src/gromacs/gmxlib/nonbonded/nb_kernel_avx_256_double/nb_kernel_ElecRFCut_VdwLJSw_GeomW3W3_avx_256_double.c

   1 /*
   2  * This file is part of the GROMACS molecular simulation package.
   3  *
   4  * Copyright (c) 2012,2013,2014,2015,2017, 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_double 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_double.h"
  48
  49 /*
  50  * Gromacs nonbonded kernel:   nb_kernel_ElecRFCut_VdwLJSw_GeomW3W3_VF_avx_256_double
  51  * Electrostatics interaction: ReactionField
  52  * VdW interaction:            LennardJones
  53  * Geometry:                   Water3-Water3
  54  * Calculate force/pot:        PotentialAndForce
  55  */
  56 void
  57 nb_kernel_ElecRFCut_VdwLJSw_GeomW3W3_VF_avx_256_double
  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 refer to j loop unrolling done with AVX, e.g. for the four 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              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
  75     int              jnrlistE,jnrlistF,jnrlistG,jnrlistH;
  76     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
  77     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  78     real             rcutoff_scalar;
  79     real             *shiftvec,*fshift,*x,*f;
  80     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD;
  81     real             scratch[4*DIM];
  82     __m256d          tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
  83     real *           vdwioffsetptr0;
  84     __m256d          ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  85     real *           vdwioffsetptr1;
  86     __m256d          ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
  87     real *           vdwioffsetptr2;
  88     __m256d          ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
  89     int              vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D;
  90     __m256d          jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  91     int              vdwjidx1A,vdwjidx1B,vdwjidx1C,vdwjidx1D;
  92     __m256d          jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
  93     int              vdwjidx2A,vdwjidx2B,vdwjidx2C,vdwjidx2D;
  94     __m256d          jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
  95     __m256d          dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
  96     __m256d          dx01,dy01,dz01,rsq01,rinv01,rinvsq01,r01,qq01,c6_01,c12_01;
  97     __m256d          dx02,dy02,dz02,rsq02,rinv02,rinvsq02,r02,qq02,c6_02,c12_02;
  98     __m256d          dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
  99     __m256d          dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11;
 100     __m256d          dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12;
 101     __m256d          dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
 102     __m256d          dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21;
 103     __m256d          dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22;
 104     __m256d          velec,felec,velecsum,facel,crf,krf,krf2;
 105     real             *charge;
 106     int              nvdwtype;
 107     __m256d          rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
 108     int              *vdwtype;
 109     real             *vdwparam;
 110     __m256d          one_sixth   = _mm256_set1_pd(1.0/6.0);
 111     __m256d          one_twelfth = _mm256_set1_pd(1.0/12.0);
 112     __m256d          rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
 113     real             rswitch_scalar,d_scalar;
 114     __m256d          dummy_mask,cutoff_mask;
 115     __m128           tmpmask0,tmpmask1;
 116     __m256d          signbit = _mm256_castsi256_pd( _mm256_set1_epi32(0x80000000) );
 117     __m256d          one     = _mm256_set1_pd(1.0);
 118     __m256d          two     = _mm256_set1_pd(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_pd(fr->ic->epsfac);
 131     charge           = mdatoms->chargeA;
 132     krf              = _mm256_set1_pd(fr->ic->k_rf);
 133     krf2             = _mm256_set1_pd(fr->ic->k_rf*2.0);
 134     crf              = _mm256_set1_pd(fr->ic->c_rf);
 135     nvdwtype         = fr->ntype;
 136     vdwparam         = fr->nbfp;
 137     vdwtype          = mdatoms->typeA;
 138
 139     /* Setup water-specific parameters */
 140     inr              = nlist->iinr[0];
 141     iq0              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+0]));
 142     iq1              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+1]));
 143     iq2              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+2]));
 144     vdwioffsetptr0   = vdwparam+2*nvdwtype*vdwtype[inr+0];
 145
 146     jq0              = _mm256_set1_pd(charge[inr+0]);
 147     jq1              = _mm256_set1_pd(charge[inr+1]);
 148     jq2              = _mm256_set1_pd(charge[inr+2]);
 149     vdwjidx0A        = 2*vdwtype[inr+0];
 150     qq00             = _mm256_mul_pd(iq0,jq0);
 151     c6_00            = _mm256_set1_pd(vdwioffsetptr0[vdwjidx0A]);
 152     c12_00           = _mm256_set1_pd(vdwioffsetptr0[vdwjidx0A+1]);
 153     qq01             = _mm256_mul_pd(iq0,jq1);
 154     qq02             = _mm256_mul_pd(iq0,jq2);
 155     qq10             = _mm256_mul_pd(iq1,jq0);
 156     qq11             = _mm256_mul_pd(iq1,jq1);
 157     qq12             = _mm256_mul_pd(iq1,jq2);
 158     qq20             = _mm256_mul_pd(iq2,jq0);
 159     qq21             = _mm256_mul_pd(iq2,jq1);
 160     qq22             = _mm256_mul_pd(iq2,jq2);
 161
 162     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 163     rcutoff_scalar   = fr->ic->rcoulomb;
 164     rcutoff          = _mm256_set1_pd(rcutoff_scalar);
 165     rcutoff2         = _mm256_mul_pd(rcutoff,rcutoff);
 166
 167     rswitch_scalar   = fr->ic->rvdw_switch;
 168     rswitch          = _mm256_set1_pd(rswitch_scalar);
 169     /* Setup switch parameters */
 170     d_scalar         = rcutoff_scalar-rswitch_scalar;
 171     d                = _mm256_set1_pd(d_scalar);
 172     swV3             = _mm256_set1_pd(-10.0/(d_scalar*d_scalar*d_scalar));
 173     swV4             = _mm256_set1_pd( 15.0/(d_scalar*d_scalar*d_scalar*d_scalar));
 174     swV5             = _mm256_set1_pd( -6.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
 175     swF2             = _mm256_set1_pd(-30.0/(d_scalar*d_scalar*d_scalar));
 176     swF3             = _mm256_set1_pd( 60.0/(d_scalar*d_scalar*d_scalar*d_scalar));
 177     swF4             = _mm256_set1_pd(-30.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
 178
 179     /* Avoid stupid compiler warnings */
 180     jnrA = jnrB = jnrC = jnrD = 0;
 181     j_coord_offsetA = 0;
 182     j_coord_offsetB = 0;
 183     j_coord_offsetC = 0;
 184     j_coord_offsetD = 0;
 185
 186     outeriter        = 0;
 187     inneriter        = 0;
 188
 189     for(iidx=0;iidx<4*DIM;iidx++)
 190     {
 191         scratch[iidx] = 0.0;
 192     }
 193
 194     /* Start outer loop over neighborlists */
 195     for(iidx=0; iidx<nri; iidx++)
 196     {
 197         /* Load shift vector for this list */
 198         i_shift_offset   = DIM*shiftidx[iidx];
 199
 200         /* Load limits for loop over neighbors */
 201         j_index_start    = jindex[iidx];
 202         j_index_end      = jindex[iidx+1];
 203
 204         /* Get outer coordinate index */
 205         inr              = iinr[iidx];
 206         i_coord_offset   = DIM*inr;
 207
 208         /* Load i particle coords and add shift vector */
 209         gmx_mm256_load_shift_and_3rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset,
 210                                                     &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2);
 211
 212         fix0             = _mm256_setzero_pd();
 213         fiy0             = _mm256_setzero_pd();
 214         fiz0             = _mm256_setzero_pd();
 215         fix1             = _mm256_setzero_pd();
 216         fiy1             = _mm256_setzero_pd();
 217         fiz1             = _mm256_setzero_pd();
 218         fix2             = _mm256_setzero_pd();
 219         fiy2             = _mm256_setzero_pd();
 220         fiz2             = _mm256_setzero_pd();
 221
 222         /* Reset potential sums */
 223         velecsum         = _mm256_setzero_pd();
 224         vvdwsum          = _mm256_setzero_pd();
 225
 226         /* Start inner kernel loop */
 227         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4)
 228         {
 229
 230             /* Get j neighbor index, and coordinate index */
 231             jnrA             = jjnr[jidx];
 232             jnrB             = jjnr[jidx+1];
 233             jnrC             = jjnr[jidx+2];
 234             jnrD             = jjnr[jidx+3];
 235             j_coord_offsetA  = DIM*jnrA;
 236             j_coord_offsetB  = DIM*jnrB;
 237             j_coord_offsetC  = DIM*jnrC;
 238             j_coord_offsetD  = DIM*jnrD;
 239
 240             /* load j atom coordinates */
 241             gmx_mm256_load_3rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
 242                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 243                                               &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,&jy2,&jz2);
 244
 245             /* Calculate displacement vector */
 246             dx00             = _mm256_sub_pd(ix0,jx0);
 247             dy00             = _mm256_sub_pd(iy0,jy0);
 248             dz00             = _mm256_sub_pd(iz0,jz0);
 249             dx01             = _mm256_sub_pd(ix0,jx1);
 250             dy01             = _mm256_sub_pd(iy0,jy1);
 251             dz01             = _mm256_sub_pd(iz0,jz1);
 252             dx02             = _mm256_sub_pd(ix0,jx2);
 253             dy02             = _mm256_sub_pd(iy0,jy2);
 254             dz02             = _mm256_sub_pd(iz0,jz2);
 255             dx10             = _mm256_sub_pd(ix1,jx0);
 256             dy10             = _mm256_sub_pd(iy1,jy0);
 257             dz10             = _mm256_sub_pd(iz1,jz0);
 258             dx11             = _mm256_sub_pd(ix1,jx1);
 259             dy11             = _mm256_sub_pd(iy1,jy1);
 260             dz11             = _mm256_sub_pd(iz1,jz1);
 261             dx12             = _mm256_sub_pd(ix1,jx2);
 262             dy12             = _mm256_sub_pd(iy1,jy2);
 263             dz12             = _mm256_sub_pd(iz1,jz2);
 264             dx20             = _mm256_sub_pd(ix2,jx0);
 265             dy20             = _mm256_sub_pd(iy2,jy0);
 266             dz20             = _mm256_sub_pd(iz2,jz0);
 267             dx21             = _mm256_sub_pd(ix2,jx1);
 268             dy21             = _mm256_sub_pd(iy2,jy1);
 269             dz21             = _mm256_sub_pd(iz2,jz1);
 270             dx22             = _mm256_sub_pd(ix2,jx2);
 271             dy22             = _mm256_sub_pd(iy2,jy2);
 272             dz22             = _mm256_sub_pd(iz2,jz2);
 273
 274             /* Calculate squared distance and things based on it */
 275             rsq00            = gmx_mm256_calc_rsq_pd(dx00,dy00,dz00);
 276             rsq01            = gmx_mm256_calc_rsq_pd(dx01,dy01,dz01);
 277             rsq02            = gmx_mm256_calc_rsq_pd(dx02,dy02,dz02);
 278             rsq10            = gmx_mm256_calc_rsq_pd(dx10,dy10,dz10);
 279             rsq11            = gmx_mm256_calc_rsq_pd(dx11,dy11,dz11);
 280             rsq12            = gmx_mm256_calc_rsq_pd(dx12,dy12,dz12);
 281             rsq20            = gmx_mm256_calc_rsq_pd(dx20,dy20,dz20);
 282             rsq21            = gmx_mm256_calc_rsq_pd(dx21,dy21,dz21);
 283             rsq22            = gmx_mm256_calc_rsq_pd(dx22,dy22,dz22);
 284
 285             rinv00           = avx256_invsqrt_d(rsq00);
 286             rinv01           = avx256_invsqrt_d(rsq01);
 287             rinv02           = avx256_invsqrt_d(rsq02);
 288             rinv10           = avx256_invsqrt_d(rsq10);
 289             rinv11           = avx256_invsqrt_d(rsq11);
 290             rinv12           = avx256_invsqrt_d(rsq12);
 291             rinv20           = avx256_invsqrt_d(rsq20);
 292             rinv21           = avx256_invsqrt_d(rsq21);
 293             rinv22           = avx256_invsqrt_d(rsq22);
 294
 295             rinvsq00         = _mm256_mul_pd(rinv00,rinv00);
 296             rinvsq01         = _mm256_mul_pd(rinv01,rinv01);
 297             rinvsq02         = _mm256_mul_pd(rinv02,rinv02);
 298             rinvsq10         = _mm256_mul_pd(rinv10,rinv10);
 299             rinvsq11         = _mm256_mul_pd(rinv11,rinv11);
 300             rinvsq12         = _mm256_mul_pd(rinv12,rinv12);
 301             rinvsq20         = _mm256_mul_pd(rinv20,rinv20);
 302             rinvsq21         = _mm256_mul_pd(rinv21,rinv21);
 303             rinvsq22         = _mm256_mul_pd(rinv22,rinv22);
 304
 305             fjx0             = _mm256_setzero_pd();
 306             fjy0             = _mm256_setzero_pd();
