src/gromacs/gmxlib/nonbonded/nb_kernel_avx_256_double/nb_kernel_ElecRFCut_VdwNone_GeomW4W4_avx_256_double.c

   1 /*
   2  * This file is part of the GROMACS molecular simulation package.
   3  *
   4  * Copyright (c) 2012,2013,2014, by the GROMACS development team, led by
   5  * Mark Abraham, David van der Spoel, Berk Hess, and Erik Lindahl,
   6  * and including many others, as listed in the AUTHORS file in the
   7  * top-level source directory and at http://www.gromacs.org.
   8  *
   9  * GROMACS is free software; you can redistribute it and/or
  10  * modify it under the terms of the GNU Lesser General Public License
  11  * as published by the Free Software Foundation; either version 2.1
  12  * of the License, or (at your option) any later version.
  13  *
  14  * GROMACS is distributed in the hope that it will be useful,
  15  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  16  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  17  * Lesser General Public License for more details.
  18  *
  19  * You should have received a copy of the GNU Lesser General Public
  20  * License along with GROMACS; if not, see
  21  * http://www.gnu.org/licenses, or write to the Free Software Foundation,
  22  * Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA.
  23  *
  24  * If you want to redistribute modifications to GROMACS, please
  25  * consider that scientific software is very special. Version
  26  * control is crucial - bugs must be traceable. We will be happy to
  27  * consider code for inclusion in the official distribution, but
  28  * derived work must not be called official GROMACS. Details are found
  29  * in the README & COPYING files - if they are missing, get the
  30  * official version at http://www.gromacs.org.
  31  *
  32  * To help us fund GROMACS development, we humbly ask that you cite
  33  * the research papers on the package. Check out http://www.gromacs.org.
  34  */
  35 /*
  36  * Note: this file was generated by the GROMACS avx_256_double kernel generator.
  37  */
  38 #include "config.h"
  39
  40 #include <math.h>
  41
  42 #include "../nb_kernel.h"
  43 #include "types/simple.h"
  44 #include "gromacs/math/vec.h"
  45 #include "nrnb.h"
  46
  47 #include "gromacs/simd/math_x86_avx_256_double.h"
  48 #include "kernelutil_x86_avx_256_double.h"
  49
  50 /*
  51  * Gromacs nonbonded kernel:   nb_kernel_ElecRFCut_VdwNone_GeomW4W4_VF_avx_256_double
  52  * Electrostatics interaction: ReactionField
  53  * VdW interaction:            None
  54  * Geometry:                   Water4-Water4
  55  * Calculate force/pot:        PotentialAndForce
  56  */
  57 void
  58 nb_kernel_ElecRFCut_VdwNone_GeomW4W4_VF_avx_256_double
  59                     (t_nblist                    * gmx_restrict       nlist,
  60                      rvec                        * gmx_restrict          xx,
  61                      rvec                        * gmx_restrict          ff,
  62                      t_forcerec                  * gmx_restrict          fr,
  63                      t_mdatoms                   * gmx_restrict     mdatoms,
  64                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
  65                      t_nrnb                      * gmx_restrict        nrnb)
  66 {
  67     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
  68      * just 0 for non-waters.
  69      * Suffixes A,B,C,D refer to j loop unrolling done with AVX, e.g. for the four different
  70      * jnr indices corresponding to data put in the four positions in the SIMD register.
  71      */
  72     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
  73     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
  74     int              jnrA,jnrB,jnrC,jnrD;
  75     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
  76     int              jnrlistE,jnrlistF,jnrlistG,jnrlistH;
  77     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
  78     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  79     real             rcutoff_scalar;
  80     real             *shiftvec,*fshift,*x,*f;
  81     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD;
  82     real             scratch[4*DIM];
  83     __m256d          tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
  84     real *           vdwioffsetptr1;
  85     __m256d          ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
  86     real *           vdwioffsetptr2;
  87     __m256d          ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
  88     real *           vdwioffsetptr3;
  89     __m256d          ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
  90     int              vdwjidx1A,vdwjidx1B,vdwjidx1C,vdwjidx1D;
  91     __m256d          jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
  92     int              vdwjidx2A,vdwjidx2B,vdwjidx2C,vdwjidx2D;
  93     __m256d          jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
  94     int              vdwjidx3A,vdwjidx3B,vdwjidx3C,vdwjidx3D;
  95     __m256d          jx3,jy3,jz3,fjx3,fjy3,fjz3,jq3,isaj3;
  96     __m256d          dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11;
  97     __m256d          dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12;
  98     __m256d          dx13,dy13,dz13,rsq13,rinv13,rinvsq13,r13,qq13,c6_13,c12_13;
  99     __m256d          dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21;
 100     __m256d          dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22;
 101     __m256d          dx23,dy23,dz23,rsq23,rinv23,rinvsq23,r23,qq23,c6_23,c12_23;
 102     __m256d          dx31,dy31,dz31,rsq31,rinv31,rinvsq31,r31,qq31,c6_31,c12_31;
 103     __m256d          dx32,dy32,dz32,rsq32,rinv32,rinvsq32,r32,qq32,c6_32,c12_32;
 104     __m256d          dx33,dy33,dz33,rsq33,rinv33,rinvsq33,r33,qq33,c6_33,c12_33;
 105     __m256d          velec,felec,velecsum,facel,crf,krf,krf2;
 106     real             *charge;
 107     __m256d          dummy_mask,cutoff_mask;
 108     __m128           tmpmask0,tmpmask1;
 109     __m256d          signbit = _mm256_castsi256_pd( _mm256_set1_epi32(0x80000000) );
 110     __m256d          one     = _mm256_set1_pd(1.0);
 111     __m256d          two     = _mm256_set1_pd(2.0);
 112     x                = xx[0];
 113     f                = ff[0];
 114
 115     nri              = nlist->nri;
 116     iinr             = nlist->iinr;
 117     jindex           = nlist->jindex;
 118     jjnr             = nlist->jjnr;
 119     shiftidx         = nlist->shift;
 120     gid              = nlist->gid;
 121     shiftvec         = fr->shift_vec[0];
 122     fshift           = fr->fshift[0];
 123     facel            = _mm256_set1_pd(fr->epsfac);
 124     charge           = mdatoms->chargeA;
 125     krf              = _mm256_set1_pd(fr->ic->k_rf);
 126     krf2             = _mm256_set1_pd(fr->ic->k_rf*2.0);
 127     crf              = _mm256_set1_pd(fr->ic->c_rf);
 128
 129     /* Setup water-specific parameters */
 130     inr              = nlist->iinr[0];
 131     iq1              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+1]));
 132     iq2              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+2]));
 133     iq3              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+3]));
 134
 135     jq1              = _mm256_set1_pd(charge[inr+1]);
 136     jq2              = _mm256_set1_pd(charge[inr+2]);
 137     jq3              = _mm256_set1_pd(charge[inr+3]);
 138     qq11             = _mm256_mul_pd(iq1,jq1);
 139     qq12             = _mm256_mul_pd(iq1,jq2);
 140     qq13             = _mm256_mul_pd(iq1,jq3);
 141     qq21             = _mm256_mul_pd(iq2,jq1);
 142     qq22             = _mm256_mul_pd(iq2,jq2);
 143     qq23             = _mm256_mul_pd(iq2,jq3);
 144     qq31             = _mm256_mul_pd(iq3,jq1);
 145     qq32             = _mm256_mul_pd(iq3,jq2);
 146     qq33             = _mm256_mul_pd(iq3,jq3);
 147
 148     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 149     rcutoff_scalar   = fr->rcoulomb;
 150     rcutoff          = _mm256_set1_pd(rcutoff_scalar);
 151     rcutoff2         = _mm256_mul_pd(rcutoff,rcutoff);
 152
 153     /* Avoid stupid compiler warnings */
 154     jnrA = jnrB = jnrC = jnrD = 0;
 155     j_coord_offsetA = 0;
 156     j_coord_offsetB = 0;
 157     j_coord_offsetC = 0;
 158     j_coord_offsetD = 0;
 159
 160     outeriter        = 0;
 161     inneriter        = 0;
 162
 163     for(iidx=0;iidx<4*DIM;iidx++)
 164     {
 165         scratch[iidx] = 0.0;
 166     }
 167
 168     /* Start outer loop over neighborlists */
 169     for(iidx=0; iidx<nri; iidx++)
 170     {
 171         /* Load shift vector for this list */
 172         i_shift_offset   = DIM*shiftidx[iidx];
 173
 174         /* Load limits for loop over neighbors */
 175         j_index_start    = jindex[iidx];
 176         j_index_end      = jindex[iidx+1];
 177
 178         /* Get outer coordinate index */
 179         inr              = iinr[iidx];
 180         i_coord_offset   = DIM*inr;
 181
 182         /* Load i particle coords and add shift vector */
 183         gmx_mm256_load_shift_and_3rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset+DIM,
 184                                                     &ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
 185
 186         fix1             = _mm256_setzero_pd();
 187         fiy1             = _mm256_setzero_pd();
 188         fiz1             = _mm256_setzero_pd();
 189         fix2             = _mm256_setzero_pd();
 190         fiy2             = _mm256_setzero_pd();
 191         fiz2             = _mm256_setzero_pd();
 192         fix3             = _mm256_setzero_pd();
 193         fiy3             = _mm256_setzero_pd();
 194         fiz3             = _mm256_setzero_pd();
 195
 196         /* Reset potential sums */
 197         velecsum         = _mm256_setzero_pd();
 198
 199         /* Start inner kernel loop */
 200         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4)
 201         {
 202
 203             /* Get j neighbor index, and coordinate index */
 204             jnrA             = jjnr[jidx];
 205             jnrB             = jjnr[jidx+1];
 206             jnrC             = jjnr[jidx+2];
 207             jnrD             = jjnr[jidx+3];
 208             j_coord_offsetA  = DIM*jnrA;
 209             j_coord_offsetB  = DIM*jnrB;
 210             j_coord_offsetC  = DIM*jnrC;
 211             j_coord_offsetD  = DIM*jnrD;
 212
 213             /* load j atom coordinates */
 214             gmx_mm256_load_3rvec_4ptr_swizzle_pd(x+j_coord_offsetA+DIM,x+j_coord_offsetB+DIM,
 215                                                  x+j_coord_offsetC+DIM,x+j_coord_offsetD+DIM,
 216                                                  &jx1,&jy1,&jz1,&jx2,&jy2,&jz2,&jx3,&jy3,&jz3);
 217
 218             /* Calculate displacement vector */
 219             dx11             = _mm256_sub_pd(ix1,jx1);
 220             dy11             = _mm256_sub_pd(iy1,jy1);
 221             dz11             = _mm256_sub_pd(iz1,jz1);
 222             dx12             = _mm256_sub_pd(ix1,jx2);
 223             dy12             = _mm256_sub_pd(iy1,jy2);
 224             dz12             = _mm256_sub_pd(iz1,jz2);
 225             dx13             = _mm256_sub_pd(ix1,jx3);
 226             dy13             = _mm256_sub_pd(iy1,jy3);
 227             dz13             = _mm256_sub_pd(iz1,jz3);
 228             dx21             = _mm256_sub_pd(ix2,jx1);
 229             dy21             = _mm256_sub_pd(iy2,jy1);
 230             dz21             = _mm256_sub_pd(iz2,jz1);
 231             dx22             = _mm256_sub_pd(ix2,jx2);
 232             dy22             = _mm256_sub_pd(iy2,jy2);
 233             dz22             = _mm256_sub_pd(iz2,jz2);
 234             dx23             = _mm256_sub_pd(ix2,jx3);
 235             dy23             = _mm256_sub_pd(iy2,jy3);
 236             dz23             = _mm256_sub_pd(iz2,jz3);
 237             dx31             = _mm256_sub_pd(ix3,jx1);
 238             dy31             = _mm256_sub_pd(iy3,jy1);
 239             dz31             = _mm256_sub_pd(iz3,jz1);
 240             dx32             = _mm256_sub_pd(ix3,jx2);
