_avx_256_single/nb_kernel_ElecEwSw_VdwNone_GeomW3W3_avx_256_single.cpp

   1 /*
   2  * This file is part of the GROMACS molecular simulation package.
   3  *
   4  * Copyright (c) 2012,2013,2014,2015,2017,2018, by the GROMACS development team, led by
   5  * Mark Abraham, David van der Spoel, Berk Hess, and Erik Lindahl,
   6  * and including many others, as listed in the AUTHORS file in the
   7  * top-level source directory and at http://www.gromacs.org.
   8  *
   9  * GROMACS is free software; you can redistribute it and/or
  10  * modify it under the terms of the GNU Lesser General Public License
  11  * as published by the Free Software Foundation; either version 2.1
  12  * of the License, or (at your option) any later version.
  13  *
  14  * GROMACS is distributed in the hope that it will be useful,
  15  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  16  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  17  * Lesser General Public License for more details.
  18  *
  19  * You should have received a copy of the GNU Lesser General Public
  20  * License along with GROMACS; if not, see
  21  * http://www.gnu.org/licenses, or write to the Free Software Foundation,
  22  * Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA.
  23  *
  24  * If you want to redistribute modifications to GROMACS, please
  25  * consider that scientific software is very special. Version
  26  * control is crucial - bugs must be traceable. We will be happy to
  27  * consider code for inclusion in the official distribution, but
  28  * derived work must not be called official GROMACS. Details are found
  29  * in the README & COPYING files - if they are missing, get the
  30  * official version at http://www.gromacs.org.
  31  *
  32  * To help us fund GROMACS development, we humbly ask that you cite
  33  * the research papers on the package. Check out http://www.gromacs.org.
  34  */
  35 /*
  36  * Note: this file was generated by the GROMACS avx_256_single kernel generator.
  37  */
  38 #include "gmxpre.h"
  39
  40 #include "config.h"
  41
  42 #include <math.h>
  43
  44 #include "../nb_kernel.h"
  45 #include "gromacs/gmxlib/nrnb.h"
  46
  47 #include "kernelutil_x86_avx_256_single.h"
  48
  49 /*
  50  * Gromacs nonbonded kernel:   nb_kernel_ElecEwSw_VdwNone_GeomW3W3_VF_avx_256_single
  51  * Electrostatics interaction: Ewald
  52  * VdW interaction:            None
  53  * Geometry:                   Water3-Water3
  54  * Calculate force/pot:        PotentialAndForce
  55  */
  56 void
  57 nb_kernel_ElecEwSw_VdwNone_GeomW3W3_VF_avx_256_single
  58                     (t_nblist                    * gmx_restrict       nlist,
  59                      rvec                        * gmx_restrict          xx,
  60                      rvec                        * gmx_restrict          ff,
  61                      struct t_forcerec           * gmx_restrict          fr,
  62                      t_mdatoms                   * gmx_restrict     mdatoms,
  63                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
  64                      t_nrnb                      * gmx_restrict        nrnb)
  65 {
  66     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
  67      * just 0 for non-waters.
  68      * Suffixes A,B,C,D,E,F,G,H refer to j loop unrolling done with AVX, e.g. for the eight different
  69      * jnr indices corresponding to data put in the four positions in the SIMD register.
  70      */
  71     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
  72     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
  73     int              jnrA,jnrB,jnrC,jnrD;
  74     int              jnrE,jnrF,jnrG,jnrH;
  75     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
  76     int              jnrlistE,jnrlistF,jnrlistG,jnrlistH;
  77     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
  78     int              j_coord_offsetE,j_coord_offsetF,j_coord_offsetG,j_coord_offsetH;
  79     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  80     real             rcutoff_scalar;
  81     real             *shiftvec,*fshift,*x,*f;
  82     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD,*fjptrE,*fjptrF,*fjptrG,*fjptrH;
  83     real             scratch[4*DIM];
  84     __m256           tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
  85     real *           vdwioffsetptr0;
  86     __m256           ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  87     real *           vdwioffsetptr1;
  88     __m256           ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
  89     real *           vdwioffsetptr2;
  90     __m256           ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
  91     int              vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D,vdwjidx0E,vdwjidx0F,vdwjidx0G,vdwjidx0H;
  92     __m256           jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  93     int              vdwjidx1A,vdwjidx1B,vdwjidx1C,vdwjidx1D,vdwjidx1E,vdwjidx1F,vdwjidx1G,vdwjidx1H;
  94     __m256           jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
  95     int              vdwjidx2A,vdwjidx2B,vdwjidx2C,vdwjidx2D,vdwjidx2E,vdwjidx2F,vdwjidx2G,vdwjidx2H;
  96     __m256           jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
  97     __m256           dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
  98     __m256           dx01,dy01,dz01,rsq01,rinv01,rinvsq01,r01,qq01,c6_01,c12_01;
  99     __m256           dx02,dy02,dz02,rsq02,rinv02,rinvsq02,r02,qq02,c6_02,c12_02;
 100     __m256           dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
 101     __m256           dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11;
 102     __m256           dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12;
 103     __m256           dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
 104     __m256           dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21;
 105     __m256           dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22;
 106     __m256           velec,felec,velecsum,facel,crf,krf,krf2;
 107     real             *charge;
 108     __m256i          ewitab;
 109     __m128i          ewitab_lo,ewitab_hi;
 110     __m256           ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
 111     __m256           beta,beta2,beta3,zeta2,pmecorrF,pmecorrV,rinv3;
 112     real             *ewtab;
 113     __m256           rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
 114     real             rswitch_scalar,d_scalar;
 115     __m256           dummy_mask,cutoff_mask;
 116     __m256           signbit = _mm256_castsi256_ps( _mm256_set1_epi32(0x80000000) );
 117     __m256           one     = _mm256_set1_ps(1.0);
 118     __m256           two     = _mm256_set1_ps(2.0);
 119     x                = xx[0];
 120     f                = ff[0];
 121
 122     nri              = nlist->nri;
 123     iinr             = nlist->iinr;
 124     jindex           = nlist->jindex;
 125     jjnr             = nlist->jjnr;
 126     shiftidx         = nlist->shift;
 127     gid              = nlist->gid;
 128     shiftvec         = fr->shift_vec[0];
 129     fshift           = fr->fshift[0];
 130     facel            = _mm256_set1_ps(fr->ic->epsfac);
 131     charge           = mdatoms->chargeA;
 132
 133     sh_ewald         = _mm256_set1_ps(fr->ic->sh_ewald);
 134     beta             = _mm256_set1_ps(fr->ic->ewaldcoeff_q);
 135     beta2            = _mm256_mul_ps(beta,beta);
 136     beta3            = _mm256_mul_ps(beta,beta2);
 137
 138     ewtab            = fr->ic->tabq_coul_FDV0;
 139     ewtabscale       = _mm256_set1_ps(fr->ic->tabq_scale);
 140     ewtabhalfspace   = _mm256_set1_ps(0.5/fr->ic->tabq_scale);
 141
 142     /* Setup water-specific parameters */
 143     inr              = nlist->iinr[0];
 144     iq0              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+0]));
 145     iq1              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+1]));
 146     iq2              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+2]));
 147
 148     jq0              = _mm256_set1_ps(charge[inr+0]);
 149     jq1              = _mm256_set1_ps(charge[inr+1]);
 150     jq2              = _mm256_set1_ps(charge[inr+2]);
 151     qq00             = _mm256_mul_ps(iq0,jq0);
 152     qq01             = _mm256_mul_ps(iq0,jq1);
 153     qq02             = _mm256_mul_ps(iq0,jq2);
 154     qq10             = _mm256_mul_ps(iq1,jq0);
 155     qq11             = _mm256_mul_ps(iq1,jq1);
 156     qq12             = _mm256_mul_ps(iq1,jq2);
 157     qq20             = _mm256_mul_ps(iq2,jq0);
 158     qq21             = _mm256_mul_ps(iq2,jq1);
 159     qq22             = _mm256_mul_ps(iq2,jq2);
 160
 161     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 162     rcutoff_scalar   = fr->ic->rcoulomb;
 163     rcutoff          = _mm256_set1_ps(rcutoff_scalar);
 164     rcutoff2         = _mm256_mul_ps(rcutoff,rcutoff);
 165
 166     rswitch_scalar   = fr->ic->rcoulomb_switch;
 167     rswitch          = _mm256_set1_ps(rswitch_scalar);
 168     /* Setup switch parameters */
 169     d_scalar         = rcutoff_scalar-rswitch_scalar;
 170     d                = _mm256_set1_ps(d_scalar);
 171     swV3             = _mm256_set1_ps(-10.0/(d_scalar*d_scalar*d_scalar));
 172     swV4             = _mm256_set1_ps( 15.0/(d_scalar*d_scalar*d_scalar*d_scalar));
 173     swV5             = _mm256_set1_ps( -6.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
 174     swF2             = _mm256_set1_ps(-30.0/(d_scalar*d_scalar*d_scalar));
 175     swF3             = _mm256_set1_ps( 60.0/(d_scalar*d_scalar*d_scalar*d_scalar));
 176     swF4             = _mm256_set1_ps(-30.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
 177
 178     /* Avoid stupid compiler warnings */
 179     jnrA = jnrB = jnrC = jnrD = jnrE = jnrF = jnrG = jnrH = 0;
 180     j_coord_offsetA = 0;
 181     j_coord_offsetB = 0;
 182     j_coord_offsetC = 0;
 183     j_coord_offsetD = 0;
 184     j_coord_offsetE = 0;
 185     j_coord_offsetF = 0;
 186     j_coord_offsetG = 0;
 187     j_coord_offsetH = 0;
 188
 189     outeriter        = 0;
 190     inneriter        = 0;
 191
 192     for(iidx=0;iidx<4*DIM;iidx++)
 193     {
 194         scratch[iidx] = 0.0;
 195     }
 196
 197     /* Start outer loop over neighborlists */
 198     for(iidx=0; iidx<nri; iidx++)
 199     {
 200         /* Load shift vector for this list */
 201         i_shift_offset   = DIM*shiftidx[iidx];
 202
 203         /* Load limits for loop over neighbors */
 204         j_index_start    = jindex[iidx];
 205         j_index_end      = jindex[iidx+1];
 206
 207         /* Get outer coordinate index */
 208         inr              = iinr[iidx];
 209         i_coord_offset   = DIM*inr;
 210
 211         /* Load i particle coords and add shift vector */
 212         gmx_mm256_load_shift_and_3rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,
 213                                                     &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2);
 214
 215         fix0             = _mm256_setzero_ps();
 216         fiy0             = _mm256_setzero_ps();
 217         fiz0             = _mm256_setzero_ps();
 218         fix1             = _mm256_setzero_ps();
 219         fiy1             = _mm256_setzero_ps();
 220         fiz1             = _mm256_setzero_ps();
 221         fix2             = _mm256_setzero_ps();
 222         fiy2             = _mm256_setzero_ps();
 223         fiz2             = _mm256_setzero_ps();
 224
 225         /* Reset potential sums */
 226         velecsum         = _mm256_setzero_ps();
 227
 228         /* Start inner kernel loop */
 229         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+7]>=0; jidx+=8)
 230         {
 231
 232             /* Get j neighbor index, and coordinate index */
 233             jnrA             = jjnr[jidx];
 234             jnrB             = jjnr[jidx+1];
 235             jnrC             = jjnr[jidx+2];
 236             jnrD             = jjnr[jidx+3];
 237             jnrE             = jjnr[jidx+4];
 238             jnrF             = jjnr[jidx+5];
 239             jnrG             = jjnr[jidx+6];
 240             jnrH             = jjnr[jidx+7];
 241             j_coord_offsetA  = DIM*jnrA;
 242             j_coord_offsetB  = DIM*jnrB;
 243             j_coord_offsetC  = DIM*jnrC;
 244             j_coord_offsetD  = DIM*jnrD;
 245             j_coord_offsetE  = DIM*jnrE;
 246             j_coord_offsetF  = DIM*jnrF;
 247             j_coord_offsetG  = DIM*jnrG;
 248             j_coord_offsetH  = DIM*jnrH;
 249
 250             /* load j atom coordinates */
 251             gmx_mm256_load_3rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 252                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 253                                                  x+j_coord_offsetE,x+j_coord_offsetF,
 254                                                  x+j_coord_offsetG,x+j_coord_offsetH,
 255                                               &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,&jy2,&jz2);
 256
 257             /* Calculate displacement vector */
 258             dx00             = _mm256_sub_ps(ix0,jx0);
 259             dy00             = _mm256_sub_ps(iy0,jy0);
 260             dz00             = _mm256_sub_ps(iz0,jz0);
 261             dx01             = _mm256_sub_ps(ix0,jx1);
 262             dy01             = _mm256_sub_ps(iy0,jy1);
 263             dz01             = _mm256_sub_ps(iz0,jz1);
 264             dx02             = _mm256_sub_ps(ix0,jx2);
 265             dy02             = _mm256_sub_ps(iy0,jy2);
 266             dz02             = _mm256_sub_ps(iz0,jz2);
 267             dx10             = _mm256_sub_ps(ix1,jx0);
 268             dy10             = _mm256_sub_ps(iy1,jy0);
 269             dz10             = _mm256_sub_ps(iz1,jz0);
 270             dx11             = _mm256_sub_ps(ix1,jx1);
 271             dy11             = _mm256_sub_ps(iy1,jy1);
 272             dz11             = _mm256_sub_ps(iz1,jz1);
 273             dx12             = _mm256_sub_ps(ix1,jx2);
 274             dy12             = _mm256_sub_ps(iy1,jy2);
 275             dz12             = _mm256_sub_ps(iz1,jz2);
 276             dx20             = _mm256_sub_ps(ix2,jx0);
 277             dy20             = _mm256_sub_ps(iy2,jy0);
 278             dz20             = _mm256_sub_ps(iz2,jz0);
 279             dx21             = _mm256_sub_ps(ix2,jx1);
 280             dy21             = _mm256_sub_ps(iy2,jy1);
 281             dz21             = _mm256_sub_ps(iz2,jz1);
 282             dx22             = _mm256_sub_ps(ix2,jx2);
 283             dy22             = _mm256_sub_ps(iy2,jy2);
 284             dz22             = _mm256_sub_ps(iz2,jz2);
 285
 286             /* Calculate squared distance and things based on it */
 287             rsq00            = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
 288             rsq01            = gmx_mm256_calc_rsq_ps(dx01,dy01,dz01);
 289             rsq02            = gmx_mm256_calc_rsq_ps(dx02,dy02,dz02);
 290             rsq10            = gmx_mm256_calc_rsq_ps(dx10,dy10,dz10);
 291             rsq11            = gmx_mm256_calc_rsq_ps(dx11,dy11,dz11);
 292             rsq12            = gmx_mm256_calc_rsq_ps(dx12,dy12,dz12);
 293             rsq20            = gmx_mm256_calc_rsq_ps(dx20,dy20,dz20);
 294             rsq21            = gmx_mm256_calc_rsq_ps(dx21,dy21,dz21);
 295             rsq22            = gmx_mm256_calc_rsq_ps(dx22,dy22,dz22);
 296
 297             rinv00           = avx256_invsqrt_f(rsq00);
 298             rinv01           = avx256_invsqrt_f(rsq01);
 299             rinv02           = avx256_invsqrt_f(rsq02);
 300             rinv10           = avx256_invsqrt_f(rsq10);
 301             rinv11           = avx256_invsqrt_f(rsq11);
 302             rinv12           = avx256_invsqrt_f(rsq12);
 303             rinv20           = avx256_invsqrt_f(rsq20);
 304             rinv21           = avx256_invsqrt_f(rsq21);
 305             rinv22           = avx256_invsqrt_f(rsq22);
 306
 307             rinvsq00         = _mm256_mul_ps(rinv00,rinv00);
 308             rinvsq01         = _mm256_mul_ps(rinv01,rinv01);
 309             rinvsq02         = _mm256_mul_ps(rinv02,rinv02);
 310             rinvsq10         = _mm256_mul_ps(rinv10,rinv10);
 311             rinvsq11         = _mm256_mul_ps(rinv11,rinv11);
 312             rinvsq12         = _mm256_mul_ps(rinv12,rinv12);
 313             rinvsq20         = _mm256_mul_ps(rinv20,rinv20);
 314             rinvsq21         = _mm256_mul_ps(rinv21,rinv21);
 315             rinvsq22         = _mm256_mul_ps(rinv22,rinv22);
 316
 317             fjx0             = _mm256_setzero_ps();
 318             fjy0             = _mm256_setzero_ps();
 319             fjz0             = _mm256_setzero_ps();
 320             fjx1             = _mm256_setzero_ps();
 321             fjy1             = _mm256_setzero_ps();
 322             fjz1             = _mm256_setzero_ps();
 323             fjx2             = _mm256_setzero_ps();
 324             fjy2             = _mm256_setzero_ps();
 325             fjz2             = _mm256_setzero_ps();
 326
 327             /**************************
 328              * CALCULATE INTERACTIONS *
 329              **************************/
 330
 331             if (gmx_mm256_any_lt(rsq00,rcutoff2))
 332             {
 333
 334             r00              = _mm256_mul_ps(rsq00,rinv00);
 335
 336             /* EWALD ELECTROSTATICS */
 337
 338             /* Analytical PME correction */
 339             zeta2            = _mm256_mul_ps(beta2,rsq00);
 340             rinv3            = _mm256_mul_ps(rinvsq00,rinv00);
 341             pmecorrF         = avx256_pmecorrF_f(zeta2);
 342             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 343             felec            = _mm256_mul_ps(qq00,felec);
 344             pmecorrV         = avx256_pmecorrV_f(zeta2);
 345             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 346             velec            = _mm256_sub_ps(rinv00,pmecorrV);
 347             velec            = _mm256_mul_ps(qq00,velec);
 348
 349             d                = _mm256_sub_ps(r00,rswitch);
 350             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 351             d2               = _mm256_mul_ps(d,d);
 352             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)))))));
 353
 354             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 355
 356             /* Evaluate switch function */
 357             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 358             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv00,_mm256_mul_ps(velec,dsw)) );
 359             velec            = _mm256_mul_ps(velec,sw);
 360             cutoff_mask      = _mm256_cmp_ps(rsq00,rcutoff2,_CMP_LT_OQ);
 361
 362             /* Update potential sum for this i atom from the interaction with this j atom. */
 363             velec            = _mm256_and_ps(velec,cutoff_mask);
 364             velecsum         = _mm256_add_ps(velecsum,velec);
 365
 366             fscal            = felec;
 367
 368             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 369
 370             /* Calculate temporary vectorial force */
 371             tx               = _mm256_mul_ps(fscal,dx00);
 372             ty               = _mm256_mul_ps(fscal,dy00);
 373             tz               = _mm256_mul_ps(fscal,dz00);
 374
 375             /* Update vectorial force */
 376             fix0             = _mm256_add_ps(fix0,tx);
