src/gromacs/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecRFCut_VdwCSTab_GeomW4P1_c.c

   1 /*
   2  * This file is part of the GROMACS molecular simulation package.
   3  *
   4  * Copyright (c) 2012,2013,2014, 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 #include "config.h"
  39
  40 #include <math.h>
  41
  42 #include "../nb_kernel.h"
  43 #include "gromacs/legacyheaders/types/simple.h"
  44 #include "gromacs/math/vec.h"
  45 #include "gromacs/legacyheaders/nrnb.h"
  46
  47 /*
  48  * Gromacs nonbonded kernel:   nb_kernel_ElecRFCut_VdwCSTab_GeomW4P1_VF_c
  49  * Electrostatics interaction: ReactionField
  50  * VdW interaction:            CubicSplineTable
  51  * Geometry:                   Water4-Particle
  52  * Calculate force/pot:        PotentialAndForce
  53  */
  54 void
  55 nb_kernel_ElecRFCut_VdwCSTab_GeomW4P1_VF_c
  56                     (t_nblist                    * gmx_restrict       nlist,
  57                      rvec                        * gmx_restrict          xx,
  58                      rvec                        * gmx_restrict          ff,
  59                      t_forcerec                  * gmx_restrict          fr,
  60                      t_mdatoms                   * gmx_restrict     mdatoms,
  61                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
  62                      t_nrnb                      * gmx_restrict        nrnb)
  63 {
  64     int              i_shift_offset,i_coord_offset,j_coord_offset;
  65     int              j_index_start,j_index_end;
  66     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
  67     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
  68     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  69     real             *shiftvec,*fshift,*x,*f;
  70     int              vdwioffset0;
  71     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  72     int              vdwioffset1;
  73     real             ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
  74     int              vdwioffset2;
  75     real             ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
  76     int              vdwioffset3;
  77     real             ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
  78     int              vdwjidx0;
  79     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  80     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
  81     real             dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10;
  82     real             dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20;
  83     real             dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30,cexp1_30,cexp2_30;
  84     real             velec,felec,velecsum,facel,crf,krf,krf2;
  85     real             *charge;
  86     int              nvdwtype;
  87     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
  88     int              *vdwtype;
  89     real             *vdwparam;
  90     int              vfitab;
  91     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
  92     real             *vftab;
  93
  94     x                = xx[0];
  95     f                = ff[0];
  96
  97     nri              = nlist->nri;
  98     iinr             = nlist->iinr;
  99     jindex           = nlist->jindex;
 100     jjnr             = nlist->jjnr;
 101     shiftidx         = nlist->shift;
 102     gid              = nlist->gid;
 103     shiftvec         = fr->shift_vec[0];
 104     fshift           = fr->fshift[0];
 105     facel            = fr->epsfac;
 106     charge           = mdatoms->chargeA;
 107     krf              = fr->ic->k_rf;
 108     krf2             = krf*2.0;
 109     crf              = fr->ic->c_rf;
 110     nvdwtype         = fr->ntype;
 111     vdwparam         = fr->nbfp;
 112     vdwtype          = mdatoms->typeA;
 113
 114     vftab            = kernel_data->table_vdw->data;
 115     vftabscale       = kernel_data->table_vdw->scale;
 116
 117     /* Setup water-specific parameters */
 118     inr              = nlist->iinr[0];
 119     iq1              = facel*charge[inr+1];
 120     iq2              = facel*charge[inr+2];
 121     iq3              = facel*charge[inr+3];
 122     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 123
