src/gromacs/gmxlib/nonbonded/nb_kernel_avx_256_single/nb_kernel_template_avx_256_single.pre

   1 /* #if 0 */
   2 /*
   3  * This file is part of the GROMACS molecular simulation package.
   4  *
   5  * Copyright (c) 2012,2013, by the GROMACS development team, led by
   6  * Mark Abraham, David van der Spoel, Berk Hess, and Erik Lindahl,
   7  * and including many others, as listed in the AUTHORS file in the
   8  * top-level source directory and at http://www.gromacs.org.
   9  *
  10  * GROMACS is free software; you can redistribute it and/or
  11  * modify it under the terms of the GNU Lesser General Public License
  12  * as published by the Free Software Foundation; either version 2.1
  13  * of the License, or (at your option) any later version.
  14  *
  15  * GROMACS is distributed in the hope that it will be useful,
  16  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  17  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  18  * Lesser General Public License for more details.
  19  *
  20  * You should have received a copy of the GNU Lesser General Public
  21  * License along with GROMACS; if not, see
  22  * http://www.gnu.org/licenses, or write to the Free Software Foundation,
  23  * Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA.
  24  *
  25  * If you want to redistribute modifications to GROMACS, please
  26  * consider that scientific software is very special. Version
  27  * control is crucial - bugs must be traceable. We will be happy to
  28  * consider code for inclusion in the official distribution, but
  29  * derived work must not be called official GROMACS. Details are found
  30  * in the README & COPYING files - if they are missing, get the
  31  * official version at http://www.gromacs.org.
  32  *
  33  * To help us fund GROMACS development, we humbly ask that you cite
  34  * the research papers on the package. Check out http://www.gromacs.org.
  35  */
  36 #error This file must be processed with the Gromacs pre-preprocessor
  37 /* #endif */
  38 /* #if INCLUDE_HEADER */
  39 #ifdef HAVE_CONFIG_H
  40 #include <config.h>
  41 #endif
  42
  43 #include <math.h>
  44
  45 #include "../nb_kernel.h"
  46 #include "types/simple.h"
  47 #include "vec.h"
  48 #include "nrnb.h"
  49
  50 #include "gromacs/simd/math_x86_avx_256_single.h"
  51 #include "kernelutil_x86_avx_256_single.h"
  52 /* #endif */
  53
  54 /* ## List of variables set by the generating script:                                    */
  55 /* ##                                                                                    */
  56 /* ## Setttings that apply to the entire kernel:                                         */
  57 /* ## KERNEL_ELEC:           String, choice for electrostatic interactions               */
  58 /* ## KERNEL_VDW:            String, choice for van der Waals interactions               */
  59 /* ## KERNEL_NAME:           String, name of this kernel                                 */
  60 /* ## KERNEL_VF:             String telling if we calculate potential, force, or both    */
  61 /* ## GEOMETRY_I/GEOMETRY_J: String, name of each geometry, e.g. 'Water3' or '1Particle' */
  62 /* ##                                                                                    */
  63 /* ## Setttings that apply to particles in the outer (I) or inner (J) loops:             */
  64 /* ## PARTICLES_I[]/         Arrays with lists of i/j particles to use in kernel. It is  */
  65 /* ## PARTICLES_J[]:         just [0] for particle geometry, but can be longer for water */
  66 /* ## PARTICLES_ELEC_I[]/    Arrays with lists of i/j particle that have electrostatics  */
  67 /* ## PARTICLES_ELEC_J[]:    interactions that should be calculated in this kernel.      */
  68 /* ## PARTICLES_VDW_I[]/     Arrays with the list of i/j particle that have VdW          */
  69 /* ## PARTICLES_VDW_J[]:     interactions that should be calculated in this kernel.      */
  70 /* ##                                                                                    */
  71 /* ## Setttings for pairs of interactions (e.g. 2nd i particle against 1st j particle)   */
  72 /* ## PAIRS_IJ[]:            Array with (i,j) tuples of pairs for which interactions     */
  73 /* ##                        should be calculated in this kernel. Zero-charge particles  */
  74 /* ##                        do not have interactions with particles without vdw, and    */
  75 /* ##                        Vdw-only interactions are not evaluated in a no-vdw-kernel. */
  76 /* ## INTERACTION_FLAGS[][]: 2D matrix, dimension e.g. 3*3 for water-water interactions. */
  77 /* ##                        For each i-j pair, the element [I][J] is a list of strings  */
  78 /* ##                        defining properties/flags of this interaction. Examples     */
  79 /* ##                        include 'electrostatics'/'vdw' if that type of interaction  */
  80 /* ##                        should be evaluated, 'rsq'/'rinv'/'rinvsq' if those values  */
  81 /* ##                        are needed, and 'exactcutoff' or 'shift','switch' to        */
  82 /* ##                        decide if the force/potential should be modified. This way  */
  83 /* ##                        we only calculate values absolutely needed for each case.   */
  84
  85 /* ## Calculate the size and offset for (merged/interleaved) table data */
  86
  87 /*
  88  * Gromacs nonbonded kernel:   {KERNEL_NAME}
  89  * Electrostatics interaction: {KERNEL_ELEC}
  90  * VdW interaction:            {KERNEL_VDW}
  91  * Geometry:                   {GEOMETRY_I}-{GEOMETRY_J}
  92  * Calculate force/pot:        {KERNEL_VF}
  93  */
  94 void
  95 {KERNEL_NAME}
  96                     (t_nblist                    * gmx_restrict       nlist,
  97                      rvec                        * gmx_restrict          xx,
  98                      rvec                        * gmx_restrict          ff,
  99                      t_forcerec                  * gmx_restrict          fr,
 100                      t_mdatoms                   * gmx_restrict     mdatoms,
 101                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 102                      t_nrnb                      * gmx_restrict        nrnb)
 103 {
 104     /* ## Not all variables are used for all kernels, but any optimizing compiler fixes that, */
 105     /* ## so there is no point in going to extremes to exclude variables that are not needed. */
 106     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
 107      * just 0 for non-waters.
 108      * Suffixes A,B,C,D,E,F,G,H refer to j loop unrolling done with AVX, e.g. for the eight different
 109      * jnr indices corresponding to data put in the four positions in the SIMD register.
