src/gromacs/gmxlib/nonbonded/nb_kernel_sse4_1_single/nb_kernel_template_sse4_1_single.pre

   1 /* #if 0 */
   2 /*
   3  * This file is part of the GROMACS molecular simulation package.
   4  *
   5  * Copyright (c) 2012,2013,2014, 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 #include "config.h"
  40
  41 #include <math.h>
  42
  43 #include "../nb_kernel.h"
  44 #include "gromacs/legacyheaders/types/simple.h"
  45 #include "gromacs/math/vec.h"
  46 #include "gromacs/legacyheaders/nrnb.h"
  47
  48 #include "gromacs/simd/math_x86_sse4_1_single.h"
  49 #include "kernelutil_x86_sse4_1_single.h"
  50 /* #endif */
  51
  52 /* ## List of variables set by the generating script:                                    */
  53 /* ##                                                                                    */
  54 /* ## Setttings that apply to the entire kernel:                                         */
  55 /* ## KERNEL_ELEC:           String, choice for electrostatic interactions               */
  56 /* ## KERNEL_VDW:            String, choice for van der Waals interactions               */
  57 /* ## KERNEL_NAME:           String, name of this kernel                                 */
  58 /* ## KERNEL_VF:             String telling if we calculate potential, force, or both    */
  59 /* ## GEOMETRY_I/GEOMETRY_J: String, name of each geometry, e.g. 'Water3' or '1Particle' */
  60 /* ##                                                                                    */
  61 /* ## Setttings that apply to particles in the outer (I) or inner (J) loops:             */
  62 /* ## PARTICLES_I[]/         Arrays with lists of i/j particles to use in kernel. It is  */
  63 /* ## PARTICLES_J[]:         just [0] for particle geometry, but can be longer for water */
  64 /* ## PARTICLES_ELEC_I[]/    Arrays with lists of i/j particle that have electrostatics  */
  65 /* ## PARTICLES_ELEC_J[]:    interactions that should be calculated in this kernel.      */
  66 /* ## PARTICLES_VDW_I[]/     Arrays with the list of i/j particle that have VdW          */
  67 /* ## PARTICLES_VDW_J[]:     interactions that should be calculated in this kernel.      */
  68 /* ##                                                                                    */
  69 /* ## Setttings for pairs of interactions (e.g. 2nd i particle against 1st j particle)   */
  70 /* ## PAIRS_IJ[]:            Array with (i,j) tuples of pairs for which interactions     */
  71 /* ##                        should be calculated in this kernel. Zero-charge particles  */
  72 /* ##                        do not have interactions with particles without vdw, and    */
  73 /* ##                        Vdw-only interactions are not evaluated in a no-vdw-kernel. */
  74 /* ## INTERACTION_FLAGS[][]: 2D matrix, dimension e.g. 3*3 for water-water interactions. */
  75 /* ##                        For each i-j pair, the element [I][J] is a list of strings  */
  76 /* ##                        defining properties/flags of this interaction. Examples     */
  77 /* ##                        include 'electrostatics'/'vdw' if that type of interaction  */
  78 /* ##                        should be evaluated, 'rsq'/'rinv'/'rinvsq' if those values  */
  79 /* ##                        are needed, and 'exactcutoff' or 'shift','switch' to        */
  80 /* ##                        decide if the force/potential should be modified. This way  */
  81 /* ##                        we only calculate values absolutely needed for each case.   */
  82
  83 /* ## Calculate the size and offset for (merged/interleaved) table data */
  84
  85 /*
  86  * Gromacs nonbonded kernel:   {KERNEL_NAME}
  87  * Electrostatics interaction: {KERNEL_ELEC}
  88  * VdW interaction:            {KERNEL_VDW}
  89  * Geometry:                   {GEOMETRY_I}-{GEOMETRY_J}
  90  * Calculate force/pot:        {KERNEL_VF}
  91  */
  92 void
  93 {KERNEL_NAME}
  94                     (t_nblist                    * gmx_restrict       nlist,
  95                      rvec                        * gmx_restrict          xx,
  96                      rvec                        * gmx_restrict          ff,
  97                      t_forcerec                  * gmx_restrict          fr,
  98                      t_mdatoms                   * gmx_restrict     mdatoms,
  99                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 100                      t_nrnb                      * gmx_restrict        nrnb)
 101 {
 102     /* ## Not all variables are used for all kernels, but any optimizing compiler fixes that, */
 103     /* ## so there is no point in going to extremes to exclude variables that are not needed. */
 104     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
 105      * just 0 for non-waters.
 106      * Suffixes A,B,C,D refer to j loop unrolling done with SSE, e.g. for the four different
 107      * jnr indices corresponding to data put in the four positions in the SIMD register.
 108      */
 109     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
 110     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
 111     int              jnrA,jnrB,jnrC,jnrD;
 112     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
 113     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
 114     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 115     real             rcutoff_scalar;
 116     real             *shiftvec,*fshift,*x,*f;
 117     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD;
 118     real             scratch[4*DIM];
 119     __m128           tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
 120     /* #for I in PARTICLES_I */
 121     int              vdwioffset{I};
 122     __m128           ix{I},iy{I},iz{I},fix{I},fiy{I},fiz{I},iq{I},isai{I};
 123     /* #endfor */
 124     /* #for J in PARTICLES_J */
 125     int              vdwjidx{J}A,vdwjidx{J}B,vdwjidx{J}C,vdwjidx{J}D;
 126     __m128           jx{J},jy{J},jz{J},fjx{J},fjy{J},fjz{J},jq{J},isaj{J};
 127     /* #endfor */
 128     /* #for I,J in PAIRS_IJ */
 129     __m128           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};
 130     /* #endfor */
 131     /* #if KERNEL_ELEC != 'None' */
 132     __m128           velec,felec,velecsum,facel,crf,krf,krf2;
 133     real             *charge;
 134     /* #endif */
 135     /* #if 'GeneralizedBorn' in KERNEL_ELEC */
 136     __m128i          gbitab;
 137     __m128           vgb,fgb,vgbsum,dvdasum,gbscale,gbtabscale,isaprod,gbqqfactor,gbinvepsdiff,gbeps,dvdatmp;
 138     __m128           minushalf = _mm_set1_ps(-0.5);
 139     real             *invsqrta,*dvda,*gbtab;
 140     /* #endif */
 141     /* #if KERNEL_VDW != 'None' */
 142     int              nvdwtype;
 143     __m128           rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
 144     int              *vdwtype;
 145     real             *vdwparam;
 146     __m128           one_sixth   = _mm_set1_ps(1.0/6.0);
 147     __m128           one_twelfth = _mm_set1_ps(1.0/12.0);
 148     /* #endif */
 149     /* #if 'Table' in KERNEL_ELEC or 'GeneralizedBorn' in KERNEL_ELEC or 'Table' in KERNEL_VDW */
 150     __m128i          vfitab;
 151     __m128i          ifour       = _mm_set1_epi32(4);
 152     __m128           rt,vfeps,vftabscale,Y,F,G,H,Heps,Fp,VV,FF;
 153     real             *vftab;
 154     /* #endif */
 155     /* #if 'LJEwald' in KERNEL_VDW */
 156     /* #for I,J in PAIRS_IJ */
 157     __m128           c6grid_{I}{J};
 158     /* #endfor */
 159     __m128           ewclj,ewclj2,ewclj6,ewcljrsq,poly,exponent,f6A,f6B,sh_lj_ewald;
 160     real             *vdwgridparam;
 161     __m128           one_half  = _mm_set1_ps(0.5);
 162     __m128           minus_one = _mm_set1_ps(-1.0);
