src/gromacs/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecEw_VdwLJEw_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 "types/simple.h"
  44 #include "gromacs/math/vec.h"
  45 #include "nrnb.h"
  46
  47 /*
  48  * Gromacs nonbonded kernel:   nb_kernel_ElecEw_VdwLJEw_GeomW4P1_VF_c
  49  * Electrostatics interaction: Ewald
  50  * VdW interaction:            LJEwald
  51  * Geometry:                   Water4-Particle
  52  * Calculate force/pot:        PotentialAndForce
  53  */
  54 void
  55 nb_kernel_ElecEw_VdwLJEw_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     real             c6grid_00;
  91     real             c6grid_10;
  92     real             c6grid_20;
  93     real             c6grid_30;
  94     real             ewclj,ewclj2,ewclj6,ewcljrsq,poly,exponent,sh_lj_ewald;
  95     real             *vdwgridparam;
  96     int              ewitab;
  97     real             ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
  98     real             *ewtab;
  99
 100     x                = xx[0];
 101     f                = ff[0];
 102
 103     nri              = nlist->nri;
 104     iinr             = nlist->iinr;
 105     jindex           = nlist->jindex;
 106     jjnr             = nlist->jjnr;
 107     shiftidx         = nlist->shift;
 108     gid              = nlist->gid;
 109     shiftvec         = fr->shift_vec[0];
 110     fshift           = fr->fshift[0];
 111     facel            = fr->epsfac;
 112     charge           = mdatoms->chargeA;
 113     nvdwtype         = fr->ntype;
 114     vdwparam         = fr->nbfp;
 115     vdwtype          = mdatoms->typeA;
 116     vdwgridparam     = fr->ljpme_c6grid;
 117     ewclj            = fr->ewaldcoeff_lj;
 118     sh_lj_ewald      = fr->ic->sh_lj_ewald;
 119     ewclj2           = ewclj*ewclj;
 120     ewclj6           = ewclj2*ewclj2*ewclj2;
 121
 122     sh_ewald         = fr->ic->sh_ewald;
 123     ewtab            = fr->ic->tabq_coul_FDV0;
 124     ewtabscale       = fr->ic->tabq_scale;
 125     ewtabhalfspace   = 0.5/ewtabscale;
 126
 127     /* Setup water-specific parameters */
 128     inr              = nlist->iinr[0];
 129     iq1              = facel*charge[inr+1];
 130     iq2              = facel*charge[inr+2];
 131     iq3              = facel*charge[inr+3];
 132     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 133
 134     outeriter        = 0;
 135     inneriter        = 0;
 136
 137     /* Start outer loop over neighborlists */
 138     for(iidx=0; iidx<nri; iidx++)
 139     {
 140         /* Load shift vector for this list */
 141         i_shift_offset   = DIM*shiftidx[iidx];
 142         shX              = shiftvec[i_shift_offset+XX];
 143         shY              = shiftvec[i_shift_offset+YY];
 144         shZ              = shiftvec[i_shift_offset+ZZ];
 145
 146         /* Load limits for loop over neighbors */
 147         j_index_start    = jindex[iidx];
 148         j_index_end      = jindex[iidx+1];
 149
 150         /* Get outer coordinate index */
 151         inr              = iinr[iidx];
 152         i_coord_offset   = DIM*inr;
 153
 154         /* Load i particle coords and add shift vector */
 155         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 156         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 157         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 158         ix1              = shX + x[i_coord_offset+DIM*1+XX];
 159         iy1              = shY + x[i_coord_offset+DIM*1+YY];
 160         iz1              = shZ + x[i_coord_offset+DIM*1+ZZ];
 161         ix2              = shX + x[i_coord_offset+DIM*2+XX];
 162         iy2              = shY + x[i_coord_offset+DIM*2+YY];
 163         iz2              = shZ + x[i_coord_offset+DIM*2+ZZ];
 164         ix3              = shX + x[i_coord_offset+DIM*3+XX];
 165         iy3              = shY + x[i_coord_offset+DIM*3+YY];
 166         iz3              = shZ + x[i_coord_offset+DIM*3+ZZ];
 167
 168         fix0             = 0.0;
 169         fiy0             = 0.0;
 170         fiz0             = 0.0;
 171         fix1             = 0.0;
 172         fiy1             = 0.0;
 173         fiz1             = 0.0;
 174         fix2             = 0.0;
