src/gmxlib/nonbonded/nb_kernel_avx_256_single/nb_kernel_template_avx_256_single.pre

   1 /* #if 0 */
   2 #error This file must be processed with the Gromacs pre-preprocessor
   3 /* #endif */
   4 /* #if INCLUDE_HEADER */
   5 #ifdef HAVE_CONFIG_H
   6 #include <config.h>
   7 #endif
   8
   9 #include <math.h>
  10
  11 #include "../nb_kernel.h"
  12 #include "types/simple.h"
  13 #include "vec.h"
  14 #include "nrnb.h"
  15
  16 #include "gmx_math_x86_avx_256_single.h"
  17 #include "kernelutil_x86_avx_256_single.h"
  18 /* #endif */
  19
  20 /* ## List of variables set by the generating script:                                    */
  21 /* ##                                                                                    */
  22 /* ## Setttings that apply to the entire kernel:                                         */
  23 /* ## KERNEL_ELEC:           String, choice for electrostatic interactions               */
  24 /* ## KERNEL_VDW:            String, choice for van der Waals interactions               */
  25 /* ## KERNEL_NAME:           String, name of this kernel                                 */
  26 /* ## KERNEL_VF:             String telling if we calculate potential, force, or both    */
  27 /* ## GEOMETRY_I/GEOMETRY_J: String, name of each geometry, e.g. 'Water3' or '1Particle' */
  28 /* ##                                                                                    */
  29 /* ## Setttings that apply to particles in the outer (I) or inner (J) loops:             */
  30 /* ## PARTICLES_I[]/         Arrays with lists of i/j particles to use in kernel. It is  */
  31 /* ## PARTICLES_J[]:         just [0] for particle geometry, but can be longer for water */
  32 /* ## PARTICLES_ELEC_I[]/    Arrays with lists of i/j particle that have electrostatics  */
  33 /* ## PARTICLES_ELEC_J[]:    interactions that should be calculated in this kernel.      */
  34 /* ## PARTICLES_VDW_I[]/     Arrays with the list of i/j particle that have VdW          */
  35 /* ## PARTICLES_VDW_J[]:     interactions that should be calculated in this kernel.      */
  36 /* ##                                                                                    */
  37 /* ## Setttings for pairs of interactions (e.g. 2nd i particle against 1st j particle)   */
  38 /* ## PAIRS_IJ[]:            Array with (i,j) tuples of pairs for which interactions     */
  39 /* ##                        should be calculated in this kernel. Zero-charge particles  */
  40 /* ##                        do not have interactions with particles without vdw, and    */
  41 /* ##                        Vdw-only interactions are not evaluated in a no-vdw-kernel. */
  42 /* ## INTERACTION_FLAGS[][]: 2D matrix, dimension e.g. 3*3 for water-water interactions. */
  43 /* ##                        For each i-j pair, the element [I][J] is a list of strings  */
  44 /* ##                        defining properties/flags of this interaction. Examples     */
  45 /* ##                        include 'electrostatics'/'vdw' if that type of interaction  */
  46 /* ##                        should be evaluated, 'rsq'/'rinv'/'rinvsq' if those values  */
  47 /* ##                        are needed, and 'exactcutoff' or 'shift','switch' to        */
  48 /* ##                        decide if the force/potential should be modified. This way  */
  49 /* ##                        we only calculate values absolutely needed for each case.   */
  50
  51 /* ## Calculate the size and offset for (merged/interleaved) table data */
  52
  53 /*
  54  * Gromacs nonbonded kernel:   {KERNEL_NAME}
  55  * Electrostatics interaction: {KERNEL_ELEC}
  56  * VdW interaction:            {KERNEL_VDW}
  57  * Geometry:                   {GEOMETRY_I}-{GEOMETRY_J}
  58  * Calculate force/pot:        {KERNEL_VF}
  59  */
  60 void
  61 {KERNEL_NAME}
  62                     (t_nblist * gmx_restrict                nlist,
  63                      rvec * gmx_restrict                    xx,
  64                      rvec * gmx_restrict                    ff,
  65                      t_forcerec * gmx_restrict              fr,
  66                      t_mdatoms * gmx_restrict               mdatoms,
  67                      nb_kernel_data_t * gmx_restrict        kernel_data,
  68                      t_nrnb * gmx_restrict                  nrnb)
  69 {
  70     /* ## Not all variables are used for all kernels, but any optimizing compiler fixes that, */
  71     /* ## so there is no point in going to extremes to exclude variables that are not needed. */
  72     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
  73      * just 0 for non-waters.
  74      * Suffixes A,B,C,D,E,F,G,H refer to j loop unrolling done with AVX, e.g. for the eight different
  75      * jnr indices corresponding to data put in the four positions in the SIMD register.
  76      */
  77     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
  78     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
  79     int              jnrA,jnrB,jnrC,jnrD;
  80     int              jnrE,jnrF,jnrG,jnrH;
  81     int              jnrlistA,jnrlistB,jnrlistC,jnrlistD;
  82     int              jnrlistE,jnrlistF,jnrlistG,jnrlistH;
  83     int              j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
  84     int              j_coord_offsetE,j_coord_offsetF,j_coord_offsetG,j_coord_offsetH;
  85     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  86     real             rcutoff_scalar;
  87     real             *shiftvec,*fshift,*x,*f;
  88     real             *fjptrA,*fjptrB,*fjptrC,*fjptrD,*fjptrE,*fjptrF,*fjptrG,*fjptrH;
  89     real             scratch[4*DIM];
  90     __m256           tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
  91     /* #for I in PARTICLES_I */
  92     real *           vdwioffsetptr{I};
  93     __m256           ix{I},iy{I},iz{I},fix{I},fiy{I},fiz{I},iq{I},isai{I};
  94     /* #endfor */
  95     /* #for J in PARTICLES_J */
  96     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;
  97     __m256           jx{J},jy{J},jz{J},fjx{J},fjy{J},fjz{J},jq{J},isaj{J};
  98     /* #endfor */
  99     /* #for I,J in PAIRS_IJ */
 100     __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};
 101     /* #endfor */
 102     /* #if KERNEL_ELEC != 'None' */
 103     __m256           velec,felec,velecsum,facel,crf,krf,krf2;
 104     real             *charge;
 105     /* #endif */
 106     /* #if 'GeneralizedBorn' in KERNEL_ELEC */
 107     __m256i          gbitab;
 108     __m128i          gbitab_lo,gbitab_hi;
 109     __m256           vgb,fgb,vgbsum,dvdasum,gbscale,gbtabscale,isaprod,gbqqfactor,gbinvepsdiff,gbeps,dvdatmp;
 110     __m256           minushalf = _mm256_set1_ps(-0.5);
 111     real             *invsqrta,*dvda,*gbtab;
 112     /* #endif */
 113     /* #if KERNEL_VDW != 'None' */
 114     int              nvdwtype;
 115     __m256           rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
 116     int              *vdwtype;
 117     real             *vdwparam;
 118     __m256           one_sixth   = _mm256_set1_ps(1.0/6.0);
 119     __m256           one_twelfth = _mm256_set1_ps(1.0/12.0);
 120     /* #endif */
 121     /* #if 'Table' in KERNEL_ELEC or 'GeneralizedBorn' in KERNEL_ELEC or 'Table' in KERNEL_VDW */
 122     __m256i          vfitab;
 123     __m128i          vfitab_lo,vfitab_hi;
 124     __m128i          ifour       = _mm_set1_epi32(4);
 125     __m256           rt,vfeps,vftabscale,Y,F,G,H,Heps,Fp,VV,FF;
 126     real             *vftab;
 127     /* #endif */
 128     /* #if 'Ewald' in KERNEL_ELEC */
 129     __m256i          ewitab;
 130     __m128i          ewitab_lo,ewitab_hi;
 131     __m256           ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
 132     __m256           beta,beta2,beta3,zeta2,pmecorrF,pmecorrV,rinv3;
 133     real             *ewtab;
 134     /* #endif */
 135     /* #if 'PotentialSwitch' in [KERNEL_MOD_ELEC,KERNEL_MOD_VDW] */
 136     __m256           rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
 137     real             rswitch_scalar,d_scalar;
 138     /* #endif */
 139     __m256           dummy_mask,cutoff_mask;
 140     __m256           signbit = _mm256_castsi256_ps( _mm256_set1_epi32(0x80000000) );
 141     __m256           one     = _mm256_set1_ps(1.0);
 142     __m256           two     = _mm256_set1_ps(2.0);
 143     x                = xx[0];
 144     f                = ff[0];
 145
 146     nri              = nlist->nri;
 147     iinr             = nlist->iinr;
