src/gromacs/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecCoul_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 "types/simple.h"
  44 #include "gromacs/math/vec.h"
  45 #include "nrnb.h"
  46
  47 /*
  48  * Gromacs nonbonded kernel:   nb_kernel_ElecCoul_VdwCSTab_GeomW4P1_VF_c
  49  * Electrostatics interaction: Coulomb
  50  * VdW interaction:            CubicSplineTable
  51  * Geometry:                   Water4-Particle
  52  * Calculate force/pot:        PotentialAndForce
  53  */
  54 void
  55 nb_kernel_ElecCoul_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     nvdwtype         = fr->ntype;
 108     vdwparam         = fr->nbfp;
 109     vdwtype          = mdatoms->typeA;
 110
 111     vftab            = kernel_data->table_vdw->data;
 112     vftabscale       = kernel_data->table_vdw->scale;
 113
 114     /* Setup water-specific parameters */
 115     inr              = nlist->iinr[0];
 116     iq1              = facel*charge[inr+1];
 117     iq2              = facel*charge[inr+2];
 118     iq3              = facel*charge[inr+3];
 119     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 120
 121     outeriter        = 0;
 122     inneriter        = 0;
 123
 124     /* Start outer loop over neighborlists */
 125     for(iidx=0; iidx<nri; iidx++)
 126     {
 127         /* Load shift vector for this list */
 128         i_shift_offset   = DIM*shiftidx[iidx];
 129         shX              = shiftvec[i_shift_offset+XX];
 130         shY              = shiftvec[i_shift_offset+YY];
 131         shZ              = shiftvec[i_shift_offset+ZZ];
 132
 133         /* Load limits for loop over neighbors */
 134         j_index_start    = jindex[iidx];
 135         j_index_end      = jindex[iidx+1];
 136
 137         /* Get outer coordinate index */
 138         inr              = iinr[iidx];
 139         i_coord_offset   = DIM*inr;
 140
 141         /* Load i particle coords and add shift vector */
 142         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 143         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 144         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 145         ix1              = shX + x[i_coord_offset+DIM*1+XX];
 146         iy1              = shY + x[i_coord_offset+DIM*1+YY];
 147         iz1              = shZ + x[i_coord_offset+DIM*1+ZZ];
 148         ix2              = shX + x[i_coord_offset+DIM*2+XX];
 149         iy2              = shY + x[i_coord_offset+DIM*2+YY];
 150         iz2              = shZ + x[i_coord_offset+DIM*2+ZZ];
 151         ix3              = shX + x[i_coord_offset+DIM*3+XX];
 152         iy3              = shY + x[i_coord_offset+DIM*3+YY];
 153         iz3              = shZ + x[i_coord_offset+DIM*3+ZZ];
 154
 155         fix0             = 0.0;
 156         fiy0             = 0.0;
 157         fiz0             = 0.0;
 158         fix1             = 0.0;
 159         fiy1             = 0.0;
 160         fiz1             = 0.0;
 161         fix2             = 0.0;
 162         fiy2             = 0.0;
 163         fiz2             = 0.0;
 164         fix3             = 0.0;
 165         fiy3             = 0.0;
 166         fiz3             = 0.0;
 167
 168         /* Reset potential sums */
 169         velecsum         = 0.0;
 170         vvdwsum          = 0.0;
 171
 172         /* Start inner kernel loop */
 173         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 174         {
 175             /* Get j neighbor index, and coordinate index */
 176             jnr              = jjnr[jidx];
 177             j_coord_offset   = DIM*jnr;
 178
 179             /* load j atom coordinates */
 180             jx0              = x[j_coord_offset+DIM*0+XX];
 181             jy0              = x[j_coord_offset+DIM*0+YY];
 182             jz0              = x[j_coord_offset+DIM*0+ZZ];
 183
 184             /* Calculate displacement vector */
 185             dx00             = ix0 - jx0;
 186             dy00             = iy0 - jy0;
 187             dz00             = iz0 - jz0;
 188             dx10             = ix1 - jx0;
 189             dy10             = iy1 - jy0;
 190             dz10             = iz1 - jz0;
 191             dx20             = ix2 - jx0;
 192             dy20             = iy2 - jy0;
 193             dz20             = iz2 - jz0;
 194             dx30             = ix3 - jx0;
 195             dy30             = iy3 - jy0;
 196             dz30             = iz3 - jz0;
 197
 198             /* Calculate squared distance and things based on it */
 199             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 200             rsq10            = dx10*dx10+dy10*dy10+dz10*dz10;