 175         fiy2             = 0.0;
 176         fiz2             = 0.0;
 177         fix3             = 0.0;
 178         fiy3             = 0.0;
 179         fiz3             = 0.0;
 180
 181         /* Reset potential sums */
 182         velecsum         = 0.0;
 183         vvdwsum          = 0.0;
 184
 185         /* Start inner kernel loop */
 186         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 187         {
 188             /* Get j neighbor index, and coordinate index */
 189             jnr              = jjnr[jidx];
 190             j_coord_offset   = DIM*jnr;
 191
 192             /* load j atom coordinates */
 193             jx0              = x[j_coord_offset+DIM*0+XX];
 194             jy0              = x[j_coord_offset+DIM*0+YY];
 195             jz0              = x[j_coord_offset+DIM*0+ZZ];
 196
 197             /* Calculate displacement vector */
 198             dx00             = ix0 - jx0;
 199             dy00             = iy0 - jy0;
 200             dz00             = iz0 - jz0;
 201             dx10             = ix1 - jx0;
 202             dy10             = iy1 - jy0;
 203             dz10             = iz1 - jz0;
 204             dx20             = ix2 - jx0;
 205             dy20             = iy2 - jy0;
 206             dz20             = iz2 - jz0;
 207             dx30             = ix3 - jx0;
 208             dy30             = iy3 - jy0;
 209             dz30             = iz3 - jz0;
 210
 211             /* Calculate squared distance and things based on it */
 212             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 213             rsq10            = dx10*dx10+dy10*dy10+dz10*dz10;
 214             rsq20            = dx20*dx20+dy20*dy20+dz20*dz20;
 215             rsq30            = dx30*dx30+dy30*dy30+dz30*dz30;
 216
 217             rinv00           = gmx_invsqrt(rsq00);
 218             rinv10           = gmx_invsqrt(rsq10);
 219             rinv20           = gmx_invsqrt(rsq20);
 220             rinv30           = gmx_invsqrt(rsq30);
 221
 222             rinvsq00         = rinv00*rinv00;
 223             rinvsq10         = rinv10*rinv10;
 224             rinvsq20         = rinv20*rinv20;
 225             rinvsq30         = rinv30*rinv30;
 226
 227             /* Load parameters for j particles */
 228             jq0              = charge[jnr+0];
 229             vdwjidx0         = 2*vdwtype[jnr+0];
 230
 231             /**************************
 232              * CALCULATE INTERACTIONS *
 233              **************************/
 234
 235             r00              = rsq00*rinv00;
 236
 237             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 238             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 239             c6grid_00        = vdwgridparam[vdwioffset0+vdwjidx0];
 240
 241             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 242             ewcljrsq         = ewclj2*rsq00;
 243             exponent         = exp(-ewcljrsq);
 244             poly             = exponent*(1.0 + ewcljrsq + ewcljrsq*ewcljrsq*0.5);
 245             vvdw6            = (c6_00-c6grid_00*(1.0-poly))*rinvsix;
 246             vvdw12           = c12_00*rinvsix*rinvsix;
 247             vvdw             = vvdw12*(1.0/12.0) - vvdw6*(1.0/6.0);
 248             fvdw             = (vvdw12 - vvdw6 - c6grid_00*(1.0/6.0)*exponent*ewclj6)*rinvsq00;
 249
 250             /* Update potential sums from outer loop */
 251             vvdwsum         += vvdw;
 252
 253             fscal            = fvdw;
 254
 255             /* Calculate temporary vectorial force */
 256             tx               = fscal*dx00;
 257             ty               = fscal*dy00;
 258             tz               = fscal*dz00;
 259
 260             /* Update vectorial force */
 261             fix0            += tx;
 262             fiy0            += ty;
 263             fiz0            += tz;
 264             f[j_coord_offset+DIM*0+XX] -= tx;
 265             f[j_coord_offset+DIM*0+YY] -= ty;
 266             f[j_coord_offset+DIM*0+ZZ] -= tz;
 267
 268             /**************************
 269              * CALCULATE INTERACTIONS *
 270              **************************/
 271
 272             r10              = rsq10*rinv10;