 148     jindex           = nlist->jindex;
 149     jjnr             = nlist->jjnr;
 150     shiftidx         = nlist->shift;
 151     gid              = nlist->gid;
 152     shiftvec         = fr->shift_vec[0];
 153     fshift           = fr->fshift[0];
 154     /* #if KERNEL_ELEC != 'None' */
 155     facel            = _mm256_set1_ps(fr->epsfac);
 156     charge           = mdatoms->chargeA;
 157     /*     #if 'ReactionField' in KERNEL_ELEC */
 158     krf              = _mm256_set1_ps(fr->ic->k_rf);
 159     krf2             = _mm256_set1_ps(fr->ic->k_rf*2.0);
 160     crf              = _mm256_set1_ps(fr->ic->c_rf);
 161     /*     #endif */
 162     /* #endif */
 163     /* #if KERNEL_VDW != 'None' */
 164     nvdwtype         = fr->ntype;
 165     vdwparam         = fr->nbfp;
 166     vdwtype          = mdatoms->typeA;
 167     /* #endif */
 168
 169     /* #if 'Table' in KERNEL_ELEC and 'Table' in KERNEL_VDW */
 170     vftab            = kernel_data->table_elec_vdw->data;
 171     vftabscale       = _mm256_set1_ps(kernel_data->table_elec_vdw->scale);
 172     /* #elif 'Table' in KERNEL_ELEC */
 173     vftab            = kernel_data->table_elec->data;
 174     vftabscale       = _mm256_set1_ps(kernel_data->table_elec->scale);
 175     /* #elif 'Table' in KERNEL_VDW */
 176     vftab            = kernel_data->table_vdw->data;
 177     vftabscale       = _mm256_set1_ps(kernel_data->table_vdw->scale);
 178     /* #endif */
 179
 180     /* #if 'Ewald' in KERNEL_ELEC */
 181     sh_ewald         = _mm256_set1_ps(fr->ic->sh_ewald);
 182     beta             = _mm256_set1_ps(fr->ic->ewaldcoeff);
 183     beta2            = _mm256_mul_ps(beta,beta);
 184     beta3            = _mm256_mul_ps(beta,beta2);
 185
 186     /*     #if KERNEL_VF=='Force' and KERNEL_MOD_ELEC!='PotentialSwitch' */
 187     ewtab            = fr->ic->tabq_coul_F;
 188     ewtabscale       = _mm256_set1_ps(fr->ic->tabq_scale);
 189     ewtabhalfspace   = _mm256_set1_ps(0.5/fr->ic->tabq_scale);
 190     /*     #else */
 191     ewtab            = fr->ic->tabq_coul_FDV0;
 192     ewtabscale       = _mm256_set1_ps(fr->ic->tabq_scale);
 193     ewtabhalfspace   = _mm256_set1_ps(0.5/fr->ic->tabq_scale);
 194      /*     #endif */
 195     /* #endif */
 196
 197     /* #if KERNEL_ELEC=='GeneralizedBorn' */
 198     invsqrta         = fr->invsqrta;
 199     dvda             = fr->dvda;
 200     gbtabscale       = _mm256_set1_ps(fr->gbtab.scale);
 201     gbtab            = fr->gbtab.data;
 202     gbinvepsdiff     = _mm256_set1_ps((1.0/fr->epsilon_r) - (1.0/fr->gb_epsilon_solvent));
 203     /* #endif */
 204
 205     /* #if 'Water' in GEOMETRY_I */
 206     /* Setup water-specific parameters */
 207     inr              = nlist->iinr[0];
 208     /*     #for I in PARTICLES_ELEC_I */
 209     iq{I}              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+{I}]));
 210     /*     #endfor */
 211     /*     #for I in PARTICLES_VDW_I */
 212     vdwioffsetptr{I}   = vdwparam+2*nvdwtype*vdwtype[inr+{I}];
 213     /*     #endfor */
 214     /* #endif */
 215
 216     /* #if 'Water' in GEOMETRY_J */
 217     /*     #for J in PARTICLES_ELEC_J */
 218     jq{J}              = _mm256_set1_ps(charge[inr+{J}]);
 219     /*     #endfor */
 220     /*     #for J in PARTICLES_VDW_J */
 221     vdwjidx{J}A        = 2*vdwtype[inr+{J}];
 222     /*     #endfor */
 223     /*     #for I,J in PAIRS_IJ */
 224     /*         #if 'electrostatics' in INTERACTION_FLAGS[I][J] */
 225     qq{I}{J}             = _mm256_mul_ps(iq{I},jq{J});
 226     /*         #endif */
 227     /*         #if 'vdw' in INTERACTION_FLAGS[I][J] */
 228     c6_{I}{J}            = _mm256_set1_ps(vdwioffsetptr{I}[vdwjidx{J}A]);
 229     c12_{I}{J}           = _mm256_set1_ps(vdwioffsetptr{I}[vdwjidx{J}A+1]);
 230     /*         #endif */
 231     /*     #endfor */
 232     /* #endif */
 233
 234     /* #if KERNEL_MOD_ELEC!='None' or KERNEL_MOD_VDW!='None' */
 235     /*     #if KERNEL_ELEC!='None' */
 236     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 237     rcutoff_scalar   = fr->rcoulomb;
 238     /*     #else */
 239     rcutoff_scalar   = fr->rvdw;
 240     /*     #endif */
 241     rcutoff          = _mm256_set1_ps(rcutoff_scalar);
 242     rcutoff2         = _mm256_mul_ps(rcutoff,rcutoff);
 243     /* #endif */
 244
 245     /* #if KERNEL_MOD_VDW=='PotentialShift' */
 246     sh_vdw_invrcut6  = _mm256_set1_ps(fr->ic->sh_invrc6);
 247     rvdw             = _mm256_set1_ps(fr->rvdw);
 248     /* #endif */
 249
 250     /* #if 'PotentialSwitch' in [KERNEL_MOD_ELEC,KERNEL_MOD_VDW] */
 251     /*     #if KERNEL_MOD_ELEC=='PotentialSwitch'  */
 252     rswitch_scalar   = fr->rcoulomb_switch;
 253     rswitch          = _mm256_set1_ps(rswitch_scalar);
 254     /*     #else */
 255     rswitch_scalar   = fr->rvdw_switch;
 256     rswitch          = _mm256_set1_ps(rswitch_scalar);
 257     /*     #endif */
 258     /* Setup switch parameters */
 259     d_scalar         = rcutoff_scalar-rswitch_scalar;
 260     d                = _mm256_set1_ps(d_scalar);
 261     swV3             = _mm256_set1_ps(-10.0/(d_scalar*d_scalar*d_scalar));
 262     swV4             = _mm256_set1_ps( 15.0/(d_scalar*d_scalar*d_scalar*d_scalar));
 263     swV5             = _mm256_set1_ps( -6.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
 264     /*     #if 'Force' in KERNEL_VF */
 265     swF2             = _mm256_set1_ps(-30.0/(d_scalar*d_scalar*d_scalar));
 266     swF3             = _mm256_set1_ps( 60.0/(d_scalar*d_scalar*d_scalar*d_scalar));
 267     swF4             = _mm256_set1_ps(-30.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
 268     /*     #endif */
 269     /* #endif */
 270
 271     /* Avoid stupid compiler warnings */
 272     jnrA = jnrB = jnrC = jnrD = jnrE = jnrF = jnrG = jnrH = 0;
 273     j_coord_offsetA = 0;
 274     j_coord_offsetB = 0;
 275     j_coord_offsetC = 0;
 276     j_coord_offsetD = 0;
 277     j_coord_offsetE = 0;
 278     j_coord_offsetF = 0;
 279     j_coord_offsetG = 0;
 280     j_coord_offsetH = 0;
 281
 282     /* ## Keep track of the floating point operations we issue for reporting! */
 283     /* #define OUTERFLOPS 0 */
 284     outeriter        = 0;
 285     inneriter        = 0;
 286
 287     for(iidx=0;iidx<4*DIM;iidx++)
 288     {
 289         scratch[iidx] = 0.0;
 290     }
 291
 292     /* Start outer loop over neighborlists */
 293     for(iidx=0; iidx<nri; iidx++)
 294     {
 295         /* Load shift vector for this list */
 296         i_shift_offset   = DIM*shiftidx[iidx];
 297
 298         /* Load limits for loop over neighbors */
 299         j_index_start    = jindex[iidx];
 300         j_index_end      = jindex[iidx+1];
 301
 302         /* Get outer coordinate index */
 303         inr              = iinr[iidx];
 304         i_coord_offset   = DIM*inr;
 305
 306         /* Load i particle coords and add shift vector */
 307         /* #if GEOMETRY_I == 'Particle' */
 308         gmx_mm256_load_shift_and_1rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,&ix0,&iy0,&iz0);
 309         /* #elif GEOMETRY_I == 'Water3' */
 310         gmx_mm256_load_shift_and_3rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,
 311                                                     &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2);
 312         /* #elif GEOMETRY_I == 'Water4' */
 313         /*     #if 0 in PARTICLES_I                 */
 314         gmx_mm256_load_shift_and_4rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,
 315                                                     &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
 316         /*     #else                                */
 317         gmx_mm256_load_shift_and_3rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset+DIM,
 318                                                     &ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
 319         /*     #endif                               */
 320         /* #endif                                   */
 321
 322         /* #if 'Force' in KERNEL_VF */
 323         /*     #for I in PARTICLES_I */
 324         fix{I}             = _mm256_setzero_ps();
 325         fiy{I}             = _mm256_setzero_ps();
 326         fiz{I}             = _mm256_setzero_ps();
 327         /*     #endfor */
 328         /* #endif */
 329
 330         /* ## For water we already preloaded parameters at the start of the kernel */
 331         /* #if not 'Water' in GEOMETRY_I */
 332         /* Load parameters for i particles */
