src/gromacs/gmxlib/nonbonded/nb_kernel_avx_256_double/nb_kernel_ElecCoul_VdwLJ_GeomW4P1_avx_256_double.cpp

   1 /*
   2  * This file is part of the GROMACS molecular simulation package.
   3  *
   4  * Copyright (c) 2012,2013,2014,2015,2017,2018, by the GROMACS development team, led by
   5  * Mark Abraham, David van der Spoel, Berk Hess, and Erik Lindahl,
   6  * and including many others, as listed in the AUTHORS file in the
   7  * top-level source directory and at http://www.gromacs.org.
   8  *
   9  * GROMACS is free software; you can redistribute it and/or
  10  * modify it under the terms of the GNU Lesser General Public License
  11  * as published by the Free Software Foundation; either version 2.1
  12  * of the License, or (at your option) any later version.
  13  *
  14  * GROMACS is distributed in the hope that it will be useful,
  15  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  16  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  17  * Lesser General Public License for more details.
  18  *
  19  * You should have received a copy of the GNU Lesser General Public
  20  * License along with GROMACS; if not, see
  21  * http://www.gnu.org/licenses, or write to the Free Software Foundation,
  22  * Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA.
  23  *
  24  * If you want to redistribute modifications to GROMACS, please
  25  * consider that scientific software is very special. Version
  26  * control is crucial - bugs must be traceable. We will be happy to
  27  * consider code for inclusion in the official distribution, but
  28  * derived work must not be called official GROMACS. Details are found
  29  * in the README & COPYING files - if they are missing, get the
  30  * official version at http://www.gromacs.org.
  31  *
  32  * To help us fund GROMACS development, we humbly ask that you cite
  33  * the research papers on the package. Check out http://www.gromacs.org.
  34  */
  35 /*
  36  * Note: this file was generated by the GROMACS avx_256_double kernel generator.
  37  */
  38 #include "gmxpre.h"
  39
  40 #include "config.h"
  41
  42 #include <math.h>
  43
  44 #include "../nb_kernel.h"
  45 #include "gromacs/gmxlib/nrnb.h"
  46
  47 #include "kernelutil_x86_avx_256_double.h"
  48
  49 /*
  50  * Gromacs nonbonded kernel:   nb_kernel_ElecCoul_VdwLJ_GeomW4P1_VF_avx_256_double
  51  * Electrostatics interaction: Coulomb
  52  * VdW interaction:            LennardJones
  53  * Geometry:                   Water4-Particle
  54  * Calculate force/pot:        PotentialAndForce
  55  */
  56 void
  57 nb_kernel_ElecCoul_VdwLJ_GeomW4P1_VF_avx_256_double
  58                     (t_nblist                    * gmx_restrict       nlist,
  59                      rvec                        * gmx_restrict          xx,
  60                      rvec                        * gmx_restrict          ff,
  61                      struct t_forcerec           * gmx_restrict          fr,
  62                      t_mdatoms                   * gmx_restrict     mdatoms,
  63                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
  64                      t_nrnb                      * gmx_restrict        nrnb)
  65 {
  66     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
  67      * just 0 for non-waters.
  68      * Suffixes A,B,C,D refer to j loop unrolling done with AVX, e.g. for the four different
  69      * jnr indices corresponding to data put in the four positions in the SIMD register.
  70      */
  71     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
  72     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
  73     int              jnrA,jnrB,jnrC,jnrD;
  74     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
  75     int              jnrlistE,jnrlistF,jnrlistG,jnrlistH;
  76     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
  77     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  78     real             rcutoff_scalar;
  79     real             *shiftvec,*fshift,*x,*f;
  80     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD;
  81     real             scratch[4*DIM];
  82     __m256d          tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
  83     real *           vdwioffsetptr0;
  84     __m256d          ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  85     real *           vdwioffsetptr1;
  86     __m256d          ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
  87     real *           vdwioffsetptr2;
  88     __m256d          ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
  89     real *           vdwioffsetptr3;
  90     __m256d          ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
  91     int              vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D;
  92     __m256d          jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  93     __m256d          dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
  94     __m256d          dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
  95     __m256d          dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
  96     __m256d          dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30;
  97     __m256d          velec,felec,velecsum,facel,crf,krf,krf2;
  98     real             *charge;
  99     int              nvdwtype;
 100     __m256d          rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
 101     int              *vdwtype;
 102     real             *vdwparam;
 103     __m256d          one_sixth   = _mm256_set1_pd(1.0/6.0);
 104     __m256d          one_twelfth = _mm256_set1_pd(1.0/12.0);
 105     __m256d          dummy_mask,cutoff_mask;
 106     __m128           tmpmask0,tmpmask1;
 107     __m256d          signbit = _mm256_castsi256_pd( _mm256_set1_epi32(0x80000000) );
 108     __m256d          one     = _mm256_set1_pd(1.0);
 109     __m256d          two     = _mm256_set1_pd(2.0);
 110     x                = xx[0];
 111     f                = ff[0];
 112
 113     nri              = nlist->nri;
 114     iinr             = nlist->iinr;
 115     jindex           = nlist->jindex;
 116     jjnr             = nlist->jjnr;
 117     shiftidx         = nlist->shift;
 118     gid              = nlist->gid;
 119     shiftvec         = fr->shift_vec[0];
 120     fshift           = fr->fshift[0];
