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 "gmxpre.h"
  39
  40 #include "config.h"
  41
  42 #include <math.h>
  43
  44 #include "../nb_kernel.h"
  45 #include "gromacs/legacyheaders/types/simple.h"
  46 #include "gromacs/math/vec.h"
  47 #include "gromacs/legacyheaders/nrnb.h"
  48
  49 /*
  50  * Gromacs nonbonded kernel:   nb_kernel_ElecCoul_VdwCSTab_GeomW4P1_VF_c
  51  * Electrostatics interaction: Coulomb
  52  * VdW interaction:            CubicSplineTable
  53  * Geometry:                   Water4-Particle
  54  * Calculate force/pot:        PotentialAndForce
  55  */
  56 void
  57 nb_kernel_ElecCoul_VdwCSTab_GeomW4P1_VF_c
  58                     (t_nblist                    * gmx_restrict       nlist,
  59                      rvec                        * gmx_restrict          xx,
  60                      rvec                        * gmx_restrict          ff,
  61                      t_forcerec                  * gmx_restrict          fr,
  62                      t_mdatoms                   * gmx_restrict     mdatoms,
  63                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
  64                      t_nrnb                      * gmx_restrict        nrnb)
  65 {
  66     int              i_shift_offset,i_coord_offset,j_coord_offset;
  67     int              j_index_start,j_index_end;
  68     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
  69     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
  70     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  71     real             *shiftvec,*fshift,*x,*f;
  72     int              vdwioffset0;
  73     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  74     int              vdwioffset1;
  75     real             ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
  76     int              vdwioffset2;
  77     real             ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
  78     int              vdwioffset3;
  79     real             ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
  80     int              vdwjidx0;
  81     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  82     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
  83     real             dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10;
  84     real             dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20;
  85     real             dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30,cexp1_30,cexp2_30;
  86     real             velec,felec,velecsum,facel,crf,krf,krf2;
  87     real             *charge;
  88     int              nvdwtype;
  89     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
  90     int              *vdwtype;
  91     real             *vdwparam;
  92     int              vfitab;
  93     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
  94     real             *vftab;
  95
  96     x                = xx[0];
  97     f                = ff[0];
  98
  99     nri              = nlist->nri;
 100     iinr             = nlist->iinr;
 101     jindex           = nlist->jindex;
 102     jjnr             = nlist->jjnr;
 103     shiftidx         = nlist->shift;
 104     gid              = nlist->gid;
 105     shiftvec         = fr->shift_vec[0];
 106     fshift           = fr->fshift[0];
 107     facel            = fr->epsfac;
 108     charge           = mdatoms->chargeA;
 109     nvdwtype         = fr->ntype;
 110     vdwparam         = fr->nbfp;
 111     vdwtype          = mdatoms->typeA;
 112
 113     vftab            = kernel_data->table_vdw->data;
 114     vftabscale       = kernel_data->table_vdw->scale;
 115
 116     /* Setup water-specific parameters */
 117     inr              = nlist->iinr[0];
 118     iq1              = facel*charge[inr+1];
 119     iq2              = facel*charge[inr+2];
 120     iq3              = facel*charge[inr+3];
 121     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 122
 123     outeriter        = 0;
 124     inneriter        = 0;
 125
 126     /* Start outer loop over neighborlists */
 127     for(iidx=0; iidx<nri; iidx++)
 128     {
 129         /* Load shift vector for this list */
 130         i_shift_offset   = DIM*shiftidx[iidx];
 131         shX              = shiftvec[i_shift_offset+XX];