 273
 274             qq10             = iq1*jq0;
 275
 276             /* EWALD ELECTROSTATICS */
 277
 278             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 279             ewrt             = r10*ewtabscale;
 280             ewitab           = ewrt;
 281             eweps            = ewrt-ewitab;
 282             ewitab           = 4*ewitab;
 283             felec            = ewtab[ewitab]+eweps*ewtab[ewitab+1];
 284             velec            = qq10*(rinv10-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
 285             felec            = qq10*rinv10*(rinvsq10-felec);
 286
 287             /* Update potential sums from outer loop */
 288             velecsum        += velec;
 289
 290             fscal            = felec;
 291
 292             /* Calculate temporary vectorial force */
 293             tx               = fscal*dx10;
 294             ty               = fscal*dy10;
 295             tz               = fscal*dz10;
 296
 297             /* Update vectorial force */
 298             fix1            += tx;
 299             fiy1            += ty;
 300             fiz1            += tz;
 301             f[j_coord_offset+DIM*0+XX] -= tx;
 302             f[j_coord_offset+DIM*0+YY] -= ty;
 303             f[j_coord_offset+DIM*0+ZZ] -= tz;
 304
 305             /**************************
 306              * CALCULATE INTERACTIONS *
 307              **************************/
 308
 309             r20              = rsq20*rinv20;
 310
 311             qq20             = iq2*jq0;
 312
 313             /* EWALD ELECTROSTATICS */
 314
 315             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 316             ewrt             = r20*ewtabscale;
 317             ewitab           = ewrt;
 318             eweps            = ewrt-ewitab;
 319             ewitab           = 4*ewitab;
 320             felec            = ewtab[ewitab]+eweps*ewtab[ewitab+1];
 321             velec            = qq20*(rinv20-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
 322             felec            = qq20*rinv20*(rinvsq20-felec);
 323
 324             /* Update potential sums from outer loop */
 325             velecsum        += velec;
 326
 327             fscal            = felec;
 328
 329             /* Calculate temporary vectorial force */
 330             tx               = fscal*dx20;
 331             ty               = fscal*dy20;
 332             tz               = fscal*dz20;
 333
 334             /* Update vectorial force */
 335             fix2            += tx;
 336             fiy2            += ty;
 337             fiz2            += tz;
 338             f[j_coord_offset+DIM*0+XX] -= tx;
 339             f[j_coord_offset+DIM*0+YY] -= ty;
 340             f[j_coord_offset+DIM*0+ZZ] -= tz;
 341
 342             /**************************
 343              * CALCULATE INTERACTIONS *
 344              **************************/
 345
 346             r30              = rsq30*rinv30;
 347
 348             qq30             = iq3*jq0;
 349
 350             /* EWALD ELECTROSTATICS */
 351
 352             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 353             ewrt             = r30*ewtabscale;
 354             ewitab           = ewrt;
 355             eweps            = ewrt-ewitab;
 356             ewitab           = 4*ewitab;
 357             felec            = ewtab[ewitab]+eweps*ewtab[ewitab+1];
 358             velec            = qq30*(rinv30-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
 359             felec            = qq30*rinv30*(rinvsq30-felec);
 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             /* Inner loop uses 172 flops */
 380         }
 381         /* End of innermost loop */
 382
 383         tx = ty = tz = 0;
 384         f[i_coord_offset+DIM*0+XX] += fix0;
 385         f[i_coord_offset+DIM*0+YY] += fiy0;
 386         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 387         tx                         += fix0;
 388         ty                         += fiy0;
 389         tz                         += fiz0;
 390         f[i_coord_offset+DIM*1+XX] += fix1;
 391         f[i_coord_offset+DIM*1+YY] += fiy1;
 392         f[i_coord_offset+DIM*1+ZZ] += fiz1;
