src/gromacs/gmxlib/nonbonded/nb_kernel_sparc64_hpc_ace_double/nb_kernel_template_sparc64_hpc_ace_double.pre

   1 /* #if 0 */
   2 /*
   3  * This file is part of the GROMACS molecular simulation package.
   4  *
   5  * Copyright (c) 2012,2013,2014,2015,2017,2018, 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 "gmxpre.h"
  40
  41 #include "config.h"
  42
  43 #include <math.h>
  44
  45 #include "../nb_kernel.h"
  46 #include "gromacs/gmxlib/nrnb.h"
  47
  48 #include "kernelutil_sparc64_hpc_ace_double.h"
  49 /* #endif */
  50
  51 /* ## List of variables set by the generating script:                                    */
  52 /* ##                                                                                    */
  53 /* ## Setttings that apply to the entire kernel:                                         */
  54 /* ## KERNEL_ELEC:           String, choice for electrostatic interactions               */
  55 /* ## KERNEL_VDW:            String, choice for van der Waals interactions               */
  56 /* ## KERNEL_NAME:           String, name of this kernel                                 */
  57 /* ## KERNEL_VF:             String telling if we calculate potential, force, or both    */
  58 /* ## GEOMETRY_I/GEOMETRY_J: String, name of each geometry, e.g. 'Water3' or '1Particle' */
  59 /* ##                                                                                    */
  60 /* ## Setttings that apply to particles in the outer (I) or inner (J) loops:             */
  61 /* ## PARTICLES_I[]/         Arrays with lists of i/j particles to use in kernel. It is  */
  62 /* ## PARTICLES_J[]:         just [0] for particle geometry, but can be longer for water */
  63 /* ## PARTICLES_ELEC_I[]/    Arrays with lists of i/j particle that have electrostatics  */
  64 /* ## PARTICLES_ELEC_J[]:    interactions that should be calculated in this kernel.      */
  65 /* ## PARTICLES_VDW_I[]/     Arrays with the list of i/j particle that have VdW          */
  66 /* ## PARTICLES_VDW_J[]:     interactions that should be calculated in this kernel.      */
  67 /* ##                                                                                    */
  68 /* ## Setttings for pairs of interactions (e.g. 2nd i particle against 1st j particle)   */
  69 /* ## PAIRS_IJ[]:            Array with (i,j) tuples of pairs for which interactions     */
  70 /* ##                        should be calculated in this kernel. Zero-charge particles  */
  71 /* ##                        do not have interactions with particles without vdw, and    */
  72 /* ##                        Vdw-only interactions are not evaluated in a no-vdw-kernel. */
  73 /* ## INTERACTION_FLAGS[][]: 2D matrix, dimension e.g. 3*3 for water-water interactions. */
  74 /* ##                        For each i-j pair, the element [I][J] is a list of strings  */
  75 /* ##                        defining properties/flags of this interaction. Examples     */
  76 /* ##                        include 'electrostatics'/'vdw' if that type of interaction  */
  77 /* ##                        should be evaluated, 'rsq'/'rinv'/'rinvsq' if those values  */
  78 /* ##                        are needed, and 'exactcutoff' or 'shift','switch' to        */
  79 /* ##                        decide if the force/potential should be modified. This way  */
  80 /* ##                        we only calculate values absolutely needed for each case.   */
  81
  82 /* ## Calculate the size and offset for (merged/interleaved) table data */
  83
  84 /*
  85  * Gromacs nonbonded kernel:   {KERNEL_NAME}
  86  * Electrostatics interaction: {KERNEL_ELEC}
  87  * VdW interaction:            {KERNEL_VDW}
  88  * Geometry:                   {GEOMETRY_I}-{GEOMETRY_J}
  89  * Calculate force/pot:        {KERNEL_VF}
  90  */
  91 void
  92 {KERNEL_NAME}
  93                     (t_nblist                    * gmx_restrict       nlist,
  94                      rvec                        * gmx_restrict          xx,
  95                      rvec                        * gmx_restrict          ff,
  96                      struct t_forcerec           * gmx_restrict          fr,
  97                      t_mdatoms                   * gmx_restrict     mdatoms,
  98                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
  99                      t_nrnb                      * gmx_restrict        nrnb)
 100 {
 101     /* ## Not all variables are used for all kernels, but any optimizing compiler fixes that, */
 102     /* ## so there is no point in going to extremes to exclude variables that are not needed. */
 103     /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
 104      * just 0 for non-waters.
 105      * Suffixes A,B refer to j loop unrolling done with double precision SIMD, e.g. for the two different
 106      * jnr indices corresponding to data put in the four positions in the SIMD register.
 107      */
 108     int              i_shift_offset,i_coord_offset,outeriter,inneriter;
 109     int              j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
 110     int              jnrA,jnrB;
 111     int              j_coord_offsetA,j_coord_offsetB;
 112     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 113     real             rcutoff_scalar;
 114     real             *shiftvec,*fshift,*x,*f;
 115     _fjsp_v2r8       tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
 116     /* #for I in PARTICLES_I */
 117     int              vdwioffset{I};
 118     _fjsp_v2r8       ix{I},iy{I},iz{I},fix{I},fiy{I},fiz{I},iq{I},isai{I};
 119     /* #endfor */
 120     /* #for J in PARTICLES_J */
 121     int              vdwjidx{J}A,vdwjidx{J}B;
 122     _fjsp_v2r8       jx{J},jy{J},jz{J},fjx{J},fjy{J},fjz{J},jq{J},isaj{J};
 123     /* #endfor */
 124     /* #for I,J in PAIRS_IJ */
 125     _fjsp_v2r8       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};
 126     /* #endfor */
 127     /* #if KERNEL_ELEC != 'None' */
 128     _fjsp_v2r8       velec,felec,velecsum,facel,crf,krf,krf2;
 129     real             *charge;
 130     /* #endif */
 131     /* #if KERNEL_VDW != 'None' */
 132     int              nvdwtype;
 133     _fjsp_v2r8       rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
 134     int              *vdwtype;
 135     real             *vdwparam;
 136     _fjsp_v2r8       one_sixth   = gmx_fjsp_set1_v2r8(1.0/6.0);
 137     _fjsp_v2r8       one_twelfth = gmx_fjsp_set1_v2r8(1.0/12.0);
 138     /* #endif */
 139     /* #if 'Table' in KERNEL_ELEC or 'Table' in KERNEL_VDW */
 140     _fjsp_v2r8       rt,vfeps,vftabscale,Y,F,G,H,Heps,Fp,VV,FF,twovfeps;
 141     real             *vftab;
 142     /* #endif */
 143     /* #if 'LJEwald' in KERNEL_VDW */
 144     /* #for I,J in PAIRS_IJ */
 145     _fjsp_v2r8           c6grid_{I}{J};
 146     /* #endfor */
 147     real                 *vdwgridparam;
 148     _fjsp_v2r8           ewclj,ewclj2,ewclj6,ewcljrsq,poly,exponent,f6A,f6B,sh_lj_ewald;
 149     _fjsp_v2r8           one_half = gmx_fjsp_set1_v2r8(0.5);
 150     _fjsp_v2r8           minus_one = gmx_fjsp_set1_v2r8(-1.0);
 151     /* #endif */
 152     /* #if 'Ewald' in KERNEL_ELEC */
 153     _fjsp_v2r8       ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
 154     real             *ewtab;
 155     /* #endif */