 110      */
 111     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
 112     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
 113     int              jnrA,jnrB,jnrC,jnrD;
 114     int              jnrE,jnrF,jnrG,jnrH;
 115     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
 116     int              jnrlistE,jnrlistF,jnrlistG,jnrlistH;
 117     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
 118     int              j_coord_offsetE,j_coord_offsetF,j_coord_offsetG,j_coord_offsetH;
 119     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 120     real             rcutoff_scalar;
 121     real             *shiftvec,*fshift,*x,*f;
 122     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD,*fjptrE,*fjptrF,*fjptrG,*fjptrH;
 123     real             scratch[4*DIM];
 124     __m256           tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
 125     /* #for I in PARTICLES_I */
 126     real *           vdwioffsetptr{I};
 127     __m256           ix{I},iy{I},iz{I},fix{I},fiy{I},fiz{I},iq{I},isai{I};
 128     /* #endfor */
 129     /* #for J in PARTICLES_J */
 130     int              vdwjidx{J}A,vdwjidx{J}B,vdwjidx{J}C,vdwjidx{J}D,vdwjidx{J}E,vdwjidx{J}F,vdwjidx{J}G,vdwjidx{J}H;
 131     __m256           jx{J},jy{J},jz{J},fjx{J},fjy{J},fjz{J},jq{J},isaj{J};
 132     /* #endfor */
 133     /* #for I,J in PAIRS_IJ */
 134     __m256           dx{I}{J},dy{I}{J},dz{I}{J},rsq{I}{J},rinv{I}{J},rinvsq{I}{J},r{I}{J},qq{I}{J},c6_{I}{J},c12_{I}{J};
 135     /* #endfor */
 136     /* #if KERNEL_ELEC != 'None' */
 137     __m256           velec,felec,velecsum,facel,crf,krf,krf2;
 138     real             *charge;
 139     /* #endif */
 140     /* #if 'GeneralizedBorn' in KERNEL_ELEC */
 141     __m256i          gbitab;
 142     __m128i          gbitab_lo,gbitab_hi;
 143     __m256           vgb,fgb,vgbsum,dvdasum,gbscale,gbtabscale,isaprod,gbqqfactor,gbinvepsdiff,gbeps,dvdatmp;
 144     __m256           minushalf = _mm256_set1_ps(-0.5);
 145     real             *invsqrta,*dvda,*gbtab;
 146     /* #endif */
 147     /* #if KERNEL_VDW != 'None' */
 148     int              nvdwtype;
 149     __m256           rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
 150     int              *vdwtype;
 151     real             *vdwparam;
 152     __m256           one_sixth   = _mm256_set1_ps(1.0/6.0);
 153     __m256           one_twelfth = _mm256_set1_ps(1.0/12.0);
 154     /* #endif */
 155     /* #if 'Table' in KERNEL_ELEC or 'GeneralizedBorn' in KERNEL_ELEC or 'Table' in KERNEL_VDW */
 156     __m256i          vfitab;
 157     __m128i          vfitab_lo,vfitab_hi;
 158     __m128i          ifour       = _mm_set1_epi32(4);
 159     __m256           rt,vfeps,vftabscale,Y,F,G,H,Heps,Fp,VV,FF;
 160     real             *vftab;
 161     /* #endif */
 162     /* #if 'Ewald' in KERNEL_ELEC */
 163     __m256i          ewitab;
 164     __m128i          ewitab_lo,ewitab_hi;
 165     __m256           ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
 166     __m256           beta,beta2,beta3,zeta2,pmecorrF,pmecorrV,rinv3;
 167     real             *ewtab;
 168     /* #endif */
 169     /* #if 'PotentialSwitch' in [KERNEL_MOD_ELEC,KERNEL_MOD_VDW] */
 170     __m256           rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
 171     real             rswitch_scalar,d_scalar;
 172     /* #endif */
 173     __m256           dummy_mask,cutoff_mask;
 174     __m256           signbit = _mm256_castsi256_ps( _mm256_set1_epi32(0x80000000) );
 175     __m256           one     = _mm256_set1_ps(1.0);
 176     __m256           two     = _mm256_set1_ps(2.0);
 177     x                = xx[0];
 178     f                = ff[0];
 179
 180     nri              = nlist->nri;
 181     iinr             = nlist->iinr;
 182     jindex           = nlist->jindex;
 183     jjnr             = nlist->jjnr;
 184     shiftidx         = nlist->shift;
 185     gid              = nlist->gid;
 186     shiftvec         = fr->shift_vec[0];
 187     fshift           = fr->fshift[0];
 188     /* #if KERNEL_ELEC != 'None' */
 189     facel            = _mm256_set1_ps(fr->epsfac);
 190     charge           = mdatoms->chargeA;
 191     /*     #if 'ReactionField' in KERNEL_ELEC */
 192     krf              = _mm256_set1_ps(fr->ic->k_rf);
 193     krf2             = _mm256_set1_ps(fr->ic->k_rf*2.0);
 194     crf              = _mm256_set1_ps(fr->ic->c_rf);
 195     /*     #endif */
 196     /* #endif */
 197     /* #if KERNEL_VDW != 'None' */
 198     nvdwtype         = fr->ntype;
 199     vdwparam         = fr->nbfp;
 200     vdwtype          = mdatoms->typeA;
 201     /* #endif */
 202
 203     /* #if 'Table' in KERNEL_ELEC and 'Table' in KERNEL_VDW */
 204     vftab            = kernel_data->table_elec_vdw->data;
 205     vftabscale       = _mm256_set1_ps(kernel_data->table_elec_vdw->scale);
 206     /* #elif 'Table' in KERNEL_ELEC */
 207     vftab            = kernel_data->table_elec->data;
 208     vftabscale       = _mm256_set1_ps(kernel_data->table_elec->scale);
 209     /* #elif 'Table' in KERNEL_VDW */
 210     vftab            = kernel_data->table_vdw->data;
 211     vftabscale       = _mm256_set1_ps(kernel_data->table_vdw->scale);
 212     /* #endif */
 213
 214     /* #if 'Ewald' in KERNEL_ELEC */
 215     sh_ewald         = _mm256_set1_ps(fr->ic->sh_ewald);
 216     beta             = _mm256_set1_ps(fr->ic->ewaldcoeff);
 217     beta2            = _mm256_mul_ps(beta,beta);
 218     beta3            = _mm256_mul_ps(beta,beta2);
 219
 220     /*     #if KERNEL_VF=='Force' and KERNEL_MOD_ELEC!='PotentialSwitch' */
 221     ewtab            = fr->ic->tabq_coul_F;
 222     ewtabscale       = _mm256_set1_ps(fr->ic->tabq_scale);
 223     ewtabhalfspace   = _mm256_set1_ps(0.5/fr->ic->tabq_scale);
 224     /*     #else */
 225     ewtab            = fr->ic->tabq_coul_FDV0;
 226     ewtabscale       = _mm256_set1_ps(fr->ic->tabq_scale);
 227     ewtabhalfspace   = _mm256_set1_ps(0.5/fr->ic->tabq_scale);
 228      /*     #endif */
 229     /* #endif */
 230
 231     /* #if KERNEL_ELEC=='GeneralizedBorn' */
 232     invsqrta         = fr->invsqrta;
 233     dvda             = fr->dvda;
 234     gbtabscale       = _mm256_set1_ps(fr->gbtab.scale);
 235     gbtab            = fr->gbtab.data;
 236     gbinvepsdiff     = _mm256_set1_ps((1.0/fr->epsilon_r) - (1.0/fr->gb_epsilon_solvent));
 237     /* #endif */
 238
 239     /* #if 'Water' in GEOMETRY_I */
 240     /* Setup water-specific parameters */
 241     inr              = nlist->iinr[0];
 242     /*     #for I in PARTICLES_ELEC_I */
 243     iq{I}              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+{I}]));
 244     /*     #endfor */
 245     /*     #for I in PARTICLES_VDW_I */
 246     vdwioffsetptr{I}   = vdwparam+2*nvdwtype*vdwtype[inr+{I}];
 247     /*     #endfor */
 248     /* #endif */
 249
 250     /* #if 'Water' in GEOMETRY_J */
 251     /*     #for J in PARTICLES_ELEC_J */
 252     jq{J}              = _mm256_set1_ps(charge[inr+{J}]);
 253     /*     #endfor */
 254     /*     #for J in PARTICLES_VDW_J */
 255     vdwjidx{J}A        = 2*vdwtype[inr+{J}];
 256     /*     #endfor */
 257     /*     #for I,J in PAIRS_IJ */
 258     /*         #if 'electrostatics' in INTERACTION_FLAGS[I][J] */
 259     qq{I}{J}             = _mm256_mul_ps(iq{I},jq{J});
 260     /*         #endif */
 261     /*         #if 'vdw' in INTERACTION_FLAGS[I][J] */
 262     c6_{I}{J}            = _mm256_set1_ps(vdwioffsetptr{I}[vdwjidx{J}A]);
 263     c12_{I}{J}           = _mm256_set1_ps(vdwioffsetptr{I}[vdwjidx{J}A+1]);
 264     /*         #endif */
 265     /*     #endfor */
 266     /* #endif */
 267
 268     /* #if KERNEL_MOD_ELEC!='None' or KERNEL_MOD_VDW!='None' */
 269     /*     #if KERNEL_ELEC!='None' */
 270     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 271     rcutoff_scalar   = fr->rcoulomb;
 272     /*     #else */
 273     rcutoff_scalar   = fr->rvdw;
 274     /*     #endif */
 275     rcutoff          = _mm256_set1_ps(rcutoff_scalar);
 276     rcutoff2         = _mm256_mul_ps(rcutoff,rcutoff);
 277     /* #endif */
 278
 279     /* #if KERNEL_MOD_VDW=='PotentialShift' */
 280     sh_vdw_invrcut6  = _mm256_set1_ps(fr->ic->sh_invrc6);
 281     rvdw             = _mm256_set1_ps(fr->rvdw);
 282     /* #endif */
 283
 284     /* #if 'PotentialSwitch' in [KERNEL_MOD_ELEC,KERNEL_MOD_VDW] */
 285     /*     #if KERNEL_MOD_ELEC=='PotentialSwitch'  */
 286     rswitch_scalar   = fr->rcoulomb_switch;
 287     rswitch          = _mm256_set1_ps(rswitch_scalar);
 288     /*     #else */
 289     rswitch_scalar   = fr->rvdw_switch;
 290     rswitch          = _mm256_set1_ps(rswitch_scalar);
 291     /*     #endif */
 292     /* Setup switch parameters */
 293     d_scalar         = rcutoff_scalar-rswitch_scalar;
 294     d                = _mm256_set1_ps(d_scalar);
 295     swV3             = _mm256_set1_ps(-10.0/(d_scalar*d_scalar*d_scalar));
 296     swV4             = _mm256_set1_ps( 15.0/(d_scalar*d_scalar*d_scalar*d_scalar));
 297     swV5             = _mm256_set1_ps( -6.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
 298     /*     #if 'Force' in KERNEL_VF */
 299     swF2             = _mm256_set1_ps(-30.0/(d_scalar*d_scalar*d_scalar));
 300     swF3             = _mm256_set1_ps( 60.0/(d_scalar*d_scalar*d_scalar*d_scalar));
 301     swF4             = _mm256_set1_ps(-30.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
 302     /*     #endif */
 303     /* #endif */
 304
 305     /* Avoid stupid compiler warnings */
 306     jnrA = jnrB = jnrC = jnrD = jnrE = jnrF = jnrG = jnrH = 0;
 307     j_coord_offsetA = 0;
 308     j_coord_offsetB = 0;
 309     j_coord_offsetC = 0;
 310     j_coord_offsetD = 0;
 311     j_coord_offsetE = 0;
 312     j_coord_offsetF = 0;
 313     j_coord_offsetG = 0;