 201             rsq20            = dx20*dx20+dy20*dy20+dz20*dz20;
 202             rsq30            = dx30*dx30+dy30*dy30+dz30*dz30;
 203
 204             rinv00           = gmx_invsqrt(rsq00);
 205             rinv10           = gmx_invsqrt(rsq10);
 206             rinv20           = gmx_invsqrt(rsq20);
 207             rinv30           = gmx_invsqrt(rsq30);
 208
 209             rinvsq10         = rinv10*rinv10;
 210             rinvsq20         = rinv20*rinv20;
 211             rinvsq30         = rinv30*rinv30;
 212
 213             /* Load parameters for j particles */
 214             jq0              = charge[jnr+0];
 215             vdwjidx0         = 2*vdwtype[jnr+0];
 216
 217             /**************************
 218              * CALCULATE INTERACTIONS *
 219              **************************/
 220
 221             r00              = rsq00*rinv00;
 222
 223             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 224             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 225
 226             /* Calculate table index by multiplying r with table scale and truncate to integer */
 227             rt               = r00*vftabscale;
 228             vfitab           = rt;
 229             vfeps            = rt-vfitab;
 230             vfitab           = 2*4*vfitab;
 231
 232             /* CUBIC SPLINE TABLE DISPERSION */
 233             vfitab          += 0;
 234             Y                = vftab[vfitab];
 235             F                = vftab[vfitab+1];
 236             Geps             = vfeps*vftab[vfitab+2];
 237             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 238             Fp               = F+Geps+Heps2;
 239             VV               = Y+vfeps*Fp;
 240             vvdw6            = c6_00*VV;
 241             FF               = Fp+Geps+2.0*Heps2;
 242             fvdw6            = c6_00*FF;
 243
 244             /* CUBIC SPLINE TABLE REPULSION */
 245             Y                = vftab[vfitab+4];
 246             F                = vftab[vfitab+5];
 247             Geps             = vfeps*vftab[vfitab+6];
 248             Heps2            = vfeps*vfeps*vftab[vfitab+7];
 249             Fp               = F+Geps+Heps2;
 250             VV               = Y+vfeps*Fp;
 251             vvdw12           = c12_00*VV;
 252             FF               = Fp+Geps+2.0*Heps2;
 253             fvdw12           = c12_00*FF;
 254             vvdw             = vvdw12+vvdw6;
 255             fvdw             = -(fvdw6+fvdw12)*vftabscale*rinv00;
 256
 257             /* Update potential sums from outer loop */
 258             vvdwsum         += vvdw;
 259
 260             fscal            = fvdw;
 261
 262             /* Calculate temporary vectorial force */
 263             tx               = fscal*dx00;
 264             ty               = fscal*dy00;
 265             tz               = fscal*dz00;
 266
 267             /* Update vectorial force */
 268             fix0            += tx;
 269             fiy0            += ty;
 270             fiz0            += tz;
 271             f[j_coord_offset+DIM*0+XX] -= tx;
 272             f[j_coord_offset+DIM*0+YY] -= ty;
 273             f[j_coord_offset+DIM*0+ZZ] -= tz;
 274
 275             /**************************
 276              * CALCULATE INTERACTIONS *
 277              **************************/
 278
 279             qq10             = iq1*jq0;
 280
 281             /* COULOMB ELECTROSTATICS */
 282             velec            = qq10*rinv10;
 283             felec            = velec*rinvsq10;
 284
 285             /* Update potential sums from outer loop */
 286             velecsum        += velec;
 287
 288             fscal            = felec;
 289
 290             /* Calculate temporary vectorial force */
 291             tx               = fscal*dx10;
 292             ty               = fscal*dy10;
 293             tz               = fscal*dz10;
 294
 295             /* Update vectorial force */
 296             fix1            += tx;
 297             fiy1            += ty;
 298             fiz1            += tz;
 299             f[j_coord_offset+DIM*0+XX] -= tx;
 300             f[j_coord_offset+DIM*0+YY] -= ty;
 301             f[j_coord_offset+DIM*0+ZZ] -= tz;
 302
 303             /**************************
 304              * CALCULATE INTERACTIONS *
 305              **************************/
 306
 307             qq20             = iq2*jq0;
 308
 309             /* COULOMB ELECTROSTATICS */
 310             velec            = qq20*rinv20;