 124     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 125     rcutoff          = fr->rcoulomb;
 126     rcutoff2         = rcutoff*rcutoff;
 127
 128     outeriter        = 0;
 129     inneriter        = 0;
 130
 131     /* Start outer loop over neighborlists */
 132     for(iidx=0; iidx<nri; iidx++)
 133     {
 134         /* Load shift vector for this list */
 135         i_shift_offset   = DIM*shiftidx[iidx];
 136         shX              = shiftvec[i_shift_offset+XX];
 137         shY              = shiftvec[i_shift_offset+YY];
 138         shZ              = shiftvec[i_shift_offset+ZZ];
 139
 140         /* Load limits for loop over neighbors */
 141         j_index_start    = jindex[iidx];
 142         j_index_end      = jindex[iidx+1];
 143
 144         /* Get outer coordinate index */
 145         inr              = iinr[iidx];
 146         i_coord_offset   = DIM*inr;
 147
 148         /* Load i particle coords and add shift vector */
 149         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 150         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 151         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 152         ix1              = shX + x[i_coord_offset+DIM*1+XX];
 153         iy1              = shY + x[i_coord_offset+DIM*1+YY];
 154         iz1              = shZ + x[i_coord_offset+DIM*1+ZZ];
 155         ix2              = shX + x[i_coord_offset+DIM*2+XX];
 156         iy2              = shY + x[i_coord_offset+DIM*2+YY];
 157         iz2              = shZ + x[i_coord_offset+DIM*2+ZZ];
 158         ix3              = shX + x[i_coord_offset+DIM*3+XX];
 159         iy3              = shY + x[i_coord_offset+DIM*3+YY];
 160         iz3              = shZ + x[i_coord_offset+DIM*3+ZZ];
 161
 162         fix0             = 0.0;
 163         fiy0             = 0.0;
 164         fiz0             = 0.0;
 165         fix1             = 0.0;
 166         fiy1             = 0.0;
 167         fiz1             = 0.0;
 168         fix2             = 0.0;
 169         fiy2             = 0.0;
 170         fiz2             = 0.0;
 171         fix3             = 0.0;
 172         fiy3             = 0.0;
 173         fiz3             = 0.0;
 174
 175         /* Reset potential sums */
 176         velecsum         = 0.0;
 177         vvdwsum          = 0.0;
 178
 179         /* Start inner kernel loop */
 180         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 181         {
 182             /* Get j neighbor index, and coordinate index */
 183             jnr              = jjnr[jidx];
 184             j_coord_offset   = DIM*jnr;
 185
 186             /* load j atom coordinates */
 187             jx0              = x[j_coord_offset+DIM*0+XX];
 188             jy0              = x[j_coord_offset+DIM*0+YY];
 189             jz0              = x[j_coord_offset+DIM*0+ZZ];
 190
 191             /* Calculate displacement vector */
 192             dx00             = ix0 - jx0;
 193             dy00             = iy0 - jy0;
 194             dz00             = iz0 - jz0;
 195             dx10             = ix1 - jx0;
 196             dy10             = iy1 - jy0;
 197             dz10             = iz1 - jz0;
 198             dx20             = ix2 - jx0;
 199             dy20             = iy2 - jy0;
 200             dz20             = iz2 - jz0;
 201             dx30             = ix3 - jx0;
 202             dy30             = iy3 - jy0;
 203             dz30             = iz3 - jz0;
 204
 205             /* Calculate squared distance and things based on it */
 206             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 207             rsq10            = dx10*dx10+dy10*dy10+dz10*dz10;
 208             rsq20            = dx20*dx20+dy20*dy20+dz20*dz20;
 209             rsq30            = dx30*dx30+dy30*dy30+dz30*dz30;
 210
 211             rinv00           = gmx_invsqrt(rsq00);
 212             rinv10           = gmx_invsqrt(rsq10);
 213             rinv20           = gmx_invsqrt(rsq20);
 214             rinv30           = gmx_invsqrt(rsq30);
 215
 216             rinvsq10         = rinv10*rinv10;
 217             rinvsq20         = rinv20*rinv20;
 218             rinvsq30         = rinv30*rinv30;
 219
 220             /* Load parameters for j particles */