 132         shY              = shiftvec[i_shift_offset+YY];
 133         shZ              = shiftvec[i_shift_offset+ZZ];
 134
 135         /* Load limits for loop over neighbors */
 136         j_index_start    = jindex[iidx];
 137         j_index_end      = jindex[iidx+1];
 138
 139         /* Get outer coordinate index */
 140         inr              = iinr[iidx];
 141         i_coord_offset   = DIM*inr;
 142
 143         /* Load i particle coords and add shift vector */
 144         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 145         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 146         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 147         ix1              = shX + x[i_coord_offset+DIM*1+XX];
 148         iy1              = shY + x[i_coord_offset+DIM*1+YY];
 149         iz1              = shZ + x[i_coord_offset+DIM*1+ZZ];
 150         ix2              = shX + x[i_coord_offset+DIM*2+XX];
 151         iy2              = shY + x[i_coord_offset+DIM*2+YY];
 152         iz2              = shZ + x[i_coord_offset+DIM*2+ZZ];
 153         ix3              = shX + x[i_coord_offset+DIM*3+XX];
 154         iy3              = shY + x[i_coord_offset+DIM*3+YY];
 155         iz3              = shZ + x[i_coord_offset+DIM*3+ZZ];
 156
 157         fix0             = 0.0;
 158         fiy0             = 0.0;
 159         fiz0             = 0.0;
 160         fix1             = 0.0;
 161         fiy1             = 0.0;
 162         fiz1             = 0.0;
 163         fix2             = 0.0;
 164         fiy2             = 0.0;
 165         fiz2             = 0.0;
 166         fix3             = 0.0;
 167         fiy3             = 0.0;
 168         fiz3             = 0.0;
 169
 170         /* Reset potential sums */
 171         velecsum         = 0.0;
 172         vvdwsum          = 0.0;
 173
 174         /* Start inner kernel loop */
 175         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 176         {
 177             /* Get j neighbor index, and coordinate index */
 178             jnr              = jjnr[jidx];
 179             j_coord_offset   = DIM*jnr;
 180
 181             /* load j atom coordinates */
 182             jx0              = x[j_coord_offset+DIM*0+XX];
 183             jy0              = x[j_coord_offset+DIM*0+YY];
 184             jz0              = x[j_coord_offset+DIM*0+ZZ];
 185
 186             /* Calculate displacement vector */
 187             dx00             = ix0 - jx0;
 188             dy00             = iy0 - jy0;
 189             dz00             = iz0 - jz0;
 190             dx10             = ix1 - jx0;
 191             dy10             = iy1 - jy0;
 192             dz10             = iz1 - jz0;
 193             dx20             = ix2 - jx0;
 194             dy20             = iy2 - jy0;
 195             dz20             = iz2 - jz0;
 196             dx30             = ix3 - jx0;
 197             dy30             = iy3 - jy0;
 198             dz30             = iz3 - jz0;
 199
 200             /* Calculate squared distance and things based on it */
 201             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 202             rsq10            = dx10*dx10+dy10*dy10+dz10*dz10;
 203             rsq20            = dx20*dx20+dy20*dy20+dz20*dz20;
 204             rsq30            = dx30*dx30+dy30*dy30+dz30*dz30;
 205
 206             rinv00           = gmx_invsqrt(rsq00);
 207             rinv10           = gmx_invsqrt(rsq10);
 208             rinv20           = gmx_invsqrt(rsq20);
 209             rinv30           = gmx_invsqrt(rsq30);
 210
 211             rinvsq10         = rinv10*rinv10;
 212             rinvsq20         = rinv20*rinv20;
 213             rinvsq30         = rinv30*rinv30;
 214
 215             /* Load parameters for j particles */
 216             jq0              = charge[jnr+0];
 217             vdwjidx0         = 2*vdwtype[jnr+0];
 218
 219             /**************************
 220              * CALCULATE INTERACTIONS *
 221              **************************/
 222
 223             r00              = rsq00*rinv00;
 224
 225             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 226             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 227