 311             felec            = velec*rinvsq20;
 312
 313             /* Update potential sums from outer loop */
 314             velecsum        += velec;
 315
 316             fscal            = felec;
 317
 318             /* Calculate temporary vectorial force */
 319             tx               = fscal*dx20;
 320             ty               = fscal*dy20;
 321             tz               = fscal*dz20;
 322
 323             /* Update vectorial force */
 324             fix2            += tx;
 325             fiy2            += ty;
 326             fiz2            += tz;
 327             f[j_coord_offset+DIM*0+XX] -= tx;
 328             f[j_coord_offset+DIM*0+YY] -= ty;
 329             f[j_coord_offset+DIM*0+ZZ] -= tz;
 330
 331             /**************************
 332              * CALCULATE INTERACTIONS *
 333              **************************/
 334
 335             qq30             = iq3*jq0;
 336
 337             /* COULOMB ELECTROSTATICS */
 338             velec            = qq30*rinv30;
 339             felec            = velec*rinvsq30;
 340
 341             /* Update potential sums from outer loop */
 342             velecsum        += velec;
 343
 344             fscal            = felec;
 345
 346             /* Calculate temporary vectorial force */
 347             tx               = fscal*dx30;
 348             ty               = fscal*dy30;
 349             tz               = fscal*dz30;
 350
 351             /* Update vectorial force */
 352             fix3            += tx;
 353             fiy3            += ty;
 354             fiz3            += tz;
 355             f[j_coord_offset+DIM*0+XX] -= tx;
 356             f[j_coord_offset+DIM*0+YY] -= ty;
 357             f[j_coord_offset+DIM*0+ZZ] -= tz;
 358
 359             /* Inner loop uses 139 flops */
 360         }
 361         /* End of innermost loop */
 362
 363         tx = ty = tz = 0;
 364         f[i_coord_offset+DIM*0+XX] += fix0;
 365         f[i_coord_offset+DIM*0+YY] += fiy0;
 366         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 367         tx                         += fix0;
 368         ty                         += fiy0;
 369         tz                         += fiz0;
 370         f[i_coord_offset+DIM*1+XX] += fix1;
 371         f[i_coord_offset+DIM*1+YY] += fiy1;
 372         f[i_coord_offset+DIM*1+ZZ] += fiz1;
 373         tx                         += fix1;
 374         ty                         += fiy1;
 375         tz                         += fiz1;
 376         f[i_coord_offset+DIM*2+XX] += fix2;
 377         f[i_coord_offset+DIM*2+YY] += fiy2;
 378         f[i_coord_offset+DIM*2+ZZ] += fiz2;
 379         tx                         += fix2;
 380         ty                         += fiy2;
 381         tz                         += fiz2;
 382         f[i_coord_offset+DIM*3+XX] += fix3;
 383         f[i_coord_offset+DIM*3+YY] += fiy3;
 384         f[i_coord_offset+DIM*3+ZZ] += fiz3;
 385         tx                         += fix3;
 386         ty                         += fiy3;
 387         tz                         += fiz3;
 388         fshift[i_shift_offset+XX]  += tx;
 389         fshift[i_shift_offset+YY]  += ty;
 390         fshift[i_shift_offset+ZZ]  += tz;
 391
 392         ggid                        = gid[iidx];
 393         /* Update potential energies */
 394         kernel_data->energygrp_elec[ggid] += velecsum;
 395         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 396
 397         /* Increment number of inner iterations */
 398         inneriter                  += j_index_end - j_index_start;
 399
 400         /* Outer loop uses 41 flops */
 401     }
 402
 403     /* Increment number of outer iterations */
 404     outeriter        += nri;
 405
 406     /* Update outer/inner flops */
 407
 408     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_VF,outeriter*41 + inneriter*139);
 409 }
 410 /*
 411  * Gromacs nonbonded kernel:   nb_kernel_ElecCoul_VdwCSTab_GeomW4P1_F_c
 412  * Electrostatics interaction: Coulomb
 413  * VdW interaction:            CubicSplineTable
 414  * Geometry:                   Water4-Particle
 415  * Calculate force/pot:        Force
 416  */
 417 void
 418 nb_kernel_ElecCoul_VdwCSTab_GeomW4P1_F_c
 419                     (t_nblist                    * gmx_restrict       nlist,
 420                      rvec                        * gmx_restrict          xx,
 421                      rvec                        * gmx_restrict          ff,
 422                      t_forcerec                  * gmx_restrict          fr,