 314     j_coord_offsetH = 0;
 315
 316     /* ## Keep track of the floating point operations we issue for reporting! */
 317     /* #define OUTERFLOPS 0 */
 318     outeriter        = 0;
 319     inneriter        = 0;
 320
 321     for(iidx=0;iidx<4*DIM;iidx++)
 322     {
 323         scratch[iidx] = 0.0;
 324     }
 325
 326     /* Start outer loop over neighborlists */
 327     for(iidx=0; iidx<nri; iidx++)
 328     {
 329         /* Load shift vector for this list */
 330         i_shift_offset   = DIM*shiftidx[iidx];
 331
 332         /* Load limits for loop over neighbors */
 333         j_index_start    = jindex[iidx];
 334         j_index_end      = jindex[iidx+1];
 335
 336         /* Get outer coordinate index */
 337         inr              = iinr[iidx];
 338         i_coord_offset   = DIM*inr;
 339
 340         /* Load i particle coords and add shift vector */
 341         /* #if GEOMETRY_I == 'Particle' */
 342         gmx_mm256_load_shift_and_1rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,&ix0,&iy0,&iz0);
 343         /* #elif GEOMETRY_I == 'Water3' */
 344         gmx_mm256_load_shift_and_3rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,
 345                                                     &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2);
 346         /* #elif GEOMETRY_I == 'Water4' */
 347         /*     #if 0 in PARTICLES_I                 */
 348         gmx_mm256_load_shift_and_4rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,
 349                                                     &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
 350         /*     #else                                */
 351         gmx_mm256_load_shift_and_3rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset+DIM,
 352                                                     &ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
 353         /*     #endif                               */
 354         /* #endif                                   */
 355
 356         /* #if 'Force' in KERNEL_VF */
 357         /*     #for I in PARTICLES_I */
 358         fix{I}             = _mm256_setzero_ps();
 359         fiy{I}             = _mm256_setzero_ps();
 360         fiz{I}             = _mm256_setzero_ps();
 361         /*     #endfor */
 362         /* #endif */
 363
 364         /* ## For water we already preloaded parameters at the start of the kernel */
 365         /* #if not 'Water' in GEOMETRY_I */
 366         /* Load parameters for i particles */
 367         /*     #for I in PARTICLES_ELEC_I */
 368         iq{I}              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+{I}]));
 369         /*         #define OUTERFLOPS OUTERFLOPS+1 */
 370         /*         #if KERNEL_ELEC=='GeneralizedBorn' */
 371         isai{I}            = _mm256_set1_ps(invsqrta[inr+{I}]);
 372         /*         #endif */
 373         /*     #endfor */
 374         /*     #for I in PARTICLES_VDW_I */
 375         vdwioffsetptr{I}   = vdwparam+2*nvdwtype*vdwtype[inr+{I}];
 376         /*     #endfor */
 377         /* #endif */
 378
 379         /* #if 'Potential' in KERNEL_VF */
 380         /* Reset potential sums */
 381         /*     #if KERNEL_ELEC != 'None' */
 382         velecsum         = _mm256_setzero_ps();
 383         /*     #endif */
 384         /*     #if 'GeneralizedBorn' in KERNEL_ELEC */
 385         vgbsum           = _mm256_setzero_ps();
 386         /*     #endif */
 387         /*     #if KERNEL_VDW != 'None' */
 388         vvdwsum          = _mm256_setzero_ps();
 389         /*     #endif */
 390         /* #endif */
 391         /*     #if 'GeneralizedBorn' in KERNEL_ELEC and 'Force' in KERNEL_VF */
 392         dvdasum          = _mm256_setzero_ps();
 393         /*     #endif */
 394
 395         /* #for ROUND in ['Loop','Epilogue'] */
 396
 397         /* #if ROUND =='Loop' */
 398         /* Start inner kernel loop */
 399         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+7]>=0; jidx+=8)
 400         {
 401         /* ## First round is normal loop (next statement resets indentation) */
 402         /*     #if 0 */
 403         }
 404         /*     #endif */
 405         /* #else */
 406         if(jidx<j_index_end)
 407         {
 408         /* ## Second round is epilogue */
 409         /* #endif */
 410         /* #define INNERFLOPS 0 */
 411
 412             /* Get j neighbor index, and coordinate index */
 413             /* #if ROUND =='Loop' */
 414             jnrA             = jjnr[jidx];
 415             jnrB             = jjnr[jidx+1];
 416             jnrC             = jjnr[jidx+2];
 417             jnrD             = jjnr[jidx+3];
 418             jnrE             = jjnr[jidx+4];
 419             jnrF             = jjnr[jidx+5];
 420             jnrG             = jjnr[jidx+6];
 421             jnrH             = jjnr[jidx+7];
 422             /* #else */
 423             jnrlistA         = jjnr[jidx];
 424             jnrlistB         = jjnr[jidx+1];
 425             jnrlistC         = jjnr[jidx+2];
 426             jnrlistD         = jjnr[jidx+3];
 427             jnrlistE         = jjnr[jidx+4];
 428             jnrlistF         = jjnr[jidx+5];
 429             jnrlistG         = jjnr[jidx+6];
 430             jnrlistH         = jjnr[jidx+7];
 431             /* Sign of each element will be negative for non-real atoms.
 432              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
 433              * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
 434              */
 435             dummy_mask = gmx_mm256_set_m128(gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx+4)),_mm_setzero_si128())),
 436                                             gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128())));
 437
 438             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
 439             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
 440             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
 441             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
 442             jnrE       = (jnrlistE>=0) ? jnrlistE : 0;
 443             jnrF       = (jnrlistF>=0) ? jnrlistF : 0;
 444             jnrG       = (jnrlistG>=0) ? jnrlistG : 0;
 445             jnrH       = (jnrlistH>=0) ? jnrlistH : 0;
 446             /* #endif */
 447             j_coord_offsetA  = DIM*jnrA;
 448             j_coord_offsetB  = DIM*jnrB;
 449             j_coord_offsetC  = DIM*jnrC;
 450             j_coord_offsetD  = DIM*jnrD;
 451             j_coord_offsetE  = DIM*jnrE;
 452             j_coord_offsetF  = DIM*jnrF;
 453             j_coord_offsetG  = DIM*jnrG;
 454             j_coord_offsetH  = DIM*jnrH;
 455
 456             /* load j atom coordinates */
 457             /* #if GEOMETRY_J == 'Particle'             */
 458             gmx_mm256_load_1rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 459                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 460                                                  x+j_coord_offsetE,x+j_coord_offsetF,
 461                                                  x+j_coord_offsetG,x+j_coord_offsetH,
 462                                                  &jx0,&jy0,&jz0);
 463             /* #elif GEOMETRY_J == 'Water3'             */
 464             gmx_mm256_load_3rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 465                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 466                                                  x+j_coord_offsetE,x+j_coord_offsetF,
 467                                                  x+j_coord_offsetG,x+j_coord_offsetH,
 468                                               &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,&jy2,&jz2);
 469             /* #elif GEOMETRY_J == 'Water4'             */
 470             /*     #if 0 in PARTICLES_J                 */
 471             gmx_mm256_load_4rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 472                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 473                                                  x+j_coord_offsetE,x+j_coord_offsetF,
 474                                                  x+j_coord_offsetG,x+j_coord_offsetH,
 475                                                  &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,
 476                                                  &jy2,&jz2,&jx3,&jy3,&jz3);
 477             /*     #else                                */
 478             gmx_mm256_load_3rvec_8ptr_swizzle_ps(x+j_coord_offsetA+DIM,x+j_coord_offsetB+DIM,
 479                                                  x+j_coord_offsetC+DIM,x+j_coord_offsetD+DIM,
 480                                                  x+j_coord_offsetE+DIM,x+j_coord_offsetF+DIM,
 481                                                  x+j_coord_offsetG+DIM,x+j_coord_offsetH+DIM,
 482                                                  &jx1,&jy1,&jz1,&jx2,&jy2,&jz2,&jx3,&jy3,&jz3);
 483             /*     #endif                               */
 484             /* #endif                                   */
 485
 486             /* Calculate displacement vector */
 487             /* #for I,J in PAIRS_IJ */
 488             dx{I}{J}             = _mm256_sub_ps(ix{I},jx{J});
 489             dy{I}{J}             = _mm256_sub_ps(iy{I},jy{J});
 490             dz{I}{J}             = _mm256_sub_ps(iz{I},jz{J});
 491             /*     #define INNERFLOPS INNERFLOPS+3 */
 492             /* #endfor */
 493
 494             /* Calculate squared distance and things based on it */
 495             /* #for I,J in PAIRS_IJ */
 496             rsq{I}{J}            = gmx_mm256_calc_rsq_ps(dx{I}{J},dy{I}{J},dz{I}{J});
 497             /*     #define INNERFLOPS INNERFLOPS+5 */