 156     /* #if 'PotentialSwitch' in [KERNEL_MOD_ELEC,KERNEL_MOD_VDW] */
 157     _fjsp_v2r8       rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
 158     real             rswitch_scalar,d_scalar;
 159     /* #endif */
 160     _fjsp_v2r8       itab_tmp;
 161     _fjsp_v2r8       dummy_mask,cutoff_mask;
 162     _fjsp_v2r8       one     = gmx_fjsp_set1_v2r8(1.0);
 163     _fjsp_v2r8       two     = gmx_fjsp_set1_v2r8(2.0);
 164     union { _fjsp_v2r8 simd; long long int i[2]; } vfconv,gbconv,ewconv;
 165
 166     x                = xx[0];
 167     f                = ff[0];
 168
 169     nri              = nlist->nri;
 170     iinr             = nlist->iinr;
 171     jindex           = nlist->jindex;
 172     jjnr             = nlist->jjnr;
 173     shiftidx         = nlist->shift;
 174     gid              = nlist->gid;
 175     shiftvec         = fr->shift_vec[0];
 176     fshift           = fr->fshift[0];
 177     /* #if KERNEL_ELEC != 'None' */
 178     facel            = gmx_fjsp_set1_v2r8(fr->ic->epsfac);
 179     charge           = mdatoms->chargeA;
 180     /*     #if 'ReactionField' in KERNEL_ELEC */
 181     krf              = gmx_fjsp_set1_v2r8(fr->ic->k_rf);
 182     krf2             = gmx_fjsp_set1_v2r8(fr->ic->k_rf*2.0);
 183     crf              = gmx_fjsp_set1_v2r8(fr->ic->c_rf);
 184     /*     #endif */
 185     /* #endif */
 186     /* #if KERNEL_VDW != 'None' */
 187     nvdwtype         = fr->ntype;
 188     vdwparam         = fr->nbfp;
 189     vdwtype          = mdatoms->typeA;
 190     /* #endif */
 191     /* #if 'LJEwald' in KERNEL_VDW */
 192     vdwgridparam     = fr->ljpme_c6grid;
 193     sh_lj_ewald      = gmx_fjsp_set1_v2r8(fr->ic->sh_lj_ewald);
 194     ewclj            = gmx_fjsp_set1_v2r8(fr->ic->ewaldcoeff_lj);
 195     ewclj2           = _fjsp_mul_v2r8(minus_one,_fjsp_mul_v2r8(ewclj,ewclj));
 196     /* #endif */
 197
 198     /* #if 'Table' in KERNEL_ELEC and 'Table' in KERNEL_VDW */
 199     vftab            = kernel_data->table_elec_vdw->data;
 200     vftabscale       = gmx_fjsp_set1_v2r8(kernel_data->table_elec_vdw->scale);
 201     /* #elif 'Table' in KERNEL_ELEC */
 202     vftab            = kernel_data->table_elec->data;
 203     vftabscale       = gmx_fjsp_set1_v2r8(kernel_data->table_elec->scale);
 204     /* #elif 'Table' in KERNEL_VDW */
 205     vftab            = kernel_data->table_vdw->data;
 206     vftabscale       = gmx_fjsp_set1_v2r8(kernel_data->table_vdw->scale);
 207     /* #endif */
 208
 209     /* #if 'Ewald' in KERNEL_ELEC */
 210     sh_ewald         = gmx_fjsp_set1_v2r8(fr->ic->sh_ewald);
 211     /*     #if KERNEL_VF=='Force' and KERNEL_MOD_ELEC!='PotentialSwitch' */
 212     ewtab            = fr->ic->tabq_coul_F;
 213     ewtabscale       = gmx_fjsp_set1_v2r8(fr->ic->tabq_scale);
 214     ewtabhalfspace   = gmx_fjsp_set1_v2r8(0.5/fr->ic->tabq_scale);
 215     /*     #else */
 216     ewtab            = fr->ic->tabq_coul_FDV0;
 217     ewtabscale       = gmx_fjsp_set1_v2r8(fr->ic->tabq_scale);
 218     ewtabhalfspace   = gmx_fjsp_set1_v2r8(0.5/fr->ic->tabq_scale);
 219      /*     #endif */
 220     /* #endif */
 221
 222     /* #if 'Water' in GEOMETRY_I */
 223     /* Setup water-specific parameters */
 224     inr              = nlist->iinr[0];
 225     /*     #for I in PARTICLES_ELEC_I */
 226     iq{I}              = _fjsp_mul_v2r8(facel,gmx_fjsp_set1_v2r8(charge[inr+{I}]));
 227     /*     #endfor */
 228     /*     #for I in PARTICLES_VDW_I */
 229     vdwioffset{I}      = 2*nvdwtype*vdwtype[inr+{I}];
 230     /*     #endfor */
 231     /* #endif */
 232
 233     /* #if 'Water' in GEOMETRY_J */
 234     /*     #for J in PARTICLES_ELEC_J */
 235     jq{J}              = gmx_fjsp_set1_v2r8(charge[inr+{J}]);
 236     /*     #endfor */
 237     /*     #for J in PARTICLES_VDW_J */
 238     vdwjidx{J}A        = 2*vdwtype[inr+{J}];
 239     /*     #endfor */
 240     /*     #for I,J in PAIRS_IJ */
 241     /*         #if 'electrostatics' in INTERACTION_FLAGS[I][J] */
 242     qq{I}{J}             = _fjsp_mul_v2r8(iq{I},jq{J});
 243     /*         #endif */
 244     /*         #if 'vdw' in INTERACTION_FLAGS[I][J] */
 245     /*             #if 'LJEwald' in KERNEL_VDW */
 246     c6_{I}{J}            = gmx_fjsp_set1_v2r8(vdwparam[vdwioffset{I}+vdwjidx{J}A]);
 247     c12_{I}{J}           = gmx_fjsp_set1_v2r8(vdwparam[vdwioffset{I}+vdwjidx{J}A+1]);
 248     c6grid_{I}{J}        = gmx_fjsp_set1_v2r8(vdwgridparam[vdwioffset{I}+vdwjidx{J}A]);
 249     /*             #else */
 250     c6_{I}{J}            = gmx_fjsp_set1_v2r8(vdwparam[vdwioffset{I}+vdwjidx{J}A]);
 251     c12_{I}{J}           = gmx_fjsp_set1_v2r8(vdwparam[vdwioffset{I}+vdwjidx{J}A+1]);
 252     /*             #endif */
 253     /*         #endif */
 254     /*     #endfor */
 255     /* #endif */
 256
 257     /* #if KERNEL_MOD_ELEC!='None' or KERNEL_MOD_VDW!='None' */
 258     /*     #if KERNEL_ELEC!='None' */
 259     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 260     rcutoff_scalar   = fr->ic->rcoulomb;
 261     /*     #else */
 262     rcutoff_scalar   = fr->ic->rvdw;
 263     /*     #endif */
 264     rcutoff          = gmx_fjsp_set1_v2r8(rcutoff_scalar);
 265     rcutoff2         = _fjsp_mul_v2r8(rcutoff,rcutoff);
 266     /* #endif */
 267
 268     /* #if KERNEL_MOD_VDW=='PotentialShift' */
 269     sh_vdw_invrcut6  = gmx_fjsp_set1_v2r8(fr->ic->sh_invrc6);
 270     rvdw             = gmx_fjsp_set1_v2r8(fr->ic->rvdw);
 271     /* #endif */
 272
 273     /* #if 'PotentialSwitch' in [KERNEL_MOD_ELEC,KERNEL_MOD_VDW] */
 274     /*     #if KERNEL_MOD_ELEC=='PotentialSwitch'  */
 275     rswitch_scalar   = fr->ic->rcoulomb_switch;
 276     rswitch          = gmx_fjsp_set1_v2r8(rswitch_scalar);
 277     /*     #else */
 278     rswitch_scalar   = fr->ic->rvdw_switch;
 279     rswitch          = gmx_fjsp_set1_v2r8(rswitch_scalar);
 280     /*     #endif */
 281     /* Setup switch parameters */
 282     d_scalar         = rcutoff_scalar-rswitch_scalar;
 283     d                = gmx_fjsp_set1_v2r8(d_scalar);
 284     swV3             = gmx_fjsp_set1_v2r8(-10.0/(d_scalar*d_scalar*d_scalar));
 285     swV4             = gmx_fjsp_set1_v2r8( 15.0/(d_scalar*d_scalar*d_scalar*d_scalar));
 286     swV5             = gmx_fjsp_set1_v2r8( -6.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
 287     /*     #if 'Force' in KERNEL_VF */
 288     swF2             = gmx_fjsp_set1_v2r8(-30.0/(d_scalar*d_scalar*d_scalar));
 289     swF3             = gmx_fjsp_set1_v2r8( 60.0/(d_scalar*d_scalar*d_scalar*d_scalar));
 290     swF4             = gmx_fjsp_set1_v2r8(-30.0/(d_scalar*d_scalar*d_scalar*d_scalar*d_scalar));
 291     /*     #endif */
 292     /* #endif */
 293
 294     /* Avoid stupid compiler warnings */
 295     jnrA = jnrB = 0;
 296     j_coord_offsetA = 0;
 297     j_coord_offsetB = 0;
 298
 299     /* ## Keep track of the floating point operations we issue for reporting! */
 300     /* #define OUTERFLOPS 0 */
 301     outeriter        = 0;
 302     inneriter        = 0;
 303
 304     /* Start outer loop over neighborlists */
 305     for(iidx=0; iidx<nri; iidx++)
 306     {
 307         /* Load shift vector for this list */
 308         i_shift_offset   = DIM*shiftidx[iidx];
 309
 310         /* Load limits for loop over neighbors */
 311         j_index_start    = jindex[iidx];
 312         j_index_end      = jindex[iidx+1];