 241             dy32             = _mm256_sub_pd(iy3,jy2);
 242             dz32             = _mm256_sub_pd(iz3,jz2);
 243             dx33             = _mm256_sub_pd(ix3,jx3);
 244             dy33             = _mm256_sub_pd(iy3,jy3);
 245             dz33             = _mm256_sub_pd(iz3,jz3);
 246
 247             /* Calculate squared distance and things based on it */
 248             rsq11            = gmx_mm256_calc_rsq_pd(dx11,dy11,dz11);
 249             rsq12            = gmx_mm256_calc_rsq_pd(dx12,dy12,dz12);
 250             rsq13            = gmx_mm256_calc_rsq_pd(dx13,dy13,dz13);
 251             rsq21            = gmx_mm256_calc_rsq_pd(dx21,dy21,dz21);
 252             rsq22            = gmx_mm256_calc_rsq_pd(dx22,dy22,dz22);
 253             rsq23            = gmx_mm256_calc_rsq_pd(dx23,dy23,dz23);
 254             rsq31            = gmx_mm256_calc_rsq_pd(dx31,dy31,dz31);
 255             rsq32            = gmx_mm256_calc_rsq_pd(dx32,dy32,dz32);
 256             rsq33            = gmx_mm256_calc_rsq_pd(dx33,dy33,dz33);
 257
 258             rinv11           = gmx_mm256_invsqrt_pd(rsq11);
 259             rinv12           = gmx_mm256_invsqrt_pd(rsq12);
 260             rinv13           = gmx_mm256_invsqrt_pd(rsq13);
 261             rinv21           = gmx_mm256_invsqrt_pd(rsq21);
 262             rinv22           = gmx_mm256_invsqrt_pd(rsq22);
 263             rinv23           = gmx_mm256_invsqrt_pd(rsq23);
 264             rinv31           = gmx_mm256_invsqrt_pd(rsq31);
 265             rinv32           = gmx_mm256_invsqrt_pd(rsq32);
 266             rinv33           = gmx_mm256_invsqrt_pd(rsq33);
 267
 268             rinvsq11         = _mm256_mul_pd(rinv11,rinv11);
 269             rinvsq12         = _mm256_mul_pd(rinv12,rinv12);
 270             rinvsq13         = _mm256_mul_pd(rinv13,rinv13);
 271             rinvsq21         = _mm256_mul_pd(rinv21,rinv21);
 272             rinvsq22         = _mm256_mul_pd(rinv22,rinv22);
 273             rinvsq23         = _mm256_mul_pd(rinv23,rinv23);
 274             rinvsq31         = _mm256_mul_pd(rinv31,rinv31);
 275             rinvsq32         = _mm256_mul_pd(rinv32,rinv32);
 276             rinvsq33         = _mm256_mul_pd(rinv33,rinv33);
 277
 278             fjx1             = _mm256_setzero_pd();
 279             fjy1             = _mm256_setzero_pd();
 280             fjz1             = _mm256_setzero_pd();
 281             fjx2             = _mm256_setzero_pd();
 282             fjy2             = _mm256_setzero_pd();
 283             fjz2             = _mm256_setzero_pd();
 284             fjx3             = _mm256_setzero_pd();
 285             fjy3             = _mm256_setzero_pd();
 286             fjz3             = _mm256_setzero_pd();
 287
 288             /**************************
 289              * CALCULATE INTERACTIONS *
 290              **************************/
 291
 292             if (gmx_mm256_any_lt(rsq11,rcutoff2))
 293             {
 294
 295             /* REACTION-FIELD ELECTROSTATICS */
 296             velec            = _mm256_mul_pd(qq11,_mm256_sub_pd(_mm256_add_pd(rinv11,_mm256_mul_pd(krf,rsq11)),crf));
 297             felec            = _mm256_mul_pd(qq11,_mm256_sub_pd(_mm256_mul_pd(rinv11,rinvsq11),krf2));
 298
 299             cutoff_mask      = _mm256_cmp_pd(rsq11,rcutoff2,_CMP_LT_OQ);
 300
 301             /* Update potential sum for this i atom from the interaction with this j atom. */
 302             velec            = _mm256_and_pd(velec,cutoff_mask);
 303             velecsum         = _mm256_add_pd(velecsum,velec);
 304
 305             fscal            = felec;
 306
 307             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 308
 309             /* Calculate temporary vectorial force */
 310             tx               = _mm256_mul_pd(fscal,dx11);
 311             ty               = _mm256_mul_pd(fscal,dy11);
 312             tz               = _mm256_mul_pd(fscal,dz11);
 313
 314             /* Update vectorial force */
 315             fix1             = _mm256_add_pd(fix1,tx);
 316             fiy1             = _mm256_add_pd(fiy1,ty);
 317             fiz1             = _mm256_add_pd(fiz1,tz);
 318
 319             fjx1             = _mm256_add_pd(fjx1,tx);
 320             fjy1             = _mm256_add_pd(fjy1,ty);
 321             fjz1             = _mm256_add_pd(fjz1,tz);
 322
 323             }
 324
 325             /**************************
 326              * CALCULATE INTERACTIONS *
 327              **************************/
 328
 329             if (gmx_mm256_any_lt(rsq12,rcutoff2))
 330             {
 331
 332             /* REACTION-FIELD ELECTROSTATICS */
 333             velec            = _mm256_mul_pd(qq12,_mm256_sub_pd(_mm256_add_pd(rinv12,_mm256_mul_pd(krf,rsq12)),crf));
 334             felec            = _mm256_mul_pd(qq12,_mm256_sub_pd(_mm256_mul_pd(rinv12,rinvsq12),krf2));
 335
 336             cutoff_mask      = _mm256_cmp_pd(rsq12,rcutoff2,_CMP_LT_OQ);
 337
 338             /* Update potential sum for this i atom from the interaction with this j atom. */
 339             velec            = _mm256_and_pd(velec,cutoff_mask);
 340             velecsum         = _mm256_add_pd(velecsum,velec);
 341
 342             fscal            = felec;
 343
 344             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 345
 346             /* Calculate temporary vectorial force */
 347             tx               = _mm256_mul_pd(fscal,dx12);
 348             ty               = _mm256_mul_pd(fscal,dy12);
 349             tz               = _mm256_mul_pd(fscal,dz12);
 350
 351             /* Update vectorial force */
 352             fix1             = _mm256_add_pd(fix1,tx);
 353             fiy1             = _mm256_add_pd(fiy1,ty);
 354             fiz1             = _mm256_add_pd(fiz1,tz);
 355
 356             fjx2             = _mm256_add_pd(fjx2,tx);
 357             fjy2             = _mm256_add_pd(fjy2,ty);
 358             fjz2             = _mm256_add_pd(fjz2,tz);
 359
 360             }
 361
 362             /**************************
 363              * CALCULATE INTERACTIONS *
 364              **************************/
 365
 366             if (gmx_mm256_any_lt(rsq13,rcutoff2))
 367             {
 368
 369             /* REACTION-FIELD ELECTROSTATICS */
 370             velec            = _mm256_mul_pd(qq13,_mm256_sub_pd(_mm256_add_pd(rinv13,_mm256_mul_pd(krf,rsq13)),crf));
 371             felec            = _mm256_mul_pd(qq13,_mm256_sub_pd(_mm256_mul_pd(rinv13,rinvsq13),krf2));
 372
 373             cutoff_mask      = _mm256_cmp_pd(rsq13,rcutoff2,_CMP_LT_OQ);
 374
 375             /* Update potential sum for this i atom from the interaction with this j atom. */
 376             velec            = _mm256_and_pd(velec,cutoff_mask);
 377             velecsum         = _mm256_add_pd(velecsum,velec);
 378
 379             fscal            = felec;
 380
 381             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 382
 383             /* Calculate temporary vectorial force */
 384             tx               = _mm256_mul_pd(fscal,dx13);
 385             ty               = _mm256_mul_pd(fscal,dy13);
 386             tz               = _mm256_mul_pd(fscal,dz13);
 387
 388             /* Update vectorial force */
 389             fix1             = _mm256_add_pd(fix1,tx);
 390             fiy1             = _mm256_add_pd(fiy1,ty);
 391             fiz1             = _mm256_add_pd(fiz1,tz);
 392
 393             fjx3             = _mm256_add_pd(fjx3,tx);
 394             fjy3             = _mm256_add_pd(fjy3,ty);
 395             fjz3             = _mm256_add_pd(fjz3,tz);
 396
 397             }
 398
 399             /**************************
 400              * CALCULATE INTERACTIONS *
 401              **************************/
 402
 403             if (gmx_mm256_any_lt(rsq21,rcutoff2))
 404             {
 405
 406             /* REACTION-FIELD ELECTROSTATICS */
 407             velec            = _mm256_mul_pd(qq21,_mm256_sub_pd(_mm256_add_pd(rinv21,_mm256_mul_pd(krf,rsq21)),crf));
 408             felec            = _mm256_mul_pd(qq21,_mm256_sub_pd(_mm256_mul_pd(rinv21,rinvsq21),krf2));
 409
 410             cutoff_mask      = _mm256_cmp_pd(rsq21,rcutoff2,_CMP_LT_OQ);
 411
 412             /* Update potential sum for this i atom from the interaction with this j atom. */
 413             velec            = _mm256_and_pd(velec,cutoff_mask);
 414             velecsum         = _mm256_add_pd(velecsum,velec);
 415
 416             fscal            = felec;
 417
 418             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 419
 420             /* Calculate temporary vectorial force */
 421             tx               = _mm256_mul_pd(fscal,dx21);
 422             ty               = _mm256_mul_pd(fscal,dy21);
 423             tz               = _mm256_mul_pd(fscal,dz21);
 424
 425             /* Update vectorial force */
 426             fix2             = _mm256_add_pd(fix2,tx);
 427             fiy2             = _mm256_add_pd(fiy2,ty);
 428             fiz2             = _mm256_add_pd(fiz2,tz);
 429
 430             fjx1             = _mm256_add_pd(fjx1,tx);
 431             fjy1             = _mm256_add_pd(fjy1,ty);
 432             fjz1             = _mm256_add_pd(fjz1,tz);
 433
 434             }
 435
 436             /**************************
 437              * CALCULATE INTERACTIONS *
 438              **************************/
 439
 440             if (gmx_mm256_any_lt(rsq22,rcutoff2))
 441             {
 442
 443             /* REACTION-FIELD ELECTROSTATICS */
 444             velec            = _mm256_mul_pd(qq22,_mm256_sub_pd(_mm256_add_pd(rinv22,_mm256_mul_pd(krf,rsq22)),crf));
 445             felec            = _mm256_mul_pd(qq22,_mm256_sub_pd(_mm256_mul_pd(rinv22,rinvsq22),krf2));
 446
 447             cutoff_mask      = _mm256_cmp_pd(rsq22,rcutoff2,_CMP_LT_OQ);
 448
 449             /* Update potential sum for this i atom from the interaction with this j atom. */
 450             velec            = _mm256_and_pd(velec,cutoff_mask);
 451             velecsum         = _mm256_add_pd(velecsum,velec);
 452
 453             fscal            = felec;
 454
 455             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 456
 457             /* Calculate temporary vectorial force */
 458             tx               = _mm256_mul_pd(fscal,dx22);
 459             ty               = _mm256_mul_pd(fscal,dy22);
 460             tz               = _mm256_mul_pd(fscal,dz22);
 461
 462             /* Update vectorial force */
 463             fix2             = _mm256_add_pd(fix2,tx);
 464             fiy2             = _mm256_add_pd(fiy2,ty);
 465             fiz2             = _mm256_add_pd(fiz2,tz);
 466
 467             fjx2             = _mm256_add_pd(fjx2,tx);
 468             fjy2             = _mm256_add_pd(fjy2,ty);
 469             fjz2             = _mm256_add_pd(fjz2,tz);
 470
 471             }
 472
 473             /**************************
 474              * CALCULATE INTERACTIONS *
 475              **************************/
 476
 477             if (gmx_mm256_any_lt(rsq23,rcutoff2))
 478             {
 479
 480             /* REACTION-FIELD ELECTROSTATICS */
 481             velec            = _mm256_mul_pd(qq23,_mm256_sub_pd(_mm256_add_pd(rinv23,_mm256_mul_pd(krf,rsq23)),crf));
 482             felec            = _mm256_mul_pd(qq23,_mm256_sub_pd(_mm256_mul_pd(rinv23,rinvsq23),krf2));
 483
 484             cutoff_mask      = _mm256_cmp_pd(rsq23,rcutoff2,_CMP_LT_OQ);
 485
 486             /* Update potential sum for this i atom from the interaction with this j atom. */
 487             velec            = _mm256_and_pd(velec,cutoff_mask);
 488             velecsum         = _mm256_add_pd(velecsum,velec);
 489
 490             fscal            = felec;
 491
 492             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 493
 494             /* Calculate temporary vectorial force */
 495             tx               = _mm256_mul_pd(fscal,dx23);
 496             ty               = _mm256_mul_pd(fscal,dy23);
 497             tz               = _mm256_mul_pd(fscal,dz23);
 498
 499             /* Update vectorial force */