 228             /* Calculate table index by multiplying r with table scale and truncate to integer */
 229             rt               = r00*vftabscale;
 230             vfitab           = rt;
 231             vfeps            = rt-vfitab;
 232             vfitab           = 2*4*vfitab;
 233
 234             /* CUBIC SPLINE TABLE DISPERSION */
 235             vfitab          += 0;
 236             Y                = vftab[vfitab];
 237             F                = vftab[vfitab+1];
 238             Geps             = vfeps*vftab[vfitab+2];
 239             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 240             Fp               = F+Geps+Heps2;
 241             VV               = Y+vfeps*Fp;
 242             vvdw6            = c6_00*VV;
 243             FF               = Fp+Geps+2.0*Heps2;
 244             fvdw6            = c6_00*FF;
 245
 246             /* CUBIC SPLINE TABLE REPULSION */
 247             Y                = vftab[vfitab+4];
 248             F                = vftab[vfitab+5];
 249             Geps             = vfeps*vftab[vfitab+6];
 250             Heps2            = vfeps*vfeps*vftab[vfitab+7];
 251             Fp               = F+Geps+Heps2;
 252             VV               = Y+vfeps*Fp;
 253             vvdw12           = c12_00*VV;
 254             FF               = Fp+Geps+2.0*Heps2;
 255             fvdw12           = c12_00*FF;
 256             vvdw             = vvdw12+vvdw6;
 257             fvdw             = -(fvdw6+fvdw12)*vftabscale*rinv00;
 258
 259             /* Update potential sums from outer loop */
 260             vvdwsum         += vvdw;
 261
 262             fscal            = fvdw;
 263
 264             /* Calculate temporary vectorial force */
 265             tx               = fscal*dx00;
 266             ty               = fscal*dy00;
 267             tz               = fscal*dz00;
 268
 269             /* Update vectorial force */
 270             fix0            += tx;
 271             fiy0            += ty;
 272             fiz0            += tz;
 273             f[j_coord_offset+DIM*0+XX] -= tx;
 274             f[j_coord_offset+DIM*0+YY] -= ty;
 275             f[j_coord_offset+DIM*0+ZZ] -= tz;
 276
 277             /**************************
 278              * CALCULATE INTERACTIONS *
 279              **************************/
 280
 281             qq10             = iq1*jq0;
 282
 283             /* COULOMB ELECTROSTATICS */
 284             velec            = qq10*rinv10;
 285             felec            = velec*rinvsq10;
 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             qq20             = iq2*jq0;
 310
 311             /* COULOMB ELECTROSTATICS */
 312             velec            = qq20*rinv20;
 313             felec            = velec*rinvsq20;
 314
 315             /* Update potential sums from outer loop */
 316             velecsum        += velec;
 317
 318             fscal            = felec;
 319
 320             /* Calculate temporary vectorial force */
 321             tx               = fscal*dx20;
 322             ty               = fscal*dy20;
 323             tz               = fscal*dz20;
 324
 325             /* Update vectorial force */
 326             fix2            += tx;
 327             fiy2            += ty;
 328             fiz2            += tz;
 329             f[j_coord_offset+DIM*0+XX] -= tx;
 330             f[j_coord_offset+DIM*0+YY] -= ty;
 331             f[j_coord_offset+DIM*0+ZZ] -= tz;
 332
 333             /**************************
 334              * CALCULATE INTERACTIONS *
 335              **************************/
 336
 337             qq30             = iq3*jq0;
 338
 339             /* COULOMB ELECTROSTATICS */
 340             velec            = qq30*rinv30;
 341             felec            = velec*rinvsq30;
 342
 343             /* Update potential sums from outer loop */
 344             velecsum        += velec;
 345
 346             fscal            = felec;
 347
 348             /* Calculate temporary vectorial force */
 349             tx               = fscal*dx30;
 350             ty               = fscal*dy30;
 351             tz               = fscal*dz30;
 352