 163     /* #endif */
 164     /* #if 'Ewald' in KERNEL_ELEC */
 165     __m128i          ewitab;
 166     __m128           ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
 167     real             *ewtab;
 168     /* #endif */
 169     /* #if 'PotentialSwitch' in [KERNEL_MOD_ELEC,KERNEL_MOD_VDW] */
 170     __m128           rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
 171     real             rswitch_scalar,d_scalar;
 172     /* #endif */
 173     __m128           dummy_mask,cutoff_mask;
 174     __m128           signbit = _mm_castsi128_ps( _mm_set1_epi32(0x80000000) );
 175     __m128           one     = _mm_set1_ps(1.0);
 176     __m128           two     = _mm_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            = _mm_set1_ps(fr->epsfac);
 190     charge           = mdatoms->chargeA;
 191     /*     #if 'ReactionField' in KERNEL_ELEC */
 192     krf              = _mm_set1_ps(fr->ic->k_rf);
 193     krf2             = _mm_set1_ps(fr->ic->k_rf*2.0);
 194     crf              = _mm_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     /* #if 'LJEwald' in KERNEL_VDW */
 203     vdwgridparam     = fr->ljpme_c6grid;
 204     sh_lj_ewald      = _mm_set1_ps(fr->ic->sh_lj_ewald);
 205     ewclj            = _mm_set1_ps(fr->ewaldcoeff_lj);
 206     ewclj2           = _mm_mul_ps(minus_one,_mm_mul_ps(ewclj,ewclj));
 207     /* #endif */
 208
 209     /* #if 'Table' in KERNEL_ELEC and 'Table' in KERNEL_VDW */
 210     vftab            = kernel_data->table_elec_vdw->data;
 211     vftabscale       = _mm_set1_ps(kernel_data->table_elec_vdw->scale);
 212     /* #elif 'Table' in KERNEL_ELEC */
 213     vftab            = kernel_data->table_elec->data;
 214     vftabscale       = _mm_set1_ps(kernel_data->table_elec->scale);
 215     /* #elif 'Table' in KERNEL_VDW */
 216     vftab            = kernel_data->table_vdw->data;
 217     vftabscale       = _mm_set1_ps(kernel_data->table_vdw->scale);
 218     /* #endif */
 219
 220     /* #if 'Ewald' in KERNEL_ELEC */
 221     sh_ewald         = _mm_set1_ps(fr->ic->sh_ewald);
 222     /*     #if KERNEL_VF=='Force' and KERNEL_MOD_ELEC!='PotentialSwitch' */
 223     ewtab            = fr->ic->tabq_coul_F;
 224     ewtabscale       = _mm_set1_ps(fr->ic->tabq_scale);
 225     ewtabhalfspace   = _mm_set1_ps(0.5/fr->ic->tabq_scale);
 226     /*     #else */
 227     ewtab            = fr->ic->tabq_coul_FDV0;
 228     ewtabscale       = _mm_set1_ps(fr->ic->tabq_scale);
 229     ewtabhalfspace   = _mm_set1_ps(0.5/fr->ic->tabq_scale);
 230      /*     #endif */
 231     /* #endif */
 232
 233     /* #if KERNEL_ELEC=='GeneralizedBorn' */
 234     invsqrta         = fr->invsqrta;
 235     dvda             = fr->dvda;
 236     gbtabscale       = _mm_set1_ps(fr->gbtab.scale);
 237     gbtab            = fr->gbtab.data;
 238     gbinvepsdiff     = _mm_set1_ps((1.0/fr->epsilon_r) - (1.0/fr->gb_epsilon_solvent));
 239     /* #endif */
 240
 241     /* #if 'Water' in GEOMETRY_I */
 242     /* Setup water-specific parameters */
 243     inr              = nlist->iinr[0];
 244     /*     #for I in PARTICLES_ELEC_I */
 245     iq{I}              = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+{I}]));
 246     /*     #endfor */
 247     /*     #for I in PARTICLES_VDW_I */
 248     vdwioffset{I}      = 2*nvdwtype*vdwtype[inr+{I}];
 249     /*     #endfor */
 250     /* #endif */
 251
 252     /* #if 'Water' in GEOMETRY_J */
 253     /*     #for J in PARTICLES_ELEC_J */
 254     jq{J}              = _mm_set1_ps(charge[inr+{J}]);
 255     /*     #endfor */
 256     /*     #for J in PARTICLES_VDW_J */
 257     vdwjidx{J}A        = 2*vdwtype[inr+{J}];
 258     /*     #endfor */
 259     /*     #for I,J in PAIRS_IJ */
 260     /*         #if 'electrostatics' in INTERACTION_FLAGS[I][J] */
 261     qq{I}{J}             = _mm_mul_ps(iq{I},jq{J});
 262     /*         #endif */
 263     /*         #if 'vdw' in INTERACTION_FLAGS[I][J] */
 264     /*             #if 'LJEwald' in KERNEL_VDW */
 265     c6_{I}{J}            = _mm_set1_ps(vdwparam[vdwioffset{I}+vdwjidx{J}A]);
 266     c12_{I}{J}           = _mm_set1_ps(vdwparam[vdwioffset{I}+vdwjidx{J}A+1]);
 267     c6grid_{I}{J}        = _mm_set1_ps(vdwgridparam[vdwioffset{I}+vdwjidx{J}A]);
 268     /*             #else */
 269     c6_{I}{J}            = _mm_set1_ps(vdwparam[vdwioffset{I}+vdwjidx{J}A]);
 270     c12_{I}{J}           = _mm_set1_ps(vdwparam[vdwioffset{I}+vdwjidx{J}A+1]);
 271     /*             #endif */
 272     /*         #endif */
 273     /*     #endfor */
 274     /* #endif */
 275
 276     /* #if KERNEL_MOD_ELEC!='None' or KERNEL_MOD_VDW!='None' */
 277     /*     #if KERNEL_ELEC!='None' */
 278     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 279     rcutoff_scalar   = fr->rcoulomb;
 280     /*     #else */
 281     rcutoff_scalar   = fr->rvdw;
 282     /*     #endif */
 283     rcutoff          = _mm_set1_ps(rcutoff_scalar);
 284     rcutoff2         = _mm_mul_ps(rcutoff,rcutoff);
 285     /* #endif */
 286
 287     /* #if KERNEL_MOD_VDW=='PotentialShift' */
 288     sh_vdw_invrcut6  = _mm_set1_ps(fr->ic->sh_invrc6);
 289     rvdw             = _mm_set1_ps(fr->rvdw);
 290     /* #endif */
 291
 292     /* #if 'PotentialSwitch' in [KERNEL_MOD_ELEC,KERNEL_MOD_VDW] */
 293     /*     #if KERNEL_MOD_ELEC=='PotentialSwitch'  */
 294     rswitch_scalar   = fr->rcoulomb_switch;
 295     rswitch          = _mm_set1_ps(rswitch_scalar);
 296     /*     #else */
 297     rswitch_scalar   = fr->rvdw_switch;
 298     rswitch          = _mm_set1_ps(rswitch_scalar);
 299     /*     #endif */
 300     /* Setup switch parameters */
 301     d_scalar         = rcutoff_scalar-rswitch_scalar;
 302     d                = _mm_set1_ps(d_scalar);
 303     swV3             = _mm_set1_ps(-10.0/(d_scalar*d_scalar*d_scalar));
 304     swV4             = _mm_set1_ps( 15.0/(d_scalar*d_scalar*d_scalar*d_scalar));
 305     swV5             = _mm_set1_ps( -6.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
 306     /*     #if 'Force' in KERNEL_VF */
 307     swF2             = _mm_set1_ps(-30.0/(d_scalar*d_scalar*d_scalar));
 308     swF3             = _mm_set1_ps( 60.0/(d_scalar*d_scalar*d_scalar*d_scalar));
 309     swF4             = _mm_set1_ps(-30.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
 310     /*     #endif */
 311     /* #endif */
 312
 313     /* Avoid stupid compiler warnings */
 314     jnrA = jnrB = jnrC = jnrD = 0;
 315     j_coord_offsetA = 0;
 316     j_coord_offsetB = 0;
 317     j_coord_offsetC = 0;
 318     j_coord_offsetD = 0;
 319
 320     /* ## Keep track of the floating point operations we issue for reporting! */
 321     /* #define OUTERFLOPS 0 */
 322     outeriter        = 0;
 323     inneriter        = 0;
 324
 325     for(iidx=0;iidx<4*DIM;iidx++)
 326     {
 327         scratch[iidx] = 0.0;
 328     }
 329
 330     /* Start outer loop over neighborlists */
 331     for(iidx=0; iidx<nri; iidx++)
 332     {
 333         /* Load shift vector for this list */
 334         i_shift_offset   = DIM*shiftidx[iidx];
 335
 336         /* Load limits for loop over neighbors */
 337         j_index_start    = jindex[iidx];
 338         j_index_end      = jindex[iidx+1];
 339
 340         /* Get outer coordinate index */
 341         inr              = iinr[iidx];
 342         i_coord_offset   = DIM*inr;
 343
 344         /* Load i particle coords and add shift vector */
 345         /* #if GEOMETRY_I == 'Particle' */
 346         gmx_mm_load_shift_and_1rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,&ix0,&iy0,&iz0);
 347         /* #elif GEOMETRY_I == 'Water3' */
 348         gmx_mm_load_shift_and_3rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,
 349                                                  &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2);