 500             fix2             = _mm256_add_pd(fix2,tx);
 501             fiy2             = _mm256_add_pd(fiy2,ty);
 502             fiz2             = _mm256_add_pd(fiz2,tz);
 503
 504             fjx3             = _mm256_add_pd(fjx3,tx);
 505             fjy3             = _mm256_add_pd(fjy3,ty);
 506             fjz3             = _mm256_add_pd(fjz3,tz);
 507
 508             }
 509
 510             /**************************
 511              * CALCULATE INTERACTIONS *
 512              **************************/
 513
 514             if (gmx_mm256_any_lt(rsq31,rcutoff2))
 515             {
 516
 517             /* REACTION-FIELD ELECTROSTATICS */
 518             velec            = _mm256_mul_pd(qq31,_mm256_sub_pd(_mm256_add_pd(rinv31,_mm256_mul_pd(krf,rsq31)),crf));
 519             felec            = _mm256_mul_pd(qq31,_mm256_sub_pd(_mm256_mul_pd(rinv31,rinvsq31),krf2));
 520
 521             cutoff_mask      = _mm256_cmp_pd(rsq31,rcutoff2,_CMP_LT_OQ);
 522
 523             /* Update potential sum for this i atom from the interaction with this j atom. */
 524             velec            = _mm256_and_pd(velec,cutoff_mask);
 525             velecsum         = _mm256_add_pd(velecsum,velec);
 526
 527             fscal            = felec;
 528
 529             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 530
 531             /* Calculate temporary vectorial force */
 532             tx               = _mm256_mul_pd(fscal,dx31);
 533             ty               = _mm256_mul_pd(fscal,dy31);
 534             tz               = _mm256_mul_pd(fscal,dz31);
 535
 536             /* Update vectorial force */
 537             fix3             = _mm256_add_pd(fix3,tx);
 538             fiy3             = _mm256_add_pd(fiy3,ty);
 539             fiz3             = _mm256_add_pd(fiz3,tz);
 540
 541             fjx1             = _mm256_add_pd(fjx1,tx);
 542             fjy1             = _mm256_add_pd(fjy1,ty);
 543             fjz1             = _mm256_add_pd(fjz1,tz);
 544
 545             }
 546
 547             /**************************
 548              * CALCULATE INTERACTIONS *
 549              **************************/
 550
 551             if (gmx_mm256_any_lt(rsq32,rcutoff2))
 552             {
 553
 554             /* REACTION-FIELD ELECTROSTATICS */
 555             velec            = _mm256_mul_pd(qq32,_mm256_sub_pd(_mm256_add_pd(rinv32,_mm256_mul_pd(krf,rsq32)),crf));
 556             felec            = _mm256_mul_pd(qq32,_mm256_sub_pd(_mm256_mul_pd(rinv32,rinvsq32),krf2));
 557
 558             cutoff_mask      = _mm256_cmp_pd(rsq32,rcutoff2,_CMP_LT_OQ);
 559
 560             /* Update potential sum for this i atom from the interaction with this j atom. */
 561             velec            = _mm256_and_pd(velec,cutoff_mask);
 562             velecsum         = _mm256_add_pd(velecsum,velec);
 563
 564             fscal            = felec;
 565
 566             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 567
 568             /* Calculate temporary vectorial force */
 569             tx               = _mm256_mul_pd(fscal,dx32);
 570             ty               = _mm256_mul_pd(fscal,dy32);
 571             tz               = _mm256_mul_pd(fscal,dz32);
 572
 573             /* Update vectorial force */
 574             fix3             = _mm256_add_pd(fix3,tx);
 575             fiy3             = _mm256_add_pd(fiy3,ty);
 576             fiz3             = _mm256_add_pd(fiz3,tz);
 577
 578             fjx2             = _mm256_add_pd(fjx2,tx);
 579             fjy2             = _mm256_add_pd(fjy2,ty);
 580             fjz2             = _mm256_add_pd(fjz2,tz);
 581
 582             }
 583
 584             /**************************
 585              * CALCULATE INTERACTIONS *
 586              **************************/
 587
 588             if (gmx_mm256_any_lt(rsq33,rcutoff2))
 589             {
 590
 591             /* REACTION-FIELD ELECTROSTATICS */
 592             velec            = _mm256_mul_pd(qq33,_mm256_sub_pd(_mm256_add_pd(rinv33,_mm256_mul_pd(krf,rsq33)),crf));
 593             felec            = _mm256_mul_pd(qq33,_mm256_sub_pd(_mm256_mul_pd(rinv33,rinvsq33),krf2));
 594
 595             cutoff_mask      = _mm256_cmp_pd(rsq33,rcutoff2,_CMP_LT_OQ);
 596
 597             /* Update potential sum for this i atom from the interaction with this j atom. */
 598             velec            = _mm256_and_pd(velec,cutoff_mask);
 599             velecsum         = _mm256_add_pd(velecsum,velec);
 600
 601             fscal            = felec;
 602
 603             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 604
 605             /* Calculate temporary vectorial force */
 606             tx               = _mm256_mul_pd(fscal,dx33);
 607             ty               = _mm256_mul_pd(fscal,dy33);
 608             tz               = _mm256_mul_pd(fscal,dz33);
 609
 610             /* Update vectorial force */
 611             fix3             = _mm256_add_pd(fix3,tx);
 612             fiy3             = _mm256_add_pd(fiy3,ty);
 613             fiz3             = _mm256_add_pd(fiz3,tz);
 614
 615             fjx3             = _mm256_add_pd(fjx3,tx);
 616             fjy3             = _mm256_add_pd(fjy3,ty);
 617             fjz3             = _mm256_add_pd(fjz3,tz);
 618
 619             }
 620
 621             fjptrA             = f+j_coord_offsetA;
 622             fjptrB             = f+j_coord_offsetB;
 623             fjptrC             = f+j_coord_offsetC;
 624             fjptrD             = f+j_coord_offsetD;
 625
 626             gmx_mm256_decrement_3rvec_4ptr_swizzle_pd(fjptrA+DIM,fjptrB+DIM,fjptrC+DIM,fjptrD+DIM,
 627                                                       fjx1,fjy1,fjz1,fjx2,fjy2,fjz2,fjx3,fjy3,fjz3);
 628
 629             /* Inner loop uses 324 flops */
 630         }
 631
 632         if(jidx<j_index_end)
 633         {
 634
 635             /* Get j neighbor index, and coordinate index */
 636             jnrlistA         = jjnr[jidx];
 637             jnrlistB         = jjnr[jidx+1];
 638             jnrlistC         = jjnr[jidx+2];
 639             jnrlistD         = jjnr[jidx+3];
 640             /* Sign of each element will be negative for non-real atoms.
 641              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
 642              * so use it as val = _mm_andnot_pd(mask,val) to clear dummy entries.
 643              */
 644             tmpmask0 = gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128()));
 645
 646             tmpmask1 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(3,3,2,2));
 647             tmpmask0 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(1,1,0,0));
 648             dummy_mask = _mm256_castps_pd(gmx_mm256_set_m128(tmpmask1,tmpmask0));
 649
 650             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
 651             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
 652             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
 653             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
 654             j_coord_offsetA  = DIM*jnrA;
 655             j_coord_offsetB  = DIM*jnrB;
 656             j_coord_offsetC  = DIM*jnrC;
 657             j_coord_offsetD  = DIM*jnrD;
 658
 659             /* load j atom coordinates */
 660             gmx_mm256_load_3rvec_4ptr_swizzle_pd(x+j_coord_offsetA+DIM,x+j_coord_offsetB+DIM,
 661                                                  x+j_coord_offsetC+DIM,x+j_coord_offsetD+DIM,
 662                                                  &jx1,&jy1,&jz1,&jx2,&jy2,&jz2,&jx3,&jy3,&jz3);
 663
 664             /* Calculate displacement vector */
 665             dx11             = _mm256_sub_pd(ix1,jx1);
 666             dy11             = _mm256_sub_pd(iy1,jy1);
 667             dz11             = _mm256_sub_pd(iz1,jz1);
 668             dx12             = _mm256_sub_pd(ix1,jx2);
 669             dy12             = _mm256_sub_pd(iy1,jy2);
 670             dz12             = _mm256_sub_pd(iz1,jz2);
 671             dx13             = _mm256_sub_pd(ix1,jx3);
 672             dy13             = _mm256_sub_pd(iy1,jy3);
 673             dz13             = _mm256_sub_pd(iz1,jz3);
 674             dx21             = _mm256_sub_pd(ix2,jx1);
 675             dy21             = _mm256_sub_pd(iy2,jy1);
 676             dz21             = _mm256_sub_pd(iz2,jz1);
 677             dx22             = _mm256_sub_pd(ix2,jx2);
 678             dy22             = _mm256_sub_pd(iy2,jy2);
 679             dz22             = _mm256_sub_pd(iz2,jz2);
 680             dx23             = _mm256_sub_pd(ix2,jx3);
 681             dy23             = _mm256_sub_pd(iy2,jy3);
 682             dz23             = _mm256_sub_pd(iz2,jz3);
 683             dx31             = _mm256_sub_pd(ix3,jx1);
 684             dy31             = _mm256_sub_pd(iy3,jy1);
 685             dz31             = _mm256_sub_pd(iz3,jz1);
 686             dx32             = _mm256_sub_pd(ix3,jx2);
 687             dy32             = _mm256_sub_pd(iy3,jy2);
 688             dz32             = _mm256_sub_pd(iz3,jz2);
 689             dx33             = _mm256_sub_pd(ix3,jx3);
 690             dy33             = _mm256_sub_pd(iy3,jy3);
 691             dz33             = _mm256_sub_pd(iz3,jz3);
 692
 693             /* Calculate squared distance and things based on it */
 694             rsq11            = gmx_mm256_calc_rsq_pd(dx11,dy11,dz11);
 695             rsq12            = gmx_mm256_calc_rsq_pd(dx12,dy12,dz12);
 696             rsq13            = gmx_mm256_calc_rsq_pd(dx13,dy13,dz13);
 697             rsq21            = gmx_mm256_calc_rsq_pd(dx21,dy21,dz21);
 698             rsq22            = gmx_mm256_calc_rsq_pd(dx22,dy22,dz22);
 699             rsq23            = gmx_mm256_calc_rsq_pd(dx23,dy23,dz23);
 700             rsq31            = gmx_mm256_calc_rsq_pd(dx31,dy31,dz31);
 701             rsq32            = gmx_mm256_calc_rsq_pd(dx32,dy32,dz32);
 702             rsq33            = gmx_mm256_calc_rsq_pd(dx33,dy33,dz33);
 703
 704             rinv11           = gmx_mm256_invsqrt_pd(rsq11);
 705             rinv12           = gmx_mm256_invsqrt_pd(rsq12);
 706             rinv13           = gmx_mm256_invsqrt_pd(rsq13);
 707             rinv21           = gmx_mm256_invsqrt_pd(rsq21);
 708             rinv22           = gmx_mm256_invsqrt_pd(rsq22);
 709             rinv23           = gmx_mm256_invsqrt_pd(rsq23);
 710             rinv31           = gmx_mm256_invsqrt_pd(rsq31);
 711             rinv32           = gmx_mm256_invsqrt_pd(rsq32);
 712             rinv33           = gmx_mm256_invsqrt_pd(rsq33);
 713
 714             rinvsq11         = _mm256_mul_pd(rinv11,rinv11);
 715             rinvsq12         = _mm256_mul_pd(rinv12,rinv12);
 716             rinvsq13         = _mm256_mul_pd(rinv13,rinv13);
 717             rinvsq21         = _mm256_mul_pd(rinv21,rinv21);
 718             rinvsq22         = _mm256_mul_pd(rinv22,rinv22);
 719             rinvsq23         = _mm256_mul_pd(rinv23,rinv23);
 720             rinvsq31         = _mm256_mul_pd(rinv31,rinv31);
 721             rinvsq32         = _mm256_mul_pd(rinv32,rinv32);
 722             rinvsq33         = _mm256_mul_pd(rinv33,rinv33);
 723
 724             fjx1             = _mm256_setzero_pd();
 725             fjy1             = _mm256_setzero_pd();
 726             fjz1             = _mm256_setzero_pd();
 727             fjx2             = _mm256_setzero_pd();
 728             fjy2             = _mm256_setzero_pd();
 729             fjz2             = _mm256_setzero_pd();
 730             fjx3             = _mm256_setzero_pd();
 731             fjy3             = _mm256_setzero_pd();
 732             fjz3             = _mm256_setzero_pd();
 733
 734             /**************************
 735              * CALCULATE INTERACTIONS *
 736              **************************/
 737
 738             if (gmx_mm256_any_lt(rsq11,rcutoff2))
 739             {
 740
 741             /* REACTION-FIELD ELECTROSTATICS */
 742             velec            = _mm256_mul_pd(qq11,_mm256_sub_pd(_mm256_add_pd(rinv11,_mm256_mul_pd(krf,rsq11)),crf));
 743             felec            = _mm256_mul_pd(qq11,_mm256_sub_pd(_mm256_mul_pd(rinv11,rinvsq11),krf2));
