src/gromacs/gmxlib/nonbonded/nb_kernel_avx_256_double/nb_kernel_ElecRFCut_VdwNone_GeomW3P1_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_ElecRFCut_VdwNone_GeomW3P1_VF_avx_256_double
  51  * Electrostatics interaction: ReactionField
  52  * VdW interaction:            None
  53  * Geometry:                   Water3-Particle
  54  * Calculate force/pot:        PotentialAndForce
  55  */
  56 void
  57 nb_kernel_ElecRFCut_VdwNone_GeomW3P1_VF_avx_256_double
  58                     (t_nblist                    * gmx_restrict       nlist,
  59                      rvec                        * gmx_restrict          xx,
  60                      rvec                        * gmx_restrict          ff,
  61                      struct t_forcerec           * gmx_restrict          fr,
  62                      t_mdatoms                   * gmx_restrict     mdatoms,
  63                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
  64                      t_nrnb                      * gmx_restrict        nrnb)
  65 {
  66     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
  67      * just 0 for non-waters.
  68      * Suffixes A,B,C,D refer to j loop unrolling done with AVX, e.g. for the four different
  69      * jnr indices corresponding to data put in the four positions in the SIMD register.
  70      */
  71     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
  72     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
  73     int              jnrA,jnrB,jnrC,jnrD;
  74     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
  75     int              jnrlistE,jnrlistF,jnrlistG,jnrlistH;
  76     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
  77     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  78     real             rcutoff_scalar;
  79     real             *shiftvec,*fshift,*x,*f;
  80     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD;
  81     real             scratch[4*DIM];
  82     __m256d          tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
  83     real *           vdwioffsetptr0;
  84     __m256d          ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  85     real *           vdwioffsetptr1;
  86     __m256d          ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
  87     real *           vdwioffsetptr2;
  88     __m256d          ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
  89     int              vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D;
  90     __m256d          jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  91     __m256d          dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
  92     __m256d          dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
  93     __m256d          dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
  94     __m256d          velec,felec,velecsum,facel,crf,krf,krf2;
  95     real             *charge;
  96     __m256d          dummy_mask,cutoff_mask;
  97     __m128           tmpmask0,tmpmask1;
  98     __m256d          signbit = _mm256_castsi256_pd( _mm256_set1_epi32(0x80000000) );
  99     __m256d          one     = _mm256_set1_pd(1.0);
 100     __m256d          two     = _mm256_set1_pd(2.0);
 101     x                = xx[0];
 102     f                = ff[0];
 103
 104     nri              = nlist->nri;
 105     iinr             = nlist->iinr;
 106     jindex           = nlist->jindex;
 107     jjnr             = nlist->jjnr;
 108     shiftidx         = nlist->shift;
 109     gid              = nlist->gid;
 110     shiftvec         = fr->shift_vec[0];
 111     fshift           = fr->fshift[0];
 112     facel            = _mm256_set1_pd(fr->ic->epsfac);
 113     charge           = mdatoms->chargeA;
 114     krf              = _mm256_set1_pd(fr->ic->k_rf);
 115     krf2             = _mm256_set1_pd(fr->ic->k_rf*2.0);
 116     crf              = _mm256_set1_pd(fr->ic->c_rf);
 117
 118     /* Setup water-specific parameters */
 119     inr              = nlist->iinr[0];
 120     iq0              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+0]));
 121     iq1              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+1]));
 122     iq2              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+2]));
 123
 124     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 125     rcutoff_scalar   = fr->ic->rcoulomb;
 126     rcutoff          = _mm256_set1_pd(rcutoff_scalar);
 127     rcutoff2         = _mm256_mul_pd(rcutoff,rcutoff);
 128
 129     /* Avoid stupid compiler warnings */
 130     jnrA = jnrB = jnrC = jnrD = 0;
 131     j_coord_offsetA = 0;
 132     j_coord_offsetB = 0;
 133     j_coord_offsetC = 0;
 134     j_coord_offsetD = 0;
 135
 136     outeriter        = 0;
 137     inneriter        = 0;
 138
 139     for(iidx=0;iidx<4*DIM;iidx++)
 140     {
 141         scratch[iidx] = 0.0;
 142     }
 143
 144     /* Start outer loop over neighborlists */
 145     for(iidx=0; iidx<nri; iidx++)
 146     {
 147         /* Load shift vector for this list */
 148         i_shift_offset   = DIM*shiftidx[iidx];
 149
 150         /* Load limits for loop over neighbors */
 151         j_index_start    = jindex[iidx];
 152         j_index_end      = jindex[iidx+1];
 153
 154         /* Get outer coordinate index */
 155         inr              = iinr[iidx];
 156         i_coord_offset   = DIM*inr;
 157
 158         /* Load i particle coords and add shift vector */
 159         gmx_mm256_load_shift_and_3rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset,
 160                                                     &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2);
 161
 162         fix0             = _mm256_setzero_pd();
 163         fiy0             = _mm256_setzero_pd();
 164         fiz0             = _mm256_setzero_pd();
 165         fix1             = _mm256_setzero_pd();
 166         fiy1             = _mm256_setzero_pd();
 167         fiz1             = _mm256_setzero_pd();