 350         /* #elif GEOMETRY_I == 'Water4' */
 351         /*     #if 0 in PARTICLES_I                 */
 352         gmx_mm_load_shift_and_4rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,
 353                                                  &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
 354         /*     #else                                */
 355         gmx_mm_load_shift_and_3rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset+DIM,
 356                                                  &ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
 357         /*     #endif                               */
 358         /* #endif                                   */
 359
 360         /* #if 'Force' in KERNEL_VF */
 361         /*     #for I in PARTICLES_I */
 362         fix{I}             = _mm_setzero_ps();
 363         fiy{I}             = _mm_setzero_ps();
 364         fiz{I}             = _mm_setzero_ps();
 365         /*     #endfor */
 366         /* #endif */
 367
 368         /* ## For water we already preloaded parameters at the start of the kernel */
 369         /* #if not 'Water' in GEOMETRY_I */
 370         /* Load parameters for i particles */
 371         /*     #for I in PARTICLES_ELEC_I */
 372         iq{I}              = _mm_mul_ps(facel,_mm_load1_ps(charge+inr+{I}));
 373         /*         #define OUTERFLOPS OUTERFLOPS+1 */
 374         /*         #if KERNEL_ELEC=='GeneralizedBorn' */
 375         isai{I}            = _mm_load1_ps(invsqrta+inr+{I});
 376         /*         #endif */
 377         /*     #endfor */
 378         /*     #for I in PARTICLES_VDW_I */
 379         vdwioffset{I}      = 2*nvdwtype*vdwtype[inr+{I}];
 380         /*     #endfor */
 381         /* #endif */
 382
 383         /* #if 'Potential' in KERNEL_VF */
 384         /* Reset potential sums */
 385         /*     #if KERNEL_ELEC != 'None' */
 386         velecsum         = _mm_setzero_ps();
 387         /*     #endif */
 388         /*     #if 'GeneralizedBorn' in KERNEL_ELEC */
 389         vgbsum           = _mm_setzero_ps();
 390         /*     #endif */
 391         /*     #if KERNEL_VDW != 'None' */
 392         vvdwsum          = _mm_setzero_ps();
 393         /*     #endif */
 394         /* #endif */
 395         /*     #if 'GeneralizedBorn' in KERNEL_ELEC and 'Force' in KERNEL_VF */
 396         dvdasum          = _mm_setzero_ps();
 397         /*     #endif */
 398
 399         /* #for ROUND in ['Loop','Epilogue'] */
 400
 401         /* #if ROUND =='Loop' */
 402         /* Start inner kernel loop */
 403         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4)
 404         {
 405         /* ## First round is normal loop (next statement resets indentation) */
 406         /*     #if 0 */
 407         }
 408         /*     #endif */
 409         /* #else */
 410         if(jidx<j_index_end)
 411         {
 412         /* ## Second round is epilogue */
 413         /* #endif */
 414         /* #define INNERFLOPS 0 */
 415
 416             /* Get j neighbor index, and coordinate index */
 417             /* #if ROUND =='Loop' */
 418             jnrA             = jjnr[jidx];
 419             jnrB             = jjnr[jidx+1];
 420             jnrC             = jjnr[jidx+2];
 421             jnrD             = jjnr[jidx+3];
 422             /* #else */
 423             jnrlistA         = jjnr[jidx];
 424             jnrlistB         = jjnr[jidx+1];
 425             jnrlistC         = jjnr[jidx+2];
 426             jnrlistD         = jjnr[jidx+3];
 427             /* Sign of each element will be negative for non-real atoms.
 428              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
 429              * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
 430              */
 431             dummy_mask = gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128()));
 432             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
 433             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
 434             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
 435             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
 436             /* #endif */
 437             j_coord_offsetA  = DIM*jnrA;
 438             j_coord_offsetB  = DIM*jnrB;
 439             j_coord_offsetC  = DIM*jnrC;
 440             j_coord_offsetD  = DIM*jnrD;
 441
 442             /* load j atom coordinates */
 443             /* #if GEOMETRY_J == 'Particle'             */
 444             gmx_mm_load_1rvec_4ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 445                                               x+j_coord_offsetC,x+j_coord_offsetD,
 446                                               &jx0,&jy0,&jz0);
 447             /* #elif GEOMETRY_J == 'Water3'             */
 448             gmx_mm_load_3rvec_4ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 449                                               x+j_coord_offsetC,x+j_coord_offsetD,
 450                                               &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,&jy2,&jz2);
 451             /* #elif GEOMETRY_J == 'Water4'             */
 452             /*     #if 0 in PARTICLES_J                 */
 453             gmx_mm_load_4rvec_4ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 454                                               x+j_coord_offsetC,x+j_coord_offsetD,
 455                                               &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,
 456                                               &jy2,&jz2,&jx3,&jy3,&jz3);
 457             /*     #else                                */
 458             gmx_mm_load_3rvec_4ptr_swizzle_ps(x+j_coord_offsetA+DIM,x+j_coord_offsetB+DIM,
 459                                               x+j_coord_offsetC+DIM,x+j_coord_offsetD+DIM,
 460                                               &jx1,&jy1,&jz1,&jx2,&jy2,&jz2,&jx3,&jy3,&jz3);
 461             /*     #endif                               */
 462             /* #endif                                   */
 463
 464             /* Calculate displacement vector */
 465             /* #for I,J in PAIRS_IJ */
 466             dx{I}{J}             = _mm_sub_ps(ix{I},jx{J});
 467             dy{I}{J}             = _mm_sub_ps(iy{I},jy{J});
 468             dz{I}{J}             = _mm_sub_ps(iz{I},jz{J});
 469             /*     #define INNERFLOPS INNERFLOPS+3 */
 470             /* #endfor */
 471
 472             /* Calculate squared distance and things based on it */
 473             /* #for I,J in PAIRS_IJ */
 474             rsq{I}{J}            = gmx_mm_calc_rsq_ps(dx{I}{J},dy{I}{J},dz{I}{J});
 475             /*     #define INNERFLOPS INNERFLOPS+5 */
 476             /* #endfor */
 477
 478             /* #for I,J in PAIRS_IJ */
 479             /*     #if 'rinv' in INTERACTION_FLAGS[I][J] */
 480             rinv{I}{J}           = gmx_mm_invsqrt_ps(rsq{I}{J});
 481             /*         #define INNERFLOPS INNERFLOPS+5 */
 482             /*     #endif */
 483             /* #endfor */
 484
 485             /* #for I,J in PAIRS_IJ */
 486             /*     #if 'rinvsq' in INTERACTION_FLAGS[I][J] */
 487             /*         # if 'rinv' not in INTERACTION_FLAGS[I][J] */
 488             rinvsq{I}{J}         = gmx_mm_inv_ps(rsq{I}{J});
 489             /*             #define INNERFLOPS INNERFLOPS+4 */
 490             /*         #else */
 491             rinvsq{I}{J}         = _mm_mul_ps(rinv{I}{J},rinv{I}{J});
 492             /*             #define INNERFLOPS INNERFLOPS+1 */
 493             /*         #endif */
 494             /*     #endif */
 495             /* #endfor */
 496
 497             /* #if not 'Water' in GEOMETRY_J */