 377             fiy0             = _mm256_add_ps(fiy0,ty);
 378             fiz0             = _mm256_add_ps(fiz0,tz);
 379
 380             fjx0             = _mm256_add_ps(fjx0,tx);
 381             fjy0             = _mm256_add_ps(fjy0,ty);
 382             fjz0             = _mm256_add_ps(fjz0,tz);
 383
 384             }
 385
 386             /**************************
 387              * CALCULATE INTERACTIONS *
 388              **************************/
 389
 390             if (gmx_mm256_any_lt(rsq01,rcutoff2))
 391             {
 392
 393             r01              = _mm256_mul_ps(rsq01,rinv01);
 394
 395             /* EWALD ELECTROSTATICS */
 396
 397             /* Analytical PME correction */
 398             zeta2            = _mm256_mul_ps(beta2,rsq01);
 399             rinv3            = _mm256_mul_ps(rinvsq01,rinv01);
 400             pmecorrF         = avx256_pmecorrF_f(zeta2);
 401             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 402             felec            = _mm256_mul_ps(qq01,felec);
 403             pmecorrV         = avx256_pmecorrV_f(zeta2);
 404             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 405             velec            = _mm256_sub_ps(rinv01,pmecorrV);
 406             velec            = _mm256_mul_ps(qq01,velec);
 407
 408             d                = _mm256_sub_ps(r01,rswitch);
 409             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 410             d2               = _mm256_mul_ps(d,d);
 411             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)))))));
 412
 413             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 414
 415             /* Evaluate switch function */
 416             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 417             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv01,_mm256_mul_ps(velec,dsw)) );
 418             velec            = _mm256_mul_ps(velec,sw);
 419             cutoff_mask      = _mm256_cmp_ps(rsq01,rcutoff2,_CMP_LT_OQ);
 420
 421             /* Update potential sum for this i atom from the interaction with this j atom. */
 422             velec            = _mm256_and_ps(velec,cutoff_mask);
 423             velecsum         = _mm256_add_ps(velecsum,velec);
 424
 425             fscal            = felec;
 426
 427             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 428
 429             /* Calculate temporary vectorial force */
 430             tx               = _mm256_mul_ps(fscal,dx01);
 431             ty               = _mm256_mul_ps(fscal,dy01);
 432             tz               = _mm256_mul_ps(fscal,dz01);
 433
 434             /* Update vectorial force */
 435             fix0             = _mm256_add_ps(fix0,tx);
 436             fiy0             = _mm256_add_ps(fiy0,ty);
 437             fiz0             = _mm256_add_ps(fiz0,tz);
 438
 439             fjx1             = _mm256_add_ps(fjx1,tx);
 440             fjy1             = _mm256_add_ps(fjy1,ty);
 441             fjz1             = _mm256_add_ps(fjz1,tz);
 442
 443             }
 444
 445             /**************************
 446              * CALCULATE INTERACTIONS *
 447              **************************/
 448
 449             if (gmx_mm256_any_lt(rsq02,rcutoff2))
 450             {
 451
 452             r02              = _mm256_mul_ps(rsq02,rinv02);
 453
 454             /* EWALD ELECTROSTATICS */
 455
 456             /* Analytical PME correction */
 457             zeta2            = _mm256_mul_ps(beta2,rsq02);
 458             rinv3            = _mm256_mul_ps(rinvsq02,rinv02);
 459             pmecorrF         = avx256_pmecorrF_f(zeta2);
 460             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 461             felec            = _mm256_mul_ps(qq02,felec);
 462             pmecorrV         = avx256_pmecorrV_f(zeta2);
 463             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 464             velec            = _mm256_sub_ps(rinv02,pmecorrV);
 465             velec            = _mm256_mul_ps(qq02,velec);
 466
 467             d                = _mm256_sub_ps(r02,rswitch);
 468             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 469             d2               = _mm256_mul_ps(d,d);
 470             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)))))));
 471
 472             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 473
 474             /* Evaluate switch function */
 475             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 476             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv02,_mm256_mul_ps(velec,dsw)) );
 477             velec            = _mm256_mul_ps(velec,sw);
 478             cutoff_mask      = _mm256_cmp_ps(rsq02,rcutoff2,_CMP_LT_OQ);
 479
 480             /* Update potential sum for this i atom from the interaction with this j atom. */
 481             velec            = _mm256_and_ps(velec,cutoff_mask);
 482             velecsum         = _mm256_add_ps(velecsum,velec);
 483
 484             fscal            = felec;
 485
 486             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 487
 488             /* Calculate temporary vectorial force */
 489             tx               = _mm256_mul_ps(fscal,dx02);
 490             ty               = _mm256_mul_ps(fscal,dy02);
 491             tz               = _mm256_mul_ps(fscal,dz02);
 492
 493             /* Update vectorial force */
 494             fix0             = _mm256_add_ps(fix0,tx);
 495             fiy0             = _mm256_add_ps(fiy0,ty);
 496             fiz0             = _mm256_add_ps(fiz0,tz);
 497
 498             fjx2             = _mm256_add_ps(fjx2,tx);
 499             fjy2             = _mm256_add_ps(fjy2,ty);
 500             fjz2             = _mm256_add_ps(fjz2,tz);
 501
 502             }
 503
 504             /**************************
 505              * CALCULATE INTERACTIONS *
 506              **************************/
 507
 508             if (gmx_mm256_any_lt(rsq10,rcutoff2))
 509             {
 510
 511             r10              = _mm256_mul_ps(rsq10,rinv10);
 512
 513             /* EWALD ELECTROSTATICS */
 514
 515             /* Analytical PME correction */
 516             zeta2            = _mm256_mul_ps(beta2,rsq10);
 517             rinv3            = _mm256_mul_ps(rinvsq10,rinv10);
 518             pmecorrF         = avx256_pmecorrF_f(zeta2);
 519             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 520             felec            = _mm256_mul_ps(qq10,felec);
 521             pmecorrV         = avx256_pmecorrV_f(zeta2);
 522             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 523             velec            = _mm256_sub_ps(rinv10,pmecorrV);
 524             velec            = _mm256_mul_ps(qq10,velec);
 525
 526             d                = _mm256_sub_ps(r10,rswitch);
 527             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 528             d2               = _mm256_mul_ps(d,d);
 529             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)))))));
 530
 531             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 532
 533             /* Evaluate switch function */
 534             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 535             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv10,_mm256_mul_ps(velec,dsw)) );
 536             velec            = _mm256_mul_ps(velec,sw);
 537             cutoff_mask      = _mm256_cmp_ps(rsq10,rcutoff2,_CMP_LT_OQ);
 538
 539             /* Update potential sum for this i atom from the interaction with this j atom. */
 540             velec            = _mm256_and_ps(velec,cutoff_mask);
 541             velecsum         = _mm256_add_ps(velecsum,velec);
 542
 543             fscal            = felec;
 544
 545             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 546
 547             /* Calculate temporary vectorial force */
 548             tx               = _mm256_mul_ps(fscal,dx10);
 549             ty               = _mm256_mul_ps(fscal,dy10);
 550             tz               = _mm256_mul_ps(fscal,dz10);
 551
 552             /* Update vectorial force */
 553             fix1             = _mm256_add_ps(fix1,tx);
 554             fiy1             = _mm256_add_ps(fiy1,ty);
 555             fiz1             = _mm256_add_ps(fiz1,tz);
 556
 557             fjx0             = _mm256_add_ps(fjx0,tx);
 558             fjy0             = _mm256_add_ps(fjy0,ty);
 559             fjz0             = _mm256_add_ps(fjz0,tz);
 560
 561             }
 562
 563             /**************************
 564              * CALCULATE INTERACTIONS *
 565              **************************/
 566
 567             if (gmx_mm256_any_lt(rsq11,rcutoff2))
 568             {
 569
 570             r11              = _mm256_mul_ps(rsq11,rinv11);
 571
 572             /* EWALD ELECTROSTATICS */
 573
 574             /* Analytical PME correction */
 575             zeta2            = _mm256_mul_ps(beta2,rsq11);
 576             rinv3            = _mm256_mul_ps(rinvsq11,rinv11);
 577             pmecorrF         = avx256_pmecorrF_f(zeta2);
 578             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 579             felec            = _mm256_mul_ps(qq11,felec);
 580             pmecorrV         = avx256_pmecorrV_f(zeta2);
 581             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 582             velec            = _mm256_sub_ps(rinv11,pmecorrV);
 583             velec            = _mm256_mul_ps(qq11,velec);
 584
 585             d                = _mm256_sub_ps(r11,rswitch);
 586             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 587             d2               = _mm256_mul_ps(d,d);
 588             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)))))));
 589
 590             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 591
 592             /* Evaluate switch function */
 593             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 594             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv11,_mm256_mul_ps(velec,dsw)) );
 595             velec            = _mm256_mul_ps(velec,sw);
 596             cutoff_mask      = _mm256_cmp_ps(rsq11,rcutoff2,_CMP_LT_OQ);
 597
 598             /* Update potential sum for this i atom from the interaction with this j atom. */
 599             velec            = _mm256_and_ps(velec,cutoff_mask);
 600             velecsum         = _mm256_add_ps(velecsum,velec);
 601
 602             fscal            = felec;
 603
 604             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 605
 606             /* Calculate temporary vectorial force */
 607             tx               = _mm256_mul_ps(fscal,dx11);
 608             ty               = _mm256_mul_ps(fscal,dy11);
 609             tz               = _mm256_mul_ps(fscal,dz11);
 610
 611             /* Update vectorial force */
 612             fix1             = _mm256_add_ps(fix1,tx);
 613             fiy1             = _mm256_add_ps(fiy1,ty);
 614             fiz1             = _mm256_add_ps(fiz1,tz);
 615
 616             fjx1             = _mm256_add_ps(fjx1,tx);
 617             fjy1             = _mm256_add_ps(fjy1,ty);
 618             fjz1             = _mm256_add_ps(fjz1,tz);
 619
 620             }
 621
 622             /**************************
 623              * CALCULATE INTERACTIONS *
 624              **************************/
 625
 626             if (gmx_mm256_any_lt(rsq12,rcutoff2))
 627             {
 628
 629             r12              = _mm256_mul_ps(rsq12,rinv12);
 630
 631             /* EWALD ELECTROSTATICS */
 632
 633             /* Analytical PME correction */
 634             zeta2            = _mm256_mul_ps(beta2,rsq12);
 635             rinv3            = _mm256_mul_ps(rinvsq12,rinv12);
 636             pmecorrF         = avx256_pmecorrF_f(zeta2);
 637             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 638             felec            = _mm256_mul_ps(qq12,felec);
 639             pmecorrV         = avx256_pmecorrV_f(zeta2);
 640             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 641             velec            = _mm256_sub_ps(rinv12,pmecorrV);
 642             velec            = _mm256_mul_ps(qq12,velec);
 643
 644             d                = _mm256_sub_ps(r12,rswitch);
 645             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 646             d2               = _mm256_mul_ps(d,d);
 647             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)))))));
 648
 649             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 650
 651             /* Evaluate switch function */
 652             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 653             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv12,_mm256_mul_ps(velec,dsw)) );
 654             velec            = _mm256_mul_ps(velec,sw);
 655             cutoff_mask      = _mm256_cmp_ps(rsq12,rcutoff2,_CMP_LT_OQ);
 656
 657             /* Update potential sum for this i atom from the interaction with this j atom. */
 658             velec            = _mm256_and_ps(velec,cutoff_mask);
 659             velecsum         = _mm256_add_ps(velecsum,velec);
 660
 661             fscal            = felec;
 662
 663             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 664
 665             /* Calculate temporary vectorial force */
 666             tx               = _mm256_mul_ps(fscal,dx12);
 667             ty               = _mm256_mul_ps(fscal,dy12);
 668             tz               = _mm256_mul_ps(fscal,dz12);
 669
 670             /* Update vectorial force */
 671             fix1             = _mm256_add_ps(fix1,tx);
 672             fiy1             = _mm256_add_ps(fiy1,ty);
 673             fiz1             = _mm256_add_ps(fiz1,tz);
 674
 675             fjx2             = _mm256_add_ps(fjx2,tx);
 676             fjy2             = _mm256_add_ps(fjy2,ty);
 677             fjz2             = _mm256_add_ps(fjz2,tz);
 678
 679             }
 680
 681             /**************************
 682              * CALCULATE INTERACTIONS *
 683              **************************/
 684
 685             if (gmx_mm256_any_lt(rsq20,rcutoff2))
 686             {
 687
 688             r20              = _mm256_mul_ps(rsq20,rinv20);
 689
 690             /* EWALD ELECTROSTATICS */
 691
 692             /* Analytical PME correction */
 693             zeta2            = _mm256_mul_ps(beta2,rsq20);
 694             rinv3            = _mm256_mul_ps(rinvsq20,rinv20);
 695             pmecorrF         = avx256_pmecorrF_f(zeta2);
 696             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 697             felec            = _mm256_mul_ps(qq20,felec);
 698             pmecorrV         = avx256_pmecorrV_f(zeta2);
 699             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 700             velec            = _mm256_sub_ps(rinv20,pmecorrV);
 701             velec            = _mm256_mul_ps(qq20,velec);
 702
 703             d                = _mm256_sub_ps(r20,rswitch);
 704             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 705             d2               = _mm256_mul_ps(d,d);
 706             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)))))));
 707
 708             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 709
 710             /* Evaluate switch function */
 711             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 712             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv20,_mm256_mul_ps(velec,dsw)) );
 713             velec            = _mm256_mul_ps(velec,sw);
 714             cutoff_mask      = _mm256_cmp_ps(rsq20,rcutoff2,_CMP_LT_OQ);
 715
 716             /* Update potential sum for this i atom from the interaction with this j atom. */
 717             velec            = _mm256_and_ps(velec,cutoff_mask);
 718             velecsum         = _mm256_add_ps(velecsum,velec);
 719
 720             fscal            = felec;
 721
 722             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 723
 724             /* Calculate temporary vectorial force */
 725             tx               = _mm256_mul_ps(fscal,dx20);
 726             ty               = _mm256_mul_ps(fscal,dy20);
 727             tz               = _mm256_mul_ps(fscal,dz20);
 728
 729             /* Update vectorial force */
 730             fix2             = _mm256_add_ps(fix2,tx);
 731             fiy2             = _mm256_add_ps(fiy2,ty);
 732             fiz2             = _mm256_add_ps(fiz2,tz);
 733
 734             fjx0             = _mm256_add_ps(fjx0,tx);
 735             fjy0             = _mm256_add_ps(fjy0,ty);
 736             fjz0             = _mm256_add_ps(fjz0,tz);
 737
 738             }
 739
 740             /**************************
 741              * CALCULATE INTERACTIONS *
 742              **************************/
 743
 744             if (gmx_mm256_any_lt(rsq21,rcutoff2))
 745             {
 746
 747             r21              = _mm256_mul_ps(rsq21,rinv21);
 748
 749             /* EWALD ELECTROSTATICS */
 750
 751             /* Analytical PME correction */
 752             zeta2            = _mm256_mul_ps(beta2,rsq21);
 753             rinv3            = _mm256_mul_ps(rinvsq21,rinv21);
 754             pmecorrF         = avx256_pmecorrF_f(zeta2);
 755             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 756             felec            = _mm256_mul_ps(qq21,felec);
 757             pmecorrV         = avx256_pmecorrV_f(zeta2);
 758             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 759             velec            = _mm256_sub_ps(rinv21,pmecorrV);
 760             velec            = _mm256_mul_ps(qq21,velec);
 761
 762             d                = _mm256_sub_ps(r21,rswitch);
 763             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 764             d2               = _mm256_mul_ps(d,d);
 765             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)))))));
 766
 767             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 768
 769             /* Evaluate switch function */
 770             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 771             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv21,_mm256_mul_ps(velec,dsw)) );
 772             velec            = _mm256_mul_ps(velec,sw);
 773             cutoff_mask      = _mm256_cmp_ps(rsq21,rcutoff2,_CMP_LT_OQ);
 774
 775             /* Update potential sum for this i atom from the interaction with this j atom. */
 776             velec            = _mm256_and_ps(velec,cutoff_mask);
 777             velecsum         = _mm256_add_ps(velecsum,velec);
 778
 779             fscal            = felec;
 780
 781             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 782
 783             /* Calculate temporary vectorial force */
 784             tx               = _mm256_mul_ps(fscal,dx21);
 785             ty               = _mm256_mul_ps(fscal,dy21);
 786             tz               = _mm256_mul_ps(fscal,dz21);
 787
 788             /* Update vectorial force */
 789             fix2             = _mm256_add_ps(fix2,tx);
 790             fiy2             = _mm256_add_ps(fiy2,ty);
 791             fiz2             = _mm256_add_ps(fiz2,tz);
 792
 793             fjx1             = _mm256_add_ps(fjx1,tx);
 794             fjy1             = _mm256_add_ps(fjy1,ty);
 795             fjz1             = _mm256_add_ps(fjz1,tz);
 796
 797             }
 798
 799             /**************************
 800              * CALCULATE INTERACTIONS *
 801              **************************/