 168         fix2             = _mm256_setzero_pd();
 169         fiy2             = _mm256_setzero_pd();
 170         fiz2             = _mm256_setzero_pd();
 171
 172         /* Reset potential sums */
 173         velecsum         = _mm256_setzero_pd();
 174
 175         /* Start inner kernel loop */
 176         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4)
 177         {
 178
 179             /* Get j neighbor index, and coordinate index */
 180             jnrA             = jjnr[jidx];
 181             jnrB             = jjnr[jidx+1];
 182             jnrC             = jjnr[jidx+2];
 183             jnrD             = jjnr[jidx+3];
 184             j_coord_offsetA  = DIM*jnrA;
 185             j_coord_offsetB  = DIM*jnrB;
 186             j_coord_offsetC  = DIM*jnrC;
 187             j_coord_offsetD  = DIM*jnrD;
 188
 189             /* load j atom coordinates */
 190             gmx_mm256_load_1rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
 191                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 192                                                  &jx0,&jy0,&jz0);
 193
 194             /* Calculate displacement vector */
 195             dx00             = _mm256_sub_pd(ix0,jx0);
 196             dy00             = _mm256_sub_pd(iy0,jy0);
 197             dz00             = _mm256_sub_pd(iz0,jz0);
 198             dx10             = _mm256_sub_pd(ix1,jx0);
 199             dy10             = _mm256_sub_pd(iy1,jy0);
 200             dz10             = _mm256_sub_pd(iz1,jz0);
 201             dx20             = _mm256_sub_pd(ix2,jx0);
 202             dy20             = _mm256_sub_pd(iy2,jy0);
 203             dz20             = _mm256_sub_pd(iz2,jz0);
 204
 205             /* Calculate squared distance and things based on it */
 206             rsq00            = gmx_mm256_calc_rsq_pd(dx00,dy00,dz00);
 207             rsq10            = gmx_mm256_calc_rsq_pd(dx10,dy10,dz10);
 208             rsq20            = gmx_mm256_calc_rsq_pd(dx20,dy20,dz20);
 209
 210             rinv00           = avx256_invsqrt_d(rsq00);
 211             rinv10           = avx256_invsqrt_d(rsq10);
 212             rinv20           = avx256_invsqrt_d(rsq20);
 213
 214             rinvsq00         = _mm256_mul_pd(rinv00,rinv00);
 215             rinvsq10         = _mm256_mul_pd(rinv10,rinv10);
 216             rinvsq20         = _mm256_mul_pd(rinv20,rinv20);
 217
 218             /* Load parameters for j particles */
 219             jq0              = gmx_mm256_load_4real_swizzle_pd(charge+jnrA+0,charge+jnrB+0,
 220                                                                  charge+jnrC+0,charge+jnrD+0);
 221
 222             fjx0             = _mm256_setzero_pd();
 223             fjy0             = _mm256_setzero_pd();
 224             fjz0             = _mm256_setzero_pd();
 225
 226             /**************************
 227              * CALCULATE INTERACTIONS *
 228              **************************/
 229
 230             if (gmx_mm256_any_lt(rsq00,rcutoff2))
 231             {
 232
 233             /* Compute parameters for interactions between i and j atoms */
 234             qq00             = _mm256_mul_pd(iq0,jq0);
 235
 236             /* REACTION-FIELD ELECTROSTATICS */
 237             velec            = _mm256_mul_pd(qq00,_mm256_sub_pd(_mm256_add_pd(rinv00,_mm256_mul_pd(krf,rsq00)),crf));
 238             felec            = _mm256_mul_pd(qq00,_mm256_sub_pd(_mm256_mul_pd(rinv00,rinvsq00),krf2));
 239
 240             cutoff_mask      = _mm256_cmp_pd(rsq00,rcutoff2,_CMP_LT_OQ);
 241
 242             /* Update potential sum for this i atom from the interaction with this j atom. */
 243             velec            = _mm256_and_pd(velec,cutoff_mask);
 244             velecsum         = _mm256_add_pd(velecsum,velec);
 245
 246             fscal            = felec;
 247
 248             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 249
 250             /* Calculate temporary vectorial force */
 251             tx               = _mm256_mul_pd(fscal,dx00);
 252             ty               = _mm256_mul_pd(fscal,dy00);
 253             tz               = _mm256_mul_pd(fscal,dz00);
 254
 255             /* Update vectorial force */
 256             fix0             = _mm256_add_pd(fix0,tx);
 257             fiy0             = _mm256_add_pd(fiy0,ty);
 258             fiz0             = _mm256_add_pd(fiz0,tz);
 259
 260             fjx0             = _mm256_add_pd(fjx0,tx);
 261             fjy0             = _mm256_add_pd(fjy0,ty);
 262             fjz0             = _mm256_add_pd(fjz0,tz);
 263
 264             }
 265
 266             /**************************
 267              * CALCULATE INTERACTIONS *
 268              **************************/
 269
 270             if (gmx_mm256_any_lt(rsq10,rcutoff2))
 271             {
 272
 273             /* Compute parameters for interactions between i and j atoms */
 274             qq10             = _mm256_mul_pd(iq1,jq0);
 275
 276             /* REACTION-FIELD ELECTROSTATICS */
 277             velec            = _mm256_mul_pd(qq10,_mm256_sub_pd(_mm256_add_pd(rinv10,_mm256_mul_pd(krf,rsq10)),crf));
 278             felec            = _mm256_mul_pd(qq10,_mm256_sub_pd(_mm256_mul_pd(rinv10,rinvsq10),krf2));
 279
 280             cutoff_mask      = _mm256_cmp_pd(rsq10,rcutoff2,_CMP_LT_OQ);
 281
 282             /* Update potential sum for this i atom from the interaction with this j atom. */
 283             velec            = _mm256_and_pd(velec,cutoff_mask);
 284             velecsum         = _mm256_add_pd(velecsum,velec);
 285
 286             fscal            = felec;
 287
 288             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 289
 290             /* Calculate temporary vectorial force */
 291             tx               = _mm256_mul_pd(fscal,dx10);
 292             ty               = _mm256_mul_pd(fscal,dy10);