 498             /* Load parameters for j particles */
 499             /*     #for J in PARTICLES_ELEC_J */
 500             jq{J}              = gmx_mm_load_4real_swizzle_ps(charge+jnrA+{J},charge+jnrB+{J},
 501                                                               charge+jnrC+{J},charge+jnrD+{J});
 502             /*         #if KERNEL_ELEC=='GeneralizedBorn' */
 503             isaj{J}            = gmx_mm_load_4real_swizzle_ps(invsqrta+jnrA+{J},invsqrta+jnrB+{J},
 504                                                               invsqrta+jnrC+{J},invsqrta+jnrD+{J});
 505             /*         #endif */
 506             /*     #endfor */
 507             /*     #for J in PARTICLES_VDW_J */
 508             vdwjidx{J}A        = 2*vdwtype[jnrA+{J}];
 509             vdwjidx{J}B        = 2*vdwtype[jnrB+{J}];
 510             vdwjidx{J}C        = 2*vdwtype[jnrC+{J}];
 511             vdwjidx{J}D        = 2*vdwtype[jnrD+{J}];
 512             /*     #endfor */
 513             /* #endif */
 514
 515             /* #if 'Force' in KERNEL_VF and not 'Particle' in GEOMETRY_I */
 516             /*     #for J in PARTICLES_J */
 517             fjx{J}             = _mm_setzero_ps();
 518             fjy{J}             = _mm_setzero_ps();
 519             fjz{J}             = _mm_setzero_ps();
 520             /*     #endfor */
 521             /* #endif */
 522
 523             /* #for I,J in PAIRS_IJ */
 524
 525             /**************************
 526              * CALCULATE INTERACTIONS *
 527              **************************/
 528
 529             /*     ## Note special check for TIP4P-TIP4P. Since we are cutting of all hydrogen interactions we also cut the LJ-only O-O interaction */
 530             /*     #if 'exactcutoff' in INTERACTION_FLAGS[I][J] or (GEOMETRY_I=='Water4' and GEOMETRY_J=='Water4' and 'exactcutoff' in INTERACTION_FLAGS[1][1]) */
 531             /*         ## We always calculate rinv/rinvsq above to enable pipelineing in compilers (performance tested on x86) */
 532             if (gmx_mm_any_lt(rsq{I}{J},rcutoff2))
 533             {
 534                 /*     #if 0    ## this and the next two lines is a hack to maintain auto-indentation in template file */
 535             }
 536             /*         #endif */
 537             /*         #define INNERFLOPS INNERFLOPS+1 */
 538             /*     #endif */
 539
 540             /*     #if 'r' in INTERACTION_FLAGS[I][J] */
 541             r{I}{J}              = _mm_mul_ps(rsq{I}{J},rinv{I}{J});
 542             /*         #if ROUND == 'Epilogue' */
 543             r{I}{J}              = _mm_andnot_ps(dummy_mask,r{I}{J});
 544             /*             #define INNERFLOPS INNERFLOPS+1 */
 545             /*         #endif */
 546             /*         #define INNERFLOPS INNERFLOPS+1 */
 547             /*     #endif */
 548
 549             /*     ## For water geometries we already loaded parameters at the start of the kernel */
 550             /*     #if not 'Water' in GEOMETRY_J */
 551             /* Compute parameters for interactions between i and j atoms */
 552             /*         #if 'electrostatics' in INTERACTION_FLAGS[I][J] */
 553             qq{I}{J}             = _mm_mul_ps(iq{I},jq{J});
 554             /*             #define INNERFLOPS INNERFLOPS+1 */
 555             /*         #endif */
 556             /*         #if 'vdw' in INTERACTION_FLAGS[I][J] */
 557             gmx_mm_load_4pair_swizzle_ps(vdwparam+vdwioffset{I}+vdwjidx{J}A,
 558                                          vdwparam+vdwioffset{I}+vdwjidx{J}B,
 559                                          vdwparam+vdwioffset{I}+vdwjidx{J}C,
 560                                          vdwparam+vdwioffset{I}+vdwjidx{J}D,
 561                                          &c6_{I}{J},&c12_{I}{J});
 562
 563             /*         #if 'LJEwald' in KERNEL_VDW */
 564             c6grid_{I}{J}       = gmx_mm_load_4real_swizzle_ps(vdwgridparam+vdwioffset{I}+vdwjidx{J}A,
 565                                                                vdwgridparam+vdwioffset{I}+vdwjidx{J}B,
 566                                                                vdwgridparam+vdwioffset{I}+vdwjidx{J}C,
 567                                                                vdwgridparam+vdwioffset{I}+vdwjidx{J}D);
 568             /*          #endif */
 569             /*         #endif */
 570             /*     #endif */
 571
 572             /*     #if 'table' in INTERACTION_FLAGS[I][J] */
 573             /* Calculate table index by multiplying r with table scale and truncate to integer */
 574             rt               = _mm_mul_ps(r{I}{J},vftabscale);
 575             vfitab           = _mm_cvttps_epi32(rt);
 576             vfeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
 577             /*         #define INNERFLOPS INNERFLOPS+4                          */
 578             /*         #if 'Table' in KERNEL_ELEC and 'Table' in KERNEL_VDW     */
 579             /*             ## 3 tables, 4 bytes per point: multiply index by 12 */
 580             vfitab           = _mm_slli_epi32(_mm_add_epi32(vfitab,_mm_slli_epi32(vfitab,1)),2);
 581             /*         #elif 'Table' in KERNEL_ELEC                             */
 582             /*             ## 1 table, 4 bytes per point: multiply index by 4   */
 583             vfitab           = _mm_slli_epi32(vfitab,2);
 584             /*         #elif 'Table' in KERNEL_VDW                              */
 585             /*             ## 2 tables, 4 bytes per point: multiply index by 8  */
 586             vfitab           = _mm_slli_epi32(vfitab,3);
 587             /*         #endif                                                   */
 588             /*     #endif */
 589
 590             /*     ## ELECTROSTATIC INTERACTIONS */
 591             /*     #if 'electrostatics' in INTERACTION_FLAGS[I][J] */
 592
 593             /*         #if KERNEL_ELEC=='Coulomb' */
 594
 595             /* COULOMB ELECTROSTATICS */
 596             velec            = _mm_mul_ps(qq{I}{J},rinv{I}{J});
 597             /*             #define INNERFLOPS INNERFLOPS+1 */
 598             /*             #if 'Force' in KERNEL_VF */
 599             felec            = _mm_mul_ps(velec,rinvsq{I}{J});
 600             /*                 #define INNERFLOPS INNERFLOPS+2 */
 601             /*             #endif */
 602
 603             /*         #elif KERNEL_ELEC=='ReactionField' */
 604
 605             /* REACTION-FIELD ELECTROSTATICS */
 606             /*             #if 'Potential' in KERNEL_VF */
 607             velec            = _mm_mul_ps(qq{I}{J},_mm_sub_ps(_mm_add_ps(rinv{I}{J},_mm_mul_ps(krf,rsq{I}{J})),crf));
 608             /*                 #define INNERFLOPS INNERFLOPS+4 */
 609             /*             #endif */
 610             /*             #if 'Force' in KERNEL_VF */
 611             felec            = _mm_mul_ps(qq{I}{J},_mm_sub_ps(_mm_mul_ps(rinv{I}{J},rinvsq{I}{J}),krf2));
 612             /*                 #define INNERFLOPS INNERFLOPS+3 */
 613             /*             #endif */
 614
 615             /*         #elif KERNEL_ELEC=='GeneralizedBorn' */
 616
 617             /* GENERALIZED BORN AND COULOMB ELECTROSTATICS */
 618             isaprod          = _mm_mul_ps(isai{I},isaj{J});
 619             gbqqfactor       = _mm_xor_ps(signbit,_mm_mul_ps(qq{I}{J},_mm_mul_ps(isaprod,gbinvepsdiff)));
 620             gbscale          = _mm_mul_ps(isaprod,gbtabscale);
 621             /*             #define INNERFLOPS INNERFLOPS+5 */
 622
 623             /* Calculate generalized born table index - this is a separate table from the normal one,
 624              * but we use the same procedure by multiplying r with scale and truncating to integer.