 423                      t_mdatoms                   * gmx_restrict     mdatoms,
 424                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 425                      t_nrnb                      * gmx_restrict        nrnb)
 426 {
 427     int              i_shift_offset,i_coord_offset,j_coord_offset;
 428     int              j_index_start,j_index_end;
 429     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 430     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 431     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 432     real             *shiftvec,*fshift,*x,*f;
 433     int              vdwioffset0;
 434     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 435     int              vdwioffset1;
 436     real             ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
 437     int              vdwioffset2;
 438     real             ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
 439     int              vdwioffset3;
 440     real             ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
 441     int              vdwjidx0;
 442     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 443     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 444     real             dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10;
 445     real             dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20;
 446     real             dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30,cexp1_30,cexp2_30;
 447     real             velec,felec,velecsum,facel,crf,krf,krf2;
 448     real             *charge;
 449     int              nvdwtype;
 450     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 451     int              *vdwtype;
 452     real             *vdwparam;
 453     int              vfitab;
 454     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
 455     real             *vftab;
 456
 457     x                = xx[0];
 458     f                = ff[0];
 459
 460     nri              = nlist->nri;
 461     iinr             = nlist->iinr;
 462     jindex           = nlist->jindex;
 463     jjnr             = nlist->jjnr;
 464     shiftidx         = nlist->shift;
 465     gid              = nlist->gid;
 466     shiftvec         = fr->shift_vec[0];
 467     fshift           = fr->fshift[0];
 468     facel            = fr->epsfac;
 469     charge           = mdatoms->chargeA;
 470     nvdwtype         = fr->ntype;
 471     vdwparam         = fr->nbfp;
 472     vdwtype          = mdatoms->typeA;
 473
 474     vftab            = kernel_data->table_vdw->data;
 475     vftabscale       = kernel_data->table_vdw->scale;
 476
 477     /* Setup water-specific parameters */
 478     inr              = nlist->iinr[0];
 479     iq1              = facel*charge[inr+1];
 480     iq2              = facel*charge[inr+2];
 481     iq3              = facel*charge[inr+3];
 482     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 483
 484     outeriter        = 0;
 485     inneriter        = 0;
 486
 487     /* Start outer loop over neighborlists */
 488     for(iidx=0; iidx<nri; iidx++)
 489     {
 490         /* Load shift vector for this list */
 491         i_shift_offset   = DIM*shiftidx[iidx];
 492         shX              = shiftvec[i_shift_offset+XX];
 493         shY              = shiftvec[i_shift_offset+YY];
 494         shZ              = shiftvec[i_shift_offset+ZZ];
 495
 496         /* Load limits for loop over neighbors */
 497         j_index_start    = jindex[iidx];
 498         j_index_end      = jindex[iidx+1];
 499
 500         /* Get outer coordinate index */
 501         inr              = iinr[iidx];
 502         i_coord_offset   = DIM*inr;
 503
 504         /* Load i particle coords and add shift vector */
 505         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 506         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 507         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 508         ix1              = shX + x[i_coord_offset+DIM*1+XX];
 509         iy1              = shY + x[i_coord_offset+DIM*1+YY];
 510         iz1              = shZ + x[i_coord_offset+DIM*1+ZZ];
 511         ix2              = shX + x[i_coord_offset+DIM*2+XX];
 512         iy2              = shY + x[i_coord_offset+DIM*2+YY];
 513         iz2              = shZ + x[i_coord_offset+DIM*2+ZZ];
 514         ix3              = shX + x[i_coord_offset+DIM*3+XX];
 515         iy3              = shY + x[i_coord_offset+DIM*3+YY];
 516         iz3              = shZ + x[i_coord_offset+DIM*3+ZZ];
 517
 518         fix0             = 0.0;
 519         fiy0             = 0.0;
