src/gromacs/gmxlib/nonbonded/nb_kernel_avx_256_single/nb_kernel_ElecRFCut_VdwCSTab_GeomW4P1_avx_256_single.c

   1 /*
   2  * This file is part of the GROMACS molecular simulation package.
   3  *
   4  * Copyright (c) 2012,2013, by the GROMACS development team, led by
   5  * Mark Abraham, David van der Spoel, Berk Hess, and Erik Lindahl,
   6  * and including many others, as listed in the AUTHORS file in the
   7  * top-level source directory and at http://www.gromacs.org.
   8  *
   9  * GROMACS is free software; you can redistribute it and/or
  10  * modify it under the terms of the GNU Lesser General Public License
  11  * as published by the Free Software Foundation; either version 2.1
  12  * of the License, or (at your option) any later version.
  13  *
  14  * GROMACS is distributed in the hope that it will be useful,
  15  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  16  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  17  * Lesser General Public License for more details.
  18  *
  19  * You should have received a copy of the GNU Lesser General Public
  20  * License along with GROMACS; if not, see
  21  * http://www.gnu.org/licenses, or write to the Free Software Foundation,
  22  * Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA.
  23  *
  24  * If you want to redistribute modifications to GROMACS, please
  25  * consider that scientific software is very special. Version
  26  * control is crucial - bugs must be traceable. We will be happy to
  27  * consider code for inclusion in the official distribution, but
  28  * derived work must not be called official GROMACS. Details are found
  29  * in the README & COPYING files - if they are missing, get the
  30  * official version at http://www.gromacs.org.
  31  *
  32  * To help us fund GROMACS development, we humbly ask that you cite
  33  * the research papers on the package. Check out http://www.gromacs.org.
  34  */
  35 /*
  36  * Note: this file was generated by the GROMACS avx_256_single kernel generator.
  37  */
  38 #ifdef HAVE_CONFIG_H
  39 #include <config.h>
  40 #endif
  41
  42 #include <math.h>
  43
  44 #include "../nb_kernel.h"
  45 #include "types/simple.h"
  46 #include "vec.h"
  47 #include "nrnb.h"
  48
  49 #include "gromacs/simd/math_x86_avx_256_single.h"
  50 #include "kernelutil_x86_avx_256_single.h"
  51
  52 /*
  53  * Gromacs nonbonded kernel:   nb_kernel_ElecRFCut_VdwCSTab_GeomW4P1_VF_avx_256_single
  54  * Electrostatics interaction: ReactionField
  55  * VdW interaction:            CubicSplineTable
  56  * Geometry:                   Water4-Particle
  57  * Calculate force/pot:        PotentialAndForce
  58  */
  59 void
  60 nb_kernel_ElecRFCut_VdwCSTab_GeomW4P1_VF_avx_256_single
  61                     (t_nblist                    * gmx_restrict       nlist,
  62                      rvec                        * gmx_restrict          xx,
  63                      rvec                        * gmx_restrict          ff,
  64                      t_forcerec                  * gmx_restrict          fr,
  65                      t_mdatoms                   * gmx_restrict     mdatoms,
  66                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
  67                      t_nrnb                      * gmx_restrict        nrnb)
  68 {
  69     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
  70      * just 0 for non-waters.
  71      * Suffixes A,B,C,D,E,F,G,H refer to j loop unrolling done with AVX, e.g. for the eight different
  72      * jnr indices corresponding to data put in the four positions in the SIMD register.
  73      */
  74     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
  75     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
  76     int              jnrA,jnrB,jnrC,jnrD;
  77     int              jnrE,jnrF,jnrG,jnrH;
  78     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
  79     int              jnrlistE,jnrlistF,jnrlistG,jnrlistH;
  80     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
  81     int              j_coord_offsetE,j_coord_offsetF,j_coord_offsetG,j_coord_offsetH;
  82     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  83     real             rcutoff_scalar;
  84     real             *shiftvec,*fshift,*x,*f;
  85     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD,*fjptrE,*fjptrF,*fjptrG,*fjptrH;
  86     real             scratch[4*DIM];
  87     __m256           tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
  88     real *           vdwioffsetptr0;
  89     __m256           ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  90     real *           vdwioffsetptr1;
  91     __m256           ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
  92     real *           vdwioffsetptr2;
  93     __m256           ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
  94     real *           vdwioffsetptr3;
  95     __m256           ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
  96     int              vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D,vdwjidx0E,vdwjidx0F,vdwjidx0G,vdwjidx0H;
  97     __m256           jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  98     __m256           dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
  99     __m256           dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
 100     __m256           dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
 101     __m256           dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30;
 102     __m256           velec,felec,velecsum,facel,crf,krf,krf2;
 103     real             *charge;
 104     int              nvdwtype;
 105     __m256           rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
 106     int              *vdwtype;
 107     real             *vdwparam;
 108     __m256           one_sixth   = _mm256_set1_ps(1.0/6.0);
 109     __m256           one_twelfth = _mm256_set1_ps(1.0/12.0);
 110     __m256i          vfitab;
 111     __m128i          vfitab_lo,vfitab_hi;
 112     __m128i          ifour       = _mm_set1_epi32(4);
 113     __m256           rt,vfeps,vftabscale,Y,F,G,H,Heps,Fp,VV,FF;
 114     real             *vftab;
 115     __m256           dummy_mask,cutoff_mask;
 116     __m256           signbit = _mm256_castsi256_ps( _mm256_set1_epi32(0x80000000) );
 117     __m256           one     = _mm256_set1_ps(1.0);
 118     __m256           two     = _mm256_set1_ps(2.0);
 119     x                = xx[0];
 120     f                = ff[0];
 121
 122     nri              = nlist->nri;
 123     iinr             = nlist->iinr;
 124     jindex           = nlist->jindex;
 125     jjnr             = nlist->jjnr;
 126     shiftidx         = nlist->shift;
 127     gid              = nlist->gid;
 128     shiftvec         = fr->shift_vec[0];
 129     fshift           = fr->fshift[0];
 130     facel            = _mm256_set1_ps(fr->epsfac);
 131     charge           = mdatoms->chargeA;
 132     krf              = _mm256_set1_ps(fr->ic->k_rf);
 133     krf2             = _mm256_set1_ps(fr->ic->k_rf*2.0);
 134     crf              = _mm256_set1_ps(fr->ic->c_rf);
 135     nvdwtype         = fr->ntype;
 136     vdwparam         = fr->nbfp;
 137     vdwtype          = mdatoms->typeA;
 138
 139     vftab            = kernel_data->table_vdw->data;
 140     vftabscale       = _mm256_set1_ps(kernel_data->table_vdw->scale);
 141
 142     /* Setup water-specific parameters */
 143     inr              = nlist->iinr[0];
 144     iq1              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+1]));
 145     iq2              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+2]));
 146     iq3              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+3]));
 147     vdwioffsetptr0   = vdwparam+2*nvdwtype*vdwtype[inr+0];
 148
 149     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 150     rcutoff_scalar   = fr->rcoulomb;
 151     rcutoff          = _mm256_set1_ps(rcutoff_scalar);
 152     rcutoff2         = _mm256_mul_ps(rcutoff,rcutoff);
 153
 154     /* Avoid stupid compiler warnings */
 155     jnrA = jnrB = jnrC = jnrD = jnrE = jnrF = jnrG = jnrH = 0;
 156     j_coord_offsetA = 0;
 157     j_coord_offsetB = 0;
 158     j_coord_offsetC = 0;
 159     j_coord_offsetD = 0;
 160     j_coord_offsetE = 0;
 161     j_coord_offsetF = 0;
 162     j_coord_offsetG = 0;
 163     j_coord_offsetH = 0;
 164
 165     outeriter        = 0;
 166     inneriter        = 0;
 167
 168     for(iidx=0;iidx<4*DIM;iidx++)
 169     {
 170         scratch[iidx] = 0.0;
 171     }
 172
 173     /* Start outer loop over neighborlists */
 174     for(iidx=0; iidx<nri; iidx++)
 175     {
 176         /* Load shift vector for this list */
 177         i_shift_offset   = DIM*shiftidx[iidx];
 178
 179         /* Load limits for loop over neighbors */
 180         j_index_start    = jindex[iidx];
 181         j_index_end      = jindex[iidx+1];
 182
 183         /* Get outer coordinate index */
 184         inr              = iinr[iidx];
 185         i_coord_offset   = DIM*inr;
 186
 187         /* Load i particle coords and add shift vector */
 188         gmx_mm256_load_shift_and_4rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,
 189                                                     &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
 190
 191         fix0             = _mm256_setzero_ps();
 192         fiy0             = _mm256_setzero_ps();
 193         fiz0             = _mm256_setzero_ps();
 194         fix1             = _mm256_setzero_ps();
 195         fiy1             = _mm256_setzero_ps();
 196         fiz1             = _mm256_setzero_ps();
 197         fix2             = _mm256_setzero_ps();
 198         fiy2             = _mm256_setzero_ps();
 199         fiz2             = _mm256_setzero_ps();
 200         fix3             = _mm256_setzero_ps();
 201         fiy3             = _mm256_setzero_ps();
 202         fiz3             = _mm256_setzero_ps();
 203
 204         /* Reset potential sums */
 205         velecsum         = _mm256_setzero_ps();
 206         vvdwsum          = _mm256_setzero_ps();
 207
 208         /* Start inner kernel loop */
 209         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+7]>=0; jidx+=8)