 307             fjz0             = _mm256_setzero_pd();
 308             fjx1             = _mm256_setzero_pd();
 309             fjy1             = _mm256_setzero_pd();
 310             fjz1             = _mm256_setzero_pd();
 311             fjx2             = _mm256_setzero_pd();
 312             fjy2             = _mm256_setzero_pd();
 313             fjz2             = _mm256_setzero_pd();
 314
 315             /**************************
 316              * CALCULATE INTERACTIONS *
 317              **************************/
 318
 319             if (gmx_mm256_any_lt(rsq00,rcutoff2))
 320             {
 321
 322             r00              = _mm256_mul_pd(rsq00,rinv00);
 323
 324             /* REACTION-FIELD ELECTROSTATICS */
 325             velec            = _mm256_mul_pd(qq00,_mm256_sub_pd(_mm256_add_pd(rinv00,_mm256_mul_pd(krf,rsq00)),crf));
 326             felec            = _mm256_mul_pd(qq00,_mm256_sub_pd(_mm256_mul_pd(rinv00,rinvsq00),krf2));
 327
 328             /* LENNARD-JONES DISPERSION/REPULSION */
 329
 330             rinvsix          = _mm256_mul_pd(_mm256_mul_pd(rinvsq00,rinvsq00),rinvsq00);
 331             vvdw6            = _mm256_mul_pd(c6_00,rinvsix);
 332             vvdw12           = _mm256_mul_pd(c12_00,_mm256_mul_pd(rinvsix,rinvsix));
 333             vvdw             = _mm256_sub_pd( _mm256_mul_pd(vvdw12,one_twelfth) , _mm256_mul_pd(vvdw6,one_sixth) );
 334             fvdw             = _mm256_mul_pd(_mm256_sub_pd(vvdw12,vvdw6),rinvsq00);
 335
 336             d                = _mm256_sub_pd(r00,rswitch);
 337             d                = _mm256_max_pd(d,_mm256_setzero_pd());
 338             d2               = _mm256_mul_pd(d,d);
 339             sw               = _mm256_add_pd(one,_mm256_mul_pd(d2,_mm256_mul_pd(d,_mm256_add_pd(swV3,_mm256_mul_pd(d,_mm256_add_pd(swV4,_mm256_mul_pd(d,swV5)))))));
 340
 341             dsw              = _mm256_mul_pd(d2,_mm256_add_pd(swF2,_mm256_mul_pd(d,_mm256_add_pd(swF3,_mm256_mul_pd(d,swF4)))));
 342
 343             /* Evaluate switch function */
 344             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 345             fvdw             = _mm256_sub_pd( _mm256_mul_pd(fvdw,sw) , _mm256_mul_pd(rinv00,_mm256_mul_pd(vvdw,dsw)) );
 346             vvdw             = _mm256_mul_pd(vvdw,sw);
 347             cutoff_mask      = _mm256_cmp_pd(rsq00,rcutoff2,_CMP_LT_OQ);
 348
 349             /* Update potential sum for this i atom from the interaction with this j atom. */
 350             velec            = _mm256_and_pd(velec,cutoff_mask);
 351             velecsum         = _mm256_add_pd(velecsum,velec);
 352             vvdw             = _mm256_and_pd(vvdw,cutoff_mask);
 353             vvdwsum          = _mm256_add_pd(vvdwsum,vvdw);
 354
 355             fscal            = _mm256_add_pd(felec,fvdw);
 356
 357             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 358
 359             /* Calculate temporary vectorial force */
 360             tx               = _mm256_mul_pd(fscal,dx00);
 361             ty               = _mm256_mul_pd(fscal,dy00);
 362             tz               = _mm256_mul_pd(fscal,dz00);
 363
 364             /* Update vectorial force */
 365             fix0             = _mm256_add_pd(fix0,tx);
 366             fiy0             = _mm256_add_pd(fiy0,ty);
 367             fiz0             = _mm256_add_pd(fiz0,tz);
 368
 369             fjx0             = _mm256_add_pd(fjx0,tx);
 370             fjy0             = _mm256_add_pd(fjy0,ty);
 371             fjz0             = _mm256_add_pd(fjz0,tz);
 372
 373             }
 374
 375             /**************************
 376              * CALCULATE INTERACTIONS *
 377              **************************/
 378
 379             if (gmx_mm256_any_lt(rsq01,rcutoff2))
 380             {
 381
 382             /* REACTION-FIELD ELECTROSTATICS */
 383             velec            = _mm256_mul_pd(qq01,_mm256_sub_pd(_mm256_add_pd(rinv01,_mm256_mul_pd(krf,rsq01)),crf));
 384             felec            = _mm256_mul_pd(qq01,_mm256_sub_pd(_mm256_mul_pd(rinv01,rinvsq01),krf2));
 385
 386             cutoff_mask      = _mm256_cmp_pd(rsq01,rcutoff2,_CMP_LT_OQ);
 387
 388             /* Update potential sum for this i atom from the interaction with this j atom. */
 389             velec            = _mm256_and_pd(velec,cutoff_mask);
 390             velecsum         = _mm256_add_pd(velecsum,velec);
 391
 392             fscal            = felec;
 393
 394             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 395
 396             /* Calculate temporary vectorial force */
 397             tx               = _mm256_mul_pd(fscal,dx01);
 398             ty               = _mm256_mul_pd(fscal,dy01);
 399             tz               = _mm256_mul_pd(fscal,dz01);
 400
 401             /* Update vectorial force */
 402             fix0             = _mm256_add_pd(fix0,tx);
 403             fiy0             = _mm256_add_pd(fiy0,ty);
 404             fiz0             = _mm256_add_pd(fiz0,tz);
 405
 406             fjx1             = _mm256_add_pd(fjx1,tx);
 407             fjy1             = _mm256_add_pd(fjy1,ty);
 408             fjz1             = _mm256_add_pd(fjz1,tz);
 409
 410             }
 411
 412             /**************************
 413              * CALCULATE INTERACTIONS *
 414              **************************/
 415
 416             if (gmx_mm256_any_lt(rsq02,rcutoff2))
 417             {
 418
 419             /* REACTION-FIELD ELECTROSTATICS */
 420             velec            = _mm256_mul_pd(qq02,_mm256_sub_pd(_mm256_add_pd(rinv02,_mm256_mul_pd(krf,rsq02)),crf));
 421             felec            = _mm256_mul_pd(qq02,_mm256_sub_pd(_mm256_mul_pd(rinv02,rinvsq02),krf2));
 422
 423             cutoff_mask      = _mm256_cmp_pd(rsq02,rcutoff2,_CMP_LT_OQ);
 424
 425             /* Update potential sum for this i atom from the interaction with this j atom. */
 426             velec            = _mm256_and_pd(velec,cutoff_mask);
 427             velecsum         = _mm256_add_pd(velecsum,velec);
 428
 429             fscal            = felec;
 430
 431             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 432
 433             /* Calculate temporary vectorial force */
 434             tx               = _mm256_mul_pd(fscal,dx02);
 435             ty               = _mm256_mul_pd(fscal,dy02);
 436             tz               = _mm256_mul_pd(fscal,dz02);
 437
 438             /* Update vectorial force */
 439             fix0             = _mm256_add_pd(fix0,tx);
 440             fiy0             = _mm256_add_pd(fiy0,ty);
 441             fiz0             = _mm256_add_pd(fiz0,tz);
 442
 443             fjx2             = _mm256_add_pd(fjx2,tx);
 444             fjy2             = _mm256_add_pd(fjy2,ty);
 445             fjz2             = _mm256_add_pd(fjz2,tz);
 446
 447             }
 448
 449             /**************************
 450              * CALCULATE INTERACTIONS *
 451              **************************/
 452
 453             if (gmx_mm256_any_lt(rsq10,rcutoff2))
 454             {
 455
 456             /* REACTION-FIELD ELECTROSTATICS */
 457             velec            = _mm256_mul_pd(qq10,_mm256_sub_pd(_mm256_add_pd(rinv10,_mm256_mul_pd(krf,rsq10)),crf));
 458             felec            = _mm256_mul_pd(qq10,_mm256_sub_pd(_mm256_mul_pd(rinv10,rinvsq10),krf2));
 459
 460             cutoff_mask      = _mm256_cmp_pd(rsq10,rcutoff2,_CMP_LT_OQ);
 461
 462             /* Update potential sum for this i atom from the interaction with this j atom. */
 463             velec            = _mm256_and_pd(velec,cutoff_mask);
 464             velecsum         = _mm256_add_pd(velecsum,velec);
 465
 466             fscal            = felec;
 467
 468             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 469
 470             /* Calculate temporary vectorial force */
 471             tx               = _mm256_mul_pd(fscal,dx10);
 472             ty               = _mm256_mul_pd(fscal,dy10);
 473             tz               = _mm256_mul_pd(fscal,dz10);
 474
 475             /* Update vectorial force */
 476             fix1             = _mm256_add_pd(fix1,tx);
 477             fiy1             = _mm256_add_pd(fiy1,ty);
 478             fiz1             = _mm256_add_pd(fiz1,tz);
 479
 480             fjx0             = _mm256_add_pd(fjx0,tx);
 481             fjy0             = _mm256_add_pd(fjy0,ty);
 482             fjz0             = _mm256_add_pd(fjz0,tz);
 483
 484             }
 485
 486             /**************************
 487              * CALCULATE INTERACTIONS *
 488              **************************/
 489
 490             if (gmx_mm256_any_lt(rsq11,rcutoff2))
 491             {
 492
 493             /* REACTION-FIELD ELECTROSTATICS */
 494             velec            = _mm256_mul_pd(qq11,_mm256_sub_pd(_mm256_add_pd(rinv11,_mm256_mul_pd(krf,rsq11)),crf));
 495             felec            = _mm256_mul_pd(qq11,_mm256_sub_pd(_mm256_mul_pd(rinv11,rinvsq11),krf2));
 496
 497             cutoff_mask      = _mm256_cmp_pd(rsq11,rcutoff2,_CMP_LT_OQ);
 498
 499             /* Update potential sum for this i atom from the interaction with this j atom. */
 500             velec            = _mm256_and_pd(velec,cutoff_mask);
 501             velecsum         = _mm256_add_pd(velecsum,velec);
 502
 503             fscal            = felec;
 504
 505             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 506
 507             /* Calculate temporary vectorial force */
 508             tx               = _mm256_mul_pd(fscal,dx11);
 509             ty               = _mm256_mul_pd(fscal,dy11);
 510             tz               = _mm256_mul_pd(fscal,dz11);
 511
 512             /* Update vectorial force */
 513             fix1             = _mm256_add_pd(fix1,tx);
 514             fiy1             = _mm256_add_pd(fiy1,ty);
 515             fiz1             = _mm256_add_pd(fiz1,tz);
 516
 517             fjx1             = _mm256_add_pd(fjx1,tx);
 518             fjy1             = _mm256_add_pd(fjy1,ty);
 519             fjz1             = _mm256_add_pd(fjz1,tz);
 520
 521             }
 522
 523             /**************************
 524              * CALCULATE INTERACTIONS *
 525              **************************/
 526
 527             if (gmx_mm256_any_lt(rsq12,rcutoff2))
 528             {
 529
 530             /* REACTION-FIELD ELECTROSTATICS */
 531             velec            = _mm256_mul_pd(qq12,_mm256_sub_pd(_mm256_add_pd(rinv12,_mm256_mul_pd(krf,rsq12)),crf));
 532             felec            = _mm256_mul_pd(qq12,_mm256_sub_pd(_mm256_mul_pd(rinv12,rinvsq12),krf2));
 533
 534             cutoff_mask      = _mm256_cmp_pd(rsq12,rcutoff2,_CMP_LT_OQ);
 535
 536             /* Update potential sum for this i atom from the interaction with this j atom. */
 537             velec            = _mm256_and_pd(velec,cutoff_mask);
 538             velecsum         = _mm256_add_pd(velecsum,velec);
 539
 540             fscal            = felec;
 541
 542             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 543
 544             /* Calculate temporary vectorial force */
 545             tx               = _mm256_mul_pd(fscal,dx12);
 546             ty               = _mm256_mul_pd(fscal,dy12);
 547             tz               = _mm256_mul_pd(fscal,dz12);
 548
 549             /* Update vectorial force */
 550             fix1             = _mm256_add_pd(fix1,tx);
 551             fiy1             = _mm256_add_pd(fiy1,ty);
 552             fiz1             = _mm256_add_pd(fiz1,tz);
 553
 554             fjx2             = _mm256_add_pd(fjx2,tx);
 555             fjy2             = _mm256_add_pd(fjy2,ty);
 556             fjz2             = _mm256_add_pd(fjz2,tz);
 557
 558             }
 559
 560             /**************************
 561              * CALCULATE INTERACTIONS *
 562              **************************/
 563
 564             if (gmx_mm256_any_lt(rsq20,rcutoff2))
 565             {
 566
 567             /* REACTION-FIELD ELECTROSTATICS */
 568             velec            = _mm256_mul_pd(qq20,_mm256_sub_pd(_mm256_add_pd(rinv20,_mm256_mul_pd(krf,rsq20)),crf));
 569             felec            = _mm256_mul_pd(qq20,_mm256_sub_pd(_mm256_mul_pd(rinv20,rinvsq20),krf2));
 570
 571             cutoff_mask      = _mm256_cmp_pd(rsq20,rcutoff2,_CMP_LT_OQ);
 572
 573             /* Update potential sum for this i atom from the interaction with this j atom. */
 574             velec            = _mm256_and_pd(velec,cutoff_mask);
 575             velecsum         = _mm256_add_pd(velecsum,velec);
 576
 577             fscal            = felec;
 578
 579             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 580
 581             /* Calculate temporary vectorial force */
 582             tx               = _mm256_mul_pd(fscal,dx20);
 583             ty               = _mm256_mul_pd(fscal,dy20);