 353             /* Update vectorial force */
 354             fix3            += tx;
 355             fiy3            += ty;
 356             fiz3            += tz;
 357             f[j_coord_offset+DIM*0+XX] -= tx;
 358             f[j_coord_offset+DIM*0+YY] -= ty;
 359             f[j_coord_offset+DIM*0+ZZ] -= tz;
 360
 361             /* Inner loop uses 139 flops */
 362         }
 363         /* End of innermost loop */
 364
 365         tx = ty = tz = 0;
 366         f[i_coord_offset+DIM*0+XX] += fix0;
 367         f[i_coord_offset+DIM*0+YY] += fiy0;
 368         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 369         tx                         += fix0;
 370         ty                         += fiy0;
 371         tz                         += fiz0;
 372         f[i_coord_offset+DIM*1+XX] += fix1;
 373         f[i_coord_offset+DIM*1+YY] += fiy1;
 374         f[i_coord_offset+DIM*1+ZZ] += fiz1;
 375         tx                         += fix1;
 376         ty                         += fiy1;
 377         tz                         += fiz1;
 378         f[i_coord_offset+DIM*2+XX] += fix2;
 379         f[i_coord_offset+DIM*2+YY] += fiy2;
 380         f[i_coord_offset+DIM*2+ZZ] += fiz2;
 381         tx                         += fix2;
 382         ty                         += fiy2;
 383         tz                         += fiz2;
 384         f[i_coord_offset+DIM*3+XX] += fix3;
 385         f[i_coord_offset+DIM*3+YY] += fiy3;
 386         f[i_coord_offset+DIM*3+ZZ] += fiz3;
 387         tx                         += fix3;
 388         ty                         += fiy3;
 389         tz                         += fiz3;
 390         fshift[i_shift_offset+XX]  += tx;
 391         fshift[i_shift_offset+YY]  += ty;
 392         fshift[i_shift_offset+ZZ]  += tz;
 393
 394         ggid                        = gid[iidx];
 395         /* Update potential energies */
 396         kernel_data->energygrp_elec[ggid] += velecsum;
 397         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 398
 399         /* Increment number of inner iterations */
 400         inneriter                  += j_index_end - j_index_start;
 401
 402         /* Outer loop uses 41 flops */
 403     }
 404
 405     /* Increment number of outer iterations */
 406     outeriter        += nri;
 407
 408     /* Update outer/inner flops */
 409
 410     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_VF,outeriter*41 + inneriter*139);
 411 }
 412 /*
 413  * Gromacs nonbonded kernel:   nb_kernel_ElecCoul_VdwCSTab_GeomW4P1_F_c
 414  * Electrostatics interaction: Coulomb
 415  * VdW interaction:            CubicSplineTable
 416  * Geometry:                   Water4-Particle
 417  * Calculate force/pot:        Force
 418  */
 419 void
 420 nb_kernel_ElecCoul_VdwCSTab_GeomW4P1_F_c
 421                     (t_nblist                    * gmx_restrict       nlist,
 422                      rvec                        * gmx_restrict          xx,
 423                      rvec                        * gmx_restrict          ff,
 424                      t_forcerec                  * gmx_restrict          fr,
 425                      t_mdatoms                   * gmx_restrict     mdatoms,
 426                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 427                      t_nrnb                      * gmx_restrict        nrnb)
 428 {
 429     int              i_shift_offset,i_coord_offset,j_coord_offset;
 430     int              j_index_start,j_index_end;
 431     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 432     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 433     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 434     real             *shiftvec,*fshift,*x,*f;
 435     int              vdwioffset0;
 436     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 437     int              vdwioffset1;
 438     real             ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
 439     int              vdwioffset2;
 440     real             ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
 441     int              vdwioffset3;
 442     real             ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
 443     int              vdwjidx0;
