src/gromacs/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecCSTab_VdwBham_GeomW4P1_c.c

   1 /*
   2  * This file is part of the GROMACS molecular simulation package.
   3  *
   4  * Copyright (c) 2012,2013, by the GROMACS development team, led by
   5  * Mark Abraham, David van der Spoel, Berk Hess, and Erik Lindahl,
   6  * and including many others, as listed in the AUTHORS file in the
   7  * top-level source directory and at http://www.gromacs.org.
   8  *
   9  * GROMACS is free software; you can redistribute it and/or
  10  * modify it under the terms of the GNU Lesser General Public License
  11  * as published by the Free Software Foundation; either version 2.1
  12  * of the License, or (at your option) any later version.
  13  *
  14  * GROMACS is distributed in the hope that it will be useful,
  15  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  16  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  17  * Lesser General Public License for more details.
  18  *
  19  * You should have received a copy of the GNU Lesser General Public
  20  * License along with GROMACS; if not, see
  21  * http://www.gnu.org/licenses, or write to the Free Software Foundation,
  22  * Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA.
  23  *
  24  * If you want to redistribute modifications to GROMACS, please
  25  * consider that scientific software is very special. Version
  26  * control is crucial - bugs must be traceable. We will be happy to
  27  * consider code for inclusion in the official distribution, but
  28  * derived work must not be called official GROMACS. Details are found
  29  * in the README & COPYING files - if they are missing, get the
  30  * official version at http://www.gromacs.org.
  31  *
  32  * To help us fund GROMACS development, we humbly ask that you cite
  33  * the research papers on the package. Check out http://www.gromacs.org.
  34  */
  35 /*
  36  * Note: this file was generated by the GROMACS c kernel generator.
  37  */
  38 #ifdef HAVE_CONFIG_H
  39 #include <config.h>
  40 #endif
  41
  42 #include <math.h>
  43
  44 #include "../nb_kernel.h"
  45 #include "types/simple.h"
  46 #include "vec.h"
  47 #include "nrnb.h"
  48
  49 /*
  50  * Gromacs nonbonded kernel:   nb_kernel_ElecCSTab_VdwBham_GeomW4P1_VF_c
  51  * Electrostatics interaction: CubicSplineTable
  52  * VdW interaction:            Buckingham
  53  * Geometry:                   Water4-Particle
  54  * Calculate force/pot:        PotentialAndForce
  55  */
  56 void
  57 nb_kernel_ElecCSTab_VdwBham_GeomW4P1_VF_c
  58                     (t_nblist                    * gmx_restrict       nlist,
  59                      rvec                        * gmx_restrict          xx,
  60                      rvec                        * gmx_restrict          ff,
  61                      t_forcerec                  * gmx_restrict          fr,
  62                      t_mdatoms                   * gmx_restrict     mdatoms,
  63                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
  64                      t_nrnb                      * gmx_restrict        nrnb)
  65 {
  66     int              i_shift_offset,i_coord_offset,j_coord_offset;
  67     int              j_index_start,j_index_end;
  68     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
  69     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
  70     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  71     real             *shiftvec,*fshift,*x,*f;
  72     int              vdwioffset0;
  73     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  74     int              vdwioffset1;
  75     real             ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
  76     int              vdwioffset2;
  77     real             ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
  78     int              vdwioffset3;
  79     real             ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
  80     int              vdwjidx0;
  81     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  82     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
  83     real             dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10;
  84     real             dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20;