 498             /* #endfor */
 499
 500             /* #for I,J in PAIRS_IJ */
 501             /*     #if 'rinv' in INTERACTION_FLAGS[I][J] */
 502             rinv{I}{J}           = gmx_mm256_invsqrt_ps(rsq{I}{J});
 503             /*         #define INNERFLOPS INNERFLOPS+5 */
 504             /*     #endif */
 505             /* #endfor */
 506
 507             /* #for I,J in PAIRS_IJ */
 508             /*     #if 'rinvsq' in INTERACTION_FLAGS[I][J] */
 509             /*         # if 'rinv' not in INTERACTION_FLAGS[I][J] */
 510             rinvsq{I}{J}         = gmx_mm256_inv_ps(rsq{I}{J});
 511             /*             #define INNERFLOPS INNERFLOPS+4 */
 512             /*         #else */
 513             rinvsq{I}{J}         = _mm256_mul_ps(rinv{I}{J},rinv{I}{J});
 514             /*             #define INNERFLOPS INNERFLOPS+1 */
 515             /*         #endif */
 516             /*     #endif */
 517             /* #endfor */
 518
 519             /* #if not 'Water' in GEOMETRY_J */
 520             /* Load parameters for j particles */
 521             /*     #for J in PARTICLES_ELEC_J */
 522             jq{J}              = gmx_mm256_load_8real_swizzle_ps(charge+jnrA+{J},charge+jnrB+{J},
 523                                                                  charge+jnrC+{J},charge+jnrD+{J},
 524                                                                  charge+jnrE+{J},charge+jnrF+{J},
 525                                                                  charge+jnrG+{J},charge+jnrH+{J});
 526             /*         #if KERNEL_ELEC=='GeneralizedBorn' */
 527             isaj{J}            = gmx_mm256_load_8real_swizzle_ps(invsqrta+jnrA+{J},invsqrta+jnrB+{J},
 528                                                                  invsqrta+jnrC+{J},invsqrta+jnrD+{J},
 529                                                                  invsqrta+jnrE+{J},invsqrta+jnrF+{J},
 530                                                                  invsqrta+jnrG+{J},invsqrta+jnrH+{J});
 531             /*         #endif */
 532             /*     #endfor */
 533             /*     #for J in PARTICLES_VDW_J */
 534             vdwjidx{J}A        = 2*vdwtype[jnrA+{J}];
 535             vdwjidx{J}B        = 2*vdwtype[jnrB+{J}];
 536             vdwjidx{J}C        = 2*vdwtype[jnrC+{J}];
 537             vdwjidx{J}D        = 2*vdwtype[jnrD+{J}];
 538             vdwjidx{J}E        = 2*vdwtype[jnrE+{J}];
 539             vdwjidx{J}F        = 2*vdwtype[jnrF+{J}];
 540             vdwjidx{J}G        = 2*vdwtype[jnrG+{J}];
 541             vdwjidx{J}H        = 2*vdwtype[jnrH+{J}];
 542             /*     #endfor */
 543             /* #endif */
 544
 545             /* #if 'Force' in KERNEL_VF and not 'Particle' in GEOMETRY_I */
 546             /*     #for J in PARTICLES_J */
 547             fjx{J}             = _mm256_setzero_ps();
 548             fjy{J}             = _mm256_setzero_ps();
 549             fjz{J}             = _mm256_setzero_ps();
 550             /*     #endfor */
 551             /* #endif */
 552
 553             /* #for I,J in PAIRS_IJ */
 554
 555             /**************************
 556              * CALCULATE INTERACTIONS *
 557              **************************/
 558
 559             /*     ## Note special check for TIP4P-TIP4P. Since we are cutting of all hydrogen interactions we also cut the LJ-only O-O interaction */
 560             /*     #if 'exactcutoff' in INTERACTION_FLAGS[I][J] or (GEOMETRY_I=='Water4' and GEOMETRY_J=='Water4' and 'exactcutoff' in INTERACTION_FLAGS[1][1]) */
 561             /*         ## We always calculate rinv/rinvsq above to enable pipelineing in compilers (performance tested on x86) */
 562             if (gmx_mm256_any_lt(rsq{I}{J},rcutoff2))
 563             {
 564                 /*     #if 0    ## this and the next two lines is a hack to maintain auto-indentation in template file */
 565             }
 566             /*         #endif */
 567             /*         #define INNERFLOPS INNERFLOPS+1 */
 568             /*     #endif */
 569
 570             /*     #if 'r' in INTERACTION_FLAGS[I][J] */
 571             r{I}{J}              = _mm256_mul_ps(rsq{I}{J},rinv{I}{J});
 572             /*         #if ROUND == 'Epilogue' */
 573             r{I}{J}              = _mm256_andnot_ps(dummy_mask,r{I}{J});
 574             /*             #define INNERFLOPS INNERFLOPS+1 */
 575             /*         #endif */
 576             /*         #define INNERFLOPS INNERFLOPS+1 */
 577             /*     #endif */
 578
 579             /*     ## For water geometries we already loaded parameters at the start of the kernel */
 580             /*     #if not 'Water' in GEOMETRY_J */
 581             /* Compute parameters for interactions between i and j atoms */
 582             /*         #if 'electrostatics' in INTERACTION_FLAGS[I][J] */
 583             qq{I}{J}             = _mm256_mul_ps(iq{I},jq{J});
 584             /*             #define INNERFLOPS INNERFLOPS+1 */
 585             /*         #endif */
 586             /*         #if 'vdw' in INTERACTION_FLAGS[I][J] */
 587             gmx_mm256_load_8pair_swizzle_ps(vdwioffsetptr{I}+vdwjidx{J}A,
 588                                             vdwioffsetptr{I}+vdwjidx{J}B,
 589                                             vdwioffsetptr{I}+vdwjidx{J}C,
 590                                             vdwioffsetptr{I}+vdwjidx{J}D,
 591                                             vdwioffsetptr{I}+vdwjidx{J}E,
 592                                             vdwioffsetptr{I}+vdwjidx{J}F,
 593                                             vdwioffsetptr{I}+vdwjidx{J}G,
 594                                             vdwioffsetptr{I}+vdwjidx{J}H,
 595                                             &c6_{I}{J},&c12_{I}{J});
 596             /*         #endif */
 597             /*     #endif */
 598
 599             /*     #if 'table' in INTERACTION_FLAGS[I][J] */
 600             /* Calculate table index by multiplying r with table scale and truncate to integer */
 601             rt               = _mm256_mul_ps(r{I}{J},vftabscale);
 602             vfitab           = _mm256_cvttps_epi32(rt);
 603             vfeps            = _mm256_sub_ps(rt,_mm256_round_ps(rt, _MM_FROUND_FLOOR));
 604             /*         #define INNERFLOPS INNERFLOPS+4                          */
 605             /*         AVX1 does not support 256-bit integer operations, so now we go to 128-bit mode... */
 606             vfitab_lo        = _mm256_extractf128_si256(vfitab,0x0);
 607             vfitab_hi        = _mm256_extractf128_si256(vfitab,0x1);
 608             /*         #if 'Table' in KERNEL_ELEC and 'Table' in KERNEL_VDW     */
 609             /*             ## 3 tables, 4 bytes per point: multiply index by 12 */
 610             vfitab_lo        = _mm_slli_epi32(_mm_add_epi32(vfitab_lo,_mm_slli_epi32(vfitab_lo,1)),2);
 611             vfitab_hi        = _mm_slli_epi32(_mm_add_epi32(vfitab_hi,_mm_slli_epi32(vfitab_hi,1)),2);
 612             /*         #elif 'Table' in KERNEL_ELEC                             */
 613             /*             ## 1 table, 4 bytes per point: multiply index by 4   */
 614             vfitab_lo        = _mm_slli_epi32(vfitab_lo,2);
 615             vfitab_hi        = _mm_slli_epi32(vfitab_hi,2);
 616             /*         #elif 'Table' in KERNEL_VDW                              */
 617             /*             ## 2 tables, 4 bytes per point: multiply index by 8  */
 618             vfitab_lo        = _mm_slli_epi32(vfitab_lo,3);
 619             vfitab_hi        = _mm_slli_epi32(vfitab_hi,3);
 620             /*         #endif                                                   */
 621             /*     #endif */
 622
 623             /*     ## ELECTROSTATIC INTERACTIONS */
 624             /*     #if 'electrostatics' in INTERACTION_FLAGS[I][J] */
 625
 626             /*         #if KERNEL_ELEC=='Coulomb' */
 627
 628             /* COULOMB ELECTROSTATICS */
 629             velec            = _mm256_mul_ps(qq{I}{J},rinv{I}{J});
 630             /*             #define INNERFLOPS INNERFLOPS+1 */
 631             /*             #if 'Force' in KERNEL_VF */
 632             felec            = _mm256_mul_ps(velec,rinvsq{I}{J});
 633             /*                 #define INNERFLOPS INNERFLOPS+1 */
 634             /*             #endif */
 635
 636             /*         #elif KERNEL_ELEC=='ReactionField' */
 637
 638             /* REACTION-FIELD ELECTROSTATICS */
 639             /*             #if 'Potential' in KERNEL_VF */
 640             velec            = _mm256_mul_ps(qq{I}{J},_mm256_sub_ps(_mm256_add_ps(rinv{I}{J},_mm256_mul_ps(krf,rsq{I}{J})),crf));
 641             /*                 #define INNERFLOPS INNERFLOPS+4 */
 642             /*             #endif */
 643             /*             #if 'Force' in KERNEL_VF */
 644             felec            = _mm256_mul_ps(qq{I}{J},_mm256_sub_ps(_mm256_mul_ps(rinv{I}{J},rinvsq{I}{J}),krf2));
 645             /*                 #define INNERFLOPS INNERFLOPS+3 */
 646             /*             #endif */
 647
 648             /*         #elif KERNEL_ELEC=='GeneralizedBorn' */
 649
 650             /* GENERALIZED BORN AND COULOMB ELECTROSTATICS */
 651             isaprod          = _mm256_mul_ps(isai{I},isaj{J});
 652             gbqqfactor       = _mm256_xor_ps(signbit,_mm256_mul_ps(qq{I}{J},_mm256_mul_ps(isaprod,gbinvepsdiff)));
 653             gbscale          = _mm256_mul_ps(isaprod,gbtabscale);
 654             /*             #define INNERFLOPS INNERFLOPS+5 */
 655
 656             /* Calculate generalized born table index - this is a separate table from the normal one,
 657              * but we use the same procedure by multiplying r with scale and truncating to integer.