 121     facel            = _mm256_set1_pd(fr->ic->epsfac);
 122     charge           = mdatoms->chargeA;
 123     nvdwtype         = fr->ntype;
 124     vdwparam         = fr->nbfp;
 125     vdwtype          = mdatoms->typeA;
 126
 127     /* Setup water-specific parameters */
 128     inr              = nlist->iinr[0];
 129     iq1              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+1]));
 130     iq2              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+2]));
 131     iq3              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+3]));
 132     vdwioffsetptr0   = vdwparam+2*nvdwtype*vdwtype[inr+0];
 133
 134     /* Avoid stupid compiler warnings */
 135     jnrA = jnrB = jnrC = jnrD = 0;
 136     j_coord_offsetA = 0;
 137     j_coord_offsetB = 0;
 138     j_coord_offsetC = 0;
 139     j_coord_offsetD = 0;
 140
 141     outeriter        = 0;
 142     inneriter        = 0;
 143
 144     for(iidx=0;iidx<4*DIM;iidx++)
 145     {
 146         scratch[iidx] = 0.0;
 147     }
 148
 149     /* Start outer loop over neighborlists */
 150     for(iidx=0; iidx<nri; iidx++)
 151     {
 152         /* Load shift vector for this list */
 153         i_shift_offset   = DIM*shiftidx[iidx];
 154
 155         /* Load limits for loop over neighbors */
 156         j_index_start    = jindex[iidx];
 157         j_index_end      = jindex[iidx+1];
 158
 159         /* Get outer coordinate index */
 160         inr              = iinr[iidx];
 161         i_coord_offset   = DIM*inr;
 162
 163         /* Load i particle coords and add shift vector */
 164         gmx_mm256_load_shift_and_4rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset,
 165                                                     &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
 166
 167         fix0             = _mm256_setzero_pd();
 168         fiy0             = _mm256_setzero_pd();
 169         fiz0             = _mm256_setzero_pd();
 170         fix1             = _mm256_setzero_pd();
 171         fiy1             = _mm256_setzero_pd();
 172         fiz1             = _mm256_setzero_pd();
 173         fix2             = _mm256_setzero_pd();
 174         fiy2             = _mm256_setzero_pd();
 175         fiz2             = _mm256_setzero_pd();
 176         fix3             = _mm256_setzero_pd();
 177         fiy3             = _mm256_setzero_pd();
 178         fiz3             = _mm256_setzero_pd();
 179
 180         /* Reset potential sums */
 181         velecsum         = _mm256_setzero_pd();
 182         vvdwsum          = _mm256_setzero_pd();
 183
 184         /* Start inner kernel loop */
 185         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4)
 186         {
 187
 188             /* Get j neighbor index, and coordinate index */
 189             jnrA             = jjnr[jidx];
 190             jnrB             = jjnr[jidx+1];
 191             jnrC             = jjnr[jidx+2];
 192             jnrD             = jjnr[jidx+3];
 193             j_coord_offsetA  = DIM*jnrA;
 194             j_coord_offsetB  = DIM*jnrB;
 195             j_coord_offsetC  = DIM*jnrC;
 196             j_coord_offsetD  = DIM*jnrD;
 197
 198             /* load j atom coordinates */
 199             gmx_mm256_load_1rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
 200                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 201                                                  &jx0,&jy0,&jz0);
 202
 203             /* Calculate displacement vector */
 204             dx00             = _mm256_sub_pd(ix0,jx0);
 205             dy00             = _mm256_sub_pd(iy0,jy0);
 206             dz00             = _mm256_sub_pd(iz0,jz0);
 207             dx10             = _mm256_sub_pd(ix1,jx0);
 208             dy10             = _mm256_sub_pd(iy1,jy0);
 209             dz10             = _mm256_sub_pd(iz1,jz0);
 210             dx20             = _mm256_sub_pd(ix2,jx0);
 211             dy20             = _mm256_sub_pd(iy2,jy0);
 212             dz20             = _mm256_sub_pd(iz2,jz0);
 213             dx30             = _mm256_sub_pd(ix3,jx0);
 214             dy30             = _mm256_sub_pd(iy3,jy0);
 215             dz30             = _mm256_sub_pd(iz3,jz0);
 216
 217             /* Calculate squared distance and things based on it */
 218             rsq00            = gmx_mm256_calc_rsq_pd(dx00,dy00,dz00);
 219             rsq10            = gmx_mm256_calc_rsq_pd(dx10,dy10,dz10);
 220             rsq20            = gmx_mm256_calc_rsq_pd(dx20,dy20,dz20);
 221             rsq30            = gmx_mm256_calc_rsq_pd(dx30,dy30,dz30);
 222
 223             rinv10           = avx256_invsqrt_d(rsq10);
 224             rinv20           = avx256_invsqrt_d(rsq20);
 225             rinv30           = avx256_invsqrt_d(rsq30);
 226
 227             rinvsq00         = avx256_inv_d(rsq00);
 228             rinvsq10         = _mm256_mul_pd(rinv10,rinv10);
 229             rinvsq20         = _mm256_mul_pd(rinv20,rinv20);
 230             rinvsq30         = _mm256_mul_pd(rinv30,rinv30);
 231
 232             /* Load parameters for j particles */
 233             jq0              = gmx_mm256_load_4real_swizzle_pd(charge+jnrA+0,charge+jnrB+0,
 234                                                                  charge+jnrC+0,charge+jnrD+0);
 235             vdwjidx0A        = 2*vdwtype[jnrA+0];
 236             vdwjidx0B        = 2*vdwtype[jnrB+0];
 237             vdwjidx0C        = 2*vdwtype[jnrC+0];
 238             vdwjidx0D        = 2*vdwtype[jnrD+0];
 239
 240             fjx0             = _mm256_setzero_pd();
 241             fjy0             = _mm256_setzero_pd();
 242             fjz0             = _mm256_setzero_pd();
 243
 244             /**************************
 245              * CALCULATE INTERACTIONS *
 246              **************************/
 247
 248             /* Compute parameters for interactions between i and j atoms */
 249             gmx_mm256_load_4pair_swizzle_pd(vdwioffsetptr0+vdwjidx0A,
 250                                             vdwioffsetptr0+vdwjidx0B,
 251                                             vdwioffsetptr0+vdwjidx0C,
 252                                             vdwioffsetptr0+vdwjidx0D,
 253                                             &c6_00,&c12_00);
 254
 255             /* LENNARD-JONES DISPERSION/REPULSION */
 256