 210         {
 211
 212             /* Get j neighbor index, and coordinate index */
 213             jnrA             = jjnr[jidx];
 214             jnrB             = jjnr[jidx+1];
 215             jnrC             = jjnr[jidx+2];
 216             jnrD             = jjnr[jidx+3];
 217             jnrE             = jjnr[jidx+4];
 218             jnrF             = jjnr[jidx+5];
 219             jnrG             = jjnr[jidx+6];
 220             jnrH             = jjnr[jidx+7];
 221             j_coord_offsetA  = DIM*jnrA;
 222             j_coord_offsetB  = DIM*jnrB;
 223             j_coord_offsetC  = DIM*jnrC;
 224             j_coord_offsetD  = DIM*jnrD;
 225             j_coord_offsetE  = DIM*jnrE;
 226             j_coord_offsetF  = DIM*jnrF;
 227             j_coord_offsetG  = DIM*jnrG;
 228             j_coord_offsetH  = DIM*jnrH;
 229
 230             /* load j atom coordinates */
 231             gmx_mm256_load_1rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 232                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 233                                                  x+j_coord_offsetE,x+j_coord_offsetF,
 234                                                  x+j_coord_offsetG,x+j_coord_offsetH,
 235                                                  &jx0,&jy0,&jz0);
 236
 237             /* Calculate displacement vector */
 238             dx00             = _mm256_sub_ps(ix0,jx0);
 239             dy00             = _mm256_sub_ps(iy0,jy0);
 240             dz00             = _mm256_sub_ps(iz0,jz0);
 241             dx10             = _mm256_sub_ps(ix1,jx0);
 242             dy10             = _mm256_sub_ps(iy1,jy0);
 243             dz10             = _mm256_sub_ps(iz1,jz0);
 244             dx20             = _mm256_sub_ps(ix2,jx0);
 245             dy20             = _mm256_sub_ps(iy2,jy0);
 246             dz20             = _mm256_sub_ps(iz2,jz0);
 247             dx30             = _mm256_sub_ps(ix3,jx0);
 248             dy30             = _mm256_sub_ps(iy3,jy0);
 249             dz30             = _mm256_sub_ps(iz3,jz0);
 250
 251             /* Calculate squared distance and things based on it */
 252             rsq00            = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
 253             rsq10            = gmx_mm256_calc_rsq_ps(dx10,dy10,dz10);
 254             rsq20            = gmx_mm256_calc_rsq_ps(dx20,dy20,dz20);
 255             rsq30            = gmx_mm256_calc_rsq_ps(dx30,dy30,dz30);
 256
 257             rinv00           = gmx_mm256_invsqrt_ps(rsq00);
 258             rinv10           = gmx_mm256_invsqrt_ps(rsq10);
 259             rinv20           = gmx_mm256_invsqrt_ps(rsq20);
 260             rinv30           = gmx_mm256_invsqrt_ps(rsq30);
 261
 262             rinvsq10         = _mm256_mul_ps(rinv10,rinv10);
 263             rinvsq20         = _mm256_mul_ps(rinv20,rinv20);
 264             rinvsq30         = _mm256_mul_ps(rinv30,rinv30);
 265
 266             /* Load parameters for j particles */
 267             jq0              = gmx_mm256_load_8real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
 268                                                                  charge+jnrC+0,charge+jnrD+0,
 269                                                                  charge+jnrE+0,charge+jnrF+0,
 270                                                                  charge+jnrG+0,charge+jnrH+0);
 271             vdwjidx0A        = 2*vdwtype[jnrA+0];
 272             vdwjidx0B        = 2*vdwtype[jnrB+0];
 273             vdwjidx0C        = 2*vdwtype[jnrC+0];
 274             vdwjidx0D        = 2*vdwtype[jnrD+0];
 275             vdwjidx0E        = 2*vdwtype[jnrE+0];
 276             vdwjidx0F        = 2*vdwtype[jnrF+0];
 277             vdwjidx0G        = 2*vdwtype[jnrG+0];
 278             vdwjidx0H        = 2*vdwtype[jnrH+0];
 279
 280             fjx0             = _mm256_setzero_ps();
 281             fjy0             = _mm256_setzero_ps();
 282             fjz0             = _mm256_setzero_ps();
 283
 284             /**************************
 285              * CALCULATE INTERACTIONS *
 286              **************************/
 287
 288             r00              = _mm256_mul_ps(rsq00,rinv00);
 289
 290             /* Compute parameters for interactions between i and j atoms */
 291             gmx_mm256_load_8pair_swizzle_ps(vdwioffsetptr0+vdwjidx0A,
 292                                             vdwioffsetptr0+vdwjidx0B,
 293                                             vdwioffsetptr0+vdwjidx0C,
 294                                             vdwioffsetptr0+vdwjidx0D,
 295                                             vdwioffsetptr0+vdwjidx0E,
 296                                             vdwioffsetptr0+vdwjidx0F,
 297                                             vdwioffsetptr0+vdwjidx0G,
 298                                             vdwioffsetptr0+vdwjidx0H,
 299                                             &c6_00,&c12_00);
 300
 301             /* Calculate table index by multiplying r with table scale and truncate to integer */
 302             rt               = _mm256_mul_ps(r00,vftabscale);
 303             vfitab           = _mm256_cvttps_epi32(rt);
 304             vfeps            = _mm256_sub_ps(rt,_mm256_round_ps(rt, _MM_FROUND_FLOOR));
 305             /*         AVX1 does not support 256-bit integer operations, so now we go to 128-bit mode... */
 306             vfitab_lo        = _mm256_extractf128_si256(vfitab,0x0);
 307             vfitab_hi        = _mm256_extractf128_si256(vfitab,0x1);
 308             vfitab_lo        = _mm_slli_epi32(vfitab_lo,3);
 309             vfitab_hi        = _mm_slli_epi32(vfitab_hi,3);
 310
 311             /* CUBIC SPLINE TABLE DISPERSION */
 312             Y                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,0)),
 313                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,0)));
 314             F                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,1)),
 315                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,1)));
 316             G                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,2)),
 317                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,2)));
 318             H                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,3)),
 319                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,3)));
 320             GMX_MM256_HALFTRANSPOSE4_PS(Y,F,G,H);
 321             Heps             = _mm256_mul_ps(vfeps,H);
 322             Fp               = _mm256_add_ps(F,_mm256_mul_ps(vfeps,_mm256_add_ps(G,Heps)));
 323             VV               = _mm256_add_ps(Y,_mm256_mul_ps(vfeps,Fp));
 324             vvdw6            = _mm256_mul_ps(c6_00,VV);
 325             FF               = _mm256_add_ps(Fp,_mm256_mul_ps(vfeps,_mm256_add_ps(G,_mm256_add_ps(Heps,Heps))));
 326             fvdw6            = _mm256_mul_ps(c6_00,FF);
 327
 328             /* CUBIC SPLINE TABLE REPULSION */
 329             vfitab_lo        = _mm_add_epi32(vfitab_lo,ifour);
 330             vfitab_hi        = _mm_add_epi32(vfitab_hi,ifour);
 331             Y                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,0)),
 332                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,0)));
 333             F                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,1)),
 334                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,1)));
 335             G                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,2)),
 336                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,2)));
 337             H                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,3)),
 338                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,3)));
 339             GMX_MM256_HALFTRANSPOSE4_PS(Y,F,G,H);
 340             Heps             = _mm256_mul_ps(vfeps,H);
 341             Fp               = _mm256_add_ps(F,_mm256_mul_ps(vfeps,_mm256_add_ps(G,Heps)));
 342             VV               = _mm256_add_ps(Y,_mm256_mul_ps(vfeps,Fp));
 343             vvdw12           = _mm256_mul_ps(c12_00,VV);
 344             FF               = _mm256_add_ps(Fp,_mm256_mul_ps(vfeps,_mm256_add_ps(G,_mm256_add_ps(Heps,Heps))));
 345             fvdw12           = _mm256_mul_ps(c12_00,FF);
 346             vvdw             = _mm256_add_ps(vvdw12,vvdw6);
 347             fvdw             = _mm256_xor_ps(signbit,_mm256_mul_ps(_mm256_add_ps(fvdw6,fvdw12),_mm256_mul_ps(vftabscale,rinv00)));
 348
 349             /* Update potential sum for this i atom from the interaction with this j atom. */
 350             vvdwsum          = _mm256_add_ps(vvdwsum,vvdw);
 351
 352             fscal            = fvdw;
 353
 354             /* Calculate temporary vectorial force */
 355             tx               = _mm256_mul_ps(fscal,dx00);
 356             ty               = _mm256_mul_ps(fscal,dy00);
 357             tz               = _mm256_mul_ps(fscal,dz00);
 358
 359             /* Update vectorial force */
 360             fix0             = _mm256_add_ps(fix0,tx);
 361             fiy0             = _mm256_add_ps(fiy0,ty);
 362             fiz0             = _mm256_add_ps(fiz0,tz);
 363
 364             fjx0             = _mm256_add_ps(fjx0,tx);
 365             fjy0             = _mm256_add_ps(fjy0,ty);
 366             fjz0             = _mm256_add_ps(fjz0,tz);
 367
 368             /**************************
 369              * CALCULATE INTERACTIONS *
 370              **************************/
 371
 372             if (gmx_mm256_any_lt(rsq10,rcutoff2))
 373             {
 374
 375             /* Compute parameters for interactions between i and j atoms */
 376             qq10             = _mm256_mul_ps(iq1,jq0);
 377
 378             /* REACTION-FIELD ELECTROSTATICS */
 379             velec            = _mm256_mul_ps(qq10,_mm256_sub_ps(_mm256_add_ps(rinv10,_mm256_mul_ps(krf,rsq10)),crf));
 380             felec            = _mm256_mul_ps(qq10,_mm256_sub_ps(_mm256_mul_ps(rinv10,rinvsq10),krf2));
 381
 382             cutoff_mask      = _mm256_cmp_ps(rsq10,rcutoff2,_CMP_LT_OQ);
 383