 802
 803             if (gmx_mm256_any_lt(rsq22,rcutoff2))
 804             {
 805
 806             r22              = _mm256_mul_ps(rsq22,rinv22);
 807
 808             /* EWALD ELECTROSTATICS */
 809
 810             /* Analytical PME correction */
 811             zeta2            = _mm256_mul_ps(beta2,rsq22);
 812             rinv3            = _mm256_mul_ps(rinvsq22,rinv22);
 813             pmecorrF         = avx256_pmecorrF_f(zeta2);
 814             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 815             felec            = _mm256_mul_ps(qq22,felec);
 816             pmecorrV         = avx256_pmecorrV_f(zeta2);
 817             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 818             velec            = _mm256_sub_ps(rinv22,pmecorrV);
 819             velec            = _mm256_mul_ps(qq22,velec);
 820
 821             d                = _mm256_sub_ps(r22,rswitch);
 822             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 823             d2               = _mm256_mul_ps(d,d);
 824             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)))))));
 825
 826             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 827
 828             /* Evaluate switch function */
 829             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 830             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv22,_mm256_mul_ps(velec,dsw)) );
 831             velec            = _mm256_mul_ps(velec,sw);
 832             cutoff_mask      = _mm256_cmp_ps(rsq22,rcutoff2,_CMP_LT_OQ);
 833
 834             /* Update potential sum for this i atom from the interaction with this j atom. */
 835             velec            = _mm256_and_ps(velec,cutoff_mask);
 836             velecsum         = _mm256_add_ps(velecsum,velec);
 837
 838             fscal            = felec;
 839
 840             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 841
 842             /* Calculate temporary vectorial force */
 843             tx               = _mm256_mul_ps(fscal,dx22);
 844             ty               = _mm256_mul_ps(fscal,dy22);
 845             tz               = _mm256_mul_ps(fscal,dz22);
 846
 847             /* Update vectorial force */
 848             fix2             = _mm256_add_ps(fix2,tx);
 849             fiy2             = _mm256_add_ps(fiy2,ty);
 850             fiz2             = _mm256_add_ps(fiz2,tz);
 851
 852             fjx2             = _mm256_add_ps(fjx2,tx);
 853             fjy2             = _mm256_add_ps(fjy2,ty);
 854             fjz2             = _mm256_add_ps(fjz2,tz);
 855
 856             }
 857
 858             fjptrA             = f+j_coord_offsetA;
 859             fjptrB             = f+j_coord_offsetB;
 860             fjptrC             = f+j_coord_offsetC;
 861             fjptrD             = f+j_coord_offsetD;
 862             fjptrE             = f+j_coord_offsetE;
 863             fjptrF             = f+j_coord_offsetF;
 864             fjptrG             = f+j_coord_offsetG;
 865             fjptrH             = f+j_coord_offsetH;
 866
 867             gmx_mm256_decrement_3rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,
 868                                                       fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2);
 869
 870             /* Inner loop uses 972 flops */
 871         }
 872
 873         if(jidx<j_index_end)
 874         {
 875
 876             /* Get j neighbor index, and coordinate index */
 877             jnrlistA         = jjnr[jidx];
 878             jnrlistB         = jjnr[jidx+1];
 879             jnrlistC         = jjnr[jidx+2];
 880             jnrlistD         = jjnr[jidx+3];
 881             jnrlistE         = jjnr[jidx+4];
 882             jnrlistF         = jjnr[jidx+5];
 883             jnrlistG         = jjnr[jidx+6];
 884             jnrlistH         = jjnr[jidx+7];
 885             /* Sign of each element will be negative for non-real atoms.
 886              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
 887              * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
 888              */
 889             dummy_mask = gmx_mm256_set_m128(gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx+4)),_mm_setzero_si128())),
 890                                             gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128())));
 891
 892             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
 893             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
 894             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
 895             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
 896             jnrE       = (jnrlistE>=0) ? jnrlistE : 0;
 897             jnrF       = (jnrlistF>=0) ? jnrlistF : 0;
 898             jnrG       = (jnrlistG>=0) ? jnrlistG : 0;
 899             jnrH       = (jnrlistH>=0) ? jnrlistH : 0;
 900             j_coord_offsetA  = DIM*jnrA;
 901             j_coord_offsetB  = DIM*jnrB;
 902             j_coord_offsetC  = DIM*jnrC;
 903             j_coord_offsetD  = DIM*jnrD;
 904             j_coord_offsetE  = DIM*jnrE;
 905             j_coord_offsetF  = DIM*jnrF;
 906             j_coord_offsetG  = DIM*jnrG;
 907             j_coord_offsetH  = DIM*jnrH;
 908
 909             /* load j atom coordinates */
 910             gmx_mm256_load_3rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 911                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 912                                                  x+j_coord_offsetE,x+j_coord_offsetF,
 913                                                  x+j_coord_offsetG,x+j_coord_offsetH,
 914                                               &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,&jy2,&jz2);
 915
 916             /* Calculate displacement vector */
 917             dx00             = _mm256_sub_ps(ix0,jx0);
 918             dy00             = _mm256_sub_ps(iy0,jy0);
 919             dz00             = _mm256_sub_ps(iz0,jz0);
 920             dx01             = _mm256_sub_ps(ix0,jx1);
 921             dy01             = _mm256_sub_ps(iy0,jy1);
 922             dz01             = _mm256_sub_ps(iz0,jz1);
 923             dx02             = _mm256_sub_ps(ix0,jx2);
 924             dy02             = _mm256_sub_ps(iy0,jy2);
 925             dz02             = _mm256_sub_ps(iz0,jz2);
 926             dx10             = _mm256_sub_ps(ix1,jx0);
 927             dy10             = _mm256_sub_ps(iy1,jy0);
 928             dz10             = _mm256_sub_ps(iz1,jz0);
 929             dx11             = _mm256_sub_ps(ix1,jx1);
 930             dy11             = _mm256_sub_ps(iy1,jy1);
 931             dz11             = _mm256_sub_ps(iz1,jz1);
 932             dx12             = _mm256_sub_ps(ix1,jx2);
 933             dy12             = _mm256_sub_ps(iy1,jy2);
 934             dz12             = _mm256_sub_ps(iz1,jz2);
 935             dx20             = _mm256_sub_ps(ix2,jx0);
 936             dy20             = _mm256_sub_ps(iy2,jy0);
 937             dz20             = _mm256_sub_ps(iz2,jz0);
 938             dx21             = _mm256_sub_ps(ix2,jx1);
 939             dy21             = _mm256_sub_ps(iy2,jy1);
 940             dz21             = _mm256_sub_ps(iz2,jz1);
 941             dx22             = _mm256_sub_ps(ix2,jx2);
 942             dy22             = _mm256_sub_ps(iy2,jy2);
 943             dz22             = _mm256_sub_ps(iz2,jz2);
 944
 945             /* Calculate squared distance and things based on it */
 946             rsq00            = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
 947             rsq01            = gmx_mm256_calc_rsq_ps(dx01,dy01,dz01);
 948             rsq02            = gmx_mm256_calc_rsq_ps(dx02,dy02,dz02);
 949             rsq10            = gmx_mm256_calc_rsq_ps(dx10,dy10,dz10);
 950             rsq11            = gmx_mm256_calc_rsq_ps(dx11,dy11,dz11);
 951             rsq12            = gmx_mm256_calc_rsq_ps(dx12,dy12,dz12);
 952             rsq20            = gmx_mm256_calc_rsq_ps(dx20,dy20,dz20);
 953             rsq21            = gmx_mm256_calc_rsq_ps(dx21,dy21,dz21);
 954             rsq22            = gmx_mm256_calc_rsq_ps(dx22,dy22,dz22);
 955
 956             rinv00           = avx256_invsqrt_f(rsq00);
 957             rinv01           = avx256_invsqrt_f(rsq01);
 958             rinv02           = avx256_invsqrt_f(rsq02);
 959             rinv10           = avx256_invsqrt_f(rsq10);
 960             rinv11           = avx256_invsqrt_f(rsq11);
 961             rinv12           = avx256_invsqrt_f(rsq12);
 962             rinv20           = avx256_invsqrt_f(rsq20);
 963             rinv21           = avx256_invsqrt_f(rsq21);
 964             rinv22           = avx256_invsqrt_f(rsq22);
 965
 966             rinvsq00         = _mm256_mul_ps(rinv00,rinv00);
 967             rinvsq01         = _mm256_mul_ps(rinv01,rinv01);
 968             rinvsq02         = _mm256_mul_ps(rinv02,rinv02);
 969             rinvsq10         = _mm256_mul_ps(rinv10,rinv10);
 970             rinvsq11         = _mm256_mul_ps(rinv11,rinv11);
 971             rinvsq12         = _mm256_mul_ps(rinv12,rinv12);
 972             rinvsq20         = _mm256_mul_ps(rinv20,rinv20);
 973             rinvsq21         = _mm256_mul_ps(rinv21,rinv21);
 974             rinvsq22         = _mm256_mul_ps(rinv22,rinv22);
 975
 976             fjx0             = _mm256_setzero_ps();
 977             fjy0             = _mm256_setzero_ps();
 978             fjz0             = _mm256_setzero_ps();
 979             fjx1             = _mm256_setzero_ps();
 980             fjy1             = _mm256_setzero_ps();
 981             fjz1             = _mm256_setzero_ps();
 982             fjx2             = _mm256_setzero_ps();
 983             fjy2             = _mm256_setzero_ps();
 984             fjz2             = _mm256_setzero_ps();
 985
 986             /**************************
 987              * CALCULATE INTERACTIONS *
 988              **************************/
 989
 990             if (gmx_mm256_any_lt(rsq00,rcutoff2))
 991             {
 992
 993             r00              = _mm256_mul_ps(rsq00,rinv00);
 994             r00              = _mm256_andnot_ps(dummy_mask,r00);
 995
 996             /* EWALD ELECTROSTATICS */
 997
 998             /* Analytical PME correction */
 999             zeta2            = _mm256_mul_ps(beta2,rsq00);
1000             rinv3            = _mm256_mul_ps(rinvsq00,rinv00);
1001             pmecorrF         = avx256_pmecorrF_f(zeta2);
1002             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1003             felec            = _mm256_mul_ps(qq00,felec);
1004             pmecorrV         = avx256_pmecorrV_f(zeta2);
1005             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1006             velec            = _mm256_sub_ps(rinv00,pmecorrV);
1007             velec            = _mm256_mul_ps(qq00,velec);
1008
1009             d                = _mm256_sub_ps(r00,rswitch);
1010             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1011             d2               = _mm256_mul_ps(d,d);
1012             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)))))));
1013
1014             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1015
1016             /* Evaluate switch function */
1017             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1018             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv00,_mm256_mul_ps(velec,dsw)) );
1019             velec            = _mm256_mul_ps(velec,sw);
1020             cutoff_mask      = _mm256_cmp_ps(rsq00,rcutoff2,_CMP_LT_OQ);
1021
1022             /* Update potential sum for this i atom from the interaction with this j atom. */
1023             velec            = _mm256_and_ps(velec,cutoff_mask);
1024             velec            = _mm256_andnot_ps(dummy_mask,velec);
1025             velecsum         = _mm256_add_ps(velecsum,velec);
1026
1027             fscal            = felec;
1028
1029             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1030
1031             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1032
1033             /* Calculate temporary vectorial force */
1034             tx               = _mm256_mul_ps(fscal,dx00);
1035             ty               = _mm256_mul_ps(fscal,dy00);
1036             tz               = _mm256_mul_ps(fscal,dz00);
1037
1038             /* Update vectorial force */
1039             fix0             = _mm256_add_ps(fix0,tx);
1040             fiy0             = _mm256_add_ps(fiy0,ty);
1041             fiz0             = _mm256_add_ps(fiz0,tz);
1042
1043             fjx0             = _mm256_add_ps(fjx0,tx);
1044             fjy0             = _mm256_add_ps(fjy0,ty);
1045             fjz0             = _mm256_add_ps(fjz0,tz);
1046
1047             }
1048
1049             /**************************
1050              * CALCULATE INTERACTIONS *
1051              **************************/
1052
1053             if (gmx_mm256_any_lt(rsq01,rcutoff2))
1054             {
1055
1056             r01              = _mm256_mul_ps(rsq01,rinv01);
1057             r01              = _mm256_andnot_ps(dummy_mask,r01);
1058
1059             /* EWALD ELECTROSTATICS */
1060
1061             /* Analytical PME correction */
1062             zeta2            = _mm256_mul_ps(beta2,rsq01);
1063             rinv3            = _mm256_mul_ps(rinvsq01,rinv01);
1064             pmecorrF         = avx256_pmecorrF_f(zeta2);
1065             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1066             felec            = _mm256_mul_ps(qq01,felec);
1067             pmecorrV         = avx256_pmecorrV_f(zeta2);
1068             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1069             velec            = _mm256_sub_ps(rinv01,pmecorrV);
1070             velec            = _mm256_mul_ps(qq01,velec);
1071
1072             d                = _mm256_sub_ps(r01,rswitch);
1073             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1074             d2               = _mm256_mul_ps(d,d);
1075             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)))))));
1076
1077             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1078
1079             /* Evaluate switch function */
1080             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1081             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv01,_mm256_mul_ps(velec,dsw)) );
1082             velec            = _mm256_mul_ps(velec,sw);
1083             cutoff_mask      = _mm256_cmp_ps(rsq01,rcutoff2,_CMP_LT_OQ);
1084
1085             /* Update potential sum for this i atom from the interaction with this j atom. */
1086             velec            = _mm256_and_ps(velec,cutoff_mask);
1087             velec            = _mm256_andnot_ps(dummy_mask,velec);
1088             velecsum         = _mm256_add_ps(velecsum,velec);
1089
1090             fscal            = felec;
1091
1092             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1093
1094             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1095
1096             /* Calculate temporary vectorial force */
1097             tx               = _mm256_mul_ps(fscal,dx01);
1098             ty               = _mm256_mul_ps(fscal,dy01);
1099             tz               = _mm256_mul_ps(fscal,dz01);
1100
1101             /* Update vectorial force */
1102             fix0             = _mm256_add_ps(fix0,tx);
1103             fiy0             = _mm256_add_ps(fiy0,ty);
1104             fiz0             = _mm256_add_ps(fiz0,tz);
1105
1106             fjx1             = _mm256_add_ps(fjx1,tx);
1107             fjy1             = _mm256_add_ps(fjy1,ty);
1108             fjz1             = _mm256_add_ps(fjz1,tz);
1109
1110             }
1111
1112             /**************************
1113              * CALCULATE INTERACTIONS *
1114              **************************/
1115
1116             if (gmx_mm256_any_lt(rsq02,rcutoff2))
1117             {
1118
1119             r02              = _mm256_mul_ps(rsq02,rinv02);
1120             r02              = _mm256_andnot_ps(dummy_mask,r02);
1121
1122             /* EWALD ELECTROSTATICS */
1123
1124             /* Analytical PME correction */
1125             zeta2            = _mm256_mul_ps(beta2,rsq02);
1126             rinv3            = _mm256_mul_ps(rinvsq02,rinv02);
1127             pmecorrF         = avx256_pmecorrF_f(zeta2);
1128             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1129             felec            = _mm256_mul_ps(qq02,felec);
1130             pmecorrV         = avx256_pmecorrV_f(zeta2);
1131             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1132             velec            = _mm256_sub_ps(rinv02,pmecorrV);
1133             velec            = _mm256_mul_ps(qq02,velec);
1134
1135             d                = _mm256_sub_ps(r02,rswitch);
1136             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1137             d2               = _mm256_mul_ps(d,d);
1138             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)))))));
1139
1140             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1141
1142             /* Evaluate switch function */
1143             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1144             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv02,_mm256_mul_ps(velec,dsw)) );
1145             velec            = _mm256_mul_ps(velec,sw);
1146             cutoff_mask      = _mm256_cmp_ps(rsq02,rcutoff2,_CMP_LT_OQ);
1147
1148             /* Update potential sum for this i atom from the interaction with this j atom. */
1149             velec            = _mm256_and_ps(velec,cutoff_mask);
1150             velec            = _mm256_andnot_ps(dummy_mask,velec);
1151             velecsum         = _mm256_add_ps(velecsum,velec);
1152
1153             fscal            = felec;
1154
1155             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1156
1157             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1158
1159             /* Calculate temporary vectorial force */
1160             tx               = _mm256_mul_ps(fscal,dx02);
1161             ty               = _mm256_mul_ps(fscal,dy02);
1162             tz               = _mm256_mul_ps(fscal,dz02);
1163
1164             /* Update vectorial force */
1165             fix0             = _mm256_add_ps(fix0,tx);
1166             fiy0             = _mm256_add_ps(fiy0,ty);
1167             fiz0             = _mm256_add_ps(fiz0,tz);
1168
1169             fjx2             = _mm256_add_ps(fjx2,tx);
1170             fjy2             = _mm256_add_ps(fjy2,ty);
1171             fjz2             = _mm256_add_ps(fjz2,tz);
1172
1173             }
1174
1175             /**************************
1176              * CALCULATE INTERACTIONS *
1177              **************************/
1178
1179             if (gmx_mm256_any_lt(rsq10,rcutoff2))
1180             {
1181
1182             r10              = _mm256_mul_ps(rsq10,rinv10);
1183             r10              = _mm256_andnot_ps(dummy_mask,r10);
1184
1185             /* EWALD ELECTROSTATICS */
1186
1187             /* Analytical PME correction */
1188             zeta2            = _mm256_mul_ps(beta2,rsq10);
1189             rinv3            = _mm256_mul_ps(rinvsq10,rinv10);
1190             pmecorrF         = avx256_pmecorrF_f(zeta2);
1191             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1192             felec            = _mm256_mul_ps(qq10,felec);
1193             pmecorrV         = avx256_pmecorrV_f(zeta2);