 393         tx                         += fix1;
 394         ty                         += fiy1;
 395         tz                         += fiz1;
 396         f[i_coord_offset+DIM*2+XX] += fix2;
 397         f[i_coord_offset+DIM*2+YY] += fiy2;
 398         f[i_coord_offset+DIM*2+ZZ] += fiz2;
 399         tx                         += fix2;
 400         ty                         += fiy2;
 401         tz                         += fiz2;
 402         f[i_coord_offset+DIM*3+XX] += fix3;
 403         f[i_coord_offset+DIM*3+YY] += fiy3;
 404         f[i_coord_offset+DIM*3+ZZ] += fiz3;
 405         tx                         += fix3;
 406         ty                         += fiy3;
 407         tz                         += fiz3;
 408         fshift[i_shift_offset+XX]  += tx;
 409         fshift[i_shift_offset+YY]  += ty;
 410         fshift[i_shift_offset+ZZ]  += tz;
 411
 412         ggid                        = gid[iidx];
 413         /* Update potential energies */
 414         kernel_data->energygrp_elec[ggid] += velecsum;
 415         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 416
 417         /* Increment number of inner iterations */
 418         inneriter                  += j_index_end - j_index_start;
 419
 420         /* Outer loop uses 41 flops */
 421     }
 422
 423     /* Increment number of outer iterations */
 424     outeriter        += nri;
 425
 426     /* Update outer/inner flops */
 427
 428     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_VF,outeriter*41 + inneriter*172);
 429 }
 430 /*
 431  * Gromacs nonbonded kernel:   nb_kernel_ElecEw_VdwLJEw_GeomW4P1_F_c
 432  * Electrostatics interaction: Ewald
 433  * VdW interaction:            LJEwald
 434  * Geometry:                   Water4-Particle
 435  * Calculate force/pot:        Force
 436  */
 437 void
 438 nb_kernel_ElecEw_VdwLJEw_GeomW4P1_F_c
 439                     (t_nblist                    * gmx_restrict       nlist,
 440                      rvec                        * gmx_restrict          xx,
 441                      rvec                        * gmx_restrict          ff,
 442                      t_forcerec                  * gmx_restrict          fr,
 443                      t_mdatoms                   * gmx_restrict     mdatoms,
 444                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 445                      t_nrnb                      * gmx_restrict        nrnb)
 446 {
 447     int              i_shift_offset,i_coord_offset,j_coord_offset;
 448     int              j_index_start,j_index_end;
 449     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 450     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 451     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 452     real             *shiftvec,*fshift,*x,*f;
 453     int              vdwioffset0;
 454     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 455     int              vdwioffset1;
 456     real             ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
 457     int              vdwioffset2;
 458     real             ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
 459     int              vdwioffset3;
 460     real             ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
 461     int              vdwjidx0;
 462     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 463     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 464     real             dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10;
 465     real             dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20;
 466     real             dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30,cexp1_30,cexp2_30;
 467     real             velec,felec,velecsum,facel,crf,krf,krf2;
 468     real             *charge;
 469     int              nvdwtype;
 470     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 471     int              *vdwtype;
 472     real             *vdwparam;
 473     real             c6grid_00;
 474     real             c6grid_10;
 475     real             c6grid_20;
 476     real             c6grid_30;
 477     real             ewclj,ewclj2,ewclj6,ewcljrsq,poly,exponent,sh_lj_ewald;
 478     real             *vdwgridparam;