 221             jq0              = charge[jnr+0];
 222             vdwjidx0         = 2*vdwtype[jnr+0];
 223
 224             /**************************
 225              * CALCULATE INTERACTIONS *
 226              **************************/
 227
 228             r00              = rsq00*rinv00;
 229
 230             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 231             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 232
 233             /* Calculate table index by multiplying r with table scale and truncate to integer */
 234             rt               = r00*vftabscale;
 235             vfitab           = rt;
 236             vfeps            = rt-vfitab;
 237             vfitab           = 2*4*vfitab;
 238
 239             /* CUBIC SPLINE TABLE DISPERSION */
 240             vfitab          += 0;
 241             Y                = vftab[vfitab];
 242             F                = vftab[vfitab+1];
 243             Geps             = vfeps*vftab[vfitab+2];
 244             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 245             Fp               = F+Geps+Heps2;
 246             VV               = Y+vfeps*Fp;
 247             vvdw6            = c6_00*VV;
 248             FF               = Fp+Geps+2.0*Heps2;
 249             fvdw6            = c6_00*FF;
 250
 251             /* CUBIC SPLINE TABLE REPULSION */
 252             Y                = vftab[vfitab+4];
 253             F                = vftab[vfitab+5];
 254             Geps             = vfeps*vftab[vfitab+6];
 255             Heps2            = vfeps*vfeps*vftab[vfitab+7];
 256             Fp               = F+Geps+Heps2;
 257             VV               = Y+vfeps*Fp;
 258             vvdw12           = c12_00*VV;
 259             FF               = Fp+Geps+2.0*Heps2;
 260             fvdw12           = c12_00*FF;
 261             vvdw             = vvdw12+vvdw6;
 262             fvdw             = -(fvdw6+fvdw12)*vftabscale*rinv00;
 263
 264             /* Update potential sums from outer loop */
 265             vvdwsum         += vvdw;
 266
 267             fscal            = fvdw;
 268
 269             /* Calculate temporary vectorial force */
 270             tx               = fscal*dx00;
 271             ty               = fscal*dy00;
 272             tz               = fscal*dz00;
 273
 274             /* Update vectorial force */
 275             fix0            += tx;
 276             fiy0            += ty;
 277             fiz0            += tz;
 278             f[j_coord_offset+DIM*0+XX] -= tx;
 279             f[j_coord_offset+DIM*0+YY] -= ty;
 280             f[j_coord_offset+DIM*0+ZZ] -= tz;
 281
 282             /**************************
 283              * CALCULATE INTERACTIONS *
 284              **************************/
 285
 286             if (rsq10<rcutoff2)
 287             {
 288
 289             qq10             = iq1*jq0;
 290
 291             /* REACTION-FIELD ELECTROSTATICS */
 292             velec            = qq10*(rinv10+krf*rsq10-crf);
 293             felec            = qq10*(rinv10*rinvsq10-krf2);
 294
 295             /* Update potential sums from outer loop */
 296             velecsum        += velec;
 297
 298             fscal            = felec;
 299
 300             /* Calculate temporary vectorial force */
 301             tx               = fscal*dx10;
 302             ty               = fscal*dy10;
 303             tz               = fscal*dz10;
 304
 305             /* Update vectorial force */
 306             fix1            += tx;
 307             fiy1            += ty;
 308             fiz1            += tz;
 309             f[j_coord_offset+DIM*0+XX] -= tx;
 310             f[j_coord_offset+DIM*0+YY] -= ty;
 311             f[j_coord_offset+DIM*0+ZZ] -= tz;
 312
 313             }
 314
 315             /**************************
 316              * CALCULATE INTERACTIONS *
 317              **************************/
 318
 319             if (rsq20<rcutoff2)
 320             {
 321
 322             qq20             = iq2*jq0;
 323
 324             /* REACTION-FIELD ELECTROSTATICS */
 325             velec            = qq20*(rinv20+krf*rsq20-crf);
 326             felec            = qq20*(rinv20*rinvsq20-krf2);
 327
 328             /* Update potential sums from outer loop */