  85     real             dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30,cexp1_30,cexp2_30;
  86     real             velec,felec,velecsum,facel,crf,krf,krf2;
  87     real             *charge;
  88     int              nvdwtype;
  89     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
  90     int              *vdwtype;
  91     real             *vdwparam;
  92     int              vfitab;
  93     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
  94     real             *vftab;
  95
  96     x                = xx[0];
  97     f                = ff[0];
  98
  99     nri              = nlist->nri;
 100     iinr             = nlist->iinr;
 101     jindex           = nlist->jindex;
 102     jjnr             = nlist->jjnr;
 103     shiftidx         = nlist->shift;
 104     gid              = nlist->gid;
 105     shiftvec         = fr->shift_vec[0];
 106     fshift           = fr->fshift[0];
 107     facel            = fr->epsfac;
 108     charge           = mdatoms->chargeA;
 109     nvdwtype         = fr->ntype;
 110     vdwparam         = fr->nbfp;
 111     vdwtype          = mdatoms->typeA;
 112
 113     vftab            = kernel_data->table_elec->data;
 114     vftabscale       = kernel_data->table_elec->scale;
 115
 116     /* Setup water-specific parameters */
 117     inr              = nlist->iinr[0];
 118     iq1              = facel*charge[inr+1];
 119     iq2              = facel*charge[inr+2];
 120     iq3              = facel*charge[inr+3];
 121     vdwioffset0      = 3*nvdwtype*vdwtype[inr+0];
 122
 123     outeriter        = 0;
 124     inneriter        = 0;
 125
 126     /* Start outer loop over neighborlists */
 127     for(iidx=0; iidx<nri; iidx++)
 128     {
 129         /* Load shift vector for this list */
 130         i_shift_offset   = DIM*shiftidx[iidx];
 131         shX              = shiftvec[i_shift_offset+XX];
 132         shY              = shiftvec[i_shift_offset+YY];
 133         shZ              = shiftvec[i_shift_offset+ZZ];
 134
 135         /* Load limits for loop over neighbors */
 136         j_index_start    = jindex[iidx];
 137         j_index_end      = jindex[iidx+1];
 138
 139         /* Get outer coordinate index */
 140         inr              = iinr[iidx];
 141         i_coord_offset   = DIM*inr;
 142
 143         /* Load i particle coords and add shift vector */
 144         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 145         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 146         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 147         ix1              = shX + x[i_coord_offset+DIM*1+XX];
 148         iy1              = shY + x[i_coord_offset+DIM*1+YY];
 149         iz1              = shZ + x[i_coord_offset+DIM*1+ZZ];
 150         ix2              = shX + x[i_coord_offset+DIM*2+XX];
 151         iy2              = shY + x[i_coord_offset+DIM*2+YY];
 152         iz2              = shZ + x[i_coord_offset+DIM*2+ZZ];
 153         ix3              = shX + x[i_coord_offset+DIM*3+XX];
 154         iy3              = shY + x[i_coord_offset+DIM*3+YY];
 155         iz3              = shZ + x[i_coord_offset+DIM*3+ZZ];
 156
 157         fix0             = 0.0;
 158         fiy0             = 0.0;
 159         fiz0             = 0.0;
 160         fix1             = 0.0;
 161         fiy1             = 0.0;
 162         fiz1             = 0.0;
 163         fix2             = 0.0;
 164         fiy2             = 0.0;
 165         fiz2             = 0.0;
 166         fix3             = 0.0;
 167         fiy3             = 0.0;
 168         fiz3             = 0.0;
 169
 170         /* Reset potential sums */
 171         velecsum         = 0.0;
 172         vvdwsum          = 0.0;
 173
 174         /* Start inner kernel loop */
 175         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 176         {
 177             /* Get j neighbor index, and coordinate index */
 178             jnr              = jjnr[jidx];
 179             j_coord_offset   = DIM*jnr;
 180
 181             /* load j atom coordinates */
 182             jx0              = x[j_coord_offset+DIM*0+XX];
 183             jy0              = x[j_coord_offset+DIM*0+YY];
 184             jz0              = x[j_coord_offset+DIM*0+ZZ];
 185
 186             /* Calculate displacement vector */