 293             tz               = _mm256_mul_pd(fscal,dz10);
 294
 295             /* Update vectorial force */
 296             fix1             = _mm256_add_pd(fix1,tx);
 297             fiy1             = _mm256_add_pd(fiy1,ty);
 298             fiz1             = _mm256_add_pd(fiz1,tz);
 299
 300             fjx0             = _mm256_add_pd(fjx0,tx);
 301             fjy0             = _mm256_add_pd(fjy0,ty);
 302             fjz0             = _mm256_add_pd(fjz0,tz);
 303
 304             }
 305
 306             /**************************
 307              * CALCULATE INTERACTIONS *
 308              **************************/
 309
 310             if (gmx_mm256_any_lt(rsq20,rcutoff2))
 311             {
 312
 313             /* Compute parameters for interactions between i and j atoms */
 314             qq20             = _mm256_mul_pd(iq2,jq0);
 315
 316             /* REACTION-FIELD ELECTROSTATICS */
 317             velec            = _mm256_mul_pd(qq20,_mm256_sub_pd(_mm256_add_pd(rinv20,_mm256_mul_pd(krf,rsq20)),crf));
 318             felec            = _mm256_mul_pd(qq20,_mm256_sub_pd(_mm256_mul_pd(rinv20,rinvsq20),krf2));
 319
 320             cutoff_mask      = _mm256_cmp_pd(rsq20,rcutoff2,_CMP_LT_OQ);
 321
 322             /* Update potential sum for this i atom from the interaction with this j atom. */
 323             velec            = _mm256_and_pd(velec,cutoff_mask);
 324             velecsum         = _mm256_add_pd(velecsum,velec);
 325
 326             fscal            = felec;
 327
 328             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 329
 330             /* Calculate temporary vectorial force */
 331             tx               = _mm256_mul_pd(fscal,dx20);
 332             ty               = _mm256_mul_pd(fscal,dy20);
 333             tz               = _mm256_mul_pd(fscal,dz20);
 334
 335             /* Update vectorial force */
 336             fix2             = _mm256_add_pd(fix2,tx);
 337             fiy2             = _mm256_add_pd(fiy2,ty);
 338             fiz2             = _mm256_add_pd(fiz2,tz);
 339
 340             fjx0             = _mm256_add_pd(fjx0,tx);
 341             fjy0             = _mm256_add_pd(fjy0,ty);
 342             fjz0             = _mm256_add_pd(fjz0,tz);
 343
 344             }
 345
 346             fjptrA             = f+j_coord_offsetA;
 347             fjptrB             = f+j_coord_offsetB;
 348             fjptrC             = f+j_coord_offsetC;
 349             fjptrD             = f+j_coord_offsetD;
 350
 351             gmx_mm256_decrement_1rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0);
 352
 353             /* Inner loop uses 111 flops */
 354         }
 355
 356         if(jidx<j_index_end)
 357         {
 358
 359             /* Get j neighbor index, and coordinate index */
 360             jnrlistA         = jjnr[jidx];
 361             jnrlistB         = jjnr[jidx+1];
 362             jnrlistC         = jjnr[jidx+2];
 363             jnrlistD         = jjnr[jidx+3];
 364             /* Sign of each element will be negative for non-real atoms.
 365              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
 366              * so use it as val = _mm_andnot_pd(mask,val) to clear dummy entries.
 367              */
 368             tmpmask0 = gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128()));
 369
 370             tmpmask1 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(3,3,2,2));
 371             tmpmask0 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(1,1,0,0));
 372             dummy_mask = _mm256_castps_pd(gmx_mm256_set_m128(tmpmask1,tmpmask0));
 373
 374             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
 375             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
 376             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
 377             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
 378             j_coord_offsetA  = DIM*jnrA;
 379             j_coord_offsetB  = DIM*jnrB;
 380             j_coord_offsetC  = DIM*jnrC;
 381             j_coord_offsetD  = DIM*jnrD;
 382
 383             /* load j atom coordinates */
 384             gmx_mm256_load_1rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
 385                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 386                                                  &jx0,&jy0,&jz0);
 387
 388             /* Calculate displacement vector */
 389             dx00             = _mm256_sub_pd(ix0,jx0);
 390             dy00             = _mm256_sub_pd(iy0,jy0);
 391             dz00             = _mm256_sub_pd(iz0,jz0);
 392             dx10             = _mm256_sub_pd(ix1,jx0);
 393             dy10             = _mm256_sub_pd(iy1,jy0);
 394             dz10             = _mm256_sub_pd(iz1,jz0);
 395             dx20             = _mm256_sub_pd(ix2,jx0);
 396             dy20             = _mm256_sub_pd(iy2,jy0);
 397             dz20             = _mm256_sub_pd(iz2,jz0);
 398
 399             /* Calculate squared distance and things based on it */
 400             rsq00            = gmx_mm256_calc_rsq_pd(dx00,dy00,dz00);
 401             rsq10            = gmx_mm256_calc_rsq_pd(dx10,dy10,dz10);
 402             rsq20            = gmx_mm256_calc_rsq_pd(dx20,dy20,dz20);
 403
 404             rinv00           = avx256_invsqrt_d(rsq00);
 405             rinv10           = avx256_invsqrt_d(rsq10);
 406             rinv20           = avx256_invsqrt_d(rsq20);
 407
 408             rinvsq00         = _mm256_mul_pd(rinv00,rinv00);
 409             rinvsq10         = _mm256_mul_pd(rinv10,rinv10);
 410             rinvsq20         = _mm256_mul_pd(rinv20,rinv20);
 411
 412             /* Load parameters for j particles */
 413             jq0              = gmx_mm256_load_4real_swizzle_pd(charge+jnrA+0,charge+jnrB+0,
 414                                                                  charge+jnrC+0,charge+jnrD+0);