 384             /* Update potential sum for this i atom from the interaction with this j atom. */
 385             velec            = _mm256_and_ps(velec,cutoff_mask);
 386             velecsum         = _mm256_add_ps(velecsum,velec);
 387
 388             fscal            = felec;
 389
 390             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 391
 392             /* Calculate temporary vectorial force */
 393             tx               = _mm256_mul_ps(fscal,dx10);
 394             ty               = _mm256_mul_ps(fscal,dy10);
 395             tz               = _mm256_mul_ps(fscal,dz10);
 396
 397             /* Update vectorial force */
 398             fix1             = _mm256_add_ps(fix1,tx);
 399             fiy1             = _mm256_add_ps(fiy1,ty);
 400             fiz1             = _mm256_add_ps(fiz1,tz);
 401
 402             fjx0             = _mm256_add_ps(fjx0,tx);
 403             fjy0             = _mm256_add_ps(fjy0,ty);
 404             fjz0             = _mm256_add_ps(fjz0,tz);
 405
 406             }
 407
 408             /**************************
 409              * CALCULATE INTERACTIONS *
 410              **************************/
 411
 412             if (gmx_mm256_any_lt(rsq20,rcutoff2))
 413             {
 414
 415             /* Compute parameters for interactions between i and j atoms */
 416             qq20             = _mm256_mul_ps(iq2,jq0);
 417
 418             /* REACTION-FIELD ELECTROSTATICS */
 419             velec            = _mm256_mul_ps(qq20,_mm256_sub_ps(_mm256_add_ps(rinv20,_mm256_mul_ps(krf,rsq20)),crf));
 420             felec            = _mm256_mul_ps(qq20,_mm256_sub_ps(_mm256_mul_ps(rinv20,rinvsq20),krf2));
 421
 422             cutoff_mask      = _mm256_cmp_ps(rsq20,rcutoff2,_CMP_LT_OQ);
 423
 424             /* Update potential sum for this i atom from the interaction with this j atom. */
 425             velec            = _mm256_and_ps(velec,cutoff_mask);
 426             velecsum         = _mm256_add_ps(velecsum,velec);
 427
 428             fscal            = felec;
 429
 430             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 431
 432             /* Calculate temporary vectorial force */
 433             tx               = _mm256_mul_ps(fscal,dx20);
 434             ty               = _mm256_mul_ps(fscal,dy20);
 435             tz               = _mm256_mul_ps(fscal,dz20);
 436
 437             /* Update vectorial force */
 438             fix2             = _mm256_add_ps(fix2,tx);
 439             fiy2             = _mm256_add_ps(fiy2,ty);
 440             fiz2             = _mm256_add_ps(fiz2,tz);
 441
 442             fjx0             = _mm256_add_ps(fjx0,tx);
 443             fjy0             = _mm256_add_ps(fjy0,ty);
 444             fjz0             = _mm256_add_ps(fjz0,tz);
 445
 446             }
 447
 448             /**************************
 449              * CALCULATE INTERACTIONS *
 450              **************************/
 451
 452             if (gmx_mm256_any_lt(rsq30,rcutoff2))
 453             {
 454
 455             /* Compute parameters for interactions between i and j atoms */
 456             qq30             = _mm256_mul_ps(iq3,jq0);
 457
 458             /* REACTION-FIELD ELECTROSTATICS */
 459             velec            = _mm256_mul_ps(qq30,_mm256_sub_ps(_mm256_add_ps(rinv30,_mm256_mul_ps(krf,rsq30)),crf));
 460             felec            = _mm256_mul_ps(qq30,_mm256_sub_ps(_mm256_mul_ps(rinv30,rinvsq30),krf2));
 461
 462             cutoff_mask      = _mm256_cmp_ps(rsq30,rcutoff2,_CMP_LT_OQ);
 463
 464             /* Update potential sum for this i atom from the interaction with this j atom. */
 465             velec            = _mm256_and_ps(velec,cutoff_mask);
 466             velecsum         = _mm256_add_ps(velecsum,velec);
 467
 468             fscal            = felec;
 469
 470             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 471
 472             /* Calculate temporary vectorial force */
 473             tx               = _mm256_mul_ps(fscal,dx30);
 474             ty               = _mm256_mul_ps(fscal,dy30);
 475             tz               = _mm256_mul_ps(fscal,dz30);
 476
 477             /* Update vectorial force */
 478             fix3             = _mm256_add_ps(fix3,tx);
 479             fiy3             = _mm256_add_ps(fiy3,ty);
 480             fiz3             = _mm256_add_ps(fiz3,tz);
 481
 482             fjx0             = _mm256_add_ps(fjx0,tx);
 483             fjy0             = _mm256_add_ps(fjy0,ty);
 484             fjz0             = _mm256_add_ps(fjz0,tz);
 485
 486             }
 487
 488             fjptrA             = f+j_coord_offsetA;
 489             fjptrB             = f+j_coord_offsetB;
 490             fjptrC             = f+j_coord_offsetC;
 491             fjptrD             = f+j_coord_offsetD;
 492             fjptrE             = f+j_coord_offsetE;
 493             fjptrF             = f+j_coord_offsetF;
 494             fjptrG             = f+j_coord_offsetG;
 495             fjptrH             = f+j_coord_offsetH;
 496
 497             gmx_mm256_decrement_1rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,fjx0,fjy0,fjz0);
 498
 499             /* Inner loop uses 167 flops */
 500         }
 501
 502         if(jidx<j_index_end)
 503         {
 504
 505             /* Get j neighbor index, and coordinate index */
 506             jnrlistA         = jjnr[jidx];
 507             jnrlistB         = jjnr[jidx+1];
 508             jnrlistC         = jjnr[jidx+2];
 509             jnrlistD         = jjnr[jidx+3];
 510             jnrlistE         = jjnr[jidx+4];
 511             jnrlistF         = jjnr[jidx+5];
 512             jnrlistG         = jjnr[jidx+6];
 513             jnrlistH         = jjnr[jidx+7];
 514             /* Sign of each element will be negative for non-real atoms.
 515              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
 516              * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
 517              */
 518             dummy_mask = gmx_mm256_set_m128(gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx+4)),_mm_setzero_si128())),
 519                                             gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128())));
 520
 521             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
 522             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
 523             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
 524             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
 525             jnrE       = (jnrlistE>=0) ? jnrlistE : 0;
 526             jnrF       = (jnrlistF>=0) ? jnrlistF : 0;
 527             jnrG       = (jnrlistG>=0) ? jnrlistG : 0;
 528             jnrH       = (jnrlistH>=0) ? jnrlistH : 0;
 529             j_coord_offsetA  = DIM*jnrA;
 530             j_coord_offsetB  = DIM*jnrB;
 531             j_coord_offsetC  = DIM*jnrC;
 532             j_coord_offsetD  = DIM*jnrD;
 533             j_coord_offsetE  = DIM*jnrE;
 534             j_coord_offsetF  = DIM*jnrF;
 535             j_coord_offsetG  = DIM*jnrG;
 536             j_coord_offsetH  = DIM*jnrH;
 537
 538             /* load j atom coordinates */
 539             gmx_mm256_load_1rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 540                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 541                                                  x+j_coord_offsetE,x+j_coord_offsetF,
 542                                                  x+j_coord_offsetG,x+j_coord_offsetH,
 543                                                  &jx0,&jy0,&jz0);
 544
 545             /* Calculate displacement vector */
 546             dx00             = _mm256_sub_ps(ix0,jx0);
 547             dy00             = _mm256_sub_ps(iy0,jy0);
 548             dz00             = _mm256_sub_ps(iz0,jz0);
 549             dx10             = _mm256_sub_ps(ix1,jx0);
 550             dy10             = _mm256_sub_ps(iy1,jy0);
 551             dz10             = _mm256_sub_ps(iz1,jz0);
 552             dx20             = _mm256_sub_ps(ix2,jx0);
 553             dy20             = _mm256_sub_ps(iy2,jy0);
 554             dz20             = _mm256_sub_ps(iz2,jz0);
 555             dx30             = _mm256_sub_ps(ix3,jx0);
 556             dy30             = _mm256_sub_ps(iy3,jy0);
 557             dz30             = _mm256_sub_ps(iz3,jz0);
 558
 559             /* Calculate squared distance and things based on it */
 560             rsq00            = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
 561             rsq10            = gmx_mm256_calc_rsq_ps(dx10,dy10,dz10);
 562             rsq20            = gmx_mm256_calc_rsq_ps(dx20,dy20,dz20);
 563             rsq30            = gmx_mm256_calc_rsq_ps(dx30,dy30,dz30);
 564
 565             rinv00           = gmx_mm256_invsqrt_ps(rsq00);
 566             rinv10           = gmx_mm256_invsqrt_ps(rsq10);
 567             rinv20           = gmx_mm256_invsqrt_ps(rsq20);
 568             rinv30           = gmx_mm256_invsqrt_ps(rsq30);
 569
 570             rinvsq10         = _mm256_mul_ps(rinv10,rinv10);
 571             rinvsq20         = _mm256_mul_ps(rinv20,rinv20);
 572             rinvsq30         = _mm256_mul_ps(rinv30,rinv30);
 573
 574             /* Load parameters for j particles */
 575             jq0              = gmx_mm256_load_8real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
 576                                                                  charge+jnrC+0,charge+jnrD+0,
 577                                                                  charge+jnrE+0,charge+jnrF+0,
 578                                                                  charge+jnrG+0,charge+jnrH+0);
 579             vdwjidx0A        = 2*vdwtype[jnrA+0];
 580             vdwjidx0B        = 2*vdwtype[jnrB+0];
 581             vdwjidx0C        = 2*vdwtype[jnrC+0];
 582             vdwjidx0D        = 2*vdwtype[jnrD+0];
 583             vdwjidx0E        = 2*vdwtype[jnrE+0];
 584             vdwjidx0F        = 2*vdwtype[jnrF+0];
 585             vdwjidx0G        = 2*vdwtype[jnrG+0];
 586             vdwjidx0H        = 2*vdwtype[jnrH+0];
 587
 588             fjx0             = _mm256_setzero_ps();