 625              */
 626             rt               = _mm_mul_ps(r{I}{J},gbscale);
 627             gbitab           = _mm_cvttps_epi32(rt);
 628             gbeps            = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
 629             gbitab           = _mm_slli_epi32(gbitab,2);
 630             Y                = _mm_load_ps( gbtab + gmx_mm_extract_epi32(gbitab,0) );
 631             F                = _mm_load_ps( gbtab + gmx_mm_extract_epi32(gbitab,1) );
 632             G                = _mm_load_ps( gbtab + gmx_mm_extract_epi32(gbitab,2) );
 633             H                = _mm_load_ps( gbtab + gmx_mm_extract_epi32(gbitab,3) );
 634             _MM_TRANSPOSE4_PS(Y,F,G,H);
 635             Heps             = _mm_mul_ps(gbeps,H);
 636             Fp               = _mm_add_ps(F,_mm_mul_ps(gbeps,_mm_add_ps(G,Heps)));
 637             VV               = _mm_add_ps(Y,_mm_mul_ps(gbeps,Fp));
 638             vgb              = _mm_mul_ps(gbqqfactor,VV);
 639             /*             #define INNERFLOPS INNERFLOPS+10 */
 640
 641             /*             #if 'Force' in KERNEL_VF */
 642             FF               = _mm_add_ps(Fp,_mm_mul_ps(gbeps,_mm_add_ps(G,_mm_add_ps(Heps,Heps))));
 643             fgb              = _mm_mul_ps(gbqqfactor,_mm_mul_ps(FF,gbscale));
 644             dvdatmp          = _mm_mul_ps(minushalf,_mm_add_ps(vgb,_mm_mul_ps(fgb,r{I}{J})));
 645             /*                 #if ROUND == 'Epilogue' */
 646             dvdatmp          = _mm_andnot_ps(dummy_mask,dvdatmp);
 647             /*                 #endif */
 648             dvdasum          = _mm_add_ps(dvdasum,dvdatmp);
 649             /*                 #if ROUND == 'Loop' */
 650             fjptrA           = dvda+jnrA;
 651             fjptrB           = dvda+jnrB;
 652             fjptrC           = dvda+jnrC;
 653             fjptrD           = dvda+jnrD;
 654             /*                 #else */
 655             /* 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. */
 656             fjptrA             = (jnrlistA>=0) ? dvda+jnrA : scratch;
 657             fjptrB             = (jnrlistB>=0) ? dvda+jnrB : scratch;
 658             fjptrC             = (jnrlistC>=0) ? dvda+jnrC : scratch;
 659             fjptrD             = (jnrlistD>=0) ? dvda+jnrD : scratch;
 660             /*                 #endif */
 661             gmx_mm_increment_4real_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,_mm_mul_ps(dvdatmp,_mm_mul_ps(isaj{J},isaj{J})));
 662             /*                 #define INNERFLOPS INNERFLOPS+13 */
 663             /*             #endif */
 664             velec            = _mm_mul_ps(qq{I}{J},rinv{I}{J});
 665             /*                 #define INNERFLOPS INNERFLOPS+1 */
 666             /*             #if 'Force' in KERNEL_VF */
 667             felec            = _mm_mul_ps(_mm_sub_ps(_mm_mul_ps(velec,rinv{I}{J}),fgb),rinv{I}{J});
 668             /*                 #define INNERFLOPS INNERFLOPS+3 */
 669             /*             #endif */
 670
 671             /*         #elif KERNEL_ELEC=='Ewald' */
 672             /* EWALD ELECTROSTATICS */
 673
 674             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 675             ewrt             = _mm_mul_ps(r{I}{J},ewtabscale);
 676             ewitab           = _mm_cvttps_epi32(ewrt);
 677             eweps            = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR));
 678             /*             #define INNERFLOPS INNERFLOPS+4 */
 679             /*             #if 'Potential' in KERNEL_VF or KERNEL_MOD_ELEC=='PotentialSwitch' */
 680             ewitab           = _mm_slli_epi32(ewitab,2);
 681             ewtabF           = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0) );
 682             ewtabD           = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1) );
 683             ewtabV           = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2) );
 684             ewtabFn          = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3) );
 685             _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn);
 686             felec            = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD));
 687             /*                 #define INNERFLOPS INNERFLOPS+2 */
 688             /*                 #if KERNEL_MOD_ELEC=='PotentialShift' */
 689             velec            = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec)));
 690             velec            = _mm_mul_ps(qq{I}{J},_mm_sub_ps(_mm_sub_ps(rinv{I}{J},sh_ewald),velec));
 691             /*                     #define INNERFLOPS INNERFLOPS+7 */
 692             /*                 #else */
 693             velec            = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec)));
 694             velec            = _mm_mul_ps(qq{I}{J},_mm_sub_ps(rinv{I}{J},velec));
 695             /*                     #define INNERFLOPS INNERFLOPS+6 */
 696             /*                 #endif */
 697             /*                 #if 'Force' in KERNEL_VF */
 698             felec            = _mm_mul_ps(_mm_mul_ps(qq{I}{J},rinv{I}{J}),_mm_sub_ps(rinvsq{I}{J},felec));
 699             /*                      #define INNERFLOPS INNERFLOPS+3 */
 700             /*                 #endif */
 701             /*             #elif KERNEL_VF=='Force' */
 702             gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0),ewtab + gmx_mm_extract_epi32(ewitab,1),
 703                                          ewtab + gmx_mm_extract_epi32(ewitab,2),ewtab + gmx_mm_extract_epi32(ewitab,3),
 704                                          &ewtabF,&ewtabFn);
 705             felec            = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn));
 706             felec            = _mm_mul_ps(_mm_mul_ps(qq{I}{J},rinv{I}{J}),_mm_sub_ps(rinvsq{I}{J},felec));
 707             /*                 #define INNERFLOPS INNERFLOPS+7 */
 708             /*             #endif */
 709
 710             /*         #elif KERNEL_ELEC=='CubicSplineTable' */
 711
 712             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 713             Y                = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,0) );
 714             F                = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,1) );
 715             G                = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,2) );
 716             H                = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,3) );
 717             _MM_TRANSPOSE4_PS(Y,F,G,H);
 718             Heps             = _mm_mul_ps(vfeps,H);
 719             Fp               = _mm_add_ps(F,_mm_mul_ps(vfeps,_mm_add_ps(G,Heps)));
 720             /*             #define INNERFLOPS INNERFLOPS+4 */
 721             /*             #if 'Potential' in KERNEL_VF */
 722             VV               = _mm_add_ps(Y,_mm_mul_ps(vfeps,Fp));
 723             velec            = _mm_mul_ps(qq{I}{J},VV);
 724             /*                 #define INNERFLOPS INNERFLOPS+3 */
 725             /*             #endif */
 726             /*             #if 'Force' in KERNEL_VF */
 727             FF               = _mm_add_ps(Fp,_mm_mul_ps(vfeps,_mm_add_ps(G,_mm_add_ps(Heps,Heps))));
 728             felec            = _mm_xor_ps(signbit,_mm_mul_ps(_mm_mul_ps(qq{I}{J},FF),_mm_mul_ps(vftabscale,rinv{I}{J})));
 729             /*                 #define INNERFLOPS INNERFLOPS+7 */
 730             /*             #endif */
 731             /*         #endif */
 732             /*         ## End of check for electrostatics interaction forms */
 733             /*     #endif */
 734             /*     ## END OF ELECTROSTATIC INTERACTION CHECK FOR PAIR I-J */
 735
 736             /*     #if 'vdw' in INTERACTION_FLAGS[I][J] */
 737
 738             /*         #if KERNEL_VDW=='LennardJones' */
 739
 740             /* LENNARD-JONES DISPERSION/REPULSION */
 741
 742             rinvsix          = _mm_mul_ps(_mm_mul_ps(rinvsq{I}{J},rinvsq{I}{J}),rinvsq{I}{J});
 743             /*             #define INNERFLOPS INNERFLOPS+2 */
 744             /*             #if 'Potential' in KERNEL_VF or KERNEL_MOD_VDW=='PotentialSwitch' */
 745             vvdw6            = _mm_mul_ps(c6_{I}{J},rinvsix);
 746             vvdw12           = _mm_mul_ps(c12_{I}{J},_mm_mul_ps(rinvsix,rinvsix));
 747             /*                 #define INNERFLOPS INNERFLOPS+3 */