1194             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1195             velec            = _mm256_sub_ps(rinv10,pmecorrV);
1196             velec            = _mm256_mul_ps(qq10,velec);
1197
1198             d                = _mm256_sub_ps(r10,rswitch);
1199             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1200             d2               = _mm256_mul_ps(d,d);
1201             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)))))));
1202
1203             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1204
1205             /* Evaluate switch function */
1206             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1207             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv10,_mm256_mul_ps(velec,dsw)) );
1208             velec            = _mm256_mul_ps(velec,sw);
1209             cutoff_mask      = _mm256_cmp_ps(rsq10,rcutoff2,_CMP_LT_OQ);
1210
1211             /* Update potential sum for this i atom from the interaction with this j atom. */
1212             velec            = _mm256_and_ps(velec,cutoff_mask);
1213             velec            = _mm256_andnot_ps(dummy_mask,velec);
1214             velecsum         = _mm256_add_ps(velecsum,velec);
1215
1216             fscal            = felec;
1217
1218             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1219
1220             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1221
1222             /* Calculate temporary vectorial force */
1223             tx               = _mm256_mul_ps(fscal,dx10);
1224             ty               = _mm256_mul_ps(fscal,dy10);
1225             tz               = _mm256_mul_ps(fscal,dz10);
1226
1227             /* Update vectorial force */
1228             fix1             = _mm256_add_ps(fix1,tx);
1229             fiy1             = _mm256_add_ps(fiy1,ty);
1230             fiz1             = _mm256_add_ps(fiz1,tz);
1231
1232             fjx0             = _mm256_add_ps(fjx0,tx);
1233             fjy0             = _mm256_add_ps(fjy0,ty);
1234             fjz0             = _mm256_add_ps(fjz0,tz);
1235
1236             }
1237
1238             /**************************
1239              * CALCULATE INTERACTIONS *
1240              **************************/
1241
1242             if (gmx_mm256_any_lt(rsq11,rcutoff2))
1243             {
1244
1245             r11              = _mm256_mul_ps(rsq11,rinv11);
1246             r11              = _mm256_andnot_ps(dummy_mask,r11);
1247
1248             /* EWALD ELECTROSTATICS */
1249
1250             /* Analytical PME correction */
1251             zeta2            = _mm256_mul_ps(beta2,rsq11);
1252             rinv3            = _mm256_mul_ps(rinvsq11,rinv11);
1253             pmecorrF         = avx256_pmecorrF_f(zeta2);
1254             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1255             felec            = _mm256_mul_ps(qq11,felec);
1256             pmecorrV         = avx256_pmecorrV_f(zeta2);
1257             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1258             velec            = _mm256_sub_ps(rinv11,pmecorrV);
1259             velec            = _mm256_mul_ps(qq11,velec);
1260
1261             d                = _mm256_sub_ps(r11,rswitch);
1262             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1263             d2               = _mm256_mul_ps(d,d);
1264             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)))))));
1265
1266             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1267
1268             /* Evaluate switch function */
1269             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1270             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv11,_mm256_mul_ps(velec,dsw)) );
1271             velec            = _mm256_mul_ps(velec,sw);
1272             cutoff_mask      = _mm256_cmp_ps(rsq11,rcutoff2,_CMP_LT_OQ);
1273
1274             /* Update potential sum for this i atom from the interaction with this j atom. */
1275             velec            = _mm256_and_ps(velec,cutoff_mask);
1276             velec            = _mm256_andnot_ps(dummy_mask,velec);
1277             velecsum         = _mm256_add_ps(velecsum,velec);
1278
1279             fscal            = felec;
1280
1281             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1282
1283             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1284
1285             /* Calculate temporary vectorial force */
1286             tx               = _mm256_mul_ps(fscal,dx11);
1287             ty               = _mm256_mul_ps(fscal,dy11);
1288             tz               = _mm256_mul_ps(fscal,dz11);
1289
1290             /* Update vectorial force */
1291             fix1             = _mm256_add_ps(fix1,tx);
1292             fiy1             = _mm256_add_ps(fiy1,ty);
1293             fiz1             = _mm256_add_ps(fiz1,tz);
1294
1295             fjx1             = _mm256_add_ps(fjx1,tx);
1296             fjy1             = _mm256_add_ps(fjy1,ty);
1297             fjz1             = _mm256_add_ps(fjz1,tz);
1298
1299             }
1300
1301             /**************************
1302              * CALCULATE INTERACTIONS *
1303              **************************/
1304
1305             if (gmx_mm256_any_lt(rsq12,rcutoff2))
1306             {
1307
1308             r12              = _mm256_mul_ps(rsq12,rinv12);
1309             r12              = _mm256_andnot_ps(dummy_mask,r12);
1310
1311             /* EWALD ELECTROSTATICS */
1312
1313             /* Analytical PME correction */
1314             zeta2            = _mm256_mul_ps(beta2,rsq12);
1315             rinv3            = _mm256_mul_ps(rinvsq12,rinv12);
1316             pmecorrF         = avx256_pmecorrF_f(zeta2);
1317             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1318             felec            = _mm256_mul_ps(qq12,felec);
1319             pmecorrV         = avx256_pmecorrV_f(zeta2);
1320             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1321             velec            = _mm256_sub_ps(rinv12,pmecorrV);
1322             velec            = _mm256_mul_ps(qq12,velec);
1323
1324             d                = _mm256_sub_ps(r12,rswitch);
1325             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1326             d2               = _mm256_mul_ps(d,d);
1327             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)))))));
1328
1329             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1330
1331             /* Evaluate switch function */
1332             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1333             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv12,_mm256_mul_ps(velec,dsw)) );
1334             velec            = _mm256_mul_ps(velec,sw);
1335             cutoff_mask      = _mm256_cmp_ps(rsq12,rcutoff2,_CMP_LT_OQ);
1336
1337             /* Update potential sum for this i atom from the interaction with this j atom. */
1338             velec            = _mm256_and_ps(velec,cutoff_mask);
1339             velec            = _mm256_andnot_ps(dummy_mask,velec);
1340             velecsum         = _mm256_add_ps(velecsum,velec);
1341
1342             fscal            = felec;
1343
1344             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1345
1346             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1347
1348             /* Calculate temporary vectorial force */
1349             tx               = _mm256_mul_ps(fscal,dx12);
1350             ty               = _mm256_mul_ps(fscal,dy12);
1351             tz               = _mm256_mul_ps(fscal,dz12);
1352
1353             /* Update vectorial force */
1354             fix1             = _mm256_add_ps(fix1,tx);
1355             fiy1             = _mm256_add_ps(fiy1,ty);
1356             fiz1             = _mm256_add_ps(fiz1,tz);
1357
1358             fjx2             = _mm256_add_ps(fjx2,tx);
1359             fjy2             = _mm256_add_ps(fjy2,ty);
1360             fjz2             = _mm256_add_ps(fjz2,tz);
1361
1362             }
1363
1364             /**************************
1365              * CALCULATE INTERACTIONS *
1366              **************************/
1367
1368             if (gmx_mm256_any_lt(rsq20,rcutoff2))
1369             {
1370
1371             r20              = _mm256_mul_ps(rsq20,rinv20);
1372             r20              = _mm256_andnot_ps(dummy_mask,r20);
1373
1374             /* EWALD ELECTROSTATICS */
1375
1376             /* Analytical PME correction */
1377             zeta2            = _mm256_mul_ps(beta2,rsq20);
1378             rinv3            = _mm256_mul_ps(rinvsq20,rinv20);
1379             pmecorrF         = avx256_pmecorrF_f(zeta2);
1380             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1381             felec            = _mm256_mul_ps(qq20,felec);
1382             pmecorrV         = avx256_pmecorrV_f(zeta2);
1383             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1384             velec            = _mm256_sub_ps(rinv20,pmecorrV);
1385             velec            = _mm256_mul_ps(qq20,velec);
1386
1387             d                = _mm256_sub_ps(r20,rswitch);
1388             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1389             d2               = _mm256_mul_ps(d,d);
1390             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)))))));
1391
1392             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1393
1394             /* Evaluate switch function */
1395             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1396             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv20,_mm256_mul_ps(velec,dsw)) );
1397             velec            = _mm256_mul_ps(velec,sw);
1398             cutoff_mask      = _mm256_cmp_ps(rsq20,rcutoff2,_CMP_LT_OQ);
1399
1400             /* Update potential sum for this i atom from the interaction with this j atom. */
1401             velec            = _mm256_and_ps(velec,cutoff_mask);
1402             velec            = _mm256_andnot_ps(dummy_mask,velec);
1403             velecsum         = _mm256_add_ps(velecsum,velec);
1404
1405             fscal            = felec;
1406
1407             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1408
1409             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1410
1411             /* Calculate temporary vectorial force */
1412             tx               = _mm256_mul_ps(fscal,dx20);
1413             ty               = _mm256_mul_ps(fscal,dy20);
1414             tz               = _mm256_mul_ps(fscal,dz20);
1415
1416             /* Update vectorial force */
1417             fix2             = _mm256_add_ps(fix2,tx);
1418             fiy2             = _mm256_add_ps(fiy2,ty);
1419             fiz2             = _mm256_add_ps(fiz2,tz);
1420
1421             fjx0             = _mm256_add_ps(fjx0,tx);
1422             fjy0             = _mm256_add_ps(fjy0,ty);
1423             fjz0             = _mm256_add_ps(fjz0,tz);
1424
1425             }
1426
1427             /**************************
1428              * CALCULATE INTERACTIONS *
1429              **************************/
1430
1431             if (gmx_mm256_any_lt(rsq21,rcutoff2))
1432             {
1433
1434             r21              = _mm256_mul_ps(rsq21,rinv21);
1435             r21              = _mm256_andnot_ps(dummy_mask,r21);
1436
1437             /* EWALD ELECTROSTATICS */
1438
1439             /* Analytical PME correction */
1440             zeta2            = _mm256_mul_ps(beta2,rsq21);
1441             rinv3            = _mm256_mul_ps(rinvsq21,rinv21);
1442             pmecorrF         = avx256_pmecorrF_f(zeta2);
1443             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1444             felec            = _mm256_mul_ps(qq21,felec);
1445             pmecorrV         = avx256_pmecorrV_f(zeta2);
1446             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1447             velec            = _mm256_sub_ps(rinv21,pmecorrV);
1448             velec            = _mm256_mul_ps(qq21,velec);
1449
1450             d                = _mm256_sub_ps(r21,rswitch);
1451             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1452             d2               = _mm256_mul_ps(d,d);
1453             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)))))));
1454
1455             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1456
1457             /* Evaluate switch function */
1458             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1459             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv21,_mm256_mul_ps(velec,dsw)) );
1460             velec            = _mm256_mul_ps(velec,sw);
1461             cutoff_mask      = _mm256_cmp_ps(rsq21,rcutoff2,_CMP_LT_OQ);
1462
1463             /* Update potential sum for this i atom from the interaction with this j atom. */
1464             velec            = _mm256_and_ps(velec,cutoff_mask);
1465             velec            = _mm256_andnot_ps(dummy_mask,velec);
1466             velecsum         = _mm256_add_ps(velecsum,velec);
1467
1468             fscal            = felec;
1469
1470             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1471
1472             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1473
1474             /* Calculate temporary vectorial force */
1475             tx               = _mm256_mul_ps(fscal,dx21);
1476             ty               = _mm256_mul_ps(fscal,dy21);
1477             tz               = _mm256_mul_ps(fscal,dz21);
1478
1479             /* Update vectorial force */
1480             fix2             = _mm256_add_ps(fix2,tx);
1481             fiy2             = _mm256_add_ps(fiy2,ty);
1482             fiz2             = _mm256_add_ps(fiz2,tz);
1483
1484             fjx1             = _mm256_add_ps(fjx1,tx);
1485             fjy1             = _mm256_add_ps(fjy1,ty);
1486             fjz1             = _mm256_add_ps(fjz1,tz);
1487
1488             }
1489
1490             /**************************
1491              * CALCULATE INTERACTIONS *
1492              **************************/
1493
1494             if (gmx_mm256_any_lt(rsq22,rcutoff2))
1495             {
1496
1497             r22              = _mm256_mul_ps(rsq22,rinv22);
1498             r22              = _mm256_andnot_ps(dummy_mask,r22);
1499
1500             /* EWALD ELECTROSTATICS */
1501
1502             /* Analytical PME correction */
1503             zeta2            = _mm256_mul_ps(beta2,rsq22);
1504             rinv3            = _mm256_mul_ps(rinvsq22,rinv22);
1505             pmecorrF         = avx256_pmecorrF_f(zeta2);
1506             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1507             felec            = _mm256_mul_ps(qq22,felec);
1508             pmecorrV         = avx256_pmecorrV_f(zeta2);
1509             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1510             velec            = _mm256_sub_ps(rinv22,pmecorrV);
1511             velec            = _mm256_mul_ps(qq22,velec);
1512
1513             d                = _mm256_sub_ps(r22,rswitch);
1514             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1515             d2               = _mm256_mul_ps(d,d);
1516             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)))))));
1517
1518             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1519
1520             /* Evaluate switch function */
1521             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1522             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv22,_mm256_mul_ps(velec,dsw)) );
1523             velec            = _mm256_mul_ps(velec,sw);
1524             cutoff_mask      = _mm256_cmp_ps(rsq22,rcutoff2,_CMP_LT_OQ);
1525
1526             /* Update potential sum for this i atom from the interaction with this j atom. */
1527             velec            = _mm256_and_ps(velec,cutoff_mask);
1528             velec            = _mm256_andnot_ps(dummy_mask,velec);
1529             velecsum         = _mm256_add_ps(velecsum,velec);
1530
1531             fscal            = felec;
1532
1533             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1534
1535             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1536
1537             /* Calculate temporary vectorial force */
1538             tx               = _mm256_mul_ps(fscal,dx22);
1539             ty               = _mm256_mul_ps(fscal,dy22);
1540             tz               = _mm256_mul_ps(fscal,dz22);
1541
1542             /* Update vectorial force */
1543             fix2             = _mm256_add_ps(fix2,tx);
1544             fiy2             = _mm256_add_ps(fiy2,ty);
1545             fiz2             = _mm256_add_ps(fiz2,tz);
1546
1547             fjx2             = _mm256_add_ps(fjx2,tx);
1548             fjy2             = _mm256_add_ps(fjy2,ty);
1549             fjz2             = _mm256_add_ps(fjz2,tz);
1550
1551             }
1552
1553             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
1554             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
1555             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
1556             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
1557             fjptrE             = (jnrlistE>=0) ? f+j_coord_offsetE : scratch;
1558             fjptrF             = (jnrlistF>=0) ? f+j_coord_offsetF : scratch;
1559             fjptrG             = (jnrlistG>=0) ? f+j_coord_offsetG : scratch;
1560             fjptrH             = (jnrlistH>=0) ? f+j_coord_offsetH : scratch;
1561
1562             gmx_mm256_decrement_3rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,
1563                                                       fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2);
1564
1565             /* Inner loop uses 981 flops */
1566         }
1567
1568         /* End of innermost loop */
1569
1570         gmx_mm256_update_iforce_3atom_swizzle_ps(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,
1571                                                  f+i_coord_offset,fshift+i_shift_offset);
1572
1573         ggid                        = gid[iidx];
1574         /* Update potential energies */
1575         gmx_mm256_update_1pot_ps(velecsum,kernel_data->energygrp_elec+ggid);
1576
1577         /* Increment number of inner iterations */
1578         inneriter                  += j_index_end - j_index_start;
1579
1580         /* Outer loop uses 19 flops */
1581     }
1582
1583     /* Increment number of outer iterations */
1584     outeriter        += nri;
1585
1586     /* Update outer/inner flops */
1587
1588     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_W3W3_VF,outeriter*19 + inneriter*981);
1589 }
1590 /*
1591  * Gromacs nonbonded kernel:   nb_kernel_ElecEwSw_VdwNone_GeomW3W3_F_avx_256_single
1592  * Electrostatics interaction: Ewald
1593  * VdW interaction:            None
1594  * Geometry:                   Water3-Water3
1595  * Calculate force/pot:        Force
1596  */
1597 void
1598 nb_kernel_ElecEwSw_VdwNone_GeomW3W3_F_avx_256_single
1599                     (t_nblist                    * gmx_restrict       nlist,
1600                      rvec                        * gmx_restrict          xx,
1601                      rvec                        * gmx_restrict          ff,
1602                      struct t_forcerec           * gmx_restrict          fr,
1603                      t_mdatoms                   * gmx_restrict     mdatoms,
1604                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
1605                      t_nrnb                      * gmx_restrict        nrnb)
1606 {
1607     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
1608      * just 0 for non-waters.