 333         /*     #for I in PARTICLES_ELEC_I */
 334         iq{I}              = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+{I}]));
 335         /*         #define OUTERFLOPS OUTERFLOPS+1 */
 336         /*         #if KERNEL_ELEC=='GeneralizedBorn' */
 337         isai{I}            = _mm256_set1_ps(invsqrta[inr+{I}]);
 338         /*         #endif */
 339         /*     #endfor */
 340         /*     #for I in PARTICLES_VDW_I */
 341         vdwioffsetptr{I}   = vdwparam+2*nvdwtype*vdwtype[inr+{I}];
 342         /*     #endfor */
 343         /* #endif */
 344
 345         /* #if 'Potential' in KERNEL_VF */
 346         /* Reset potential sums */
 347         /*     #if KERNEL_ELEC != 'None' */
 348         velecsum         = _mm256_setzero_ps();
 349         /*     #endif */
 350         /*     #if 'GeneralizedBorn' in KERNEL_ELEC */
 351         vgbsum           = _mm256_setzero_ps();
 352         /*     #endif */
 353         /*     #if KERNEL_VDW != 'None' */
 354         vvdwsum          = _mm256_setzero_ps();
 355         /*     #endif */
 356         /* #endif */
 357         /*     #if 'GeneralizedBorn' in KERNEL_ELEC and 'Force' in KERNEL_VF */
 358         dvdasum          = _mm256_setzero_ps();
 359         /*     #endif */
 360
 361         /* #for ROUND in ['Loop','Epilogue'] */
 362
 363         /* #if ROUND =='Loop' */
 364         /* Start inner kernel loop */
 365         for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+7]>=0; jidx+=8)
 366         {
 367         /* ## First round is normal loop (next statement resets indentation) */
 368         /*     #if 0 */
 369         }
 370         /*     #endif */
 371         /* #else */
 372         if(jidx<j_index_end)
 373         {
 374         /* ## Second round is epilogue */
 375         /* #endif */
 376         /* #define INNERFLOPS 0 */
 377
 378             /* Get j neighbor index, and coordinate index */
 379             /* #if ROUND =='Loop' */
 380             jnrA             = jjnr[jidx];
 381             jnrB             = jjnr[jidx+1];
 382             jnrC             = jjnr[jidx+2];
 383             jnrD             = jjnr[jidx+3];
 384             jnrE             = jjnr[jidx+4];
 385             jnrF             = jjnr[jidx+5];
 386             jnrG             = jjnr[jidx+6];
 387             jnrH             = jjnr[jidx+7];
 388             /* #else */
 389             jnrlistA         = jjnr[jidx];
 390             jnrlistB         = jjnr[jidx+1];
 391             jnrlistC         = jjnr[jidx+2];
 392             jnrlistD         = jjnr[jidx+3];
 393             jnrlistE         = jjnr[jidx+4];
 394             jnrlistF         = jjnr[jidx+5];
 395             jnrlistG         = jjnr[jidx+6];
 396             jnrlistH         = jjnr[jidx+7];
 397             /* Sign of each element will be negative for non-real atoms.
 398              * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
 399              * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
 400              */
 401             dummy_mask = gmx_mm256_set_m128(gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx+4)),_mm_setzero_si128())),
 402                                             gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128())));
 403
 404             jnrA       = (jnrlistA>=0) ? jnrlistA : 0;
 405             jnrB       = (jnrlistB>=0) ? jnrlistB : 0;
 406             jnrC       = (jnrlistC>=0) ? jnrlistC : 0;
 407             jnrD       = (jnrlistD>=0) ? jnrlistD : 0;
 408             jnrE       = (jnrlistE>=0) ? jnrlistE : 0;
 409             jnrF       = (jnrlistF>=0) ? jnrlistF : 0;
 410             jnrG       = (jnrlistG>=0) ? jnrlistG : 0;
 411             jnrH       = (jnrlistH>=0) ? jnrlistH : 0;
 412             /* #endif */
 413             j_coord_offsetA  = DIM*jnrA;
 414             j_coord_offsetB  = DIM*jnrB;
 415             j_coord_offsetC  = DIM*jnrC;
 416             j_coord_offsetD  = DIM*jnrD;
 417             j_coord_offsetE  = DIM*jnrE;
 418             j_coord_offsetF  = DIM*jnrF;
 419             j_coord_offsetG  = DIM*jnrG;
 420             j_coord_offsetH  = DIM*jnrH;
 421
 422             /* load j atom coordinates */
 423             /* #if GEOMETRY_J == 'Particle'             */
 424             gmx_mm256_load_1rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 425                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 426                                                  x+j_coord_offsetE,x+j_coord_offsetF,
 427                                                  x+j_coord_offsetG,x+j_coord_offsetH,
 428                                                  &jx0,&jy0,&jz0);
 429             /* #elif GEOMETRY_J == 'Water3'             */
 430             gmx_mm256_load_3rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 431                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 432                                                  x+j_coord_offsetE,x+j_coord_offsetF,
 433                                                  x+j_coord_offsetG,x+j_coord_offsetH,
 434                                               &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,&jy2,&jz2);
 435             /* #elif GEOMETRY_J == 'Water4'             */
 436             /*     #if 0 in PARTICLES_J                 */
 437             gmx_mm256_load_4rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
 438                                                  x+j_coord_offsetC,x+j_coord_offsetD,
 439                                                  x+j_coord_offsetE,x+j_coord_offsetF,
 440                                                  x+j_coord_offsetG,x+j_coord_offsetH,
 441                                                  &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,
 442                                                  &jy2,&jz2,&jx3,&jy3,&jz3);
 443             /*     #else                                */
 444             gmx_mm256_load_3rvec_8ptr_swizzle_ps(x+j_coord_offsetA+DIM,x+j_coord_offsetB+DIM,
 445                                                  x+j_coord_offsetC+DIM,x+j_coord_offsetD+DIM,
 446                                                  x+j_coord_offsetE+DIM,x+j_coord_offsetF+DIM,
 447                                                  x+j_coord_offsetG+DIM,x+j_coord_offsetH+DIM,
 448                                                  &jx1,&jy1,&jz1,&jx2,&jy2,&jz2,&jx3,&jy3,&jz3);
 449             /*     #endif                               */
 450             /* #endif                                   */
 451
 452             /* Calculate displacement vector */
 453             /* #for I,J in PAIRS_IJ */
 454             dx{I}{J}             = _mm256_sub_ps(ix{I},jx{J});
 455             dy{I}{J}             = _mm256_sub_ps(iy{I},jy{J});
 456             dz{I}{J}             = _mm256_sub_ps(iz{I},jz{J});
 457             /*     #define INNERFLOPS INNERFLOPS+3 */
 458             /* #endfor */
 459
 460             /* Calculate squared distance and things based on it */
 461             /* #for I,J in PAIRS_IJ */
 462             rsq{I}{J}            = gmx_mm256_calc_rsq_ps(dx{I}{J},dy{I}{J},dz{I}{J});
 463             /*     #define INNERFLOPS INNERFLOPS+5 */
 464             /* #endfor */
 465
 466             /* #for I,J in PAIRS_IJ */
 467             /*     #if 'rinv' in INTERACTION_FLAGS[I][J] */
 468             rinv{I}{J}           = gmx_mm256_invsqrt_ps(rsq{I}{J});
 469             /*         #define INNERFLOPS INNERFLOPS+5 */
 470             /*     #endif */
 471             /* #endfor */
 472
 473             /* #for I,J in PAIRS_IJ */
 474             /*     #if 'rinvsq' in INTERACTION_FLAGS[I][J] */
 475             /*         # if 'rinv' not in INTERACTION_FLAGS[I][J] */
 476             rinvsq{I}{J}         = gmx_mm256_inv_ps(rsq{I}{J});
 477             /*             #define INNERFLOPS INNERFLOPS+4 */
 478             /*         #else */
 479             rinvsq{I}{J}         = _mm256_mul_ps(rinv{I}{J},rinv{I}{J});
 480             /*             #define INNERFLOPS INNERFLOPS+1 */
 481             /*         #endif */
 482             /*     #endif */
 483             /* #endfor */
 484
 485             /* #if not 'Water' in GEOMETRY_J */
 486             /* Load parameters for j particles */
 487             /*     #for J in PARTICLES_ELEC_J */
 488             jq{J}              = gmx_mm256_load_8real_swizzle_ps(charge+jnrA+{J},charge+jnrB+{J},
 489                                                                  charge+jnrC+{J},charge+jnrD+{J},