 584             tz               = _mm256_mul_pd(fscal,dz20);
 585
 586             /* Update vectorial force */
 587             fix2             = _mm256_add_pd(fix2,tx);
 588             fiy2             = _mm256_add_pd(fiy2,ty);
 589             fiz2             = _mm256_add_pd(fiz2,tz);
 590
 591             fjx0             = _mm256_add_pd(fjx0,tx);
 592             fjy0             = _mm256_add_pd(fjy0,ty);
 593             fjz0             = _mm256_add_pd(fjz0,tz);
 594
 595             }
 596
 597             /**************************
 598              * CALCULATE INTERACTIONS *
 599              **************************/
 600
 601             if (gmx_mm256_any_lt(rsq21,rcutoff2))
 602             {
 603
 604             /* REACTION-FIELD ELECTROSTATICS */
 605             velec            = _mm256_mul_pd(qq21,_mm256_sub_pd(_mm256_add_pd(rinv21,_mm256_mul_pd(krf,rsq21)),crf));
 606             felec            = _mm256_mul_pd(qq21,_mm256_sub_pd(_mm256_mul_pd(rinv21,rinvsq21),krf2));
 607
 608             cutoff_mask      = _mm256_cmp_pd(rsq21,rcutoff2,_CMP_LT_OQ);
 609
 610             /* Update potential sum for this i atom from the interaction with this j atom. */
 611             velec            = _mm256_and_pd(velec,cutoff_mask);
 612             velecsum         = _mm256_add_pd(velecsum,velec);
 613
 614             fscal            = felec;
 615
 616             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 617
 618             /* Calculate temporary vectorial force */
 619             tx               = _mm256_mul_pd(fscal,dx21);
 620             ty               = _mm256_mul_pd(fscal,dy21);
 621             tz               = _mm256_mul_pd(fscal,dz21);
 622
 623             /* Update vectorial force */
 624             fix2             = _mm256_add_pd(fix2,tx);
 625             fiy2             = _mm256_add_pd(fiy2,ty);
 626             fiz2             = _mm256_add_pd(fiz2,tz);
 627
 628             fjx1             = _mm256_add_pd(fjx1,tx);
 629             fjy1             = _mm256_add_pd(fjy1,ty);
 630             fjz1             = _mm256_add_pd(fjz1,tz);
 631
 632             }
 633
 634             /**************************
 635              * CALCULATE INTERACTIONS *
 636              **************************/
 637
 638             if (gmx_mm256_any_lt(rsq22,rcutoff2))
 639             {
 640
 641             /* REACTION-FIELD ELECTROSTATICS */
 642             velec            = _mm256_mul_pd(qq22,_mm256_sub_pd(_mm256_add_pd(rinv22,_mm256_mul_pd(krf,rsq22)),crf));
 643             felec            = _mm256_mul_pd(qq22,_mm256_sub_pd(_mm256_mul_pd(rinv22,rinvsq22),krf2));
 644
 645             cutoff_mask      = _mm256_cmp_pd(rsq22,rcutoff2,_CMP_LT_OQ);
 646
 647             /* Update potential sum for this i atom from the interaction with this j atom. */
 648             velec            = _mm256_and_pd(velec,cutoff_mask);
 649             velecsum         = _mm256_add_pd(velecsum,velec);
 650
 651             fscal            = felec;
 652
 653             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 654
 655             /* Calculate temporary vectorial force */
 656             tx               = _mm256_mul_pd(fscal,dx22);
 657             ty               = _mm256_mul_pd(fscal,dy22);
 658             tz               = _mm256_mul_pd(fscal,dz22);
 659
 660             /* Update vectorial force */
 661             fix2             = _mm256_add_pd(fix2,tx);
 662             fiy2             = _mm256_add_pd(fiy2,ty);
 663             fiz2             = _mm256_add_pd(fiz2,tz);
 664
 665             fjx2             = _mm256_add_pd(fjx2,tx);
 666             fjy2             = _mm256_add_pd(fjy2,ty);
 667             fjz2             = _mm256_add_pd(fjz2,tz);
 668
 669             }
 670
 671             fjptrA             = f+j_coord_offsetA;
 672             fjptrB             = f+j_coord_offsetB;
 673             fjptrC             = f+j_coord_offsetC;
 674             fjptrD             = f+j_coord_offsetD;
 675
 676             gmx_mm256_decrement_3rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,
 677                                                       fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2);
 678
 679             /* Inner loop uses 358 flops */
 680         }
 681
 682         if(jidx<j_index_end)
 683         {
 684
 685             /* Get j neighbor index, and coordinate index */
 686             jnrlistA         = jjnr[jidx];
 687             jnrlistB         = jjnr[jidx+1];
 688             jnrlistC         = jjnr[jidx+2];
 689             jnrlistD         = jjnr[jidx+3];
 690             /* Sign of each element will be negative for non-real atoms.
 691              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
 692              * so use it as val = _mm_andnot_pd(mask,val) to clear dummy entries.
 693              */
 694             tmpmask0 = gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128()));
 695
 696             tmpmask1 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(3,3,2,2));
 697             tmpmask0 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(1,1,0,0));
 698             dummy_mask = _mm256_castps_pd(gmx_mm256_set_m128(tmpmask1,tmpmask0));
 699
 700             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
 701             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
 702             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
 703             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
 704             j_coord_offsetA  = DIM*jnrA;
 705             j_coord_offsetB  = DIM*jnrB;
 706             j_coord_offsetC  = DIM*jnrC;
 707             j_coord_offsetD  = DIM*jnrD;
 708
 709             /* load j atom coordinates */
 710             gmx_mm256_load_3rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
 711                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 712                                               &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,&jy2,&jz2);
 713
 714             /* Calculate displacement vector */
 715             dx00             = _mm256_sub_pd(ix0,jx0);
 716             dy00             = _mm256_sub_pd(iy0,jy0);
 717             dz00             = _mm256_sub_pd(iz0,jz0);
 718             dx01             = _mm256_sub_pd(ix0,jx1);
 719             dy01             = _mm256_sub_pd(iy0,jy1);
 720             dz01             = _mm256_sub_pd(iz0,jz1);
 721             dx02             = _mm256_sub_pd(ix0,jx2);
 722             dy02             = _mm256_sub_pd(iy0,jy2);
 723             dz02             = _mm256_sub_pd(iz0,jz2);
 724             dx10             = _mm256_sub_pd(ix1,jx0);
 725             dy10             = _mm256_sub_pd(iy1,jy0);
 726             dz10             = _mm256_sub_pd(iz1,jz0);
 727             dx11             = _mm256_sub_pd(ix1,jx1);
 728             dy11             = _mm256_sub_pd(iy1,jy1);
 729             dz11             = _mm256_sub_pd(iz1,jz1);
 730             dx12             = _mm256_sub_pd(ix1,jx2);
 731             dy12             = _mm256_sub_pd(iy1,jy2);
 732             dz12             = _mm256_sub_pd(iz1,jz2);
 733             dx20             = _mm256_sub_pd(ix2,jx0);
 734             dy20             = _mm256_sub_pd(iy2,jy0);
 735             dz20             = _mm256_sub_pd(iz2,jz0);
 736             dx21             = _mm256_sub_pd(ix2,jx1);
 737             dy21             = _mm256_sub_pd(iy2,jy1);
 738             dz21             = _mm256_sub_pd(iz2,jz1);
 739             dx22             = _mm256_sub_pd(ix2,jx2);
 740             dy22             = _mm256_sub_pd(iy2,jy2);
 741             dz22             = _mm256_sub_pd(iz2,jz2);
 742
 743             /* Calculate squared distance and things based on it */
 744             rsq00            = gmx_mm256_calc_rsq_pd(dx00,dy00,dz00);
 745             rsq01            = gmx_mm256_calc_rsq_pd(dx01,dy01,dz01);
 746             rsq02            = gmx_mm256_calc_rsq_pd(dx02,dy02,dz02);
 747             rsq10            = gmx_mm256_calc_rsq_pd(dx10,dy10,dz10);
 748             rsq11            = gmx_mm256_calc_rsq_pd(dx11,dy11,dz11);
 749             rsq12            = gmx_mm256_calc_rsq_pd(dx12,dy12,dz12);
 750             rsq20            = gmx_mm256_calc_rsq_pd(dx20,dy20,dz20);
 751             rsq21            = gmx_mm256_calc_rsq_pd(dx21,dy21,dz21);
 752             rsq22            = gmx_mm256_calc_rsq_pd(dx22,dy22,dz22);
 753
 754             rinv00           = avx256_invsqrt_d(rsq00);
 755             rinv01           = avx256_invsqrt_d(rsq01);
 756             rinv02           = avx256_invsqrt_d(rsq02);
 757             rinv10           = avx256_invsqrt_d(rsq10);
 758             rinv11           = avx256_invsqrt_d(rsq11);
 759             rinv12           = avx256_invsqrt_d(rsq12);
 760             rinv20           = avx256_invsqrt_d(rsq20);
 761             rinv21           = avx256_invsqrt_d(rsq21);
 762             rinv22           = avx256_invsqrt_d(rsq22);
 763
 764             rinvsq00         = _mm256_mul_pd(rinv00,rinv00);
 765             rinvsq01         = _mm256_mul_pd(rinv01,rinv01);
 766             rinvsq02         = _mm256_mul_pd(rinv02,rinv02);
 767             rinvsq10         = _mm256_mul_pd(rinv10,rinv10);
 768             rinvsq11         = _mm256_mul_pd(rinv11,rinv11);
 769             rinvsq12         = _mm256_mul_pd(rinv12,rinv12);
 770             rinvsq20         = _mm256_mul_pd(rinv20,rinv20);
 771             rinvsq21         = _mm256_mul_pd(rinv21,rinv21);
 772             rinvsq22         = _mm256_mul_pd(rinv22,rinv22);
 773
 774             fjx0             = _mm256_setzero_pd();
 775             fjy0             = _mm256_setzero_pd();
 776             fjz0             = _mm256_setzero_pd();
 777             fjx1             = _mm256_setzero_pd();
 778             fjy1             = _mm256_setzero_pd();
 779             fjz1             = _mm256_setzero_pd();
 780             fjx2             = _mm256_setzero_pd();
 781             fjy2             = _mm256_setzero_pd();
 782             fjz2             = _mm256_setzero_pd();
 783
 784             /**************************
 785              * CALCULATE INTERACTIONS *
 786              **************************/
 787
 788             if (gmx_mm256_any_lt(rsq00,rcutoff2))
 789             {
 790
 791             r00              = _mm256_mul_pd(rsq00,rinv00);
 792             r00              = _mm256_andnot_pd(dummy_mask,r00);
 793
 794             /* REACTION-FIELD ELECTROSTATICS */
 795             velec            = _mm256_mul_pd(qq00,_mm256_sub_pd(_mm256_add_pd(rinv00,_mm256_mul_pd(krf,rsq00)),crf));
 796             felec            = _mm256_mul_pd(qq00,_mm256_sub_pd(_mm256_mul_pd(rinv00,rinvsq00),krf2));
 797
 798             /* LENNARD-JONES DISPERSION/REPULSION */
 799
 800             rinvsix          = _mm256_mul_pd(_mm256_mul_pd(rinvsq00,rinvsq00),rinvsq00);
 801             vvdw6            = _mm256_mul_pd(c6_00,rinvsix);
 802             vvdw12           = _mm256_mul_pd(c12_00,_mm256_mul_pd(rinvsix,rinvsix));
 803             vvdw             = _mm256_sub_pd( _mm256_mul_pd(vvdw12,one_twelfth) , _mm256_mul_pd(vvdw6,one_sixth) );
 804             fvdw             = _mm256_mul_pd(_mm256_sub_pd(vvdw12,vvdw6),rinvsq00);
 805
 806             d                = _mm256_sub_pd(r00,rswitch);
 807             d                = _mm256_max_pd(d,_mm256_setzero_pd());
 808             d2               = _mm256_mul_pd(d,d);
 809             sw               = _mm256_add_pd(one,_mm256_mul_pd(d2,_mm256_mul_pd(d,_mm256_add_pd(swV3,_mm256_mul_pd(d,_mm256_add_pd(swV4,_mm256_mul_pd(d,swV5)))))));
 810
 811             dsw              = _mm256_mul_pd(d2,_mm256_add_pd(swF2,_mm256_mul_pd(d,_mm256_add_pd(swF3,_mm256_mul_pd(d,swF4)))));
 812
 813             /* Evaluate switch function */
 814             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 815             fvdw             = _mm256_sub_pd( _mm256_mul_pd(fvdw,sw) , _mm256_mul_pd(rinv00,_mm256_mul_pd(vvdw,dsw)) );
 816             vvdw             = _mm256_mul_pd(vvdw,sw);
 817             cutoff_mask      = _mm256_cmp_pd(rsq00,rcutoff2,_CMP_LT_OQ);
 818
 819             /* Update potential sum for this i atom from the interaction with this j atom. */
 820             velec            = _mm256_and_pd(velec,cutoff_mask);
 821             velec            = _mm256_andnot_pd(dummy_mask,velec);
 822             velecsum         = _mm256_add_pd(velecsum,velec);
 823             vvdw             = _mm256_and_pd(vvdw,cutoff_mask);
 824             vvdw             = _mm256_andnot_pd(dummy_mask,vvdw);
 825             vvdwsum          = _mm256_add_pd(vvdwsum,vvdw);
 826
 827             fscal            = _mm256_add_pd(felec,fvdw);
 828
 829             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 830
 831             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
 832