 415
 416             fjx0             = _mm256_setzero_pd();
 417             fjy0             = _mm256_setzero_pd();
 418             fjz0             = _mm256_setzero_pd();
 419
 420             /**************************
 421              * CALCULATE INTERACTIONS *
 422              **************************/
 423
 424             if (gmx_mm256_any_lt(rsq00,rcutoff2))
 425             {
 426
 427             /* Compute parameters for interactions between i and j atoms */
 428             qq00             = _mm256_mul_pd(iq0,jq0);
 429
 430             /* REACTION-FIELD ELECTROSTATICS */
 431             velec            = _mm256_mul_pd(qq00,_mm256_sub_pd(_mm256_add_pd(rinv00,_mm256_mul_pd(krf,rsq00)),crf));
 432             felec            = _mm256_mul_pd(qq00,_mm256_sub_pd(_mm256_mul_pd(rinv00,rinvsq00),krf2));
 433
 434             cutoff_mask      = _mm256_cmp_pd(rsq00,rcutoff2,_CMP_LT_OQ);
 435
 436             /* Update potential sum for this i atom from the interaction with this j atom. */
 437             velec            = _mm256_and_pd(velec,cutoff_mask);
 438             velec            = _mm256_andnot_pd(dummy_mask,velec);
 439             velecsum         = _mm256_add_pd(velecsum,velec);
 440
 441             fscal            = felec;
 442
 443             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 444
 445             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
 446
 447             /* Calculate temporary vectorial force */
 448             tx               = _mm256_mul_pd(fscal,dx00);
 449             ty               = _mm256_mul_pd(fscal,dy00);
 450             tz               = _mm256_mul_pd(fscal,dz00);
 451
 452             /* Update vectorial force */
 453             fix0             = _mm256_add_pd(fix0,tx);
 454             fiy0             = _mm256_add_pd(fiy0,ty);
 455             fiz0             = _mm256_add_pd(fiz0,tz);
 456
 457             fjx0             = _mm256_add_pd(fjx0,tx);
 458             fjy0             = _mm256_add_pd(fjy0,ty);
 459             fjz0             = _mm256_add_pd(fjz0,tz);
 460
 461             }
 462
 463             /**************************
 464              * CALCULATE INTERACTIONS *
 465              **************************/
 466
 467             if (gmx_mm256_any_lt(rsq10,rcutoff2))
 468             {
 469
 470             /* Compute parameters for interactions between i and j atoms */
 471             qq10             = _mm256_mul_pd(iq1,jq0);
 472
 473             /* REACTION-FIELD ELECTROSTATICS */
 474             velec            = _mm256_mul_pd(qq10,_mm256_sub_pd(_mm256_add_pd(rinv10,_mm256_mul_pd(krf,rsq10)),crf));
 475             felec            = _mm256_mul_pd(qq10,_mm256_sub_pd(_mm256_mul_pd(rinv10,rinvsq10),krf2));
 476
 477             cutoff_mask      = _mm256_cmp_pd(rsq10,rcutoff2,_CMP_LT_OQ);
 478
 479             /* Update potential sum for this i atom from the interaction with this j atom. */
 480             velec            = _mm256_and_pd(velec,cutoff_mask);
 481             velec            = _mm256_andnot_pd(dummy_mask,velec);
 482             velecsum         = _mm256_add_pd(velecsum,velec);
 483
 484             fscal            = felec;
 485
 486             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 487
 488             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
 489
 490             /* Calculate temporary vectorial force */
 491             tx               = _mm256_mul_pd(fscal,dx10);
 492             ty               = _mm256_mul_pd(fscal,dy10);
 493             tz               = _mm256_mul_pd(fscal,dz10);
 494
 495             /* Update vectorial force */
 496             fix1             = _mm256_add_pd(fix1,tx);
 497             fiy1             = _mm256_add_pd(fiy1,ty);
 498             fiz1             = _mm256_add_pd(fiz1,tz);
 499
 500             fjx0             = _mm256_add_pd(fjx0,tx);
 501             fjy0             = _mm256_add_pd(fjy0,ty);
 502             fjz0             = _mm256_add_pd(fjz0,tz);
 503
 504             }
 505
 506             /**************************
 507              * CALCULATE INTERACTIONS *
 508              **************************/
 509
 510             if (gmx_mm256_any_lt(rsq20,rcutoff2))
 511             {
 512
 513             /* Compute parameters for interactions between i and j atoms */
 514             qq20             = _mm256_mul_pd(iq2,jq0);
 515
 516             /* REACTION-FIELD ELECTROSTATICS */
 517             velec            = _mm256_mul_pd(qq20,_mm256_sub_pd(_mm256_add_pd(rinv20,_mm256_mul_pd(krf,rsq20)),crf));
 518             felec            = _mm256_mul_pd(qq20,_mm256_sub_pd(_mm256_mul_pd(rinv20,rinvsq20),krf2));
 519
 520             cutoff_mask      = _mm256_cmp_pd(rsq20,rcutoff2,_CMP_LT_OQ);
 521
 522             /* Update potential sum for this i atom from the interaction with this j atom. */
 523             velec            = _mm256_and_pd(velec,cutoff_mask);
 524             velec            = _mm256_andnot_pd(dummy_mask,velec);
 525             velecsum         = _mm256_add_pd(velecsum,velec);
 526
 527             fscal            = felec;
 528
 529             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 530
 531             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
 532
 533             /* Calculate temporary vectorial force */
 534             tx               = _mm256_mul_pd(fscal,dx20);
 535             ty               = _mm256_mul_pd(fscal,dy20);
 536             tz               = _mm256_mul_pd(fscal,dz20);
 537
 538             /* Update vectorial force */
 539             fix2             = _mm256_add_pd(fix2,tx);
 540             fiy2             = _mm256_add_pd(fiy2,ty);
 541             fiz2             = _mm256_add_pd(fiz2,tz);
 542
 543             fjx0             = _mm256_add_pd(fjx0,tx);
 544             fjy0             = _mm256_add_pd(fjy0,ty);