 257             rinvsix          = _mm256_mul_pd(_mm256_mul_pd(rinvsq00,rinvsq00),rinvsq00);
 258             vvdw6            = _mm256_mul_pd(c6_00,rinvsix);
 259             vvdw12           = _mm256_mul_pd(c12_00,_mm256_mul_pd(rinvsix,rinvsix));
 260             vvdw             = _mm256_sub_pd( _mm256_mul_pd(vvdw12,one_twelfth) , _mm256_mul_pd(vvdw6,one_sixth) );
 261             fvdw             = _mm256_mul_pd(_mm256_sub_pd(vvdw12,vvdw6),rinvsq00);
 262
 263             /* Update potential sum for this i atom from the interaction with this j atom. */
 264             vvdwsum          = _mm256_add_pd(vvdwsum,vvdw);
 265
 266             fscal            = fvdw;
 267
 268             /* Calculate temporary vectorial force */
 269             tx               = _mm256_mul_pd(fscal,dx00);
 270             ty               = _mm256_mul_pd(fscal,dy00);
 271             tz               = _mm256_mul_pd(fscal,dz00);
 272
 273             /* Update vectorial force */
 274             fix0             = _mm256_add_pd(fix0,tx);
 275             fiy0             = _mm256_add_pd(fiy0,ty);
 276             fiz0             = _mm256_add_pd(fiz0,tz);
 277
 278             fjx0             = _mm256_add_pd(fjx0,tx);
 279             fjy0             = _mm256_add_pd(fjy0,ty);
 280             fjz0             = _mm256_add_pd(fjz0,tz);
 281
 282             /**************************
 283              * CALCULATE INTERACTIONS *
 284              **************************/
 285
 286             /* Compute parameters for interactions between i and j atoms */
 287             qq10             = _mm256_mul_pd(iq1,jq0);
 288
 289             /* COULOMB ELECTROSTATICS */
 290             velec            = _mm256_mul_pd(qq10,rinv10);
 291             felec            = _mm256_mul_pd(velec,rinvsq10);
 292
 293             /* Update potential sum for this i atom from the interaction with this j atom. */
 294             velecsum         = _mm256_add_pd(velecsum,velec);
 295
 296             fscal            = felec;
 297
 298             /* Calculate temporary vectorial force */
 299             tx               = _mm256_mul_pd(fscal,dx10);
 300             ty               = _mm256_mul_pd(fscal,dy10);
 301             tz               = _mm256_mul_pd(fscal,dz10);
 302
 303             /* Update vectorial force */
 304             fix1             = _mm256_add_pd(fix1,tx);
 305             fiy1             = _mm256_add_pd(fiy1,ty);
 306             fiz1             = _mm256_add_pd(fiz1,tz);
 307
 308             fjx0             = _mm256_add_pd(fjx0,tx);
 309             fjy0             = _mm256_add_pd(fjy0,ty);
 310             fjz0             = _mm256_add_pd(fjz0,tz);
 311
 312             /**************************
 313              * CALCULATE INTERACTIONS *
 314              **************************/
 315
 316             /* Compute parameters for interactions between i and j atoms */
 317             qq20             = _mm256_mul_pd(iq2,jq0);
 318
 319             /* COULOMB ELECTROSTATICS */
 320             velec            = _mm256_mul_pd(qq20,rinv20);
 321             felec            = _mm256_mul_pd(velec,rinvsq20);
 322
 323             /* Update potential sum for this i atom from the interaction with this j atom. */
 324             velecsum         = _mm256_add_pd(velecsum,velec);
 325
 326             fscal            = felec;
 327
 328             /* Calculate temporary vectorial force */
 329             tx               = _mm256_mul_pd(fscal,dx20);
 330             ty               = _mm256_mul_pd(fscal,dy20);
 331             tz               = _mm256_mul_pd(fscal,dz20);
 332
 333             /* Update vectorial force */
 334             fix2             = _mm256_add_pd(fix2,tx);
 335             fiy2             = _mm256_add_pd(fiy2,ty);
 336             fiz2             = _mm256_add_pd(fiz2,tz);
 337
 338             fjx0             = _mm256_add_pd(fjx0,tx);
 339             fjy0             = _mm256_add_pd(fjy0,ty);
 340             fjz0             = _mm256_add_pd(fjz0,tz);
 341
 342             /**************************
 343              * CALCULATE INTERACTIONS *
 344              **************************/
 345
 346             /* Compute parameters for interactions between i and j atoms */
 347             qq30             = _mm256_mul_pd(iq3,jq0);
 348
 349             /* COULOMB ELECTROSTATICS */
 350             velec            = _mm256_mul_pd(qq30,rinv30);
 351             felec            = _mm256_mul_pd(velec,rinvsq30);
 352
 353             /* Update potential sum for this i atom from the interaction with this j atom. */
 354             velecsum         = _mm256_add_pd(velecsum,velec);
 355
 356             fscal            = felec;
 357
 358             /* Calculate temporary vectorial force */
 359             tx               = _mm256_mul_pd(fscal,dx30);
 360             ty               = _mm256_mul_pd(fscal,dy30);
 361             tz               = _mm256_mul_pd(fscal,dz30);
 362
 363             /* Update vectorial force */
 364             fix3             = _mm256_add_pd(fix3,tx);
 365             fiy3             = _mm256_add_pd(fiy3,ty);
 366             fiz3             = _mm256_add_pd(fiz3,tz);
 367
 368             fjx0             = _mm256_add_pd(fjx0,tx);
 369             fjy0             = _mm256_add_pd(fjy0,ty);
 370             fjz0             = _mm256_add_pd(fjz0,tz);
 371
 372             fjptrA             = f+j_coord_offsetA;
 373             fjptrB             = f+j_coord_offsetB;
 374             fjptrC             = f+j_coord_offsetC;
 375             fjptrD             = f+j_coord_offsetD;
 376
 377             gmx_mm256_decrement_1rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0);
 378
 379             /* Inner loop uses 116 flops */
 380         }
 381
 382         if(jidx<j_index_end)
 383         {
 384
 385             /* Get j neighbor index, and coordinate index */
 386             jnrlistA         = jjnr[jidx];
 387             jnrlistB         = jjnr[jidx+1];
 388             jnrlistC         = jjnr[jidx+2];
 389             jnrlistD         = jjnr[jidx+3];
 390             /* Sign of each element will be negative for non-real atoms.
 391              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
 392              * so use it as val = _mm_andnot_pd(mask,val) to clear dummy entries.