 833             /* Calculate temporary vectorial force */
 834             tx               = _mm256_mul_pd(fscal,dx00);
 835             ty               = _mm256_mul_pd(fscal,dy00);
 836             tz               = _mm256_mul_pd(fscal,dz00);
 837
 838             /* Update vectorial force */
 839             fix0             = _mm256_add_pd(fix0,tx);
 840             fiy0             = _mm256_add_pd(fiy0,ty);
 841             fiz0             = _mm256_add_pd(fiz0,tz);
 842
 843             fjx0             = _mm256_add_pd(fjx0,tx);
 844             fjy0             = _mm256_add_pd(fjy0,ty);
 845             fjz0             = _mm256_add_pd(fjz0,tz);
 846
 847             }
 848
 849             /**************************
 850              * CALCULATE INTERACTIONS *
 851              **************************/
 852
 853             if (gmx_mm256_any_lt(rsq01,rcutoff2))
 854             {
 855
 856             /* REACTION-FIELD ELECTROSTATICS */
 857             velec            = _mm256_mul_pd(qq01,_mm256_sub_pd(_mm256_add_pd(rinv01,_mm256_mul_pd(krf,rsq01)),crf));
 858             felec            = _mm256_mul_pd(qq01,_mm256_sub_pd(_mm256_mul_pd(rinv01,rinvsq01),krf2));
 859
 860             cutoff_mask      = _mm256_cmp_pd(rsq01,rcutoff2,_CMP_LT_OQ);
 861
 862             /* Update potential sum for this i atom from the interaction with this j atom. */
 863             velec            = _mm256_and_pd(velec,cutoff_mask);
 864             velec            = _mm256_andnot_pd(dummy_mask,velec);
 865             velecsum         = _mm256_add_pd(velecsum,velec);
 866
 867             fscal            = felec;
 868
 869             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 870
 871             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
 872
 873             /* Calculate temporary vectorial force */
 874             tx               = _mm256_mul_pd(fscal,dx01);
 875             ty               = _mm256_mul_pd(fscal,dy01);
 876             tz               = _mm256_mul_pd(fscal,dz01);
 877
 878             /* Update vectorial force */
 879             fix0             = _mm256_add_pd(fix0,tx);
 880             fiy0             = _mm256_add_pd(fiy0,ty);
 881             fiz0             = _mm256_add_pd(fiz0,tz);
 882
 883             fjx1             = _mm256_add_pd(fjx1,tx);
 884             fjy1             = _mm256_add_pd(fjy1,ty);
 885             fjz1             = _mm256_add_pd(fjz1,tz);
 886
 887             }
 888
 889             /**************************
 890              * CALCULATE INTERACTIONS *
 891              **************************/
 892
 893             if (gmx_mm256_any_lt(rsq02,rcutoff2))
 894             {
 895
 896             /* REACTION-FIELD ELECTROSTATICS */
 897             velec            = _mm256_mul_pd(qq02,_mm256_sub_pd(_mm256_add_pd(rinv02,_mm256_mul_pd(krf,rsq02)),crf));
 898             felec            = _mm256_mul_pd(qq02,_mm256_sub_pd(_mm256_mul_pd(rinv02,rinvsq02),krf2));
 899
 900             cutoff_mask      = _mm256_cmp_pd(rsq02,rcutoff2,_CMP_LT_OQ);
 901
 902             /* Update potential sum for this i atom from the interaction with this j atom. */
 903             velec            = _mm256_and_pd(velec,cutoff_mask);
 904             velec            = _mm256_andnot_pd(dummy_mask,velec);
 905             velecsum         = _mm256_add_pd(velecsum,velec);
 906
 907             fscal            = felec;
 908
 909             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 910
 911             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
 912
 913             /* Calculate temporary vectorial force */
 914             tx               = _mm256_mul_pd(fscal,dx02);
 915             ty               = _mm256_mul_pd(fscal,dy02);
 916             tz               = _mm256_mul_pd(fscal,dz02);
 917
 918             /* Update vectorial force */
 919             fix0             = _mm256_add_pd(fix0,tx);
 920             fiy0             = _mm256_add_pd(fiy0,ty);
 921             fiz0             = _mm256_add_pd(fiz0,tz);
 922
 923             fjx2             = _mm256_add_pd(fjx2,tx);
 924             fjy2             = _mm256_add_pd(fjy2,ty);
 925             fjz2             = _mm256_add_pd(fjz2,tz);
 926
 927             }
 928
 929             /**************************
 930              * CALCULATE INTERACTIONS *
 931              **************************/
 932
 933             if (gmx_mm256_any_lt(rsq10,rcutoff2))
 934             {
 935
 936             /* REACTION-FIELD ELECTROSTATICS */
 937             velec            = _mm256_mul_pd(qq10,_mm256_sub_pd(_mm256_add_pd(rinv10,_mm256_mul_pd(krf,rsq10)),crf));
 938             felec            = _mm256_mul_pd(qq10,_mm256_sub_pd(_mm256_mul_pd(rinv10,rinvsq10),krf2));
 939
 940             cutoff_mask      = _mm256_cmp_pd(rsq10,rcutoff2,_CMP_LT_OQ);
 941
 942             /* Update potential sum for this i atom from the interaction with this j atom. */
 943             velec            = _mm256_and_pd(velec,cutoff_mask);
 944             velec            = _mm256_andnot_pd(dummy_mask,velec);
 945             velecsum         = _mm256_add_pd(velecsum,velec);
 946
 947             fscal            = felec;
 948
 949             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 950
 951             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
 952
 953             /* Calculate temporary vectorial force */
 954             tx               = _mm256_mul_pd(fscal,dx10);
 955             ty               = _mm256_mul_pd(fscal,dy10);
 956             tz               = _mm256_mul_pd(fscal,dz10);
 957
 958             /* Update vectorial force */
 959             fix1             = _mm256_add_pd(fix1,tx);
 960             fiy1             = _mm256_add_pd(fiy1,ty);
 961             fiz1             = _mm256_add_pd(fiz1,tz);
 962
 963             fjx0             = _mm256_add_pd(fjx0,tx);
 964             fjy0             = _mm256_add_pd(fjy0,ty);
 965             fjz0             = _mm256_add_pd(fjz0,tz);
 966
 967             }
 968
 969             /**************************
 970              * CALCULATE INTERACTIONS *
 971              **************************/
 972
 973             if (gmx_mm256_any_lt(rsq11,rcutoff2))
 974             {
 975
 976             /* REACTION-FIELD ELECTROSTATICS */
 977             velec            = _mm256_mul_pd(qq11,_mm256_sub_pd(_mm256_add_pd(rinv11,_mm256_mul_pd(krf,rsq11)),crf));
 978             felec            = _mm256_mul_pd(qq11,_mm256_sub_pd(_mm256_mul_pd(rinv11,rinvsq11),krf2));
 979
 980             cutoff_mask      = _mm256_cmp_pd(rsq11,rcutoff2,_CMP_LT_OQ);
 981
 982             /* Update potential sum for this i atom from the interaction with this j atom. */
 983             velec            = _mm256_and_pd(velec,cutoff_mask);
 984             velec            = _mm256_andnot_pd(dummy_mask,velec);
 985             velecsum         = _mm256_add_pd(velecsum,velec);
 986
 987             fscal            = felec;
 988
 989             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 990
 991             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
 992
 993             /* Calculate temporary vectorial force */
 994             tx               = _mm256_mul_pd(fscal,dx11);
 995             ty               = _mm256_mul_pd(fscal,dy11);
 996             tz               = _mm256_mul_pd(fscal,dz11);
 997
 998             /* Update vectorial force */
 999             fix1             = _mm256_add_pd(fix1,tx);
1000             fiy1             = _mm256_add_pd(fiy1,ty);
1001             fiz1             = _mm256_add_pd(fiz1,tz);
1002
1003             fjx1             = _mm256_add_pd(fjx1,tx);
1004             fjy1             = _mm256_add_pd(fjy1,ty);
1005             fjz1             = _mm256_add_pd(fjz1,tz);
1006
1007             }
1008
1009             /**************************
1010              * CALCULATE INTERACTIONS *
1011              **************************/
1012
1013             if (gmx_mm256_any_lt(rsq12,rcutoff2))
1014             {
1015
1016             /* REACTION-FIELD ELECTROSTATICS */
1017             velec            = _mm256_mul_pd(qq12,_mm256_sub_pd(_mm256_add_pd(rinv12,_mm256_mul_pd(krf,rsq12)),crf));
1018             felec            = _mm256_mul_pd(qq12,_mm256_sub_pd(_mm256_mul_pd(rinv12,rinvsq12),krf2));
1019
1020             cutoff_mask      = _mm256_cmp_pd(rsq12,rcutoff2,_CMP_LT_OQ);
1021
1022             /* Update potential sum for this i atom from the interaction with this j atom. */
1023             velec            = _mm256_and_pd(velec,cutoff_mask);
1024             velec            = _mm256_andnot_pd(dummy_mask,velec);
1025             velecsum         = _mm256_add_pd(velecsum,velec);
1026
1027             fscal            = felec;
1028
1029             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1030
1031             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
1032
1033             /* Calculate temporary vectorial force */
1034             tx               = _mm256_mul_pd(fscal,dx12);
1035             ty               = _mm256_mul_pd(fscal,dy12);
1036             tz               = _mm256_mul_pd(fscal,dz12);
1037
1038             /* Update vectorial force */
1039             fix1             = _mm256_add_pd(fix1,tx);
1040             fiy1             = _mm256_add_pd(fiy1,ty);
1041             fiz1             = _mm256_add_pd(fiz1,tz);
1042
1043             fjx2             = _mm256_add_pd(fjx2,tx);
1044             fjy2             = _mm256_add_pd(fjy2,ty);
1045             fjz2             = _mm256_add_pd(fjz2,tz);
1046
1047             }
1048
1049             /**************************
1050              * CALCULATE INTERACTIONS *
1051              **************************/
1052
1053             if (gmx_mm256_any_lt(rsq20,rcutoff2))
1054             {
1055
1056             /* REACTION-FIELD ELECTROSTATICS */
1057             velec            = _mm256_mul_pd(qq20,_mm256_sub_pd(_mm256_add_pd(rinv20,_mm256_mul_pd(krf,rsq20)),crf));
1058             felec            = _mm256_mul_pd(qq20,_mm256_sub_pd(_mm256_mul_pd(rinv20,rinvsq20),krf2));
1059
1060             cutoff_mask      = _mm256_cmp_pd(rsq20,rcutoff2,_CMP_LT_OQ);
1061
1062             /* Update potential sum for this i atom from the interaction with this j atom. */
1063             velec            = _mm256_and_pd(velec,cutoff_mask);
1064             velec            = _mm256_andnot_pd(dummy_mask,velec);
1065             velecsum         = _mm256_add_pd(velecsum,velec);
1066
1067             fscal            = felec;
1068
1069             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1070
1071             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
1072
1073             /* Calculate temporary vectorial force */
1074             tx               = _mm256_mul_pd(fscal,dx20);
1075             ty               = _mm256_mul_pd(fscal,dy20);
1076             tz               = _mm256_mul_pd(fscal,dz20);
1077
1078             /* Update vectorial force */
1079             fix2             = _mm256_add_pd(fix2,tx);
1080             fiy2             = _mm256_add_pd(fiy2,ty);
1081             fiz2             = _mm256_add_pd(fiz2,tz);
1082
1083             fjx0             = _mm256_add_pd(fjx0,tx);
1084             fjy0             = _mm256_add_pd(fjy0,ty);
1085             fjz0             = _mm256_add_pd(fjz0,tz);
1086
1087             }
1088
1089             /**************************
1090              * CALCULATE INTERACTIONS *
1091              **************************/
1092
1093             if (gmx_mm256_any_lt(rsq21,rcutoff2))
1094             {
1095
1096             /* REACTION-FIELD ELECTROSTATICS */
1097             velec            = _mm256_mul_pd(qq21,_mm256_sub_pd(_mm256_add_pd(rinv21,_mm256_mul_pd(krf,rsq21)),crf));
1098             felec            = _mm256_mul_pd(qq21,_mm256_sub_pd(_mm256_mul_pd(rinv21,rinvsq21),krf2));
1099
1100             cutoff_mask      = _mm256_cmp_pd(rsq21,rcutoff2,_CMP_LT_OQ);
1101
1102             /* Update potential sum for this i atom from the interaction with this j atom. */
1103             velec            = _mm256_and_pd(velec,cutoff_mask);
1104             velec            = _mm256_andnot_pd(dummy_mask,velec);
1105             velecsum         = _mm256_add_pd(velecsum,velec);
1106
1107             fscal            = felec;
1108
1109             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1110
1111             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
1112
1113             /* Calculate temporary vectorial force */
1114             tx               = _mm256_mul_pd(fscal,dx21);
1115             ty               = _mm256_mul_pd(fscal,dy21);
1116             tz               = _mm256_mul_pd(fscal,dz21);
1117
1118             /* Update vectorial force */
1119             fix2             = _mm256_add_pd(fix2,tx);
1120             fiy2             = _mm256_add_pd(fiy2,ty);