 329             velecsum        += velec;
 330
 331             fscal            = felec;
 332
 333             /* Calculate temporary vectorial force */
 334             tx               = fscal*dx20;
 335             ty               = fscal*dy20;
 336             tz               = fscal*dz20;
 337
 338             /* Update vectorial force */
 339             fix2            += tx;
 340             fiy2            += ty;
 341             fiz2            += tz;
 342             f[j_coord_offset+DIM*0+XX] -= tx;
 343             f[j_coord_offset+DIM*0+YY] -= ty;
 344             f[j_coord_offset+DIM*0+ZZ] -= tz;
 345
 346             }
 347
 348             /**************************
 349              * CALCULATE INTERACTIONS *
 350              **************************/
 351
 352             if (rsq30<rcutoff2)
 353             {
 354
 355             qq30             = iq3*jq0;
 356
 357             /* REACTION-FIELD ELECTROSTATICS */
 358             velec            = qq30*(rinv30+krf*rsq30-crf);
 359             felec            = qq30*(rinv30*rinvsq30-krf2);
 360
 361             /* Update potential sums from outer loop */
 362             velecsum        += velec;
 363
 364             fscal            = felec;
 365
 366             /* Calculate temporary vectorial force */
 367             tx               = fscal*dx30;
 368             ty               = fscal*dy30;
 369             tz               = fscal*dz30;
 370
 371             /* Update vectorial force */
 372             fix3            += tx;
 373             fiy3            += ty;
 374             fiz3            += tz;
 375             f[j_coord_offset+DIM*0+XX] -= tx;
 376             f[j_coord_offset+DIM*0+YY] -= ty;
 377             f[j_coord_offset+DIM*0+ZZ] -= tz;
 378
 379             }
 380
 381             /* Inner loop uses 151 flops */
 382         }
 383         /* End of innermost loop */
 384
 385         tx = ty = tz = 0;
 386         f[i_coord_offset+DIM*0+XX] += fix0;
 387         f[i_coord_offset+DIM*0+YY] += fiy0;
 388         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 389         tx                         += fix0;
 390         ty                         += fiy0;
 391         tz                         += fiz0;
 392         f[i_coord_offset+DIM*1+XX] += fix1;
 393         f[i_coord_offset+DIM*1+YY] += fiy1;
 394         f[i_coord_offset+DIM*1+ZZ] += fiz1;
 395         tx                         += fix1;
 396         ty                         += fiy1;
 397         tz                         += fiz1;
 398         f[i_coord_offset+DIM*2+XX] += fix2;
 399         f[i_coord_offset+DIM*2+YY] += fiy2;
 400         f[i_coord_offset+DIM*2+ZZ] += fiz2;
 401         tx                         += fix2;
 402         ty                         += fiy2;
 403         tz                         += fiz2;
 404         f[i_coord_offset+DIM*3+XX] += fix3;
 405         f[i_coord_offset+DIM*3+YY] += fiy3;
 406         f[i_coord_offset+DIM*3+ZZ] += fiz3;
 407         tx                         += fix3;
 408         ty                         += fiy3;
 409         tz                         += fiz3;
 410         fshift[i_shift_offset+XX]  += tx;
 411         fshift[i_shift_offset+YY]  += ty;
 412         fshift[i_shift_offset+ZZ]  += tz;
 413
 414         ggid                        = gid[iidx];
 415         /* Update potential energies */
 416         kernel_data->energygrp_elec[ggid] += velecsum;
 417         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 418
 419         /* Increment number of inner iterations */
 420         inneriter                  += j_index_end - j_index_start;
 421
 422         /* Outer loop uses 41 flops */
 423     }
 424
 425     /* Increment number of outer iterations */
 426     outeriter        += nri;
 427
 428     /* Update outer/inner flops */
 429
 430     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_VF,outeriter*41 + inneriter*151);
 431 }
 432 /*
 433  * Gromacs nonbonded kernel:   nb_kernel_ElecRFCut_VdwCSTab_GeomW4P1_F_c
 434  * Electrostatics interaction: ReactionField
 435  * VdW interaction:            CubicSplineTable
 436  * Geometry:                   Water4-Particle
 437  * Calculate force/pot:        Force
 438  */
 439 void