 520         fiz0             = 0.0;
 521         fix1             = 0.0;
 522         fiy1             = 0.0;
 523         fiz1             = 0.0;
 524         fix2             = 0.0;
 525         fiy2             = 0.0;
 526         fiz2             = 0.0;
 527         fix3             = 0.0;
 528         fiy3             = 0.0;
 529         fiz3             = 0.0;
 530
 531         /* Start inner kernel loop */
 532         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 533         {
 534             /* Get j neighbor index, and coordinate index */
 535             jnr              = jjnr[jidx];
 536             j_coord_offset   = DIM*jnr;
 537
 538             /* load j atom coordinates */
 539             jx0              = x[j_coord_offset+DIM*0+XX];
 540             jy0              = x[j_coord_offset+DIM*0+YY];
 541             jz0              = x[j_coord_offset+DIM*0+ZZ];
 542
 543             /* Calculate displacement vector */
 544             dx00             = ix0 - jx0;
 545             dy00             = iy0 - jy0;
 546             dz00             = iz0 - jz0;
 547             dx10             = ix1 - jx0;
 548             dy10             = iy1 - jy0;
 549             dz10             = iz1 - jz0;
 550             dx20             = ix2 - jx0;
 551             dy20             = iy2 - jy0;
 552             dz20             = iz2 - jz0;
 553             dx30             = ix3 - jx0;
 554             dy30             = iy3 - jy0;
 555             dz30             = iz3 - jz0;
 556
 557             /* Calculate squared distance and things based on it */
 558             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 559             rsq10            = dx10*dx10+dy10*dy10+dz10*dz10;
 560             rsq20            = dx20*dx20+dy20*dy20+dz20*dz20;
 561             rsq30            = dx30*dx30+dy30*dy30+dz30*dz30;
 562
 563             rinv00           = gmx_invsqrt(rsq00);
 564             rinv10           = gmx_invsqrt(rsq10);
 565             rinv20           = gmx_invsqrt(rsq20);
 566             rinv30           = gmx_invsqrt(rsq30);
 567
 568             rinvsq10         = rinv10*rinv10;
 569             rinvsq20         = rinv20*rinv20;
 570             rinvsq30         = rinv30*rinv30;
 571
 572             /* Load parameters for j particles */
 573             jq0              = charge[jnr+0];
 574             vdwjidx0         = 2*vdwtype[jnr+0];
 575
 576             /**************************
 577              * CALCULATE INTERACTIONS *
 578              **************************/
 579
 580             r00              = rsq00*rinv00;
 581
 582             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 583             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 584
 585             /* Calculate table index by multiplying r with table scale and truncate to integer */
 586             rt               = r00*vftabscale;
 587             vfitab           = rt;
 588             vfeps            = rt-vfitab;
 589             vfitab           = 2*4*vfitab;
 590
 591             /* CUBIC SPLINE TABLE DISPERSION */
 592             vfitab          += 0;
 593             F                = vftab[vfitab+1];
 594             Geps             = vfeps*vftab[vfitab+2];
 595             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 596             Fp               = F+Geps+Heps2;
 597             FF               = Fp+Geps+2.0*Heps2;
 598             fvdw6            = c6_00*FF;
 599
 600             /* CUBIC SPLINE TABLE REPULSION */
 601             F                = vftab[vfitab+5];
 602             Geps             = vfeps*vftab[vfitab+6];
 603             Heps2            = vfeps*vfeps*vftab[vfitab+7];
 604             Fp               = F+Geps+Heps2;
 605             FF               = Fp+Geps+2.0*Heps2;
 606             fvdw12           = c12_00*FF;
 607             fvdw             = -(fvdw6+fvdw12)*vftabscale*rinv00;
 608
 609             fscal            = fvdw;
 610
 611             /* Calculate temporary vectorial force */
 612             tx               = fscal*dx00;
 613             ty               = fscal*dy00;
 614             tz               = fscal*dz00;
 615
 616             /* Update vectorial force */
 617             fix0            += tx;
 618             fiy0            += ty;
 619             fiz0            += tz;
 620             f[j_coord_offset+DIM*0+XX] -= tx;
 621             f[j_coord_offset+DIM*0+YY] -= ty;
 622             f[j_coord_offset+DIM*0+ZZ] -= tz;
 623
 624             /**************************