 187             dx00             = ix0 - jx0;
 188             dy00             = iy0 - jy0;
 189             dz00             = iz0 - jz0;
 190             dx10             = ix1 - jx0;
 191             dy10             = iy1 - jy0;
 192             dz10             = iz1 - jz0;
 193             dx20             = ix2 - jx0;
 194             dy20             = iy2 - jy0;
 195             dz20             = iz2 - jz0;
 196             dx30             = ix3 - jx0;
 197             dy30             = iy3 - jy0;
 198             dz30             = iz3 - jz0;
 199
 200             /* Calculate squared distance and things based on it */
 201             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 202             rsq10            = dx10*dx10+dy10*dy10+dz10*dz10;
 203             rsq20            = dx20*dx20+dy20*dy20+dz20*dz20;
 204             rsq30            = dx30*dx30+dy30*dy30+dz30*dz30;
 205
 206             rinv00           = gmx_invsqrt(rsq00);
 207             rinv10           = gmx_invsqrt(rsq10);
 208             rinv20           = gmx_invsqrt(rsq20);
 209             rinv30           = gmx_invsqrt(rsq30);
 210
 211             rinvsq00         = rinv00*rinv00;
 212
 213             /* Load parameters for j particles */
 214             jq0              = charge[jnr+0];
 215             vdwjidx0         = 3*vdwtype[jnr+0];
 216
 217             /**************************
 218              * CALCULATE INTERACTIONS *
 219              **************************/
 220
 221             r00              = rsq00*rinv00;
 222
 223             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 224             cexp1_00         = vdwparam[vdwioffset0+vdwjidx0+1];
 225             cexp2_00         = vdwparam[vdwioffset0+vdwjidx0+2];
 226
 227             /* BUCKINGHAM DISPERSION/REPULSION */
 228             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 229             vvdw6            = c6_00*rinvsix;
 230             br               = cexp2_00*r00;
 231             vvdwexp          = cexp1_00*exp(-br);
 232             vvdw             = vvdwexp - vvdw6*(1.0/6.0);
 233             fvdw             = (br*vvdwexp-vvdw6)*rinvsq00;
 234
 235             /* Update potential sums from outer loop */
 236             vvdwsum         += vvdw;
 237
 238             fscal            = fvdw;
 239
 240             /* Calculate temporary vectorial force */
 241             tx               = fscal*dx00;
 242             ty               = fscal*dy00;
 243             tz               = fscal*dz00;
 244
 245             /* Update vectorial force */
 246             fix0            += tx;
 247             fiy0            += ty;
 248             fiz0            += tz;
 249             f[j_coord_offset+DIM*0+XX] -= tx;
 250             f[j_coord_offset+DIM*0+YY] -= ty;
 251             f[j_coord_offset+DIM*0+ZZ] -= tz;
 252
 253             /**************************
 254              * CALCULATE INTERACTIONS *
 255              **************************/
 256
 257             r10              = rsq10*rinv10;
 258
 259             qq10             = iq1*jq0;
 260
 261             /* Calculate table index by multiplying r with table scale and truncate to integer */
 262             rt               = r10*vftabscale;
 263             vfitab           = rt;
 264             vfeps            = rt-vfitab;
 265             vfitab           = 1*4*vfitab;
 266
 267             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 268             Y                = vftab[vfitab];
 269             F                = vftab[vfitab+1];
 270             Geps             = vfeps*vftab[vfitab+2];
 271             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 272             Fp               = F+Geps+Heps2;
 273             VV               = Y+vfeps*Fp;
 274             velec            = qq10*VV;
 275             FF               = Fp+Geps+2.0*Heps2;
 276             felec            = -qq10*FF*vftabscale*rinv10;
 277
 278             /* Update potential sums from outer loop */
 279             velecsum        += velec;
 280
 281             fscal            = felec;
 282
 283             /* Calculate temporary vectorial force */
 284             tx               = fscal*dx10;
 285             ty               = fscal*dy10;
 286             tz               = fscal*dz10;