 490                                                                  charge+jnrE+{J},charge+jnrF+{J},
 491                                                                  charge+jnrG+{J},charge+jnrH+{J});
 492             /*         #if KERNEL_ELEC=='GeneralizedBorn' */
 493             isaj{J}            = gmx_mm256_load_8real_swizzle_ps(invsqrta+jnrA+{J},invsqrta+jnrB+{J},
 494                                                                  invsqrta+jnrC+{J},invsqrta+jnrD+{J},
 495                                                                  invsqrta+jnrE+{J},invsqrta+jnrF+{J},
 496                                                                  invsqrta+jnrG+{J},invsqrta+jnrH+{J});
 497             /*         #endif */
 498             /*     #endfor */
 499             /*     #for J in PARTICLES_VDW_J */
 500             vdwjidx{J}A        = 2*vdwtype[jnrA+{J}];
 501             vdwjidx{J}B        = 2*vdwtype[jnrB+{J}];
 502             vdwjidx{J}C        = 2*vdwtype[jnrC+{J}];
 503             vdwjidx{J}D        = 2*vdwtype[jnrD+{J}];
 504             vdwjidx{J}E        = 2*vdwtype[jnrE+{J}];
 505             vdwjidx{J}F        = 2*vdwtype[jnrF+{J}];
 506             vdwjidx{J}G        = 2*vdwtype[jnrG+{J}];
 507             vdwjidx{J}H        = 2*vdwtype[jnrH+{J}];
 508             /*     #endfor */
 509             /* #endif */
 510
 511             /* #if 'Force' in KERNEL_VF and not 'Particle' in GEOMETRY_I */
 512             /*     #for J in PARTICLES_J */
 513             fjx{J}             = _mm256_setzero_ps();
 514             fjy{J}             = _mm256_setzero_ps();
 515             fjz{J}             = _mm256_setzero_ps();
 516             /*     #endfor */
 517             /* #endif */
 518
 519             /* #for I,J in PAIRS_IJ */
 520
 521             /**************************
 522              * CALCULATE INTERACTIONS *
 523              **************************/
 524
 525             /*     ## Note special check for TIP4P-TIP4P. Since we are cutting of all hydrogen interactions we also cut the LJ-only O-O interaction */
 526             /*     #if 'exactcutoff' in INTERACTION_FLAGS[I][J] or (GEOMETRY_I=='Water4' and GEOMETRY_J=='Water4' and 'exactcutoff' in INTERACTION_FLAGS[1][1]) */
 527             /*         ## We always calculate rinv/rinvsq above to enable pipelineing in compilers (performance tested on x86) */
 528             if (gmx_mm256_any_lt(rsq{I}{J},rcutoff2))
 529             {
 530                 /*     #if 0    ## this and the next two lines is a hack to maintain auto-indentation in template file */
 531             }
 532             /*         #endif */
 533             /*         #define INNERFLOPS INNERFLOPS+1 */
 534             /*     #endif */
 535
 536             /*     #if 'r' in INTERACTION_FLAGS[I][J] */
 537             r{I}{J}              = _mm256_mul_ps(rsq{I}{J},rinv{I}{J});
 538             /*         #if ROUND == 'Epilogue' */
 539             r{I}{J}              = _mm256_andnot_ps(dummy_mask,r{I}{J});
 540             /*             #define INNERFLOPS INNERFLOPS+1 */
 541             /*         #endif */
 542             /*         #define INNERFLOPS INNERFLOPS+1 */
 543             /*     #endif */
 544
 545             /*     ## For water geometries we already loaded parameters at the start of the kernel */
 546             /*     #if not 'Water' in GEOMETRY_J */
 547             /* Compute parameters for interactions between i and j atoms */
 548             /*         #if 'electrostatics' in INTERACTION_FLAGS[I][J] */
 549             qq{I}{J}             = _mm256_mul_ps(iq{I},jq{J});
 550             /*             #define INNERFLOPS INNERFLOPS+1 */
 551             /*         #endif */
 552             /*         #if 'vdw' in INTERACTION_FLAGS[I][J] */
 553             gmx_mm256_load_8pair_swizzle_ps(vdwioffsetptr{I}+vdwjidx{J}A,
 554                                             vdwioffsetptr{I}+vdwjidx{J}B,
 555                                             vdwioffsetptr{I}+vdwjidx{J}C,
 556                                             vdwioffsetptr{I}+vdwjidx{J}D,
 557                                             vdwioffsetptr{I}+vdwjidx{J}E,
 558                                             vdwioffsetptr{I}+vdwjidx{J}F,
 559                                             vdwioffsetptr{I}+vdwjidx{J}G,
 560                                             vdwioffsetptr{I}+vdwjidx{J}H,
 561                                             &c6_{I}{J},&c12_{I}{J});
 562             /*         #endif */
 563             /*     #endif */
 564
 565             /*     #if 'table' in INTERACTION_FLAGS[I][J] */
 566             /* Calculate table index by multiplying r with table scale and truncate to integer */
 567             rt               = _mm256_mul_ps(r{I}{J},vftabscale);
 568             vfitab           = _mm256_cvttps_epi32(rt);
 569             vfeps            = _mm256_sub_ps(rt,_mm256_round_ps(rt, _MM_FROUND_FLOOR));
 570             /*         #define INNERFLOPS INNERFLOPS+4                          */
 571             /*         AVX1 does not support 256-bit integer operations, so now we go to 128-bit mode... */
 572             vfitab_lo        = _mm256_extractf128_si256(vfitab,0x0);
 573             vfitab_hi        = _mm256_extractf128_si256(vfitab,0x1);
 574             /*         #if 'Table' in KERNEL_ELEC and 'Table' in KERNEL_VDW     */
 575             /*             ## 3 tables, 4 bytes per point: multiply index by 12 */
 576             vfitab_lo        = _mm_slli_epi32(_mm_add_epi32(vfitab_lo,_mm_slli_epi32(vfitab_lo,1)),2);
 577             vfitab_hi        = _mm_slli_epi32(_mm_add_epi32(vfitab_hi,_mm_slli_epi32(vfitab_hi,1)),2);
 578             /*         #elif 'Table' in KERNEL_ELEC                             */
 579             /*             ## 1 table, 4 bytes per point: multiply index by 4   */
 580             vfitab_lo        = _mm_slli_epi32(vfitab_lo,2);
 581             vfitab_hi        = _mm_slli_epi32(vfitab_hi,2);
 582             /*         #elif 'Table' in KERNEL_VDW                              */
 583             /*             ## 2 tables, 4 bytes per point: multiply index by 8  */
 584             vfitab_lo        = _mm_slli_epi32(vfitab_lo,3);
 585             vfitab_hi        = _mm_slli_epi32(vfitab_hi,3);
 586             /*         #endif                                                   */
 587             /*     #endif */
 588
 589             /*     ## ELECTROSTATIC INTERACTIONS */
 590             /*     #if 'electrostatics' in INTERACTION_FLAGS[I][J] */
 591
 592             /*         #if KERNEL_ELEC=='Coulomb' */
 593
 594             /* COULOMB ELECTROSTATICS */
 595             velec            = _mm256_mul_ps(qq{I}{J},rinv{I}{J});
 596             /*             #define INNERFLOPS INNERFLOPS+1 */
 597             /*             #if 'Force' in KERNEL_VF */
 598             felec            = _mm256_mul_ps(velec,rinvsq{I}{J});
 599             /*                 #define INNERFLOPS INNERFLOPS+1 */
 600             /*             #endif */
 601
 602             /*         #elif KERNEL_ELEC=='ReactionField' */
 603
 604             /* REACTION-FIELD ELECTROSTATICS */
 605             /*             #if 'Potential' in KERNEL_VF */
 606             velec            = _mm256_mul_ps(qq{I}{J},_mm256_sub_ps(_mm256_add_ps(rinv{I}{J},_mm256_mul_ps(krf,rsq{I}{J})),crf));
 607             /*                 #define INNERFLOPS INNERFLOPS+4 */
 608             /*             #endif */
 609             /*             #if 'Force' in KERNEL_VF */
 610             felec            = _mm256_mul_ps(qq{I}{J},_mm256_sub_ps(_mm256_mul_ps(rinv{I}{J},rinvsq{I}{J}),krf2));
 611             /*                 #define INNERFLOPS INNERFLOPS+3 */
 612             /*             #endif */
 613
 614             /*         #elif KERNEL_ELEC=='GeneralizedBorn' */
 615
 616             /* GENERALIZED BORN AND COULOMB ELECTROSTATICS */
 617             isaprod          = _mm256_mul_ps(isai{I},isaj{J});
 618             gbqqfactor       = _mm256_xor_ps(signbit,_mm256_mul_ps(qq{I}{J},_mm256_mul_ps(isaprod,gbinvepsdiff)));
 619             gbscale          = _mm256_mul_ps(isaprod,gbtabscale);
 620             /*             #define INNERFLOPS INNERFLOPS+5 */
 621
 622             /* Calculate generalized born table index - this is a separate table from the normal one,
 623              * but we use the same procedure by multiplying r with scale and truncating to integer.