 313
 314         /* Get outer coordinate index */
 315         inr              = iinr[iidx];
 316         i_coord_offset   = DIM*inr;
 317
 318         /* Load i particle coords and add shift vector */
 319         /* #if GEOMETRY_I == 'Particle' */
 320         gmx_fjsp_load_shift_and_1rvec_broadcast_v2r8(shiftvec+i_shift_offset,x+i_coord_offset,&ix0,&iy0,&iz0);
 321         /* #elif GEOMETRY_I == 'Water3' */
 322         gmx_fjsp_load_shift_and_3rvec_broadcast_v2r8(shiftvec+i_shift_offset,x+i_coord_offset,
 323                                                  &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2);
 324         /* #elif GEOMETRY_I == 'Water4' */
 325         /*     #if 0 in PARTICLES_I                 */
 326         gmx_fjsp_load_shift_and_4rvec_broadcast_v2r8(shiftvec+i_shift_offset,x+i_coord_offset,
 327                                                  &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
 328         /*     #else                                */
 329         gmx_fjsp_load_shift_and_3rvec_broadcast_v2r8(shiftvec+i_shift_offset,x+i_coord_offset+DIM,
 330                                                  &ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
 331         /*     #endif                               */
 332         /* #endif                                   */
 333
 334         /* #if 'Force' in KERNEL_VF */
 335         /*     #for I in PARTICLES_I */
 336         fix{I}             = _fjsp_setzero_v2r8();
 337         fiy{I}             = _fjsp_setzero_v2r8();
 338         fiz{I}             = _fjsp_setzero_v2r8();
 339         /*     #endfor */
 340         /* #endif */
 341
 342         /* ## For water we already preloaded parameters at the start of the kernel */
 343         /* #if not 'Water' in GEOMETRY_I */
 344         /* Load parameters for i particles */
 345         /*     #for I in PARTICLES_ELEC_I */
 346         iq{I}              = _fjsp_mul_v2r8(facel,gmx_fjsp_load1_v2r8(charge+inr+{I}));
 347         /*         #define OUTERFLOPS OUTERFLOPS+1 */
 348         /*     #endfor */
 349         /*     #for I in PARTICLES_VDW_I */
 350         vdwioffset{I}      = 2*nvdwtype*vdwtype[inr+{I}];
 351         /*     #endfor */
 352         /* #endif */
 353
 354         /* #if 'Potential' in KERNEL_VF */
 355         /* Reset potential sums */
 356         /*     #if KERNEL_ELEC != 'None' */
 357         velecsum         = _fjsp_setzero_v2r8();
 358         /*     #endif */
 359         /*     #if KERNEL_VDW != 'None' */
 360         vvdwsum          = _fjsp_setzero_v2r8();
 361         /*     #endif */
 362         /* #endif */
 363
 364         /* #for ROUND in ['Loop','Epilogue'] */
 365
 366         /* #if ROUND =='Loop' */
 367         /* Start inner kernel loop */
 368         for(jidx=j_index_start; jidx<j_index_end-1; jidx+=2)
 369         {
 370         /* ## First round is normal loop (next statement resets indentation) */
 371         /*     #if 0 */
 372         }
 373         /*     #endif */
 374         /* #else */
 375         if(jidx<j_index_end)
 376         {
 377         /* ## Second round is epilogue */
 378         /* #endif */
 379         /* #define INNERFLOPS 0 */
 380
 381             /* #if ROUND =='Loop' */
 382             /* Get j neighbor index, and coordinate index */
 383             jnrA             = jjnr[jidx];
 384             jnrB             = jjnr[jidx+1];
 385             j_coord_offsetA  = DIM*jnrA;
 386             j_coord_offsetB  = DIM*jnrB;
 387
 388             /* load j atom coordinates */
 389             /*     #if GEOMETRY_J == 'Particle'             */
 390             gmx_fjsp_load_1rvec_2ptr_swizzle_v2r8(x+j_coord_offsetA,x+j_coord_offsetB,
 391                                               &jx0,&jy0,&jz0);
 392             /*     #elif GEOMETRY_J == 'Water3'             */
 393             gmx_fjsp_load_3rvec_2ptr_swizzle_v2r8(x+j_coord_offsetA,x+j_coord_offsetB,
 394                                               &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,&jy2,&jz2);
 395             /*     #elif GEOMETRY_J == 'Water4'             */
 396             /*         #if 0 in PARTICLES_J                 */
 397             gmx_fjsp_load_4rvec_2ptr_swizzle_v2r8(x+j_coord_offsetA,x+j_coord_offsetB,
 398                                               &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,
 399                                               &jy2,&jz2,&jx3,&jy3,&jz3);
 400             /*         #else                                */
 401             gmx_fjsp_load_3rvec_2ptr_swizzle_v2r8(x+j_coord_offsetA+DIM,x+j_coord_offsetB+DIM,
 402                                               &jx1,&jy1,&jz1,&jx2,&jy2,&jz2,&jx3,&jy3,&jz3);
 403             /*         #endif                               */
 404             /*     #endif                                   */
 405             /* #else */
 406             jnrA             = jjnr[jidx];
 407             j_coord_offsetA  = DIM*jnrA;
 408
 409             /* load j atom coordinates */
 410             /*     #if GEOMETRY_J == 'Particle'             */
 411             gmx_fjsp_load_1rvec_1ptr_swizzle_v2r8(x+j_coord_offsetA,
 412                                               &jx0,&jy0,&jz0);
 413             /*     #elif GEOMETRY_J == 'Water3'             */
 414             gmx_fjsp_load_3rvec_1ptr_swizzle_v2r8(x+j_coord_offsetA,
 415                                               &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,&jy2,&jz2);
 416             /*     #elif GEOMETRY_J == 'Water4'             */
 417             /*         #if 0 in PARTICLES_J                 */
 418             gmx_fjsp_load_4rvec_1ptr_swizzle_v2r8(x+j_coord_offsetA,
 419                                               &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,
 420                                               &jy2,&jz2,&jx3,&jy3,&jz3);
 421             /*         #else                                */
 422             gmx_fjsp_load_3rvec_1ptr_swizzle_v2r8(x+j_coord_offsetA+DIM,
 423                                               &jx1,&jy1,&jz1,&jx2,&jy2,&jz2,&jx3,&jy3,&jz3);
 424             /*         #endif                               */
 425             /*     #endif                                   */
 426             /* #endif */
 427
 428             /* Calculate displacement vector */
 429             /* #for I,J in PAIRS_IJ */
 430             dx{I}{J}             = _fjsp_sub_v2r8(ix{I},jx{J});
 431             dy{I}{J}             = _fjsp_sub_v2r8(iy{I},jy{J});
 432             dz{I}{J}             = _fjsp_sub_v2r8(iz{I},jz{J});
 433             /*     #define INNERFLOPS INNERFLOPS+3 */
 434             /* #endfor */
 435
 436             /* Calculate squared distance and things based on it */
 437             /* #for I,J in PAIRS_IJ */
 438             rsq{I}{J}            = gmx_fjsp_calc_rsq_v2r8(dx{I}{J},dy{I}{J},dz{I}{J});
 439             /*     #define INNERFLOPS INNERFLOPS+5 */
 440             /* #endfor */
 441
 442             /* #for I,J in PAIRS_IJ */
 443             /*     #if 'rinv' in INTERACTION_FLAGS[I][J] */
 444             rinv{I}{J}           = gmx_fjsp_invsqrt_v2r8(rsq{I}{J});
 445             /*         #define INNERFLOPS INNERFLOPS+5 */
 446             /*     #endif */
 447             /* #endfor */