1609      * Suffixes A,B,C,D,E,F,G,H refer to j loop unrolling done with AVX, e.g. for the eight different
1610      * jnr indices corresponding to data put in the four positions in the SIMD register.
1611      */
1612     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
1613     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
1614     int              jnrA,jnrB,jnrC,jnrD;
1615     int              jnrE,jnrF,jnrG,jnrH;
1616     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
1617     int              jnrlistE,jnrlistF,jnrlistG,jnrlistH;
1618     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
1619     int              j_coord_offsetE,j_coord_offsetF,j_coord_offsetG,j_coord_offsetH;
1620     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
1621     real             rcutoff_scalar;
1622     real             *shiftvec,*fshift,*x,*f;
1623     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD,*fjptrE,*fjptrF,*fjptrG,*fjptrH;
1624     real             scratch[4*DIM];
1625     __m256           tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
1626     real *           vdwioffsetptr0;
1627     __m256           ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
1628     real *           vdwioffsetptr1;
1629     __m256           ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
1630     real *           vdwioffsetptr2;
1631     __m256           ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
1632     int              vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D,vdwjidx0E,vdwjidx0F,vdwjidx0G,vdwjidx0H;
1633     __m256           jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
1634     int              vdwjidx1A,vdwjidx1B,vdwjidx1C,vdwjidx1D,vdwjidx1E,vdwjidx1F,vdwjidx1G,vdwjidx1H;
1635     __m256           jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
1636     int              vdwjidx2A,vdwjidx2B,vdwjidx2C,vdwjidx2D,vdwjidx2E,vdwjidx2F,vdwjidx2G,vdwjidx2H;
1637     __m256           jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
1638     __m256           dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
1639     __m256           dx01,dy01,dz01,rsq01,rinv01,rinvsq01,r01,qq01,c6_01,c12_01;
1640     __m256           dx02,dy02,dz02,rsq02,rinv02,rinvsq02,r02,qq02,c6_02,c12_02;
1641     __m256           dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
1642     __m256           dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11;
1643     __m256           dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12;
1644     __m256           dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
1645     __m256           dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21;
1646     __m256           dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22;
1647     __m256           velec,felec,velecsum,facel,crf,krf,krf2;
1648     real             *charge;
1649     __m256i          ewitab;
1650     __m128i          ewitab_lo,ewitab_hi;
1651     __m256           ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
1652     __m256           beta,beta2,beta3,zeta2,pmecorrF,pmecorrV,rinv3;
1653     real             *ewtab;
1654     __m256           rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
1655     real             rswitch_scalar,d_scalar;
1656     __m256           dummy_mask,cutoff_mask;
1657     __m256           signbit = _mm256_castsi256_ps( _mm256_set1_epi32(0x80000000) );
1658     __m256           one     = _mm256_set1_ps(1.0);
1659     __m256           two     = _mm256_set1_ps(2.0);
1660     x                = xx[0];
1661     f                = ff[0];
1662
1663     nri              = nlist->nri;
1664     iinr             = nlist->iinr;
1665     jindex           = nlist->jindex;
1666     jjnr             = nlist->jjnr;
1667     shiftidx         = nlist->shift;
1668     gid              = nlist->gid;
1669     shiftvec         = fr->shift_vec[0];
1670     fshift           = fr->fshift[0];
1671     facel            = _mm256_set1_ps(fr->ic->epsfac);
1672     charge           = mdatoms->chargeA;
1673
1674     sh_ewald         = _mm256_set1_ps(fr->ic->sh_ewald);
1675     beta             = _mm256_set1_ps(fr->ic->ewaldcoeff_q);
1676     beta2            = _mm256_mul_ps(beta,beta);
1677     beta3            = _mm256_mul_ps(beta,beta2);
1678
1679     ewtab            = fr->ic->tabq_coul_FDV0;
1680     ewtabscale       = _mm256_set1_ps(fr->ic->tabq_scale);
1681     ewtabhalfspace   = _mm256_set1_ps(0.5/fr->ic->tabq_scale);
1682
1683     /* Setup water-specific parameters */
1684     inr              = nlist->iinr[0];
1685     iq0              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+0]));
1686     iq1              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+1]));
1687     iq2              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+2]));
1688
1689     jq0              = _mm256_set1_ps(charge[inr+0]);
1690     jq1              = _mm256_set1_ps(charge[inr+1]);
1691     jq2              = _mm256_set1_ps(charge[inr+2]);
1692     qq00             = _mm256_mul_ps(iq0,jq0);
1693     qq01             = _mm256_mul_ps(iq0,jq1);
1694     qq02             = _mm256_mul_ps(iq0,jq2);
1695     qq10             = _mm256_mul_ps(iq1,jq0);
1696     qq11             = _mm256_mul_ps(iq1,jq1);
1697     qq12             = _mm256_mul_ps(iq1,jq2);
1698     qq20             = _mm256_mul_ps(iq2,jq0);
1699     qq21             = _mm256_mul_ps(iq2,jq1);
1700     qq22             = _mm256_mul_ps(iq2,jq2);
1701
1702     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
1703     rcutoff_scalar   = fr->ic->rcoulomb;
1704     rcutoff          = _mm256_set1_ps(rcutoff_scalar);
1705     rcutoff2         = _mm256_mul_ps(rcutoff,rcutoff);
1706
1707     rswitch_scalar   = fr->ic->rcoulomb_switch;
1708     rswitch          = _mm256_set1_ps(rswitch_scalar);
1709     /* Setup switch parameters */
1710     d_scalar         = rcutoff_scalar-rswitch_scalar;
1711     d                = _mm256_set1_ps(d_scalar);
1712     swV3             = _mm256_set1_ps(-10.0/(d_scalar*d_scalar*d_scalar));
1713     swV4             = _mm256_set1_ps( 15.0/(d_scalar*d_scalar*d_scalar*d_scalar));
1714     swV5             = _mm256_set1_ps( -6.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
1715     swF2             = _mm256_set1_ps(-30.0/(d_scalar*d_scalar*d_scalar));
1716     swF3             = _mm256_set1_ps( 60.0/(d_scalar*d_scalar*d_scalar*d_scalar));
1717     swF4             = _mm256_set1_ps(-30.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
1718
1719     /* Avoid stupid compiler warnings */
1720     jnrA = jnrB = jnrC = jnrD = jnrE = jnrF = jnrG = jnrH = 0;
1721     j_coord_offsetA = 0;
1722     j_coord_offsetB = 0;
1723     j_coord_offsetC = 0;
1724     j_coord_offsetD = 0;
1725     j_coord_offsetE = 0;
1726     j_coord_offsetF = 0;
1727     j_coord_offsetG = 0;
1728     j_coord_offsetH = 0;
1729
1730     outeriter        = 0;
1731     inneriter        = 0;
1732
1733     for(iidx=0;iidx<4*DIM;iidx++)
1734     {
1735         scratch[iidx] = 0.0;
1736     }
1737
1738     /* Start outer loop over neighborlists */
1739     for(iidx=0; iidx<nri; iidx++)
1740     {
1741         /* Load shift vector for this list */
1742         i_shift_offset   = DIM*shiftidx[iidx];
1743
1744         /* Load limits for loop over neighbors */
1745         j_index_start    = jindex[iidx];
1746         j_index_end      = jindex[iidx+1];
1747
1748         /* Get outer coordinate index */
1749         inr              = iinr[iidx];
1750         i_coord_offset   = DIM*inr;
1751
1752         /* Load i particle coords and add shift vector */
1753         gmx_mm256_load_shift_and_3rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,
1754                                                     &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2);
1755
1756         fix0             = _mm256_setzero_ps();
1757         fiy0             = _mm256_setzero_ps();
1758         fiz0             = _mm256_setzero_ps();
1759         fix1             = _mm256_setzero_ps();
1760         fiy1             = _mm256_setzero_ps();
1761         fiz1             = _mm256_setzero_ps();
1762         fix2             = _mm256_setzero_ps();
1763         fiy2             = _mm256_setzero_ps();
1764         fiz2             = _mm256_setzero_ps();
1765
1766         /* Start inner kernel loop */
1767         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+7]>=0; jidx+=8)
1768         {
1769
1770             /* Get j neighbor index, and coordinate index */
1771             jnrA             = jjnr[jidx];
1772             jnrB             = jjnr[jidx+1];
1773             jnrC             = jjnr[jidx+2];
1774             jnrD             = jjnr[jidx+3];
1775             jnrE             = jjnr[jidx+4];
1776             jnrF             = jjnr[jidx+5];
1777             jnrG             = jjnr[jidx+6];
1778             jnrH             = jjnr[jidx+7];
1779             j_coord_offsetA  = DIM*jnrA;
1780             j_coord_offsetB  = DIM*jnrB;
1781             j_coord_offsetC  = DIM*jnrC;
1782             j_coord_offsetD  = DIM*jnrD;
1783             j_coord_offsetE  = DIM*jnrE;
1784             j_coord_offsetF  = DIM*jnrF;
1785             j_coord_offsetG  = DIM*jnrG;
1786             j_coord_offsetH  = DIM*jnrH;
1787
1788             /* load j atom coordinates */
1789             gmx_mm256_load_3rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
1790                                                  x+j_coord_offsetC,x+j_coord_offsetD,
1791                                                  x+j_coord_offsetE,x+j_coord_offsetF,
1792                                                  x+j_coord_offsetG,x+j_coord_offsetH,
1793                                               &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,&jy2,&jz2);
1794
1795             /* Calculate displacement vector */
1796             dx00             = _mm256_sub_ps(ix0,jx0);
1797             dy00             = _mm256_sub_ps(iy0,jy0);
1798             dz00             = _mm256_sub_ps(iz0,jz0);
1799             dx01             = _mm256_sub_ps(ix0,jx1);
1800             dy01             = _mm256_sub_ps(iy0,jy1);
1801             dz01             = _mm256_sub_ps(iz0,jz1);
1802             dx02             = _mm256_sub_ps(ix0,jx2);
1803             dy02             = _mm256_sub_ps(iy0,jy2);
1804             dz02             = _mm256_sub_ps(iz0,jz2);
1805             dx10             = _mm256_sub_ps(ix1,jx0);
1806             dy10             = _mm256_sub_ps(iy1,jy0);
1807             dz10             = _mm256_sub_ps(iz1,jz0);
1808             dx11             = _mm256_sub_ps(ix1,jx1);
1809             dy11             = _mm256_sub_ps(iy1,jy1);
1810             dz11             = _mm256_sub_ps(iz1,jz1);
1811             dx12             = _mm256_sub_ps(ix1,jx2);
1812             dy12             = _mm256_sub_ps(iy1,jy2);
1813             dz12             = _mm256_sub_ps(iz1,jz2);
1814             dx20             = _mm256_sub_ps(ix2,jx0);
1815             dy20             = _mm256_sub_ps(iy2,jy0);
1816             dz20             = _mm256_sub_ps(iz2,jz0);
1817             dx21             = _mm256_sub_ps(ix2,jx1);
1818             dy21             = _mm256_sub_ps(iy2,jy1);
1819             dz21             = _mm256_sub_ps(iz2,jz1);
1820             dx22             = _mm256_sub_ps(ix2,jx2);
1821             dy22             = _mm256_sub_ps(iy2,jy2);
1822             dz22             = _mm256_sub_ps(iz2,jz2);
1823
1824             /* Calculate squared distance and things based on it */
1825             rsq00            = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
1826             rsq01            = gmx_mm256_calc_rsq_ps(dx01,dy01,dz01);
1827             rsq02            = gmx_mm256_calc_rsq_ps(dx02,dy02,dz02);
1828             rsq10            = gmx_mm256_calc_rsq_ps(dx10,dy10,dz10);
1829             rsq11            = gmx_mm256_calc_rsq_ps(dx11,dy11,dz11);
1830             rsq12            = gmx_mm256_calc_rsq_ps(dx12,dy12,dz12);
1831             rsq20            = gmx_mm256_calc_rsq_ps(dx20,dy20,dz20);
1832             rsq21            = gmx_mm256_calc_rsq_ps(dx21,dy21,dz21);
1833             rsq22            = gmx_mm256_calc_rsq_ps(dx22,dy22,dz22);
1834
1835             rinv00           = avx256_invsqrt_f(rsq00);
1836             rinv01           = avx256_invsqrt_f(rsq01);
1837             rinv02           = avx256_invsqrt_f(rsq02);
1838             rinv10           = avx256_invsqrt_f(rsq10);
1839             rinv11           = avx256_invsqrt_f(rsq11);
1840             rinv12           = avx256_invsqrt_f(rsq12);
1841             rinv20           = avx256_invsqrt_f(rsq20);
1842             rinv21           = avx256_invsqrt_f(rsq21);
1843             rinv22           = avx256_invsqrt_f(rsq22);
1844
1845             rinvsq00         = _mm256_mul_ps(rinv00,rinv00);
1846             rinvsq01         = _mm256_mul_ps(rinv01,rinv01);
1847             rinvsq02         = _mm256_mul_ps(rinv02,rinv02);
1848             rinvsq10         = _mm256_mul_ps(rinv10,rinv10);
1849             rinvsq11         = _mm256_mul_ps(rinv11,rinv11);
1850             rinvsq12         = _mm256_mul_ps(rinv12,rinv12);
1851             rinvsq20         = _mm256_mul_ps(rinv20,rinv20);
1852             rinvsq21         = _mm256_mul_ps(rinv21,rinv21);
1853             rinvsq22         = _mm256_mul_ps(rinv22,rinv22);
1854
1855             fjx0             = _mm256_setzero_ps();
1856             fjy0             = _mm256_setzero_ps();
1857             fjz0             = _mm256_setzero_ps();
1858             fjx1             = _mm256_setzero_ps();
1859             fjy1             = _mm256_setzero_ps();
1860             fjz1             = _mm256_setzero_ps();
1861             fjx2             = _mm256_setzero_ps();
1862             fjy2             = _mm256_setzero_ps();
1863             fjz2             = _mm256_setzero_ps();
1864
1865             /**************************
1866              * CALCULATE INTERACTIONS *
1867              **************************/
1868
1869             if (gmx_mm256_any_lt(rsq00,rcutoff2))
1870             {
1871
1872             r00              = _mm256_mul_ps(rsq00,rinv00);
1873
1874             /* EWALD ELECTROSTATICS */
1875
1876             /* Analytical PME correction */
1877             zeta2            = _mm256_mul_ps(beta2,rsq00);
1878             rinv3            = _mm256_mul_ps(rinvsq00,rinv00);
1879             pmecorrF         = avx256_pmecorrF_f(zeta2);
1880             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1881             felec            = _mm256_mul_ps(qq00,felec);
1882             pmecorrV         = avx256_pmecorrV_f(zeta2);
1883             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1884             velec            = _mm256_sub_ps(rinv00,pmecorrV);
1885             velec            = _mm256_mul_ps(qq00,velec);
1886
1887             d                = _mm256_sub_ps(r00,rswitch);
1888             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1889             d2               = _mm256_mul_ps(d,d);
1890             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)))))));
1891
1892             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1893
1894             /* Evaluate switch function */
1895             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1896             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv00,_mm256_mul_ps(velec,dsw)) );
1897             cutoff_mask      = _mm256_cmp_ps(rsq00,rcutoff2,_CMP_LT_OQ);
1898
1899             fscal            = felec;
1900
1901             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1902
1903             /* Calculate temporary vectorial force */
1904             tx               = _mm256_mul_ps(fscal,dx00);
1905             ty               = _mm256_mul_ps(fscal,dy00);
1906             tz               = _mm256_mul_ps(fscal,dz00);
1907
1908             /* Update vectorial force */
1909             fix0             = _mm256_add_ps(fix0,tx);
1910             fiy0             = _mm256_add_ps(fiy0,ty);
1911             fiz0             = _mm256_add_ps(fiz0,tz);
1912
1913             fjx0             = _mm256_add_ps(fjx0,tx);
1914             fjy0             = _mm256_add_ps(fjy0,ty);
1915             fjz0             = _mm256_add_ps(fjz0,tz);
1916
1917             }
1918
1919             /**************************
1920              * CALCULATE INTERACTIONS *
1921              **************************/
1922
1923             if (gmx_mm256_any_lt(rsq01,rcutoff2))
1924             {
1925
1926             r01              = _mm256_mul_ps(rsq01,rinv01);
1927
1928             /* EWALD ELECTROSTATICS */
1929
1930             /* Analytical PME correction */
1931             zeta2            = _mm256_mul_ps(beta2,rsq01);
1932             rinv3            = _mm256_mul_ps(rinvsq01,rinv01);
1933             pmecorrF         = avx256_pmecorrF_f(zeta2);
1934             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1935             felec            = _mm256_mul_ps(qq01,felec);