 393              */
 394             tmpmask0 = gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128()));
 395
 396             tmpmask1 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(3,3,2,2));
 397             tmpmask0 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(1,1,0,0));
 398             dummy_mask = _mm256_castps_pd(gmx_mm256_set_m128(tmpmask1,tmpmask0));
 399
 400             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
 401             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
 402             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
 403             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
 404             j_coord_offsetA  = DIM*jnrA;
 405             j_coord_offsetB  = DIM*jnrB;
 406             j_coord_offsetC  = DIM*jnrC;
 407             j_coord_offsetD  = DIM*jnrD;
 408
 409             /* load j atom coordinates */
 410             gmx_mm256_load_1rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
 411                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 412                                                  &jx0,&jy0,&jz0);
 413
 414             /* Calculate displacement vector */
 415             dx00             = _mm256_sub_pd(ix0,jx0);
 416             dy00             = _mm256_sub_pd(iy0,jy0);
 417             dz00             = _mm256_sub_pd(iz0,jz0);
 418             dx10             = _mm256_sub_pd(ix1,jx0);
 419             dy10             = _mm256_sub_pd(iy1,jy0);
 420             dz10             = _mm256_sub_pd(iz1,jz0);
 421             dx20             = _mm256_sub_pd(ix2,jx0);
 422             dy20             = _mm256_sub_pd(iy2,jy0);
 423             dz20             = _mm256_sub_pd(iz2,jz0);
 424             dx30             = _mm256_sub_pd(ix3,jx0);
 425             dy30             = _mm256_sub_pd(iy3,jy0);
 426             dz30             = _mm256_sub_pd(iz3,jz0);
 427
 428             /* Calculate squared distance and things based on it */
 429             rsq00            = gmx_mm256_calc_rsq_pd(dx00,dy00,dz00);
 430             rsq10            = gmx_mm256_calc_rsq_pd(dx10,dy10,dz10);
 431             rsq20            = gmx_mm256_calc_rsq_pd(dx20,dy20,dz20);
 432             rsq30            = gmx_mm256_calc_rsq_pd(dx30,dy30,dz30);
 433
 434             rinv10           = avx256_invsqrt_d(rsq10);
 435             rinv20           = avx256_invsqrt_d(rsq20);
 436             rinv30           = avx256_invsqrt_d(rsq30);
 437
 438             rinvsq00         = avx256_inv_d(rsq00);
 439             rinvsq10         = _mm256_mul_pd(rinv10,rinv10);
 440             rinvsq20         = _mm256_mul_pd(rinv20,rinv20);
 441             rinvsq30         = _mm256_mul_pd(rinv30,rinv30);
 442
 443             /* Load parameters for j particles */
 444             jq0              = gmx_mm256_load_4real_swizzle_pd(charge+jnrA+0,charge+jnrB+0,
 445                                                                  charge+jnrC+0,charge+jnrD+0);
 446             vdwjidx0A        = 2*vdwtype[jnrA+0];
 447             vdwjidx0B        = 2*vdwtype[jnrB+0];
 448             vdwjidx0C        = 2*vdwtype[jnrC+0];
 449             vdwjidx0D        = 2*vdwtype[jnrD+0];
 450
 451             fjx0             = _mm256_setzero_pd();
 452             fjy0             = _mm256_setzero_pd();
 453             fjz0             = _mm256_setzero_pd();
 454
 455             /**************************
 456              * CALCULATE INTERACTIONS *
 457              **************************/
 458
 459             /* Compute parameters for interactions between i and j atoms */
 460             gmx_mm256_load_4pair_swizzle_pd(vdwioffsetptr0+vdwjidx0A,
 461                                             vdwioffsetptr0+vdwjidx0B,
 462                                             vdwioffsetptr0+vdwjidx0C,
 463                                             vdwioffsetptr0+vdwjidx0D,
 464                                             &c6_00,&c12_00);
 465
 466             /* LENNARD-JONES DISPERSION/REPULSION */
 467
 468             rinvsix          = _mm256_mul_pd(_mm256_mul_pd(rinvsq00,rinvsq00),rinvsq00);
 469             vvdw6            = _mm256_mul_pd(c6_00,rinvsix);
 470             vvdw12           = _mm256_mul_pd(c12_00,_mm256_mul_pd(rinvsix,rinvsix));
 471             vvdw             = _mm256_sub_pd( _mm256_mul_pd(vvdw12,one_twelfth) , _mm256_mul_pd(vvdw6,one_sixth) );
 472             fvdw             = _mm256_mul_pd(_mm256_sub_pd(vvdw12,vvdw6),rinvsq00);
 473
 474             /* Update potential sum for this i atom from the interaction with this j atom. */
 475             vvdw             = _mm256_andnot_pd(dummy_mask,vvdw);
 476             vvdwsum          = _mm256_add_pd(vvdwsum,vvdw);
 477
 478             fscal            = fvdw;
 479
 480             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
 481
 482             /* Calculate temporary vectorial force */
 483             tx               = _mm256_mul_pd(fscal,dx00);
 484             ty               = _mm256_mul_pd(fscal,dy00);
 485             tz               = _mm256_mul_pd(fscal,dz00);
 486
 487             /* Update vectorial force */
 488             fix0             = _mm256_add_pd(fix0,tx);
 489             fiy0             = _mm256_add_pd(fiy0,ty);
 490             fiz0             = _mm256_add_pd(fiz0,tz);
 491
 492             fjx0             = _mm256_add_pd(fjx0,tx);
 493             fjy0             = _mm256_add_pd(fjy0,ty);
 494             fjz0             = _mm256_add_pd(fjz0,tz);
 495
 496             /**************************
 497              * CALCULATE INTERACTIONS *
 498              **************************/
 499
 500             /* Compute parameters for interactions between i and j atoms */
 501             qq10             = _mm256_mul_pd(iq1,jq0);
 502
 503             /* COULOMB ELECTROSTATICS */
 504             velec            = _mm256_mul_pd(qq10,rinv10);
 505             felec            = _mm256_mul_pd(velec,rinvsq10);
 506
 507             /* Update potential sum for this i atom from the interaction with this j atom. */
 508             velec            = _mm256_andnot_pd(dummy_mask,velec);
 509             velecsum         = _mm256_add_pd(velecsum,velec);
 510
 511             fscal            = felec;
 512
 513             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
 514
 515             /* Calculate temporary vectorial force */
 516             tx               = _mm256_mul_pd(fscal,dx10);
 517             ty               = _mm256_mul_pd(fscal,dy10);
 518             tz               = _mm256_mul_pd(fscal,dz10);
 519
 520             /* Update vectorial force */
 521             fix1             = _mm256_add_pd(fix1,tx);
 522             fiy1             = _mm256_add_pd(fiy1,ty);
 523             fiz1             = _mm256_add_pd(fiz1,tz);
 524
 525             fjx0             = _mm256_add_pd(fjx0,tx);
 526             fjy0             = _mm256_add_pd(fjy0,ty);
 527             fjz0             = _mm256_add_pd(fjz0,tz);
 528
 529             /**************************
 530              * CALCULATE INTERACTIONS *
 531              **************************/
 532
 533             /* Compute parameters for interactions between i and j atoms */
 534             qq20             = _mm256_mul_pd(iq2,jq0);
 535
 536             /* COULOMB ELECTROSTATICS */
 537             velec            = _mm256_mul_pd(qq20,rinv20);