 479     int              ewitab;
 480     real             ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
 481     real             *ewtab;
 482
 483     x                = xx[0];
 484     f                = ff[0];
 485
 486     nri              = nlist->nri;
 487     iinr             = nlist->iinr;
 488     jindex           = nlist->jindex;
 489     jjnr             = nlist->jjnr;
 490     shiftidx         = nlist->shift;
 491     gid              = nlist->gid;
 492     shiftvec         = fr->shift_vec[0];
 493     fshift           = fr->fshift[0];
 494     facel            = fr->epsfac;
 495     charge           = mdatoms->chargeA;
 496     nvdwtype         = fr->ntype;
 497     vdwparam         = fr->nbfp;
 498     vdwtype          = mdatoms->typeA;
 499     vdwgridparam     = fr->ljpme_c6grid;
 500     ewclj            = fr->ewaldcoeff_lj;
 501     sh_lj_ewald      = fr->ic->sh_lj_ewald;
 502     ewclj2           = ewclj*ewclj;
 503     ewclj6           = ewclj2*ewclj2*ewclj2;
 504
 505     sh_ewald         = fr->ic->sh_ewald;
 506     ewtab            = fr->ic->tabq_coul_F;
 507     ewtabscale       = fr->ic->tabq_scale;
 508     ewtabhalfspace   = 0.5/ewtabscale;
 509
 510     /* Setup water-specific parameters */
 511     inr              = nlist->iinr[0];
 512     iq1              = facel*charge[inr+1];
 513     iq2              = facel*charge[inr+2];
 514     iq3              = facel*charge[inr+3];
 515     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 516
 517     outeriter        = 0;
 518     inneriter        = 0;
 519
 520     /* Start outer loop over neighborlists */
 521     for(iidx=0; iidx<nri; iidx++)
 522     {
 523         /* Load shift vector for this list */
 524         i_shift_offset   = DIM*shiftidx[iidx];
 525         shX              = shiftvec[i_shift_offset+XX];
 526         shY              = shiftvec[i_shift_offset+YY];
 527         shZ              = shiftvec[i_shift_offset+ZZ];
 528
 529         /* Load limits for loop over neighbors */
 530         j_index_start    = jindex[iidx];
 531         j_index_end      = jindex[iidx+1];
 532
 533         /* Get outer coordinate index */
 534         inr              = iinr[iidx];
 535         i_coord_offset   = DIM*inr;
 536
 537         /* Load i particle coords and add shift vector */
 538         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 539         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 540         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 541         ix1              = shX + x[i_coord_offset+DIM*1+XX];
 542         iy1              = shY + x[i_coord_offset+DIM*1+YY];
 543         iz1              = shZ + x[i_coord_offset+DIM*1+ZZ];
 544         ix2              = shX + x[i_coord_offset+DIM*2+XX];
 545         iy2              = shY + x[i_coord_offset+DIM*2+YY];
 546         iz2              = shZ + x[i_coord_offset+DIM*2+ZZ];
 547         ix3              = shX + x[i_coord_offset+DIM*3+XX];
 548         iy3              = shY + x[i_coord_offset+DIM*3+YY];
 549         iz3              = shZ + x[i_coord_offset+DIM*3+ZZ];
 550
 551         fix0             = 0.0;
 552         fiy0             = 0.0;
 553         fiz0             = 0.0;
 554         fix1             = 0.0;
 555         fiy1             = 0.0;
 556         fiz1             = 0.0;
 557         fix2             = 0.0;
 558         fiy2             = 0.0;
 559         fiz2             = 0.0;
 560         fix3             = 0.0;
 561         fiy3             = 0.0;
 562         fiz3             = 0.0;
 563
 564         /* Start inner kernel loop */
 565         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 566         {
 567             /* Get j neighbor index, and coordinate index */
 568             jnr              = jjnr[jidx];
 569             j_coord_offset   = DIM*jnr;
 570
 571             /* load j atom coordinates */
 572             jx0              = x[j_coord_offset+DIM*0+XX];
 573             jy0              = x[j_coord_offset+DIM*0+YY];
 574             jz0              = x[j_coord_offset+DIM*0+ZZ];
 575
 576             /* Calculate displacement vector */
 577             dx00             = ix0 - jx0;
 578             dy00             = iy0 - jy0;