 624              */
 625             rt               = _mm256_mul_ps(r{I}{J},gbscale);
 626             gbitab           = _mm256_cvttps_epi32(rt);
 627             gbeps            = _mm256_sub_ps(rt,_mm256_round_ps(rt, _MM_FROUND_FLOOR));
 628             /*         AVX1 does not support 256-bit integer operations, so now we go to 128-bit mode... */
 629             gbitab_lo        = _mm256_extractf128_si256(gbitab,0x0);
 630             gbitab_hi        = _mm256_extractf128_si256(gbitab,0x1);
 631             gbitab_lo        = _mm_slli_epi32(gbitab_lo,2);
 632             gbitab_hi        = _mm_slli_epi32(gbitab_hi,2);
 633             Y                = gmx_mm256_set_m128(_mm_load_ps(gbtab + _mm_extract_epi32(gbitab_hi,0)),
 634                                                   _mm_load_ps(gbtab + _mm_extract_epi32(gbitab_lo,0)));
 635             F                = gmx_mm256_set_m128(_mm_load_ps(gbtab + _mm_extract_epi32(gbitab_hi,1)),
 636                                                   _mm_load_ps(gbtab + _mm_extract_epi32(gbitab_lo,1)));
 637             G                = gmx_mm256_set_m128(_mm_load_ps(gbtab + _mm_extract_epi32(gbitab_hi,2)),
 638                                                   _mm_load_ps(gbtab + _mm_extract_epi32(gbitab_lo,2)));
 639             H                = gmx_mm256_set_m128(_mm_load_ps(gbtab + _mm_extract_epi32(gbitab_hi,3)),
 640                                                   _mm_load_ps(gbtab + _mm_extract_epi32(gbitab_lo,3)));
 641             GMX_MM256_HALFTRANSPOSE4_PS(Y,F,G,H);
 642             Heps             = _mm256_mul_ps(gbeps,H);
 643             Fp               = _mm256_add_ps(F,_mm256_mul_ps(gbeps,_mm256_add_ps(G,Heps)));
 644             VV               = _mm256_add_ps(Y,_mm256_mul_ps(gbeps,Fp));
 645             vgb              = _mm256_mul_ps(gbqqfactor,VV);
 646             /*             #define INNERFLOPS INNERFLOPS+10 */
 647
 648             /*             #if 'Force' in KERNEL_VF */
 649             FF               = _mm256_add_ps(Fp,_mm256_mul_ps(gbeps,_mm256_add_ps(G,_mm256_add_ps(Heps,Heps))));
 650             fgb              = _mm256_mul_ps(gbqqfactor,_mm256_mul_ps(FF,gbscale));
 651             dvdatmp          = _mm256_mul_ps(minushalf,_mm256_add_ps(vgb,_mm256_mul_ps(fgb,r{I}{J})));
 652             dvdasum          = _mm256_add_ps(dvdasum,dvdatmp);
 653             /*                 #if ROUND == 'Loop' */
 654             fjptrA           = dvda+jnrA;
 655             fjptrB           = dvda+jnrB;
 656             fjptrC           = dvda+jnrC;
 657             fjptrD           = dvda+jnrD;
 658             fjptrE           = dvda+jnrE;
 659             fjptrF           = dvda+jnrF;
 660             fjptrG           = dvda+jnrG;
 661             fjptrH           = dvda+jnrH;
 662             /*                 #else */
 663             /* 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. */
 664             fjptrA             = (jnrlistA>=0) ? dvda+jnrA : scratch;
 665             fjptrB             = (jnrlistB>=0) ? dvda+jnrB : scratch;
 666             fjptrC             = (jnrlistC>=0) ? dvda+jnrC : scratch;
 667             fjptrD             = (jnrlistD>=0) ? dvda+jnrD : scratch;
 668             fjptrE             = (jnrlistE>=0) ? dvda+jnrE : scratch;
 669             fjptrF             = (jnrlistF>=0) ? dvda+jnrF : scratch;
 670             fjptrG             = (jnrlistG>=0) ? dvda+jnrG : scratch;
 671             fjptrH             = (jnrlistH>=0) ? dvda+jnrH : scratch;
 672             /*                 #endif */
 673             gmx_mm256_increment_8real_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,
 674                                                  _mm256_mul_ps(dvdatmp,_mm256_mul_ps(isaj{J},isaj{J})));
 675             /*                 #define INNERFLOPS INNERFLOPS+12 */
 676             /*             #endif */
 677             velec            = _mm256_mul_ps(qq{I}{J},rinv{I}{J});
 678             /*                 #define INNERFLOPS INNERFLOPS+1 */
 679             /*             #if 'Force' in KERNEL_VF */
 680             felec            = _mm256_mul_ps(_mm256_sub_ps(_mm256_mul_ps(velec,rinv{I}{J}),fgb),rinv{I}{J});
 681             /*                 #define INNERFLOPS INNERFLOPS+3 */
 682             /*             #endif */
 683
 684             /*         #elif KERNEL_ELEC=='Ewald' */
 685             /* EWALD ELECTROSTATICS */
 686
 687             /* Analytical PME correction */
 688             zeta2            = _mm256_mul_ps(beta2,rsq{I}{J});
 689             /*             #if 'Force' in KERNEL_VF */
 690             rinv3            = _mm256_mul_ps(rinvsq{I}{J},rinv{I}{J});
 691             pmecorrF         = gmx_mm256_pmecorrF_ps(zeta2);
 692             felec            = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
 693             felec            = _mm256_mul_ps(qq{I}{J},felec);
 694             /*                 #define INNERFLOPS INNERFLOPS+31 */
 695             /*             #endif */
 696             /*             #if 'Potential' in KERNEL_VF or KERNEL_MOD_ELEC=='PotentialSwitch' */
 697             pmecorrV         = gmx_mm256_pmecorrV_ps(zeta2);
 698             pmecorrV         = _mm256_mul_ps(pmecorrV,beta);
 699             /*                 #define INNERFLOPS INNERFLOPS+27 */
 700             /*                 #if KERNEL_MOD_ELEC=='PotentialShift' */
 701             velec            = _mm256_sub_ps(_mm256_sub_ps(rinv{I}{J},sh_ewald),pmecorrV);
 702             /*                     #define INNERFLOPS INNERFLOPS+21 */
 703             /*                 #else */
 704             velec            = _mm256_sub_ps(rinv{I}{J},pmecorrV);
 705             /*                 #endif */
 706             velec            = _mm256_mul_ps(qq{I}{J},velec);
 707             /*             #endif */
 708
 709             /*         #elif KERNEL_ELEC=='CubicSplineTable' */
 710
 711             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 712             Y                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,0)),
 713                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,0)));
 714             F                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,1)),
 715                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,1)));
 716             G                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,2)),
 717                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,2)));
 718             H                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,3)),
 719                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,3)));
 720             GMX_MM256_HALFTRANSPOSE4_PS(Y,F,G,H);
 721             Heps             = _mm256_mul_ps(vfeps,H);
 722             Fp               = _mm256_add_ps(F,_mm256_mul_ps(vfeps,_mm256_add_ps(G,Heps)));
 723             /*             #define INNERFLOPS INNERFLOPS+4 */
 724             /*             #if 'Potential' in KERNEL_VF */
 725             VV               = _mm256_add_ps(Y,_mm256_mul_ps(vfeps,Fp));
 726             velec            = _mm256_mul_ps(qq{I}{J},VV);
 727             /*                 #define INNERFLOPS INNERFLOPS+3 */
 728             /*             #endif */
 729             /*             #if 'Force' in KERNEL_VF */
 730             FF               = _mm256_add_ps(Fp,_mm256_mul_ps(vfeps,_mm256_add_ps(G,_mm256_add_ps(Heps,Heps))));
 731             felec            = _mm256_xor_ps(signbit,_mm256_mul_ps(_mm256_mul_ps(qq{I}{J},FF),_mm256_mul_ps(vftabscale,rinv{I}{J})));
 732             /*                 #define INNERFLOPS INNERFLOPS+7 */
 733             /*             #endif */
 734             /*         #endif */
 735             /*         ## End of check for electrostatics interaction forms */
 736             /*     #endif */
 737             /*     ## END OF ELECTROSTATIC INTERACTION CHECK FOR PAIR I-J */
 738
 739             /*     #if 'vdw' in INTERACTION_FLAGS[I][J] */
 740
 741             /*         #if KERNEL_VDW=='LennardJones' */
 742
 743             /* LENNARD-JONES DISPERSION/REPULSION */
 744
 745             rinvsix          = _mm256_mul_ps(_mm256_mul_ps(rinvsq{I}{J},rinvsq{I}{J}),rinvsq{I}{J});