 538             felec            = _mm256_mul_pd(velec,rinvsq20);
 539
 540             /* Update potential sum for this i atom from the interaction with this j atom. */
 541             velec            = _mm256_andnot_pd(dummy_mask,velec);
 542             velecsum         = _mm256_add_pd(velecsum,velec);
 543
 544             fscal            = felec;
 545
 546             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
 547
 548             /* Calculate temporary vectorial force */
 549             tx               = _mm256_mul_pd(fscal,dx20);
 550             ty               = _mm256_mul_pd(fscal,dy20);
 551             tz               = _mm256_mul_pd(fscal,dz20);
 552
 553             /* Update vectorial force */
 554             fix2             = _mm256_add_pd(fix2,tx);
 555             fiy2             = _mm256_add_pd(fiy2,ty);
 556             fiz2             = _mm256_add_pd(fiz2,tz);
 557
 558             fjx0             = _mm256_add_pd(fjx0,tx);
 559             fjy0             = _mm256_add_pd(fjy0,ty);
 560             fjz0             = _mm256_add_pd(fjz0,tz);
 561
 562             /**************************
 563              * CALCULATE INTERACTIONS *
 564              **************************/
 565
 566             /* Compute parameters for interactions between i and j atoms */
 567             qq30             = _mm256_mul_pd(iq3,jq0);
 568
 569             /* COULOMB ELECTROSTATICS */
 570             velec            = _mm256_mul_pd(qq30,rinv30);
 571             felec            = _mm256_mul_pd(velec,rinvsq30);
 572
 573             /* Update potential sum for this i atom from the interaction with this j atom. */
 574             velec            = _mm256_andnot_pd(dummy_mask,velec);
 575             velecsum         = _mm256_add_pd(velecsum,velec);
 576
 577             fscal            = felec;
 578
 579             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
 580
 581             /* Calculate temporary vectorial force */
 582             tx               = _mm256_mul_pd(fscal,dx30);
 583             ty               = _mm256_mul_pd(fscal,dy30);
 584             tz               = _mm256_mul_pd(fscal,dz30);
 585
 586             /* Update vectorial force */
 587             fix3             = _mm256_add_pd(fix3,tx);
 588             fiy3             = _mm256_add_pd(fiy3,ty);
 589             fiz3             = _mm256_add_pd(fiz3,tz);
 590
 591             fjx0             = _mm256_add_pd(fjx0,tx);
 592             fjy0             = _mm256_add_pd(fjy0,ty);
 593             fjz0             = _mm256_add_pd(fjz0,tz);
 594
 595             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
 596             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
 597             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
 598             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
 599
 600             gmx_mm256_decrement_1rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0);
 601
 602             /* Inner loop uses 116 flops */
 603         }
 604
 605         /* End of innermost loop */
 606
 607         gmx_mm256_update_iforce_4atom_swizzle_pd(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
 608                                                  f+i_coord_offset,fshift+i_shift_offset);
 609
 610         ggid                        = gid[iidx];
 611         /* Update potential energies */
 612         gmx_mm256_update_1pot_pd(velecsum,kernel_data->energygrp_elec+ggid);
 613         gmx_mm256_update_1pot_pd(vvdwsum,kernel_data->energygrp_vdw+ggid);
 614
 615         /* Increment number of inner iterations */
 616         inneriter                  += j_index_end - j_index_start;
 617
 618         /* Outer loop uses 26 flops */
 619     }
 620
 621     /* Increment number of outer iterations */
 622     outeriter        += nri;
 623
 624     /* Update outer/inner flops */
 625
 626     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_VF,outeriter*26 + inneriter*116);
 627 }
 628 /*
 629  * Gromacs nonbonded kernel:   nb_kernel_ElecCoul_VdwLJ_GeomW4P1_F_avx_256_double
 630  * Electrostatics interaction: Coulomb
 631  * VdW interaction:            LennardJones
 632  * Geometry:                   Water4-Particle
 633  * Calculate force/pot:        Force
 634  */
 635 void
 636 nb_kernel_ElecCoul_VdwLJ_GeomW4P1_F_avx_256_double
 637                     (t_nblist                    * gmx_restrict       nlist,
 638                      rvec                        * gmx_restrict          xx,
 639                      rvec                        * gmx_restrict          ff,
 640                      struct t_forcerec           * gmx_restrict          fr,
 641                      t_mdatoms                   * gmx_restrict     mdatoms,
 642                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 643                      t_nrnb                      * gmx_restrict        nrnb)
 644 {
 645     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
 646      * just 0 for non-waters.
 647      * Suffixes A,B,C,D refer to j loop unrolling done with AVX, e.g. for the four different
 648      * jnr indices corresponding to data put in the four positions in the SIMD register.
 649      */
 650     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
 651     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
 652     int              jnrA,jnrB,jnrC,jnrD;
 653     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
 654     int              jnrlistE,jnrlistF,jnrlistG,jnrlistH;
 655     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
 656     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 657     real             rcutoff_scalar;
 658     real             *shiftvec,*fshift,*x,*f;
 659     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD;
 660     real             scratch[4*DIM];
 661     __m256d          tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
 662     real *           vdwioffsetptr0;
 663     __m256d          ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 664     real *           vdwioffsetptr1;
 665     __m256d          ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
 666     real *           vdwioffsetptr2;
 667     __m256d          ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
 668     real *           vdwioffsetptr3;
 669     __m256d          ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
 670     int              vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D;
 671     __m256d          jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 672     __m256d          dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
 673     __m256d          dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
 674     __m256d          dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
 675     __m256d          dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30;
 676     __m256d          velec,felec,velecsum,facel,crf,krf,krf2;
 677     real             *charge;
 678     int              nvdwtype;
 679     __m256d          rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
 680     int              *vdwtype;
 681     real             *vdwparam;
 682     __m256d          one_sixth   = _mm256_set1_pd(1.0/6.0);
 683     __m256d          one_twelfth = _mm256_set1_pd(1.0/12.0);
 684     __m256d          dummy_mask,cutoff_mask;
 685     __m128           tmpmask0,tmpmask1;