 658              */
 659             rt               = _mm256_mul_ps(r{I}{J},gbscale);
 660             gbitab           = _mm256_cvttps_epi32(rt);
 661             gbeps            = _mm256_sub_ps(rt,_mm256_round_ps(rt, _MM_FROUND_FLOOR));
 662             /*         AVX1 does not support 256-bit integer operations, so now we go to 128-bit mode... */
 663             gbitab_lo        = _mm256_extractf128_si256(gbitab,0x0);
 664             gbitab_hi        = _mm256_extractf128_si256(gbitab,0x1);
 665             gbitab_lo        = _mm_slli_epi32(gbitab_lo,2);
 666             gbitab_hi        = _mm_slli_epi32(gbitab_hi,2);
 667             Y                = gmx_mm256_set_m128(_mm_load_ps(gbtab + _mm_extract_epi32(gbitab_hi,0)),
 668                                                   _mm_load_ps(gbtab + _mm_extract_epi32(gbitab_lo,0)));
 669             F                = gmx_mm256_set_m128(_mm_load_ps(gbtab + _mm_extract_epi32(gbitab_hi,1)),
 670                                                   _mm_load_ps(gbtab + _mm_extract_epi32(gbitab_lo,1)));
 671             G                = gmx_mm256_set_m128(_mm_load_ps(gbtab + _mm_extract_epi32(gbitab_hi,2)),
 672                                                   _mm_load_ps(gbtab + _mm_extract_epi32(gbitab_lo,2)));
 673             H                = gmx_mm256_set_m128(_mm_load_ps(gbtab + _mm_extract_epi32(gbitab_hi,3)),
 674                                                   _mm_load_ps(gbtab + _mm_extract_epi32(gbitab_lo,3)));
 675             GMX_MM256_HALFTRANSPOSE4_PS(Y,F,G,H);
 676             Heps             = _mm256_mul_ps(gbeps,H);
 677             Fp               = _mm256_add_ps(F,_mm256_mul_ps(gbeps,_mm256_add_ps(G,Heps)));
 678             VV               = _mm256_add_ps(Y,_mm256_mul_ps(gbeps,Fp));
 679             vgb              = _mm256_mul_ps(gbqqfactor,VV);
 680             /*             #define INNERFLOPS INNERFLOPS+10 */
 681
 682             /*             #if 'Force' in KERNEL_VF */
 683             FF               = _mm256_add_ps(Fp,_mm256_mul_ps(gbeps,_mm256_add_ps(G,_mm256_add_ps(Heps,Heps))));
 684             fgb              = _mm256_mul_ps(gbqqfactor,_mm256_mul_ps(FF,gbscale));
 685             dvdatmp          = _mm256_mul_ps(minushalf,_mm256_add_ps(vgb,_mm256_mul_ps(fgb,r{I}{J})));
 686             /*                 #if ROUND == 'Epilogue' */
 687             dvdatmp          = _mm256_andnot_ps(dummy_mask,dvdatmp);
 688             /*                 #endif */
 689             dvdasum          = _mm256_add_ps(dvdasum,dvdatmp);
 690             /*                 #if ROUND == 'Loop' */
 691             fjptrA           = dvda+jnrA;
 692             fjptrB           = dvda+jnrB;
 693             fjptrC           = dvda+jnrC;
 694             fjptrD           = dvda+jnrD;
 695             fjptrE           = dvda+jnrE;
 696             fjptrF           = dvda+jnrF;
 697             fjptrG           = dvda+jnrG;
 698             fjptrH           = dvda+jnrH;
 699             /*                 #else */
 700             /* The pointers to scratch make sure that this code with compilers that take gmx_restrict seriously (e.g. icc 13) really can't screw things up. */
 701             fjptrA             = (jnrlistA>=0) ? dvda+jnrA : scratch;
 702             fjptrB             = (jnrlistB>=0) ? dvda+jnrB : scratch;
 703             fjptrC             = (jnrlistC>=0) ? dvda+jnrC : scratch;
 704             fjptrD             = (jnrlistD>=0) ? dvda+jnrD : scratch;
 705             fjptrE             = (jnrlistE>=0) ? dvda+jnrE : scratch;
 706             fjptrF             = (jnrlistF>=0) ? dvda+jnrF : scratch;
 707             fjptrG             = (jnrlistG>=0) ? dvda+jnrG : scratch;
 708             fjptrH             = (jnrlistH>=0) ? dvda+jnrH : scratch;
 709             /*                 #endif */
 710             gmx_mm256_increment_8real_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,
 711                                                  _mm256_mul_ps(dvdatmp,_mm256_mul_ps(isaj{J},isaj{J})));
 712             /*                 #define INNERFLOPS INNERFLOPS+12 */
 713             /*             #endif */
 714             velec            = _mm256_mul_ps(qq{I}{J},rinv{I}{J});
 715             /*                 #define INNERFLOPS INNERFLOPS+1 */
 716             /*             #if 'Force' in KERNEL_VF */
 717             felec            = _mm256_mul_ps(_mm256_sub_ps(_mm256_mul_ps(velec,rinv{I}{J}),fgb),rinv{I}{J});
 718             /*                 #define INNERFLOPS INNERFLOPS+3 */
 719             /*             #endif */
 720
 721             /*         #elif KERNEL_ELEC=='Ewald' */
 722             /* EWALD ELECTROSTATICS */
 723
 724             /* Analytical PME correction */
 725             zeta2            = _mm256_mul_ps(beta2,rsq{I}{J});
 726             /*             #if 'Force' in KERNEL_VF */
 727             rinv3            = _mm256_mul_ps(rinvsq{I}{J},rinv{I}{J});
 728             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
 729             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 730             felec            = _mm256_mul_ps(qq{I}{J},felec);
 731             /*                 #define INNERFLOPS INNERFLOPS+31 */
 732             /*             #endif */
 733             /*             #if 'Potential' in KERNEL_VF or KERNEL_MOD_ELEC=='PotentialSwitch' */
 734             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
 735             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 736             /*                 #define INNERFLOPS INNERFLOPS+27 */
 737             /*                 #if KERNEL_MOD_ELEC=='PotentialShift' */
 738             velec            = _mm256_sub_ps(_mm256_sub_ps(rinv{I}{J},sh_ewald),pmecorrV);
 739             /*                     #define INNERFLOPS INNERFLOPS+21 */
 740             /*                 #else */
 741             velec            = _mm256_sub_ps(rinv{I}{J},pmecorrV);
 742             /*                 #endif */
 743             velec            = _mm256_mul_ps(qq{I}{J},velec);
 744             /*             #endif */
 745
 746             /*         #elif KERNEL_ELEC=='CubicSplineTable' */
 747
 748             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 749             Y                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,0)),
 750                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,0)));
 751             F                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,1)),
 752                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,1)));
 753             G                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,2)),
 754                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,2)));
 755             H                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,3)),
 756                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,3)));
 757             GMX_MM256_HALFTRANSPOSE4_PS(Y,F,G,H);
 758             Heps             = _mm256_mul_ps(vfeps,H);
 759             Fp               = _mm256_add_ps(F,_mm256_mul_ps(vfeps,_mm256_add_ps(G,Heps)));
 760             /*             #define INNERFLOPS INNERFLOPS+4 */
 761             /*             #if 'Potential' in KERNEL_VF */
 762             VV               = _mm256_add_ps(Y,_mm256_mul_ps(vfeps,Fp));
 763             velec            = _mm256_mul_ps(qq{I}{J},VV);
 764             /*                 #define INNERFLOPS INNERFLOPS+3 */
 765             /*             #endif */
 766             /*             #if 'Force' in KERNEL_VF */
 767             FF               = _mm256_add_ps(Fp,_mm256_mul_ps(vfeps,_mm256_add_ps(G,_mm256_add_ps(Heps,Heps))));
 768             felec            = _mm256_xor_ps(signbit,_mm256_mul_ps(_mm256_mul_ps(qq{I}{J},FF),_mm256_mul_ps(vftabscale,rinv{I}{J})));
 769             /*                 #define INNERFLOPS INNERFLOPS+7 */
 770             /*             #endif */
 771             /*         #endif */
 772             /*         ## End of check for electrostatics interaction forms */
 773             /*     #endif */
 774             /*     ## END OF ELECTROSTATIC INTERACTION CHECK FOR PAIR I-J */
 775
 776             /*     #if 'vdw' in INTERACTION_FLAGS[I][J] */
 777