 746             /*             #define INNERFLOPS INNERFLOPS+2 */
 747             /*             #if 'Potential' in KERNEL_VF or KERNEL_MOD_VDW=='PotentialSwitch' */
 748             vvdw6            = _mm256_mul_ps(c6_{I}{J},rinvsix);
 749             vvdw12           = _mm256_mul_ps(c12_{I}{J},_mm256_mul_ps(rinvsix,rinvsix));
 750             /*                 #define INNERFLOPS INNERFLOPS+3 */
 751             /*                 #if KERNEL_MOD_VDW=='PotentialShift' */
 752             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) ,
 753                                           _mm256_mul_ps( _mm256_sub_ps(vvdw6,_mm256_mul_ps(c6_{I}{J},sh_vdw_invrcut6)),one_sixth));
 754             /*                     #define INNERFLOPS INNERFLOPS+8 */
 755             /*                 #else */
 756             vvdw             = _mm256_sub_ps( _mm256_mul_ps(vvdw12,one_twelfth) , _mm256_mul_ps(vvdw6,one_sixth) );
 757             /*                     #define INNERFLOPS INNERFLOPS+3 */
 758             /*                 #endif */
 759             /*                 ## Check for force inside potential check, i.e. this means we already did the potential part */
 760             /*                 #if 'Force' in KERNEL_VF */
 761             fvdw             = _mm256_mul_ps(_mm256_sub_ps(vvdw12,vvdw6),rinvsq{I}{J});
 762             /*                     #define INNERFLOPS INNERFLOPS+2 */
 763             /*                 #endif */
 764             /*             #elif KERNEL_VF=='Force' */
 765             /*                 ## Force-only LennardJones makes it possible to save 1 flop (they do add up...) */
 766             fvdw             = _mm256_mul_ps(_mm256_sub_ps(_mm256_mul_ps(c12_{I}{J},rinvsix),c6_{I}{J}),_mm256_mul_ps(rinvsix,rinvsq{I}{J}));
 767             /*                 #define INNERFLOPS INNERFLOPS+4 */
 768             /*             #endif */
 769
 770             /*         #elif KERNEL_VDW=='CubicSplineTable' */
 771
 772             /* CUBIC SPLINE TABLE DISPERSION */
 773             /*             #if 'Table' in KERNEL_ELEC */
 774             vfitab_lo        = _mm_add_epi32(vfitab_lo,ifour);
 775             vfitab_hi        = _mm_add_epi32(vfitab_hi,ifour);
 776             /*             #endif                     */
 777             Y                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,0)),
 778                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,0)));
 779             F                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,1)),
 780                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,1)));
 781             G                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,2)),
 782                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,2)));
 783             H                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,3)),
 784                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,3)));
 785             GMX_MM256_HALFTRANSPOSE4_PS(Y,F,G,H);
 786             Heps             = _mm256_mul_ps(vfeps,H);
 787             Fp               = _mm256_add_ps(F,_mm256_mul_ps(vfeps,_mm256_add_ps(G,Heps)));
 788             /*             #define INNERFLOPS INNERFLOPS+4 */
 789             /*             #if 'Potential' in KERNEL_VF */
 790             VV               = _mm256_add_ps(Y,_mm256_mul_ps(vfeps,Fp));
 791             vvdw6            = _mm256_mul_ps(c6_{I}{J},VV);
 792             /*                 #define INNERFLOPS INNERFLOPS+3 */
 793             /*             #endif */
 794             /*             #if 'Force' in KERNEL_VF */
 795             FF               = _mm256_add_ps(Fp,_mm256_mul_ps(vfeps,_mm256_add_ps(G,_mm256_add_ps(Heps,Heps))));
 796             fvdw6            = _mm256_mul_ps(c6_{I}{J},FF);
 797             /*                 #define INNERFLOPS INNERFLOPS+4 */
 798             /*             #endif */
 799
 800             /* CUBIC SPLINE TABLE REPULSION */
 801             vfitab_lo        = _mm_add_epi32(vfitab_lo,ifour);
 802             vfitab_hi        = _mm_add_epi32(vfitab_hi,ifour);
 803             Y                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,0)),
 804                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,0)));
 805             F                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,1)),
 806                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,1)));
 807             G                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,2)),
 808                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,2)));
 809             H                = gmx_mm256_set_m128(_mm_load_ps(vftab + _mm_extract_epi32(vfitab_hi,3)),
 810                                                   _mm_load_ps(vftab + _mm_extract_epi32(vfitab_lo,3)));
 811             GMX_MM256_HALFTRANSPOSE4_PS(Y,F,G,H);
 812             Heps             = _mm256_mul_ps(vfeps,H);
 813             Fp               = _mm256_add_ps(F,_mm256_mul_ps(vfeps,_mm256_add_ps(G,Heps)));
 814             /*             #define INNERFLOPS INNERFLOPS+4 */
 815             /*             #if 'Potential' in KERNEL_VF */
 816             VV               = _mm256_add_ps(Y,_mm256_mul_ps(vfeps,Fp));
 817             vvdw12           = _mm256_mul_ps(c12_{I}{J},VV);
 818             /*                 #define INNERFLOPS INNERFLOPS+3 */
 819             /*             #endif */
 820             /*             #if 'Force' in KERNEL_VF */
 821             FF               = _mm256_add_ps(Fp,_mm256_mul_ps(vfeps,_mm256_add_ps(G,_mm256_add_ps(Heps,Heps))));
 822             fvdw12           = _mm256_mul_ps(c12_{I}{J},FF);
 823             /*                 #define INNERFLOPS INNERFLOPS+5 */
 824             /*             #endif */
 825             /*             #if 'Potential' in KERNEL_VF */
 826             vvdw             = _mm256_add_ps(vvdw12,vvdw6);
 827             /*                 #define INNERFLOPS INNERFLOPS+1 */
 828             /*             #endif */
 829             /*             #if 'Force' in KERNEL_VF */
 830             fvdw             = _mm256_xor_ps(signbit,_mm256_mul_ps(_mm256_add_ps(fvdw6,fvdw12),_mm256_mul_ps(vftabscale,rinv{I}{J})));
 831             /*                 #define INNERFLOPS INNERFLOPS+4 */
 832             /*             #endif */
 833             /*         #endif */
 834             /*         ## End of check for vdw interaction forms */
 835             /*     #endif */
 836             /*     ## END OF VDW INTERACTION CHECK FOR PAIR I-J */
 837
 838             /*     #if 'switch' in INTERACTION_FLAGS[I][J] */
 839             d                = _mm256_sub_ps(r{I}{J},rswitch);
 840             d                = _mm256_max_ps(d,_mm256_setzero_ps());
 841             d2               = _mm256_mul_ps(d,d);
 842             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)))))));
 843             /*         #define INNERFLOPS INNERFLOPS+10 */
 844
 845             /*         #if 'Force' in KERNEL_VF */
 846             dsw              = _mm256_mul_ps(d2,_mm256_add_ps(swF2,_mm256_mul_ps(d,_mm256_add_ps(swF3,_mm256_mul_ps(d,swF4)))));
 847             /*             #define INNERFLOPS INNERFLOPS+5 */
 848             /*         #endif */
 849
 850             /* Evaluate switch function */
 851             /*         #if 'Force' in KERNEL_VF */
 852             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 853             /*             #if 'electrostatics' in INTERACTION_FLAGS[I][J] and KERNEL_MOD_ELEC=='PotentialSwitch' */
 854             felec            = _mm256_sub_ps( _mm256_mul_ps(felec,sw) , _mm256_mul_ps(rinv{I}{J},_mm256_mul_ps(velec,dsw)) );
 855             /*                 #define INNERFLOPS INNERFLOPS+4 */
 856             /*             #endif */
 857             /*             #if 'vdw' in INTERACTION_FLAGS[I][J] and KERNEL_MOD_VDW=='PotentialSwitch' */
 858             fvdw             = _mm256_sub_ps( _mm256_mul_ps(fvdw,sw) , _mm256_mul_ps(rinv{I}{J},_mm256_mul_ps(vvdw,dsw)) );
 859             /*                 #define INNERFLOPS INNERFLOPS+4 */
 860             /*             #endif */
 861             /*         #endif */