 748             /*                 #if KERNEL_MOD_VDW=='PotentialShift' */
 749             vvdw             = _mm_sub_ps(_mm_mul_ps( _mm_sub_ps(vvdw12 , _mm_mul_ps(c12_{I}{J},_mm_mul_ps(sh_vdw_invrcut6,sh_vdw_invrcut6))), one_twelfth) ,
 750                                           _mm_mul_ps( _mm_sub_ps(vvdw6,_mm_mul_ps(c6_{I}{J},sh_vdw_invrcut6)),one_sixth));
 751             /*                     #define INNERFLOPS INNERFLOPS+8 */
 752             /*                 #else */
 753             vvdw             = _mm_sub_ps( _mm_mul_ps(vvdw12,one_twelfth) , _mm_mul_ps(vvdw6,one_sixth) );
 754             /*                     #define INNERFLOPS INNERFLOPS+3 */
 755             /*                 #endif */
 756             /*                 ## Check for force inside potential check, i.e. this means we already did the potential part */
 757             /*                 #if 'Force' in KERNEL_VF */
 758             fvdw             = _mm_mul_ps(_mm_sub_ps(vvdw12,vvdw6),rinvsq{I}{J});
 759             /*                     #define INNERFLOPS INNERFLOPS+2 */
 760             /*                 #endif */
 761             /*             #elif KERNEL_VF=='Force' */
 762             /*                 ## Force-only LennardJones makes it possible to save 1 flop (they do add up...) */
 763             fvdw             = _mm_mul_ps(_mm_sub_ps(_mm_mul_ps(c12_{I}{J},rinvsix),c6_{I}{J}),_mm_mul_ps(rinvsix,rinvsq{I}{J}));
 764             /*                 #define INNERFLOPS INNERFLOPS+4 */
 765             /*             #endif */
 766
 767             /*         #elif KERNEL_VDW=='CubicSplineTable' */
 768
 769             /* CUBIC SPLINE TABLE DISPERSION */
 770             /*             #if 'Table' in KERNEL_ELEC */
 771             vfitab           = _mm_add_epi32(vfitab,ifour);
 772             /*             #endif                     */
 773             Y                = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,0) );
 774             F                = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,1) );
 775             G                = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,2) );
 776             H                = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,3) );
 777             _MM_TRANSPOSE4_PS(Y,F,G,H);
 778             Heps             = _mm_mul_ps(vfeps,H);
 779             Fp               = _mm_add_ps(F,_mm_mul_ps(vfeps,_mm_add_ps(G,Heps)));
 780             /*             #define INNERFLOPS INNERFLOPS+4 */
 781             /*             #if 'Potential' in KERNEL_VF */
 782             VV               = _mm_add_ps(Y,_mm_mul_ps(vfeps,Fp));
 783             vvdw6            = _mm_mul_ps(c6_{I}{J},VV);
 784             /*                 #define INNERFLOPS INNERFLOPS+3 */
 785             /*             #endif */
 786             /*             #if 'Force' in KERNEL_VF */
 787             FF               = _mm_add_ps(Fp,_mm_mul_ps(vfeps,_mm_add_ps(G,_mm_add_ps(Heps,Heps))));
 788             fvdw6            = _mm_mul_ps(c6_{I}{J},FF);
 789             /*                 #define INNERFLOPS INNERFLOPS+4 */
 790             /*             #endif */
 791
 792             /* CUBIC SPLINE TABLE REPULSION */
 793             vfitab           = _mm_add_epi32(vfitab,ifour);
 794             Y                = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,0) );
 795             F                = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,1) );
 796             G                = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,2) );
 797             H                = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,3) );
 798             _MM_TRANSPOSE4_PS(Y,F,G,H);
 799             Heps             = _mm_mul_ps(vfeps,H);
 800             Fp               = _mm_add_ps(F,_mm_mul_ps(vfeps,_mm_add_ps(G,Heps)));
 801             /*             #define INNERFLOPS INNERFLOPS+4 */
 802             /*             #if 'Potential' in KERNEL_VF */
 803             VV               = _mm_add_ps(Y,_mm_mul_ps(vfeps,Fp));
 804             vvdw12           = _mm_mul_ps(c12_{I}{J},VV);
 805             /*                 #define INNERFLOPS INNERFLOPS+3 */
 806             /*             #endif */
 807             /*             #if 'Force' in KERNEL_VF */
 808             FF               = _mm_add_ps(Fp,_mm_mul_ps(vfeps,_mm_add_ps(G,_mm_add_ps(Heps,Heps))));
 809             fvdw12           = _mm_mul_ps(c12_{I}{J},FF);
 810             /*                 #define INNERFLOPS INNERFLOPS+5 */
 811             /*             #endif */
 812             /*             #if 'Potential' in KERNEL_VF */
 813             vvdw             = _mm_add_ps(vvdw12,vvdw6);
 814             /*                 #define INNERFLOPS INNERFLOPS+1 */
 815             /*             #endif */
 816             /*             #if 'Force' in KERNEL_VF */
 817             fvdw             = _mm_xor_ps(signbit,_mm_mul_ps(_mm_add_ps(fvdw6,fvdw12),_mm_mul_ps(vftabscale,rinv{I}{J})));
 818             /*                 #define INNERFLOPS INNERFLOPS+4 */
 819             /*             #endif */
 820
 821             /*         #elif KERNEL_VDW=='LJEwald' */
 822
 823             /* Analytical LJ-PME */
 824             rinvsix          = _mm_mul_ps(_mm_mul_ps(rinvsq{I}{J},rinvsq{I}{J}),rinvsq{I}{J});
 825             ewcljrsq         = _mm_mul_ps(ewclj2,rsq{I}{J});
 826             ewclj6           = _mm_mul_ps(ewclj2,_mm_mul_ps(ewclj2,ewclj2));
 827             exponent         = gmx_simd_exp_r(ewcljrsq);
 828             /* poly = exp(-(beta*r)^2) * (1 + (beta*r)^2 + (beta*r)^4 /2) */
 829             poly             = _mm_mul_ps(exponent,_mm_add_ps(_mm_sub_ps(one,ewcljrsq),_mm_mul_ps(_mm_mul_ps(ewcljrsq,ewcljrsq),one_half)));
 830             /*             #define INNERFLOPS INNERFLOPS+11 */
 831             /*             #if 'Potential' in KERNEL_VF or KERNEL_MOD_VDW=='PotentialSwitch' */
 832             /* vvdw6 = [C6 - C6grid * (1-poly)]/r6 */
 833             vvdw6            = _mm_mul_ps(_mm_sub_ps(c6_{I}{J},_mm_mul_ps(c6grid_{I}{J},_mm_sub_ps(one,poly))),rinvsix);
 834             vvdw12           = _mm_mul_ps(c12_{I}{J},_mm_mul_ps(rinvsix,rinvsix));
 835             /*                 #define INNERFLOPS INNERFLOPS+6 */
 836             /*                 #if KERNEL_MOD_VDW=='PotentialShift' */
 837             vvdw             = _mm_sub_ps(_mm_mul_ps( _mm_sub_ps(vvdw12 , _mm_mul_ps(c12_{I}{J},_mm_mul_ps(sh_vdw_invrcut6,sh_vdw_invrcut6))),one_twelfth),
 838                                _mm_mul_ps( _mm_sub_ps(vvdw6,_mm_add_ps(_mm_mul_ps(c6_{I}{J},sh_vdw_invrcut6),_mm_mul_ps(c6grid_{I}{J},sh_lj_ewald))),one_sixth));
 839             /*                 #define INNERFLOPS INNERFLOPS+10 */
 840             /*                 #else */
 841             vvdw             = _mm_sub_ps(_mm_mul_ps(vvdw12,one_twelfth),_mm_mul_ps(vvdw6,one_sixth));
 842             /*                 #define INNERFLOPS INNERFLOPS+3 */
 843             /*                 #endif */
 844             /*                 ## Check for force inside potential check, i.e. this means we already did the potential part */
 845             /*                 #if 'Force' in KERNEL_VF */
 846             /* fvdw = vvdw12/r - (vvdw6/r + (C6grid * exponent * beta^6)/r) */
 847             fvdw             = _mm_mul_ps(_mm_sub_ps(vvdw12,_mm_sub_ps(vvdw6,_mm_mul_ps(_mm_mul_ps(c6grid_{I}{J},one_sixth),_mm_mul_ps(exponent,ewclj6)))),rinvsq{I}{J});
 848             /*                 #define INNERFLOPS INNERFLOPS+6 */
 849             /*                 #endif */
 850             /*             #elif KERNEL_VF=='Force' */
 851             /* f6A = 6 * C6grid * (1 - poly) */
 852             f6A              = _mm_mul_ps(c6grid_{I}{J},_mm_sub_ps(one,poly));
 853             /* f6B = C6grid * exponent * beta^6 */
 854             f6B              = _mm_mul_ps(_mm_mul_ps(c6grid_{I}{J},one_sixth),_mm_mul_ps(exponent,ewclj6));
 855             /* fvdw = 12*C12/r13 - ((6*C6 - f6A)/r6 + f6B)/r */
 856             fvdw              = _mm_mul_ps(_mm_add_ps(_mm_mul_ps(_mm_sub_ps(_mm_mul_ps(c12_{I}{J},rinvsix),_mm_sub_ps(c6_{I}{J},f6A)),rinvsix),f6B),rinvsq{I}{J});
 857             /*                 #define INNERFLOPS INNERFLOPS+11 */
 858             /*             #endif */
 859             /*         #endif */