1936             pmecorrV         = avx256_pmecorrV_f(zeta2);
1937             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1938             velec            = _mm256_sub_ps(rinv01,pmecorrV);
1939             velec            = _mm256_mul_ps(qq01,velec);
1940
1941             d                = _mm256_sub_ps(r01,rswitch);
1942             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1943             d2               = _mm256_mul_ps(d,d);
1944             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)))))));
1945
1946             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
1947
1948             /* Evaluate switch function */
1949             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1950             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv01,_mm256_mul_ps(velec,dsw)) );
1951             cutoff_mask      = _mm256_cmp_ps(rsq01,rcutoff2,_CMP_LT_OQ);
1952
1953             fscal            = felec;
1954
1955             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1956
1957             /* Calculate temporary vectorial force */
1958             tx               = _mm256_mul_ps(fscal,dx01);
1959             ty               = _mm256_mul_ps(fscal,dy01);
1960             tz               = _mm256_mul_ps(fscal,dz01);
1961
1962             /* Update vectorial force */
1963             fix0             = _mm256_add_ps(fix0,tx);
1964             fiy0             = _mm256_add_ps(fiy0,ty);
1965             fiz0             = _mm256_add_ps(fiz0,tz);
1966
1967             fjx1             = _mm256_add_ps(fjx1,tx);
1968             fjy1             = _mm256_add_ps(fjy1,ty);
1969             fjz1             = _mm256_add_ps(fjz1,tz);
1970
1971             }
1972
1973             /**************************
1974              * CALCULATE INTERACTIONS *
1975              **************************/
1976
1977             if (gmx_mm256_any_lt(rsq02,rcutoff2))
1978             {
1979
1980             r02              = _mm256_mul_ps(rsq02,rinv02);
1981
1982             /* EWALD ELECTROSTATICS */
1983
1984             /* Analytical PME correction */
1985             zeta2            = _mm256_mul_ps(beta2,rsq02);
1986             rinv3            = _mm256_mul_ps(rinvsq02,rinv02);
1987             pmecorrF         = avx256_pmecorrF_f(zeta2);
1988             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1989             felec            = _mm256_mul_ps(qq02,felec);
1990             pmecorrV         = avx256_pmecorrV_f(zeta2);
1991             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
1992             velec            = _mm256_sub_ps(rinv02,pmecorrV);
1993             velec            = _mm256_mul_ps(qq02,velec);
1994
1995             d                = _mm256_sub_ps(r02,rswitch);
1996             d                = _mm256_max_ps(d,_mm256_setzero_ps());
1997             d2               = _mm256_mul_ps(d,d);
1998             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)))))));
1999
2000             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2001
2002             /* Evaluate switch function */
2003             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2004             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv02,_mm256_mul_ps(velec,dsw)) );
2005             cutoff_mask      = _mm256_cmp_ps(rsq02,rcutoff2,_CMP_LT_OQ);
2006
2007             fscal            = felec;
2008
2009             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2010
2011             /* Calculate temporary vectorial force */
2012             tx               = _mm256_mul_ps(fscal,dx02);
2013             ty               = _mm256_mul_ps(fscal,dy02);
2014             tz               = _mm256_mul_ps(fscal,dz02);
2015
2016             /* Update vectorial force */
2017             fix0             = _mm256_add_ps(fix0,tx);
2018             fiy0             = _mm256_add_ps(fiy0,ty);
2019             fiz0             = _mm256_add_ps(fiz0,tz);
2020
2021             fjx2             = _mm256_add_ps(fjx2,tx);
2022             fjy2             = _mm256_add_ps(fjy2,ty);
2023             fjz2             = _mm256_add_ps(fjz2,tz);
2024
2025             }
2026
2027             /**************************
2028              * CALCULATE INTERACTIONS *
2029              **************************/
2030
2031             if (gmx_mm256_any_lt(rsq10,rcutoff2))
2032             {
2033
2034             r10              = _mm256_mul_ps(rsq10,rinv10);
2035
2036             /* EWALD ELECTROSTATICS */
2037
2038             /* Analytical PME correction */
2039             zeta2            = _mm256_mul_ps(beta2,rsq10);
2040             rinv3            = _mm256_mul_ps(rinvsq10,rinv10);
2041             pmecorrF         = avx256_pmecorrF_f(zeta2);
2042             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2043             felec            = _mm256_mul_ps(qq10,felec);
2044             pmecorrV         = avx256_pmecorrV_f(zeta2);
2045             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2046             velec            = _mm256_sub_ps(rinv10,pmecorrV);
2047             velec            = _mm256_mul_ps(qq10,velec);
2048
2049             d                = _mm256_sub_ps(r10,rswitch);
2050             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2051             d2               = _mm256_mul_ps(d,d);
2052             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)))))));
2053
2054             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2055
2056             /* Evaluate switch function */
2057             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2058             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv10,_mm256_mul_ps(velec,dsw)) );
2059             cutoff_mask      = _mm256_cmp_ps(rsq10,rcutoff2,_CMP_LT_OQ);
2060
2061             fscal            = felec;
2062
2063             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2064
2065             /* Calculate temporary vectorial force */
2066             tx               = _mm256_mul_ps(fscal,dx10);
2067             ty               = _mm256_mul_ps(fscal,dy10);
2068             tz               = _mm256_mul_ps(fscal,dz10);
2069
2070             /* Update vectorial force */
2071             fix1             = _mm256_add_ps(fix1,tx);
2072             fiy1             = _mm256_add_ps(fiy1,ty);
2073             fiz1             = _mm256_add_ps(fiz1,tz);
2074
2075             fjx0             = _mm256_add_ps(fjx0,tx);
2076             fjy0             = _mm256_add_ps(fjy0,ty);
2077             fjz0             = _mm256_add_ps(fjz0,tz);
2078
2079             }
2080
2081             /**************************
2082              * CALCULATE INTERACTIONS *
2083              **************************/
2084
2085             if (gmx_mm256_any_lt(rsq11,rcutoff2))
2086             {
2087
2088             r11              = _mm256_mul_ps(rsq11,rinv11);
2089
2090             /* EWALD ELECTROSTATICS */
2091
2092             /* Analytical PME correction */
2093             zeta2            = _mm256_mul_ps(beta2,rsq11);
2094             rinv3            = _mm256_mul_ps(rinvsq11,rinv11);
2095             pmecorrF         = avx256_pmecorrF_f(zeta2);
2096             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2097             felec            = _mm256_mul_ps(qq11,felec);
2098             pmecorrV         = avx256_pmecorrV_f(zeta2);
2099             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2100             velec            = _mm256_sub_ps(rinv11,pmecorrV);
2101             velec            = _mm256_mul_ps(qq11,velec);
2102
2103             d                = _mm256_sub_ps(r11,rswitch);
2104             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2105             d2               = _mm256_mul_ps(d,d);
2106             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)))))));
2107
2108             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2109
2110             /* Evaluate switch function */
2111             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2112             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv11,_mm256_mul_ps(velec,dsw)) );
2113             cutoff_mask      = _mm256_cmp_ps(rsq11,rcutoff2,_CMP_LT_OQ);
2114
2115             fscal            = felec;
2116
2117             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2118
2119             /* Calculate temporary vectorial force */
2120             tx               = _mm256_mul_ps(fscal,dx11);
2121             ty               = _mm256_mul_ps(fscal,dy11);
2122             tz               = _mm256_mul_ps(fscal,dz11);
2123
2124             /* Update vectorial force */
2125             fix1             = _mm256_add_ps(fix1,tx);
2126             fiy1             = _mm256_add_ps(fiy1,ty);
2127             fiz1             = _mm256_add_ps(fiz1,tz);
2128
2129             fjx1             = _mm256_add_ps(fjx1,tx);
2130             fjy1             = _mm256_add_ps(fjy1,ty);
2131             fjz1             = _mm256_add_ps(fjz1,tz);
2132
2133             }
2134
2135             /**************************
2136              * CALCULATE INTERACTIONS *
2137              **************************/
2138
2139             if (gmx_mm256_any_lt(rsq12,rcutoff2))
2140             {
2141
2142             r12              = _mm256_mul_ps(rsq12,rinv12);
2143
2144             /* EWALD ELECTROSTATICS */
2145
2146             /* Analytical PME correction */
2147             zeta2            = _mm256_mul_ps(beta2,rsq12);
2148             rinv3            = _mm256_mul_ps(rinvsq12,rinv12);
2149             pmecorrF         = avx256_pmecorrF_f(zeta2);
2150             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2151             felec            = _mm256_mul_ps(qq12,felec);
2152             pmecorrV         = avx256_pmecorrV_f(zeta2);
2153             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2154             velec            = _mm256_sub_ps(rinv12,pmecorrV);
2155             velec            = _mm256_mul_ps(qq12,velec);
2156
2157             d                = _mm256_sub_ps(r12,rswitch);
2158             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2159             d2               = _mm256_mul_ps(d,d);
2160             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)))))));
2161
2162             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2163
2164             /* Evaluate switch function */
2165             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2166             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv12,_mm256_mul_ps(velec,dsw)) );
2167             cutoff_mask      = _mm256_cmp_ps(rsq12,rcutoff2,_CMP_LT_OQ);
2168
2169             fscal            = felec;
2170
2171             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2172
2173             /* Calculate temporary vectorial force */
2174             tx               = _mm256_mul_ps(fscal,dx12);
2175             ty               = _mm256_mul_ps(fscal,dy12);
2176             tz               = _mm256_mul_ps(fscal,dz12);
2177
2178             /* Update vectorial force */
2179             fix1             = _mm256_add_ps(fix1,tx);
2180             fiy1             = _mm256_add_ps(fiy1,ty);
2181             fiz1             = _mm256_add_ps(fiz1,tz);
2182
2183             fjx2             = _mm256_add_ps(fjx2,tx);
2184             fjy2             = _mm256_add_ps(fjy2,ty);
2185             fjz2             = _mm256_add_ps(fjz2,tz);
2186
2187             }
2188
2189             /**************************
2190              * CALCULATE INTERACTIONS *
2191              **************************/
2192
2193             if (gmx_mm256_any_lt(rsq20,rcutoff2))
2194             {
2195
2196             r20              = _mm256_mul_ps(rsq20,rinv20);
2197
2198             /* EWALD ELECTROSTATICS */
2199
2200             /* Analytical PME correction */
2201             zeta2            = _mm256_mul_ps(beta2,rsq20);
2202             rinv3            = _mm256_mul_ps(rinvsq20,rinv20);
2203             pmecorrF         = avx256_pmecorrF_f(zeta2);
2204             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2205             felec            = _mm256_mul_ps(qq20,felec);
2206             pmecorrV         = avx256_pmecorrV_f(zeta2);
2207             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2208             velec            = _mm256_sub_ps(rinv20,pmecorrV);
2209             velec            = _mm256_mul_ps(qq20,velec);
2210
2211             d                = _mm256_sub_ps(r20,rswitch);
2212             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2213             d2               = _mm256_mul_ps(d,d);
2214             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)))))));
2215
2216             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2217
2218             /* Evaluate switch function */
2219             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2220             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv20,_mm256_mul_ps(velec,dsw)) );
2221             cutoff_mask      = _mm256_cmp_ps(rsq20,rcutoff2,_CMP_LT_OQ);
2222
2223             fscal            = felec;
2224
2225             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2226
2227             /* Calculate temporary vectorial force */
2228             tx               = _mm256_mul_ps(fscal,dx20);
2229             ty               = _mm256_mul_ps(fscal,dy20);
2230             tz               = _mm256_mul_ps(fscal,dz20);
2231
2232             /* Update vectorial force */
2233             fix2             = _mm256_add_ps(fix2,tx);
2234             fiy2             = _mm256_add_ps(fiy2,ty);
2235             fiz2             = _mm256_add_ps(fiz2,tz);
2236
2237             fjx0             = _mm256_add_ps(fjx0,tx);
2238             fjy0             = _mm256_add_ps(fjy0,ty);
2239             fjz0             = _mm256_add_ps(fjz0,tz);
2240
2241             }
2242
2243             /**************************
2244              * CALCULATE INTERACTIONS *
2245              **************************/
2246
2247             if (gmx_mm256_any_lt(rsq21,rcutoff2))
2248             {
2249
2250             r21              = _mm256_mul_ps(rsq21,rinv21);
2251
2252             /* EWALD ELECTROSTATICS */
2253
2254             /* Analytical PME correction */
2255             zeta2            = _mm256_mul_ps(beta2,rsq21);
2256             rinv3            = _mm256_mul_ps(rinvsq21,rinv21);
2257             pmecorrF         = avx256_pmecorrF_f(zeta2);
2258             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2259             felec            = _mm256_mul_ps(qq21,felec);
2260             pmecorrV         = avx256_pmecorrV_f(zeta2);
2261             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2262             velec            = _mm256_sub_ps(rinv21,pmecorrV);
2263             velec            = _mm256_mul_ps(qq21,velec);
2264
2265             d                = _mm256_sub_ps(r21,rswitch);
2266             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2267             d2               = _mm256_mul_ps(d,d);
2268             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)))))));
2269
2270             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2271
2272             /* Evaluate switch function */
2273             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2274             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv21,_mm256_mul_ps(velec,dsw)) );
2275             cutoff_mask      = _mm256_cmp_ps(rsq21,rcutoff2,_CMP_LT_OQ);
2276
2277             fscal            = felec;
2278
2279             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2280
2281             /* Calculate temporary vectorial force */
2282             tx               = _mm256_mul_ps(fscal,dx21);
2283             ty               = _mm256_mul_ps(fscal,dy21);
2284             tz               = _mm256_mul_ps(fscal,dz21);
2285
2286             /* Update vectorial force */
2287             fix2             = _mm256_add_ps(fix2,tx);
2288             fiy2             = _mm256_add_ps(fiy2,ty);
2289             fiz2             = _mm256_add_ps(fiz2,tz);
2290
2291             fjx1             = _mm256_add_ps(fjx1,tx);
2292             fjy1             = _mm256_add_ps(fjy1,ty);
2293             fjz1             = _mm256_add_ps(fjz1,tz);
2294
2295             }
2296
2297             /**************************
2298              * CALCULATE INTERACTIONS *
2299              **************************/
2300
2301             if (gmx_mm256_any_lt(rsq22,rcutoff2))
2302             {
2303
2304             r22              = _mm256_mul_ps(rsq22,rinv22);
2305
2306             /* EWALD ELECTROSTATICS */
2307
2308             /* Analytical PME correction */
2309             zeta2            = _mm256_mul_ps(beta2,rsq22);
2310             rinv3            = _mm256_mul_ps(rinvsq22,rinv22);
2311             pmecorrF         = avx256_pmecorrF_f(zeta2);