 440 nb_kernel_ElecRFCut_VdwCSTab_GeomW4P1_F_c
 441                     (t_nblist                    * gmx_restrict       nlist,
 442                      rvec                        * gmx_restrict          xx,
 443                      rvec                        * gmx_restrict          ff,
 444                      t_forcerec                  * gmx_restrict          fr,
 445                      t_mdatoms                   * gmx_restrict     mdatoms,
 446                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 447                      t_nrnb                      * gmx_restrict        nrnb)
 448 {
 449     int              i_shift_offset,i_coord_offset,j_coord_offset;
 450     int              j_index_start,j_index_end;
 451     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 452     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 453     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 454     real             *shiftvec,*fshift,*x,*f;
 455     int              vdwioffset0;
 456     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 457     int              vdwioffset1;
 458     real             ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
 459     int              vdwioffset2;
 460     real             ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
 461     int              vdwioffset3;
 462     real             ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
 463     int              vdwjidx0;
 464     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 465     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 466     real             dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10;
 467     real             dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20;
 468     real             dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30,cexp1_30,cexp2_30;
 469     real             velec,felec,velecsum,facel,crf,krf,krf2;
 470     real             *charge;
 471     int              nvdwtype;
 472     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 473     int              *vdwtype;
 474     real             *vdwparam;
 475     int              vfitab;
 476     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
 477     real             *vftab;
 478
 479     x                = xx[0];
 480     f                = ff[0];
 481
 482     nri              = nlist->nri;
 483     iinr             = nlist->iinr;
 484     jindex           = nlist->jindex;
 485     jjnr             = nlist->jjnr;
 486     shiftidx         = nlist->shift;
 487     gid              = nlist->gid;
 488     shiftvec         = fr->shift_vec[0];
 489     fshift           = fr->fshift[0];
 490     facel            = fr->epsfac;
 491     charge           = mdatoms->chargeA;
 492     krf              = fr->ic->k_rf;
 493     krf2             = krf*2.0;
 494     crf              = fr->ic->c_rf;
 495     nvdwtype         = fr->ntype;
 496     vdwparam         = fr->nbfp;
 497     vdwtype          = mdatoms->typeA;
 498
 499     vftab            = kernel_data->table_vdw->data;
 500     vftabscale       = kernel_data->table_vdw->scale;
 501
 502     /* Setup water-specific parameters */
 503     inr              = nlist->iinr[0];
 504     iq1              = facel*charge[inr+1];
 505     iq2              = facel*charge[inr+2];
 506     iq3              = facel*charge[inr+3];
 507     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 508
 509     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 510     rcutoff          = fr->rcoulomb;
 511     rcutoff2         = rcutoff*rcutoff;
 512
 513     outeriter        = 0;
 514     inneriter        = 0;
 515
 516     /* Start outer loop over neighborlists */
 517     for(iidx=0; iidx<nri; iidx++)
 518     {
 519         /* Load shift vector for this list */
 520         i_shift_offset   = DIM*shiftidx[iidx];
 521         shX              = shiftvec[i_shift_offset+XX];
 522         shY              = shiftvec[i_shift_offset+YY];
 523         shZ              = shiftvec[i_shift_offset+ZZ];
 524
 525         /* Load limits for loop over neighbors */
 526         j_index_start    = jindex[iidx];
 527         j_index_end      = jindex[iidx+1];
 528
 529         /* Get outer coordinate index */
 530         inr              = iinr[iidx];