 686     __m256d          signbit = _mm256_castsi256_pd( _mm256_set1_epi32(0x80000000) );
 687     __m256d          one     = _mm256_set1_pd(1.0);
 688     __m256d          two     = _mm256_set1_pd(2.0);
 689     x                = xx[0];
 690     f                = ff[0];
 691
 692     nri              = nlist->nri;
 693     iinr             = nlist->iinr;
 694     jindex           = nlist->jindex;
 695     jjnr             = nlist->jjnr;
 696     shiftidx         = nlist->shift;
 697     gid              = nlist->gid;
 698     shiftvec         = fr->shift_vec[0];
 699     fshift           = fr->fshift[0];
 700     facel            = _mm256_set1_pd(fr->ic->epsfac);
 701     charge           = mdatoms->chargeA;
 702     nvdwtype         = fr->ntype;
 703     vdwparam         = fr->nbfp;
 704     vdwtype          = mdatoms->typeA;
 705
 706     /* Setup water-specific parameters */
 707     inr              = nlist->iinr[0];
 708     iq1              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+1]));
 709     iq2              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+2]));
 710     iq3              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+3]));
 711     vdwioffsetptr0   = vdwparam+2*nvdwtype*vdwtype[inr+0];
 712
 713     /* Avoid stupid compiler warnings */
 714     jnrA = jnrB = jnrC = jnrD = 0;
 715     j_coord_offsetA = 0;
 716     j_coord_offsetB = 0;
 717     j_coord_offsetC = 0;
 718     j_coord_offsetD = 0;
 719
 720     outeriter        = 0;
 721     inneriter        = 0;
 722
 723     for(iidx=0;iidx<4*DIM;iidx++)
 724     {
 725         scratch[iidx] = 0.0;
 726     }
 727
 728     /* Start outer loop over neighborlists */
 729     for(iidx=0; iidx<nri; iidx++)
 730     {
 731         /* Load shift vector for this list */
 732         i_shift_offset   = DIM*shiftidx[iidx];
 733
 734         /* Load limits for loop over neighbors */
 735         j_index_start    = jindex[iidx];
 736         j_index_end      = jindex[iidx+1];
 737
 738         /* Get outer coordinate index */
 739         inr              = iinr[iidx];
 740         i_coord_offset   = DIM*inr;
 741
 742         /* Load i particle coords and add shift vector */
 743         gmx_mm256_load_shift_and_4rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset,
 744                                                     &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
 745
 746         fix0             = _mm256_setzero_pd();
 747         fiy0             = _mm256_setzero_pd();
 748         fiz0             = _mm256_setzero_pd();
 749         fix1             = _mm256_setzero_pd();
 750         fiy1             = _mm256_setzero_pd();
 751         fiz1             = _mm256_setzero_pd();
 752         fix2             = _mm256_setzero_pd();
 753         fiy2             = _mm256_setzero_pd();
 754         fiz2             = _mm256_setzero_pd();
 755         fix3             = _mm256_setzero_pd();
 756         fiy3             = _mm256_setzero_pd();
 757         fiz3             = _mm256_setzero_pd();
 758
 759         /* Start inner kernel loop */
 760         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4)
 761         {
 762
 763             /* Get j neighbor index, and coordinate index */
 764             jnrA             = jjnr[jidx];
 765             jnrB             = jjnr[jidx+1];
 766             jnrC             = jjnr[jidx+2];
 767             jnrD             = jjnr[jidx+3];
 768             j_coord_offsetA  = DIM*jnrA;
 769             j_coord_offsetB  = DIM*jnrB;
 770             j_coord_offsetC  = DIM*jnrC;
 771             j_coord_offsetD  = DIM*jnrD;
 772
 773             /* load j atom coordinates */
 774             gmx_mm256_load_1rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
 775                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 776                                                  &jx0,&jy0,&jz0);
 777
 778             /* Calculate displacement vector */
 779             dx00             = _mm256_sub_pd(ix0,jx0);
 780             dy00             = _mm256_sub_pd(iy0,jy0);
 781             dz00             = _mm256_sub_pd(iz0,jz0);
 782             dx10             = _mm256_sub_pd(ix1,jx0);
 783             dy10             = _mm256_sub_pd(iy1,jy0);
 784             dz10             = _mm256_sub_pd(iz1,jz0);
 785             dx20             = _mm256_sub_pd(ix2,jx0);
 786             dy20             = _mm256_sub_pd(iy2,jy0);
 787             dz20             = _mm256_sub_pd(iz2,jz0);
 788             dx30             = _mm256_sub_pd(ix3,jx0);
 789             dy30             = _mm256_sub_pd(iy3,jy0);
 790             dz30             = _mm256_sub_pd(iz3,jz0);
 791
 792             /* Calculate squared distance and things based on it */
 793             rsq00            = gmx_mm256_calc_rsq_pd(dx00,dy00,dz00);
 794             rsq10            = gmx_mm256_calc_rsq_pd(dx10,dy10,dz10);
 795             rsq20            = gmx_mm256_calc_rsq_pd(dx20,dy20,dz20);
 796             rsq30            = gmx_mm256_calc_rsq_pd(dx30,dy30,dz30);
 797
 798             rinv10           = avx256_invsqrt_d(rsq10);
 799             rinv20           = avx256_invsqrt_d(rsq20);
 800             rinv30           = avx256_invsqrt_d(rsq30);
 801
 802             rinvsq00         = avx256_inv_d(rsq00);
 803             rinvsq10         = _mm256_mul_pd(rinv10,rinv10);
 804             rinvsq20         = _mm256_mul_pd(rinv20,rinv20);
 805             rinvsq30         = _mm256_mul_pd(rinv30,rinv30);
 806
 807             /* Load parameters for j particles */
 808             jq0              = gmx_mm256_load_4real_swizzle_pd(charge+jnrA+0,charge+jnrB+0,
 809                                                                  charge+jnrC+0,charge+jnrD+0);
 810             vdwjidx0A        = 2*vdwtype[jnrA+0];
 811             vdwjidx0B        = 2*vdwtype[jnrB+0];
 812             vdwjidx0C        = 2*vdwtype[jnrC+0];
 813             vdwjidx0D        = 2*vdwtype[jnrD+0];
 814
 815             fjx0             = _mm256_setzero_pd();
 816             fjy0             = _mm256_setzero_pd();
 817             fjz0             = _mm256_setzero_pd();
 818
 819             /**************************
 820              * CALCULATE INTERACTIONS *
 821              **************************/
 822
 823             /* Compute parameters for interactions between i and j atoms */
 824             gmx_mm256_load_4pair_swizzle_pd(vdwioffsetptr0+vdwjidx0A,
 825                                             vdwioffsetptr0+vdwjidx0B,
 826                                             vdwioffsetptr0+vdwjidx0C,
 827                                             vdwioffsetptr0+vdwjidx0D,
 828                                             &c6_00,&c12_00);
 829
 830             /* LENNARD-JONES DISPERSION/REPULSION */
 831
 832             rinvsix          = _mm256_mul_pd(_mm256_mul_pd(rinvsq00,rinvsq00),rinvsq00);
 833             fvdw             = _mm256_mul_pd(_mm256_sub_pd(_mm256_mul_pd(c12_00,rinvsix),c6_00),_mm256_mul_pd(rinvsix,rinvsq00));
 834
 835             fscal            = fvdw;
 836
 837             /* Calculate temporary vectorial force */
 838             tx               = _mm256_mul_pd(fscal,dx00);
 839             ty               = _mm256_mul_pd(fscal,dy00);
 840             tz               = _mm256_mul_pd(fscal,dz00);
 841
 842             /* Update vectorial force */