 545             fjz0             = _mm256_add_pd(fjz0,tz);
 546
 547             }
 548
 549             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
 550             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
 551             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
 552             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
 553
 554             gmx_mm256_decrement_1rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0);
 555
 556             /* Inner loop uses 111 flops */
 557         }
 558
 559         /* End of innermost loop */
 560
 561         gmx_mm256_update_iforce_3atom_swizzle_pd(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,
 562                                                  f+i_coord_offset,fshift+i_shift_offset);
 563
 564         ggid                        = gid[iidx];
 565         /* Update potential energies */
 566         gmx_mm256_update_1pot_pd(velecsum,kernel_data->energygrp_elec+ggid);
 567
 568         /* Increment number of inner iterations */
 569         inneriter                  += j_index_end - j_index_start;
 570
 571         /* Outer loop uses 19 flops */
 572     }
 573
 574     /* Increment number of outer iterations */
 575     outeriter        += nri;
 576
 577     /* Update outer/inner flops */
 578
 579     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_W3_VF,outeriter*19 + inneriter*111);
 580 }
 581 /*
 582  * Gromacs nonbonded kernel:   nb_kernel_ElecRFCut_VdwNone_GeomW3P1_F_avx_256_double
 583  * Electrostatics interaction: ReactionField
 584  * VdW interaction:            None
 585  * Geometry:                   Water3-Particle
 586  * Calculate force/pot:        Force
 587  */
 588 void
 589 nb_kernel_ElecRFCut_VdwNone_GeomW3P1_F_avx_256_double
 590                     (t_nblist                    * gmx_restrict       nlist,
 591                      rvec                        * gmx_restrict          xx,
 592                      rvec                        * gmx_restrict          ff,
 593                      struct t_forcerec           * gmx_restrict          fr,
 594                      t_mdatoms                   * gmx_restrict     mdatoms,
 595                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 596                      t_nrnb                      * gmx_restrict        nrnb)
 597 {
 598     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
 599      * just 0 for non-waters.
 600      * Suffixes A,B,C,D refer to j loop unrolling done with AVX, e.g. for the four different
 601      * jnr indices corresponding to data put in the four positions in the SIMD register.
 602      */
 603     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
 604     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
 605     int              jnrA,jnrB,jnrC,jnrD;
 606     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
 607     int              jnrlistE,jnrlistF,jnrlistG,jnrlistH;
 608     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
 609     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 610     real             rcutoff_scalar;
 611     real             *shiftvec,*fshift,*x,*f;
 612     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD;
 613     real             scratch[4*DIM];
 614     __m256d          tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
 615     real *           vdwioffsetptr0;
 616     __m256d          ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 617     real *           vdwioffsetptr1;
 618     __m256d          ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
 619     real *           vdwioffsetptr2;
 620     __m256d          ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
 621     int              vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D;
 622     __m256d          jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 623     __m256d          dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
 624     __m256d          dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
 625     __m256d          dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
 626     __m256d          velec,felec,velecsum,facel,crf,krf,krf2;
 627     real             *charge;
 628     __m256d          dummy_mask,cutoff_mask;
 629     __m128           tmpmask0,tmpmask1;
 630     __m256d          signbit = _mm256_castsi256_pd( _mm256_set1_epi32(0x80000000) );
 631     __m256d          one     = _mm256_set1_pd(1.0);
 632     __m256d          two     = _mm256_set1_pd(2.0);
 633     x                = xx[0];
 634     f                = ff[0];
 635
 636     nri              = nlist->nri;
 637     iinr             = nlist->iinr;
 638     jindex           = nlist->jindex;
 639     jjnr             = nlist->jjnr;
 640     shiftidx         = nlist->shift;
 641     gid              = nlist->gid;
 642     shiftvec         = fr->shift_vec[0];
 643     fshift           = fr->fshift[0];
 644     facel            = _mm256_set1_pd(fr->ic->epsfac);
 645     charge           = mdatoms->chargeA;
 646     krf              = _mm256_set1_pd(fr->ic->k_rf);
 647     krf2             = _mm256_set1_pd(fr->ic->k_rf*2.0);
 648     crf              = _mm256_set1_pd(fr->ic->c_rf);
 649
 650     /* Setup water-specific parameters */
 651     inr              = nlist->iinr[0];
 652     iq0              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+0]));
 653     iq1              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+1]));
 654     iq2              = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+2]));
 655
 656     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 657     rcutoff_scalar   = fr->ic->rcoulomb;
 658     rcutoff          = _mm256_set1_pd(rcutoff_scalar);
 659     rcutoff2         = _mm256_mul_pd(rcutoff,rcutoff);
 660
 661     /* Avoid stupid compiler warnings */
 662     jnrA = jnrB = jnrC = jnrD = 0;
 663     j_coord_offsetA = 0;
 664     j_coord_offsetB = 0;