 444     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 445     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 446     real             dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10;
 447     real             dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20;
 448     real             dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30,cexp1_30,cexp2_30;
 449     real             velec,felec,velecsum,facel,crf,krf,krf2;
 450     real             *charge;
 451     int              nvdwtype;
 452     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 453     int              *vdwtype;
 454     real             *vdwparam;
 455     int              vfitab;
 456     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
 457     real             *vftab;
 458
 459     x                = xx[0];
 460     f                = ff[0];
 461
 462     nri              = nlist->nri;
 463     iinr             = nlist->iinr;
 464     jindex           = nlist->jindex;
 465     jjnr             = nlist->jjnr;
 466     shiftidx         = nlist->shift;
 467     gid              = nlist->gid;
 468     shiftvec         = fr->shift_vec[0];
 469     fshift           = fr->fshift[0];
 470     facel            = fr->epsfac;
 471     charge           = mdatoms->chargeA;
 472     nvdwtype         = fr->ntype;
 473     vdwparam         = fr->nbfp;
 474     vdwtype          = mdatoms->typeA;
 475
 476     vftab            = kernel_data->table_vdw->data;
 477     vftabscale       = kernel_data->table_vdw->scale;
 478
 479     /* Setup water-specific parameters */
 480     inr              = nlist->iinr[0];
 481     iq1              = facel*charge[inr+1];
 482     iq2              = facel*charge[inr+2];
 483     iq3              = facel*charge[inr+3];
 484     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 485
 486     outeriter        = 0;
 487     inneriter        = 0;
 488
 489     /* Start outer loop over neighborlists */
 490     for(iidx=0; iidx<nri; iidx++)
 491     {
 492         /* Load shift vector for this list */
 493         i_shift_offset   = DIM*shiftidx[iidx];
 494         shX              = shiftvec[i_shift_offset+XX];
 495         shY              = shiftvec[i_shift_offset+YY];
 496         shZ              = shiftvec[i_shift_offset+ZZ];
 497
 498         /* Load limits for loop over neighbors */
 499         j_index_start    = jindex[iidx];
 500         j_index_end      = jindex[iidx+1];
 501
 502         /* Get outer coordinate index */
 503         inr              = iinr[iidx];
 504         i_coord_offset   = DIM*inr;
 505
 506         /* Load i particle coords and add shift vector */
 507         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 508         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 509         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 510         ix1              = shX + x[i_coord_offset+DIM*1+XX];
 511         iy1              = shY + x[i_coord_offset+DIM*1+YY];
 512         iz1              = shZ + x[i_coord_offset+DIM*1+ZZ];
 513         ix2              = shX + x[i_coord_offset+DIM*2+XX];
 514         iy2              = shY + x[i_coord_offset+DIM*2+YY];
 515         iz2              = shZ + x[i_coord_offset+DIM*2+ZZ];
 516         ix3              = shX + x[i_coord_offset+DIM*3+XX];
 517         iy3              = shY + x[i_coord_offset+DIM*3+YY];
 518         iz3              = shZ + x[i_coord_offset+DIM*3+ZZ];
 519
 520         fix0             = 0.0;
 521         fiy0             = 0.0;
 522         fiz0             = 0.0;
 523         fix1             = 0.0;
 524         fiy1             = 0.0;
 525         fiz1             = 0.0;
 526         fix2             = 0.0;
 527         fiy2             = 0.0;
 528         fiz2             = 0.0;
 529         fix3             = 0.0;
 530         fiy3             = 0.0;
 531         fiz3             = 0.0;
 532
 533         /* Start inner kernel loop */
 534         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 535         {
 536             /* Get j neighbor index, and coordinate index */
 537             jnr              = jjnr[jidx];
 538             j_coord_offset   = DIM*jnr;
 539
 540             /* load j atom coordinates */