 589             fjy0             = _mm256_setzero_ps();
 590             fjz0             = _mm256_setzero_ps();
 591
 592             /**************************
 593              * CALCULATE INTERACTIONS *
 594              **************************/
 595
 596             r00              = _mm256_mul_ps(rsq00,rinv00);
 597             r00              = _mm256_andnot_ps(dummy_mask,r00);
 598
 599             /* Compute parameters for interactions between i and j atoms */
 600             gmx_mm256_load_8pair_swizzle_ps(vdwioffsetptr0+vdwjidx0A,
 601                                             vdwioffsetptr0+vdwjidx0B,
 602                                             vdwioffsetptr0+vdwjidx0C,
 603                                             vdwioffsetptr0+vdwjidx0D,
 604                                             vdwioffsetptr0+vdwjidx0E,
 605                                             vdwioffsetptr0+vdwjidx0F,
 606                                             vdwioffsetptr0+vdwjidx0G,
 607                                             vdwioffsetptr0+vdwjidx0H,
 608                                             &c6_00,&c12_00);
 609
 610             /* Calculate table index by multiplying r with table scale and truncate to integer */
 611             rt               = _mm256_mul_ps(r00,vftabscale);
 612             vfitab           = _mm256_cvttps_epi32(rt);
 613             vfeps            = _mm256_sub_ps(rt,_mm256_round_ps(rt, _MM_FROUND_FLOOR));
 614             /*         AVX1 does not support 256-bit integer operations, so now we go to 128-bit mode... */
 615             vfitab_lo        = _mm256_extractf128_si256(vfitab,0x0);
 616             vfitab_hi        = _mm256_extractf128_si256(vfitab,0x1);
 617             vfitab_lo        = _mm_slli_epi32(vfitab_lo,3);
 618             vfitab_hi        = _mm_slli_epi32(vfitab_hi,3);
 619
 620             /* CUBIC SPLINE TABLE DISPERSION */
 621             Y                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,0)),
 622                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,0)));
 623             F                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,1)),
 624                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,1)));
 625             G                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,2)),
 626                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,2)));
 627             H                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,3)),
 628                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,3)));
 629             GMX_MM256_HALFTRANSPOSE4_PS(Y,F,G,H);
 630             Heps             = _mm256_mul_ps(vfeps,H);
 631             Fp               = _mm256_add_ps(F,_mm256_mul_ps(vfeps,_mm256_add_ps(G,Heps)));
 632             VV               = _mm256_add_ps(Y,_mm256_mul_ps(vfeps,Fp));
 633             vvdw6            = _mm256_mul_ps(c6_00,VV);
 634             FF               = _mm256_add_ps(Fp,_mm256_mul_ps(vfeps,_mm256_add_ps(G,_mm256_add_ps(Heps,Heps))));
 635             fvdw6            = _mm256_mul_ps(c6_00,FF);
 636
 637             /* CUBIC SPLINE TABLE REPULSION */
 638             vfitab_lo        = _mm_add_epi32(vfitab_lo,ifour);
 639             vfitab_hi        = _mm_add_epi32(vfitab_hi,ifour);
 640             Y                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,0)),
 641                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,0)));
 642             F                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,1)),
 643                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,1)));
 644             G                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,2)),
 645                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,2)));
 646             H                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,3)),
 647                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,3)));
 648             GMX_MM256_HALFTRANSPOSE4_PS(Y,F,G,H);
 649             Heps             = _mm256_mul_ps(vfeps,H);
 650             Fp               = _mm256_add_ps(F,_mm256_mul_ps(vfeps,_mm256_add_ps(G,Heps)));
 651             VV               = _mm256_add_ps(Y,_mm256_mul_ps(vfeps,Fp));
 652             vvdw12           = _mm256_mul_ps(c12_00,VV);
 653             FF               = _mm256_add_ps(Fp,_mm256_mul_ps(vfeps,_mm256_add_ps(G,_mm256_add_ps(Heps,Heps))));
 654             fvdw12           = _mm256_mul_ps(c12_00,FF);
 655             vvdw             = _mm256_add_ps(vvdw12,vvdw6);
 656             fvdw             = _mm256_xor_ps(signbit,_mm256_mul_ps(_mm256_add_ps(fvdw6,fvdw12),_mm256_mul_ps(vftabscale,rinv00)));
 657
 658             /* Update potential sum for this i atom from the interaction with this j atom. */
 659             vvdw             = _mm256_andnot_ps(dummy_mask,vvdw);
 660             vvdwsum          = _mm256_add_ps(vvdwsum,vvdw);
 661
 662             fscal            = fvdw;
 663
 664             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
 665
 666             /* Calculate temporary vectorial force */
 667             tx               = _mm256_mul_ps(fscal,dx00);
 668             ty               = _mm256_mul_ps(fscal,dy00);
 669             tz               = _mm256_mul_ps(fscal,dz00);
 670
 671             /* Update vectorial force */
 672             fix0             = _mm256_add_ps(fix0,tx);
 673             fiy0             = _mm256_add_ps(fiy0,ty);
 674             fiz0             = _mm256_add_ps(fiz0,tz);
 675
 676             fjx0             = _mm256_add_ps(fjx0,tx);
 677             fjy0             = _mm256_add_ps(fjy0,ty);
 678             fjz0             = _mm256_add_ps(fjz0,tz);
 679
 680             /**************************
 681              * CALCULATE INTERACTIONS *
 682              **************************/
 683
 684             if (gmx_mm256_any_lt(rsq10,rcutoff2))
 685             {
 686
 687             /* Compute parameters for interactions between i and j atoms */
 688             qq10             = _mm256_mul_ps(iq1,jq0);
 689
 690             /* REACTION-FIELD ELECTROSTATICS */
 691             velec            = _mm256_mul_ps(qq10,_mm256_sub_ps(_mm256_add_ps(rinv10,_mm256_mul_ps(krf,rsq10)),crf));
 692             felec            = _mm256_mul_ps(qq10,_mm256_sub_ps(_mm256_mul_ps(rinv10,rinvsq10),krf2));
 693
 694             cutoff_mask      = _mm256_cmp_ps(rsq10,rcutoff2,_CMP_LT_OQ);
 695
 696             /* Update potential sum for this i atom from the interaction with this j atom. */
 697             velec            = _mm256_and_ps(velec,cutoff_mask);
 698             velec            = _mm256_andnot_ps(dummy_mask,velec);
 699             velecsum         = _mm256_add_ps(velecsum,velec);
 700
 701             fscal            = felec;
 702
 703             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 704
 705             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
 706
 707             /* Calculate temporary vectorial force */
 708             tx               = _mm256_mul_ps(fscal,dx10);
 709             ty               = _mm256_mul_ps(fscal,dy10);
 710             tz               = _mm256_mul_ps(fscal,dz10);
 711
 712             /* Update vectorial force */
 713             fix1             = _mm256_add_ps(fix1,tx);
 714             fiy1             = _mm256_add_ps(fiy1,ty);
 715             fiz1             = _mm256_add_ps(fiz1,tz);
 716
 717             fjx0             = _mm256_add_ps(fjx0,tx);
 718             fjy0             = _mm256_add_ps(fjy0,ty);
 719             fjz0             = _mm256_add_ps(fjz0,tz);
 720
 721             }
 722
 723             /**************************
 724              * CALCULATE INTERACTIONS *
 725              **************************/
 726
 727             if (gmx_mm256_any_lt(rsq20,rcutoff2))
 728             {
 729
 730             /* Compute parameters for interactions between i and j atoms */
 731             qq20             = _mm256_mul_ps(iq2,jq0);
 732
 733             /* REACTION-FIELD ELECTROSTATICS */
 734             velec            = _mm256_mul_ps(qq20,_mm256_sub_ps(_mm256_add_ps(rinv20,_mm256_mul_ps(krf,rsq20)),crf));
 735             felec            = _mm256_mul_ps(qq20,_mm256_sub_ps(_mm256_mul_ps(rinv20,rinvsq20),krf2));
 736
 737             cutoff_mask      = _mm256_cmp_ps(rsq20,rcutoff2,_CMP_LT_OQ);
 738
 739             /* Update potential sum for this i atom from the interaction with this j atom. */
 740             velec            = _mm256_and_ps(velec,cutoff_mask);
 741             velec            = _mm256_andnot_ps(dummy_mask,velec);
 742             velecsum         = _mm256_add_ps(velecsum,velec);
 743
 744             fscal            = felec;
 745
 746             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 747
 748             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
 749
 750             /* Calculate temporary vectorial force */
 751             tx               = _mm256_mul_ps(fscal,dx20);
 752             ty               = _mm256_mul_ps(fscal,dy20);
 753             tz               = _mm256_mul_ps(fscal,dz20);
 754
 755             /* Update vectorial force */
 756             fix2             = _mm256_add_ps(fix2,tx);
 757             fiy2             = _mm256_add_ps(fiy2,ty);
 758             fiz2             = _mm256_add_ps(fiz2,tz);
 759
 760             fjx0             = _mm256_add_ps(fjx0,tx);
 761             fjy0             = _mm256_add_ps(fjy0,ty);
 762             fjz0             = _mm256_add_ps(fjz0,tz);
 763
 764             }
 765
 766             /**************************
 767              * CALCULATE INTERACTIONS *
 768              **************************/
 769
 770             if (gmx_mm256_any_lt(rsq30,rcutoff2))
 771             {
 772
 773             /* Compute parameters for interactions between i and j atoms */
 774             qq30             = _mm256_mul_ps(iq3,jq0);
 775
 776             /* REACTION-FIELD ELECTROSTATICS */
 777             velec            = _mm256_mul_ps(qq30,_mm256_sub_ps(_mm256_add_ps(rinv30,_mm256_mul_ps(krf,rsq30)),crf));