 778             /*         #if KERNEL_VDW=='LennardJones' */
 779
 780             /* LENNARD-JONES DISPERSION/REPULSION */
 781
 782             rinvsix          = _mm256_mul_ps(_mm256_mul_ps(rinvsq{I}{J},rinvsq{I}{J}),rinvsq{I}{J});
 783             /*             #define INNERFLOPS INNERFLOPS+2 */
 784             /*             #if 'Potential' in KERNEL_VF or KERNEL_MOD_VDW=='PotentialSwitch' */
 785             vvdw6            = _mm256_mul_ps(c6_{I}{J},rinvsix);
 786             vvdw12           = _mm256_mul_ps(c12_{I}{J},_mm256_mul_ps(rinvsix,rinvsix));
 787             /*                 #define INNERFLOPS INNERFLOPS+3 */
 788             /*                 #if KERNEL_MOD_VDW=='PotentialShift' */
 789             vvdw             = _mm256_sub_ps(_mm256_mul_ps( _mm256_sub_ps(vvdw12 , _mm256_mul_ps(c12_{I}{J},_mm256_mul_ps(sh_vdw_invrcut6,sh_vdw_invrcut6))), one_twelfth) ,
 790                                           _mm256_mul_ps( _mm256_sub_ps(vvdw6,_mm256_mul_ps(c6_{I}{J},sh_vdw_invrcut6)),one_sixth));
 791             /*                     #define INNERFLOPS INNERFLOPS+8 */
 792             /*                 #else */
 793             vvdw             = _mm256_sub_ps( _mm256_mul_ps(vvdw12,one_twelfth) , _mm256_mul_ps(vvdw6,one_sixth) );
 794             /*                     #define INNERFLOPS INNERFLOPS+3 */
 795             /*                 #endif */
 796             /*                 ## Check for force inside potential check, i.e. this means we already did the potential part */
 797             /*                 #if 'Force' in KERNEL_VF */
 798             fvdw             = _mm256_mul_ps(_mm256_sub_ps(vvdw12,vvdw6),rinvsq{I}{J});
 799             /*                     #define INNERFLOPS INNERFLOPS+2 */
 800             /*                 #endif */
 801             /*             #elif KERNEL_VF=='Force' */
 802             /*                 ## Force-only LennardJones makes it possible to save 1 flop (they do add up...) */
 803             fvdw             = _mm256_mul_ps(_mm256_sub_ps(_mm256_mul_ps(c12_{I}{J},rinvsix),c6_{I}{J}),_mm256_mul_ps(rinvsix,rinvsq{I}{J}));
 804             /*                 #define INNERFLOPS INNERFLOPS+4 */
 805             /*             #endif */
 806
 807             /*         #elif KERNEL_VDW=='CubicSplineTable' */
 808
 809             /* CUBIC SPLINE TABLE DISPERSION */
 810             /*             #if 'Table' in KERNEL_ELEC */
 811             vfitab_lo        = _mm_add_epi32(vfitab_lo,ifour);
 812             vfitab_hi        = _mm_add_epi32(vfitab_hi,ifour);
 813             /*             #endif                     */
 814             Y                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,0)),
 815                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,0)));
 816             F                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,1)),
 817                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,1)));
 818             G                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,2)),
 819                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,2)));
 820             H                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,3)),
 821                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,3)));
 822             GMX_MM256_HALFTRANSPOSE4_PS(Y,F,G,H);
 823             Heps             = _mm256_mul_ps(vfeps,H);
 824             Fp               = _mm256_add_ps(F,_mm256_mul_ps(vfeps,_mm256_add_ps(G,Heps)));
 825             /*             #define INNERFLOPS INNERFLOPS+4 */
 826             /*             #if 'Potential' in KERNEL_VF */
 827             VV               = _mm256_add_ps(Y,_mm256_mul_ps(vfeps,Fp));
 828             vvdw6            = _mm256_mul_ps(c6_{I}{J},VV);
 829             /*                 #define INNERFLOPS INNERFLOPS+3 */
 830             /*             #endif */
 831             /*             #if 'Force' in KERNEL_VF */
 832             FF               = _mm256_add_ps(Fp,_mm256_mul_ps(vfeps,_mm256_add_ps(G,_mm256_add_ps(Heps,Heps))));
 833             fvdw6            = _mm256_mul_ps(c6_{I}{J},FF);
 834             /*                 #define INNERFLOPS INNERFLOPS+4 */
 835             /*             #endif */
 836
 837             /* CUBIC SPLINE TABLE REPULSION */
 838             vfitab_lo        = _mm_add_epi32(vfitab_lo,ifour);
 839             vfitab_hi        = _mm_add_epi32(vfitab_hi,ifour);
 840             Y                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,0)),
 841                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,0)));
 842             F                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,1)),
 843                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,1)));
 844             G                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,2)),
 845                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,2)));
 846             H                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,3)),
 847                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,3)));
 848             GMX_MM256_HALFTRANSPOSE4_PS(Y,F,G,H);
 849             Heps             = _mm256_mul_ps(vfeps,H);
 850             Fp               = _mm256_add_ps(F,_mm256_mul_ps(vfeps,_mm256_add_ps(G,Heps)));
 851             /*             #define INNERFLOPS INNERFLOPS+4 */
 852             /*             #if 'Potential' in KERNEL_VF */
 853             VV               = _mm256_add_ps(Y,_mm256_mul_ps(vfeps,Fp));
 854             vvdw12           = _mm256_mul_ps(c12_{I}{J},VV);
 855             /*                 #define INNERFLOPS INNERFLOPS+3 */
 856             /*             #endif */
 857             /*             #if 'Force' in KERNEL_VF */
 858             FF               = _mm256_add_ps(Fp,_mm256_mul_ps(vfeps,_mm256_add_ps(G,_mm256_add_ps(Heps,Heps))));
 859             fvdw12           = _mm256_mul_ps(c12_{I}{J},FF);
 860             /*                 #define INNERFLOPS INNERFLOPS+5 */
 861             /*             #endif */
 862             /*             #if 'Potential' in KERNEL_VF */
 863             vvdw             = _mm256_add_ps(vvdw12,vvdw6);
 864             /*                 #define INNERFLOPS INNERFLOPS+1 */
 865             /*             #endif */
 866             /*             #if 'Force' in KERNEL_VF */
 867             fvdw             = _mm256_xor_ps(signbit,_mm256_mul_ps(_mm256_add_ps(fvdw6,fvdw12),_mm256_mul_ps(vftabscale,rinv{I}{J})));
 868             /*                 #define INNERFLOPS INNERFLOPS+4 */
 869             /*             #endif */
 870             /*         #endif */
 871             /*         ## End of check for vdw interaction forms */
 872             /*     #endif */
 873             /*     ## END OF VDW INTERACTION CHECK FOR PAIR I-J */
 874
 875             /*     #if 'switch' in INTERACTION_FLAGS[I][J] */
 876             d                = _mm256_sub_ps(r{I}{J},rswitch);
 877             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 878             d2               = _mm256_mul_ps(d,d);
 879             sw               = _mm256_add_ps(one,_mm256_mul_ps(d2,_mm256_mul_ps(d,_mm256_add_ps(swV3,_mm256_mul_ps(d,_mm256_add_ps(swV4,_mm256_mul_ps(d,swV5)))))));
 880             /*         #define INNERFLOPS INNERFLOPS+10 */
 881
 882             /*         #if 'Force' in KERNEL_VF */
 883             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 884             /*             #define INNERFLOPS INNERFLOPS+5 */
 885             /*         #endif */
 886
 887             /* Evaluate switch function */
 888             /*         #if 'Force' in KERNEL_VF */
 889             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 890             /*             #if 'electrostatics' in INTERACTION_FLAGS[I][J] and KERNEL_MOD_ELEC=='PotentialSwitch' */
 891             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv{I}{J},_mm256_mul_ps(velec,dsw)) );
 892             /*                 #define INNERFLOPS INNERFLOPS+4 */
 893             /*             #endif */
 894             /*             #if 'vdw' in INTERACTION_FLAGS[I][J] and KERNEL_MOD_VDW=='PotentialSwitch' */