 541             jx0              = x[j_coord_offset+DIM*0+XX];
 542             jy0              = x[j_coord_offset+DIM*0+YY];
 543             jz0              = x[j_coord_offset+DIM*0+ZZ];
 544
 545             /* Calculate displacement vector */
 546             dx00             = ix0 - jx0;
 547             dy00             = iy0 - jy0;
 548             dz00             = iz0 - jz0;
 549             dx10             = ix1 - jx0;
 550             dy10             = iy1 - jy0;
 551             dz10             = iz1 - jz0;
 552             dx20             = ix2 - jx0;
 553             dy20             = iy2 - jy0;
 554             dz20             = iz2 - jz0;
 555             dx30             = ix3 - jx0;
 556             dy30             = iy3 - jy0;
 557             dz30             = iz3 - jz0;
 558
 559             /* Calculate squared distance and things based on it */
 560             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 561             rsq10            = dx10*dx10+dy10*dy10+dz10*dz10;
 562             rsq20            = dx20*dx20+dy20*dy20+dz20*dz20;
 563             rsq30            = dx30*dx30+dy30*dy30+dz30*dz30;
 564
 565             rinv00           = gmx_invsqrt(rsq00);
 566             rinv10           = gmx_invsqrt(rsq10);
 567             rinv20           = gmx_invsqrt(rsq20);
 568             rinv30           = gmx_invsqrt(rsq30);
 569
 570             rinvsq10         = rinv10*rinv10;
 571             rinvsq20         = rinv20*rinv20;
 572             rinvsq30         = rinv30*rinv30;
 573
 574             /* Load parameters for j particles */
 575             jq0              = charge[jnr+0];
 576             vdwjidx0         = 2*vdwtype[jnr+0];
 577
 578             /**************************
 579              * CALCULATE INTERACTIONS *
 580              **************************/
 581
 582             r00              = rsq00*rinv00;
 583
 584             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 585             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 586
 587             /* Calculate table index by multiplying r with table scale and truncate to integer */
 588             rt               = r00*vftabscale;
 589             vfitab           = rt;
 590             vfeps            = rt-vfitab;
 591             vfitab           = 2*4*vfitab;
 592
 593             /* CUBIC SPLINE TABLE DISPERSION */
 594             vfitab          += 0;
 595             F                = vftab[vfitab+1];
 596             Geps             = vfeps*vftab[vfitab+2];
 597             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 598             Fp               = F+Geps+Heps2;
 599             FF               = Fp+Geps+2.0*Heps2;
 600             fvdw6            = c6_00*FF;
 601
 602             /* CUBIC SPLINE TABLE REPULSION */
 603             F                = vftab[vfitab+5];
 604             Geps             = vfeps*vftab[vfitab+6];
 605             Heps2            = vfeps*vfeps*vftab[vfitab+7];
 606             Fp               = F+Geps+Heps2;
 607             FF               = Fp+Geps+2.0*Heps2;
 608             fvdw12           = c12_00*FF;
 609             fvdw             = -(fvdw6+fvdw12)*vftabscale*rinv00;
 610
 611             fscal            = fvdw;
 612
 613             /* Calculate temporary vectorial force */
 614             tx               = fscal*dx00;
 615             ty               = fscal*dy00;
 616             tz               = fscal*dz00;
 617
 618             /* Update vectorial force */
 619             fix0            += tx;
 620             fiy0            += ty;
 621             fiz0            += tz;
 622             f[j_coord_offset+DIM*0+XX] -= tx;
 623             f[j_coord_offset+DIM*0+YY] -= ty;
 624             f[j_coord_offset+DIM*0+ZZ] -= tz;
 625
 626             /**************************
 627              * CALCULATE INTERACTIONS *
 628              **************************/
 629
 630             qq10             = iq1*jq0;
 631
 632             /* COULOMB ELECTROSTATICS */
 633             velec            = qq10*rinv10;
 634             felec            = velec*rinvsq10;
 635
 636             fscal            = felec;
 637
 638             /* Calculate temporary vectorial force */