 579             dz00             = iz0 - jz0;
 580             dx10             = ix1 - jx0;
 581             dy10             = iy1 - jy0;
 582             dz10             = iz1 - jz0;
 583             dx20             = ix2 - jx0;
 584             dy20             = iy2 - jy0;
 585             dz20             = iz2 - jz0;
 586             dx30             = ix3 - jx0;
 587             dy30             = iy3 - jy0;
 588             dz30             = iz3 - jz0;
 589
 590             /* Calculate squared distance and things based on it */
 591             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 592             rsq10            = dx10*dx10+dy10*dy10+dz10*dz10;
 593             rsq20            = dx20*dx20+dy20*dy20+dz20*dz20;
 594             rsq30            = dx30*dx30+dy30*dy30+dz30*dz30;
 595
 596             rinv00           = gmx_invsqrt(rsq00);
 597             rinv10           = gmx_invsqrt(rsq10);
 598             rinv20           = gmx_invsqrt(rsq20);
 599             rinv30           = gmx_invsqrt(rsq30);
 600
 601             rinvsq00         = rinv00*rinv00;
 602             rinvsq10         = rinv10*rinv10;
 603             rinvsq20         = rinv20*rinv20;
 604             rinvsq30         = rinv30*rinv30;
 605
 606             /* Load parameters for j particles */
 607             jq0              = charge[jnr+0];
 608             vdwjidx0         = 2*vdwtype[jnr+0];
 609
 610             /**************************
 611              * CALCULATE INTERACTIONS *
 612              **************************/
 613
 614             r00              = rsq00*rinv00;
 615
 616             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 617             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 618             c6grid_00        = vdwgridparam[vdwioffset0+vdwjidx0];
 619
 620             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 621             ewcljrsq         = ewclj2*rsq00;
 622             exponent         = exp(-ewcljrsq);
 623             poly             = exponent*(1.0 + ewcljrsq + ewcljrsq*ewcljrsq*0.5);
 624             fvdw             = (((c12_00*rinvsix - c6_00 + c6grid_00*(1.0-poly))*rinvsix) - c6grid_00*(1.0/6.0)*exponent*ewclj6)*rinvsq00;
 625
 626             fscal            = fvdw;
 627
 628             /* Calculate temporary vectorial force */
 629             tx               = fscal*dx00;
 630             ty               = fscal*dy00;
 631             tz               = fscal*dz00;
 632
 633             /* Update vectorial force */
 634             fix0            += tx;
 635             fiy0            += ty;
 636             fiz0            += tz;
 637             f[j_coord_offset+DIM*0+XX] -= tx;
 638             f[j_coord_offset+DIM*0+YY] -= ty;
 639             f[j_coord_offset+DIM*0+ZZ] -= tz;
 640
 641             /**************************
 642              * CALCULATE INTERACTIONS *
 643              **************************/
 644
 645             r10              = rsq10*rinv10;
 646
 647             qq10             = iq1*jq0;
 648
 649             /* EWALD ELECTROSTATICS */
 650
 651             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 652             ewrt             = r10*ewtabscale;
 653             ewitab           = ewrt;
 654             eweps            = ewrt-ewitab;
 655             felec            = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1];
 656             felec            = qq10*rinv10*(rinvsq10-felec);
 657
 658             fscal            = felec;
 659
 660             /* Calculate temporary vectorial force */
 661             tx               = fscal*dx10;
 662             ty               = fscal*dy10;
 663             tz               = fscal*dz10;
 664
 665             /* Update vectorial force */
 666             fix1            += tx;
 667             fiy1            += ty;
 668             fiz1            += tz;
 669             f[j_coord_offset+DIM*0+XX] -= tx;
 670             f[j_coord_offset+DIM*0+YY] -= ty;
 671             f[j_coord_offset+DIM*0+ZZ] -= tz;
 672
 673             /**************************
 674              * CALCULATE INTERACTIONS *
 675              **************************/
 676
 677             r20              = rsq20*rinv20;