1121             fiz2             = _mm256_add_pd(fiz2,tz);
1122
1123             fjx1             = _mm256_add_pd(fjx1,tx);
1124             fjy1             = _mm256_add_pd(fjy1,ty);
1125             fjz1             = _mm256_add_pd(fjz1,tz);
1126
1127             }
1128
1129             /**************************
1130              * CALCULATE INTERACTIONS *
1131              **************************/
1132
1133             if (gmx_mm256_any_lt(rsq22,rcutoff2))
1134             {
1135
1136             /* REACTION-FIELD ELECTROSTATICS */
1137             velec            = _mm256_mul_pd(qq22,_mm256_sub_pd(_mm256_add_pd(rinv22,_mm256_mul_pd(krf,rsq22)),crf));
1138             felec            = _mm256_mul_pd(qq22,_mm256_sub_pd(_mm256_mul_pd(rinv22,rinvsq22),krf2));
1139
1140             cutoff_mask      = _mm256_cmp_pd(rsq22,rcutoff2,_CMP_LT_OQ);
1141
1142             /* Update potential sum for this i atom from the interaction with this j atom. */
1143             velec            = _mm256_and_pd(velec,cutoff_mask);
1144             velec            = _mm256_andnot_pd(dummy_mask,velec);
1145             velecsum         = _mm256_add_pd(velecsum,velec);
1146
1147             fscal            = felec;
1148
1149             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1150
1151             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
1152
1153             /* Calculate temporary vectorial force */
1154             tx               = _mm256_mul_pd(fscal,dx22);
1155             ty               = _mm256_mul_pd(fscal,dy22);
1156             tz               = _mm256_mul_pd(fscal,dz22);
1157
1158             /* Update vectorial force */
1159             fix2             = _mm256_add_pd(fix2,tx);
1160             fiy2             = _mm256_add_pd(fiy2,ty);
1161             fiz2             = _mm256_add_pd(fiz2,tz);
1162
1163             fjx2             = _mm256_add_pd(fjx2,tx);
1164             fjy2             = _mm256_add_pd(fjy2,ty);
1165             fjz2             = _mm256_add_pd(fjz2,tz);
1166
1167             }
1168
1169             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
1170             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
1171             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
1172             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
1173
1174             gmx_mm256_decrement_3rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,
1175                                                       fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2);
1176
1177             /* Inner loop uses 359 flops */
1178         }
1179
1180         /* End of innermost loop */
1181
1182         gmx_mm256_update_iforce_3atom_swizzle_pd(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,
1183                                                  f+i_coord_offset,fshift+i_shift_offset);
1184
1185         ggid                        = gid[iidx];
1186         /* Update potential energies */
1187         gmx_mm256_update_1pot_pd(velecsum,kernel_data->energygrp_elec+ggid);
1188         gmx_mm256_update_1pot_pd(vvdwsum,kernel_data->energygrp_vdw+ggid);
1189
1190         /* Increment number of inner iterations */
1191         inneriter                  += j_index_end - j_index_start;
1192
1193         /* Outer loop uses 20 flops */
1194     }
1195
1196     /* Increment number of outer iterations */
1197     outeriter        += nri;
1198
1199     /* Update outer/inner flops */
1200
1201     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3W3_VF,outeriter*20 + inneriter*359);
1202 }
1203 /*
1204  * Gromacs nonbonded kernel:   nb_kernel_ElecRFCut_VdwLJSw_GeomW3W3_F_avx_256_double
1205  * Electrostatics interaction: ReactionField
1206  * VdW interaction:            LennardJones
1207  * Geometry:                   Water3-Water3
1208  * Calculate force/pot:        Force
1209  */
1210 void
1211 nb_kernel_ElecRFCut_VdwLJSw_GeomW3W3_F_avx_256_double
1212                     (t_nblist                    * gmx_restrict       nlist,
1213                      rvec                        * gmx_restrict          xx,
1214                      rvec                        * gmx_restrict          ff,
1215                      struct t_forcerec           * gmx_restrict          fr,
1216                      t_mdatoms                   * gmx_restrict     mdatoms,
1217                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
1218                      t_nrnb                      * gmx_restrict        nrnb)
1219 {
1220     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
1221      * just 0 for non-waters.
1222      * Suffixes A,B,C,D refer to j loop unrolling done with AVX, e.g. for the four different
1223      * jnr indices corresponding to data put in the four positions in the SIMD register.
1224      */
1225     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
1226     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
1227     int              jnrA,jnrB,jnrC,jnrD;
1228     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
1229     int              jnrlistE,jnrlistF,jnrlistG,jnrlistH;
1230     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
1231     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
1232     real             rcutoff_scalar;
1233     real             *shiftvec,*fshift,*x,*f;
1234     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD;
1235     real             scratch[4*DIM];
1236     __m256d          tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
1237     real *           vdwioffsetptr0;
1238     __m256d          ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
1239     real *           vdwioffsetptr1;
1240     __m256d          ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
1241     real *           vdwioffsetptr2;
1242     __m256d          ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
1243     int              vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D;
1244     __m256d          jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
1245     int              vdwjidx1A,vdwjidx1B,vdwjidx1C,vdwjidx1D;
1246     __m256d          jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
1247     int              vdwjidx2A,vdwjidx2B,vdwjidx2C,vdwjidx2D;
1248     __m256d          jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
1249     __m256d          dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
1250     __m256d          dx01,dy01,dz01,rsq01,rinv01,rinvsq01,r01,qq01,c6_01,c12_01;
1251     __m256d          dx02,dy02,dz02,rsq02,rinv02,rinvsq02,r02,qq02,c6_02,c12_02;
1252     __m256d          dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
1253     __m256d          dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11;
1254     __m256d          dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12;
1255     __m256d          dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
1256     __m256d          dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21;
1257     __m256d          dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22;
1258     __m256d          velec,felec,velecsum,facel,crf,krf,krf2;
1259     real             *charge;
1260     int              nvdwtype;
1261     __m256d          rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
1262     int              *vdwtype;
1263     real             *vdwparam;
1264     __m256d          one_sixth   = _mm256_set1_pd(1.0/6.0);
1265     __m256d          one_twelfth = _mm256_set1_pd(1.0/12.0);
1266     __m256d          rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
1267     real             rswitch_scalar,d_scalar;
1268     __m256d          dummy_mask,cutoff_mask;
1269     __m128           tmpmask0,tmpmask1;
1270     __m256d          signbit = _mm256_castsi256_pd( _mm256_set1_epi32(0x80000000) );
1271     __m256d          one     = _mm256_set1_pd(1.0);
1272     __m256d          two     = _mm256_set1_pd(2.0);
1273     x                = xx[0];
1274     f                = ff[0];
1275
1276     nri              = nlist->nri;
1277     iinr             = nlist->iinr;
1278     jindex           = nlist->jindex;
1279     jjnr             = nlist->jjnr;
1280     shiftidx         = nlist->shift;
1281     gid              = nlist->gid;
1282     shiftvec         = fr->shift_vec[0];
1283     fshift           = fr->fshift[0];
1284     facel            = _mm256_set1_pd(fr->ic->epsfac);
1285     charge           = mdatoms->chargeA;
1286     krf              = _mm256_set1_pd(fr->ic->k_rf);
1287     krf2             = _mm256_set1_pd(fr->ic->k_rf*2.0);
1288     crf              = _mm256_set1_pd(fr->ic->c_rf);
1289     nvdwtype         = fr->ntype;
1290     vdwparam         = fr->nbfp;
1291     vdwtype          = mdatoms->typeA;
1292
1293     /* Setup water-specific parameters */
1294     inr              = nlist->iinr[0];
1295     iq0              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+0]));
1296     iq1              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+1]));
1297     iq2              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+2]));
1298     vdwioffsetptr0   = vdwparam+2*nvdwtype*vdwtype[inr+0];
1299
1300     jq0              = _mm256_set1_pd(charge[inr+0]);
1301     jq1              = _mm256_set1_pd(charge[inr+1]);
1302     jq2              = _mm256_set1_pd(charge[inr+2]);
1303     vdwjidx0A        = 2*vdwtype[inr+0];
1304     qq00             = _mm256_mul_pd(iq0,jq0);
1305     c6_00            = _mm256_set1_pd(vdwioffsetptr0[vdwjidx0A]);
1306     c12_00           = _mm256_set1_pd(vdwioffsetptr0[vdwjidx0A+1]);
1307     qq01             = _mm256_mul_pd(iq0,jq1);
1308     qq02             = _mm256_mul_pd(iq0,jq2);
1309     qq10             = _mm256_mul_pd(iq1,jq0);
1310     qq11             = _mm256_mul_pd(iq1,jq1);
1311     qq12             = _mm256_mul_pd(iq1,jq2);
1312     qq20             = _mm256_mul_pd(iq2,jq0);
1313     qq21             = _mm256_mul_pd(iq2,jq1);
1314     qq22             = _mm256_mul_pd(iq2,jq2);
1315
1316     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
1317     rcutoff_scalar   = fr->ic->rcoulomb;
1318     rcutoff          = _mm256_set1_pd(rcutoff_scalar);
1319     rcutoff2         = _mm256_mul_pd(rcutoff,rcutoff);
1320
1321     rswitch_scalar   = fr->ic->rvdw_switch;
1322     rswitch          = _mm256_set1_pd(rswitch_scalar);
1323     /* Setup switch parameters */
1324     d_scalar         = rcutoff_scalar-rswitch_scalar;
1325     d                = _mm256_set1_pd(d_scalar);
1326     swV3             = _mm256_set1_pd(-10.0/(d_scalar*d_scalar*d_scalar));
1327     swV4             = _mm256_set1_pd( 15.0/(d_scalar*d_scalar*d_scalar*d_scalar));
1328     swV5             = _mm256_set1_pd( -6.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
1329     swF2             = _mm256_set1_pd(-30.0/(d_scalar*d_scalar*d_scalar));
1330     swF3             = _mm256_set1_pd( 60.0/(d_scalar*d_scalar*d_scalar*d_scalar));
1331     swF4             = _mm256_set1_pd(-30.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
1332
1333     /* Avoid stupid compiler warnings */
1334     jnrA = jnrB = jnrC = jnrD = 0;
1335     j_coord_offsetA = 0;
1336     j_coord_offsetB = 0;
1337     j_coord_offsetC = 0;
1338     j_coord_offsetD = 0;
1339
1340     outeriter        = 0;
1341     inneriter        = 0;
1342
1343     for(iidx=0;iidx<4*DIM;iidx++)
1344     {
1345         scratch[iidx] = 0.0;
1346     }
1347
1348     /* Start outer loop over neighborlists */
1349     for(iidx=0; iidx<nri; iidx++)
1350     {
1351         /* Load shift vector for this list */
1352         i_shift_offset   = DIM*shiftidx[iidx];
1353
1354         /* Load limits for loop over neighbors */
1355         j_index_start    = jindex[iidx];
1356         j_index_end      = jindex[iidx+1];
1357
1358         /* Get outer coordinate index */
1359         inr              = iinr[iidx];
1360         i_coord_offset   = DIM*inr;
1361
1362         /* Load i particle coords and add shift vector */
1363         gmx_mm256_load_shift_and_3rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset,
1364                                                     &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2);
1365
1366         fix0             = _mm256_setzero_pd();
1367         fiy0             = _mm256_setzero_pd();
1368         fiz0             = _mm256_setzero_pd();
1369         fix1             = _mm256_setzero_pd();
1370         fiy1             = _mm256_setzero_pd();
1371         fiz1             = _mm256_setzero_pd();
1372         fix2             = _mm256_setzero_pd();
1373         fiy2             = _mm256_setzero_pd();
1374         fiz2             = _mm256_setzero_pd();
1375
1376         /* Start inner kernel loop */
1377         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4)
1378         {
1379
1380             /* Get j neighbor index, and coordinate index */
1381             jnrA             = jjnr[jidx];
1382             jnrB             = jjnr[jidx+1];