 448
 449             /* #for I,J in PAIRS_IJ */
 450             /*     #if 'rinvsq' in INTERACTION_FLAGS[I][J] */
 451             /*         # if 'rinv' not in INTERACTION_FLAGS[I][J] */
 452             rinvsq{I}{J}         = gmx_fjsp_inv_v2r8(rsq{I}{J});
 453             /*             #define INNERFLOPS INNERFLOPS+4 */
 454             /*         #else */
 455             rinvsq{I}{J}         = _fjsp_mul_v2r8(rinv{I}{J},rinv{I}{J});
 456             /*             #define INNERFLOPS INNERFLOPS+1 */
 457             /*         #endif */
 458             /*     #endif */
 459             /* #endfor */
 460
 461             /* #if not 'Water' in GEOMETRY_J */
 462             /* Load parameters for j particles */
 463             /*     #for J in PARTICLES_ELEC_J */
 464             /*         #if ROUND =='Loop' */
 465             jq{J}              = gmx_fjsp_load_2real_swizzle_v2r8(charge+jnrA+{J},charge+jnrB+{J});
 466             /*         #else */
 467             jq{J}              = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(),charge+jnrA+{J});
 468             /*         #endif */
 469             /*     #endfor */
 470             /*     #for J in PARTICLES_VDW_J */
 471             vdwjidx{J}A        = 2*vdwtype[jnrA+{J}];
 472             /*         #if ROUND =='Loop' */
 473             vdwjidx{J}B        = 2*vdwtype[jnrB+{J}];
 474             /*         #endif */
 475             /*     #endfor */
 476             /* #endif */
 477
 478             /* #if 'Force' in KERNEL_VF and not 'Particle' in GEOMETRY_I */
 479             /*     #for J in PARTICLES_J */
 480             fjx{J}             = _fjsp_setzero_v2r8();
 481             fjy{J}             = _fjsp_setzero_v2r8();
 482             fjz{J}             = _fjsp_setzero_v2r8();
 483             /*     #endfor */
 484             /* #endif */
 485
 486             /* #for I,J in PAIRS_IJ */
 487
 488             /**************************
 489              * CALCULATE INTERACTIONS *
 490              **************************/
 491
 492             /*     #if 'exactcutoff' in INTERACTION_FLAGS[I][J] */
 493             /*         ## We always calculate rinv/rinvsq above to enable pipelineing in compilers (performance tested on x86) */
 494             if (gmx_fjsp_any_lt_v2r8(rsq{I}{J},rcutoff2))
 495             {
 496                 /*     #if 0    ## this and the next two lines is a hack to maintain auto-indentation in template file */
 497             }
 498             /*         #endif */
 499             /*         #define INNERFLOPS INNERFLOPS+1 */
 500             /*     #endif */
 501
 502             /*     #if 'r' in INTERACTION_FLAGS[I][J] */
 503             r{I}{J}              = _fjsp_mul_v2r8(rsq{I}{J},rinv{I}{J});
 504              /*         #define INNERFLOPS INNERFLOPS+1 */
 505             /*     #endif */
 506
 507             /*     ## For water geometries we already loaded parameters at the start of the kernel */
 508             /*     #if not 'Water' in GEOMETRY_J */
 509             /* Compute parameters for interactions between i and j atoms */
 510             /*         #if 'electrostatics' in INTERACTION_FLAGS[I][J] */
 511             qq{I}{J}             = _fjsp_mul_v2r8(iq{I},jq{J});
 512             /*             #define INNERFLOPS INNERFLOPS+1 */
 513             /*         #endif */
 514             /*         #if 'vdw' in INTERACTION_FLAGS[I][J] */
 515             gmx_fjsp_load_2pair_swizzle_v2r8(vdwparam+vdwioffset{I}+vdwjidx{J}A,
 516                                          vdwparam+vdwioffset{I}+vdwjidx{J}B,&c6_{I}{J},&c12_{I}{J});
 517
 518             /*             #if 'LJEwald' in KERNEL_VDW */
 519             c6grid_{I}{J}       = gmx_fjsp_load_2real_swizzle_v2r8(vdwgridparam+vdwioffset{I}+vdwjidx{J}A,
 520                                                                    vdwgridparam+vdwioffset{I}+vdwjidx{J}B);
 521             /*             #endif */
 522             /*         #endif */
 523             /*     #endif */
 524
 525             /*     #if 'table' in INTERACTION_FLAGS[I][J] */
 526             /* Calculate table index by multiplying r with table scale and truncate to integer */
 527             rt               = _fjsp_mul_v2r8(r{I}{J},vftabscale);
 528             itab_tmp         = _fjsp_dtox_v2r8(rt);
 529             vfeps            = _fjsp_sub_v2r8(rt, _fjsp_xtod_v2r8(itab_tmp));
 530             twovfeps         = _fjsp_add_v2r8(vfeps,vfeps);
 531             _fjsp_store_v2r8(&vfconv.simd,itab_tmp);
 532
 533             /*         #define INNERFLOPS INNERFLOPS+4                          */
 534             /*         #if 'Table' in KERNEL_ELEC and 'Table' in KERNEL_VDW     */
 535             /*             ## 3 tables, 4 data per point: multiply index by 12 */
 536             vfconv.i[0]     *= 12;
 537             vfconv.i[1]     *= 12;
 538             /*         #elif 'Table' in KERNEL_ELEC                             */
 539             /*             ## 1 table, 4 data per point: multiply index by 4   */
 540             vfconv.i[0]     *= 4;
 541             vfconv.i[1]     *= 4;
 542             /*         #elif 'Table' in KERNEL_VDW                              */
 543             /*             ## 2 tables, 4 data per point: multiply index by 8  */
 544             vfconv.i[0]     *= 8;
 545             vfconv.i[1]     *= 8;
 546             /*         #endif                                                   */
 547             /*     #endif */
 548
 549             /*     ## ELECTROSTATIC INTERACTIONS */
 550             /*     #if 'electrostatics' in INTERACTION_FLAGS[I][J] */
 551
 552             /*         #if KERNEL_ELEC=='Coulomb' */
 553
 554             /* COULOMB ELECTROSTATICS */
 555             velec            = _fjsp_mul_v2r8(qq{I}{J},rinv{I}{J});
 556             /*             #define INNERFLOPS INNERFLOPS+1 */
 557             /*             #if 'Force' in KERNEL_VF */
 558             felec            = _fjsp_mul_v2r8(velec,rinvsq{I}{J});
 559             /*                 #define INNERFLOPS INNERFLOPS+2 */
 560             /*             #endif */
 561
 562             /*         #elif KERNEL_ELEC=='ReactionField' */
 563
 564             /* REACTION-FIELD ELECTROSTATICS */
 565             /*             #if 'Potential' in KERNEL_VF */
 566             velec            = _fjsp_mul_v2r8(qq{I}{J},_fjsp_sub_v2r8(_fjsp_madd_v2r8(krf,rsq{I}{J},rinv{I}{J}),crf));
 567             /*                 #define INNERFLOPS INNERFLOPS+4 */
 568             /*             #endif */
 569             /*             #if 'Force' in KERNEL_VF */
 570             felec            = _fjsp_mul_v2r8(qq{I}{J},_fjsp_msub_v2r8(rinv{I}{J},rinvsq{I}{J},krf2));
 571             /*                 #define INNERFLOPS INNERFLOPS+3 */
 572             /*             #endif */
 573
 574             /*         #elif KERNEL_ELEC=='Ewald' */
 575             /* EWALD ELECTROSTATICS */
 576
 577             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 578             ewrt             = _fjsp_mul_v2r8(r{I}{J},ewtabscale);
 579             itab_tmp         = _fjsp_dtox_v2r8(ewrt);
 580             eweps            = _fjsp_sub_v2r8(ewrt,_fjsp_xtod_v2r8(itab_tmp));
 581             _fjsp_store_v2r8(&ewconv.simd,itab_tmp);
 582