 531         i_coord_offset   = DIM*inr;
 532
 533         /* Load i particle coords and add shift vector */
 534         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 535         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 536         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 537         ix1              = shX + x[i_coord_offset+DIM*1+XX];
 538         iy1              = shY + x[i_coord_offset+DIM*1+YY];
 539         iz1              = shZ + x[i_coord_offset+DIM*1+ZZ];
 540         ix2              = shX + x[i_coord_offset+DIM*2+XX];
 541         iy2              = shY + x[i_coord_offset+DIM*2+YY];
 542         iz2              = shZ + x[i_coord_offset+DIM*2+ZZ];
 543         ix3              = shX + x[i_coord_offset+DIM*3+XX];
 544         iy3              = shY + x[i_coord_offset+DIM*3+YY];
 545         iz3              = shZ + x[i_coord_offset+DIM*3+ZZ];
 546
 547         fix0             = 0.0;
 548         fiy0             = 0.0;
 549         fiz0             = 0.0;
 550         fix1             = 0.0;
 551         fiy1             = 0.0;
 552         fiz1             = 0.0;
 553         fix2             = 0.0;
 554         fiy2             = 0.0;
 555         fiz2             = 0.0;
 556         fix3             = 0.0;
 557         fiy3             = 0.0;
 558         fiz3             = 0.0;
 559
 560         /* Start inner kernel loop */
 561         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 562         {
 563             /* Get j neighbor index, and coordinate index */
 564             jnr              = jjnr[jidx];
 565             j_coord_offset   = DIM*jnr;
 566
 567             /* load j atom coordinates */
 568             jx0              = x[j_coord_offset+DIM*0+XX];
 569             jy0              = x[j_coord_offset+DIM*0+YY];
 570             jz0              = x[j_coord_offset+DIM*0+ZZ];
 571
 572             /* Calculate displacement vector */
 573             dx00             = ix0 - jx0;
 574             dy00             = iy0 - jy0;
 575             dz00             = iz0 - jz0;
 576             dx10             = ix1 - jx0;
 577             dy10             = iy1 - jy0;
 578             dz10             = iz1 - jz0;
 579             dx20             = ix2 - jx0;
 580             dy20             = iy2 - jy0;
 581             dz20             = iz2 - jz0;
 582             dx30             = ix3 - jx0;
 583             dy30             = iy3 - jy0;
 584             dz30             = iz3 - jz0;
 585
 586             /* Calculate squared distance and things based on it */
 587             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 588             rsq10            = dx10*dx10+dy10*dy10+dz10*dz10;
 589             rsq20            = dx20*dx20+dy20*dy20+dz20*dz20;
 590             rsq30            = dx30*dx30+dy30*dy30+dz30*dz30;
 591
 592             rinv00           = gmx_invsqrt(rsq00);
 593             rinv10           = gmx_invsqrt(rsq10);
 594             rinv20           = gmx_invsqrt(rsq20);
 595             rinv30           = gmx_invsqrt(rsq30);
 596
 597             rinvsq10         = rinv10*rinv10;
 598             rinvsq20         = rinv20*rinv20;
 599             rinvsq30         = rinv30*rinv30;
 600
 601             /* Load parameters for j particles */
 602             jq0              = charge[jnr+0];
 603             vdwjidx0         = 2*vdwtype[jnr+0];
 604
 605             /**************************
 606              * CALCULATE INTERACTIONS *
 607              **************************/
 608
 609             r00              = rsq00*rinv00;
 610
 611             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 612             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 613
 614             /* Calculate table index by multiplying r with table scale and truncate to integer */
 615             rt               = r00*vftabscale;
 616             vfitab           = rt;
 617             vfeps            = rt-vfitab;
 618             vfitab           = 2*4*vfitab;
 619
 620             /* CUBIC SPLINE TABLE DISPERSION */
 621             vfitab          += 0;
 622             F                = vftab[vfitab+1];
 623             Geps             = vfeps*vftab[vfitab+2];