1383             jnrC             = jjnr[jidx+2];
1384             jnrD             = jjnr[jidx+3];
1385             j_coord_offsetA  = DIM*jnrA;
1386             j_coord_offsetB  = DIM*jnrB;
1387             j_coord_offsetC  = DIM*jnrC;
1388             j_coord_offsetD  = DIM*jnrD;
1389
1390             /* load j atom coordinates */
1391             gmx_mm256_load_3rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
1392                                                  x+j_coord_offsetC,x+j_coord_offsetD,
1393                                               &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,&jy2,&jz2);
1394
1395             /* Calculate displacement vector */
1396             dx00             = _mm256_sub_pd(ix0,jx0);
1397             dy00             = _mm256_sub_pd(iy0,jy0);
1398             dz00             = _mm256_sub_pd(iz0,jz0);
1399             dx01             = _mm256_sub_pd(ix0,jx1);
1400             dy01             = _mm256_sub_pd(iy0,jy1);
1401             dz01             = _mm256_sub_pd(iz0,jz1);
1402             dx02             = _mm256_sub_pd(ix0,jx2);
1403             dy02             = _mm256_sub_pd(iy0,jy2);
1404             dz02             = _mm256_sub_pd(iz0,jz2);
1405             dx10             = _mm256_sub_pd(ix1,jx0);
1406             dy10             = _mm256_sub_pd(iy1,jy0);
1407             dz10             = _mm256_sub_pd(iz1,jz0);
1408             dx11             = _mm256_sub_pd(ix1,jx1);
1409             dy11             = _mm256_sub_pd(iy1,jy1);
1410             dz11             = _mm256_sub_pd(iz1,jz1);
1411             dx12             = _mm256_sub_pd(ix1,jx2);
1412             dy12             = _mm256_sub_pd(iy1,jy2);
1413             dz12             = _mm256_sub_pd(iz1,jz2);
1414             dx20             = _mm256_sub_pd(ix2,jx0);
1415             dy20             = _mm256_sub_pd(iy2,jy0);
1416             dz20             = _mm256_sub_pd(iz2,jz0);
1417             dx21             = _mm256_sub_pd(ix2,jx1);
1418             dy21             = _mm256_sub_pd(iy2,jy1);
1419             dz21             = _mm256_sub_pd(iz2,jz1);
1420             dx22             = _mm256_sub_pd(ix2,jx2);
1421             dy22             = _mm256_sub_pd(iy2,jy2);
1422             dz22             = _mm256_sub_pd(iz2,jz2);
1423
1424             /* Calculate squared distance and things based on it */
1425             rsq00            = gmx_mm256_calc_rsq_pd(dx00,dy00,dz00);
1426             rsq01            = gmx_mm256_calc_rsq_pd(dx01,dy01,dz01);
1427             rsq02            = gmx_mm256_calc_rsq_pd(dx02,dy02,dz02);
1428             rsq10            = gmx_mm256_calc_rsq_pd(dx10,dy10,dz10);
1429             rsq11            = gmx_mm256_calc_rsq_pd(dx11,dy11,dz11);
1430             rsq12            = gmx_mm256_calc_rsq_pd(dx12,dy12,dz12);
1431             rsq20            = gmx_mm256_calc_rsq_pd(dx20,dy20,dz20);
1432             rsq21            = gmx_mm256_calc_rsq_pd(dx21,dy21,dz21);
1433             rsq22            = gmx_mm256_calc_rsq_pd(dx22,dy22,dz22);
1434
1435             rinv00           = avx256_invsqrt_d(rsq00);
1436             rinv01           = avx256_invsqrt_d(rsq01);
1437             rinv02           = avx256_invsqrt_d(rsq02);
1438             rinv10           = avx256_invsqrt_d(rsq10);
1439             rinv11           = avx256_invsqrt_d(rsq11);
1440             rinv12           = avx256_invsqrt_d(rsq12);
1441             rinv20           = avx256_invsqrt_d(rsq20);
1442             rinv21           = avx256_invsqrt_d(rsq21);
1443             rinv22           = avx256_invsqrt_d(rsq22);
1444
1445             rinvsq00         = _mm256_mul_pd(rinv00,rinv00);
1446             rinvsq01         = _mm256_mul_pd(rinv01,rinv01);
1447             rinvsq02         = _mm256_mul_pd(rinv02,rinv02);
1448             rinvsq10         = _mm256_mul_pd(rinv10,rinv10);
1449             rinvsq11         = _mm256_mul_pd(rinv11,rinv11);
1450             rinvsq12         = _mm256_mul_pd(rinv12,rinv12);
1451             rinvsq20         = _mm256_mul_pd(rinv20,rinv20);
1452             rinvsq21         = _mm256_mul_pd(rinv21,rinv21);
1453             rinvsq22         = _mm256_mul_pd(rinv22,rinv22);
1454
1455             fjx0             = _mm256_setzero_pd();
1456             fjy0             = _mm256_setzero_pd();
1457             fjz0             = _mm256_setzero_pd();
1458             fjx1             = _mm256_setzero_pd();
1459             fjy1             = _mm256_setzero_pd();
1460             fjz1             = _mm256_setzero_pd();
1461             fjx2             = _mm256_setzero_pd();
1462             fjy2             = _mm256_setzero_pd();
1463             fjz2             = _mm256_setzero_pd();
1464
1465             /**************************
1466              * CALCULATE INTERACTIONS *
1467              **************************/
1468
1469             if (gmx_mm256_any_lt(rsq00,rcutoff2))
1470             {
1471
1472             r00              = _mm256_mul_pd(rsq00,rinv00);
1473
1474             /* REACTION-FIELD ELECTROSTATICS */
1475             felec            = _mm256_mul_pd(qq00,_mm256_sub_pd(_mm256_mul_pd(rinv00,rinvsq00),krf2));
1476
1477             /* LENNARD-JONES DISPERSION/REPULSION */
1478
1479             rinvsix          = _mm256_mul_pd(_mm256_mul_pd(rinvsq00,rinvsq00),rinvsq00);
1480             vvdw6            = _mm256_mul_pd(c6_00,rinvsix);
1481             vvdw12           = _mm256_mul_pd(c12_00,_mm256_mul_pd(rinvsix,rinvsix));
1482             vvdw             = _mm256_sub_pd( _mm256_mul_pd(vvdw12,one_twelfth) , _mm256_mul_pd(vvdw6,one_sixth) );
1483             fvdw             = _mm256_mul_pd(_mm256_sub_pd(vvdw12,vvdw6),rinvsq00);
1484
1485             d                = _mm256_sub_pd(r00,rswitch);
1486             d                = _mm256_max_pd(d,_mm256_setzero_pd());
1487             d2               = _mm256_mul_pd(d,d);
1488             sw               = _mm256_add_pd(one,_mm256_mul_pd(d2,_mm256_mul_pd(d,_mm256_add_pd(swV3,_mm256_mul_pd(d,_mm256_add_pd(swV4,_mm256_mul_pd(d,swV5)))))));
1489
1490             dsw              = _mm256_mul_pd(d2,_mm256_add_pd(swF2,_mm256_mul_pd(d,_mm256_add_pd(swF3,_mm256_mul_pd(d,swF4)))));
1491
1492             /* Evaluate switch function */
1493             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1494             fvdw             = _mm256_sub_pd( _mm256_mul_pd(fvdw,sw) , _mm256_mul_pd(rinv00,_mm256_mul_pd(vvdw,dsw)) );
1495             cutoff_mask      = _mm256_cmp_pd(rsq00,rcutoff2,_CMP_LT_OQ);
1496
1497             fscal            = _mm256_add_pd(felec,fvdw);
1498
1499             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1500
1501             /* Calculate temporary vectorial force */
1502             tx               = _mm256_mul_pd(fscal,dx00);
1503             ty               = _mm256_mul_pd(fscal,dy00);
1504             tz               = _mm256_mul_pd(fscal,dz00);
1505
1506             /* Update vectorial force */
1507             fix0             = _mm256_add_pd(fix0,tx);
1508             fiy0             = _mm256_add_pd(fiy0,ty);
1509             fiz0             = _mm256_add_pd(fiz0,tz);
1510
1511             fjx0             = _mm256_add_pd(fjx0,tx);
1512             fjy0             = _mm256_add_pd(fjy0,ty);
1513             fjz0             = _mm256_add_pd(fjz0,tz);
1514
1515             }
1516
1517             /**************************
1518              * CALCULATE INTERACTIONS *
1519              **************************/
1520
1521             if (gmx_mm256_any_lt(rsq01,rcutoff2))
1522             {
1523
1524             /* REACTION-FIELD ELECTROSTATICS */
1525             felec            = _mm256_mul_pd(qq01,_mm256_sub_pd(_mm256_mul_pd(rinv01,rinvsq01),krf2));
1526
1527             cutoff_mask      = _mm256_cmp_pd(rsq01,rcutoff2,_CMP_LT_OQ);
1528
1529             fscal            = felec;
1530
1531             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1532
1533             /* Calculate temporary vectorial force */
1534             tx               = _mm256_mul_pd(fscal,dx01);
1535             ty               = _mm256_mul_pd(fscal,dy01);
1536             tz               = _mm256_mul_pd(fscal,dz01);
1537
1538             /* Update vectorial force */
1539             fix0             = _mm256_add_pd(fix0,tx);
1540             fiy0             = _mm256_add_pd(fiy0,ty);
1541             fiz0             = _mm256_add_pd(fiz0,tz);
1542
1543             fjx1             = _mm256_add_pd(fjx1,tx);
1544             fjy1             = _mm256_add_pd(fjy1,ty);
1545             fjz1             = _mm256_add_pd(fjz1,tz);
1546
1547             }
1548
1549             /**************************
1550              * CALCULATE INTERACTIONS *
1551              **************************/
1552
1553             if (gmx_mm256_any_lt(rsq02,rcutoff2))
1554             {
1555
1556             /* REACTION-FIELD ELECTROSTATICS */
1557             felec            = _mm256_mul_pd(qq02,_mm256_sub_pd(_mm256_mul_pd(rinv02,rinvsq02),krf2));
1558
1559             cutoff_mask      = _mm256_cmp_pd(rsq02,rcutoff2,_CMP_LT_OQ);
1560
1561             fscal            = felec;
1562
1563             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1564
1565             /* Calculate temporary vectorial force */
1566             tx               = _mm256_mul_pd(fscal,dx02);
1567             ty               = _mm256_mul_pd(fscal,dy02);
1568             tz               = _mm256_mul_pd(fscal,dz02);
1569
1570             /* Update vectorial force */
1571             fix0             = _mm256_add_pd(fix0,tx);
1572             fiy0             = _mm256_add_pd(fiy0,ty);
1573             fiz0             = _mm256_add_pd(fiz0,tz);
1574
1575             fjx2             = _mm256_add_pd(fjx2,tx);
1576             fjy2             = _mm256_add_pd(fjy2,ty);
1577             fjz2             = _mm256_add_pd(fjz2,tz);
1578
1579             }
1580
1581             /**************************
1582              * CALCULATE INTERACTIONS *
1583              **************************/
1584
1585             if (gmx_mm256_any_lt(rsq10,rcutoff2))
1586             {
1587
1588             /* REACTION-FIELD ELECTROSTATICS */
1589             felec            = _mm256_mul_pd(qq10,_mm256_sub_pd(_mm256_mul_pd(rinv10,rinvsq10),krf2));
1590
1591             cutoff_mask      = _mm256_cmp_pd(rsq10,rcutoff2,_CMP_LT_OQ);
1592
1593             fscal            = felec;
1594
1595             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1596
1597             /* Calculate temporary vectorial force */
1598             tx               = _mm256_mul_pd(fscal,dx10);
1599             ty               = _mm256_mul_pd(fscal,dy10);
1600             tz               = _mm256_mul_pd(fscal,dz10);
1601
1602             /* Update vectorial force */
1603             fix1             = _mm256_add_pd(fix1,tx);
1604             fiy1             = _mm256_add_pd(fiy1,ty);
1605             fiz1             = _mm256_add_pd(fiz1,tz);
1606
1607             fjx0             = _mm256_add_pd(fjx0,tx);
1608             fjy0             = _mm256_add_pd(fjy0,ty);
1609             fjz0             = _mm256_add_pd(fjz0,tz);
1610
1611             }
1612
1613             /**************************
1614              * CALCULATE INTERACTIONS *
1615              **************************/
1616
1617             if (gmx_mm256_any_lt(rsq11,rcutoff2))
1618             {
1619
1620             /* REACTION-FIELD ELECTROSTATICS */
1621             felec            = _mm256_mul_pd(qq11,_mm256_sub_pd(_mm256_mul_pd(rinv11,rinvsq11),krf2));
1622
1623             cutoff_mask      = _mm256_cmp_pd(rsq11,rcutoff2,_CMP_LT_OQ);
1624
1625             fscal            = felec;
1626
1627             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1628
1629             /* Calculate temporary vectorial force */
1630             tx               = _mm256_mul_pd(fscal,dx11);
1631             ty               = _mm256_mul_pd(fscal,dy11);
1632             tz               = _mm256_mul_pd(fscal,dz11);
1633
1634             /* Update vectorial force */
1635             fix1             = _mm256_add_pd(fix1,tx);
1636             fiy1             = _mm256_add_pd(fiy1,ty);
1637             fiz1             = _mm256_add_pd(fiz1,tz);
1638
1639             fjx1             = _mm256_add_pd(fjx1,tx);
1640             fjy1             = _mm256_add_pd(fjy1,ty);
1641             fjz1             = _mm256_add_pd(fjz1,tz);
1642
1643             }
1644
1645             /**************************
1646              * CALCULATE INTERACTIONS *
1647              **************************/
1648
1649             if (gmx_mm256_any_lt(rsq12,rcutoff2))
1650             {
1651
1652             /* REACTION-FIELD ELECTROSTATICS */
1653             felec            = _mm256_mul_pd(qq12,_mm256_sub_pd(_mm256_mul_pd(rinv12,rinvsq12),krf2));
1654
1655             cutoff_mask      = _mm256_cmp_pd(rsq12,rcutoff2,_CMP_LT_OQ);
1656