 778             felec            = _mm256_mul_ps(qq30,_mm256_sub_ps(_mm256_mul_ps(rinv30,rinvsq30),krf2));
 779
 780             cutoff_mask      = _mm256_cmp_ps(rsq30,rcutoff2,_CMP_LT_OQ);
 781
 782             /* Update potential sum for this i atom from the interaction with this j atom. */
 783             velec            = _mm256_and_ps(velec,cutoff_mask);
 784             velec            = _mm256_andnot_ps(dummy_mask,velec);
 785             velecsum         = _mm256_add_ps(velecsum,velec);
 786
 787             fscal            = felec;
 788
 789             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 790
 791             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
 792
 793             /* Calculate temporary vectorial force */
 794             tx               = _mm256_mul_ps(fscal,dx30);
 795             ty               = _mm256_mul_ps(fscal,dy30);
 796             tz               = _mm256_mul_ps(fscal,dz30);
 797
 798             /* Update vectorial force */
 799             fix3             = _mm256_add_ps(fix3,tx);
 800             fiy3             = _mm256_add_ps(fiy3,ty);
 801             fiz3             = _mm256_add_ps(fiz3,tz);
 802
 803             fjx0             = _mm256_add_ps(fjx0,tx);
 804             fjy0             = _mm256_add_ps(fjy0,ty);
 805             fjz0             = _mm256_add_ps(fjz0,tz);
 806
 807             }
 808
 809             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
 810             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
 811             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
 812             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
 813             fjptrE             = (jnrlistE>=0) ? f+j_coord_offsetE : scratch;
 814             fjptrF             = (jnrlistF>=0) ? f+j_coord_offsetF : scratch;
 815             fjptrG             = (jnrlistG>=0) ? f+j_coord_offsetG : scratch;
 816             fjptrH             = (jnrlistH>=0) ? f+j_coord_offsetH : scratch;
 817
 818             gmx_mm256_decrement_1rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,fjx0,fjy0,fjz0);
 819
 820             /* Inner loop uses 168 flops */
 821         }
 822
 823         /* End of innermost loop */
 824
 825         gmx_mm256_update_iforce_4atom_swizzle_ps(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
 826                                                  f+i_coord_offset,fshift+i_shift_offset);
 827
 828         ggid                        = gid[iidx];
 829         /* Update potential energies */
 830         gmx_mm256_update_1pot_ps(velecsum,kernel_data->energygrp_elec+ggid);
 831         gmx_mm256_update_1pot_ps(vvdwsum,kernel_data->energygrp_vdw+ggid);
 832
 833         /* Increment number of inner iterations */
 834         inneriter                  += j_index_end - j_index_start;
 835
 836         /* Outer loop uses 26 flops */
 837     }
 838
 839     /* Increment number of outer iterations */
 840     outeriter        += nri;
 841
 842     /* Update outer/inner flops */
 843
 844     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_VF,outeriter*26 + inneriter*168);
 845 }
 846 /*
 847  * Gromacs nonbonded kernel:   nb_kernel_ElecRFCut_VdwCSTab_GeomW4P1_F_avx_256_single
 848  * Electrostatics interaction: ReactionField
 849  * VdW interaction:            CubicSplineTable
 850  * Geometry:                   Water4-Particle
 851  * Calculate force/pot:        Force
 852  */
 853 void
 854 nb_kernel_ElecRFCut_VdwCSTab_GeomW4P1_F_avx_256_single
 855                     (t_nblist                    * gmx_restrict       nlist,
 856                      rvec                        * gmx_restrict          xx,
 857                      rvec                        * gmx_restrict          ff,
 858                      t_forcerec                  * gmx_restrict          fr,
 859                      t_mdatoms                   * gmx_restrict     mdatoms,
 860                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 861                      t_nrnb                      * gmx_restrict        nrnb)
 862 {
 863     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
 864      * just 0 for non-waters.
 865      * Suffixes A,B,C,D,E,F,G,H refer to j loop unrolling done with AVX, e.g. for the eight different
 866      * jnr indices corresponding to data put in the four positions in the SIMD register.
 867      */
 868     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
 869     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
 870     int              jnrA,jnrB,jnrC,jnrD;
 871     int              jnrE,jnrF,jnrG,jnrH;
 872     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
 873     int              jnrlistE,jnrlistF,jnrlistG,jnrlistH;
 874     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
 875     int              j_coord_offsetE,j_coord_offsetF,j_coord_offsetG,j_coord_offsetH;
 876     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 877     real             rcutoff_scalar;
 878     real             *shiftvec,*fshift,*x,*f;
 879     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD,*fjptrE,*fjptrF,*fjptrG,*fjptrH;
 880     real             scratch[4*DIM];
 881     __m256           tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
 882     real *           vdwioffsetptr0;
 883     __m256           ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 884     real *           vdwioffsetptr1;
 885     __m256           ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
 886     real *           vdwioffsetptr2;
 887     __m256           ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
 888     real *           vdwioffsetptr3;
 889     __m256           ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
 890     int              vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D,vdwjidx0E,vdwjidx0F,vdwjidx0G,vdwjidx0H;
 891     __m256           jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 892     __m256           dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
 893     __m256           dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
 894     __m256           dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
 895     __m256           dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30;
 896     __m256           velec,felec,velecsum,facel,crf,krf,krf2;
 897     real             *charge;
 898     int              nvdwtype;
 899     __m256           rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
 900     int              *vdwtype;
 901     real             *vdwparam;
 902     __m256           one_sixth   = _mm256_set1_ps(1.0/6.0);
 903     __m256           one_twelfth = _mm256_set1_ps(1.0/12.0);
 904     __m256i          vfitab;
 905     __m128i          vfitab_lo,vfitab_hi;
 906     __m128i          ifour       = _mm_set1_epi32(4);
 907     __m256           rt,vfeps,vftabscale,Y,F,G,H,Heps,Fp,VV,FF;
 908     real             *vftab;
 909     __m256           dummy_mask,cutoff_mask;
 910     __m256           signbit = _mm256_castsi256_ps( _mm256_set1_epi32(0x80000000) );
 911     __m256           one     = _mm256_set1_ps(1.0);
 912     __m256           two     = _mm256_set1_ps(2.0);
 913     x                = xx[0];
 914     f                = ff[0];
 915
 916     nri              = nlist->nri;
 917     iinr             = nlist->iinr;
 918     jindex           = nlist->jindex;
 919     jjnr             = nlist->jjnr;
 920     shiftidx         = nlist->shift;
 921     gid              = nlist->gid;
 922     shiftvec         = fr->shift_vec[0];
 923     fshift           = fr->fshift[0];
 924     facel            = _mm256_set1_ps(fr->epsfac);
 925     charge           = mdatoms->chargeA;
 926     krf              = _mm256_set1_ps(fr->ic->k_rf);
 927     krf2             = _mm256_set1_ps(fr->ic->k_rf*2.0);
 928     crf              = _mm256_set1_ps(fr->ic->c_rf);
 929     nvdwtype         = fr->ntype;
 930     vdwparam         = fr->nbfp;
 931     vdwtype          = mdatoms->typeA;
 932
 933     vftab            = kernel_data->table_vdw->data;
 934     vftabscale       = _mm256_set1_ps(kernel_data->table_vdw->scale);
 935
 936     /* Setup water-specific parameters */
 937     inr              = nlist->iinr[0];
 938     iq1              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+1]));
 939     iq2              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+2]));
 940     iq3              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+3]));
 941     vdwioffsetptr0   = vdwparam+2*nvdwtype*vdwtype[inr+0];
 942
 943     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 944     rcutoff_scalar   = fr->rcoulomb;
 945     rcutoff          = _mm256_set1_ps(rcutoff_scalar);
 946     rcutoff2         = _mm256_mul_ps(rcutoff,rcutoff);
 947
 948     /* Avoid stupid compiler warnings */
 949     jnrA = jnrB = jnrC = jnrD = jnrE = jnrF = jnrG = jnrH = 0;
 950     j_coord_offsetA = 0;
 951     j_coord_offsetB = 0;
 952     j_coord_offsetC = 0;
 953     j_coord_offsetD = 0;
 954     j_coord_offsetE = 0;
 955     j_coord_offsetF = 0;
 956     j_coord_offsetG = 0;
 957     j_coord_offsetH = 0;
 958
 959     outeriter        = 0;
 960     inneriter        = 0;
 961
 962     for(iidx=0;iidx<4*DIM;iidx++)
 963     {
 964         scratch[iidx] = 0.0;
 965     }
 966
 967     /* Start outer loop over neighborlists */
 968     for(iidx=0; iidx<nri; iidx++)
 969     {
 970         /* Load shift vector for this list */
 971         i_shift_offset   = DIM*shiftidx[iidx];
 972
 973         /* Load limits for loop over neighbors */
 974         j_index_start    = jindex[iidx];
 975         j_index_end      = jindex[iidx+1];
 976
 977         /* Get outer coordinate index */
 978         inr              = iinr[iidx];
 979         i_coord_offset   = DIM*inr;
 980
 981         /* Load i particle coords and add shift vector */
 982         gmx_mm256_load_shift_and_4rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,
 983                                                     &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
 984
 985         fix0             = _mm256_setzero_ps();
 986         fiy0             = _mm256_setzero_ps();
 987         fiz0             = _mm256_setzero_ps();