 744
 745             cutoff_mask      = _mm256_cmp_pd(rsq11,rcutoff2,_CMP_LT_OQ);
 746
 747             /* Update potential sum for this i atom from the interaction with this j atom. */
 748             velec            = _mm256_and_pd(velec,cutoff_mask);
 749             velec            = _mm256_andnot_pd(dummy_mask,velec);
 750             velecsum         = _mm256_add_pd(velecsum,velec);
 751
 752             fscal            = felec;
 753
 754             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 755
 756             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
 757
 758             /* Calculate temporary vectorial force */
 759             tx               = _mm256_mul_pd(fscal,dx11);
 760             ty               = _mm256_mul_pd(fscal,dy11);
 761             tz               = _mm256_mul_pd(fscal,dz11);
 762
 763             /* Update vectorial force */
 764             fix1             = _mm256_add_pd(fix1,tx);
 765             fiy1             = _mm256_add_pd(fiy1,ty);
 766             fiz1             = _mm256_add_pd(fiz1,tz);
 767
 768             fjx1             = _mm256_add_pd(fjx1,tx);
 769             fjy1             = _mm256_add_pd(fjy1,ty);
 770             fjz1             = _mm256_add_pd(fjz1,tz);
 771
 772             }
 773
 774             /**************************
 775              * CALCULATE INTERACTIONS *
 776              **************************/
 777
 778             if (gmx_mm256_any_lt(rsq12,rcutoff2))
 779             {
 780
 781             /* REACTION-FIELD ELECTROSTATICS */
 782             velec            = _mm256_mul_pd(qq12,_mm256_sub_pd(_mm256_add_pd(rinv12,_mm256_mul_pd(krf,rsq12)),crf));
 783             felec            = _mm256_mul_pd(qq12,_mm256_sub_pd(_mm256_mul_pd(rinv12,rinvsq12),krf2));
 784
 785             cutoff_mask      = _mm256_cmp_pd(rsq12,rcutoff2,_CMP_LT_OQ);
 786
 787             /* Update potential sum for this i atom from the interaction with this j atom. */
 788             velec            = _mm256_and_pd(velec,cutoff_mask);
 789             velec            = _mm256_andnot_pd(dummy_mask,velec);
 790             velecsum         = _mm256_add_pd(velecsum,velec);
 791
 792             fscal            = felec;
 793
 794             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 795
 796             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
 797
 798             /* Calculate temporary vectorial force */
 799             tx               = _mm256_mul_pd(fscal,dx12);
 800             ty               = _mm256_mul_pd(fscal,dy12);
 801             tz               = _mm256_mul_pd(fscal,dz12);
 802
 803             /* Update vectorial force */
 804             fix1             = _mm256_add_pd(fix1,tx);
 805             fiy1             = _mm256_add_pd(fiy1,ty);
 806             fiz1             = _mm256_add_pd(fiz1,tz);
 807
 808             fjx2             = _mm256_add_pd(fjx2,tx);
 809             fjy2             = _mm256_add_pd(fjy2,ty);
 810             fjz2             = _mm256_add_pd(fjz2,tz);
 811
 812             }
 813
 814             /**************************
 815              * CALCULATE INTERACTIONS *
 816              **************************/
 817
 818             if (gmx_mm256_any_lt(rsq13,rcutoff2))
 819             {
 820
 821             /* REACTION-FIELD ELECTROSTATICS */
 822             velec            = _mm256_mul_pd(qq13,_mm256_sub_pd(_mm256_add_pd(rinv13,_mm256_mul_pd(krf,rsq13)),crf));
 823             felec            = _mm256_mul_pd(qq13,_mm256_sub_pd(_mm256_mul_pd(rinv13,rinvsq13),krf2));
 824
 825             cutoff_mask      = _mm256_cmp_pd(rsq13,rcutoff2,_CMP_LT_OQ);
 826
 827             /* Update potential sum for this i atom from the interaction with this j atom. */
 828             velec            = _mm256_and_pd(velec,cutoff_mask);
 829             velec            = _mm256_andnot_pd(dummy_mask,velec);
 830             velecsum         = _mm256_add_pd(velecsum,velec);
 831
 832             fscal            = felec;
 833
 834             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 835
 836             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
 837
 838             /* Calculate temporary vectorial force */
 839             tx               = _mm256_mul_pd(fscal,dx13);
 840             ty               = _mm256_mul_pd(fscal,dy13);
 841             tz               = _mm256_mul_pd(fscal,dz13);
 842
 843             /* Update vectorial force */
 844             fix1             = _mm256_add_pd(fix1,tx);
 845             fiy1             = _mm256_add_pd(fiy1,ty);
 846             fiz1             = _mm256_add_pd(fiz1,tz);
 847
 848             fjx3             = _mm256_add_pd(fjx3,tx);
 849             fjy3             = _mm256_add_pd(fjy3,ty);
 850             fjz3             = _mm256_add_pd(fjz3,tz);
 851
 852             }
 853
 854             /**************************
 855              * CALCULATE INTERACTIONS *
 856              **************************/
 857
 858             if (gmx_mm256_any_lt(rsq21,rcutoff2))
 859             {
 860
 861             /* REACTION-FIELD ELECTROSTATICS */
 862             velec            = _mm256_mul_pd(qq21,_mm256_sub_pd(_mm256_add_pd(rinv21,_mm256_mul_pd(krf,rsq21)),crf));
 863             felec            = _mm256_mul_pd(qq21,_mm256_sub_pd(_mm256_mul_pd(rinv21,rinvsq21),krf2));
 864
 865             cutoff_mask      = _mm256_cmp_pd(rsq21,rcutoff2,_CMP_LT_OQ);
 866
 867             /* Update potential sum for this i atom from the interaction with this j atom. */
 868             velec            = _mm256_and_pd(velec,cutoff_mask);
 869             velec            = _mm256_andnot_pd(dummy_mask,velec);
 870             velecsum         = _mm256_add_pd(velecsum,velec);
 871
 872             fscal            = felec;
 873
 874             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 875
 876             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
 877
 878             /* Calculate temporary vectorial force */
 879             tx               = _mm256_mul_pd(fscal,dx21);
 880             ty               = _mm256_mul_pd(fscal,dy21);
 881             tz               = _mm256_mul_pd(fscal,dz21);
 882
 883             /* Update vectorial force */
 884             fix2             = _mm256_add_pd(fix2,tx);
 885             fiy2             = _mm256_add_pd(fiy2,ty);
 886             fiz2             = _mm256_add_pd(fiz2,tz);
 887
 888             fjx1             = _mm256_add_pd(fjx1,tx);
 889             fjy1             = _mm256_add_pd(fjy1,ty);
 890             fjz1             = _mm256_add_pd(fjz1,tz);
 891
 892             }
 893
 894             /**************************
 895              * CALCULATE INTERACTIONS *
 896              **************************/
 897
 898             if (gmx_mm256_any_lt(rsq22,rcutoff2))
 899             {
 900
 901             /* REACTION-FIELD ELECTROSTATICS */
 902             velec            = _mm256_mul_pd(qq22,_mm256_sub_pd(_mm256_add_pd(rinv22,_mm256_mul_pd(krf,rsq22)),crf));
 903             felec            = _mm256_mul_pd(qq22,_mm256_sub_pd(_mm256_mul_pd(rinv22,rinvsq22),krf2));
 904
 905             cutoff_mask      = _mm256_cmp_pd(rsq22,rcutoff2,_CMP_LT_OQ);
 906
 907             /* Update potential sum for this i atom from the interaction with this j atom. */
 908             velec            = _mm256_and_pd(velec,cutoff_mask);
 909             velec            = _mm256_andnot_pd(dummy_mask,velec);
 910             velecsum         = _mm256_add_pd(velecsum,velec);
 911
 912             fscal            = felec;
 913
 914             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 915
 916             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
 917
 918             /* Calculate temporary vectorial force */
 919             tx               = _mm256_mul_pd(fscal,dx22);
 920             ty               = _mm256_mul_pd(fscal,dy22);
 921             tz               = _mm256_mul_pd(fscal,dz22);
 922
 923             /* Update vectorial force */
 924             fix2             = _mm256_add_pd(fix2,tx);
 925             fiy2             = _mm256_add_pd(fiy2,ty);
 926             fiz2             = _mm256_add_pd(fiz2,tz);
 927
 928             fjx2             = _mm256_add_pd(fjx2,tx);
 929             fjy2             = _mm256_add_pd(fjy2,ty);
 930             fjz2             = _mm256_add_pd(fjz2,tz);
 931
 932             }
 933
 934             /**************************
 935              * CALCULATE INTERACTIONS *
 936              **************************/
 937
 938             if (gmx_mm256_any_lt(rsq23,rcutoff2))
 939             {
 940
 941             /* REACTION-FIELD ELECTROSTATICS */
 942             velec            = _mm256_mul_pd(qq23,_mm256_sub_pd(_mm256_add_pd(rinv23,_mm256_mul_pd(krf,rsq23)),crf));
 943             felec            = _mm256_mul_pd(qq23,_mm256_sub_pd(_mm256_mul_pd(rinv23,rinvsq23),krf2));
 944
 945             cutoff_mask      = _mm256_cmp_pd(rsq23,rcutoff2,_CMP_LT_OQ);
 946
 947             /* Update potential sum for this i atom from the interaction with this j atom. */
 948             velec            = _mm256_and_pd(velec,cutoff_mask);
 949             velec            = _mm256_andnot_pd(dummy_mask,velec);
 950             velecsum         = _mm256_add_pd(velecsum,velec);
 951
 952             fscal            = felec;
 953
 954             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 955
 956             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
 957
 958             /* Calculate temporary vectorial force */
 959             tx               = _mm256_mul_pd(fscal,dx23);
 960             ty               = _mm256_mul_pd(fscal,dy23);
 961             tz               = _mm256_mul_pd(fscal,dz23);
 962
 963             /* Update vectorial force */
 964             fix2             = _mm256_add_pd(fix2,tx);
 965             fiy2             = _mm256_add_pd(fiy2,ty);
 966             fiz2             = _mm256_add_pd(fiz2,tz);
 967
 968             fjx3             = _mm256_add_pd(fjx3,tx);
 969             fjy3             = _mm256_add_pd(fjy3,ty);
 970             fjz3             = _mm256_add_pd(fjz3,tz);
 971
 972             }
 973
 974             /**************************
 975              * CALCULATE INTERACTIONS *
 976              **************************/
 977
 978             if (gmx_mm256_any_lt(rsq31,rcutoff2))
 979             {
 980
 981             /* REACTION-FIELD ELECTROSTATICS */
 982             velec            = _mm256_mul_pd(qq31,_mm256_sub_pd(_mm256_add_pd(rinv31,_mm256_mul_pd(krf,rsq31)),crf));
 983             felec            = _mm256_mul_pd(qq31,_mm256_sub_pd(_mm256_mul_pd(rinv31,rinvsq31),krf2));
 984
 985             cutoff_mask      = _mm256_cmp_pd(rsq31,rcutoff2,_CMP_LT_OQ);
 986
 987             /* Update potential sum for this i atom from the interaction with this j atom. */
 988             velec            = _mm256_and_pd(velec,cutoff_mask);
 989             velec            = _mm256_andnot_pd(dummy_mask,velec);
 990             velecsum         = _mm256_add_pd(velecsum,velec);
 991
 992             fscal            = felec;
 993
 994             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 995
 996             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
 997
 998             /* Calculate temporary vectorial force */
 999             tx               = _mm256_mul_pd(fscal,dx31);
1000             ty               = _mm256_mul_pd(fscal,dy31);
1001             tz               = _mm256_mul_pd(fscal,dz31);
1002
1003             /* Update vectorial force */
1004             fix3             = _mm256_add_pd(fix3,tx);
1005             fiy3             = _mm256_add_pd(fiy3,ty);
1006             fiz3             = _mm256_add_pd(fiz3,tz);
1007
1008             fjx1             = _mm256_add_pd(fjx1,tx);
1009             fjy1             = _mm256_add_pd(fjy1,ty);
1010             fjz1             = _mm256_add_pd(fjz1,tz);
1011
1012             }
1013
1014             /**************************
1015              * CALCULATE INTERACTIONS *
1016              **************************/