 287
 288             /* Update vectorial force */
 289             fix1            += tx;
 290             fiy1            += ty;
 291             fiz1            += tz;
 292             f[j_coord_offset+DIM*0+XX] -= tx;
 293             f[j_coord_offset+DIM*0+YY] -= ty;
 294             f[j_coord_offset+DIM*0+ZZ] -= tz;
 295
 296             /**************************
 297              * CALCULATE INTERACTIONS *
 298              **************************/
 299
 300             r20              = rsq20*rinv20;
 301
 302             qq20             = iq2*jq0;
 303
 304             /* Calculate table index by multiplying r with table scale and truncate to integer */
 305             rt               = r20*vftabscale;
 306             vfitab           = rt;
 307             vfeps            = rt-vfitab;
 308             vfitab           = 1*4*vfitab;
 309
 310             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 311             Y                = vftab[vfitab];
 312             F                = vftab[vfitab+1];
 313             Geps             = vfeps*vftab[vfitab+2];
 314             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 315             Fp               = F+Geps+Heps2;
 316             VV               = Y+vfeps*Fp;
 317             velec            = qq20*VV;
 318             FF               = Fp+Geps+2.0*Heps2;
 319             felec            = -qq20*FF*vftabscale*rinv20;
 320
 321             /* Update potential sums from outer loop */
 322             velecsum        += velec;
 323
 324             fscal            = felec;
 325
 326             /* Calculate temporary vectorial force */
 327             tx               = fscal*dx20;
 328             ty               = fscal*dy20;
 329             tz               = fscal*dz20;
 330
 331             /* Update vectorial force */
 332             fix2            += tx;
 333             fiy2            += ty;
 334             fiz2            += tz;
 335             f[j_coord_offset+DIM*0+XX] -= tx;
 336             f[j_coord_offset+DIM*0+YY] -= ty;
 337             f[j_coord_offset+DIM*0+ZZ] -= tz;
 338
 339             /**************************
 340              * CALCULATE INTERACTIONS *
 341              **************************/
 342
 343             r30              = rsq30*rinv30;
 344
 345             qq30             = iq3*jq0;
 346
 347             /* Calculate table index by multiplying r with table scale and truncate to integer */
 348             rt               = r30*vftabscale;
 349             vfitab           = rt;
 350             vfeps            = rt-vfitab;
 351             vfitab           = 1*4*vfitab;
 352
 353             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 354             Y                = vftab[vfitab];
 355             F                = vftab[vfitab+1];
 356             Geps             = vfeps*vftab[vfitab+2];
 357             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 358             Fp               = F+Geps+Heps2;
 359             VV               = Y+vfeps*Fp;
 360             velec            = qq30*VV;
 361             FF               = Fp+Geps+2.0*Heps2;
 362             felec            = -qq30*FF*vftabscale*rinv30;
 363
 364             /* Update potential sums from outer loop */
 365             velecsum        += velec;
 366
 367             fscal            = felec;
 368
 369             /* Calculate temporary vectorial force */
 370             tx               = fscal*dx30;
 371             ty               = fscal*dy30;
 372             tz               = fscal*dz30;
 373
 374             /* Update vectorial force */
 375             fix3            += tx;
 376             fiy3            += ty;
 377             fiz3            += tz;
 378             f[j_coord_offset+DIM*0+XX] -= tx;
 379             f[j_coord_offset+DIM*0+YY] -= ty;
 380             f[j_coord_offset+DIM*0+ZZ] -= tz;
 381
 382             /* Inner loop uses 187 flops */
 383         }
 384         /* End of innermost loop */
 385
 386         tx = ty = tz = 0;
 387         f[i_coord_offset+DIM*0+XX] += fix0;
 388         f[i_coord_offset+DIM*0+YY] += fiy0;
 389         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 390         tx                         += fix0;
 391         ty                         += fiy0;