 862             /*         #if 'Potential' in KERNEL_VF */
 863             /*             #if 'electrostatics' in INTERACTION_FLAGS[I][J] and KERNEL_MOD_ELEC=='PotentialSwitch' */
 864             velec            = _mm256_mul_ps(velec,sw);
 865             /*                 #define INNERFLOPS INNERFLOPS+1 */
 866             /*             #endif */
 867             /*             #if 'vdw' in INTERACTION_FLAGS[I][J] and KERNEL_MOD_VDW=='PotentialSwitch' */
 868             vvdw             = _mm256_mul_ps(vvdw,sw);
 869             /*                 #define INNERFLOPS INNERFLOPS+1 */
 870             /*             #endif */
 871             /*         #endif */
 872             /*     #endif */
 873             /*     ## Note special check for TIP4P-TIP4P. Since we are cutting of all hydrogen interactions we also cut the LJ-only O-O interaction */
 874             /*     #if 'exactcutoff' in INTERACTION_FLAGS[I][J] or (GEOMETRY_I=='Water4' and GEOMETRY_J=='Water4' and 'exactcutoff' in INTERACTION_FLAGS[1][1]) */
 875             cutoff_mask      = _mm256_cmp_ps(rsq{I}{J},rcutoff2,_CMP_LT_OQ);
 876             /*         #define INNERFLOPS INNERFLOPS+1 */
 877             /*     #endif */
 878
 879             /*     #if 'Potential' in KERNEL_VF */
 880             /* Update potential sum for this i atom from the interaction with this j atom. */
 881             /*         #if 'electrostatics' in INTERACTION_FLAGS[I][J] */
 882             /*             #if 'exactcutoff' in INTERACTION_FLAGS[I][J] */
 883             velec            = _mm256_and_ps(velec,cutoff_mask);
 884             /*                 #define INNERFLOPS INNERFLOPS+1 */
 885             /*             #endif                                       */
 886             /*             #if ROUND == 'Epilogue' */
 887             velec            = _mm256_andnot_ps(dummy_mask,velec);
 888             /*             #endif */
 889             velecsum         = _mm256_add_ps(velecsum,velec);
 890             /*             #define INNERFLOPS INNERFLOPS+1 */
 891             /*             #if KERNEL_ELEC=='GeneralizedBorn' */
 892             /*             #if 'exactcutoff' in INTERACTION_FLAGS[I][J] */
 893             vgb              = _mm256_and_ps(vgb,cutoff_mask);
 894             /*                 #define INNERFLOPS INNERFLOPS+1 */
 895             /*             #endif                                       */
 896             /*             #if ROUND == 'Epilogue' */
 897             vgb              = _mm256_andnot_ps(dummy_mask,vgb);
 898             /*             #endif */
 899             vgbsum           = _mm256_add_ps(vgbsum,vgb);
 900             /*                 #define INNERFLOPS INNERFLOPS+1 */
 901             /*             #endif */
 902             /*         #endif */
 903             /*         #if 'vdw' in INTERACTION_FLAGS[I][J] */
 904             /*     ## Note special check for TIP4P-TIP4P. Since we are cutting of all hydrogen interactions we also cut the LJ-only O-O interaction */
 905             /*     #if 'exactcutoff' in INTERACTION_FLAGS[I][J] or (GEOMETRY_I=='Water4' and GEOMETRY_J=='Water4' and 'exactcutoff' in INTERACTION_FLAGS[1][1]) */
 906             vvdw             = _mm256_and_ps(vvdw,cutoff_mask);
 907             /*                 #define INNERFLOPS INNERFLOPS+1 */
 908             /*             #endif                                       */
 909             /*             #if ROUND == 'Epilogue' */
 910             vvdw             = _mm256_andnot_ps(dummy_mask,vvdw);
 911             /*             #endif */
 912             vvdwsum          = _mm256_add_ps(vvdwsum,vvdw);
 913             /*             #define INNERFLOPS INNERFLOPS+1 */
 914             /*         #endif */
 915             /*     #endif */
 916
 917             /*     #if 'Force' in KERNEL_VF */
 918
 919             /*         #if 'electrostatics' in INTERACTION_FLAGS[I][J] and 'vdw' in INTERACTION_FLAGS[I][J] */
 920             fscal            = _mm256_add_ps(felec,fvdw);
 921             /*             #define INNERFLOPS INNERFLOPS+1 */
 922             /*         #elif 'electrostatics' in INTERACTION_FLAGS[I][J] */
 923             fscal            = felec;
 924             /*         #elif 'vdw' in INTERACTION_FLAGS[I][J] */
 925             fscal            = fvdw;
 926             /*        #endif */
 927
 928             /*     ## Note special check for TIP4P-TIP4P. Since we are cutting of all hydrogen interactions we also cut the LJ-only O-O interaction */
 929             /*     #if 'exactcutoff' in INTERACTION_FLAGS[I][J] or (GEOMETRY_I=='Water4' and GEOMETRY_J=='Water4' and 'exactcutoff' in INTERACTION_FLAGS[1][1]) */
 930             fscal            = _mm256_and_ps(fscal,cutoff_mask);
 931             /*                 #define INNERFLOPS INNERFLOPS+1 */
 932             /*             #endif                                       */
 933
 934             /*             #if ROUND == 'Epilogue' */
 935             fscal            = _mm256_andnot_ps(dummy_mask,fscal);
 936             /*             #endif */
 937
 938             /* Calculate temporary vectorial force */
 939             tx               = _mm256_mul_ps(fscal,dx{I}{J});
 940             ty               = _mm256_mul_ps(fscal,dy{I}{J});
 941             tz               = _mm256_mul_ps(fscal,dz{I}{J});
 942
 943             /* Update vectorial force */
 944             fix{I}             = _mm256_add_ps(fix{I},tx);
 945             fiy{I}             = _mm256_add_ps(fiy{I},ty);
 946             fiz{I}             = _mm256_add_ps(fiz{I},tz);
 947             /*             #define INNERFLOPS INNERFLOPS+6 */
 948
 949             /* #if GEOMETRY_I == 'Particle'             */
 950             /*     #if ROUND == 'Loop' */
 951             fjptrA             = f+j_coord_offsetA;
 952             fjptrB             = f+j_coord_offsetB;
 953             fjptrC             = f+j_coord_offsetC;
 954             fjptrD             = f+j_coord_offsetD;
 955             fjptrE             = f+j_coord_offsetE;
 956             fjptrF             = f+j_coord_offsetF;
 957             fjptrG             = f+j_coord_offsetG;
 958             fjptrH             = f+j_coord_offsetH;
 959             /*     #else */
 960             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
 961             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
 962             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
 963             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
 964             fjptrE             = (jnrlistE>=0) ? f+j_coord_offsetE : scratch;
 965             fjptrF             = (jnrlistF>=0) ? f+j_coord_offsetF : scratch;
 966             fjptrG             = (jnrlistG>=0) ? f+j_coord_offsetG : scratch;
 967             fjptrH             = (jnrlistH>=0) ? f+j_coord_offsetH : scratch;
 968             /*     #endif */
 969             gmx_mm256_decrement_1rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,tx,ty,tz);
 970             /*     #define INNERFLOPS INNERFLOPS+3      */
 971             /* #else                                    */
 972             fjx{J}             = _mm256_add_ps(fjx{J},tx);
 973             fjy{J}             = _mm256_add_ps(fjy{J},ty);
 974             fjz{J}             = _mm256_add_ps(fjz{J},tz);
 975             /*     #define INNERFLOPS INNERFLOPS+3      */
 976             /* #endif                                   */
 977
 978             /*     #endif */
 979
 980             /*     ## Note special check for TIP4P-TIP4P. Since we are cutting of all hydrogen interactions we also cut the LJ-only O-O interaction */
 981             /*     #if 'exactcutoff' in INTERACTION_FLAGS[I][J] or (GEOMETRY_I=='Water4' and GEOMETRY_J=='Water4' and 'exactcutoff' in INTERACTION_FLAGS[1][1]) */
 982             /*         #if 0    ## This and next two lines is a hack to maintain indentation in template file */
 983             {
 984                 /*     #endif */
 985             }
 986             /*     #endif */
 987             /*    ## End of check for the interaction being outside the cutoff */
 988
 989             /* #endfor */