1017
1018             if (gmx_mm256_any_lt(rsq32,rcutoff2))
1019             {
1020
1021             /* REACTION-FIELD ELECTROSTATICS */
1022             velec            = _mm256_mul_pd(qq32,_mm256_sub_pd(_mm256_add_pd(rinv32,_mm256_mul_pd(krf,rsq32)),crf));
1023             felec            = _mm256_mul_pd(qq32,_mm256_sub_pd(_mm256_mul_pd(rinv32,rinvsq32),krf2));
1024
1025             cutoff_mask      = _mm256_cmp_pd(rsq32,rcutoff2,_CMP_LT_OQ);
1026
1027             /* Update potential sum for this i atom from the interaction with this j atom. */
1028             velec            = _mm256_and_pd(velec,cutoff_mask);
1029             velec            = _mm256_andnot_pd(dummy_mask,velec);
1030             velecsum         = _mm256_add_pd(velecsum,velec);
1031
1032             fscal            = felec;
1033
1034             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1035
1036             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
1037
1038             /* Calculate temporary vectorial force */
1039             tx               = _mm256_mul_pd(fscal,dx32);
1040             ty               = _mm256_mul_pd(fscal,dy32);
1041             tz               = _mm256_mul_pd(fscal,dz32);
1042
1043             /* Update vectorial force */
1044             fix3             = _mm256_add_pd(fix3,tx);
1045             fiy3             = _mm256_add_pd(fiy3,ty);
1046             fiz3             = _mm256_add_pd(fiz3,tz);
1047
1048             fjx2             = _mm256_add_pd(fjx2,tx);
1049             fjy2             = _mm256_add_pd(fjy2,ty);
1050             fjz2             = _mm256_add_pd(fjz2,tz);
1051
1052             }
1053
1054             /**************************
1055              * CALCULATE INTERACTIONS *
1056              **************************/
1057
1058             if (gmx_mm256_any_lt(rsq33,rcutoff2))
1059             {
1060
1061             /* REACTION-FIELD ELECTROSTATICS */
1062             velec            = _mm256_mul_pd(qq33,_mm256_sub_pd(_mm256_add_pd(rinv33,_mm256_mul_pd(krf,rsq33)),crf));
1063             felec            = _mm256_mul_pd(qq33,_mm256_sub_pd(_mm256_mul_pd(rinv33,rinvsq33),krf2));
1064
1065             cutoff_mask      = _mm256_cmp_pd(rsq33,rcutoff2,_CMP_LT_OQ);
1066
1067             /* Update potential sum for this i atom from the interaction with this j atom. */
1068             velec            = _mm256_and_pd(velec,cutoff_mask);
1069             velec            = _mm256_andnot_pd(dummy_mask,velec);
1070             velecsum         = _mm256_add_pd(velecsum,velec);
1071
1072             fscal            = felec;
1073
1074             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1075
1076             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
1077
1078             /* Calculate temporary vectorial force */
1079             tx               = _mm256_mul_pd(fscal,dx33);
1080             ty               = _mm256_mul_pd(fscal,dy33);
1081             tz               = _mm256_mul_pd(fscal,dz33);
1082
1083             /* Update vectorial force */
1084             fix3             = _mm256_add_pd(fix3,tx);
1085             fiy3             = _mm256_add_pd(fiy3,ty);
1086             fiz3             = _mm256_add_pd(fiz3,tz);
1087
1088             fjx3             = _mm256_add_pd(fjx3,tx);
1089             fjy3             = _mm256_add_pd(fjy3,ty);
1090             fjz3             = _mm256_add_pd(fjz3,tz);
1091
1092             }
1093
1094             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
1095             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
1096             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
1097             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
1098
1099             gmx_mm256_decrement_3rvec_4ptr_swizzle_pd(fjptrA+DIM,fjptrB+DIM,fjptrC+DIM,fjptrD+DIM,
1100                                                       fjx1,fjy1,fjz1,fjx2,fjy2,fjz2,fjx3,fjy3,fjz3);
1101
1102             /* Inner loop uses 324 flops */
1103         }
1104
1105         /* End of innermost loop */
1106
1107         gmx_mm256_update_iforce_3atom_swizzle_pd(fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
1108                                                  f+i_coord_offset+DIM,fshift+i_shift_offset);
1109
1110         ggid                        = gid[iidx];
1111         /* Update potential energies */
1112         gmx_mm256_update_1pot_pd(velecsum,kernel_data->energygrp_elec+ggid);
1113
1114         /* Increment number of inner iterations */
1115         inneriter                  += j_index_end - j_index_start;
1116
1117         /* Outer loop uses 19 flops */
1118     }
1119
1120     /* Increment number of outer iterations */
1121     outeriter        += nri;
1122
1123     /* Update outer/inner flops */
1124
1125     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_W4W4_VF,outeriter*19 + inneriter*324);
1126 }
1127 /*
1128  * Gromacs nonbonded kernel:   nb_kernel_ElecRFCut_VdwNone_GeomW4W4_F_avx_256_double
1129  * Electrostatics interaction: ReactionField
1130  * VdW interaction:            None
1131  * Geometry:                   Water4-Water4
1132  * Calculate force/pot:        Force
1133  */
1134 void
1135 nb_kernel_ElecRFCut_VdwNone_GeomW4W4_F_avx_256_double
1136                     (t_nblist                    * gmx_restrict       nlist,
1137                      rvec                        * gmx_restrict          xx,
1138                      rvec                        * gmx_restrict          ff,
1139                      t_forcerec                  * gmx_restrict          fr,
1140                      t_mdatoms                   * gmx_restrict     mdatoms,
1141                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
1142                      t_nrnb                      * gmx_restrict        nrnb)
1143 {
1144     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
1145      * just 0 for non-waters.
1146      * Suffixes A,B,C,D refer to j loop unrolling done with AVX, e.g. for the four different
1147      * jnr indices corresponding to data put in the four positions in the SIMD register.
1148      */
1149     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
1150     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
1151     int              jnrA,jnrB,jnrC,jnrD;
1152     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
1153     int              jnrlistE,jnrlistF,jnrlistG,jnrlistH;
1154     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
1155     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
1156     real             rcutoff_scalar;
1157     real             *shiftvec,*fshift,*x,*f;
1158     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD;
1159     real             scratch[4*DIM];
1160     __m256d          tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
1161     real *           vdwioffsetptr1;
1162     __m256d          ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
1163     real *           vdwioffsetptr2;
1164     __m256d          ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
1165     real *           vdwioffsetptr3;
1166     __m256d          ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
1167     int              vdwjidx1A,vdwjidx1B,vdwjidx1C,vdwjidx1D;
1168     __m256d          jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
1169     int              vdwjidx2A,vdwjidx2B,vdwjidx2C,vdwjidx2D;
1170     __m256d          jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
1171     int              vdwjidx3A,vdwjidx3B,vdwjidx3C,vdwjidx3D;
1172     __m256d          jx3,jy3,jz3,fjx3,fjy3,fjz3,jq3,isaj3;
1173     __m256d          dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11;
1174     __m256d          dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12;
1175     __m256d          dx13,dy13,dz13,rsq13,rinv13,rinvsq13,r13,qq13,c6_13,c12_13;
1176     __m256d          dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21;
1177     __m256d          dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22;
1178     __m256d          dx23,dy23,dz23,rsq23,rinv23,rinvsq23,r23,qq23,c6_23,c12_23;
1179     __m256d          dx31,dy31,dz31,rsq31,rinv31,rinvsq31,r31,qq31,c6_31,c12_31;
1180     __m256d          dx32,dy32,dz32,rsq32,rinv32,rinvsq32,r32,qq32,c6_32,c12_32;
1181     __m256d          dx33,dy33,dz33,rsq33,rinv33,rinvsq33,r33,qq33,c6_33,c12_33;
1182     __m256d          velec,felec,velecsum,facel,crf,krf,krf2;
1183     real             *charge;
1184     __m256d          dummy_mask,cutoff_mask;
1185     __m128           tmpmask0,tmpmask1;
1186     __m256d          signbit = _mm256_castsi256_pd( _mm256_set1_epi32(0x80000000) );
1187     __m256d          one     = _mm256_set1_pd(1.0);
1188     __m256d          two     = _mm256_set1_pd(2.0);
1189     x                = xx[0];
1190     f                = ff[0];
1191
1192     nri              = nlist->nri;
1193     iinr             = nlist->iinr;
1194     jindex           = nlist->jindex;
1195     jjnr             = nlist->jjnr;
1196     shiftidx         = nlist->shift;
1197     gid              = nlist->gid;
1198     shiftvec         = fr->shift_vec[0];
1199     fshift           = fr->fshift[0];
1200     facel            = _mm256_set1_pd(fr->epsfac);
1201     charge           = mdatoms->chargeA;
1202     krf              = _mm256_set1_pd(fr->ic->k_rf);
1203     krf2             = _mm256_set1_pd(fr->ic->k_rf*2.0);
1204     crf              = _mm256_set1_pd(fr->ic->c_rf);
1205
1206     /* Setup water-specific parameters */
1207     inr              = nlist->iinr[0];
1208     iq1              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+1]));
1209     iq2              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+2]));
1210     iq3              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+3]));
1211
1212     jq1              = _mm256_set1_pd(charge[inr+1]);
1213     jq2              = _mm256_set1_pd(charge[inr+2]);
1214     jq3              = _mm256_set1_pd(charge[inr+3]);
1215     qq11             = _mm256_mul_pd(iq1,jq1);
1216     qq12             = _mm256_mul_pd(iq1,jq2);
1217     qq13             = _mm256_mul_pd(iq1,jq3);
1218     qq21             = _mm256_mul_pd(iq2,jq1);
1219     qq22             = _mm256_mul_pd(iq2,jq2);
1220     qq23             = _mm256_mul_pd(iq2,jq3);
1221     qq31             = _mm256_mul_pd(iq3,jq1);
1222     qq32             = _mm256_mul_pd(iq3,jq2);
1223     qq33             = _mm256_mul_pd(iq3,jq3);
1224
1225     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
1226     rcutoff_scalar   = fr->rcoulomb;
1227     rcutoff          = _mm256_set1_pd(rcutoff_scalar);
1228     rcutoff2         = _mm256_mul_pd(rcutoff,rcutoff);
1229
1230     /* Avoid stupid compiler warnings */
1231     jnrA = jnrB = jnrC = jnrD = 0;
1232     j_coord_offsetA = 0;
1233     j_coord_offsetB = 0;
1234     j_coord_offsetC = 0;
1235     j_coord_offsetD = 0;
1236
1237     outeriter        = 0;
1238     inneriter        = 0;
1239
1240     for(iidx=0;iidx<4*DIM;iidx++)
1241     {
1242         scratch[iidx] = 0.0;
1243     }
1244
1245     /* Start outer loop over neighborlists */
1246     for(iidx=0; iidx<nri; iidx++)
1247     {
1248         /* Load shift vector for this list */
1249         i_shift_offset   = DIM*shiftidx[iidx];
1250
1251         /* Load limits for loop over neighbors */
1252         j_index_start    = jindex[iidx];
1253         j_index_end      = jindex[iidx+1];
1254
1255         /* Get outer coordinate index */
1256         inr              = iinr[iidx];
1257         i_coord_offset   = DIM*inr;
1258
1259         /* Load i particle coords and add shift vector */
1260         gmx_mm256_load_shift_and_3rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset+DIM,
1261                                                     &ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
1262
1263         fix1             = _mm256_setzero_pd();
1264         fiy1             = _mm256_setzero_pd();
1265         fiz1             = _mm256_setzero_pd();
1266         fix2             = _mm256_setzero_pd();
1267         fiy2             = _mm256_setzero_pd();
1268         fiz2             = _mm256_setzero_pd();