 624             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 625             Fp               = F+Geps+Heps2;
 626             FF               = Fp+Geps+2.0*Heps2;
 627             fvdw6            = c6_00*FF;
 628
 629             /* CUBIC SPLINE TABLE REPULSION */
 630             F                = vftab[vfitab+5];
 631             Geps             = vfeps*vftab[vfitab+6];
 632             Heps2            = vfeps*vfeps*vftab[vfitab+7];
 633             Fp               = F+Geps+Heps2;
 634             FF               = Fp+Geps+2.0*Heps2;
 635             fvdw12           = c12_00*FF;
 636             fvdw             = -(fvdw6+fvdw12)*vftabscale*rinv00;
 637
 638             fscal            = fvdw;
 639
 640             /* Calculate temporary vectorial force */
 641             tx               = fscal*dx00;
 642             ty               = fscal*dy00;
 643             tz               = fscal*dz00;
 644
 645             /* Update vectorial force */
 646             fix0            += tx;
 647             fiy0            += ty;
 648             fiz0            += tz;
 649             f[j_coord_offset+DIM*0+XX] -= tx;
 650             f[j_coord_offset+DIM*0+YY] -= ty;
 651             f[j_coord_offset+DIM*0+ZZ] -= tz;
 652
 653             /**************************
 654              * CALCULATE INTERACTIONS *
 655              **************************/
 656
 657             if (rsq10<rcutoff2)
 658             {
 659
 660             qq10             = iq1*jq0;
 661
 662             /* REACTION-FIELD ELECTROSTATICS */
 663             felec            = qq10*(rinv10*rinvsq10-krf2);
 664
 665             fscal            = felec;
 666
 667             /* Calculate temporary vectorial force */
 668             tx               = fscal*dx10;
 669             ty               = fscal*dy10;
 670             tz               = fscal*dz10;
 671
 672             /* Update vectorial force */
 673             fix1            += tx;
 674             fiy1            += ty;
 675             fiz1            += tz;
 676             f[j_coord_offset+DIM*0+XX] -= tx;
 677             f[j_coord_offset+DIM*0+YY] -= ty;
 678             f[j_coord_offset+DIM*0+ZZ] -= tz;
 679
 680             }
 681
 682             /**************************
 683              * CALCULATE INTERACTIONS *
 684              **************************/
 685
 686             if (rsq20<rcutoff2)
 687             {
 688
 689             qq20             = iq2*jq0;
 690
 691             /* REACTION-FIELD ELECTROSTATICS */
 692             felec            = qq20*(rinv20*rinvsq20-krf2);
 693
 694             fscal            = felec;
 695
 696             /* Calculate temporary vectorial force */
 697             tx               = fscal*dx20;
 698             ty               = fscal*dy20;
 699             tz               = fscal*dz20;
 700
 701             /* Update vectorial force */
 702             fix2            += tx;
 703             fiy2            += ty;
 704             fiz2            += tz;
 705             f[j_coord_offset+DIM*0+XX] -= tx;
 706             f[j_coord_offset+DIM*0+YY] -= ty;
 707             f[j_coord_offset+DIM*0+ZZ] -= tz;
 708
 709             }
 710
 711             /**************************
 712              * CALCULATE INTERACTIONS *
 713              **************************/
 714
 715             if (rsq30<rcutoff2)
 716             {
 717
 718             qq30             = iq3*jq0;
 719
 720             /* REACTION-FIELD ELECTROSTATICS */
 721             felec            = qq30*(rinv30*rinvsq30-krf2);
 722
 723             fscal            = felec;
 724
 725             /* Calculate temporary vectorial force */
 726             tx               = fscal*dx30;
 727             ty               = fscal*dy30;
 728             tz               = fscal*dz30;
 729
 730             /* Update vectorial force */
 731             fix3            += tx;
 732             fiy3            += ty;
 733             fiz3            += tz;
 734             f[j_coord_offset+DIM*0+XX] -= tx;
 735             f[j_coord_offset+DIM*0+YY] -= ty;
 736             f[j_coord_offset+DIM*0+ZZ] -= tz;
 737
 738             }
 739
 740             /* Inner loop uses 128 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*128);
 785 }