 895             fvdw             = _mm256_sub_ps( _mm256_mul_ps(fvdw,sw) , _mm256_mul_ps(rinv{I}{J},_mm256_mul_ps(vvdw,dsw)) );
 896             /*                 #define INNERFLOPS INNERFLOPS+4 */
 897             /*             #endif */
 898             /*         #endif */
 899             /*         #if 'Potential' in KERNEL_VF */
 900             /*             #if 'electrostatics' in INTERACTION_FLAGS[I][J] and KERNEL_MOD_ELEC=='PotentialSwitch' */
 901             velec            = _mm256_mul_ps(velec,sw);
 902             /*                 #define INNERFLOPS INNERFLOPS+1 */
 903             /*             #endif */
 904             /*             #if 'vdw' in INTERACTION_FLAGS[I][J] and KERNEL_MOD_VDW=='PotentialSwitch' */
 905             vvdw             = _mm256_mul_ps(vvdw,sw);
 906             /*                 #define INNERFLOPS INNERFLOPS+1 */
 907             /*             #endif */
 908             /*         #endif */
 909             /*     #endif */
 910             /*     ## Note special check for TIP4P-TIP4P. Since we are cutting of all hydrogen interactions we also cut the LJ-only O-O interaction */
 911             /*     #if 'exactcutoff' in INTERACTION_FLAGS[I][J] or (GEOMETRY_I=='Water4' and GEOMETRY_J=='Water4' and 'exactcutoff' in INTERACTION_FLAGS[1][1]) */
 912             cutoff_mask      = _mm256_cmp_ps(rsq{I}{J},rcutoff2,_CMP_LT_OQ);
 913             /*         #define INNERFLOPS INNERFLOPS+1 */
 914             /*     #endif */
 915
 916             /*     #if 'Potential' in KERNEL_VF */
 917             /* Update potential sum for this i atom from the interaction with this j atom. */
 918             /*         #if 'electrostatics' in INTERACTION_FLAGS[I][J] */
 919             /*             #if 'exactcutoff' in INTERACTION_FLAGS[I][J] */
 920             velec            = _mm256_and_ps(velec,cutoff_mask);
 921             /*                 #define INNERFLOPS INNERFLOPS+1 */
 922             /*             #endif                                       */
 923             /*             #if ROUND == 'Epilogue' */
 924             velec            = _mm256_andnot_ps(dummy_mask,velec);
 925             /*             #endif */
 926             velecsum         = _mm256_add_ps(velecsum,velec);
 927             /*             #define INNERFLOPS INNERFLOPS+1 */
 928             /*             #if KERNEL_ELEC=='GeneralizedBorn' */
 929             /*             #if 'exactcutoff' in INTERACTION_FLAGS[I][J] */
 930             vgb              = _mm256_and_ps(vgb,cutoff_mask);
 931             /*                 #define INNERFLOPS INNERFLOPS+1 */
 932             /*             #endif                                       */
 933             /*             #if ROUND == 'Epilogue' */
 934             vgb              = _mm256_andnot_ps(dummy_mask,vgb);
 935             /*             #endif */
 936             vgbsum           = _mm256_add_ps(vgbsum,vgb);
 937             /*                 #define INNERFLOPS INNERFLOPS+1 */
 938             /*             #endif */
 939             /*         #endif */
 940             /*         #if 'vdw' in INTERACTION_FLAGS[I][J] */
 941             /*     ## Note special check for TIP4P-TIP4P. Since we are cutting of all hydrogen interactions we also cut the LJ-only O-O interaction */
 942             /*     #if 'exactcutoff' in INTERACTION_FLAGS[I][J] or (GEOMETRY_I=='Water4' and GEOMETRY_J=='Water4' and 'exactcutoff' in INTERACTION_FLAGS[1][1]) */
 943             vvdw             = _mm256_and_ps(vvdw,cutoff_mask);
 944             /*                 #define INNERFLOPS INNERFLOPS+1 */
 945             /*             #endif                                       */
 946             /*             #if ROUND == 'Epilogue' */
 947             vvdw             = _mm256_andnot_ps(dummy_mask,vvdw);
 948             /*             #endif */
 949             vvdwsum          = _mm256_add_ps(vvdwsum,vvdw);
 950             /*             #define INNERFLOPS INNERFLOPS+1 */
 951             /*         #endif */
 952             /*     #endif */
 953
 954             /*     #if 'Force' in KERNEL_VF */
 955
 956             /*         #if 'electrostatics' in INTERACTION_FLAGS[I][J] and 'vdw' in INTERACTION_FLAGS[I][J] */
 957             fscal            = _mm256_add_ps(felec,fvdw);
 958             /*             #define INNERFLOPS INNERFLOPS+1 */
 959             /*         #elif 'electrostatics' in INTERACTION_FLAGS[I][J] */
 960             fscal            = felec;
 961             /*         #elif 'vdw' in INTERACTION_FLAGS[I][J] */
 962             fscal            = fvdw;
 963             /*        #endif */
 964
 965             /*     ## Note special check for TIP4P-TIP4P. Since we are cutting of all hydrogen interactions we also cut the LJ-only O-O interaction */
 966             /*     #if 'exactcutoff' in INTERACTION_FLAGS[I][J] or (GEOMETRY_I=='Water4' and GEOMETRY_J=='Water4' and 'exactcutoff' in INTERACTION_FLAGS[1][1]) */
 967             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 968             /*                 #define INNERFLOPS INNERFLOPS+1 */
 969             /*             #endif                                       */
 970
 971             /*             #if ROUND == 'Epilogue' */
 972             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
 973             /*             #endif */
 974
 975             /* Calculate temporary vectorial force */
 976             tx               = _mm256_mul_ps(fscal,dx{I}{J});
 977             ty               = _mm256_mul_ps(fscal,dy{I}{J});
 978             tz               = _mm256_mul_ps(fscal,dz{I}{J});
 979
 980             /* Update vectorial force */
 981             fix{I}             = _mm256_add_ps(fix{I},tx);
 982             fiy{I}             = _mm256_add_ps(fiy{I},ty);
 983             fiz{I}             = _mm256_add_ps(fiz{I},tz);
 984             /*             #define INNERFLOPS INNERFLOPS+6 */
 985
 986             /* #if GEOMETRY_I == 'Particle'             */
 987             /*     #if ROUND == 'Loop' */
 988             fjptrA             = f+j_coord_offsetA;
 989             fjptrB             = f+j_coord_offsetB;
 990             fjptrC             = f+j_coord_offsetC;
 991             fjptrD             = f+j_coord_offsetD;
 992             fjptrE             = f+j_coord_offsetE;
 993             fjptrF             = f+j_coord_offsetF;
 994             fjptrG             = f+j_coord_offsetG;
 995             fjptrH             = f+j_coord_offsetH;
 996             /*     #else */
 997             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
 998             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
 999             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
1000             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
1001             fjptrE             = (jnrlistE>=0) ? f+j_coord_offsetE : scratch;
1002             fjptrF             = (jnrlistF>=0) ? f+j_coord_offsetF : scratch;
1003             fjptrG             = (jnrlistG>=0) ? f+j_coord_offsetG : scratch;
1004             fjptrH             = (jnrlistH>=0) ? f+j_coord_offsetH : scratch;
1005             /*     #endif */
1006             gmx_mm256_decrement_1rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,tx,ty,tz);
1007             /*     #define INNERFLOPS INNERFLOPS+3      */
1008             /* #else                                    */
1009             fjx{J}             = _mm256_add_ps(fjx{J},tx);
1010             fjy{J}             = _mm256_add_ps(fjy{J},ty);
1011             fjz{J}             = _mm256_add_ps(fjz{J},tz);
1012             /*     #define INNERFLOPS INNERFLOPS+3      */
1013             /* #endif                                   */
1014
1015             /*     #endif */
1016
1017             /*     ## Note special check for TIP4P-TIP4P. Since we are cutting of all hydrogen interactions we also cut the LJ-only O-O interaction */
1018             /*     #if 'exactcutoff' in INTERACTION_FLAGS[I][J] or (GEOMETRY_I=='Water4' and GEOMETRY_J=='Water4' and 'exactcutoff' in INTERACTION_FLAGS[1][1]) */
1019             /*         #if 0    ## This and next two lines is a hack to maintain indentation in template file */
1020             {
1021                 /*     #endif */
1022             }
1023             /*     #endif */