 678
 679             qq20             = iq2*jq0;
 680
 681             /* EWALD ELECTROSTATICS */
 682
 683             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 684             ewrt             = r20*ewtabscale;
 685             ewitab           = ewrt;
 686             eweps            = ewrt-ewitab;
 687             felec            = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1];
 688             felec            = qq20*rinv20*(rinvsq20-felec);
 689
 690             fscal            = felec;
 691
 692             /* Calculate temporary vectorial force */
 693             tx               = fscal*dx20;
 694             ty               = fscal*dy20;
 695             tz               = fscal*dz20;
 696
 697             /* Update vectorial force */
 698             fix2            += tx;
 699             fiy2            += ty;
 700             fiz2            += tz;
 701             f[j_coord_offset+DIM*0+XX] -= tx;
 702             f[j_coord_offset+DIM*0+YY] -= ty;
 703             f[j_coord_offset+DIM*0+ZZ] -= tz;
 704
 705             /**************************
 706              * CALCULATE INTERACTIONS *
 707              **************************/
 708
 709             r30              = rsq30*rinv30;
 710
 711             qq30             = iq3*jq0;
 712
 713             /* EWALD ELECTROSTATICS */
 714
 715             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 716             ewrt             = r30*ewtabscale;
 717             ewitab           = ewrt;
 718             eweps            = ewrt-ewitab;
 719             felec            = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1];
 720             felec            = qq30*rinv30*(rinvsq30-felec);
 721
 722             fscal            = felec;
 723
 724             /* Calculate temporary vectorial force */
 725             tx               = fscal*dx30;
 726             ty               = fscal*dy30;
 727             tz               = fscal*dz30;
 728
 729             /* Update vectorial force */
 730             fix3            += tx;
 731             fiy3            += ty;
 732             fiz3            += tz;
 733             f[j_coord_offset+DIM*0+XX] -= tx;
 734             f[j_coord_offset+DIM*0+YY] -= ty;
 735             f[j_coord_offset+DIM*0+ZZ] -= tz;
 736
 737             /* Inner loop uses 146 flops */
 738         }
 739         /* End of innermost loop */
 740
 741         tx = ty = tz = 0;
 742         f[i_coord_offset+DIM*0+XX] += fix0;
 743         f[i_coord_offset+DIM*0+YY] += fiy0;
 744         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 745         tx                         += fix0;
 746         ty                         += fiy0;
 747         tz                         += fiz0;
 748         f[i_coord_offset+DIM*1+XX] += fix1;
 749         f[i_coord_offset+DIM*1+YY] += fiy1;
 750         f[i_coord_offset+DIM*1+ZZ] += fiz1;
 751         tx                         += fix1;
 752         ty                         += fiy1;
 753         tz                         += fiz1;
 754         f[i_coord_offset+DIM*2+XX] += fix2;
 755         f[i_coord_offset+DIM*2+YY] += fiy2;
 756         f[i_coord_offset+DIM*2+ZZ] += fiz2;
 757         tx                         += fix2;
 758         ty                         += fiy2;
 759         tz                         += fiz2;
 760         f[i_coord_offset+DIM*3+XX] += fix3;
 761         f[i_coord_offset+DIM*3+YY] += fiy3;
 762         f[i_coord_offset+DIM*3+ZZ] += fiz3;
 763         tx                         += fix3;
 764         ty                         += fiy3;
 765         tz                         += fiz3;
 766         fshift[i_shift_offset+XX]  += tx;
 767         fshift[i_shift_offset+YY]  += ty;
 768         fshift[i_shift_offset+ZZ]  += tz;
 769
 770         /* Increment number of inner iterations */
 771         inneriter                  += j_index_end - j_index_start;
 772
 773         /* Outer loop uses 39 flops */
 774     }
 775
 776     /* Increment number of outer iterations */
 777     outeriter        += nri;
 778
 779     /* Update outer/inner flops */
 780
 781     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_F,outeriter*39 + inneriter*146);
 782 }