2312             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2313             felec            = _mm256_mul_ps(qq22,felec);
2314             pmecorrV         = avx256_pmecorrV_f(zeta2);
2315             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2316             velec            = _mm256_sub_ps(rinv22,pmecorrV);
2317             velec            = _mm256_mul_ps(qq22,velec);
2318
2319             d                = _mm256_sub_ps(r22,rswitch);
2320             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2321             d2               = _mm256_mul_ps(d,d);
2322             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)))))));
2323
2324             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2325
2326             /* Evaluate switch function */
2327             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2328             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv22,_mm256_mul_ps(velec,dsw)) );
2329             cutoff_mask      = _mm256_cmp_ps(rsq22,rcutoff2,_CMP_LT_OQ);
2330
2331             fscal            = felec;
2332
2333             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2334
2335             /* Calculate temporary vectorial force */
2336             tx               = _mm256_mul_ps(fscal,dx22);
2337             ty               = _mm256_mul_ps(fscal,dy22);
2338             tz               = _mm256_mul_ps(fscal,dz22);
2339
2340             /* Update vectorial force */
2341             fix2             = _mm256_add_ps(fix2,tx);
2342             fiy2             = _mm256_add_ps(fiy2,ty);
2343             fiz2             = _mm256_add_ps(fiz2,tz);
2344
2345             fjx2             = _mm256_add_ps(fjx2,tx);
2346             fjy2             = _mm256_add_ps(fjy2,ty);
2347             fjz2             = _mm256_add_ps(fjz2,tz);
2348
2349             }
2350
2351             fjptrA             = f+j_coord_offsetA;
2352             fjptrB             = f+j_coord_offsetB;
2353             fjptrC             = f+j_coord_offsetC;
2354             fjptrD             = f+j_coord_offsetD;
2355             fjptrE             = f+j_coord_offsetE;
2356             fjptrF             = f+j_coord_offsetF;
2357             fjptrG             = f+j_coord_offsetG;
2358             fjptrH             = f+j_coord_offsetH;
2359
2360             gmx_mm256_decrement_3rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,
2361                                                       fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2);
2362
2363             /* Inner loop uses 945 flops */
2364         }
2365
2366         if(jidx<j_index_end)
2367         {
2368
2369             /* Get j neighbor index, and coordinate index */
2370             jnrlistA         = jjnr[jidx];
2371             jnrlistB         = jjnr[jidx+1];
2372             jnrlistC         = jjnr[jidx+2];
2373             jnrlistD         = jjnr[jidx+3];
2374             jnrlistE         = jjnr[jidx+4];
2375             jnrlistF         = jjnr[jidx+5];
2376             jnrlistG         = jjnr[jidx+6];
2377             jnrlistH         = jjnr[jidx+7];
2378             /* Sign of each element will be negative for non-real atoms.
2379              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
2380              * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
2381              */
2382             dummy_mask = gmx_mm256_set_m128(gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx+4)),_mm_setzero_si128())),
2383                                             gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128())));
2384
2385             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
2386             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
2387             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
2388             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
2389             jnrE       = (jnrlistE>=0) ? jnrlistE : 0;
2390             jnrF       = (jnrlistF>=0) ? jnrlistF : 0;
2391             jnrG       = (jnrlistG>=0) ? jnrlistG : 0;
2392             jnrH       = (jnrlistH>=0) ? jnrlistH : 0;
2393             j_coord_offsetA  = DIM*jnrA;
2394             j_coord_offsetB  = DIM*jnrB;
2395             j_coord_offsetC  = DIM*jnrC;
2396             j_coord_offsetD  = DIM*jnrD;
2397             j_coord_offsetE  = DIM*jnrE;
2398             j_coord_offsetF  = DIM*jnrF;
2399             j_coord_offsetG  = DIM*jnrG;
2400             j_coord_offsetH  = DIM*jnrH;
2401
2402             /* load j atom coordinates */
2403             gmx_mm256_load_3rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
2404                                                  x+j_coord_offsetC,x+j_coord_offsetD,
2405                                                  x+j_coord_offsetE,x+j_coord_offsetF,
2406                                                  x+j_coord_offsetG,x+j_coord_offsetH,
2407                                               &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,&jy2,&jz2);
2408
2409             /* Calculate displacement vector */
2410             dx00             = _mm256_sub_ps(ix0,jx0);
2411             dy00             = _mm256_sub_ps(iy0,jy0);
2412             dz00             = _mm256_sub_ps(iz0,jz0);
2413             dx01             = _mm256_sub_ps(ix0,jx1);
2414             dy01             = _mm256_sub_ps(iy0,jy1);
2415             dz01             = _mm256_sub_ps(iz0,jz1);
2416             dx02             = _mm256_sub_ps(ix0,jx2);
2417             dy02             = _mm256_sub_ps(iy0,jy2);
2418             dz02             = _mm256_sub_ps(iz0,jz2);
2419             dx10             = _mm256_sub_ps(ix1,jx0);
2420             dy10             = _mm256_sub_ps(iy1,jy0);
2421             dz10             = _mm256_sub_ps(iz1,jz0);
2422             dx11             = _mm256_sub_ps(ix1,jx1);
2423             dy11             = _mm256_sub_ps(iy1,jy1);
2424             dz11             = _mm256_sub_ps(iz1,jz1);
2425             dx12             = _mm256_sub_ps(ix1,jx2);
2426             dy12             = _mm256_sub_ps(iy1,jy2);
2427             dz12             = _mm256_sub_ps(iz1,jz2);
2428             dx20             = _mm256_sub_ps(ix2,jx0);
2429             dy20             = _mm256_sub_ps(iy2,jy0);
2430             dz20             = _mm256_sub_ps(iz2,jz0);
2431             dx21             = _mm256_sub_ps(ix2,jx1);
2432             dy21             = _mm256_sub_ps(iy2,jy1);
2433             dz21             = _mm256_sub_ps(iz2,jz1);
2434             dx22             = _mm256_sub_ps(ix2,jx2);
2435             dy22             = _mm256_sub_ps(iy2,jy2);
2436             dz22             = _mm256_sub_ps(iz2,jz2);
2437
2438             /* Calculate squared distance and things based on it */
2439             rsq00            = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
2440             rsq01            = gmx_mm256_calc_rsq_ps(dx01,dy01,dz01);
2441             rsq02            = gmx_mm256_calc_rsq_ps(dx02,dy02,dz02);
2442             rsq10            = gmx_mm256_calc_rsq_ps(dx10,dy10,dz10);
2443             rsq11            = gmx_mm256_calc_rsq_ps(dx11,dy11,dz11);
2444             rsq12            = gmx_mm256_calc_rsq_ps(dx12,dy12,dz12);
2445             rsq20            = gmx_mm256_calc_rsq_ps(dx20,dy20,dz20);
2446             rsq21            = gmx_mm256_calc_rsq_ps(dx21,dy21,dz21);
2447             rsq22            = gmx_mm256_calc_rsq_ps(dx22,dy22,dz22);
2448
2449             rinv00           = avx256_invsqrt_f(rsq00);
2450             rinv01           = avx256_invsqrt_f(rsq01);
2451             rinv02           = avx256_invsqrt_f(rsq02);
2452             rinv10           = avx256_invsqrt_f(rsq10);
2453             rinv11           = avx256_invsqrt_f(rsq11);
2454             rinv12           = avx256_invsqrt_f(rsq12);
2455             rinv20           = avx256_invsqrt_f(rsq20);
2456             rinv21           = avx256_invsqrt_f(rsq21);
2457             rinv22           = avx256_invsqrt_f(rsq22);
2458
2459             rinvsq00         = _mm256_mul_ps(rinv00,rinv00);
2460             rinvsq01         = _mm256_mul_ps(rinv01,rinv01);
2461             rinvsq02         = _mm256_mul_ps(rinv02,rinv02);
2462             rinvsq10         = _mm256_mul_ps(rinv10,rinv10);
2463             rinvsq11         = _mm256_mul_ps(rinv11,rinv11);
2464             rinvsq12         = _mm256_mul_ps(rinv12,rinv12);
2465             rinvsq20         = _mm256_mul_ps(rinv20,rinv20);
2466             rinvsq21         = _mm256_mul_ps(rinv21,rinv21);
2467             rinvsq22         = _mm256_mul_ps(rinv22,rinv22);
2468
2469             fjx0             = _mm256_setzero_ps();
2470             fjy0             = _mm256_setzero_ps();
2471             fjz0             = _mm256_setzero_ps();
2472             fjx1             = _mm256_setzero_ps();
2473             fjy1             = _mm256_setzero_ps();
2474             fjz1             = _mm256_setzero_ps();
2475             fjx2             = _mm256_setzero_ps();
2476             fjy2             = _mm256_setzero_ps();
2477             fjz2             = _mm256_setzero_ps();
2478
2479             /**************************
2480              * CALCULATE INTERACTIONS *
2481              **************************/
2482
2483             if (gmx_mm256_any_lt(rsq00,rcutoff2))
2484             {
2485
2486             r00              = _mm256_mul_ps(rsq00,rinv00);
2487             r00              = _mm256_andnot_ps(dummy_mask,r00);
2488
2489             /* EWALD ELECTROSTATICS */
2490
2491             /* Analytical PME correction */
2492             zeta2            = _mm256_mul_ps(beta2,rsq00);
2493             rinv3            = _mm256_mul_ps(rinvsq00,rinv00);
2494             pmecorrF         = avx256_pmecorrF_f(zeta2);
2495             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2496             felec            = _mm256_mul_ps(qq00,felec);
2497             pmecorrV         = avx256_pmecorrV_f(zeta2);
2498             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2499             velec            = _mm256_sub_ps(rinv00,pmecorrV);
2500             velec            = _mm256_mul_ps(qq00,velec);
2501
2502             d                = _mm256_sub_ps(r00,rswitch);
2503             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2504             d2               = _mm256_mul_ps(d,d);
2505             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)))))));
2506
2507             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2508
2509             /* Evaluate switch function */
2510             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2511             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv00,_mm256_mul_ps(velec,dsw)) );
2512             cutoff_mask      = _mm256_cmp_ps(rsq00,rcutoff2,_CMP_LT_OQ);
2513
2514             fscal            = felec;
2515
2516             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2517
2518             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
2519
2520             /* Calculate temporary vectorial force */
2521             tx               = _mm256_mul_ps(fscal,dx00);
2522             ty               = _mm256_mul_ps(fscal,dy00);
2523             tz               = _mm256_mul_ps(fscal,dz00);
2524
2525             /* Update vectorial force */
2526             fix0             = _mm256_add_ps(fix0,tx);
2527             fiy0             = _mm256_add_ps(fiy0,ty);
2528             fiz0             = _mm256_add_ps(fiz0,tz);
2529
2530             fjx0             = _mm256_add_ps(fjx0,tx);
2531             fjy0             = _mm256_add_ps(fjy0,ty);
2532             fjz0             = _mm256_add_ps(fjz0,tz);
2533
2534             }
2535
2536             /**************************
2537              * CALCULATE INTERACTIONS *
2538              **************************/
2539
2540             if (gmx_mm256_any_lt(rsq01,rcutoff2))
2541             {
2542
2543             r01              = _mm256_mul_ps(rsq01,rinv01);
2544             r01              = _mm256_andnot_ps(dummy_mask,r01);
2545
2546             /* EWALD ELECTROSTATICS */
2547
2548             /* Analytical PME correction */
2549             zeta2            = _mm256_mul_ps(beta2,rsq01);
2550             rinv3            = _mm256_mul_ps(rinvsq01,rinv01);
2551             pmecorrF         = avx256_pmecorrF_f(zeta2);
2552             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2553             felec            = _mm256_mul_ps(qq01,felec);
2554             pmecorrV         = avx256_pmecorrV_f(zeta2);
2555             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2556             velec            = _mm256_sub_ps(rinv01,pmecorrV);
2557             velec            = _mm256_mul_ps(qq01,velec);
2558
2559             d                = _mm256_sub_ps(r01,rswitch);
2560             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2561             d2               = _mm256_mul_ps(d,d);
2562             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)))))));
2563
2564             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2565
2566             /* Evaluate switch function */
2567             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2568             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv01,_mm256_mul_ps(velec,dsw)) );
2569             cutoff_mask      = _mm256_cmp_ps(rsq01,rcutoff2,_CMP_LT_OQ);
2570
2571             fscal            = felec;
2572
2573             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2574
2575             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
2576
2577             /* Calculate temporary vectorial force */
2578             tx               = _mm256_mul_ps(fscal,dx01);
2579             ty               = _mm256_mul_ps(fscal,dy01);
2580             tz               = _mm256_mul_ps(fscal,dz01);
2581
2582             /* Update vectorial force */
2583             fix0             = _mm256_add_ps(fix0,tx);
2584             fiy0             = _mm256_add_ps(fiy0,ty);
2585             fiz0             = _mm256_add_ps(fiz0,tz);
2586
2587             fjx1             = _mm256_add_ps(fjx1,tx);
2588             fjy1             = _mm256_add_ps(fjy1,ty);
2589             fjz1             = _mm256_add_ps(fjz1,tz);
2590
2591             }
2592
2593             /**************************
2594              * CALCULATE INTERACTIONS *
2595              **************************/
2596
2597             if (gmx_mm256_any_lt(rsq02,rcutoff2))
2598             {
2599
2600             r02              = _mm256_mul_ps(rsq02,rinv02);
2601             r02              = _mm256_andnot_ps(dummy_mask,r02);
2602
2603             /* EWALD ELECTROSTATICS */
2604
2605             /* Analytical PME correction */
2606             zeta2            = _mm256_mul_ps(beta2,rsq02);
2607             rinv3            = _mm256_mul_ps(rinvsq02,rinv02);
2608             pmecorrF         = avx256_pmecorrF_f(zeta2);
2609             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2610             felec            = _mm256_mul_ps(qq02,felec);
2611             pmecorrV         = avx256_pmecorrV_f(zeta2);
2612             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2613             velec            = _mm256_sub_ps(rinv02,pmecorrV);
2614             velec            = _mm256_mul_ps(qq02,velec);
2615
2616             d                = _mm256_sub_ps(r02,rswitch);
2617             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2618             d2               = _mm256_mul_ps(d,d);
2619             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)))))));
2620
2621             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2622
2623             /* Evaluate switch function */
2624             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2625             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv02,_mm256_mul_ps(velec,dsw)) );
2626             cutoff_mask      = _mm256_cmp_ps(rsq02,rcutoff2,_CMP_LT_OQ);
2627
2628             fscal            = felec;
2629
2630             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2631
2632             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
2633
2634             /* Calculate temporary vectorial force */
2635             tx               = _mm256_mul_ps(fscal,dx02);
2636             ty               = _mm256_mul_ps(fscal,dy02);
2637             tz               = _mm256_mul_ps(fscal,dz02);
2638
2639             /* Update vectorial force */
2640             fix0             = _mm256_add_ps(fix0,tx);
2641             fiy0             = _mm256_add_ps(fiy0,ty);
2642             fiz0             = _mm256_add_ps(fiz0,tz);
2643
2644             fjx2             = _mm256_add_ps(fjx2,tx);
2645             fjy2             = _mm256_add_ps(fjy2,ty);
2646             fjz2             = _mm256_add_ps(fjz2,tz);
2647
2648             }
2649
2650             /**************************
2651              * CALCULATE INTERACTIONS *