1024             /*    ## End of check for the interaction being outside the cutoff */
1025
1026             /* #endfor */
1027             /* ## End of loop over i-j interaction pairs */
1028
1029             /* #if GEOMETRY_I != 'Particle' */
1030             /*     #if ROUND == 'Loop' */
1031             fjptrA             = f+j_coord_offsetA;
1032             fjptrB             = f+j_coord_offsetB;
1033             fjptrC             = f+j_coord_offsetC;
1034             fjptrD             = f+j_coord_offsetD;
1035             fjptrE             = f+j_coord_offsetE;
1036             fjptrF             = f+j_coord_offsetF;
1037             fjptrG             = f+j_coord_offsetG;
1038             fjptrH             = f+j_coord_offsetH;
1039             /*     #else */
1040             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
1041             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
1042             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
1043             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
1044             fjptrE             = (jnrlistE>=0) ? f+j_coord_offsetE : scratch;
1045             fjptrF             = (jnrlistF>=0) ? f+j_coord_offsetF : scratch;
1046             fjptrG             = (jnrlistG>=0) ? f+j_coord_offsetG : scratch;
1047             fjptrH             = (jnrlistH>=0) ? f+j_coord_offsetH : scratch;
1048             /*     #endif */
1049             /* #endif */
1050
1051             /* #if 'Water' in GEOMETRY_I and GEOMETRY_J == 'Particle' */
1052             gmx_mm256_decrement_1rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,fjx0,fjy0,fjz0);
1053             /*     #define INNERFLOPS INNERFLOPS+3      */
1054             /* #elif GEOMETRY_J == 'Water3'             */
1055             gmx_mm256_decrement_3rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,
1056                                                       fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2);
1057             /*     #define INNERFLOPS INNERFLOPS+9      */
1058             /* #elif GEOMETRY_J == 'Water4'             */
1059             /*     #if 0 in PARTICLES_J                 */
1060             gmx_mm256_decrement_4rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,
1061                                                       fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,
1062                                                       fjx2,fjy2,fjz2,fjx3,fjy3,fjz3);
1063             /*     #define INNERFLOPS INNERFLOPS+12     */
1064             /*     #else                                */
1065             gmx_mm256_decrement_3rvec_8ptr_swizzle_ps(fjptrA+DIM,fjptrB+DIM,fjptrC+DIM,fjptrD+DIM,
1066                                                       fjptrE+DIM,fjptrF+DIM,fjptrG+DIM,fjptrH+DIM,
1067                                                       fjx1,fjy1,fjz1,fjx2,fjy2,fjz2,fjx3,fjy3,fjz3);
1068             /*     #define INNERFLOPS INNERFLOPS+9      */
1069             /*     #endif                               */
1070             /* #endif                                   */
1071
1072             /* Inner loop uses {INNERFLOPS} flops */
1073         }
1074
1075         /* #endfor */
1076
1077         /* End of innermost loop */
1078
1079         /* #if 'Force' in KERNEL_VF */
1080         /*     #if GEOMETRY_I == 'Particle'            */
1081         gmx_mm256_update_iforce_1atom_swizzle_ps(fix0,fiy0,fiz0,
1082                                                  f+i_coord_offset,fshift+i_shift_offset);
1083         /*         #define OUTERFLOPS OUTERFLOPS+6     */
1084         /*     #elif GEOMETRY_I == 'Water3'            */
1085         gmx_mm256_update_iforce_3atom_swizzle_ps(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,
1086                                                  f+i_coord_offset,fshift+i_shift_offset);
1087         /*         #define OUTERFLOPS OUTERFLOPS+18    */
1088         /*     #elif GEOMETRY_I == 'Water4'            */
1089         /*         #if 0 in PARTICLES_I                */
1090         gmx_mm256_update_iforce_4atom_swizzle_ps(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
1091                                                  f+i_coord_offset,fshift+i_shift_offset);
1092         /*             #define OUTERFLOPS OUTERFLOPS+24    */
1093         /*         #else                               */
1094         gmx_mm256_update_iforce_3atom_swizzle_ps(fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
1095                                                  f+i_coord_offset+DIM,fshift+i_shift_offset);
1096         /*             #define OUTERFLOPS OUTERFLOPS+18    */
1097         /*         #endif                              */
1098         /*     #endif                                  */
1099         /* #endif                                      */
1100
1101         /* #if 'Potential' in KERNEL_VF */
1102         ggid                        = gid[iidx];
1103         /* Update potential energies */
1104         /*     #if KERNEL_ELEC != 'None' */
1105         gmx_mm256_update_1pot_ps(velecsum,kernel_data->energygrp_elec+ggid);
1106         /*         #define OUTERFLOPS OUTERFLOPS+1 */
1107         /*     #endif */
1108         /*     #if 'GeneralizedBorn' in KERNEL_ELEC */
1109         gmx_mm256_update_1pot_ps(vgbsum,kernel_data->energygrp_polarization+ggid);
1110         /*         #define OUTERFLOPS OUTERFLOPS+1 */
1111         /*     #endif */
1112         /*     #if KERNEL_VDW != 'None' */
1113         gmx_mm256_update_1pot_ps(vvdwsum,kernel_data->energygrp_vdw+ggid);
1114         /*         #define OUTERFLOPS OUTERFLOPS+1 */
1115         /*     #endif */
1116         /* #endif */
1117         /*     #if 'GeneralizedBorn' in KERNEL_ELEC and 'Force' in KERNEL_VF */
1118         dvdasum = _mm256_mul_ps(dvdasum, _mm256_mul_ps(isai{I},isai{I}));
1119         gmx_mm256_update_1pot_ps(dvdasum,dvda+inr);
1120         /*     #endif */
1121
1122         /* Increment number of inner iterations */
1123         inneriter                  += j_index_end - j_index_start;
1124
1125         /* Outer loop uses {OUTERFLOPS} flops */
1126     }
1127
1128     /* Increment number of outer iterations */
1129     outeriter        += nri;
1130
1131     /* Update outer/inner flops */
1132     /* ## NB: This is not important, it just affects the flopcount. However, since our preprocessor is */
1133     /* ## primitive and replaces aggressively even in strings inside these directives, we need to      */
1134     /* ## assemble the main part of the name (containing KERNEL/ELEC/VDW) directly in the source.      */
1135     /* #if GEOMETRY_I == 'Water3'            */
1136     /*     #define ISUFFIX '_W3'             */
1137     /* #elif GEOMETRY_I == 'Water4'          */
1138     /*     #define ISUFFIX '_W4'             */
1139     /* #else                                 */
1140     /*     #define ISUFFIX ''                */
1141     /* #endif                                */
1142     /* #if GEOMETRY_J == 'Water3'            */
1143     /*     #define JSUFFIX 'W3'              */
1144     /* #elif GEOMETRY_J == 'Water4'          */
1145     /*     #define JSUFFIX 'W4'              */
1146     /* #else                                 */
1147     /*     #define JSUFFIX ''                */
1148     /* #endif                                */
1149     /* #if 'PotentialAndForce' in KERNEL_VF  */
1150     /*     #define VFSUFFIX  '_VF'           */
1151     /* #elif 'Potential' in KERNEL_VF        */
1152     /*     #define VFSUFFIX '_V'             */
1153     /* #else                                 */
1154     /*     #define VFSUFFIX '_F'             */
1155     /* #endif                                */
1156
1157     /* #if KERNEL_ELEC != 'None' and KERNEL_VDW != 'None' */
1158     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW{ISUFFIX}{JSUFFIX}{VFSUFFIX},outeriter*{OUTERFLOPS} + inneriter*{INNERFLOPS});
1159     /* #elif KERNEL_ELEC != 'None' */
1160     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC{ISUFFIX}{JSUFFIX}{VFSUFFIX},outeriter*{OUTERFLOPS} + inneriter*{INNERFLOPS});
1161     /* #else */
1162     inc_nrnb(nrnb,eNR_NBKERNEL_VDW{ISUFFIX}{JSUFFIX}{VFSUFFIX},outeriter*{OUTERFLOPS} + inneriter*{INNERFLOPS});
1163     /* #endif  */
1164 }