1269         fix3             = _mm256_setzero_pd();
1270         fiy3             = _mm256_setzero_pd();
1271         fiz3             = _mm256_setzero_pd();
1272
1273         /* Start inner kernel loop */
1274         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4)
1275         {
1276
1277             /* Get j neighbor index, and coordinate index */
1278             jnrA             = jjnr[jidx];
1279             jnrB             = jjnr[jidx+1];
1280             jnrC             = jjnr[jidx+2];
1281             jnrD             = jjnr[jidx+3];
1282             j_coord_offsetA  = DIM*jnrA;
1283             j_coord_offsetB  = DIM*jnrB;
1284             j_coord_offsetC  = DIM*jnrC;
1285             j_coord_offsetD  = DIM*jnrD;
1286
1287             /* load j atom coordinates */
1288             gmx_mm256_load_3rvec_4ptr_swizzle_pd(x+j_coord_offsetA+DIM,x+j_coord_offsetB+DIM,
1289                                                  x+j_coord_offsetC+DIM,x+j_coord_offsetD+DIM,
1290                                                  &jx1,&jy1,&jz1,&jx2,&jy2,&jz2,&jx3,&jy3,&jz3);
1291
1292             /* Calculate displacement vector */
1293             dx11             = _mm256_sub_pd(ix1,jx1);
1294             dy11             = _mm256_sub_pd(iy1,jy1);
1295             dz11             = _mm256_sub_pd(iz1,jz1);
1296             dx12             = _mm256_sub_pd(ix1,jx2);
1297             dy12             = _mm256_sub_pd(iy1,jy2);
1298             dz12             = _mm256_sub_pd(iz1,jz2);
1299             dx13             = _mm256_sub_pd(ix1,jx3);
1300             dy13             = _mm256_sub_pd(iy1,jy3);
1301             dz13             = _mm256_sub_pd(iz1,jz3);
1302             dx21             = _mm256_sub_pd(ix2,jx1);
1303             dy21             = _mm256_sub_pd(iy2,jy1);
1304             dz21             = _mm256_sub_pd(iz2,jz1);
1305             dx22             = _mm256_sub_pd(ix2,jx2);
1306             dy22             = _mm256_sub_pd(iy2,jy2);
1307             dz22             = _mm256_sub_pd(iz2,jz2);
1308             dx23             = _mm256_sub_pd(ix2,jx3);
1309             dy23             = _mm256_sub_pd(iy2,jy3);
1310             dz23             = _mm256_sub_pd(iz2,jz3);
1311             dx31             = _mm256_sub_pd(ix3,jx1);
1312             dy31             = _mm256_sub_pd(iy3,jy1);
1313             dz31             = _mm256_sub_pd(iz3,jz1);
1314             dx32             = _mm256_sub_pd(ix3,jx2);
1315             dy32             = _mm256_sub_pd(iy3,jy2);
1316             dz32             = _mm256_sub_pd(iz3,jz2);
1317             dx33             = _mm256_sub_pd(ix3,jx3);
1318             dy33             = _mm256_sub_pd(iy3,jy3);
1319             dz33             = _mm256_sub_pd(iz3,jz3);
1320
1321             /* Calculate squared distance and things based on it */
1322             rsq11            = gmx_mm256_calc_rsq_pd(dx11,dy11,dz11);
1323             rsq12            = gmx_mm256_calc_rsq_pd(dx12,dy12,dz12);
1324             rsq13            = gmx_mm256_calc_rsq_pd(dx13,dy13,dz13);
1325             rsq21            = gmx_mm256_calc_rsq_pd(dx21,dy21,dz21);
1326             rsq22            = gmx_mm256_calc_rsq_pd(dx22,dy22,dz22);
1327             rsq23            = gmx_mm256_calc_rsq_pd(dx23,dy23,dz23);
1328             rsq31            = gmx_mm256_calc_rsq_pd(dx31,dy31,dz31);
1329             rsq32            = gmx_mm256_calc_rsq_pd(dx32,dy32,dz32);
1330             rsq33            = gmx_mm256_calc_rsq_pd(dx33,dy33,dz33);
1331
1332             rinv11           = gmx_mm256_invsqrt_pd(rsq11);
1333             rinv12           = gmx_mm256_invsqrt_pd(rsq12);
1334             rinv13           = gmx_mm256_invsqrt_pd(rsq13);
1335             rinv21           = gmx_mm256_invsqrt_pd(rsq21);
1336             rinv22           = gmx_mm256_invsqrt_pd(rsq22);
1337             rinv23           = gmx_mm256_invsqrt_pd(rsq23);
1338             rinv31           = gmx_mm256_invsqrt_pd(rsq31);
1339             rinv32           = gmx_mm256_invsqrt_pd(rsq32);
1340             rinv33           = gmx_mm256_invsqrt_pd(rsq33);
1341
1342             rinvsq11         = _mm256_mul_pd(rinv11,rinv11);
1343             rinvsq12         = _mm256_mul_pd(rinv12,rinv12);
1344             rinvsq13         = _mm256_mul_pd(rinv13,rinv13);
1345             rinvsq21         = _mm256_mul_pd(rinv21,rinv21);
1346             rinvsq22         = _mm256_mul_pd(rinv22,rinv22);
1347             rinvsq23         = _mm256_mul_pd(rinv23,rinv23);
1348             rinvsq31         = _mm256_mul_pd(rinv31,rinv31);
1349             rinvsq32         = _mm256_mul_pd(rinv32,rinv32);
1350             rinvsq33         = _mm256_mul_pd(rinv33,rinv33);
1351
1352             fjx1             = _mm256_setzero_pd();
1353             fjy1             = _mm256_setzero_pd();
1354             fjz1             = _mm256_setzero_pd();
1355             fjx2             = _mm256_setzero_pd();
1356             fjy2             = _mm256_setzero_pd();
1357             fjz2             = _mm256_setzero_pd();
1358             fjx3             = _mm256_setzero_pd();
1359             fjy3             = _mm256_setzero_pd();
1360             fjz3             = _mm256_setzero_pd();
1361
1362             /**************************
1363              * CALCULATE INTERACTIONS *
1364              **************************/
1365
1366             if (gmx_mm256_any_lt(rsq11,rcutoff2))
1367             {
1368
1369             /* REACTION-FIELD ELECTROSTATICS */
1370             felec            = _mm256_mul_pd(qq11,_mm256_sub_pd(_mm256_mul_pd(rinv11,rinvsq11),krf2));
1371
1372             cutoff_mask      = _mm256_cmp_pd(rsq11,rcutoff2,_CMP_LT_OQ);
1373
1374             fscal            = felec;
1375
1376             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1377
1378             /* Calculate temporary vectorial force */
1379             tx               = _mm256_mul_pd(fscal,dx11);
1380             ty               = _mm256_mul_pd(fscal,dy11);
1381             tz               = _mm256_mul_pd(fscal,dz11);
1382
1383             /* Update vectorial force */
1384             fix1             = _mm256_add_pd(fix1,tx);
1385             fiy1             = _mm256_add_pd(fiy1,ty);
1386             fiz1             = _mm256_add_pd(fiz1,tz);
1387
1388             fjx1             = _mm256_add_pd(fjx1,tx);
1389             fjy1             = _mm256_add_pd(fjy1,ty);
1390             fjz1             = _mm256_add_pd(fjz1,tz);
1391
1392             }
1393
1394             /**************************
1395              * CALCULATE INTERACTIONS *
1396              **************************/
1397
1398             if (gmx_mm256_any_lt(rsq12,rcutoff2))
1399             {
1400
1401             /* REACTION-FIELD ELECTROSTATICS */
1402             felec            = _mm256_mul_pd(qq12,_mm256_sub_pd(_mm256_mul_pd(rinv12,rinvsq12),krf2));
1403
1404             cutoff_mask      = _mm256_cmp_pd(rsq12,rcutoff2,_CMP_LT_OQ);
1405
1406             fscal            = felec;
1407
1408             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1409
1410             /* Calculate temporary vectorial force */
1411             tx               = _mm256_mul_pd(fscal,dx12);
1412             ty               = _mm256_mul_pd(fscal,dy12);
1413             tz               = _mm256_mul_pd(fscal,dz12);
1414
1415             /* Update vectorial force */
1416             fix1             = _mm256_add_pd(fix1,tx);
1417             fiy1             = _mm256_add_pd(fiy1,ty);
1418             fiz1             = _mm256_add_pd(fiz1,tz);
1419
1420             fjx2             = _mm256_add_pd(fjx2,tx);
1421             fjy2             = _mm256_add_pd(fjy2,ty);
1422             fjz2             = _mm256_add_pd(fjz2,tz);
1423
1424             }
1425
1426             /**************************
1427              * CALCULATE INTERACTIONS *
1428              **************************/
1429
1430             if (gmx_mm256_any_lt(rsq13,rcutoff2))
1431             {
1432
1433             /* REACTION-FIELD ELECTROSTATICS */
1434             felec            = _mm256_mul_pd(qq13,_mm256_sub_pd(_mm256_mul_pd(rinv13,rinvsq13),krf2));
1435
1436             cutoff_mask      = _mm256_cmp_pd(rsq13,rcutoff2,_CMP_LT_OQ);
1437
1438             fscal            = felec;
1439
1440             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1441
1442             /* Calculate temporary vectorial force */
1443             tx               = _mm256_mul_pd(fscal,dx13);
1444             ty               = _mm256_mul_pd(fscal,dy13);
1445             tz               = _mm256_mul_pd(fscal,dz13);
1446
1447             /* Update vectorial force */
1448             fix1             = _mm256_add_pd(fix1,tx);
1449             fiy1             = _mm256_add_pd(fiy1,ty);
1450             fiz1             = _mm256_add_pd(fiz1,tz);
1451
1452             fjx3             = _mm256_add_pd(fjx3,tx);
1453             fjy3             = _mm256_add_pd(fjy3,ty);
1454             fjz3             = _mm256_add_pd(fjz3,tz);
1455
1456             }
1457
1458             /**************************
1459              * CALCULATE INTERACTIONS *
1460              **************************/
1461
1462             if (gmx_mm256_any_lt(rsq21,rcutoff2))
1463             {
1464
1465             /* REACTION-FIELD ELECTROSTATICS */
1466             felec            = _mm256_mul_pd(qq21,_mm256_sub_pd(_mm256_mul_pd(rinv21,rinvsq21),krf2));
1467
1468             cutoff_mask      = _mm256_cmp_pd(rsq21,rcutoff2,_CMP_LT_OQ);
1469
1470             fscal            = felec;
1471
1472             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1473
1474             /* Calculate temporary vectorial force */
1475             tx               = _mm256_mul_pd(fscal,dx21);
1476             ty               = _mm256_mul_pd(fscal,dy21);
1477             tz               = _mm256_mul_pd(fscal,dz21);
1478
1479             /* Update vectorial force */
1480             fix2             = _mm256_add_pd(fix2,tx);
1481             fiy2             = _mm256_add_pd(fiy2,ty);
1482             fiz2             = _mm256_add_pd(fiz2,tz);
1483
1484             fjx1             = _mm256_add_pd(fjx1,tx);
1485             fjy1             = _mm256_add_pd(fjy1,ty);
1486             fjz1             = _mm256_add_pd(fjz1,tz);
1487
1488             }
1489
1490             /**************************
1491              * CALCULATE INTERACTIONS *
1492              **************************/
1493
1494             if (gmx_mm256_any_lt(rsq22,rcutoff2))
1495             {
1496
1497             /* REACTION-FIELD ELECTROSTATICS */
1498             felec            = _mm256_mul_pd(qq22,_mm256_sub_pd(_mm256_mul_pd(rinv22,rinvsq22),krf2));
1499
1500             cutoff_mask      = _mm256_cmp_pd(rsq22,rcutoff2,_CMP_LT_OQ);
1501
1502             fscal            = felec;
1503
1504             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1505
1506             /* Calculate temporary vectorial force */
1507             tx               = _mm256_mul_pd(fscal,dx22);
1508             ty               = _mm256_mul_pd(fscal,dy22);
1509             tz               = _mm256_mul_pd(fscal,dz22);
1510
1511             /* Update vectorial force */
1512             fix2             = _mm256_add_pd(fix2,tx);
1513             fiy2             = _mm256_add_pd(fiy2,ty);
1514             fiz2             = _mm256_add_pd(fiz2,tz);
1515
1516             fjx2             = _mm256_add_pd(fjx2,tx);
1517             fjy2             = _mm256_add_pd(fjy2,ty);
1518             fjz2             = _mm256_add_pd(fjz2,tz);
1519
1520             }
1521
1522             /**************************
1523              * CALCULATE INTERACTIONS *
1524              **************************/
1525
1526             if (gmx_mm256_any_lt(rsq23,rcutoff2))
1527             {
1528
1529             /* REACTION-FIELD ELECTROSTATICS */
1530             felec            = _mm256_mul_pd(qq23,_mm256_sub_pd(_mm256_mul_pd(rinv23,rinvsq23),krf2));
1531
1532             cutoff_mask      = _mm256_cmp_pd(rsq23,rcutoff2,_CMP_LT_OQ);
1533
1534             fscal            = felec;
1535
1536             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1537