 843             fix0             = _mm256_add_pd(fix0,tx);
 844             fiy0             = _mm256_add_pd(fiy0,ty);
 845             fiz0             = _mm256_add_pd(fiz0,tz);
 846
 847             fjx0             = _mm256_add_pd(fjx0,tx);
 848             fjy0             = _mm256_add_pd(fjy0,ty);
 849             fjz0             = _mm256_add_pd(fjz0,tz);
 850
 851             /**************************
 852              * CALCULATE INTERACTIONS *
 853              **************************/
 854
 855             /* Compute parameters for interactions between i and j atoms */
 856             qq10             = _mm256_mul_pd(iq1,jq0);
 857
 858             /* COULOMB ELECTROSTATICS */
 859             velec            = _mm256_mul_pd(qq10,rinv10);
 860             felec            = _mm256_mul_pd(velec,rinvsq10);
 861
 862             fscal            = felec;
 863
 864             /* Calculate temporary vectorial force */
 865             tx               = _mm256_mul_pd(fscal,dx10);
 866             ty               = _mm256_mul_pd(fscal,dy10);
 867             tz               = _mm256_mul_pd(fscal,dz10);
 868
 869             /* Update vectorial force */
 870             fix1             = _mm256_add_pd(fix1,tx);
 871             fiy1             = _mm256_add_pd(fiy1,ty);
 872             fiz1             = _mm256_add_pd(fiz1,tz);
 873
 874             fjx0             = _mm256_add_pd(fjx0,tx);
 875             fjy0             = _mm256_add_pd(fjy0,ty);
 876             fjz0             = _mm256_add_pd(fjz0,tz);
 877
 878             /**************************
 879              * CALCULATE INTERACTIONS *
 880              **************************/
 881
 882             /* Compute parameters for interactions between i and j atoms */
 883             qq20             = _mm256_mul_pd(iq2,jq0);
 884
 885             /* COULOMB ELECTROSTATICS */
 886             velec            = _mm256_mul_pd(qq20,rinv20);
 887             felec            = _mm256_mul_pd(velec,rinvsq20);
 888
 889             fscal            = felec;
 890
 891             /* Calculate temporary vectorial force */
 892             tx               = _mm256_mul_pd(fscal,dx20);
 893             ty               = _mm256_mul_pd(fscal,dy20);
 894             tz               = _mm256_mul_pd(fscal,dz20);
 895
 896             /* Update vectorial force */
 897             fix2             = _mm256_add_pd(fix2,tx);
 898             fiy2             = _mm256_add_pd(fiy2,ty);
 899             fiz2             = _mm256_add_pd(fiz2,tz);
 900
 901             fjx0             = _mm256_add_pd(fjx0,tx);
 902             fjy0             = _mm256_add_pd(fjy0,ty);
 903             fjz0             = _mm256_add_pd(fjz0,tz);
 904
 905             /**************************
 906              * CALCULATE INTERACTIONS *
 907              **************************/
 908
 909             /* Compute parameters for interactions between i and j atoms */
 910             qq30             = _mm256_mul_pd(iq3,jq0);
 911
 912             /* COULOMB ELECTROSTATICS */
 913             velec            = _mm256_mul_pd(qq30,rinv30);
 914             felec            = _mm256_mul_pd(velec,rinvsq30);
 915
 916             fscal            = felec;
 917
 918             /* Calculate temporary vectorial force */
 919             tx               = _mm256_mul_pd(fscal,dx30);
 920             ty               = _mm256_mul_pd(fscal,dy30);
 921             tz               = _mm256_mul_pd(fscal,dz30);
 922
 923             /* Update vectorial force */
 924             fix3             = _mm256_add_pd(fix3,tx);
 925             fiy3             = _mm256_add_pd(fiy3,ty);
 926             fiz3             = _mm256_add_pd(fiz3,tz);
 927
 928             fjx0             = _mm256_add_pd(fjx0,tx);
 929             fjy0             = _mm256_add_pd(fjy0,ty);
 930             fjz0             = _mm256_add_pd(fjz0,tz);
 931
 932             fjptrA             = f+j_coord_offsetA;
 933             fjptrB             = f+j_coord_offsetB;
 934             fjptrC             = f+j_coord_offsetC;
 935             fjptrD             = f+j_coord_offsetD;
 936
 937             gmx_mm256_decrement_1rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0);
 938
 939             /* Inner loop uses 108 flops */
 940         }
 941
 942         if(jidx<j_index_end)
 943         {
 944
 945             /* Get j neighbor index, and coordinate index */
 946             jnrlistA         = jjnr[jidx];
 947             jnrlistB         = jjnr[jidx+1];
 948             jnrlistC         = jjnr[jidx+2];
 949             jnrlistD         = jjnr[jidx+3];
 950             /* Sign of each element will be negative for non-real atoms.
 951              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
 952              * so use it as val = _mm_andnot_pd(mask,val) to clear dummy entries.
 953              */
 954             tmpmask0 = gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128()));
 955
 956             tmpmask1 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(3,3,2,2));
 957             tmpmask0 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(1,1,0,0));
 958             dummy_mask = _mm256_castps_pd(gmx_mm256_set_m128(tmpmask1,tmpmask0));
 959
 960             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
 961             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
 962             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
 963             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
 964             j_coord_offsetA  = DIM*jnrA;
 965             j_coord_offsetB  = DIM*jnrB;
 966             j_coord_offsetC  = DIM*jnrC;
 967             j_coord_offsetD  = DIM*jnrD;
 968
 969             /* load j atom coordinates */
 970             gmx_mm256_load_1rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
 971                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 972                                                  &jx0,&jy0,&jz0);
 973
 974             /* Calculate displacement vector */
 975             dx00             = _mm256_sub_pd(ix0,jx0);
 976             dy00             = _mm256_sub_pd(iy0,jy0);
 977             dz00             = _mm256_sub_pd(iz0,jz0);
 978             dx10             = _mm256_sub_pd(ix1,jx0);
 979             dy10             = _mm256_sub_pd(iy1,jy0);
 980             dz10             = _mm256_sub_pd(iz1,jz0);
 981             dx20             = _mm256_sub_pd(ix2,jx0);
 982             dy20             = _mm256_sub_pd(iy2,jy0);
 983             dz20             = _mm256_sub_pd(iz2,jz0);
 984             dx30             = _mm256_sub_pd(ix3,jx0);
 985             dy30             = _mm256_sub_pd(iy3,jy0);
 986             dz30             = _mm256_sub_pd(iz3,jz0);
 987
 988             /* Calculate squared distance and things based on it */
 989             rsq00            = gmx_mm256_calc_rsq_pd(dx00,dy00,dz00);
 990             rsq10            = gmx_mm256_calc_rsq_pd(dx10,dy10,dz10);
 991             rsq20            = gmx_mm256_calc_rsq_pd(dx20,dy20,dz20);
 992             rsq30            = gmx_mm256_calc_rsq_pd(dx30,dy30,dz30);
 993
 994             rinv10           = avx256_invsqrt_d(rsq10);
 995             rinv20           = avx256_invsqrt_d(rsq20);
 996             rinv30           = avx256_invsqrt_d(rsq30);
 997
 998             rinvsq00         = avx256_inv_d(rsq00);