 665     j_coord_offsetC = 0;
 666     j_coord_offsetD = 0;
 667
 668     outeriter        = 0;
 669     inneriter        = 0;
 670
 671     for(iidx=0;iidx<4*DIM;iidx++)
 672     {
 673         scratch[iidx] = 0.0;
 674     }
 675
 676     /* Start outer loop over neighborlists */
 677     for(iidx=0; iidx<nri; iidx++)
 678     {
 679         /* Load shift vector for this list */
 680         i_shift_offset   = DIM*shiftidx[iidx];
 681
 682         /* Load limits for loop over neighbors */
 683         j_index_start    = jindex[iidx];
 684         j_index_end      = jindex[iidx+1];
 685
 686         /* Get outer coordinate index */
 687         inr              = iinr[iidx];
 688         i_coord_offset   = DIM*inr;
 689
 690         /* Load i particle coords and add shift vector */
 691         gmx_mm256_load_shift_and_3rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset,
 692                                                     &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2);
 693
 694         fix0             = _mm256_setzero_pd();
 695         fiy0             = _mm256_setzero_pd();
 696         fiz0             = _mm256_setzero_pd();
 697         fix1             = _mm256_setzero_pd();
 698         fiy1             = _mm256_setzero_pd();
 699         fiz1             = _mm256_setzero_pd();
 700         fix2             = _mm256_setzero_pd();
 701         fiy2             = _mm256_setzero_pd();
 702         fiz2             = _mm256_setzero_pd();
 703
 704         /* Start inner kernel loop */
 705         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4)
 706         {
 707
 708             /* Get j neighbor index, and coordinate index */
 709             jnrA             = jjnr[jidx];
 710             jnrB             = jjnr[jidx+1];
 711             jnrC             = jjnr[jidx+2];
 712             jnrD             = jjnr[jidx+3];
 713             j_coord_offsetA  = DIM*jnrA;
 714             j_coord_offsetB  = DIM*jnrB;
 715             j_coord_offsetC  = DIM*jnrC;
 716             j_coord_offsetD  = DIM*jnrD;
 717
 718             /* load j atom coordinates */
 719             gmx_mm256_load_1rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
 720                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 721                                                  &jx0,&jy0,&jz0);
 722
 723             /* Calculate displacement vector */
 724             dx00             = _mm256_sub_pd(ix0,jx0);
 725             dy00             = _mm256_sub_pd(iy0,jy0);
 726             dz00             = _mm256_sub_pd(iz0,jz0);
 727             dx10             = _mm256_sub_pd(ix1,jx0);
 728             dy10             = _mm256_sub_pd(iy1,jy0);
 729             dz10             = _mm256_sub_pd(iz1,jz0);
 730             dx20             = _mm256_sub_pd(ix2,jx0);
 731             dy20             = _mm256_sub_pd(iy2,jy0);
 732             dz20             = _mm256_sub_pd(iz2,jz0);
 733
 734             /* Calculate squared distance and things based on it */
 735             rsq00            = gmx_mm256_calc_rsq_pd(dx00,dy00,dz00);
 736             rsq10            = gmx_mm256_calc_rsq_pd(dx10,dy10,dz10);
 737             rsq20            = gmx_mm256_calc_rsq_pd(dx20,dy20,dz20);
 738
 739             rinv00           = avx256_invsqrt_d(rsq00);
 740             rinv10           = avx256_invsqrt_d(rsq10);
 741             rinv20           = avx256_invsqrt_d(rsq20);
 742
 743             rinvsq00         = _mm256_mul_pd(rinv00,rinv00);
 744             rinvsq10         = _mm256_mul_pd(rinv10,rinv10);
 745             rinvsq20         = _mm256_mul_pd(rinv20,rinv20);
 746
 747             /* Load parameters for j particles */
 748             jq0              = gmx_mm256_load_4real_swizzle_pd(charge+jnrA+0,charge+jnrB+0,
 749                                                                  charge+jnrC+0,charge+jnrD+0);
 750
 751             fjx0             = _mm256_setzero_pd();
 752             fjy0             = _mm256_setzero_pd();
 753             fjz0             = _mm256_setzero_pd();
 754
 755             /**************************
 756              * CALCULATE INTERACTIONS *
 757              **************************/
 758
 759             if (gmx_mm256_any_lt(rsq00,rcutoff2))
 760             {
 761
 762             /* Compute parameters for interactions between i and j atoms */
 763             qq00             = _mm256_mul_pd(iq0,jq0);
 764
 765             /* REACTION-FIELD ELECTROSTATICS */
 766             felec            = _mm256_mul_pd(qq00,_mm256_sub_pd(_mm256_mul_pd(rinv00,rinvsq00),krf2));
 767
 768             cutoff_mask      = _mm256_cmp_pd(rsq00,rcutoff2,_CMP_LT_OQ);
 769
 770             fscal            = felec;
 771
 772             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 773
 774             /* Calculate temporary vectorial force */
 775             tx               = _mm256_mul_pd(fscal,dx00);
 776             ty               = _mm256_mul_pd(fscal,dy00);
 777             tz               = _mm256_mul_pd(fscal,dz00);
 778
 779             /* Update vectorial force */
 780             fix0             = _mm256_add_pd(fix0,tx);
 781             fiy0             = _mm256_add_pd(fiy0,ty);
 782             fiz0             = _mm256_add_pd(fiz0,tz);
 783
 784             fjx0             = _mm256_add_pd(fjx0,tx);
 785             fjy0             = _mm256_add_pd(fjy0,ty);
 786             fjz0             = _mm256_add_pd(fjz0,tz);
 787
 788             }
 789
 790             /**************************
 791              * CALCULATE INTERACTIONS *
 792              **************************/
 793
 794             if (gmx_mm256_any_lt(rsq10,rcutoff2))
 795             {
 796
 797             /* Compute parameters for interactions between i and j atoms */
 798             qq10             = _mm256_mul_pd(iq1,jq0);
 799
 800             /* REACTION-FIELD ELECTROSTATICS */