 583             /*             #define INNERFLOPS INNERFLOPS+4 */
 584             /*             #if 'Potential' in KERNEL_VF or KERNEL_MOD_ELEC=='PotentialSwitch' */
 585             ewtabF           = _fjsp_load_v2r8( ewtab + 4*ewconv.i[0] );
 586             /*                 #if ROUND == 'Loop' */
 587             ewtabD           = _fjsp_load_v2r8( ewtab + 4*ewconv.i[1] );
 588             /*                 #else */
 589             ewtabD           = _fjsp_setzero_v2r8();
 590             /*                 #endif */
 591             GMX_FJSP_TRANSPOSE2_V2R8(ewtabF,ewtabD);
 592             ewtabV           = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[0] +2);
 593             /*                 #if ROUND == 'Loop' */
 594             ewtabFn          = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(), ewtab + 4*ewconv.i[1] +2);
 595             /*                 #else */
 596             ewtabFn          = _fjsp_setzero_v2r8();
 597             /*                 #endif */
 598             GMX_FJSP_TRANSPOSE2_V2R8(ewtabV,ewtabFn);
 599             felec            = _fjsp_madd_v2r8(eweps,ewtabD,ewtabF);
 600             /*                 #define INNERFLOPS INNERFLOPS+2 */
 601             /*                 #if KERNEL_MOD_ELEC=='PotentialShift' */
 602             velec            = _fjsp_nmsub_v2r8(_fjsp_mul_v2r8(ewtabhalfspace,eweps) ,_fjsp_add_v2r8(ewtabF,felec), ewtabV);
 603             velec            = _fjsp_mul_v2r8(qq{I}{J},_fjsp_sub_v2r8(_fjsp_sub_v2r8(rinv{I}{J},sh_ewald),velec));
 604             /*                     #define INNERFLOPS INNERFLOPS+7 */
 605             /*                 #else */
 606             velec            = _fjsp_nmsub_v2r8(_fjsp_mul_v2r8(ewtabhalfspace,eweps) ,_fjsp_add_v2r8(ewtabF,felec), ewtabV);
 607             velec            = _fjsp_mul_v2r8(qq{I}{J},_fjsp_sub_v2r8(rinv{I}{J},velec));
 608             /*                     #define INNERFLOPS INNERFLOPS+6 */
 609             /*                 #endif */
 610             /*                 #if 'Force' in KERNEL_VF */
 611             felec            = _fjsp_mul_v2r8(_fjsp_mul_v2r8(qq{I}{J},rinv{I}{J}),_fjsp_sub_v2r8(rinvsq{I}{J},felec));
 612             /*                      #define INNERFLOPS INNERFLOPS+3 */
 613             /*                 #endif */
 614             /*             #elif KERNEL_VF=='Force' */
 615             /*                 #if ROUND == 'Loop' */
 616             gmx_fjsp_load_2pair_swizzle_v2r8(ewtab+ewconv.i[0],ewtab+ewconv.i[1],
 617                                          &ewtabF,&ewtabFn);
 618             /*                 #else */
 619             gmx_fjsp_load_1pair_swizzle_v2r8(ewtab+ewconv.i[0],&ewtabF,&ewtabFn);
 620             /*                 #endif */
 621             felec            = _fjsp_madd_v2r8(eweps,ewtabFn,_fjsp_nmsub_v2r8(eweps,ewtabF,ewtabF));
 622             felec            = _fjsp_mul_v2r8(_fjsp_mul_v2r8(qq{I}{J},rinv{I}{J}),_fjsp_sub_v2r8(rinvsq{I}{J},felec));
 623             /*                 #define INNERFLOPS INNERFLOPS+7 */
 624             /*             #endif */
 625
 626             /*         #elif KERNEL_ELEC=='CubicSplineTable' */
 627
 628             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 629             Y                = _fjsp_load_v2r8( vftab + vfconv.i[0] );
 630             /*             #if ROUND == 'Loop' */
 631             F                = _fjsp_load_v2r8( vftab + vfconv.i[1] );
 632             /*             #else */
 633             F                = _fjsp_setzero_v2r8();
 634             /*             #endif */
 635             GMX_FJSP_TRANSPOSE2_V2R8(Y,F);
 636             G                = _fjsp_load_v2r8( vftab + vfconv.i[0] +2);
 637             /*             #if ROUND == 'Loop' */
 638             H                = _fjsp_load_v2r8( vftab + vfconv.i[1] +2);
 639             /*             #else */
 640             H                = _fjsp_setzero_v2r8();
 641             /*             #endif */
 642             GMX_FJSP_TRANSPOSE2_V2R8(G,H);
 643             Fp               = _fjsp_madd_v2r8(vfeps,_fjsp_madd_v2r8(vfeps,H,G),F);
 644             /*             #define INNERFLOPS INNERFLOPS+4 */
 645             /*             #if 'Potential' in KERNEL_VF */
 646             VV               = _fjsp_madd_v2r8(vfeps,Fp,Y);
 647             velec            = _fjsp_mul_v2r8(qq{I}{J},VV);
 648             /*                 #define INNERFLOPS INNERFLOPS+3 */
 649             /*             #endif */
 650             /*             #if 'Force' in KERNEL_VF */
 651             FF               = _fjsp_madd_v2r8(_fjsp_madd_v2r8(twovfeps,H,G),vfeps,Fp);
 652             felec            = _fjsp_neg_v2r8(_fjsp_mul_v2r8(_fjsp_mul_v2r8(qq{I}{J},FF),_fjsp_mul_v2r8(vftabscale,rinv{I}{J})));
 653             /*                 #define INNERFLOPS INNERFLOPS+7 */
 654             /*             #endif */
 655             /*         #endif */
 656             /*         ## End of check for electrostatics interaction forms */
 657             /*     #endif */
 658             /*     ## END OF ELECTROSTATIC INTERACTION CHECK FOR PAIR I-J */
 659
 660             /*     #if 'vdw' in INTERACTION_FLAGS[I][J] */
 661
 662             /*         #if KERNEL_VDW=='LennardJones' */
 663
 664             /* LENNARD-JONES DISPERSION/REPULSION */
 665
 666             rinvsix          = _fjsp_mul_v2r8(_fjsp_mul_v2r8(rinvsq{I}{J},rinvsq{I}{J}),rinvsq{I}{J});
 667             /*             #define INNERFLOPS INNERFLOPS+2 */
 668             /*             #if 'Potential' in KERNEL_VF or KERNEL_MOD_VDW=='PotentialSwitch' */
 669             vvdw6            = _fjsp_mul_v2r8(c6_{I}{J},rinvsix);
 670             vvdw12           = _fjsp_mul_v2r8(c12_{I}{J},_fjsp_mul_v2r8(rinvsix,rinvsix));
 671             /*                 #define INNERFLOPS INNERFLOPS+3 */
 672             /*                 #if KERNEL_MOD_VDW=='PotentialShift' */
 673             vvdw             = _fjsp_msub_v2r8(_fjsp_nmsub_v2r8(c12_{I}{J},_fjsp_mul_v2r8(sh_vdw_invrcut6,sh_vdw_invrcut6),vvdw12),one_twelfth,
 674                                            _fjsp_mul_v2r8(_fjsp_nmsub_v2r8( c6_{I}{J},sh_vdw_invrcut6,vvdw6),one_sixth));
 675             /*                     #define INNERFLOPS INNERFLOPS+8 */
 676             /*                 #else */
 677             vvdw             = _fjsp_msub_v2r8( vvdw12,one_twelfth, _fjsp_mul_v2r8(vvdw6,one_sixth) );
 678             /*                     #define INNERFLOPS INNERFLOPS+3 */
 679             /*                 #endif */
 680             /*                 ## Check for force inside potential check, i.e. this means we already did the potential part */
 681             /*                 #if 'Force' in KERNEL_VF */
 682             fvdw             = _fjsp_mul_v2r8(_fjsp_sub_v2r8(vvdw12,vvdw6),rinvsq{I}{J});
 683             /*                     #define INNERFLOPS INNERFLOPS+2 */
 684             /*                 #endif */
 685             /*             #elif KERNEL_VF=='Force' */
 686             /*                 ## Force-only LennardJones makes it possible to save 1 flop (they do add up...) */
 687             fvdw             = _fjsp_mul_v2r8(_fjsp_msub_v2r8(c12_{I}{J},rinvsix,c6_{I}{J}),_fjsp_mul_v2r8(rinvsix,rinvsq{I}{J}));
 688             /*                 #define INNERFLOPS INNERFLOPS+4 */