 999             rinvsq10         = _mm256_mul_pd(rinv10,rinv10);
1000             rinvsq20         = _mm256_mul_pd(rinv20,rinv20);
1001             rinvsq30         = _mm256_mul_pd(rinv30,rinv30);
1002
1003             /* Load parameters for j particles */
1004             jq0              = gmx_mm256_load_4real_swizzle_pd(charge+jnrA+0,charge+jnrB+0,
1005                                                                  charge+jnrC+0,charge+jnrD+0);
1006             vdwjidx0A        = 2*vdwtype[jnrA+0];
1007             vdwjidx0B        = 2*vdwtype[jnrB+0];
1008             vdwjidx0C        = 2*vdwtype[jnrC+0];
1009             vdwjidx0D        = 2*vdwtype[jnrD+0];
1010
1011             fjx0             = _mm256_setzero_pd();
1012             fjy0             = _mm256_setzero_pd();
1013             fjz0             = _mm256_setzero_pd();
1014
1015             /**************************
1016              * CALCULATE INTERACTIONS *
1017              **************************/
1018
1019             /* Compute parameters for interactions between i and j atoms */
1020             gmx_mm256_load_4pair_swizzle_pd(vdwioffsetptr0+vdwjidx0A,
1021                                             vdwioffsetptr0+vdwjidx0B,
1022                                             vdwioffsetptr0+vdwjidx0C,
1023                                             vdwioffsetptr0+vdwjidx0D,
1024                                             &c6_00,&c12_00);
1025
1026             /* LENNARD-JONES DISPERSION/REPULSION */
1027
1028             rinvsix          = _mm256_mul_pd(_mm256_mul_pd(rinvsq00,rinvsq00),rinvsq00);
1029             fvdw             = _mm256_mul_pd(_mm256_sub_pd(_mm256_mul_pd(c12_00,rinvsix),c6_00),_mm256_mul_pd(rinvsix,rinvsq00));
1030
1031             fscal            = fvdw;
1032
1033             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
1034
1035             /* Calculate temporary vectorial force */
1036             tx               = _mm256_mul_pd(fscal,dx00);
1037             ty               = _mm256_mul_pd(fscal,dy00);
1038             tz               = _mm256_mul_pd(fscal,dz00);
1039
1040             /* Update vectorial force */
1041             fix0             = _mm256_add_pd(fix0,tx);
1042             fiy0             = _mm256_add_pd(fiy0,ty);
1043             fiz0             = _mm256_add_pd(fiz0,tz);
1044
1045             fjx0             = _mm256_add_pd(fjx0,tx);
1046             fjy0             = _mm256_add_pd(fjy0,ty);
1047             fjz0             = _mm256_add_pd(fjz0,tz);
1048
1049             /**************************
1050              * CALCULATE INTERACTIONS *
1051              **************************/
1052
1053             /* Compute parameters for interactions between i and j atoms */
1054             qq10             = _mm256_mul_pd(iq1,jq0);
1055
1056             /* COULOMB ELECTROSTATICS */
1057             velec            = _mm256_mul_pd(qq10,rinv10);
1058             felec            = _mm256_mul_pd(velec,rinvsq10);
1059
1060             fscal            = felec;
1061
1062             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
1063
1064             /* Calculate temporary vectorial force */
1065             tx               = _mm256_mul_pd(fscal,dx10);
1066             ty               = _mm256_mul_pd(fscal,dy10);
1067             tz               = _mm256_mul_pd(fscal,dz10);
1068
1069             /* Update vectorial force */
1070             fix1             = _mm256_add_pd(fix1,tx);
1071             fiy1             = _mm256_add_pd(fiy1,ty);
1072             fiz1             = _mm256_add_pd(fiz1,tz);
1073
1074             fjx0             = _mm256_add_pd(fjx0,tx);
1075             fjy0             = _mm256_add_pd(fjy0,ty);
1076             fjz0             = _mm256_add_pd(fjz0,tz);
1077
1078             /**************************
1079              * CALCULATE INTERACTIONS *
1080              **************************/
1081
1082             /* Compute parameters for interactions between i and j atoms */
1083             qq20             = _mm256_mul_pd(iq2,jq0);
1084
1085             /* COULOMB ELECTROSTATICS */
1086             velec            = _mm256_mul_pd(qq20,rinv20);
1087             felec            = _mm256_mul_pd(velec,rinvsq20);
1088
1089             fscal            = felec;
1090
1091             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
1092
1093             /* Calculate temporary vectorial force */
1094             tx               = _mm256_mul_pd(fscal,dx20);
1095             ty               = _mm256_mul_pd(fscal,dy20);
1096             tz               = _mm256_mul_pd(fscal,dz20);
1097
1098             /* Update vectorial force */
1099             fix2             = _mm256_add_pd(fix2,tx);
1100             fiy2             = _mm256_add_pd(fiy2,ty);
1101             fiz2             = _mm256_add_pd(fiz2,tz);
1102
1103             fjx0             = _mm256_add_pd(fjx0,tx);
1104             fjy0             = _mm256_add_pd(fjy0,ty);
1105             fjz0             = _mm256_add_pd(fjz0,tz);
1106
1107             /**************************
1108              * CALCULATE INTERACTIONS *
1109              **************************/
1110
1111             /* Compute parameters for interactions between i and j atoms */
1112             qq30             = _mm256_mul_pd(iq3,jq0);
1113
1114             /* COULOMB ELECTROSTATICS */
1115             velec            = _mm256_mul_pd(qq30,rinv30);
1116             felec            = _mm256_mul_pd(velec,rinvsq30);
1117
1118             fscal            = felec;
1119
1120             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
1121
1122             /* Calculate temporary vectorial force */
1123             tx               = _mm256_mul_pd(fscal,dx30);
1124             ty               = _mm256_mul_pd(fscal,dy30);
1125             tz               = _mm256_mul_pd(fscal,dz30);
1126
1127             /* Update vectorial force */
1128             fix3             = _mm256_add_pd(fix3,tx);
1129             fiy3             = _mm256_add_pd(fiy3,ty);
1130             fiz3             = _mm256_add_pd(fiz3,tz);
1131
1132             fjx0             = _mm256_add_pd(fjx0,tx);
1133             fjy0             = _mm256_add_pd(fjy0,ty);
1134             fjz0             = _mm256_add_pd(fjz0,tz);
1135
1136             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
1137             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
1138             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
1139             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
1140
1141             gmx_mm256_decrement_1rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0);
1142
1143             /* Inner loop uses 108 flops */
1144         }
1145
1146         /* End of innermost loop */
1147
1148         gmx_mm256_update_iforce_4atom_swizzle_pd(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
1149                                                  f+i_coord_offset,fshift+i_shift_offset);
1150
1151         /* Increment number of inner iterations */
1152         inneriter                  += j_index_end - j_index_start;
1153
1154         /* Outer loop uses 24 flops */
1155     }
1156
1157     /* Increment number of outer iterations */
1158     outeriter        += nri;
1159
1160     /* Update outer/inner flops */
1161
1162     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_F,outeriter*24 + inneriter*108);
1163 }