 801             felec            = _mm256_mul_pd(qq10,_mm256_sub_pd(_mm256_mul_pd(rinv10,rinvsq10),krf2));
 802
 803             cutoff_mask      = _mm256_cmp_pd(rsq10,rcutoff2,_CMP_LT_OQ);
 804
 805             fscal            = felec;
 806
 807             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 808
 809             /* Calculate temporary vectorial force */
 810             tx               = _mm256_mul_pd(fscal,dx10);
 811             ty               = _mm256_mul_pd(fscal,dy10);
 812             tz               = _mm256_mul_pd(fscal,dz10);
 813
 814             /* Update vectorial force */
 815             fix1             = _mm256_add_pd(fix1,tx);
 816             fiy1             = _mm256_add_pd(fiy1,ty);
 817             fiz1             = _mm256_add_pd(fiz1,tz);
 818
 819             fjx0             = _mm256_add_pd(fjx0,tx);
 820             fjy0             = _mm256_add_pd(fjy0,ty);
 821             fjz0             = _mm256_add_pd(fjz0,tz);
 822
 823             }
 824
 825             /**************************
 826              * CALCULATE INTERACTIONS *
 827              **************************/
 828
 829             if (gmx_mm256_any_lt(rsq20,rcutoff2))
 830             {
 831
 832             /* Compute parameters for interactions between i and j atoms */
 833             qq20             = _mm256_mul_pd(iq2,jq0);
 834
 835             /* REACTION-FIELD ELECTROSTATICS */
 836             felec            = _mm256_mul_pd(qq20,_mm256_sub_pd(_mm256_mul_pd(rinv20,rinvsq20),krf2));
 837
 838             cutoff_mask      = _mm256_cmp_pd(rsq20,rcutoff2,_CMP_LT_OQ);
 839
 840             fscal            = felec;
 841
 842             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 843
 844             /* Calculate temporary vectorial force */
 845             tx               = _mm256_mul_pd(fscal,dx20);
 846             ty               = _mm256_mul_pd(fscal,dy20);
 847             tz               = _mm256_mul_pd(fscal,dz20);
 848
 849             /* Update vectorial force */
 850             fix2             = _mm256_add_pd(fix2,tx);
 851             fiy2             = _mm256_add_pd(fiy2,ty);
 852             fiz2             = _mm256_add_pd(fiz2,tz);
 853
 854             fjx0             = _mm256_add_pd(fjx0,tx);
 855             fjy0             = _mm256_add_pd(fjy0,ty);
 856             fjz0             = _mm256_add_pd(fjz0,tz);
 857
 858             }
 859
 860             fjptrA             = f+j_coord_offsetA;
 861             fjptrB             = f+j_coord_offsetB;
 862             fjptrC             = f+j_coord_offsetC;
 863             fjptrD             = f+j_coord_offsetD;
 864
 865             gmx_mm256_decrement_1rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0);
 866
 867             /* Inner loop uses 93 flops */
 868         }
 869
 870         if(jidx<j_index_end)
 871         {
 872
 873             /* Get j neighbor index, and coordinate index */
 874             jnrlistA         = jjnr[jidx];
 875             jnrlistB         = jjnr[jidx+1];
 876             jnrlistC         = jjnr[jidx+2];
 877             jnrlistD         = jjnr[jidx+3];
 878             /* Sign of each element will be negative for non-real atoms.
 879              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
 880              * so use it as val = _mm_andnot_pd(mask,val) to clear dummy entries.
 881              */
 882             tmpmask0 = gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128()));
 883
 884             tmpmask1 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(3,3,2,2));
 885             tmpmask0 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(1,1,0,0));
 886             dummy_mask = _mm256_castps_pd(gmx_mm256_set_m128(tmpmask1,tmpmask0));
 887
 888             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
 889             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
 890             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
 891             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
 892             j_coord_offsetA  = DIM*jnrA;
 893             j_coord_offsetB  = DIM*jnrB;
 894             j_coord_offsetC  = DIM*jnrC;
 895             j_coord_offsetD  = DIM*jnrD;
 896
 897             /* load j atom coordinates */
 898             gmx_mm256_load_1rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
 899                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 900                                                  &jx0,&jy0,&jz0);
 901
 902             /* Calculate displacement vector */
 903             dx00             = _mm256_sub_pd(ix0,jx0);
 904             dy00             = _mm256_sub_pd(iy0,jy0);
 905             dz00             = _mm256_sub_pd(iz0,jz0);
 906             dx10             = _mm256_sub_pd(ix1,jx0);
 907             dy10             = _mm256_sub_pd(iy1,jy0);
 908             dz10             = _mm256_sub_pd(iz1,jz0);
 909             dx20             = _mm256_sub_pd(ix2,jx0);
 910             dy20             = _mm256_sub_pd(iy2,jy0);
 911             dz20             = _mm256_sub_pd(iz2,jz0);
 912
 913             /* Calculate squared distance and things based on it */
 914             rsq00            = gmx_mm256_calc_rsq_pd(dx00,dy00,dz00);
 915             rsq10            = gmx_mm256_calc_rsq_pd(dx10,dy10,dz10);
 916             rsq20            = gmx_mm256_calc_rsq_pd(dx20,dy20,dz20);
 917
 918             rinv00           = avx256_invsqrt_d(rsq00);
 919             rinv10           = avx256_invsqrt_d(rsq10);
 920             rinv20           = avx256_invsqrt_d(rsq20);
 921
 922             rinvsq00         = _mm256_mul_pd(rinv00,rinv00);
 923             rinvsq10         = _mm256_mul_pd(rinv10,rinv10);
 924             rinvsq20         = _mm256_mul_pd(rinv20,rinv20);
 925
 926             /* Load parameters for j particles */
 927             jq0              = gmx_mm256_load_4real_swizzle_pd(charge+jnrA+0,charge+jnrB+0,