 689             /*             #endif */
 690
 691             /*         #elif KERNEL_VDW=='CubicSplineTable' */
 692
 693             /* CUBIC SPLINE TABLE DISPERSION */
 694             /*             #if 'Table' in KERNEL_ELEC */
 695             vfconv.i[0]       += 4;
 696             vfconv.i[1]       += 4;
 697             /*             #endif                     */
 698             Y                = _fjsp_load_v2r8( vftab + vfconv.i[0] );
 699             /*             #if ROUND == 'Loop' */
 700             F                = _fjsp_load_v2r8( vftab + vfconv.i[1] );
 701             /*             #else */
 702             F                = _fjsp_setzero_v2r8();
 703             /*             #endif */
 704             GMX_FJSP_TRANSPOSE2_V2R8(Y,F);
 705             G                = _fjsp_load_v2r8( vftab + vfconv.i[0] + 2 );
 706             /*             #if ROUND == 'Loop' */
 707             H                = _fjsp_load_v2r8( vftab + vfconv.i[1] + 2 );
 708             /*             #else */
 709             H                = _fjsp_setzero_v2r8();
 710             /*             #endif */
 711             GMX_FJSP_TRANSPOSE2_V2R8(G,H);
 712             Fp               = _fjsp_madd_v2r8(vfeps,_fjsp_madd_v2r8(H,vfeps,G),F);
 713             /*             #define INNERFLOPS INNERFLOPS+4 */
 714             /*             #if 'Potential' in KERNEL_VF */
 715             VV               = _fjsp_madd_v2r8(vfeps,Fp,Y);
 716             vvdw6            = _fjsp_mul_v2r8(c6_{I}{J},VV);
 717             /*                 #define INNERFLOPS INNERFLOPS+3 */
 718             /*             #endif */
 719             /*             #if 'Force' in KERNEL_VF */
 720             FF               = _fjsp_madd_v2r8(vfeps,_fjsp_madd_v2r8(twovfeps,H,G),Fp);
 721             fvdw6            = _fjsp_mul_v2r8(c6_{I}{J},FF);
 722             /*                 #define INNERFLOPS INNERFLOPS+4 */
 723             /*             #endif */
 724
 725             /* CUBIC SPLINE TABLE REPULSION */
 726             Y                = _fjsp_load_v2r8( vftab + vfconv.i[0] + 4 );
 727             /*             #if ROUND == 'Loop' */
 728             F                = _fjsp_load_v2r8( vftab + vfconv.i[1] + 4 );
 729             /*             #else */
 730             F                = _fjsp_setzero_v2r8();
 731             /*             #endif */
 732             GMX_FJSP_TRANSPOSE2_V2R8(Y,F);
 733             G                = _fjsp_load_v2r8( vftab + vfconv.i[0] + 6 );
 734             /*             #if ROUND == 'Loop' */
 735             H                = _fjsp_load_v2r8( vftab + vfconv.i[1] + 6 );
 736             /*             #else */
 737             H                = _fjsp_setzero_v2r8();
 738             /*             #endif */
 739             GMX_FJSP_TRANSPOSE2_V2R8(G,H);
 740             Fp               = _fjsp_madd_v2r8(vfeps,_fjsp_madd_v2r8(H,vfeps,G),F);
 741             /*             #define INNERFLOPS INNERFLOPS+4 */
 742             /*             #if 'Potential' in KERNEL_VF */
 743             VV               = _fjsp_madd_v2r8(vfeps,Fp,Y);
 744             vvdw12           = _fjsp_mul_v2r8(c12_{I}{J},VV);
 745             /*                 #define INNERFLOPS INNERFLOPS+3 */
 746             /*             #endif */
 747             /*             #if 'Force' in KERNEL_VF */
 748             FF               = _fjsp_madd_v2r8(vfeps,_fjsp_madd_v2r8(twovfeps,H,G),Fp);
 749             fvdw12           = _fjsp_mul_v2r8(c12_{I}{J},FF);
 750             /*                 #define INNERFLOPS INNERFLOPS+5 */
 751             /*             #endif */
 752             /*             #if 'Potential' in KERNEL_VF */
 753             vvdw             = _fjsp_add_v2r8(vvdw12,vvdw6);
 754             /*                 #define INNERFLOPS INNERFLOPS+1 */
 755             /*             #endif */
 756             /*             #if 'Force' in KERNEL_VF */
 757             fvdw             = _fjsp_neg_v2r8(_fjsp_mul_v2r8(_fjsp_add_v2r8(fvdw6,fvdw12),_fjsp_mul_v2r8(vftabscale,rinv{I}{J})));
 758             /*                 #define INNERFLOPS INNERFLOPS+4 */
 759             /*             #endif */
 760
 761             /*         #elif KERNEL_VDW=='LJEwald' */
 762
 763             /* Analytical LJ-PME */
 764             rinvsix          = _fjsp_mul_v2r8(_fjsp_mul_v2r8(rinvsq{I}{J},rinvsq{I}{J}),rinvsq{I}{J});
 765             ewcljrsq         = _fjsp_mul_v2r8(ewclj2,rsq{I}{J});
 766             ewclj6           = _fjsp_mul_v2r8(ewclj2,_fjsp_mul_v2r8(ewclj2,ewclj2));
 767             exponent         = gmx_simd_exp_d(ewcljrsq);
 768             /* poly = exp(-(beta*r)^2) * (1 + (beta*r)^2 + (beta*r)^4 /2) */
 769             poly             = _fjsp_mul_v2r8(exponent,_fjsp_madd_v2r8(_fjsp_mul_v2r8(ewcljrsq,ewcljrsq),one_half,_fjsp_sub_v2r8(one,ewcljrsq)));
 770             /*                 #define INNERFLOPS INNERFLOPS+9 */
 771             /*             #if 'Potential' in KERNEL_VF or KERNEL_MOD_VDW=='PotentialSwitch' */
 772             /* vvdw6 = [C6 - C6grid * (1-poly)]/r6 */
 773             vvdw6            = _fjsp_mul_v2r8(_fjsp_madd_v2r8(c6grid_{I}{J},_fjsp_sub_v2r8(poly,one),c6_{I}{J}),rinvsix);
 774             vvdw12           = _fjsp_mul_v2r8(c12_{I}{J},_fjsp_mul_v2r8(rinvsix,rinvsix));
 775             /*                 #define INNERFLOPS INNERFLOPS+5 */
 776             /*                 #if KERNEL_MOD_VDW=='PotentialShift' */
 777             vvdw             = _fjsp_msub_v2r8(_fjsp_nmsub_v2r8(c12_{I}{J},_fjsp_mul_v2r8(sh_vdw_invrcut6,sh_vdw_invrcut6),vvdw12),one_twelfth,
 778                                _fjsp_mul_v2r8(_fjsp_sub_v2r8(vvdw6,_fjsp_madd_v2r8(c6grid_{I}{J},sh_lj_ewald,_fjsp_mul_v2r8(c6_{I}{J},sh_vdw_invrcut6))),one_sixth));
 779             /*                     #define INNERFLOPS INNERFLOPS+7 */
 780             /*                 #else */
 781             vvdw             = _fjsp_msub_v2r8(vvdw12,one_twelfth,_fjsp_mul_v2r8(vvdw6,one_sixth));
 782             /*                 #define INNERFLOPS INNERFLOPS+2 */
 783             /*                 #endif */
 784             /*                  ## Check for force inside potential check, i.e. this means we already did the potential part */
 785             /*                  #if 'Force' in KERNEL_VF */
 786             /* fvdw = vvdw12/r - (vvdw6/r + (C6grid * exponent * beta^6)/r) */
 787             fvdw             = _fjsp_mul_v2r8(_fjsp_add_v2r8(vvdw12,_fjsp_msub_v2r8(_fjsp_mul_v2r8(c6grid_{I}{J},one_sixth),_fjsp_mul_v2r8(exponent,ewclj6),vvdw6)),rinvsq{I}{J});
 788             /*                 #define INNERFLOPS INNERFLOPS+6 */
 789             /*                  #endif */
 790             /*              #elif KERNEL_VF=='Force' */
 791             /* f6A = 6 * C6grid * (1 - poly) */
 792             f6A              = _fjsp_mul_v2r8(c6grid_{I}{J},_fjsp_sub_v2r8(one,poly));
 793             /* f6B = C6grid * exponent * beta^6 */
 794             f6B              = _fjsp_mul_v2r8(_fjsp_mul_v2r8(c6grid_{I}{J},one_sixth),_fjsp_mul_v2r8(exponent,ewclj6));
 795             /* fvdw = 12*C12/r13 - ((6*C6 - f6A)/r6 + f6B)/r */
 796             fvdw              = _fjsp_mul_v2r8(_fjsp_madd_v2r8(_fjsp_msub_v2r8(c12_{I}{J},rinvsix,_fjsp_sub_v2r8(c6_{I}{J},f6A)),rinvsix,f6B),rinvsq{I}{J});