 860             /*         ## End of check for vdw interaction forms */
 861             /*     #endif */
 862             /*     ## END OF VDW INTERACTION CHECK FOR PAIR I-J */
 863
 864             /*     #if 'switch' in INTERACTION_FLAGS[I][J] */
 865             d                = _mm_sub_ps(r{I}{J},rswitch);
 866             d                = _mm_max_ps(d,_mm_setzero_ps());
 867             d2               = _mm_mul_ps(d,d);
 868             sw               = _mm_add_ps(one,_mm_mul_ps(d2,_mm_mul_ps(d,_mm_add_ps(swV3,_mm_mul_ps(d,_mm_add_ps(swV4,_mm_mul_ps(d,swV5)))))));
 869             /*         #define INNERFLOPS INNERFLOPS+10 */
 870
 871             /*         #if 'Force' in KERNEL_VF */
 872             dsw              = _mm_mul_ps(d2,_mm_add_ps(swF2,_mm_mul_ps(d,_mm_add_ps(swF3,_mm_mul_ps(d,swF4)))));
 873             /*             #define INNERFLOPS INNERFLOPS+5 */
 874             /*         #endif */
 875
 876             /* Evaluate switch function */
 877             /*         #if 'Force' in KERNEL_VF */
 878             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 879             /*             #if 'electrostatics' in INTERACTION_FLAGS[I][J] and KERNEL_MOD_ELEC=='PotentialSwitch' */
 880             felec            = _mm_sub_ps( _mm_mul_ps(felec,sw) , _mm_mul_ps(rinv{I}{J},_mm_mul_ps(velec,dsw)) );
 881             /*                 #define INNERFLOPS INNERFLOPS+4 */
 882             /*             #endif */
 883             /*             #if 'vdw' in INTERACTION_FLAGS[I][J] and KERNEL_MOD_VDW=='PotentialSwitch' */
 884             fvdw             = _mm_sub_ps( _mm_mul_ps(fvdw,sw) , _mm_mul_ps(rinv{I}{J},_mm_mul_ps(vvdw,dsw)) );
 885             /*                 #define INNERFLOPS INNERFLOPS+4 */
 886             /*             #endif */
 887             /*         #endif */
 888             /*         #if 'Potential' in KERNEL_VF */
 889             /*             #if 'electrostatics' in INTERACTION_FLAGS[I][J] and KERNEL_MOD_ELEC=='PotentialSwitch' */
 890             velec            = _mm_mul_ps(velec,sw);
 891             /*                 #define INNERFLOPS INNERFLOPS+1 */
 892             /*             #endif */
 893             /*             #if 'vdw' in INTERACTION_FLAGS[I][J] and KERNEL_MOD_VDW=='PotentialSwitch' */
 894             vvdw             = _mm_mul_ps(vvdw,sw);
 895             /*                 #define INNERFLOPS INNERFLOPS+1 */
 896             /*             #endif */
 897             /*         #endif */
 898             /*     #endif */
 899             /*     ## Note special check for TIP4P-TIP4P. Since we are cutting of all hydrogen interactions we also cut the LJ-only O-O interaction */
 900             /*     #if 'exactcutoff' in INTERACTION_FLAGS[I][J] or (GEOMETRY_I=='Water4' and GEOMETRY_J=='Water4' and 'exactcutoff' in INTERACTION_FLAGS[1][1]) */
 901             cutoff_mask      = _mm_cmplt_ps(rsq{I}{J},rcutoff2);
 902             /*         #define INNERFLOPS INNERFLOPS+1 */
 903             /*     #endif */
 904
 905             /*     #if 'Potential' in KERNEL_VF */
 906             /* Update potential sum for this i atom from the interaction with this j atom. */
 907             /*         #if 'electrostatics' in INTERACTION_FLAGS[I][J] */
 908             /*             #if 'exactcutoff' in INTERACTION_FLAGS[I][J] */
 909             velec            = _mm_and_ps(velec,cutoff_mask);
 910             /*                 #define INNERFLOPS INNERFLOPS+1 */
 911             /*             #endif                                       */
 912             /*             #if ROUND == 'Epilogue' */
 913             velec            = _mm_andnot_ps(dummy_mask,velec);
 914             /*             #endif */
 915             velecsum         = _mm_add_ps(velecsum,velec);
 916             /*             #define INNERFLOPS INNERFLOPS+1 */
 917             /*             #if KERNEL_ELEC=='GeneralizedBorn' */
 918             /*             #if 'exactcutoff' in INTERACTION_FLAGS[I][J] */
 919             vgb              = _mm_and_ps(vgb,cutoff_mask);
 920             /*                 #define INNERFLOPS INNERFLOPS+1 */
 921             /*             #endif                                       */
 922             /*             #if ROUND == 'Epilogue' */
 923             vgb              = _mm_andnot_ps(dummy_mask,vgb);
 924             /*             #endif */
 925             vgbsum           = _mm_add_ps(vgbsum,vgb);
 926             /*                 #define INNERFLOPS INNERFLOPS+1 */
 927             /*             #endif */
 928             /*         #endif */
 929             /*         #if 'vdw' in INTERACTION_FLAGS[I][J] */
 930             /*     ## Note special check for TIP4P-TIP4P. Since we are cutting of all hydrogen interactions we also cut the LJ-only O-O interaction */
 931             /*     #if 'exactcutoff' in INTERACTION_FLAGS[I][J] or (GEOMETRY_I=='Water4' and GEOMETRY_J=='Water4' and 'exactcutoff' in INTERACTION_FLAGS[1][1]) */
 932             vvdw             = _mm_and_ps(vvdw,cutoff_mask);
 933             /*                 #define INNERFLOPS INNERFLOPS+1 */
 934             /*             #endif                                       */
 935             /*             #if ROUND == 'Epilogue' */
 936             vvdw             = _mm_andnot_ps(dummy_mask,vvdw);
 937             /*             #endif */
 938             vvdwsum          = _mm_add_ps(vvdwsum,vvdw);
 939             /*             #define INNERFLOPS INNERFLOPS+1 */
 940             /*         #endif */
 941             /*     #endif */
 942
 943             /*     #if 'Force' in KERNEL_VF */
 944
 945             /*         #if 'electrostatics' in INTERACTION_FLAGS[I][J] and 'vdw' in INTERACTION_FLAGS[I][J] */
 946             fscal            = _mm_add_ps(felec,fvdw);
 947             /*             #define INNERFLOPS INNERFLOPS+1 */
 948             /*         #elif 'electrostatics' in INTERACTION_FLAGS[I][J] */
 949             fscal            = felec;
 950             /*         #elif 'vdw' in INTERACTION_FLAGS[I][J] */
 951             fscal            = fvdw;
 952             /*        #endif */
 953
 954             /*     ## Note special check for TIP4P-TIP4P. Since we are cutting of all hydrogen interactions we also cut the LJ-only O-O interaction */
 955             /*     #if 'exactcutoff' in INTERACTION_FLAGS[I][J] or (GEOMETRY_I=='Water4' and GEOMETRY_J=='Water4' and 'exactcutoff' in INTERACTION_FLAGS[1][1]) */
 956             fscal            = _mm_and_ps(fscal,cutoff_mask);
 957             /*                 #define INNERFLOPS INNERFLOPS+1 */
 958             /*             #endif                                       */
 959
 960             /*             #if ROUND == 'Epilogue' */
 961             fscal            = _mm_andnot_ps(dummy_mask,fscal);
 962             /*             #endif */
 963
 964             /* Calculate temporary vectorial force */
 965             tx               = _mm_mul_ps(fscal,dx{I}{J});
 966             ty               = _mm_mul_ps(fscal,dy{I}{J});
 967             tz               = _mm_mul_ps(fscal,dz{I}{J});
 968
 969             /* Update vectorial force */
 970             fix{I}             = _mm_add_ps(fix{I},tx);
 971             fiy{I}             = _mm_add_ps(fiy{I},ty);
 972             fiz{I}             = _mm_add_ps(fiz{I},tz);
 973             /*             #define INNERFLOPS INNERFLOPS+6 */
 974
 975             /* #if GEOMETRY_I == 'Particle'             */
 976             /*     #if ROUND == 'Loop' */
 977             fjptrA             = f+j_coord_offsetA;
 978             fjptrB             = f+j_coord_offsetB;
 979             fjptrC             = f+j_coord_offsetC;
 980             fjptrD             = f+j_coord_offsetD;
 981             /*     #else */
 982             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
 983             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
 984             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
 985             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
 986             /*     #endif */
 987             gmx_mm_decrement_1rvec_4ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,tx,ty,tz);
 988             /*     #define INNERFLOPS INNERFLOPS+3      */
 989             /* #else                                    */
 990             fjx{J}             = _mm_add_ps(fjx{J},tx);