 639             tx               = fscal*dx10;
 640             ty               = fscal*dy10;
 641             tz               = fscal*dz10;
 642
 643             /* Update vectorial force */
 644             fix1            += tx;
 645             fiy1            += ty;
 646             fiz1            += tz;
 647             f[j_coord_offset+DIM*0+XX] -= tx;
 648             f[j_coord_offset+DIM*0+YY] -= ty;
 649             f[j_coord_offset+DIM*0+ZZ] -= tz;
 650
 651             /**************************
 652              * CALCULATE INTERACTIONS *
 653              **************************/
 654
 655             qq20             = iq2*jq0;
 656
 657             /* COULOMB ELECTROSTATICS */
 658             velec            = qq20*rinv20;
 659             felec            = velec*rinvsq20;
 660
 661             fscal            = felec;
 662
 663             /* Calculate temporary vectorial force */
 664             tx               = fscal*dx20;
 665             ty               = fscal*dy20;
 666             tz               = fscal*dz20;
 667
 668             /* Update vectorial force */
 669             fix2            += tx;
 670             fiy2            += ty;
 671             fiz2            += tz;
 672             f[j_coord_offset+DIM*0+XX] -= tx;
 673             f[j_coord_offset+DIM*0+YY] -= ty;
 674             f[j_coord_offset+DIM*0+ZZ] -= tz;
 675
 676             /**************************
 677              * CALCULATE INTERACTIONS *
 678              **************************/
 679
 680             qq30             = iq3*jq0;
 681
 682             /* COULOMB ELECTROSTATICS */
 683             velec            = qq30*rinv30;
 684             felec            = velec*rinvsq30;
 685
 686             fscal            = felec;
 687
 688             /* Calculate temporary vectorial force */
 689             tx               = fscal*dx30;
 690             ty               = fscal*dy30;
 691             tz               = fscal*dz30;
 692
 693             /* Update vectorial force */
 694             fix3            += tx;
 695             fiy3            += ty;
 696             fiz3            += tz;
 697             f[j_coord_offset+DIM*0+XX] -= tx;
 698             f[j_coord_offset+DIM*0+YY] -= ty;
 699             f[j_coord_offset+DIM*0+ZZ] -= tz;
 700
 701             /* Inner loop uses 128 flops */
 702         }
 703         /* End of innermost loop */
 704
 705         tx = ty = tz = 0;
 706         f[i_coord_offset+DIM*0+XX] += fix0;
 707         f[i_coord_offset+DIM*0+YY] += fiy0;
 708         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 709         tx                         += fix0;
 710         ty                         += fiy0;
 711         tz                         += fiz0;
 712         f[i_coord_offset+DIM*1+XX] += fix1;
 713         f[i_coord_offset+DIM*1+YY] += fiy1;
 714         f[i_coord_offset+DIM*1+ZZ] += fiz1;
 715         tx                         += fix1;
 716         ty                         += fiy1;
 717         tz                         += fiz1;
 718         f[i_coord_offset+DIM*2+XX] += fix2;
 719         f[i_coord_offset+DIM*2+YY] += fiy2;
 720         f[i_coord_offset+DIM*2+ZZ] += fiz2;
 721         tx                         += fix2;
 722         ty                         += fiy2;
 723         tz                         += fiz2;
 724         f[i_coord_offset+DIM*3+XX] += fix3;
 725         f[i_coord_offset+DIM*3+YY] += fiy3;
 726         f[i_coord_offset+DIM*3+ZZ] += fiz3;
 727         tx                         += fix3;
 728         ty                         += fiy3;
 729         tz                         += fiz3;
 730         fshift[i_shift_offset+XX]  += tx;
 731         fshift[i_shift_offset+YY]  += ty;
 732         fshift[i_shift_offset+ZZ]  += tz;
 733
 734         /* Increment number of inner iterations */
 735         inneriter                  += j_index_end - j_index_start;
 736
 737         /* Outer loop uses 39 flops */
 738     }
 739
 740     /* Increment number of outer iterations */
 741     outeriter        += nri;
 742
 743     /* Update outer/inner flops */
 744
 745     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_F,outeriter*39 + inneriter*128);
 746 }