 625              * CALCULATE INTERACTIONS *
 626              **************************/
 627
 628             qq10             = iq1*jq0;
 629
 630             /* COULOMB ELECTROSTATICS */
 631             velec            = qq10*rinv10;
 632             felec            = velec*rinvsq10;
 633
 634             fscal            = felec;
 635
 636             /* Calculate temporary vectorial force */
 637             tx               = fscal*dx10;
 638             ty               = fscal*dy10;
 639             tz               = fscal*dz10;
 640
 641             /* Update vectorial force */
 642             fix1            += tx;
 643             fiy1            += ty;
 644             fiz1            += tz;
 645             f[j_coord_offset+DIM*0+XX] -= tx;
 646             f[j_coord_offset+DIM*0+YY] -= ty;
 647             f[j_coord_offset+DIM*0+ZZ] -= tz;
 648
 649             /**************************
 650              * CALCULATE INTERACTIONS *
 651              **************************/
 652
 653             qq20             = iq2*jq0;
 654
 655             /* COULOMB ELECTROSTATICS */
 656             velec            = qq20*rinv20;
 657             felec            = velec*rinvsq20;
 658
 659             fscal            = felec;
 660
 661             /* Calculate temporary vectorial force */
 662             tx               = fscal*dx20;
 663             ty               = fscal*dy20;
 664             tz               = fscal*dz20;
 665
 666             /* Update vectorial force */
 667             fix2            += tx;
 668             fiy2            += ty;
 669             fiz2            += tz;
 670             f[j_coord_offset+DIM*0+XX] -= tx;
 671             f[j_coord_offset+DIM*0+YY] -= ty;
 672             f[j_coord_offset+DIM*0+ZZ] -= tz;
 673
 674             /**************************
 675              * CALCULATE INTERACTIONS *
 676              **************************/
 677
 678             qq30             = iq3*jq0;
 679
 680             /* COULOMB ELECTROSTATICS */
 681             velec            = qq30*rinv30;
 682             felec            = velec*rinvsq30;
 683
 684             fscal            = felec;
 685
 686             /* Calculate temporary vectorial force */
 687             tx               = fscal*dx30;
 688             ty               = fscal*dy30;
 689             tz               = fscal*dz30;
 690
 691             /* Update vectorial force */
 692             fix3            += tx;
 693             fiy3            += ty;
 694             fiz3            += tz;
 695             f[j_coord_offset+DIM*0+XX] -= tx;
 696             f[j_coord_offset+DIM*0+YY] -= ty;
 697             f[j_coord_offset+DIM*0+ZZ] -= tz;
 698
 699             /* Inner loop uses 128 flops */
 700         }
 701         /* End of innermost loop */
 702
 703         tx = ty = tz = 0;
 704         f[i_coord_offset+DIM*0+XX] += fix0;
 705         f[i_coord_offset+DIM*0+YY] += fiy0;
 706         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 707         tx                         += fix0;
 708         ty                         += fiy0;
 709         tz                         += fiz0;
 710         f[i_coord_offset+DIM*1+XX] += fix1;
 711         f[i_coord_offset+DIM*1+YY] += fiy1;
 712         f[i_coord_offset+DIM*1+ZZ] += fiz1;
 713         tx                         += fix1;
 714         ty                         += fiy1;
 715         tz                         += fiz1;
 716         f[i_coord_offset+DIM*2+XX] += fix2;
 717         f[i_coord_offset+DIM*2+YY] += fiy2;
 718         f[i_coord_offset+DIM*2+ZZ] += fiz2;
 719         tx                         += fix2;
 720         ty                         += fiy2;
 721         tz                         += fiz2;
 722         f[i_coord_offset+DIM*3+XX] += fix3;
 723         f[i_coord_offset+DIM*3+YY] += fiy3;
 724         f[i_coord_offset+DIM*3+ZZ] += fiz3;
 725         tx                         += fix3;
 726         ty                         += fiy3;
 727         tz                         += fiz3;
 728         fshift[i_shift_offset+XX]  += tx;
 729         fshift[i_shift_offset+YY]  += ty;
 730         fshift[i_shift_offset+ZZ]  += tz;
 731
 732         /* Increment number of inner iterations */
 733         inneriter                  += j_index_end - j_index_start;
 734
 735         /* Outer loop uses 39 flops */
 736     }
 737
 738     /* Increment number of outer iterations */
 739     outeriter        += nri;
 740
 741     /* Update outer/inner flops */
 742
 743     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_F,outeriter*39 + inneriter*128);
 744 }