 392         tz                         += fiz0;
 393         f[i_coord_offset+DIM*1+XX] += fix1;
 394         f[i_coord_offset+DIM*1+YY] += fiy1;
 395         f[i_coord_offset+DIM*1+ZZ] += fiz1;
 396         tx                         += fix1;
 397         ty                         += fiy1;
 398         tz                         += fiz1;
 399         f[i_coord_offset+DIM*2+XX] += fix2;
 400         f[i_coord_offset+DIM*2+YY] += fiy2;
 401         f[i_coord_offset+DIM*2+ZZ] += fiz2;
 402         tx                         += fix2;
 403         ty                         += fiy2;
 404         tz                         += fiz2;
 405         f[i_coord_offset+DIM*3+XX] += fix3;
 406         f[i_coord_offset+DIM*3+YY] += fiy3;
 407         f[i_coord_offset+DIM*3+ZZ] += fiz3;
 408         tx                         += fix3;
 409         ty                         += fiy3;
 410         tz                         += fiz3;
 411         fshift[i_shift_offset+XX]  += tx;
 412         fshift[i_shift_offset+YY]  += ty;
 413         fshift[i_shift_offset+ZZ]  += tz;
 414
 415         ggid                        = gid[iidx];
 416         /* Update potential energies */
 417         kernel_data->energygrp_elec[ggid] += velecsum;
 418         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 419
 420         /* Increment number of inner iterations */
 421         inneriter                  += j_index_end - j_index_start;
 422
 423         /* Outer loop uses 41 flops */
 424     }
 425
 426     /* Increment number of outer iterations */
 427     outeriter        += nri;
 428
 429     /* Update outer/inner flops */
 430
 431     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_VF,outeriter*41 + inneriter*187);
 432 }
 433 /*
 434  * Gromacs nonbonded kernel:   nb_kernel_ElecCSTab_VdwBham_GeomW4P1_F_c
 435  * Electrostatics interaction: CubicSplineTable
 436  * VdW interaction:            Buckingham
 437  * Geometry:                   Water4-Particle
 438  * Calculate force/pot:        Force
 439  */
 440 void
 441 nb_kernel_ElecCSTab_VdwBham_GeomW4P1_F_c
 442                     (t_nblist                    * gmx_restrict       nlist,
 443                      rvec                        * gmx_restrict          xx,
 444                      rvec                        * gmx_restrict          ff,
 445                      t_forcerec                  * gmx_restrict          fr,
 446                      t_mdatoms                   * gmx_restrict     mdatoms,
 447                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 448                      t_nrnb                      * gmx_restrict        nrnb)
 449 {
 450     int              i_shift_offset,i_coord_offset,j_coord_offset;
 451     int              j_index_start,j_index_end;
 452     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 453     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 454     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 455     real             *shiftvec,*fshift,*x,*f;
 456     int              vdwioffset0;
 457     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 458     int              vdwioffset1;
 459     real             ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
 460     int              vdwioffset2;
 461     real             ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
 462     int              vdwioffset3;
 463     real             ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
 464     int              vdwjidx0;
 465     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 466     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 467     real             dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10;
 468     real             dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20;
 469     real             dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30,cexp1_30,cexp2_30;
 470     real             velec,felec,velecsum,facel,crf,krf,krf2;
 471     real             *charge;
 472     int              nvdwtype;
 473     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 474     int              *vdwtype;
 475     real             *vdwparam;
 476     int              vfitab;
 477     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
 478     real             *vftab;