1657             fscal            = felec;
1658
1659             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1660
1661             /* Calculate temporary vectorial force */
1662             tx               = _mm256_mul_pd(fscal,dx12);
1663             ty               = _mm256_mul_pd(fscal,dy12);
1664             tz               = _mm256_mul_pd(fscal,dz12);
1665
1666             /* Update vectorial force */
1667             fix1             = _mm256_add_pd(fix1,tx);
1668             fiy1             = _mm256_add_pd(fiy1,ty);
1669             fiz1             = _mm256_add_pd(fiz1,tz);
1670
1671             fjx2             = _mm256_add_pd(fjx2,tx);
1672             fjy2             = _mm256_add_pd(fjy2,ty);
1673             fjz2             = _mm256_add_pd(fjz2,tz);
1674
1675             }
1676
1677             /**************************
1678              * CALCULATE INTERACTIONS *
1679              **************************/
1680
1681             if (gmx_mm256_any_lt(rsq20,rcutoff2))
1682             {
1683
1684             /* REACTION-FIELD ELECTROSTATICS */
1685             felec            = _mm256_mul_pd(qq20,_mm256_sub_pd(_mm256_mul_pd(rinv20,rinvsq20),krf2));
1686
1687             cutoff_mask      = _mm256_cmp_pd(rsq20,rcutoff2,_CMP_LT_OQ);
1688
1689             fscal            = felec;
1690
1691             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1692
1693             /* Calculate temporary vectorial force */
1694             tx               = _mm256_mul_pd(fscal,dx20);
1695             ty               = _mm256_mul_pd(fscal,dy20);
1696             tz               = _mm256_mul_pd(fscal,dz20);
1697
1698             /* Update vectorial force */
1699             fix2             = _mm256_add_pd(fix2,tx);
1700             fiy2             = _mm256_add_pd(fiy2,ty);
1701             fiz2             = _mm256_add_pd(fiz2,tz);
1702
1703             fjx0             = _mm256_add_pd(fjx0,tx);
1704             fjy0             = _mm256_add_pd(fjy0,ty);
1705             fjz0             = _mm256_add_pd(fjz0,tz);
1706
1707             }
1708
1709             /**************************
1710              * CALCULATE INTERACTIONS *
1711              **************************/
1712
1713             if (gmx_mm256_any_lt(rsq21,rcutoff2))
1714             {
1715
1716             /* REACTION-FIELD ELECTROSTATICS */
1717             felec            = _mm256_mul_pd(qq21,_mm256_sub_pd(_mm256_mul_pd(rinv21,rinvsq21),krf2));
1718
1719             cutoff_mask      = _mm256_cmp_pd(rsq21,rcutoff2,_CMP_LT_OQ);
1720
1721             fscal            = felec;
1722
1723             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1724
1725             /* Calculate temporary vectorial force */
1726             tx               = _mm256_mul_pd(fscal,dx21);
1727             ty               = _mm256_mul_pd(fscal,dy21);
1728             tz               = _mm256_mul_pd(fscal,dz21);
1729
1730             /* Update vectorial force */
1731             fix2             = _mm256_add_pd(fix2,tx);
1732             fiy2             = _mm256_add_pd(fiy2,ty);
1733             fiz2             = _mm256_add_pd(fiz2,tz);
1734
1735             fjx1             = _mm256_add_pd(fjx1,tx);
1736             fjy1             = _mm256_add_pd(fjy1,ty);
1737             fjz1             = _mm256_add_pd(fjz1,tz);
1738
1739             }
1740
1741             /**************************
1742              * CALCULATE INTERACTIONS *
1743              **************************/
1744
1745             if (gmx_mm256_any_lt(rsq22,rcutoff2))
1746             {
1747
1748             /* REACTION-FIELD ELECTROSTATICS */
1749             felec            = _mm256_mul_pd(qq22,_mm256_sub_pd(_mm256_mul_pd(rinv22,rinvsq22),krf2));
1750
1751             cutoff_mask      = _mm256_cmp_pd(rsq22,rcutoff2,_CMP_LT_OQ);
1752
1753             fscal            = felec;
1754
1755             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1756
1757             /* Calculate temporary vectorial force */
1758             tx               = _mm256_mul_pd(fscal,dx22);
1759             ty               = _mm256_mul_pd(fscal,dy22);
1760             tz               = _mm256_mul_pd(fscal,dz22);
1761
1762             /* Update vectorial force */
1763             fix2             = _mm256_add_pd(fix2,tx);
1764             fiy2             = _mm256_add_pd(fiy2,ty);
1765             fiz2             = _mm256_add_pd(fiz2,tz);
1766
1767             fjx2             = _mm256_add_pd(fjx2,tx);
1768             fjy2             = _mm256_add_pd(fjy2,ty);
1769             fjz2             = _mm256_add_pd(fjz2,tz);
1770
1771             }
1772
1773             fjptrA             = f+j_coord_offsetA;
1774             fjptrB             = f+j_coord_offsetB;
1775             fjptrC             = f+j_coord_offsetC;
1776             fjptrD             = f+j_coord_offsetD;
1777
1778             gmx_mm256_decrement_3rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,
1779                                                       fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2);
1780
1781             /* Inner loop uses 301 flops */
1782         }
1783
1784         if(jidx<j_index_end)
1785         {
1786
1787             /* Get j neighbor index, and coordinate index */
1788             jnrlistA         = jjnr[jidx];
1789             jnrlistB         = jjnr[jidx+1];
1790             jnrlistC         = jjnr[jidx+2];
1791             jnrlistD         = jjnr[jidx+3];
1792             /* Sign of each element will be negative for non-real atoms.
1793              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
1794              * so use it as val = _mm_andnot_pd(mask,val) to clear dummy entries.
1795              */
1796             tmpmask0 = gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128()));
1797
1798             tmpmask1 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(3,3,2,2));
1799             tmpmask0 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(1,1,0,0));
1800             dummy_mask = _mm256_castps_pd(gmx_mm256_set_m128(tmpmask1,tmpmask0));
1801
1802             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
1803             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
1804             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
1805             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
1806             j_coord_offsetA  = DIM*jnrA;
1807             j_coord_offsetB  = DIM*jnrB;
1808             j_coord_offsetC  = DIM*jnrC;
1809             j_coord_offsetD  = DIM*jnrD;
1810
1811             /* load j atom coordinates */
1812             gmx_mm256_load_3rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
1813                                                  x+j_coord_offsetC,x+j_coord_offsetD,
1814                                               &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,&jy2,&jz2);
1815
1816             /* Calculate displacement vector */
1817             dx00             = _mm256_sub_pd(ix0,jx0);
1818             dy00             = _mm256_sub_pd(iy0,jy0);
1819             dz00             = _mm256_sub_pd(iz0,jz0);
1820             dx01             = _mm256_sub_pd(ix0,jx1);
1821             dy01             = _mm256_sub_pd(iy0,jy1);
1822             dz01             = _mm256_sub_pd(iz0,jz1);
1823             dx02             = _mm256_sub_pd(ix0,jx2);
1824             dy02             = _mm256_sub_pd(iy0,jy2);
1825             dz02             = _mm256_sub_pd(iz0,jz2);
1826             dx10             = _mm256_sub_pd(ix1,jx0);
1827             dy10             = _mm256_sub_pd(iy1,jy0);
1828             dz10             = _mm256_sub_pd(iz1,jz0);
1829             dx11             = _mm256_sub_pd(ix1,jx1);
1830             dy11             = _mm256_sub_pd(iy1,jy1);
1831             dz11             = _mm256_sub_pd(iz1,jz1);
1832             dx12             = _mm256_sub_pd(ix1,jx2);
1833             dy12             = _mm256_sub_pd(iy1,jy2);
1834             dz12             = _mm256_sub_pd(iz1,jz2);
1835             dx20             = _mm256_sub_pd(ix2,jx0);
1836             dy20             = _mm256_sub_pd(iy2,jy0);
1837             dz20             = _mm256_sub_pd(iz2,jz0);
1838             dx21             = _mm256_sub_pd(ix2,jx1);
1839             dy21             = _mm256_sub_pd(iy2,jy1);
1840             dz21             = _mm256_sub_pd(iz2,jz1);
1841             dx22             = _mm256_sub_pd(ix2,jx2);
1842             dy22             = _mm256_sub_pd(iy2,jy2);
1843             dz22             = _mm256_sub_pd(iz2,jz2);
1844
1845             /* Calculate squared distance and things based on it */
1846             rsq00            = gmx_mm256_calc_rsq_pd(dx00,dy00,dz00);
1847             rsq01            = gmx_mm256_calc_rsq_pd(dx01,dy01,dz01);
1848             rsq02            = gmx_mm256_calc_rsq_pd(dx02,dy02,dz02);
1849             rsq10            = gmx_mm256_calc_rsq_pd(dx10,dy10,dz10);
1850             rsq11            = gmx_mm256_calc_rsq_pd(dx11,dy11,dz11);
1851             rsq12            = gmx_mm256_calc_rsq_pd(dx12,dy12,dz12);
1852             rsq20            = gmx_mm256_calc_rsq_pd(dx20,dy20,dz20);
1853             rsq21            = gmx_mm256_calc_rsq_pd(dx21,dy21,dz21);
1854             rsq22            = gmx_mm256_calc_rsq_pd(dx22,dy22,dz22);
1855
1856             rinv00           = avx256_invsqrt_d(rsq00);
1857             rinv01           = avx256_invsqrt_d(rsq01);
1858             rinv02           = avx256_invsqrt_d(rsq02);
1859             rinv10           = avx256_invsqrt_d(rsq10);
1860             rinv11           = avx256_invsqrt_d(rsq11);
1861             rinv12           = avx256_invsqrt_d(rsq12);
1862             rinv20           = avx256_invsqrt_d(rsq20);
1863             rinv21           = avx256_invsqrt_d(rsq21);
1864             rinv22           = avx256_invsqrt_d(rsq22);
1865
1866             rinvsq00         = _mm256_mul_pd(rinv00,rinv00);
1867             rinvsq01         = _mm256_mul_pd(rinv01,rinv01);
1868             rinvsq02         = _mm256_mul_pd(rinv02,rinv02);
1869             rinvsq10         = _mm256_mul_pd(rinv10,rinv10);
1870             rinvsq11         = _mm256_mul_pd(rinv11,rinv11);
1871             rinvsq12         = _mm256_mul_pd(rinv12,rinv12);
1872             rinvsq20         = _mm256_mul_pd(rinv20,rinv20);
1873             rinvsq21         = _mm256_mul_pd(rinv21,rinv21);
1874             rinvsq22         = _mm256_mul_pd(rinv22,rinv22);
1875
1876             fjx0             = _mm256_setzero_pd();
1877             fjy0             = _mm256_setzero_pd();
1878             fjz0             = _mm256_setzero_pd();
1879             fjx1             = _mm256_setzero_pd();
1880             fjy1             = _mm256_setzero_pd();
1881             fjz1             = _mm256_setzero_pd();
1882             fjx2             = _mm256_setzero_pd();
1883             fjy2             = _mm256_setzero_pd();
1884             fjz2             = _mm256_setzero_pd();
1885
1886             /**************************
1887              * CALCULATE INTERACTIONS *
1888              **************************/
1889
1890             if (gmx_mm256_any_lt(rsq00,rcutoff2))
1891             {
1892
1893             r00              = _mm256_mul_pd(rsq00,rinv00);
1894             r00              = _mm256_andnot_pd(dummy_mask,r00);
1895
1896             /* REACTION-FIELD ELECTROSTATICS */
1897             felec            = _mm256_mul_pd(qq00,_mm256_sub_pd(_mm256_mul_pd(rinv00,rinvsq00),krf2));
1898
1899             /* LENNARD-JONES DISPERSION/REPULSION */
1900
1901             rinvsix          = _mm256_mul_pd(_mm256_mul_pd(rinvsq00,rinvsq00),rinvsq00);
1902             vvdw6            = _mm256_mul_pd(c6_00,rinvsix);
1903             vvdw12           = _mm256_mul_pd(c12_00,_mm256_mul_pd(rinvsix,rinvsix));
1904             vvdw             = _mm256_sub_pd( _mm256_mul_pd(vvdw12,one_twelfth) , _mm256_mul_pd(vvdw6,one_sixth) );
1905             fvdw             = _mm256_mul_pd(_mm256_sub_pd(vvdw12,vvdw6),rinvsq00);
1906
1907             d                = _mm256_sub_pd(r00,rswitch);
1908             d                = _mm256_max_pd(d,_mm256_setzero_pd());
1909             d2               = _mm256_mul_pd(d,d);
1910             sw               = _mm256_add_pd(one,_mm256_mul_pd(d2,_mm256_mul_pd(d,_mm256_add_pd(swV3,_mm256_mul_pd(d,_mm256_add_pd(swV4,_mm256_mul_pd(d,swV5)))))));
1911
1912             dsw              = _mm256_mul_pd(d2,_mm256_add_pd(swF2,_mm256_mul_pd(d,_mm256_add_pd(swF3,_mm256_mul_pd(d,swF4)))));
1913
1914             /* Evaluate switch function */
1915             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
1916             fvdw             = _mm256_sub_pd( _mm256_mul_pd(fvdw,sw) , _mm256_mul_pd(rinv00,_mm256_mul_pd(vvdw,dsw)) );
1917             cutoff_mask      = _mm256_cmp_pd(rsq00,rcutoff2,_CMP_LT_OQ);
1918
1919             fscal            = _mm256_add_pd(felec,fvdw);