 797             /*                 #define INNERFLOPS INNERFLOPS+12 */
 798             /*              #endif */
 799             /*         #endif */
 800             /*         ## End of check for vdw interaction forms */
 801             /*     #endif */
 802             /*     ## END OF VDW INTERACTION CHECK FOR PAIR I-J */
 803
 804             /*     #if 'switch' in INTERACTION_FLAGS[I][J] */
 805             d                = _fjsp_sub_v2r8(r{I}{J},rswitch);
 806             d                = _fjsp_max_v2r8(d,_fjsp_setzero_v2r8());
 807             d2               = _fjsp_mul_v2r8(d,d);
 808             sw               = _fjsp_add_v2r8(one,_fjsp_mul_v2r8(d2,_fjsp_mul_v2r8(d,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swV5,swV4),swV3))));
 809             /*         #define INNERFLOPS INNERFLOPS+10 */
 810
 811             /*         #if 'Force' in KERNEL_VF */
 812             dsw              = _fjsp_mul_v2r8(d2,_fjsp_madd_v2r8(d,_fjsp_madd_v2r8(d,swF4,swF3),swF2));
 813             /*             #define INNERFLOPS INNERFLOPS+5 */
 814             /*         #endif */
 815
 816             /* Evaluate switch function */
 817             /*         #if 'Force' in KERNEL_VF */
 818             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 819             /*             #if 'electrostatics' in INTERACTION_FLAGS[I][J] and KERNEL_MOD_ELEC=='PotentialSwitch' */
 820             felec            = _fjsp_msub_v2r8( felec,sw , _fjsp_mul_v2r8(rinv{I}{J},_fjsp_mul_v2r8(velec,dsw)) );
 821             /*                 #define INNERFLOPS INNERFLOPS+4 */
 822             /*             #endif */
 823             /*             #if 'vdw' in INTERACTION_FLAGS[I][J] and KERNEL_MOD_VDW=='PotentialSwitch' */
 824             fvdw             = _fjsp_msub_v2r8( fvdw,sw , _fjsp_mul_v2r8(rinv{I}{J},_fjsp_mul_v2r8(vvdw,dsw)) );
 825             /*                 #define INNERFLOPS INNERFLOPS+4 */
 826             /*             #endif */
 827             /*         #endif */
 828             /*         #if 'Potential' in KERNEL_VF */
 829             /*             #if 'electrostatics' in INTERACTION_FLAGS[I][J] and KERNEL_MOD_ELEC=='PotentialSwitch' */
 830             velec            = _fjsp_mul_v2r8(velec,sw);
 831             /*                 #define INNERFLOPS INNERFLOPS+1 */
 832             /*             #endif */
 833             /*             #if 'vdw' in INTERACTION_FLAGS[I][J] and KERNEL_MOD_VDW=='PotentialSwitch' */
 834             vvdw             = _fjsp_mul_v2r8(vvdw,sw);
 835             /*                 #define INNERFLOPS INNERFLOPS+1 */
 836             /*             #endif */
 837             /*         #endif */
 838             /*     #endif */
 839             /*     #if 'exactcutoff' in INTERACTION_FLAGS[I][J] */
 840             cutoff_mask      = _fjsp_cmplt_v2r8(rsq{I}{J},rcutoff2);
 841             /*         #define INNERFLOPS INNERFLOPS+1 */
 842             /*     #endif */
 843
 844             /*     #if 'Potential' in KERNEL_VF */
 845             /* Update potential sum for this i atom from the interaction with this j atom. */
 846             /*         #if 'electrostatics' in INTERACTION_FLAGS[I][J] */
 847             /*             #if 'exactcutoff' in INTERACTION_FLAGS[I][J] */
 848             velec            = _fjsp_and_v2r8(velec,cutoff_mask);
 849             /*                 #define INNERFLOPS INNERFLOPS+1 */
 850             /*             #endif                                       */
 851             /*             #if ROUND == 'Epilogue' */
 852             velec            = _fjsp_unpacklo_v2r8(velec,_fjsp_setzero_v2r8());
 853             /*             #endif */
 854             velecsum         = _fjsp_add_v2r8(velecsum,velec);
 855             /*             #define INNERFLOPS INNERFLOPS+1 */
 856             /*         #endif */
 857             /*         #if 'vdw' in INTERACTION_FLAGS[I][J] */
 858             /*             #if 'exactcutoff' in INTERACTION_FLAGS[I][J] */
 859             vvdw             = _fjsp_and_v2r8(vvdw,cutoff_mask);
 860             /*                 #define INNERFLOPS INNERFLOPS+1 */
 861             /*             #endif                                       */
 862             /*             #if ROUND == 'Epilogue' */
 863             vvdw             = _fjsp_unpacklo_v2r8(vvdw,_fjsp_setzero_v2r8());
 864             /*             #endif */
 865             vvdwsum          = _fjsp_add_v2r8(vvdwsum,vvdw);
 866             /*             #define INNERFLOPS INNERFLOPS+1 */
 867             /*         #endif */
 868             /*     #endif */
 869
 870             /*     #if 'Force' in KERNEL_VF */
 871
 872             /*         #if 'electrostatics' in INTERACTION_FLAGS[I][J] and 'vdw' in INTERACTION_FLAGS[I][J] */
 873             fscal            = _fjsp_add_v2r8(felec,fvdw);
 874             /*             #define INNERFLOPS INNERFLOPS+1 */
 875             /*         #elif 'electrostatics' in INTERACTION_FLAGS[I][J] */
 876             fscal            = felec;
 877             /*         #elif 'vdw' in INTERACTION_FLAGS[I][J] */
 878             fscal            = fvdw;
 879             /*        #endif */
 880
 881             /*             #if 'exactcutoff' in INTERACTION_FLAGS[I][J] */
 882             fscal            = _fjsp_and_v2r8(fscal,cutoff_mask);
 883             /*                 #define INNERFLOPS INNERFLOPS+1 */
 884             /*             #endif                                       */
 885
 886             /*             #if ROUND == 'Epilogue' */
 887             fscal            = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
 888             /*             #endif */
 889
 890             /* ## Construction of vectorial force built into FMA instructions now */
 891             /* #define INNERFLOPS INNERFLOPS+3      */
 892
 893             /* Update vectorial force */
 894             fix{I}             = _fjsp_madd_v2r8(dx{I}{J},fscal,fix{I});
 895             fiy{I}             = _fjsp_madd_v2r8(dy{I}{J},fscal,fiy{I});
 896             fiz{I}             = _fjsp_madd_v2r8(dz{I}{J},fscal,fiz{I});
 897             /*             #define INNERFLOPS INNERFLOPS+6 */
 898
 899             /* #if GEOMETRY_I == 'Particle'             */
 900             /*     #if ROUND == 'Loop' */
 901             gmx_fjsp_decrement_fma_1rvec_2ptr_swizzle_v2r8(f+j_coord_offsetA,f+j_coord_offsetB,fscal,dx{I}{J},dy{I}{J},dz{I}{J});
 902             /*     #else */
 903             gmx_fjsp_decrement_fma_1rvec_1ptr_swizzle_v2r8(f+j_coord_offsetA,fscal,dx{I}{J},dy{I}{J},dz{I}{J});
 904             /*     #endif */
 905             /*     #define INNERFLOPS INNERFLOPS+3      */
 906             /* #else                                    */
 907             fjx{J}             = _fjsp_madd_v2r8(dx{I}{J},fscal,fjx{J});
 908             fjy{J}             = _fjsp_madd_v2r8(dy{I}{J},fscal,fjy{J});
 909             fjz{J}             = _fjsp_madd_v2r8(dz{I}{J},fscal,fjz{J});
 910             /*     #define INNERFLOPS INNERFLOPS+3      */
 911             /* #endif                                   */
 912
 913             /*     #endif */
 914
 915             /*     #if 'exactcutoff' in INTERACTION_FLAGS[I][J] */
 916             /*         #if 0    ## This and next two lines is a hack to maintain indentation in template file */