 988         fix1             = _mm256_setzero_ps();
 989         fiy1             = _mm256_setzero_ps();
 990         fiz1             = _mm256_setzero_ps();
 991         fix2             = _mm256_setzero_ps();
 992         fiy2             = _mm256_setzero_ps();
 993         fiz2             = _mm256_setzero_ps();
 994         fix3             = _mm256_setzero_ps();
 995         fiy3             = _mm256_setzero_ps();
 996         fiz3             = _mm256_setzero_ps();
 997
 998         /* Start inner kernel loop */
 999         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+7]>=0; jidx+=8)
1000         {
1001
1002             /* Get j neighbor index, and coordinate index */
1003             jnrA             = jjnr[jidx];
1004             jnrB             = jjnr[jidx+1];
1005             jnrC             = jjnr[jidx+2];
1006             jnrD             = jjnr[jidx+3];
1007             jnrE             = jjnr[jidx+4];
1008             jnrF             = jjnr[jidx+5];
1009             jnrG             = jjnr[jidx+6];
1010             jnrH             = jjnr[jidx+7];
1011             j_coord_offsetA  = DIM*jnrA;
1012             j_coord_offsetB  = DIM*jnrB;
1013             j_coord_offsetC  = DIM*jnrC;
1014             j_coord_offsetD  = DIM*jnrD;
1015             j_coord_offsetE  = DIM*jnrE;
1016             j_coord_offsetF  = DIM*jnrF;
1017             j_coord_offsetG  = DIM*jnrG;
1018             j_coord_offsetH  = DIM*jnrH;
1019
1020             /* load j atom coordinates */
1021             gmx_mm256_load_1rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
1022                                                  x+j_coord_offsetC,x+j_coord_offsetD,
1023                                                  x+j_coord_offsetE,x+j_coord_offsetF,
1024                                                  x+j_coord_offsetG,x+j_coord_offsetH,
1025                                                  &jx0,&jy0,&jz0);
1026
1027             /* Calculate displacement vector */
1028             dx00             = _mm256_sub_ps(ix0,jx0);
1029             dy00             = _mm256_sub_ps(iy0,jy0);
1030             dz00             = _mm256_sub_ps(iz0,jz0);
1031             dx10             = _mm256_sub_ps(ix1,jx0);
1032             dy10             = _mm256_sub_ps(iy1,jy0);
1033             dz10             = _mm256_sub_ps(iz1,jz0);
1034             dx20             = _mm256_sub_ps(ix2,jx0);
1035             dy20             = _mm256_sub_ps(iy2,jy0);
1036             dz20             = _mm256_sub_ps(iz2,jz0);
1037             dx30             = _mm256_sub_ps(ix3,jx0);
1038             dy30             = _mm256_sub_ps(iy3,jy0);
1039             dz30             = _mm256_sub_ps(iz3,jz0);
1040
1041             /* Calculate squared distance and things based on it */
1042             rsq00            = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
1043             rsq10            = gmx_mm256_calc_rsq_ps(dx10,dy10,dz10);
1044             rsq20            = gmx_mm256_calc_rsq_ps(dx20,dy20,dz20);
1045             rsq30            = gmx_mm256_calc_rsq_ps(dx30,dy30,dz30);
1046
1047             rinv00           = gmx_mm256_invsqrt_ps(rsq00);
1048             rinv10           = gmx_mm256_invsqrt_ps(rsq10);
1049             rinv20           = gmx_mm256_invsqrt_ps(rsq20);
1050             rinv30           = gmx_mm256_invsqrt_ps(rsq30);
1051
1052             rinvsq10         = _mm256_mul_ps(rinv10,rinv10);
1053             rinvsq20         = _mm256_mul_ps(rinv20,rinv20);
1054             rinvsq30         = _mm256_mul_ps(rinv30,rinv30);
1055
1056             /* Load parameters for j particles */
1057             jq0              = gmx_mm256_load_8real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
1058                                                                  charge+jnrC+0,charge+jnrD+0,
1059                                                                  charge+jnrE+0,charge+jnrF+0,
1060                                                                  charge+jnrG+0,charge+jnrH+0);
1061             vdwjidx0A        = 2*vdwtype[jnrA+0];
1062             vdwjidx0B        = 2*vdwtype[jnrB+0];
1063             vdwjidx0C        = 2*vdwtype[jnrC+0];
1064             vdwjidx0D        = 2*vdwtype[jnrD+0];
1065             vdwjidx0E        = 2*vdwtype[jnrE+0];
1066             vdwjidx0F        = 2*vdwtype[jnrF+0];
1067             vdwjidx0G        = 2*vdwtype[jnrG+0];
1068             vdwjidx0H        = 2*vdwtype[jnrH+0];
1069
1070             fjx0             = _mm256_setzero_ps();
1071             fjy0             = _mm256_setzero_ps();
1072             fjz0             = _mm256_setzero_ps();
1073
1074             /**************************
1075              * CALCULATE INTERACTIONS *
1076              **************************/
1077
1078             r00              = _mm256_mul_ps(rsq00,rinv00);
1079
1080             /* Compute parameters for interactions between i and j atoms */
1081             gmx_mm256_load_8pair_swizzle_ps(vdwioffsetptr0+vdwjidx0A,
1082                                             vdwioffsetptr0+vdwjidx0B,
1083                                             vdwioffsetptr0+vdwjidx0C,
1084                                             vdwioffsetptr0+vdwjidx0D,
1085                                             vdwioffsetptr0+vdwjidx0E,
1086                                             vdwioffsetptr0+vdwjidx0F,
1087                                             vdwioffsetptr0+vdwjidx0G,
1088                                             vdwioffsetptr0+vdwjidx0H,
1089                                             &c6_00,&c12_00);
1090
1091             /* Calculate table index by multiplying r with table scale and truncate to integer */
1092             rt               = _mm256_mul_ps(r00,vftabscale);
1093             vfitab           = _mm256_cvttps_epi32(rt);
1094             vfeps            = _mm256_sub_ps(rt,_mm256_round_ps(rt, _MM_FROUND_FLOOR));
1095             /*         AVX1 does not support 256-bit integer operations, so now we go to 128-bit mode... */
1096             vfitab_lo        = _mm256_extractf128_si256(vfitab,0x0);
1097             vfitab_hi        = _mm256_extractf128_si256(vfitab,0x1);
1098             vfitab_lo        = _mm_slli_epi32(vfitab_lo,3);
1099             vfitab_hi        = _mm_slli_epi32(vfitab_hi,3);
1100
1101             /* CUBIC SPLINE TABLE DISPERSION */
1102             Y                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,0)),
1103                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,0)));
1104             F                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,1)),
1105                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,1)));
1106             G                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,2)),
1107                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,2)));
1108             H                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,3)),
1109                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,3)));
1110             GMX_MM256_HALFTRANSPOSE4_PS(Y,F,G,H);
1111             Heps             = _mm256_mul_ps(vfeps,H);
1112             Fp               = _mm256_add_ps(F,_mm256_mul_ps(vfeps,_mm256_add_ps(G,Heps)));
1113             FF               = _mm256_add_ps(Fp,_mm256_mul_ps(vfeps,_mm256_add_ps(G,_mm256_add_ps(Heps,Heps))));
1114             fvdw6            = _mm256_mul_ps(c6_00,FF);
1115
1116             /* CUBIC SPLINE TABLE REPULSION */
1117             vfitab_lo        = _mm_add_epi32(vfitab_lo,ifour);
1118             vfitab_hi        = _mm_add_epi32(vfitab_hi,ifour);
1119             Y                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,0)),
1120                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,0)));
1121             F                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,1)),
1122                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,1)));
1123             G                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,2)),
1124                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,2)));
1125             H                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,3)),
1126                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,3)));
1127             GMX_MM256_HALFTRANSPOSE4_PS(Y,F,G,H);
1128             Heps             = _mm256_mul_ps(vfeps,H);
1129             Fp               = _mm256_add_ps(F,_mm256_mul_ps(vfeps,_mm256_add_ps(G,Heps)));
1130             FF               = _mm256_add_ps(Fp,_mm256_mul_ps(vfeps,_mm256_add_ps(G,_mm256_add_ps(Heps,Heps))));
1131             fvdw12           = _mm256_mul_ps(c12_00,FF);
1132             fvdw             = _mm256_xor_ps(signbit,_mm256_mul_ps(_mm256_add_ps(fvdw6,fvdw12),_mm256_mul_ps(vftabscale,rinv00)));
1133
1134             fscal            = fvdw;
1135
1136             /* Calculate temporary vectorial force */
1137             tx               = _mm256_mul_ps(fscal,dx00);
1138             ty               = _mm256_mul_ps(fscal,dy00);
1139             tz               = _mm256_mul_ps(fscal,dz00);
1140
1141             /* Update vectorial force */
1142             fix0             = _mm256_add_ps(fix0,tx);
1143             fiy0             = _mm256_add_ps(fiy0,ty);
1144             fiz0             = _mm256_add_ps(fiz0,tz);
1145
1146             fjx0             = _mm256_add_ps(fjx0,tx);
1147             fjy0             = _mm256_add_ps(fjy0,ty);
1148             fjz0             = _mm256_add_ps(fjz0,tz);
1149
1150             /**************************
1151              * CALCULATE INTERACTIONS *
1152              **************************/
1153
1154             if (gmx_mm256_any_lt(rsq10,rcutoff2))
1155             {
1156
1157             /* Compute parameters for interactions between i and j atoms */
1158             qq10             = _mm256_mul_ps(iq1,jq0);
1159
1160             /* REACTION-FIELD ELECTROSTATICS */
1161             felec            = _mm256_mul_ps(qq10,_mm256_sub_ps(_mm256_mul_ps(rinv10,rinvsq10),krf2));
1162
1163             cutoff_mask      = _mm256_cmp_ps(rsq10,rcutoff2,_CMP_LT_OQ);
1164
1165             fscal            = felec;