 991             fjy{J}             = _mm_add_ps(fjy{J},ty);
 992             fjz{J}             = _mm_add_ps(fjz{J},tz);
 993             /*     #define INNERFLOPS INNERFLOPS+3      */
 994             /* #endif                                   */
 995
 996             /*     #endif */
 997
 998             /*     ## Note special check for TIP4P-TIP4P. Since we are cutting of all hydrogen interactions we also cut the LJ-only O-O interaction */
 999             /*     #if 'exactcutoff' in INTERACTION_FLAGS[I][J] or (GEOMETRY_I=='Water4' and GEOMETRY_J=='Water4' and 'exactcutoff' in INTERACTION_FLAGS[1][1]) */
1000             /*         #if 0    ## This and next two lines is a hack to maintain indentation in template file */
1001             {
1002                 /*     #endif */
1003             }
1004             /*     #endif */
1005             /*    ## End of check for the interaction being outside the cutoff */
1006
1007             /* #endfor */
1008             /* ## End of loop over i-j interaction pairs */
1009
1010             /* #if GEOMETRY_I != 'Particle' */
1011             /*     #if ROUND == 'Loop' */
1012             fjptrA             = f+j_coord_offsetA;
1013             fjptrB             = f+j_coord_offsetB;
1014             fjptrC             = f+j_coord_offsetC;
1015             fjptrD             = f+j_coord_offsetD;
1016             /*     #else */
1017             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
1018             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
1019             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
1020             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
1021             /*     #endif */
1022             /* #endif */
1023
1024             /* #if 'Water' in GEOMETRY_I and GEOMETRY_J == 'Particle' */
1025             gmx_mm_decrement_1rvec_4ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0);
1026             /* #elif GEOMETRY_J == 'Water3'               */
1027             gmx_mm_decrement_3rvec_4ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,
1028                                                    fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2);
1029             /*     #define INNERFLOPS INNERFLOPS+9      */
1030             /* #elif GEOMETRY_J == 'Water4'             */
1031             /*     #if 0 in PARTICLES_J                 */
1032             gmx_mm_decrement_4rvec_4ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,
1033                                                    fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,
1034                                                    fjx2,fjy2,fjz2,fjx3,fjy3,fjz3);
1035             /*     #define INNERFLOPS INNERFLOPS+12     */
1036             /*     #else                                */
1037             gmx_mm_decrement_3rvec_4ptr_swizzle_ps(fjptrA+DIM,fjptrB+DIM,fjptrC+DIM,fjptrD+DIM,
1038                                                    fjx1,fjy1,fjz1,fjx2,fjy2,fjz2,fjx3,fjy3,fjz3);
1039             /*     #define INNERFLOPS INNERFLOPS+9      */
1040             /*     #endif                               */
1041             /* #endif                                   */
1042
1043             /* Inner loop uses {INNERFLOPS} flops */
1044         }
1045
1046         /* #endfor */
1047
1048         /* End of innermost loop */
1049
1050         /* #if 'Force' in KERNEL_VF */
1051         /*     #if GEOMETRY_I == 'Particle'            */
1052         gmx_mm_update_iforce_1atom_swizzle_ps(fix0,fiy0,fiz0,
1053                                               f+i_coord_offset,fshift+i_shift_offset);
1054         /*         #define OUTERFLOPS OUTERFLOPS+6     */
1055         /*     #elif GEOMETRY_I == 'Water3'            */
1056         gmx_mm_update_iforce_3atom_swizzle_ps(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,
1057                                               f+i_coord_offset,fshift+i_shift_offset);
1058         /*         #define OUTERFLOPS OUTERFLOPS+18    */
1059         /*     #elif GEOMETRY_I == 'Water4'            */
1060         /*         #if 0 in PARTICLES_I                */
1061         gmx_mm_update_iforce_4atom_swizzle_ps(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
1062                                               f+i_coord_offset,fshift+i_shift_offset);
1063         /*             #define OUTERFLOPS OUTERFLOPS+24    */
1064         /*         #else                               */
1065         gmx_mm_update_iforce_3atom_swizzle_ps(fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
1066                                               f+i_coord_offset+DIM,fshift+i_shift_offset);
1067         /*             #define OUTERFLOPS OUTERFLOPS+18    */
1068         /*         #endif                              */
1069         /*     #endif                                  */
1070         /* #endif                                      */
1071
1072         /* #if 'Potential' in KERNEL_VF */
1073         ggid                        = gid[iidx];
1074         /* Update potential energies */
1075         /*     #if KERNEL_ELEC != 'None' */
1076         gmx_mm_update_1pot_ps(velecsum,kernel_data->energygrp_elec+ggid);
1077         /*         #define OUTERFLOPS OUTERFLOPS+1 */
1078         /*     #endif */
1079         /*     #if 'GeneralizedBorn' in KERNEL_ELEC */
1080         gmx_mm_update_1pot_ps(vgbsum,kernel_data->energygrp_polarization+ggid);
1081         /*         #define OUTERFLOPS OUTERFLOPS+1 */
1082         /*     #endif */
1083         /*     #if KERNEL_VDW != 'None' */
1084         gmx_mm_update_1pot_ps(vvdwsum,kernel_data->energygrp_vdw+ggid);
1085         /*         #define OUTERFLOPS OUTERFLOPS+1 */
1086         /*     #endif */
1087         /* #endif */
1088         /*     #if 'GeneralizedBorn' in KERNEL_ELEC and 'Force' in KERNEL_VF */
1089         dvdasum = _mm_mul_ps(dvdasum, _mm_mul_ps(isai{I},isai{I}));
1090         gmx_mm_update_1pot_ps(dvdasum,dvda+inr);
1091         /*     #endif */
1092
1093         /* Increment number of inner iterations */
1094         inneriter                  += j_index_end - j_index_start;
1095
1096         /* Outer loop uses {OUTERFLOPS} flops */
1097     }
1098
1099     /* Increment number of outer iterations */
1100     outeriter        += nri;
1101
1102     /* Update outer/inner flops */
1103     /* ## NB: This is not important, it just affects the flopcount. However, since our preprocessor is */
1104     /* ## primitive and replaces aggressively even in strings inside these directives, we need to      */
1105     /* ## assemble the main part of the name (containing KERNEL/ELEC/VDW) directly in the source.      */
1106     /* #if GEOMETRY_I == 'Water3'            */
1107     /*     #define ISUFFIX '_W3'             */
1108     /* #elif GEOMETRY_I == 'Water4'          */
1109     /*     #define ISUFFIX '_W4'             */
1110     /* #else                                 */
1111     /*     #define ISUFFIX ''                */
1112     /* #endif                                */
1113     /* #if GEOMETRY_J == 'Water3'            */
1114     /*     #define JSUFFIX 'W3'              */
1115     /* #elif GEOMETRY_J == 'Water4'          */
1116     /*     #define JSUFFIX 'W4'              */
1117     /* #else                                 */
1118     /*     #define JSUFFIX ''                */
1119     /* #endif                                */
1120     /* #if 'PotentialAndForce' in KERNEL_VF  */
1121     /*     #define VFSUFFIX  '_VF'           */
1122     /* #elif 'Potential' in KERNEL_VF        */
1123     /*     #define VFSUFFIX '_V'             */
1124     /* #else                                 */
1125     /*     #define VFSUFFIX '_F'             */
1126     /* #endif                                */
1127
1128     /* #if KERNEL_ELEC != 'None' and KERNEL_VDW != 'None' */
1129     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW{ISUFFIX}{JSUFFIX}{VFSUFFIX},outeriter*{OUTERFLOPS} + inneriter*{INNERFLOPS});
1130     /* #elif KERNEL_ELEC != 'None' */
1131     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC{ISUFFIX}{JSUFFIX}{VFSUFFIX},outeriter*{OUTERFLOPS} + inneriter*{INNERFLOPS});
1132     /* #else */
1133     inc_nrnb(nrnb,eNR_NBKERNEL_VDW{ISUFFIX}{JSUFFIX}{VFSUFFIX},outeriter*{OUTERFLOPS} + inneriter*{INNERFLOPS});
1134     /* #endif  */
1135 }