1538             /* Calculate temporary vectorial force */
1539             tx               = _mm256_mul_pd(fscal,dx23);
1540             ty               = _mm256_mul_pd(fscal,dy23);
1541             tz               = _mm256_mul_pd(fscal,dz23);
1542
1543             /* Update vectorial force */
1544             fix2             = _mm256_add_pd(fix2,tx);
1545             fiy2             = _mm256_add_pd(fiy2,ty);
1546             fiz2             = _mm256_add_pd(fiz2,tz);
1547
1548             fjx3             = _mm256_add_pd(fjx3,tx);
1549             fjy3             = _mm256_add_pd(fjy3,ty);
1550             fjz3             = _mm256_add_pd(fjz3,tz);
1551
1552             }
1553
1554             /**************************
1555              * CALCULATE INTERACTIONS *
1556              **************************/
1557
1558             if (gmx_mm256_any_lt(rsq31,rcutoff2))
1559             {
1560
1561             /* REACTION-FIELD ELECTROSTATICS */
1562             felec            = _mm256_mul_pd(qq31,_mm256_sub_pd(_mm256_mul_pd(rinv31,rinvsq31),krf2));
1563
1564             cutoff_mask      = _mm256_cmp_pd(rsq31,rcutoff2,_CMP_LT_OQ);
1565
1566             fscal            = felec;
1567
1568             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1569
1570             /* Calculate temporary vectorial force */
1571             tx               = _mm256_mul_pd(fscal,dx31);
1572             ty               = _mm256_mul_pd(fscal,dy31);
1573             tz               = _mm256_mul_pd(fscal,dz31);
1574
1575             /* Update vectorial force */
1576             fix3             = _mm256_add_pd(fix3,tx);
1577             fiy3             = _mm256_add_pd(fiy3,ty);
1578             fiz3             = _mm256_add_pd(fiz3,tz);
1579
1580             fjx1             = _mm256_add_pd(fjx1,tx);
1581             fjy1             = _mm256_add_pd(fjy1,ty);
1582             fjz1             = _mm256_add_pd(fjz1,tz);
1583
1584             }
1585
1586             /**************************
1587              * CALCULATE INTERACTIONS *
1588              **************************/
1589
1590             if (gmx_mm256_any_lt(rsq32,rcutoff2))
1591             {
1592
1593             /* REACTION-FIELD ELECTROSTATICS */
1594             felec            = _mm256_mul_pd(qq32,_mm256_sub_pd(_mm256_mul_pd(rinv32,rinvsq32),krf2));
1595
1596             cutoff_mask      = _mm256_cmp_pd(rsq32,rcutoff2,_CMP_LT_OQ);
1597
1598             fscal            = felec;
1599
1600             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1601
1602             /* Calculate temporary vectorial force */
1603             tx               = _mm256_mul_pd(fscal,dx32);
1604             ty               = _mm256_mul_pd(fscal,dy32);
1605             tz               = _mm256_mul_pd(fscal,dz32);
1606
1607             /* Update vectorial force */
1608             fix3             = _mm256_add_pd(fix3,tx);
1609             fiy3             = _mm256_add_pd(fiy3,ty);
1610             fiz3             = _mm256_add_pd(fiz3,tz);
1611
1612             fjx2             = _mm256_add_pd(fjx2,tx);
1613             fjy2             = _mm256_add_pd(fjy2,ty);
1614             fjz2             = _mm256_add_pd(fjz2,tz);
1615
1616             }
1617
1618             /**************************
1619              * CALCULATE INTERACTIONS *
1620              **************************/
1621
1622             if (gmx_mm256_any_lt(rsq33,rcutoff2))
1623             {
1624
1625             /* REACTION-FIELD ELECTROSTATICS */
1626             felec            = _mm256_mul_pd(qq33,_mm256_sub_pd(_mm256_mul_pd(rinv33,rinvsq33),krf2));
1627
1628             cutoff_mask      = _mm256_cmp_pd(rsq33,rcutoff2,_CMP_LT_OQ);
1629
1630             fscal            = felec;
1631
1632             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1633
1634             /* Calculate temporary vectorial force */
1635             tx               = _mm256_mul_pd(fscal,dx33);
1636             ty               = _mm256_mul_pd(fscal,dy33);
1637             tz               = _mm256_mul_pd(fscal,dz33);
1638
1639             /* Update vectorial force */
1640             fix3             = _mm256_add_pd(fix3,tx);
1641             fiy3             = _mm256_add_pd(fiy3,ty);
1642             fiz3             = _mm256_add_pd(fiz3,tz);
1643
1644             fjx3             = _mm256_add_pd(fjx3,tx);
1645             fjy3             = _mm256_add_pd(fjy3,ty);
1646             fjz3             = _mm256_add_pd(fjz3,tz);
1647
1648             }
1649
1650             fjptrA             = f+j_coord_offsetA;
1651             fjptrB             = f+j_coord_offsetB;
1652             fjptrC             = f+j_coord_offsetC;
1653             fjptrD             = f+j_coord_offsetD;
1654
1655             gmx_mm256_decrement_3rvec_4ptr_swizzle_pd(fjptrA+DIM,fjptrB+DIM,fjptrC+DIM,fjptrD+DIM,
1656                                                       fjx1,fjy1,fjz1,fjx2,fjy2,fjz2,fjx3,fjy3,fjz3);
1657
1658             /* Inner loop uses 270 flops */
1659         }
1660
1661         if(jidx<j_index_end)
1662         {
1663
1664             /* Get j neighbor index, and coordinate index */
1665             jnrlistA         = jjnr[jidx];
1666             jnrlistB         = jjnr[jidx+1];
1667             jnrlistC         = jjnr[jidx+2];
1668             jnrlistD         = jjnr[jidx+3];
1669             /* Sign of each element will be negative for non-real atoms.
1670              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
1671              * so use it as val = _mm_andnot_pd(mask,val) to clear dummy entries.
1672              */
1673             tmpmask0 = gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128()));
1674
1675             tmpmask1 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(3,3,2,2));
1676             tmpmask0 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(1,1,0,0));
1677             dummy_mask = _mm256_castps_pd(gmx_mm256_set_m128(tmpmask1,tmpmask0));
1678
1679             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
1680             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
1681             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
1682             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
1683             j_coord_offsetA  = DIM*jnrA;
1684             j_coord_offsetB  = DIM*jnrB;
1685             j_coord_offsetC  = DIM*jnrC;
1686             j_coord_offsetD  = DIM*jnrD;
1687
1688             /* load j atom coordinates */
1689             gmx_mm256_load_3rvec_4ptr_swizzle_pd(x+j_coord_offsetA+DIM,x+j_coord_offsetB+DIM,
1690                                                  x+j_coord_offsetC+DIM,x+j_coord_offsetD+DIM,
1691                                                  &jx1,&jy1,&jz1,&jx2,&jy2,&jz2,&jx3,&jy3,&jz3);
1692
1693             /* Calculate displacement vector */
1694             dx11             = _mm256_sub_pd(ix1,jx1);
1695             dy11             = _mm256_sub_pd(iy1,jy1);
1696             dz11             = _mm256_sub_pd(iz1,jz1);
1697             dx12             = _mm256_sub_pd(ix1,jx2);
1698             dy12             = _mm256_sub_pd(iy1,jy2);
1699             dz12             = _mm256_sub_pd(iz1,jz2);
1700             dx13             = _mm256_sub_pd(ix1,jx3);
1701             dy13             = _mm256_sub_pd(iy1,jy3);
1702             dz13             = _mm256_sub_pd(iz1,jz3);
1703             dx21             = _mm256_sub_pd(ix2,jx1);
1704             dy21             = _mm256_sub_pd(iy2,jy1);
1705             dz21             = _mm256_sub_pd(iz2,jz1);
1706             dx22             = _mm256_sub_pd(ix2,jx2);
1707             dy22             = _mm256_sub_pd(iy2,jy2);
1708             dz22             = _mm256_sub_pd(iz2,jz2);
1709             dx23             = _mm256_sub_pd(ix2,jx3);
1710             dy23             = _mm256_sub_pd(iy2,jy3);
1711             dz23             = _mm256_sub_pd(iz2,jz3);
1712             dx31             = _mm256_sub_pd(ix3,jx1);
1713             dy31             = _mm256_sub_pd(iy3,jy1);
1714             dz31             = _mm256_sub_pd(iz3,jz1);
1715             dx32             = _mm256_sub_pd(ix3,jx2);
1716             dy32             = _mm256_sub_pd(iy3,jy2);
1717             dz32             = _mm256_sub_pd(iz3,jz2);
1718             dx33             = _mm256_sub_pd(ix3,jx3);
1719             dy33             = _mm256_sub_pd(iy3,jy3);
1720             dz33             = _mm256_sub_pd(iz3,jz3);
1721
1722             /* Calculate squared distance and things based on it */
1723             rsq11            = gmx_mm256_calc_rsq_pd(dx11,dy11,dz11);
1724             rsq12            = gmx_mm256_calc_rsq_pd(dx12,dy12,dz12);
1725             rsq13            = gmx_mm256_calc_rsq_pd(dx13,dy13,dz13);
1726             rsq21            = gmx_mm256_calc_rsq_pd(dx21,dy21,dz21);
1727             rsq22            = gmx_mm256_calc_rsq_pd(dx22,dy22,dz22);
1728             rsq23            = gmx_mm256_calc_rsq_pd(dx23,dy23,dz23);
1729             rsq31            = gmx_mm256_calc_rsq_pd(dx31,dy31,dz31);
1730             rsq32            = gmx_mm256_calc_rsq_pd(dx32,dy32,dz32);
1731             rsq33            = gmx_mm256_calc_rsq_pd(dx33,dy33,dz33);
1732
1733             rinv11           = gmx_mm256_invsqrt_pd(rsq11);
1734             rinv12           = gmx_mm256_invsqrt_pd(rsq12);
1735             rinv13           = gmx_mm256_invsqrt_pd(rsq13);
1736             rinv21           = gmx_mm256_invsqrt_pd(rsq21);
1737             rinv22           = gmx_mm256_invsqrt_pd(rsq22);
1738             rinv23           = gmx_mm256_invsqrt_pd(rsq23);
1739             rinv31           = gmx_mm256_invsqrt_pd(rsq31);
1740             rinv32           = gmx_mm256_invsqrt_pd(rsq32);
1741             rinv33           = gmx_mm256_invsqrt_pd(rsq33);
1742
1743             rinvsq11         = _mm256_mul_pd(rinv11,rinv11);
1744             rinvsq12         = _mm256_mul_pd(rinv12,rinv12);
1745             rinvsq13         = _mm256_mul_pd(rinv13,rinv13);
1746             rinvsq21         = _mm256_mul_pd(rinv21,rinv21);
1747             rinvsq22         = _mm256_mul_pd(rinv22,rinv22);
1748             rinvsq23         = _mm256_mul_pd(rinv23,rinv23);
1749             rinvsq31         = _mm256_mul_pd(rinv31,rinv31);
1750             rinvsq32         = _mm256_mul_pd(rinv32,rinv32);
1751             rinvsq33         = _mm256_mul_pd(rinv33,rinv33);
1752
1753             fjx1             = _mm256_setzero_pd();
1754             fjy1             = _mm256_setzero_pd();
1755             fjz1             = _mm256_setzero_pd();
1756             fjx2             = _mm256_setzero_pd();
1757             fjy2             = _mm256_setzero_pd();
1758             fjz2             = _mm256_setzero_pd();
1759             fjx3             = _mm256_setzero_pd();
1760             fjy3             = _mm256_setzero_pd();
1761             fjz3             = _mm256_setzero_pd();
1762
1763             /**************************
1764              * CALCULATE INTERACTIONS *
1765              **************************/
1766
1767             if (gmx_mm256_any_lt(rsq11,rcutoff2))
1768             {
1769
1770             /* REACTION-FIELD ELECTROSTATICS */
1771             felec            = _mm256_mul_pd(qq11,_mm256_sub_pd(_mm256_mul_pd(rinv11,rinvsq11),krf2));
1772
1773             cutoff_mask      = _mm256_cmp_pd(rsq11,rcutoff2,_CMP_LT_OQ);
1774
1775             fscal            = felec;
1776
1777             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1778
1779             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
1780
1781             /* Calculate temporary vectorial force */
1782             tx               = _mm256_mul_pd(fscal,dx11);
1783             ty               = _mm256_mul_pd(fscal,dy11);
1784             tz               = _mm256_mul_pd(fscal,dz11);
1785
1786             /* Update vectorial force */
1787             fix1             = _mm256_add_pd(fix1,tx);
1788             fiy1             = _mm256_add_pd(fiy1,ty);
1789             fiz1             = _mm256_add_pd(fiz1,tz);
1790
1791             fjx1             = _mm256_add_pd(fjx1,tx);
1792             fjy1             = _mm256_add_pd(fjy1,ty);