1166
1167             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1168
1169             /* Calculate temporary vectorial force */
1170             tx               = _mm256_mul_ps(fscal,dx10);
1171             ty               = _mm256_mul_ps(fscal,dy10);
1172             tz               = _mm256_mul_ps(fscal,dz10);
1173
1174             /* Update vectorial force */
1175             fix1             = _mm256_add_ps(fix1,tx);
1176             fiy1             = _mm256_add_ps(fiy1,ty);
1177             fiz1             = _mm256_add_ps(fiz1,tz);
1178
1179             fjx0             = _mm256_add_ps(fjx0,tx);
1180             fjy0             = _mm256_add_ps(fjy0,ty);
1181             fjz0             = _mm256_add_ps(fjz0,tz);
1182
1183             }
1184
1185             /**************************
1186              * CALCULATE INTERACTIONS *
1187              **************************/
1188
1189             if (gmx_mm256_any_lt(rsq20,rcutoff2))
1190             {
1191
1192             /* Compute parameters for interactions between i and j atoms */
1193             qq20             = _mm256_mul_ps(iq2,jq0);
1194
1195             /* REACTION-FIELD ELECTROSTATICS */
1196             felec            = _mm256_mul_ps(qq20,_mm256_sub_ps(_mm256_mul_ps(rinv20,rinvsq20),krf2));
1197
1198             cutoff_mask      = _mm256_cmp_ps(rsq20,rcutoff2,_CMP_LT_OQ);
1199
1200             fscal            = felec;
1201
1202             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1203
1204             /* Calculate temporary vectorial force */
1205             tx               = _mm256_mul_ps(fscal,dx20);
1206             ty               = _mm256_mul_ps(fscal,dy20);
1207             tz               = _mm256_mul_ps(fscal,dz20);
1208
1209             /* Update vectorial force */
1210             fix2             = _mm256_add_ps(fix2,tx);
1211             fiy2             = _mm256_add_ps(fiy2,ty);
1212             fiz2             = _mm256_add_ps(fiz2,tz);
1213
1214             fjx0             = _mm256_add_ps(fjx0,tx);
1215             fjy0             = _mm256_add_ps(fjy0,ty);
1216             fjz0             = _mm256_add_ps(fjz0,tz);
1217
1218             }
1219
1220             /**************************
1221              * CALCULATE INTERACTIONS *
1222              **************************/
1223
1224             if (gmx_mm256_any_lt(rsq30,rcutoff2))
1225             {
1226
1227             /* Compute parameters for interactions between i and j atoms */
1228             qq30             = _mm256_mul_ps(iq3,jq0);
1229
1230             /* REACTION-FIELD ELECTROSTATICS */
1231             felec            = _mm256_mul_ps(qq30,_mm256_sub_ps(_mm256_mul_ps(rinv30,rinvsq30),krf2));
1232
1233             cutoff_mask      = _mm256_cmp_ps(rsq30,rcutoff2,_CMP_LT_OQ);
1234
1235             fscal            = felec;
1236
1237             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1238
1239             /* Calculate temporary vectorial force */
1240             tx               = _mm256_mul_ps(fscal,dx30);
1241             ty               = _mm256_mul_ps(fscal,dy30);
1242             tz               = _mm256_mul_ps(fscal,dz30);
1243
1244             /* Update vectorial force */
1245             fix3             = _mm256_add_ps(fix3,tx);
1246             fiy3             = _mm256_add_ps(fiy3,ty);
1247             fiz3             = _mm256_add_ps(fiz3,tz);
1248
1249             fjx0             = _mm256_add_ps(fjx0,tx);
1250             fjy0             = _mm256_add_ps(fjy0,ty);
1251             fjz0             = _mm256_add_ps(fjz0,tz);
1252
1253             }
1254
1255             fjptrA             = f+j_coord_offsetA;
1256             fjptrB             = f+j_coord_offsetB;
1257             fjptrC             = f+j_coord_offsetC;
1258             fjptrD             = f+j_coord_offsetD;
1259             fjptrE             = f+j_coord_offsetE;
1260             fjptrF             = f+j_coord_offsetF;
1261             fjptrG             = f+j_coord_offsetG;
1262             fjptrH             = f+j_coord_offsetH;
1263
1264             gmx_mm256_decrement_1rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,fjx0,fjy0,fjz0);
1265
1266             /* Inner loop uses 141 flops */
1267         }
1268
1269         if(jidx<j_index_end)
1270         {
1271
1272             /* Get j neighbor index, and coordinate index */
1273             jnrlistA         = jjnr[jidx];
1274             jnrlistB         = jjnr[jidx+1];
1275             jnrlistC         = jjnr[jidx+2];
1276             jnrlistD         = jjnr[jidx+3];
1277             jnrlistE         = jjnr[jidx+4];
1278             jnrlistF         = jjnr[jidx+5];
1279             jnrlistG         = jjnr[jidx+6];
1280             jnrlistH         = jjnr[jidx+7];
1281             /* Sign of each element will be negative for non-real atoms.
1282              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
1283              * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
1284              */
1285             dummy_mask = gmx_mm256_set_m128(gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx+4)),_mm_setzero_si128())),
1286                                             gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128())));
1287
1288             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
1289             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
1290             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
1291             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
1292             jnrE       = (jnrlistE>=0) ? jnrlistE : 0;
1293             jnrF       = (jnrlistF>=0) ? jnrlistF : 0;
1294             jnrG       = (jnrlistG>=0) ? jnrlistG : 0;
1295             jnrH       = (jnrlistH>=0) ? jnrlistH : 0;
1296             j_coord_offsetA  = DIM*jnrA;
1297             j_coord_offsetB  = DIM*jnrB;
1298             j_coord_offsetC  = DIM*jnrC;
1299             j_coord_offsetD  = DIM*jnrD;
1300             j_coord_offsetE  = DIM*jnrE;
1301             j_coord_offsetF  = DIM*jnrF;
1302             j_coord_offsetG  = DIM*jnrG;
1303             j_coord_offsetH  = DIM*jnrH;
1304
1305             /* load j atom coordinates */
1306             gmx_mm256_load_1rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
1307                                                  x+j_coord_offsetC,x+j_coord_offsetD,
1308                                                  x+j_coord_offsetE,x+j_coord_offsetF,
1309                                                  x+j_coord_offsetG,x+j_coord_offsetH,
1310                                                  &jx0,&jy0,&jz0);
1311
1312             /* Calculate displacement vector */
1313             dx00             = _mm256_sub_ps(ix0,jx0);
1314             dy00             = _mm256_sub_ps(iy0,jy0);
1315             dz00             = _mm256_sub_ps(iz0,jz0);
1316             dx10             = _mm256_sub_ps(ix1,jx0);
1317             dy10             = _mm256_sub_ps(iy1,jy0);
1318             dz10             = _mm256_sub_ps(iz1,jz0);
1319             dx20             = _mm256_sub_ps(ix2,jx0);
1320             dy20             = _mm256_sub_ps(iy2,jy0);
1321             dz20             = _mm256_sub_ps(iz2,jz0);
1322             dx30             = _mm256_sub_ps(ix3,jx0);
1323             dy30             = _mm256_sub_ps(iy3,jy0);
1324             dz30             = _mm256_sub_ps(iz3,jz0);
1325
1326             /* Calculate squared distance and things based on it */
1327             rsq00            = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
1328             rsq10            = gmx_mm256_calc_rsq_ps(dx10,dy10,dz10);
1329             rsq20            = gmx_mm256_calc_rsq_ps(dx20,dy20,dz20);
1330             rsq30            = gmx_mm256_calc_rsq_ps(dx30,dy30,dz30);
1331
1332             rinv00           = gmx_mm256_invsqrt_ps(rsq00);
1333             rinv10           = gmx_mm256_invsqrt_ps(rsq10);
1334             rinv20           = gmx_mm256_invsqrt_ps(rsq20);
1335             rinv30           = gmx_mm256_invsqrt_ps(rsq30);
1336
1337             rinvsq10         = _mm256_mul_ps(rinv10,rinv10);
1338             rinvsq20         = _mm256_mul_ps(rinv20,rinv20);
1339             rinvsq30         = _mm256_mul_ps(rinv30,rinv30);
1340
1341             /* Load parameters for j particles */
1342             jq0              = gmx_mm256_load_8real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
1343                                                                  charge+jnrC+0,charge+jnrD+0,
1344                                                                  charge+jnrE+0,charge+jnrF+0,
1345                                                                  charge+jnrG+0,charge+jnrH+0);
1346             vdwjidx0A        = 2*vdwtype[jnrA+0];
1347             vdwjidx0B        = 2*vdwtype[jnrB+0];
1348             vdwjidx0C        = 2*vdwtype[jnrC+0];
1349             vdwjidx0D        = 2*vdwtype[jnrD+0];
1350             vdwjidx0E        = 2*vdwtype[jnrE+0];
1351             vdwjidx0F        = 2*vdwtype[jnrF+0];
1352             vdwjidx0G        = 2*vdwtype[jnrG+0];
1353             vdwjidx0H        = 2*vdwtype[jnrH+0];
1354
1355             fjx0             = _mm256_setzero_ps();
1356             fjy0             = _mm256_setzero_ps();
1357             fjz0             = _mm256_setzero_ps();
1358
1359             /**************************
1360              * CALCULATE INTERACTIONS *
1361              **************************/
1362
1363             r00              = _mm256_mul_ps(rsq00,rinv00);
1364             r00              = _mm256_andnot_ps(dummy_mask,r00);
1365
1366             /* Compute parameters for interactions between i and j atoms */
1367             gmx_mm256_load_8pair_swizzle_ps(vdwioffsetptr0+vdwjidx0A,
1368                                             vdwioffsetptr0+vdwjidx0B,
1369                                             vdwioffsetptr0+vdwjidx0C,
1370                                             vdwioffsetptr0+vdwjidx0D,
1371                                             vdwioffsetptr0+vdwjidx0E,
1372                                             vdwioffsetptr0+vdwjidx0F,