 479
 480     x                = xx[0];
 481     f                = ff[0];
 482
 483     nri              = nlist->nri;
 484     iinr             = nlist->iinr;
 485     jindex           = nlist->jindex;
 486     jjnr             = nlist->jjnr;
 487     shiftidx         = nlist->shift;
 488     gid              = nlist->gid;
 489     shiftvec         = fr->shift_vec[0];
 490     fshift           = fr->fshift[0];
 491     facel            = fr->epsfac;
 492     charge           = mdatoms->chargeA;
 493     nvdwtype         = fr->ntype;
 494     vdwparam         = fr->nbfp;
 495     vdwtype          = mdatoms->typeA;
 496
 497     vftab            = kernel_data->table_elec->data;
 498     vftabscale       = kernel_data->table_elec->scale;
 499
 500     /* Setup water-specific parameters */
 501     inr              = nlist->iinr[0];
 502     iq1              = facel*charge[inr+1];
 503     iq2              = facel*charge[inr+2];
 504     iq3              = facel*charge[inr+3];
 505     vdwioffset0      = 3*nvdwtype*vdwtype[inr+0];
 506
 507     outeriter        = 0;
 508     inneriter        = 0;
 509
 510     /* Start outer loop over neighborlists */
 511     for(iidx=0; iidx<nri; iidx++)
 512     {
 513         /* Load shift vector for this list */
 514         i_shift_offset   = DIM*shiftidx[iidx];
 515         shX              = shiftvec[i_shift_offset+XX];
 516         shY              = shiftvec[i_shift_offset+YY];
 517         shZ              = shiftvec[i_shift_offset+ZZ];
 518
 519         /* Load limits for loop over neighbors */
 520         j_index_start    = jindex[iidx];
 521         j_index_end      = jindex[iidx+1];
 522
 523         /* Get outer coordinate index */
 524         inr              = iinr[iidx];
 525         i_coord_offset   = DIM*inr;
 526
 527         /* Load i particle coords and add shift vector */
 528         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 529         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 530         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 531         ix1              = shX + x[i_coord_offset+DIM*1+XX];
 532         iy1              = shY + x[i_coord_offset+DIM*1+YY];
 533         iz1              = shZ + x[i_coord_offset+DIM*1+ZZ];
 534         ix2              = shX + x[i_coord_offset+DIM*2+XX];
 535         iy2              = shY + x[i_coord_offset+DIM*2+YY];
 536         iz2              = shZ + x[i_coord_offset+DIM*2+ZZ];
 537         ix3              = shX + x[i_coord_offset+DIM*3+XX];
 538         iy3              = shY + x[i_coord_offset+DIM*3+YY];
 539         iz3              = shZ + x[i_coord_offset+DIM*3+ZZ];
 540
 541         fix0             = 0.0;
 542         fiy0             = 0.0;
 543         fiz0             = 0.0;
 544         fix1             = 0.0;
 545         fiy1             = 0.0;
 546         fiz1             = 0.0;
 547         fix2             = 0.0;
 548         fiy2             = 0.0;
 549         fiz2             = 0.0;
 550         fix3             = 0.0;
 551         fiy3             = 0.0;
 552         fiz3             = 0.0;
 553
 554         /* Start inner kernel loop */
 555         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 556         {
 557             /* Get j neighbor index, and coordinate index */
 558             jnr              = jjnr[jidx];
 559             j_coord_offset   = DIM*jnr;
 560
 561             /* load j atom coordinates */
 562             jx0              = x[j_coord_offset+DIM*0+XX];
 563             jy0              = x[j_coord_offset+DIM*0+YY];
 564             jz0              = x[j_coord_offset+DIM*0+ZZ];
 565
 566             /* Calculate displacement vector */
 567             dx00             = ix0 - jx0;
 568             dy00             = iy0 - jy0;
 569             dz00             = iz0 - jz0;
 570             dx10             = ix1 - jx0;
 571             dy10             = iy1 - jy0;
 572             dz10             = iz1 - jz0;
 573             dx20             = ix2 - jx0;
 574             dy20             = iy2 - jy0;
 575             dz20             = iz2 - jz0;
 576             dx30             = ix3 - jx0;
 577             dy30             = iy3 - jy0;
 578             dz30             = iz3 - jz0;
 579
 580             /* Calculate squared distance and things based on it */