2652              **************************/
2653
2654             if (gmx_mm256_any_lt(rsq10,rcutoff2))
2655             {
2656
2657             r10              = _mm256_mul_ps(rsq10,rinv10);
2658             r10              = _mm256_andnot_ps(dummy_mask,r10);
2659
2660             /* EWALD ELECTROSTATICS */
2661
2662             /* Analytical PME correction */
2663             zeta2            = _mm256_mul_ps(beta2,rsq10);
2664             rinv3            = _mm256_mul_ps(rinvsq10,rinv10);
2665             pmecorrF         = avx256_pmecorrF_f(zeta2);
2666             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2667             felec            = _mm256_mul_ps(qq10,felec);
2668             pmecorrV         = avx256_pmecorrV_f(zeta2);
2669             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2670             velec            = _mm256_sub_ps(rinv10,pmecorrV);
2671             velec            = _mm256_mul_ps(qq10,velec);
2672
2673             d                = _mm256_sub_ps(r10,rswitch);
2674             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2675             d2               = _mm256_mul_ps(d,d);
2676             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)))))));
2677
2678             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2679
2680             /* Evaluate switch function */
2681             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2682             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv10,_mm256_mul_ps(velec,dsw)) );
2683             cutoff_mask      = _mm256_cmp_ps(rsq10,rcutoff2,_CMP_LT_OQ);
2684
2685             fscal            = felec;
2686
2687             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2688
2689             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
2690
2691             /* Calculate temporary vectorial force */
2692             tx               = _mm256_mul_ps(fscal,dx10);
2693             ty               = _mm256_mul_ps(fscal,dy10);
2694             tz               = _mm256_mul_ps(fscal,dz10);
2695
2696             /* Update vectorial force */
2697             fix1             = _mm256_add_ps(fix1,tx);
2698             fiy1             = _mm256_add_ps(fiy1,ty);
2699             fiz1             = _mm256_add_ps(fiz1,tz);
2700
2701             fjx0             = _mm256_add_ps(fjx0,tx);
2702             fjy0             = _mm256_add_ps(fjy0,ty);
2703             fjz0             = _mm256_add_ps(fjz0,tz);
2704
2705             }
2706
2707             /**************************
2708              * CALCULATE INTERACTIONS *
2709              **************************/
2710
2711             if (gmx_mm256_any_lt(rsq11,rcutoff2))
2712             {
2713
2714             r11              = _mm256_mul_ps(rsq11,rinv11);
2715             r11              = _mm256_andnot_ps(dummy_mask,r11);
2716
2717             /* EWALD ELECTROSTATICS */
2718
2719             /* Analytical PME correction */
2720             zeta2            = _mm256_mul_ps(beta2,rsq11);
2721             rinv3            = _mm256_mul_ps(rinvsq11,rinv11);
2722             pmecorrF         = avx256_pmecorrF_f(zeta2);
2723             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2724             felec            = _mm256_mul_ps(qq11,felec);
2725             pmecorrV         = avx256_pmecorrV_f(zeta2);
2726             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2727             velec            = _mm256_sub_ps(rinv11,pmecorrV);
2728             velec            = _mm256_mul_ps(qq11,velec);
2729
2730             d                = _mm256_sub_ps(r11,rswitch);
2731             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2732             d2               = _mm256_mul_ps(d,d);
2733             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)))))));
2734
2735             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2736
2737             /* Evaluate switch function */
2738             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2739             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv11,_mm256_mul_ps(velec,dsw)) );
2740             cutoff_mask      = _mm256_cmp_ps(rsq11,rcutoff2,_CMP_LT_OQ);
2741
2742             fscal            = felec;
2743
2744             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2745
2746             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
2747
2748             /* Calculate temporary vectorial force */
2749             tx               = _mm256_mul_ps(fscal,dx11);
2750             ty               = _mm256_mul_ps(fscal,dy11);
2751             tz               = _mm256_mul_ps(fscal,dz11);
2752
2753             /* Update vectorial force */
2754             fix1             = _mm256_add_ps(fix1,tx);
2755             fiy1             = _mm256_add_ps(fiy1,ty);
2756             fiz1             = _mm256_add_ps(fiz1,tz);
2757
2758             fjx1             = _mm256_add_ps(fjx1,tx);
2759             fjy1             = _mm256_add_ps(fjy1,ty);
2760             fjz1             = _mm256_add_ps(fjz1,tz);
2761
2762             }
2763
2764             /**************************
2765              * CALCULATE INTERACTIONS *
2766              **************************/
2767
2768             if (gmx_mm256_any_lt(rsq12,rcutoff2))
2769             {
2770
2771             r12              = _mm256_mul_ps(rsq12,rinv12);
2772             r12              = _mm256_andnot_ps(dummy_mask,r12);
2773
2774             /* EWALD ELECTROSTATICS */
2775
2776             /* Analytical PME correction */
2777             zeta2            = _mm256_mul_ps(beta2,rsq12);
2778             rinv3            = _mm256_mul_ps(rinvsq12,rinv12);
2779             pmecorrF         = avx256_pmecorrF_f(zeta2);
2780             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2781             felec            = _mm256_mul_ps(qq12,felec);
2782             pmecorrV         = avx256_pmecorrV_f(zeta2);
2783             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2784             velec            = _mm256_sub_ps(rinv12,pmecorrV);
2785             velec            = _mm256_mul_ps(qq12,velec);
2786
2787             d                = _mm256_sub_ps(r12,rswitch);
2788             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2789             d2               = _mm256_mul_ps(d,d);
2790             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)))))));
2791
2792             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2793
2794             /* Evaluate switch function */
2795             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2796             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv12,_mm256_mul_ps(velec,dsw)) );
2797             cutoff_mask      = _mm256_cmp_ps(rsq12,rcutoff2,_CMP_LT_OQ);
2798
2799             fscal            = felec;
2800
2801             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2802
2803             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
2804
2805             /* Calculate temporary vectorial force */
2806             tx               = _mm256_mul_ps(fscal,dx12);
2807             ty               = _mm256_mul_ps(fscal,dy12);
2808             tz               = _mm256_mul_ps(fscal,dz12);
2809
2810             /* Update vectorial force */
2811             fix1             = _mm256_add_ps(fix1,tx);
2812             fiy1             = _mm256_add_ps(fiy1,ty);
2813             fiz1             = _mm256_add_ps(fiz1,tz);
2814
2815             fjx2             = _mm256_add_ps(fjx2,tx);
2816             fjy2             = _mm256_add_ps(fjy2,ty);
2817             fjz2             = _mm256_add_ps(fjz2,tz);
2818
2819             }
2820
2821             /**************************
2822              * CALCULATE INTERACTIONS *
2823              **************************/
2824
2825             if (gmx_mm256_any_lt(rsq20,rcutoff2))
2826             {
2827
2828             r20              = _mm256_mul_ps(rsq20,rinv20);
2829             r20              = _mm256_andnot_ps(dummy_mask,r20);
2830
2831             /* EWALD ELECTROSTATICS */
2832
2833             /* Analytical PME correction */
2834             zeta2            = _mm256_mul_ps(beta2,rsq20);
2835             rinv3            = _mm256_mul_ps(rinvsq20,rinv20);
2836             pmecorrF         = avx256_pmecorrF_f(zeta2);
2837             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2838             felec            = _mm256_mul_ps(qq20,felec);
2839             pmecorrV         = avx256_pmecorrV_f(zeta2);
2840             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2841             velec            = _mm256_sub_ps(rinv20,pmecorrV);
2842             velec            = _mm256_mul_ps(qq20,velec);
2843
2844             d                = _mm256_sub_ps(r20,rswitch);
2845             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2846             d2               = _mm256_mul_ps(d,d);
2847             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)))))));
2848
2849             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2850
2851             /* Evaluate switch function */
2852             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2853             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv20,_mm256_mul_ps(velec,dsw)) );
2854             cutoff_mask      = _mm256_cmp_ps(rsq20,rcutoff2,_CMP_LT_OQ);
2855
2856             fscal            = felec;
2857
2858             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2859
2860             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
2861
2862             /* Calculate temporary vectorial force */
2863             tx               = _mm256_mul_ps(fscal,dx20);
2864             ty               = _mm256_mul_ps(fscal,dy20);
2865             tz               = _mm256_mul_ps(fscal,dz20);
2866
2867             /* Update vectorial force */
2868             fix2             = _mm256_add_ps(fix2,tx);
2869             fiy2             = _mm256_add_ps(fiy2,ty);
2870             fiz2             = _mm256_add_ps(fiz2,tz);
2871
2872             fjx0             = _mm256_add_ps(fjx0,tx);
2873             fjy0             = _mm256_add_ps(fjy0,ty);
2874             fjz0             = _mm256_add_ps(fjz0,tz);
2875
2876             }
2877
2878             /**************************
2879              * CALCULATE INTERACTIONS *
2880              **************************/
2881
2882             if (gmx_mm256_any_lt(rsq21,rcutoff2))
2883             {
2884
2885             r21              = _mm256_mul_ps(rsq21,rinv21);
2886             r21              = _mm256_andnot_ps(dummy_mask,r21);
2887
2888             /* EWALD ELECTROSTATICS */
2889
2890             /* Analytical PME correction */
2891             zeta2            = _mm256_mul_ps(beta2,rsq21);
2892             rinv3            = _mm256_mul_ps(rinvsq21,rinv21);
2893             pmecorrF         = avx256_pmecorrF_f(zeta2);
2894             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2895             felec            = _mm256_mul_ps(qq21,felec);
2896             pmecorrV         = avx256_pmecorrV_f(zeta2);
2897             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2898             velec            = _mm256_sub_ps(rinv21,pmecorrV);
2899             velec            = _mm256_mul_ps(qq21,velec);
2900
2901             d                = _mm256_sub_ps(r21,rswitch);
2902             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2903             d2               = _mm256_mul_ps(d,d);
2904             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)))))));
2905
2906             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2907
2908             /* Evaluate switch function */
2909             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2910             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv21,_mm256_mul_ps(velec,dsw)) );
2911             cutoff_mask      = _mm256_cmp_ps(rsq21,rcutoff2,_CMP_LT_OQ);
2912
2913             fscal            = felec;
2914
2915             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2916
2917             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
2918
2919             /* Calculate temporary vectorial force */
2920             tx               = _mm256_mul_ps(fscal,dx21);
2921             ty               = _mm256_mul_ps(fscal,dy21);
2922             tz               = _mm256_mul_ps(fscal,dz21);
2923
2924             /* Update vectorial force */
2925             fix2             = _mm256_add_ps(fix2,tx);
2926             fiy2             = _mm256_add_ps(fiy2,ty);
2927             fiz2             = _mm256_add_ps(fiz2,tz);
2928
2929             fjx1             = _mm256_add_ps(fjx1,tx);
2930             fjy1             = _mm256_add_ps(fjy1,ty);
2931             fjz1             = _mm256_add_ps(fjz1,tz);
2932
2933             }
2934
2935             /**************************
2936              * CALCULATE INTERACTIONS *
2937              **************************/
2938
2939             if (gmx_mm256_any_lt(rsq22,rcutoff2))
2940             {
2941
2942             r22              = _mm256_mul_ps(rsq22,rinv22);
2943             r22              = _mm256_andnot_ps(dummy_mask,r22);
2944
2945             /* EWALD ELECTROSTATICS */
2946
2947             /* Analytical PME correction */
2948             zeta2            = _mm256_mul_ps(beta2,rsq22);
2949             rinv3            = _mm256_mul_ps(rinvsq22,rinv22);
2950             pmecorrF         = avx256_pmecorrF_f(zeta2);
2951             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
2952             felec            = _mm256_mul_ps(qq22,felec);
2953             pmecorrV         = avx256_pmecorrV_f(zeta2);
2954             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
2955             velec            = _mm256_sub_ps(rinv22,pmecorrV);
2956             velec            = _mm256_mul_ps(qq22,velec);
2957
2958             d                = _mm256_sub_ps(r22,rswitch);
2959             d                = _mm256_max_ps(d,_mm256_setzero_ps());
2960             d2               = _mm256_mul_ps(d,d);
2961             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)))))));
2962
2963             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
2964
2965             /* Evaluate switch function */
2966             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
2967             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv22,_mm256_mul_ps(velec,dsw)) );
2968             cutoff_mask      = _mm256_cmp_ps(rsq22,rcutoff2,_CMP_LT_OQ);
2969
2970             fscal            = felec;
2971
2972             fscal            = _mm256_and_ps(fscal,cutoff_mask);
2973
2974             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
2975
2976             /* Calculate temporary vectorial force */
2977             tx               = _mm256_mul_ps(fscal,dx22);
2978             ty               = _mm256_mul_ps(fscal,dy22);
2979             tz               = _mm256_mul_ps(fscal,dz22);
2980
2981             /* Update vectorial force */
2982             fix2             = _mm256_add_ps(fix2,tx);
2983             fiy2             = _mm256_add_ps(fiy2,ty);
2984             fiz2             = _mm256_add_ps(fiz2,tz);
2985
2986             fjx2             = _mm256_add_ps(fjx2,tx);
2987             fjy2             = _mm256_add_ps(fjy2,ty);
2988             fjz2             = _mm256_add_ps(fjz2,tz);
2989
2990             }
2991
2992             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
2993             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
2994             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
2995             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
2996             fjptrE             = (jnrlistE>=0) ? f+j_coord_offsetE : scratch;
2997             fjptrF             = (jnrlistF>=0) ? f+j_coord_offsetF : scratch;
2998             fjptrG             = (jnrlistG>=0) ? f+j_coord_offsetG : scratch;
2999             fjptrH             = (jnrlistH>=0) ? f+j_coord_offsetH : scratch;
3000
3001             gmx_mm256_decrement_3rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,
3002                                                       fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2);
3003
3004             /* Inner loop uses 954 flops */
3005         }
3006
3007         /* End of innermost loop */
3008
3009         gmx_mm256_update_iforce_3atom_swizzle_ps(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,
3010                                                  f+i_coord_offset,fshift+i_shift_offset);
3011
3012         /* Increment number of inner iterations */
3013         inneriter                  += j_index_end - j_index_start;
3014
3015         /* Outer loop uses 18 flops */
3016     }
3017
3018     /* Increment number of outer iterations */
3019     outeriter        += nri;
3020
3021     /* Update outer/inner flops */
3022
3023     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_W3W3_F,outeriter*18 + inneriter*954);
3024 }