1373                                             vdwioffsetptr0+vdwjidx0G,
1374                                             vdwioffsetptr0+vdwjidx0H,
1375                                             &c6_00,&c12_00);
1376
1377             /* Calculate table index by multiplying r with table scale and truncate to integer */
1378             rt               = _mm256_mul_ps(r00,vftabscale);
1379             vfitab           = _mm256_cvttps_epi32(rt);
1380             vfeps            = _mm256_sub_ps(rt,_mm256_round_ps(rt, _MM_FROUND_FLOOR));
1381             /*         AVX1 does not support 256-bit integer operations, so now we go to 128-bit mode... */
1382             vfitab_lo        = _mm256_extractf128_si256(vfitab,0x0);
1383             vfitab_hi        = _mm256_extractf128_si256(vfitab,0x1);
1384             vfitab_lo        = _mm_slli_epi32(vfitab_lo,3);
1385             vfitab_hi        = _mm_slli_epi32(vfitab_hi,3);
1386
1387             /* CUBIC SPLINE TABLE DISPERSION */
1388             Y                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,0)),
1389                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,0)));
1390             F                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,1)),
1391                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,1)));
1392             G                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,2)),
1393                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,2)));
1394             H                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,3)),
1395                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,3)));
1396             GMX_MM256_HALFTRANSPOSE4_PS(Y,F,G,H);
1397             Heps             = _mm256_mul_ps(vfeps,H);
1398             Fp               = _mm256_add_ps(F,_mm256_mul_ps(vfeps,_mm256_add_ps(G,Heps)));
1399             FF               = _mm256_add_ps(Fp,_mm256_mul_ps(vfeps,_mm256_add_ps(G,_mm256_add_ps(Heps,Heps))));
1400             fvdw6            = _mm256_mul_ps(c6_00,FF);
1401
1402             /* CUBIC SPLINE TABLE REPULSION */
1403             vfitab_lo        = _mm_add_epi32(vfitab_lo,ifour);
1404             vfitab_hi        = _mm_add_epi32(vfitab_hi,ifour);
1405             Y                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,0)),
1406                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,0)));
1407             F                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,1)),
1408                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,1)));
1409             G                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,2)),
1410                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,2)));
1411             H                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,3)),
1412                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,3)));
1413             GMX_MM256_HALFTRANSPOSE4_PS(Y,F,G,H);
1414             Heps             = _mm256_mul_ps(vfeps,H);
1415             Fp               = _mm256_add_ps(F,_mm256_mul_ps(vfeps,_mm256_add_ps(G,Heps)));
1416             FF               = _mm256_add_ps(Fp,_mm256_mul_ps(vfeps,_mm256_add_ps(G,_mm256_add_ps(Heps,Heps))));
1417             fvdw12           = _mm256_mul_ps(c12_00,FF);
1418             fvdw             = _mm256_xor_ps(signbit,_mm256_mul_ps(_mm256_add_ps(fvdw6,fvdw12),_mm256_mul_ps(vftabscale,rinv00)));
1419
1420             fscal            = fvdw;
1421
1422             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1423
1424             /* Calculate temporary vectorial force */
1425             tx               = _mm256_mul_ps(fscal,dx00);
1426             ty               = _mm256_mul_ps(fscal,dy00);
1427             tz               = _mm256_mul_ps(fscal,dz00);
1428
1429             /* Update vectorial force */
1430             fix0             = _mm256_add_ps(fix0,tx);
1431             fiy0             = _mm256_add_ps(fiy0,ty);
1432             fiz0             = _mm256_add_ps(fiz0,tz);
1433
1434             fjx0             = _mm256_add_ps(fjx0,tx);
1435             fjy0             = _mm256_add_ps(fjy0,ty);
1436             fjz0             = _mm256_add_ps(fjz0,tz);
1437
1438             /**************************
1439              * CALCULATE INTERACTIONS *
1440              **************************/
1441
1442             if (gmx_mm256_any_lt(rsq10,rcutoff2))
1443             {
1444
1445             /* Compute parameters for interactions between i and j atoms */
1446             qq10             = _mm256_mul_ps(iq1,jq0);
1447
1448             /* REACTION-FIELD ELECTROSTATICS */
1449             felec            = _mm256_mul_ps(qq10,_mm256_sub_ps(_mm256_mul_ps(rinv10,rinvsq10),krf2));
1450
1451             cutoff_mask      = _mm256_cmp_ps(rsq10,rcutoff2,_CMP_LT_OQ);
1452
1453             fscal            = felec;
1454
1455             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1456
1457             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1458
1459             /* Calculate temporary vectorial force */
1460             tx               = _mm256_mul_ps(fscal,dx10);
1461             ty               = _mm256_mul_ps(fscal,dy10);
1462             tz               = _mm256_mul_ps(fscal,dz10);
1463
1464             /* Update vectorial force */
1465             fix1             = _mm256_add_ps(fix1,tx);
1466             fiy1             = _mm256_add_ps(fiy1,ty);
1467             fiz1             = _mm256_add_ps(fiz1,tz);
1468
1469             fjx0             = _mm256_add_ps(fjx0,tx);
1470             fjy0             = _mm256_add_ps(fjy0,ty);
1471             fjz0             = _mm256_add_ps(fjz0,tz);
1472
1473             }
1474
1475             /**************************
1476              * CALCULATE INTERACTIONS *
1477              **************************/
1478
1479             if (gmx_mm256_any_lt(rsq20,rcutoff2))
1480             {
1481
1482             /* Compute parameters for interactions between i and j atoms */
1483             qq20             = _mm256_mul_ps(iq2,jq0);
1484
1485             /* REACTION-FIELD ELECTROSTATICS */
1486             felec            = _mm256_mul_ps(qq20,_mm256_sub_ps(_mm256_mul_ps(rinv20,rinvsq20),krf2));
1487
1488             cutoff_mask      = _mm256_cmp_ps(rsq20,rcutoff2,_CMP_LT_OQ);
1489
1490             fscal            = felec;
1491
1492             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1493
1494             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1495
1496             /* Calculate temporary vectorial force */
1497             tx               = _mm256_mul_ps(fscal,dx20);
1498             ty               = _mm256_mul_ps(fscal,dy20);
1499             tz               = _mm256_mul_ps(fscal,dz20);
1500
1501             /* Update vectorial force */
1502             fix2             = _mm256_add_ps(fix2,tx);
1503             fiy2             = _mm256_add_ps(fiy2,ty);
1504             fiz2             = _mm256_add_ps(fiz2,tz);
1505
1506             fjx0             = _mm256_add_ps(fjx0,tx);
1507             fjy0             = _mm256_add_ps(fjy0,ty);
1508             fjz0             = _mm256_add_ps(fjz0,tz);
1509
1510             }
1511
1512             /**************************
1513              * CALCULATE INTERACTIONS *
1514              **************************/
1515
1516             if (gmx_mm256_any_lt(rsq30,rcutoff2))
1517             {
1518
1519             /* Compute parameters for interactions between i and j atoms */
1520             qq30             = _mm256_mul_ps(iq3,jq0);
1521
1522             /* REACTION-FIELD ELECTROSTATICS */
1523             felec            = _mm256_mul_ps(qq30,_mm256_sub_ps(_mm256_mul_ps(rinv30,rinvsq30),krf2));
1524
1525             cutoff_mask      = _mm256_cmp_ps(rsq30,rcutoff2,_CMP_LT_OQ);
1526
1527             fscal            = felec;
1528
1529             fscal            = _mm256_and_ps(fscal,cutoff_mask);
1530
1531             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
1532
1533             /* Calculate temporary vectorial force */
1534             tx               = _mm256_mul_ps(fscal,dx30);
1535             ty               = _mm256_mul_ps(fscal,dy30);
1536             tz               = _mm256_mul_ps(fscal,dz30);
1537
1538             /* Update vectorial force */
1539             fix3             = _mm256_add_ps(fix3,tx);
1540             fiy3             = _mm256_add_ps(fiy3,ty);
1541             fiz3             = _mm256_add_ps(fiz3,tz);
1542
1543             fjx0             = _mm256_add_ps(fjx0,tx);
1544             fjy0             = _mm256_add_ps(fjy0,ty);
1545             fjz0             = _mm256_add_ps(fjz0,tz);
1546
1547             }
1548
1549             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
1550             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
1551             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
1552             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
1553             fjptrE             = (jnrlistE>=0) ? f+j_coord_offsetE : scratch;
1554             fjptrF             = (jnrlistF>=0) ? f+j_coord_offsetF : scratch;
1555             fjptrG             = (jnrlistG>=0) ? f+j_coord_offsetG : scratch;
1556             fjptrH             = (jnrlistH>=0) ? f+j_coord_offsetH : scratch;
1557
1558             gmx_mm256_decrement_1rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,fjx0,fjy0,fjz0);
1559
1560             /* Inner loop uses 142 flops */
1561         }
1562
1563         /* End of innermost loop */
1564
1565         gmx_mm256_update_iforce_4atom_swizzle_ps(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
1566                                                  f+i_coord_offset,fshift+i_shift_offset);
1567
1568         /* Increment number of inner iterations */
1569         inneriter                  += j_index_end - j_index_start;
1570
1571         /* Outer loop uses 24 flops */
1572     }
1573
1574     /* Increment number of outer iterations */
1575     outeriter        += nri;
1576
1577     /* Update outer/inner flops */
1578
1579     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_F,outeriter*24 + inneriter*142);
1580 }