1920
1921             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1922
1923             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
1924
1925             /* Calculate temporary vectorial force */
1926             tx               = _mm256_mul_pd(fscal,dx00);
1927             ty               = _mm256_mul_pd(fscal,dy00);
1928             tz               = _mm256_mul_pd(fscal,dz00);
1929
1930             /* Update vectorial force */
1931             fix0             = _mm256_add_pd(fix0,tx);
1932             fiy0             = _mm256_add_pd(fiy0,ty);
1933             fiz0             = _mm256_add_pd(fiz0,tz);
1934
1935             fjx0             = _mm256_add_pd(fjx0,tx);
1936             fjy0             = _mm256_add_pd(fjy0,ty);
1937             fjz0             = _mm256_add_pd(fjz0,tz);
1938
1939             }
1940
1941             /**************************
1942              * CALCULATE INTERACTIONS *
1943              **************************/
1944
1945             if (gmx_mm256_any_lt(rsq01,rcutoff2))
1946             {
1947
1948             /* REACTION-FIELD ELECTROSTATICS */
1949             felec            = _mm256_mul_pd(qq01,_mm256_sub_pd(_mm256_mul_pd(rinv01,rinvsq01),krf2));
1950
1951             cutoff_mask      = _mm256_cmp_pd(rsq01,rcutoff2,_CMP_LT_OQ);
1952
1953             fscal            = felec;
1954
1955             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1956
1957             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
1958
1959             /* Calculate temporary vectorial force */
1960             tx               = _mm256_mul_pd(fscal,dx01);
1961             ty               = _mm256_mul_pd(fscal,dy01);
1962             tz               = _mm256_mul_pd(fscal,dz01);
1963
1964             /* Update vectorial force */
1965             fix0             = _mm256_add_pd(fix0,tx);
1966             fiy0             = _mm256_add_pd(fiy0,ty);
1967             fiz0             = _mm256_add_pd(fiz0,tz);
1968
1969             fjx1             = _mm256_add_pd(fjx1,tx);
1970             fjy1             = _mm256_add_pd(fjy1,ty);
1971             fjz1             = _mm256_add_pd(fjz1,tz);
1972
1973             }
1974
1975             /**************************
1976              * CALCULATE INTERACTIONS *
1977              **************************/
1978
1979             if (gmx_mm256_any_lt(rsq02,rcutoff2))
1980             {
1981
1982             /* REACTION-FIELD ELECTROSTATICS */
1983             felec            = _mm256_mul_pd(qq02,_mm256_sub_pd(_mm256_mul_pd(rinv02,rinvsq02),krf2));
1984
1985             cutoff_mask      = _mm256_cmp_pd(rsq02,rcutoff2,_CMP_LT_OQ);
1986
1987             fscal            = felec;
1988
1989             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1990
1991             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
1992
1993             /* Calculate temporary vectorial force */
1994             tx               = _mm256_mul_pd(fscal,dx02);
1995             ty               = _mm256_mul_pd(fscal,dy02);
1996             tz               = _mm256_mul_pd(fscal,dz02);
1997
1998             /* Update vectorial force */
1999             fix0             = _mm256_add_pd(fix0,tx);
2000             fiy0             = _mm256_add_pd(fiy0,ty);
2001             fiz0             = _mm256_add_pd(fiz0,tz);
2002
2003             fjx2             = _mm256_add_pd(fjx2,tx);
2004             fjy2             = _mm256_add_pd(fjy2,ty);
2005             fjz2             = _mm256_add_pd(fjz2,tz);
2006
2007             }
2008
2009             /**************************
2010              * CALCULATE INTERACTIONS *
2011              **************************/
2012
2013             if (gmx_mm256_any_lt(rsq10,rcutoff2))
2014             {
2015
2016             /* REACTION-FIELD ELECTROSTATICS */
2017             felec            = _mm256_mul_pd(qq10,_mm256_sub_pd(_mm256_mul_pd(rinv10,rinvsq10),krf2));
2018
2019             cutoff_mask      = _mm256_cmp_pd(rsq10,rcutoff2,_CMP_LT_OQ);
2020
2021             fscal            = felec;
2022
2023             fscal            = _mm256_and_pd(fscal,cutoff_mask);
2024
2025             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
2026
2027             /* Calculate temporary vectorial force */
2028             tx               = _mm256_mul_pd(fscal,dx10);
2029             ty               = _mm256_mul_pd(fscal,dy10);
2030             tz               = _mm256_mul_pd(fscal,dz10);
2031
2032             /* Update vectorial force */
2033             fix1             = _mm256_add_pd(fix1,tx);
2034             fiy1             = _mm256_add_pd(fiy1,ty);
2035             fiz1             = _mm256_add_pd(fiz1,tz);
2036
2037             fjx0             = _mm256_add_pd(fjx0,tx);
2038             fjy0             = _mm256_add_pd(fjy0,ty);
2039             fjz0             = _mm256_add_pd(fjz0,tz);
2040
2041             }
2042
2043             /**************************
2044              * CALCULATE INTERACTIONS *
2045              **************************/
2046
2047             if (gmx_mm256_any_lt(rsq11,rcutoff2))
2048             {
2049
2050             /* REACTION-FIELD ELECTROSTATICS */
2051             felec            = _mm256_mul_pd(qq11,_mm256_sub_pd(_mm256_mul_pd(rinv11,rinvsq11),krf2));
2052
2053             cutoff_mask      = _mm256_cmp_pd(rsq11,rcutoff2,_CMP_LT_OQ);
2054
2055             fscal            = felec;
2056
2057             fscal            = _mm256_and_pd(fscal,cutoff_mask);
2058
2059             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
2060
2061             /* Calculate temporary vectorial force */
2062             tx               = _mm256_mul_pd(fscal,dx11);
2063             ty               = _mm256_mul_pd(fscal,dy11);
2064             tz               = _mm256_mul_pd(fscal,dz11);
2065
2066             /* Update vectorial force */
2067             fix1             = _mm256_add_pd(fix1,tx);
2068             fiy1             = _mm256_add_pd(fiy1,ty);
2069             fiz1             = _mm256_add_pd(fiz1,tz);
2070
2071             fjx1             = _mm256_add_pd(fjx1,tx);
2072             fjy1             = _mm256_add_pd(fjy1,ty);
2073             fjz1             = _mm256_add_pd(fjz1,tz);
2074
2075             }
2076
2077             /**************************
2078              * CALCULATE INTERACTIONS *
2079              **************************/
2080
2081             if (gmx_mm256_any_lt(rsq12,rcutoff2))
2082             {
2083
2084             /* REACTION-FIELD ELECTROSTATICS */
2085             felec            = _mm256_mul_pd(qq12,_mm256_sub_pd(_mm256_mul_pd(rinv12,rinvsq12),krf2));
2086
2087             cutoff_mask      = _mm256_cmp_pd(rsq12,rcutoff2,_CMP_LT_OQ);
2088
2089             fscal            = felec;
2090
2091             fscal            = _mm256_and_pd(fscal,cutoff_mask);
2092
2093             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
2094
2095             /* Calculate temporary vectorial force */
2096             tx               = _mm256_mul_pd(fscal,dx12);
2097             ty               = _mm256_mul_pd(fscal,dy12);
2098             tz               = _mm256_mul_pd(fscal,dz12);
2099
2100             /* Update vectorial force */
2101             fix1             = _mm256_add_pd(fix1,tx);
2102             fiy1             = _mm256_add_pd(fiy1,ty);
2103             fiz1             = _mm256_add_pd(fiz1,tz);
2104
2105             fjx2             = _mm256_add_pd(fjx2,tx);
2106             fjy2             = _mm256_add_pd(fjy2,ty);
2107             fjz2             = _mm256_add_pd(fjz2,tz);
2108
2109             }
2110
2111             /**************************
2112              * CALCULATE INTERACTIONS *
2113              **************************/
2114
2115             if (gmx_mm256_any_lt(rsq20,rcutoff2))
2116             {
2117
2118             /* REACTION-FIELD ELECTROSTATICS */
2119             felec            = _mm256_mul_pd(qq20,_mm256_sub_pd(_mm256_mul_pd(rinv20,rinvsq20),krf2));
2120
2121             cutoff_mask      = _mm256_cmp_pd(rsq20,rcutoff2,_CMP_LT_OQ);
2122
2123             fscal            = felec;
2124
2125             fscal            = _mm256_and_pd(fscal,cutoff_mask);
2126
2127             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
2128
2129             /* Calculate temporary vectorial force */
2130             tx               = _mm256_mul_pd(fscal,dx20);
2131             ty               = _mm256_mul_pd(fscal,dy20);
2132             tz               = _mm256_mul_pd(fscal,dz20);
2133
2134             /* Update vectorial force */
2135             fix2             = _mm256_add_pd(fix2,tx);
2136             fiy2             = _mm256_add_pd(fiy2,ty);
2137             fiz2             = _mm256_add_pd(fiz2,tz);
2138
2139             fjx0             = _mm256_add_pd(fjx0,tx);
2140             fjy0             = _mm256_add_pd(fjy0,ty);
2141             fjz0             = _mm256_add_pd(fjz0,tz);
2142
2143             }
2144
2145             /**************************
2146              * CALCULATE INTERACTIONS *
2147              **************************/
2148
2149             if (gmx_mm256_any_lt(rsq21,rcutoff2))
2150             {
2151
2152             /* REACTION-FIELD ELECTROSTATICS */
2153             felec            = _mm256_mul_pd(qq21,_mm256_sub_pd(_mm256_mul_pd(rinv21,rinvsq21),krf2));
2154
2155             cutoff_mask      = _mm256_cmp_pd(rsq21,rcutoff2,_CMP_LT_OQ);
2156
2157             fscal            = felec;
2158
2159             fscal            = _mm256_and_pd(fscal,cutoff_mask);
2160
2161             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
2162
2163             /* Calculate temporary vectorial force */
2164             tx               = _mm256_mul_pd(fscal,dx21);
2165             ty               = _mm256_mul_pd(fscal,dy21);
2166             tz               = _mm256_mul_pd(fscal,dz21);
2167
2168             /* Update vectorial force */
2169             fix2             = _mm256_add_pd(fix2,tx);
2170             fiy2             = _mm256_add_pd(fiy2,ty);
2171             fiz2             = _mm256_add_pd(fiz2,tz);
2172
2173             fjx1             = _mm256_add_pd(fjx1,tx);
2174             fjy1             = _mm256_add_pd(fjy1,ty);
2175             fjz1             = _mm256_add_pd(fjz1,tz);
2176
2177             }
2178
2179             /**************************
2180              * CALCULATE INTERACTIONS *
2181              **************************/
2182
2183             if (gmx_mm256_any_lt(rsq22,rcutoff2))
2184             {
2185
2186             /* REACTION-FIELD ELECTROSTATICS */
2187             felec            = _mm256_mul_pd(qq22,_mm256_sub_pd(_mm256_mul_pd(rinv22,rinvsq22),krf2));
2188
2189             cutoff_mask      = _mm256_cmp_pd(rsq22,rcutoff2,_CMP_LT_OQ);
2190
2191             fscal            = felec;
2192
2193             fscal            = _mm256_and_pd(fscal,cutoff_mask);
2194
2195             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
2196
2197             /* Calculate temporary vectorial force */
2198             tx               = _mm256_mul_pd(fscal,dx22);
2199             ty               = _mm256_mul_pd(fscal,dy22);
2200             tz               = _mm256_mul_pd(fscal,dz22);
2201
2202             /* Update vectorial force */
2203             fix2             = _mm256_add_pd(fix2,tx);
2204             fiy2             = _mm256_add_pd(fiy2,ty);
2205             fiz2             = _mm256_add_pd(fiz2,tz);
2206
2207             fjx2             = _mm256_add_pd(fjx2,tx);
2208             fjy2             = _mm256_add_pd(fjy2,ty);
2209             fjz2             = _mm256_add_pd(fjz2,tz);
2210
2211             }
2212
2213             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
2214             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
2215             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
2216             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
2217
2218             gmx_mm256_decrement_3rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,
2219                                                       fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2);
2220
2221             /* Inner loop uses 302 flops */
2222         }
2223
2224         /* End of innermost loop */
2225
2226         gmx_mm256_update_iforce_3atom_swizzle_pd(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,
2227                                                  f+i_coord_offset,fshift+i_shift_offset);
2228
2229         /* Increment number of inner iterations */
2230         inneriter                  += j_index_end - j_index_start;
2231
2232         /* Outer loop uses 18 flops */
2233     }
2234
2235     /* Increment number of outer iterations */
2236     outeriter        += nri;
2237
2238     /* Update outer/inner flops */
2239
2240     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3W3_F,outeriter*18 + inneriter*302);
2241 }