1793             fjz1             = _mm256_add_pd(fjz1,tz);
1794
1795             }
1796
1797             /**************************
1798              * CALCULATE INTERACTIONS *
1799              **************************/
1800
1801             if (gmx_mm256_any_lt(rsq12,rcutoff2))
1802             {
1803
1804             /* REACTION-FIELD ELECTROSTATICS */
1805             felec            = _mm256_mul_pd(qq12,_mm256_sub_pd(_mm256_mul_pd(rinv12,rinvsq12),krf2));
1806
1807             cutoff_mask      = _mm256_cmp_pd(rsq12,rcutoff2,_CMP_LT_OQ);
1808
1809             fscal            = felec;
1810
1811             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1812
1813             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
1814
1815             /* Calculate temporary vectorial force */
1816             tx               = _mm256_mul_pd(fscal,dx12);
1817             ty               = _mm256_mul_pd(fscal,dy12);
1818             tz               = _mm256_mul_pd(fscal,dz12);
1819
1820             /* Update vectorial force */
1821             fix1             = _mm256_add_pd(fix1,tx);
1822             fiy1             = _mm256_add_pd(fiy1,ty);
1823             fiz1             = _mm256_add_pd(fiz1,tz);
1824
1825             fjx2             = _mm256_add_pd(fjx2,tx);
1826             fjy2             = _mm256_add_pd(fjy2,ty);
1827             fjz2             = _mm256_add_pd(fjz2,tz);
1828
1829             }
1830
1831             /**************************
1832              * CALCULATE INTERACTIONS *
1833              **************************/
1834
1835             if (gmx_mm256_any_lt(rsq13,rcutoff2))
1836             {
1837
1838             /* REACTION-FIELD ELECTROSTATICS */
1839             felec            = _mm256_mul_pd(qq13,_mm256_sub_pd(_mm256_mul_pd(rinv13,rinvsq13),krf2));
1840
1841             cutoff_mask      = _mm256_cmp_pd(rsq13,rcutoff2,_CMP_LT_OQ);
1842
1843             fscal            = felec;
1844
1845             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1846
1847             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
1848
1849             /* Calculate temporary vectorial force */
1850             tx               = _mm256_mul_pd(fscal,dx13);
1851             ty               = _mm256_mul_pd(fscal,dy13);
1852             tz               = _mm256_mul_pd(fscal,dz13);
1853
1854             /* Update vectorial force */
1855             fix1             = _mm256_add_pd(fix1,tx);
1856             fiy1             = _mm256_add_pd(fiy1,ty);
1857             fiz1             = _mm256_add_pd(fiz1,tz);
1858
1859             fjx3             = _mm256_add_pd(fjx3,tx);
1860             fjy3             = _mm256_add_pd(fjy3,ty);
1861             fjz3             = _mm256_add_pd(fjz3,tz);
1862
1863             }
1864
1865             /**************************
1866              * CALCULATE INTERACTIONS *
1867              **************************/
1868
1869             if (gmx_mm256_any_lt(rsq21,rcutoff2))
1870             {
1871
1872             /* REACTION-FIELD ELECTROSTATICS */
1873             felec            = _mm256_mul_pd(qq21,_mm256_sub_pd(_mm256_mul_pd(rinv21,rinvsq21),krf2));
1874
1875             cutoff_mask      = _mm256_cmp_pd(rsq21,rcutoff2,_CMP_LT_OQ);
1876
1877             fscal            = felec;
1878
1879             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1880
1881             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
1882
1883             /* Calculate temporary vectorial force */
1884             tx               = _mm256_mul_pd(fscal,dx21);
1885             ty               = _mm256_mul_pd(fscal,dy21);
1886             tz               = _mm256_mul_pd(fscal,dz21);
1887
1888             /* Update vectorial force */
1889             fix2             = _mm256_add_pd(fix2,tx);
1890             fiy2             = _mm256_add_pd(fiy2,ty);
1891             fiz2             = _mm256_add_pd(fiz2,tz);
1892
1893             fjx1             = _mm256_add_pd(fjx1,tx);
1894             fjy1             = _mm256_add_pd(fjy1,ty);
1895             fjz1             = _mm256_add_pd(fjz1,tz);
1896
1897             }
1898
1899             /**************************
1900              * CALCULATE INTERACTIONS *
1901              **************************/
1902
1903             if (gmx_mm256_any_lt(rsq22,rcutoff2))
1904             {
1905
1906             /* REACTION-FIELD ELECTROSTATICS */
1907             felec            = _mm256_mul_pd(qq22,_mm256_sub_pd(_mm256_mul_pd(rinv22,rinvsq22),krf2));
1908
1909             cutoff_mask      = _mm256_cmp_pd(rsq22,rcutoff2,_CMP_LT_OQ);
1910
1911             fscal            = felec;
1912
1913             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1914
1915             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
1916
1917             /* Calculate temporary vectorial force */
1918             tx               = _mm256_mul_pd(fscal,dx22);
1919             ty               = _mm256_mul_pd(fscal,dy22);
1920             tz               = _mm256_mul_pd(fscal,dz22);
1921
1922             /* Update vectorial force */
1923             fix2             = _mm256_add_pd(fix2,tx);
1924             fiy2             = _mm256_add_pd(fiy2,ty);
1925             fiz2             = _mm256_add_pd(fiz2,tz);
1926
1927             fjx2             = _mm256_add_pd(fjx2,tx);
1928             fjy2             = _mm256_add_pd(fjy2,ty);
1929             fjz2             = _mm256_add_pd(fjz2,tz);
1930
1931             }
1932
1933             /**************************
1934              * CALCULATE INTERACTIONS *
1935              **************************/
1936
1937             if (gmx_mm256_any_lt(rsq23,rcutoff2))
1938             {
1939
1940             /* REACTION-FIELD ELECTROSTATICS */
1941             felec            = _mm256_mul_pd(qq23,_mm256_sub_pd(_mm256_mul_pd(rinv23,rinvsq23),krf2));
1942
1943             cutoff_mask      = _mm256_cmp_pd(rsq23,rcutoff2,_CMP_LT_OQ);
1944
1945             fscal            = felec;
1946
1947             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1948
1949             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
1950
1951             /* Calculate temporary vectorial force */
1952             tx               = _mm256_mul_pd(fscal,dx23);
1953             ty               = _mm256_mul_pd(fscal,dy23);
1954             tz               = _mm256_mul_pd(fscal,dz23);
1955
1956             /* Update vectorial force */
1957             fix2             = _mm256_add_pd(fix2,tx);
1958             fiy2             = _mm256_add_pd(fiy2,ty);
1959             fiz2             = _mm256_add_pd(fiz2,tz);
1960
1961             fjx3             = _mm256_add_pd(fjx3,tx);
1962             fjy3             = _mm256_add_pd(fjy3,ty);
1963             fjz3             = _mm256_add_pd(fjz3,tz);
1964
1965             }
1966
1967             /**************************
1968              * CALCULATE INTERACTIONS *
1969              **************************/
1970
1971             if (gmx_mm256_any_lt(rsq31,rcutoff2))
1972             {
1973
1974             /* REACTION-FIELD ELECTROSTATICS */
1975             felec            = _mm256_mul_pd(qq31,_mm256_sub_pd(_mm256_mul_pd(rinv31,rinvsq31),krf2));
1976
1977             cutoff_mask      = _mm256_cmp_pd(rsq31,rcutoff2,_CMP_LT_OQ);
1978
1979             fscal            = felec;
1980
1981             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1982
1983             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
1984
1985             /* Calculate temporary vectorial force */
1986             tx               = _mm256_mul_pd(fscal,dx31);
1987             ty               = _mm256_mul_pd(fscal,dy31);
1988             tz               = _mm256_mul_pd(fscal,dz31);
1989
1990             /* Update vectorial force */
1991             fix3             = _mm256_add_pd(fix3,tx);
1992             fiy3             = _mm256_add_pd(fiy3,ty);
1993             fiz3             = _mm256_add_pd(fiz3,tz);
1994
1995             fjx1             = _mm256_add_pd(fjx1,tx);
1996             fjy1             = _mm256_add_pd(fjy1,ty);
1997             fjz1             = _mm256_add_pd(fjz1,tz);
1998
1999             }
2000
2001             /**************************
2002              * CALCULATE INTERACTIONS *
2003              **************************/
2004
2005             if (gmx_mm256_any_lt(rsq32,rcutoff2))
2006             {
2007
2008             /* REACTION-FIELD ELECTROSTATICS */
2009             felec            = _mm256_mul_pd(qq32,_mm256_sub_pd(_mm256_mul_pd(rinv32,rinvsq32),krf2));
2010
2011             cutoff_mask      = _mm256_cmp_pd(rsq32,rcutoff2,_CMP_LT_OQ);
2012
2013             fscal            = felec;
2014
2015             fscal            = _mm256_and_pd(fscal,cutoff_mask);
2016
2017             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
2018
2019             /* Calculate temporary vectorial force */
2020             tx               = _mm256_mul_pd(fscal,dx32);
2021             ty               = _mm256_mul_pd(fscal,dy32);
2022             tz               = _mm256_mul_pd(fscal,dz32);
2023
2024             /* Update vectorial force */
2025             fix3             = _mm256_add_pd(fix3,tx);
2026             fiy3             = _mm256_add_pd(fiy3,ty);
2027             fiz3             = _mm256_add_pd(fiz3,tz);
2028
2029             fjx2             = _mm256_add_pd(fjx2,tx);
2030             fjy2             = _mm256_add_pd(fjy2,ty);
2031             fjz2             = _mm256_add_pd(fjz2,tz);
2032
2033             }
2034
2035             /**************************
2036              * CALCULATE INTERACTIONS *
2037              **************************/
2038
2039             if (gmx_mm256_any_lt(rsq33,rcutoff2))
2040             {
2041
2042             /* REACTION-FIELD ELECTROSTATICS */
2043             felec            = _mm256_mul_pd(qq33,_mm256_sub_pd(_mm256_mul_pd(rinv33,rinvsq33),krf2));
2044
2045             cutoff_mask      = _mm256_cmp_pd(rsq33,rcutoff2,_CMP_LT_OQ);
2046
2047             fscal            = felec;
2048
2049             fscal            = _mm256_and_pd(fscal,cutoff_mask);
2050
2051             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
2052
2053             /* Calculate temporary vectorial force */
2054             tx               = _mm256_mul_pd(fscal,dx33);
2055             ty               = _mm256_mul_pd(fscal,dy33);
2056             tz               = _mm256_mul_pd(fscal,dz33);
2057
2058             /* Update vectorial force */
2059             fix3             = _mm256_add_pd(fix3,tx);
2060             fiy3             = _mm256_add_pd(fiy3,ty);
2061             fiz3             = _mm256_add_pd(fiz3,tz);
2062
2063             fjx3             = _mm256_add_pd(fjx3,tx);
2064             fjy3             = _mm256_add_pd(fjy3,ty);
2065             fjz3             = _mm256_add_pd(fjz3,tz);
2066
2067             }
2068
2069             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
2070             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
2071             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
2072             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
2073
2074             gmx_mm256_decrement_3rvec_4ptr_swizzle_pd(fjptrA+DIM,fjptrB+DIM,fjptrC+DIM,fjptrD+DIM,
2075                                                       fjx1,fjy1,fjz1,fjx2,fjy2,fjz2,fjx3,fjy3,fjz3);
2076
2077             /* Inner loop uses 270 flops */
2078         }
2079
2080         /* End of innermost loop */
2081
2082         gmx_mm256_update_iforce_3atom_swizzle_pd(fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
2083                                                  f+i_coord_offset+DIM,fshift+i_shift_offset);
2084
2085         /* Increment number of inner iterations */
2086         inneriter                  += j_index_end - j_index_start;
2087
2088         /* Outer loop uses 18 flops */
2089     }
2090
2091     /* Increment number of outer iterations */
2092     outeriter        += nri;
2093
2094     /* Update outer/inner flops */
2095
2096     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_W4W4_F,outeriter*18 + inneriter*270);
2097 }