 581             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 582             rsq10            = dx10*dx10+dy10*dy10+dz10*dz10;
 583             rsq20            = dx20*dx20+dy20*dy20+dz20*dz20;
 584             rsq30            = dx30*dx30+dy30*dy30+dz30*dz30;
 585
 586             rinv00           = gmx_invsqrt(rsq00);
 587             rinv10           = gmx_invsqrt(rsq10);
 588             rinv20           = gmx_invsqrt(rsq20);
 589             rinv30           = gmx_invsqrt(rsq30);
 590
 591             rinvsq00         = rinv00*rinv00;
 592
 593             /* Load parameters for j particles */
 594             jq0              = charge[jnr+0];
 595             vdwjidx0         = 3*vdwtype[jnr+0];
 596
 597             /**************************
 598              * CALCULATE INTERACTIONS *
 599              **************************/
 600
 601             r00              = rsq00*rinv00;
 602
 603             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 604             cexp1_00         = vdwparam[vdwioffset0+vdwjidx0+1];
 605             cexp2_00         = vdwparam[vdwioffset0+vdwjidx0+2];
 606
 607             /* BUCKINGHAM DISPERSION/REPULSION */
 608             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 609             vvdw6            = c6_00*rinvsix;
 610             br               = cexp2_00*r00;
 611             vvdwexp          = cexp1_00*exp(-br);
 612             fvdw             = (br*vvdwexp-vvdw6)*rinvsq00;
 613
 614             fscal            = fvdw;
 615
 616             /* Calculate temporary vectorial force */
 617             tx               = fscal*dx00;
 618             ty               = fscal*dy00;
 619             tz               = fscal*dz00;
 620
 621             /* Update vectorial force */
 622             fix0            += tx;
 623             fiy0            += ty;
 624             fiz0            += tz;
 625             f[j_coord_offset+DIM*0+XX] -= tx;
 626             f[j_coord_offset+DIM*0+YY] -= ty;
 627             f[j_coord_offset+DIM*0+ZZ] -= tz;
 628
 629             /**************************
 630              * CALCULATE INTERACTIONS *
 631              **************************/
 632
 633             r10              = rsq10*rinv10;
 634
 635             qq10             = iq1*jq0;
 636
 637             /* Calculate table index by multiplying r with table scale and truncate to integer */
 638             rt               = r10*vftabscale;
 639             vfitab           = rt;
 640             vfeps            = rt-vfitab;
 641             vfitab           = 1*4*vfitab;
 642
 643             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 644             F                = vftab[vfitab+1];
 645             Geps             = vfeps*vftab[vfitab+2];
 646             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 647             Fp               = F+Geps+Heps2;
 648             FF               = Fp+Geps+2.0*Heps2;
 649             felec            = -qq10*FF*vftabscale*rinv10;
 650
 651             fscal            = felec;
 652
 653             /* Calculate temporary vectorial force */
 654             tx               = fscal*dx10;
 655             ty               = fscal*dy10;
 656             tz               = fscal*dz10;
 657
 658             /* Update vectorial force */
 659             fix1            += tx;
 660             fiy1            += ty;
 661             fiz1            += tz;
 662             f[j_coord_offset+DIM*0+XX] -= tx;
 663             f[j_coord_offset+DIM*0+YY] -= ty;
 664             f[j_coord_offset+DIM*0+ZZ] -= tz;
 665
 666             /**************************
 667              * CALCULATE INTERACTIONS *
 668              **************************/
 669
 670             r20              = rsq20*rinv20;
 671
 672             qq20             = iq2*jq0;
 673
 674             /* Calculate table index by multiplying r with table scale and truncate to integer */
 675             rt               = r20*vftabscale;
 676             vfitab           = rt;
 677             vfeps            = rt-vfitab;
 678             vfitab           = 1*4*vfitab;
 679
 680             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 681             F                = vftab[vfitab+1];