 917             {
 918                 /*     #endif */
 919             }
 920             /*     #endif */
 921             /*    ## End of check for the interaction being outside the cutoff */
 922
 923             /* #endfor */
 924             /* ## End of loop over i-j interaction pairs */
 925
 926             /* #if 'Water' in GEOMETRY_I and GEOMETRY_J == 'Particle' */
 927             /*     #if ROUND == 'Loop' */
 928             gmx_fjsp_decrement_1rvec_2ptr_swizzle_v2r8(f+j_coord_offsetA,f+j_coord_offsetB,fjx0,fjy0,fjz0);
 929             /*     #else */
 930             gmx_fjsp_decrement_1rvec_1ptr_swizzle_v2r8(f+j_coord_offsetA,fjx0,fjy0,fjz0);
 931             /*     #endif */
 932             /*     #define INNERFLOPS INNERFLOPS+3      */
 933             /* #elif GEOMETRY_J == 'Water3'             */
 934             /*     #if ROUND == 'Loop' */
 935             gmx_fjsp_decrement_3rvec_2ptr_swizzle_v2r8(f+j_coord_offsetA,f+j_coord_offsetB,fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2);
 936             /*     #else */
 937             gmx_fjsp_decrement_3rvec_1ptr_swizzle_v2r8(f+j_coord_offsetA,fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2);
 938             /*     #endif */
 939             /*     #define INNERFLOPS INNERFLOPS+9      */
 940             /* #elif GEOMETRY_J == 'Water4'             */
 941             /*     #if 0 in PARTICLES_J                 */
 942             /*         #if ROUND == 'Loop' */
 943             gmx_fjsp_decrement_4rvec_2ptr_swizzle_v2r8(f+j_coord_offsetA,f+j_coord_offsetB,fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2,fjx3,fjy3,fjz3);
 944             /*         #else */
 945             gmx_fjsp_decrement_4rvec_1ptr_swizzle_v2r8(f+j_coord_offsetA,fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2,fjx3,fjy3,fjz3);
 946             /*         #endif */
 947             /*         #define INNERFLOPS INNERFLOPS+12 */
 948             /*     #else                                */
 949             /*         #if ROUND == 'Loop' */
 950             gmx_fjsp_decrement_3rvec_2ptr_swizzle_v2r8(f+j_coord_offsetA+DIM,f+j_coord_offsetB+DIM,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2,fjx3,fjy3,fjz3);
 951             /*         #else */
 952             gmx_fjsp_decrement_3rvec_1ptr_swizzle_v2r8(f+j_coord_offsetA+DIM,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2,fjx3,fjy3,fjz3);
 953             /*         #endif */
 954             /*         #define INNERFLOPS INNERFLOPS+9  */
 955             /*     #endif                               */
 956             /* #endif                                   */
 957
 958             /* Inner loop uses {INNERFLOPS} flops */
 959         }
 960
 961         /* #endfor */
 962
 963         /* End of innermost loop */
 964
 965         /* #if 'Force' in KERNEL_VF */
 966         /*     #if GEOMETRY_I == 'Particle'            */
 967         gmx_fjsp_update_iforce_1atom_swizzle_v2r8(fix0,fiy0,fiz0,
 968                                               f+i_coord_offset,fshift+i_shift_offset);
 969         /*         #define OUTERFLOPS OUTERFLOPS+6     */
 970         /*     #elif GEOMETRY_I == 'Water3'            */
 971         gmx_fjsp_update_iforce_3atom_swizzle_v2r8(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,
 972                                               f+i_coord_offset,fshift+i_shift_offset);
 973         /*         #define OUTERFLOPS OUTERFLOPS+18    */
 974         /*     #elif GEOMETRY_I == 'Water4'            */
 975         /*         #if 0 in PARTICLES_I                */
 976         gmx_fjsp_update_iforce_4atom_swizzle_v2r8(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
 977                                               f+i_coord_offset,fshift+i_shift_offset);
 978         /*             #define OUTERFLOPS OUTERFLOPS+24    */
 979         /*         #else                               */
 980         gmx_fjsp_update_iforce_3atom_swizzle_v2r8(fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
 981                                               f+i_coord_offset+DIM,fshift+i_shift_offset);
 982         /*             #define OUTERFLOPS OUTERFLOPS+18    */
 983         /*         #endif                              */
 984         /*     #endif                                  */
 985         /* #endif                                      */
 986
 987         /* #if 'Potential' in KERNEL_VF */
 988         ggid                        = gid[iidx];
 989         /* Update potential energies */
 990         /*     #if KERNEL_ELEC != 'None' */
 991         gmx_fjsp_update_1pot_v2r8(velecsum,kernel_data->energygrp_elec+ggid);
 992         /*         #define OUTERFLOPS OUTERFLOPS+1 */
 993         /*     #endif */
 994         /*     #if KERNEL_VDW != 'None' */
 995         gmx_fjsp_update_1pot_v2r8(vvdwsum,kernel_data->energygrp_vdw+ggid);
 996         /*         #define OUTERFLOPS OUTERFLOPS+1 */
 997         /*     #endif */
 998         /* #endif */
 999
1000         /* Increment number of inner iterations */
1001         inneriter                  += j_index_end - j_index_start;
1002
1003         /* Outer loop uses {OUTERFLOPS} flops */
1004     }
1005
1006     /* Increment number of outer iterations */
1007     outeriter        += nri;
1008
1009     /* Update outer/inner flops */
1010     /* ## NB: This is not important, it just affects the flopcount. However, since our preprocessor is */
1011     /* ## primitive and replaces aggressively even in strings inside these directives, we need to      */
1012     /* ## assemble the main part of the name (containing KERNEL/ELEC/VDW) directly in the source.      */
1013     /* #if GEOMETRY_I == 'Water3'            */
1014     /*     #define ISUFFIX '_W3'             */
1015     /* #elif GEOMETRY_I == 'Water4'          */
1016     /*     #define ISUFFIX '_W4'             */
1017     /* #else                                 */
1018     /*     #define ISUFFIX ''                */
1019     /* #endif                                */
1020     /* #if GEOMETRY_J == 'Water3'            */
1021     /*     #define JSUFFIX 'W3'              */
1022     /* #elif GEOMETRY_J == 'Water4'          */
1023     /*     #define JSUFFIX 'W4'              */
1024     /* #else                                 */
1025     /*     #define JSUFFIX ''                */
1026     /* #endif                                */
1027     /* #if 'PotentialAndForce' in KERNEL_VF  */
1028     /*     #define VFSUFFIX  '_VF'           */
1029     /* #elif 'Potential' in KERNEL_VF        */
1030     /*     #define VFSUFFIX '_V'             */
1031     /* #else                                 */
1032     /*     #define VFSUFFIX '_F'             */
1033     /* #endif                                */
1034
1035     /* #if KERNEL_ELEC != 'None' and KERNEL_VDW != 'None' */
1036     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW{ISUFFIX}{JSUFFIX}{VFSUFFIX},outeriter*{OUTERFLOPS} + inneriter*{INNERFLOPS});
1037     /* #elif KERNEL_ELEC != 'None' */
1038     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC{ISUFFIX}{JSUFFIX}{VFSUFFIX},outeriter*{OUTERFLOPS} + inneriter*{INNERFLOPS});
1039     /* #else */
1040     inc_nrnb(nrnb,eNR_NBKERNEL_VDW{ISUFFIX}{JSUFFIX}{VFSUFFIX},outeriter*{OUTERFLOPS} + inneriter*{INNERFLOPS});
1041     /* #endif  */
1042 }