 990             /* ## End of loop over i-j interaction pairs */
 991
 992             /* #if GEOMETRY_I != 'Particle' */
 993             /*     #if ROUND == 'Loop' */
 994             fjptrA             = f+j_coord_offsetA;
 995             fjptrB             = f+j_coord_offsetB;
 996             fjptrC             = f+j_coord_offsetC;
 997             fjptrD             = f+j_coord_offsetD;
 998             fjptrE             = f+j_coord_offsetE;
 999             fjptrF             = f+j_coord_offsetF;
1000             fjptrG             = f+j_coord_offsetG;
1001             fjptrH             = f+j_coord_offsetH;
1002             /*     #else */
1003             fjptrA             = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
1004             fjptrB             = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
1005             fjptrC             = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
1006             fjptrD             = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
1007             fjptrE             = (jnrlistE>=0) ? f+j_coord_offsetE : scratch;
1008             fjptrF             = (jnrlistF>=0) ? f+j_coord_offsetF : scratch;
1009             fjptrG             = (jnrlistG>=0) ? f+j_coord_offsetG : scratch;
1010             fjptrH             = (jnrlistH>=0) ? f+j_coord_offsetH : scratch;
1011             /*     #endif */
1012             /* #endif */
1013
1014             /* #if 'Water' in GEOMETRY_I and GEOMETRY_J == 'Particle' */
1015             gmx_mm256_decrement_1rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,fjx0,fjy0,fjz0);
1016             /*     #define INNERFLOPS INNERFLOPS+3      */
1017             /* #elif GEOMETRY_J == 'Water3'             */
1018             gmx_mm256_decrement_3rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,
1019                                                       fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2);
1020             /*     #define INNERFLOPS INNERFLOPS+9      */
1021             /* #elif GEOMETRY_J == 'Water4'             */
1022             /*     #if 0 in PARTICLES_J                 */
1023             gmx_mm256_decrement_4rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,
1024                                                       fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,
1025                                                       fjx2,fjy2,fjz2,fjx3,fjy3,fjz3);
1026             /*     #define INNERFLOPS INNERFLOPS+12     */
1027             /*     #else                                */
1028             gmx_mm256_decrement_3rvec_8ptr_swizzle_ps(fjptrA+DIM,fjptrB+DIM,fjptrC+DIM,fjptrD+DIM,
1029                                                       fjptrE+DIM,fjptrF+DIM,fjptrG+DIM,fjptrH+DIM,
1030                                                       fjx1,fjy1,fjz1,fjx2,fjy2,fjz2,fjx3,fjy3,fjz3);
1031             /*     #define INNERFLOPS INNERFLOPS+9      */
1032             /*     #endif                               */
1033             /* #endif                                   */
1034
1035             /* Inner loop uses {INNERFLOPS} flops */
1036         }
1037
1038         /* #endfor */
1039
1040         /* End of innermost loop */
1041
1042         /* #if 'Force' in KERNEL_VF */
1043         /*     #if GEOMETRY_I == 'Particle'            */
1044         gmx_mm256_update_iforce_1atom_swizzle_ps(fix0,fiy0,fiz0,
1045                                                  f+i_coord_offset,fshift+i_shift_offset);
1046         /*         #define OUTERFLOPS OUTERFLOPS+6     */
1047         /*     #elif GEOMETRY_I == 'Water3'            */
1048         gmx_mm256_update_iforce_3atom_swizzle_ps(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,
1049                                                  f+i_coord_offset,fshift+i_shift_offset);
1050         /*         #define OUTERFLOPS OUTERFLOPS+18    */
1051         /*     #elif GEOMETRY_I == 'Water4'            */
1052         /*         #if 0 in PARTICLES_I                */
1053         gmx_mm256_update_iforce_4atom_swizzle_ps(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
1054                                                  f+i_coord_offset,fshift+i_shift_offset);
1055         /*             #define OUTERFLOPS OUTERFLOPS+24    */
1056         /*         #else                               */
1057         gmx_mm256_update_iforce_3atom_swizzle_ps(fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
1058                                                  f+i_coord_offset+DIM,fshift+i_shift_offset);
1059         /*             #define OUTERFLOPS OUTERFLOPS+18    */
1060         /*         #endif                              */
1061         /*     #endif                                  */
1062         /* #endif                                      */
1063
1064         /* #if 'Potential' in KERNEL_VF */
1065         ggid                        = gid[iidx];
1066         /* Update potential energies */
1067         /*     #if KERNEL_ELEC != 'None' */
1068         gmx_mm256_update_1pot_ps(velecsum,kernel_data->energygrp_elec+ggid);
1069         /*         #define OUTERFLOPS OUTERFLOPS+1 */
1070         /*     #endif */
1071         /*     #if 'GeneralizedBorn' in KERNEL_ELEC */
1072         gmx_mm256_update_1pot_ps(vgbsum,kernel_data->energygrp_polarization+ggid);
1073         /*         #define OUTERFLOPS OUTERFLOPS+1 */
1074         /*     #endif */
1075         /*     #if KERNEL_VDW != 'None' */
1076         gmx_mm256_update_1pot_ps(vvdwsum,kernel_data->energygrp_vdw+ggid);
1077         /*         #define OUTERFLOPS OUTERFLOPS+1 */
1078         /*     #endif */
1079         /* #endif */
1080         /*     #if 'GeneralizedBorn' in KERNEL_ELEC and 'Force' in KERNEL_VF */
1081         dvdasum = _mm256_mul_ps(dvdasum, _mm256_mul_ps(isai{I},isai{I}));
1082         gmx_mm256_update_1pot_ps(dvdasum,dvda+inr);
1083         /*     #endif */
1084
1085         /* Increment number of inner iterations */
1086         inneriter                  += j_index_end - j_index_start;
1087
1088         /* Outer loop uses {OUTERFLOPS} flops */
1089     }
1090
1091     /* Increment number of outer iterations */
1092     outeriter        += nri;
1093
1094     /* Update outer/inner flops */
1095     /* ## NB: This is not important, it just affects the flopcount. However, since our preprocessor is */
1096     /* ## primitive and replaces aggressively even in strings inside these directives, we need to      */
1097     /* ## assemble the main part of the name (containing KERNEL/ELEC/VDW) directly in the source.      */
1098     /* #if GEOMETRY_I == 'Water3'            */
1099     /*     #define ISUFFIX '_W3'             */
1100     /* #elif GEOMETRY_I == 'Water4'          */
1101     /*     #define ISUFFIX '_W4'             */
1102     /* #else                                 */
1103     /*     #define ISUFFIX ''                */
1104     /* #endif                                */
1105     /* #if GEOMETRY_J == 'Water3'            */
1106     /*     #define JSUFFIX 'W3'              */
1107     /* #elif GEOMETRY_J == 'Water4'          */
1108     /*     #define JSUFFIX 'W4'              */
1109     /* #else                                 */
1110     /*     #define JSUFFIX ''                */
1111     /* #endif                                */
1112     /* #if 'PotentialAndForce' in KERNEL_VF  */
1113     /*     #define VFSUFFIX  '_VF'           */
1114     /* #elif 'Potential' in KERNEL_VF        */
1115     /*     #define VFSUFFIX '_V'             */
1116     /* #else                                 */
1117     /*     #define VFSUFFIX '_F'             */
1118     /* #endif                                */
1119
1120     /* #if KERNEL_ELEC != 'None' and KERNEL_VDW != 'None' */
1121     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW{ISUFFIX}{JSUFFIX}{VFSUFFIX},outeriter*{OUTERFLOPS} + inneriter*{INNERFLOPS});
1122     /* #elif KERNEL_ELEC != 'None' */
1123     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC{ISUFFIX}{JSUFFIX}{VFSUFFIX},outeriter*{OUTERFLOPS} + inneriter*{INNERFLOPS});
1124     /* #else */
1125     inc_nrnb(nrnb,eNR_NBKERNEL_VDW{ISUFFIX}{JSUFFIX}{VFSUFFIX},outeriter*{OUTERFLOPS} + inneriter*{INNERFLOPS});
1126     /* #endif  */
1127 }