 682             Geps             = vfeps*vftab[vfitab+2];
 683             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 684             Fp               = F+Geps+Heps2;
 685             FF               = Fp+Geps+2.0*Heps2;
 686             felec            = -qq20*FF*vftabscale*rinv20;
 687
 688             fscal            = felec;
 689
 690             /* Calculate temporary vectorial force */
 691             tx               = fscal*dx20;
 692             ty               = fscal*dy20;
 693             tz               = fscal*dz20;
 694
 695             /* Update vectorial force */
 696             fix2            += tx;
 697             fiy2            += ty;
 698             fiz2            += tz;
 699             f[j_coord_offset+DIM*0+XX] -= tx;
 700             f[j_coord_offset+DIM*0+YY] -= ty;
 701             f[j_coord_offset+DIM*0+ZZ] -= tz;
 702
 703             /**************************
 704              * CALCULATE INTERACTIONS *
 705              **************************/
 706
 707             r30              = rsq30*rinv30;
 708
 709             qq30             = iq3*jq0;
 710
 711             /* Calculate table index by multiplying r with table scale and truncate to integer */
 712             rt               = r30*vftabscale;
 713             vfitab           = rt;
 714             vfeps            = rt-vfitab;
 715             vfitab           = 1*4*vfitab;
 716
 717             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 718             F                = vftab[vfitab+1];
 719             Geps             = vfeps*vftab[vfitab+2];
 720             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 721             Fp               = F+Geps+Heps2;
 722             FF               = Fp+Geps+2.0*Heps2;
 723             felec            = -qq30*FF*vftabscale*rinv30;
 724
 725             fscal            = felec;
 726
 727             /* Calculate temporary vectorial force */
 728             tx               = fscal*dx30;
 729             ty               = fscal*dy30;
 730             tz               = fscal*dz30;
 731
 732             /* Update vectorial force */
 733             fix3            += tx;
 734             fiy3            += ty;
 735             fiz3            += tz;
 736             f[j_coord_offset+DIM*0+XX] -= tx;
 737             f[j_coord_offset+DIM*0+YY] -= ty;
 738             f[j_coord_offset+DIM*0+ZZ] -= tz;
 739
 740             /* Inner loop uses 172 flops */
 741         }
 742         /* End of innermost loop */
 743
 744         tx = ty = tz = 0;
 745         f[i_coord_offset+DIM*0+XX] += fix0;
 746         f[i_coord_offset+DIM*0+YY] += fiy0;
 747         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 748         tx                         += fix0;
 749         ty                         += fiy0;
 750         tz                         += fiz0;
 751         f[i_coord_offset+DIM*1+XX] += fix1;
 752         f[i_coord_offset+DIM*1+YY] += fiy1;
 753         f[i_coord_offset+DIM*1+ZZ] += fiz1;
 754         tx                         += fix1;
 755         ty                         += fiy1;
 756         tz                         += fiz1;
 757         f[i_coord_offset+DIM*2+XX] += fix2;
 758         f[i_coord_offset+DIM*2+YY] += fiy2;
 759         f[i_coord_offset+DIM*2+ZZ] += fiz2;
 760         tx                         += fix2;
 761         ty                         += fiy2;
 762         tz                         += fiz2;
 763         f[i_coord_offset+DIM*3+XX] += fix3;
 764         f[i_coord_offset+DIM*3+YY] += fiy3;
 765         f[i_coord_offset+DIM*3+ZZ] += fiz3;
 766         tx                         += fix3;
 767         ty                         += fiy3;
 768         tz                         += fiz3;
 769         fshift[i_shift_offset+XX]  += tx;
 770         fshift[i_shift_offset+YY]  += ty;
 771         fshift[i_shift_offset+ZZ]  += tz;
 772
 773         /* Increment number of inner iterations */
 774         inneriter                  += j_index_end - j_index_start;
 775
 776         /* Outer loop uses 39 flops */
 777     }
 778
 779     /* Increment number of outer iterations */
 780     outeriter        += nri;
 781
 782     /* Update outer/inner flops */
 783
 784     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_F,outeriter*39 + inneriter*172);
 785 }