 928                                                                  charge+jnrC+0,charge+jnrD+0);
 929
 930             fjx0             = _mm256_setzero_pd();
 931             fjy0             = _mm256_setzero_pd();
 932             fjz0             = _mm256_setzero_pd();
 933
 934             /**************************
 935              * CALCULATE INTERACTIONS *
 936              **************************/
 937
 938             if (gmx_mm256_any_lt(rsq00,rcutoff2))
 939             {
 940
 941             /* Compute parameters for interactions between i and j atoms */
 942             qq00             = _mm256_mul_pd(iq0,jq0);
 943
 944             /* REACTION-FIELD ELECTROSTATICS */
 945             felec            = _mm256_mul_pd(qq00,_mm256_sub_pd(_mm256_mul_pd(rinv00,rinvsq00),krf2));
 946
 947             cutoff_mask      = _mm256_cmp_pd(rsq00,rcutoff2,_CMP_LT_OQ);
 948
 949             fscal            = felec;
 950
 951             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 952
 953             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
 954
 955             /* Calculate temporary vectorial force */
 956             tx               = _mm256_mul_pd(fscal,dx00);
 957             ty               = _mm256_mul_pd(fscal,dy00);
 958             tz               = _mm256_mul_pd(fscal,dz00);
 959
 960             /* Update vectorial force */
 961             fix0             = _mm256_add_pd(fix0,tx);
 962             fiy0             = _mm256_add_pd(fiy0,ty);
 963             fiz0             = _mm256_add_pd(fiz0,tz);
 964
 965             fjx0             = _mm256_add_pd(fjx0,tx);
 966             fjy0             = _mm256_add_pd(fjy0,ty);
 967             fjz0             = _mm256_add_pd(fjz0,tz);
 968
 969             }
 970
 971             /**************************
 972              * CALCULATE INTERACTIONS *
 973              **************************/
 974
 975             if (gmx_mm256_any_lt(rsq10,rcutoff2))
 976             {
 977
 978             /* Compute parameters for interactions between i and j atoms */
 979             qq10             = _mm256_mul_pd(iq1,jq0);
 980
 981             /* REACTION-FIELD ELECTROSTATICS */
 982             felec            = _mm256_mul_pd(qq10,_mm256_sub_pd(_mm256_mul_pd(rinv10,rinvsq10),krf2));
 983
 984             cutoff_mask      = _mm256_cmp_pd(rsq10,rcutoff2,_CMP_LT_OQ);
 985
 986             fscal            = felec;
 987
 988             fscal            = _mm256_and_pd(fscal,cutoff_mask);
 989
 990             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
 991
 992             /* Calculate temporary vectorial force */
 993             tx               = _mm256_mul_pd(fscal,dx10);
 994             ty               = _mm256_mul_pd(fscal,dy10);
 995             tz               = _mm256_mul_pd(fscal,dz10);
 996
 997             /* Update vectorial force */
 998             fix1             = _mm256_add_pd(fix1,tx);
 999             fiy1             = _mm256_add_pd(fiy1,ty);
1000             fiz1             = _mm256_add_pd(fiz1,tz);
1001
1002             fjx0             = _mm256_add_pd(fjx0,tx);
1003             fjy0             = _mm256_add_pd(fjy0,ty);
1004             fjz0             = _mm256_add_pd(fjz0,tz);
1005
1006             }
1007
1008             /**************************
1009              * CALCULATE INTERACTIONS *
1010              **************************/
1011
1012             if (gmx_mm256_any_lt(rsq20,rcutoff2))
1013             {
1014
1015             /* Compute parameters for interactions between i and j atoms */
1016             qq20             = _mm256_mul_pd(iq2,jq0);
1017
1018             /* REACTION-FIELD ELECTROSTATICS */
1019             felec            = _mm256_mul_pd(qq20,_mm256_sub_pd(_mm256_mul_pd(rinv20,rinvsq20),krf2));
1020
1021             cutoff_mask      = _mm256_cmp_pd(rsq20,rcutoff2,_CMP_LT_OQ);
1022
1023             fscal            = felec;
1024
1025             fscal            = _mm256_and_pd(fscal,cutoff_mask);
1026
1027             fscal            = _mm256_andnot_pd(dummy_mask,fscal);
1028
1029             /* Calculate temporary vectorial force */
1030             tx               = _mm256_mul_pd(fscal,dx20);
1031             ty               = _mm256_mul_pd(fscal,dy20);
1032             tz               = _mm256_mul_pd(fscal,dz20);
1033
1034             /* Update vectorial force */
1035             fix2             = _mm256_add_pd(fix2,tx);
1036             fiy2             = _mm256_add_pd(fiy2,ty);
1037             fiz2             = _mm256_add_pd(fiz2,tz);
1038
1039             fjx0             = _mm256_add_pd(fjx0,tx);
1040             fjy0             = _mm256_add_pd(fjy0,ty);
1041             fjz0             = _mm256_add_pd(fjz0,tz);
1042
1043             }
1044
1045             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
1046             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
1047             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
1048             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
1049
1050             gmx_mm256_decrement_1rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0);
1051
1052             /* Inner loop uses 93 flops */
1053         }
1054
1055         /* End of innermost loop */
1056
1057         gmx_mm256_update_iforce_3atom_swizzle_pd(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,
1058                                                  f+i_coord_offset,fshift+i_shift_offset);
1059
1060         /* Increment number of inner iterations */
1061         inneriter                  += j_index_end - j_index_start;
1062
1063         /* Outer loop uses 18 flops */
1064     }
1065
1066     /* Increment number of outer iterations */
1067     outeriter        += nri;
1068
1069     /* Update outer/inner flops */
1070
1071     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_W3_F,outeriter*18 + inneriter*93);
1072 }