src/gromacs/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecRF_VdwLJ_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_ElecRF_VdwLJ_GeomW4P1_VF_c
  51  * Electrostatics interaction: ReactionField
  52  * VdW interaction:            LennardJones
  53  * Geometry:                   Water4-Particle
  54  * Calculate force/pot:        PotentialAndForce
  55  */
  56 void
  57 nb_kernel_ElecRF_VdwLJ_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
  93     x                = xx[0];
  94     f                = ff[0];
  95
  96     nri              = nlist->nri;
  97     iinr             = nlist->iinr;
  98     jindex           = nlist->jindex;
  99     jjnr             = nlist->jjnr;
 100     shiftidx         = nlist->shift;
 101     gid              = nlist->gid;
 102     shiftvec         = fr->shift_vec[0];
 103     fshift           = fr->fshift[0];
 104     facel            = fr->epsfac;
 105     charge           = mdatoms->chargeA;
 106     krf              = fr->ic->k_rf;
 107     krf2             = krf*2.0;
 108     crf              = fr->ic->c_rf;
 109     nvdwtype         = fr->ntype;
 110     vdwparam         = fr->nbfp;
 111     vdwtype          = mdatoms->typeA;
 112
 113     /* Setup water-specific parameters */
 114     inr              = nlist->iinr[0];
 115     iq1              = facel*charge[inr+1];
 116     iq2              = facel*charge[inr+2];
 117     iq3              = facel*charge[inr+3];
 118     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 119
 120     outeriter        = 0;
 121     inneriter        = 0;
 122
 123     /* Start outer loop over neighborlists */
 124     for(iidx=0; iidx<nri; iidx++)
 125     {
 126         /* Load shift vector for this list */
 127         i_shift_offset   = DIM*shiftidx[iidx];
 128         shX              = shiftvec[i_shift_offset+XX];
 129         shY              = shiftvec[i_shift_offset+YY];
 130         shZ              = shiftvec[i_shift_offset+ZZ];
 131
 132         /* Load limits for loop over neighbors */
 133         j_index_start    = jindex[iidx];
 134         j_index_end      = jindex[iidx+1];
 135
 136         /* Get outer coordinate index */
 137         inr              = iinr[iidx];
 138         i_coord_offset   = DIM*inr;
 139
 140         /* Load i particle coords and add shift vector */
 141         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 142         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 143         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 144         ix1              = shX + x[i_coord_offset+DIM*1+XX];
 145         iy1              = shY + x[i_coord_offset+DIM*1+YY];
 146         iz1              = shZ + x[i_coord_offset+DIM*1+ZZ];
 147         ix2              = shX + x[i_coord_offset+DIM*2+XX];
 148         iy2              = shY + x[i_coord_offset+DIM*2+YY];
 149         iz2              = shZ + x[i_coord_offset+DIM*2+ZZ];
 150         ix3              = shX + x[i_coord_offset+DIM*3+XX];
 151         iy3              = shY + x[i_coord_offset+DIM*3+YY];
 152         iz3              = shZ + x[i_coord_offset+DIM*3+ZZ];
 153
 154         fix0             = 0.0;
 155         fiy0             = 0.0;
 156         fiz0             = 0.0;
 157         fix1             = 0.0;
 158         fiy1             = 0.0;
 159         fiz1             = 0.0;
 160         fix2             = 0.0;
 161         fiy2             = 0.0;
 162         fiz2             = 0.0;
 163         fix3             = 0.0;
 164         fiy3             = 0.0;
 165         fiz3             = 0.0;
 166
 167         /* Reset potential sums */
 168         velecsum         = 0.0;
 169         vvdwsum          = 0.0;
 170
 171         /* Start inner kernel loop */
 172         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 173         {
 174             /* Get j neighbor index, and coordinate index */
 175             jnr              = jjnr[jidx];
 176             j_coord_offset   = DIM*jnr;
 177
 178             /* load j atom coordinates */
 179             jx0              = x[j_coord_offset+DIM*0+XX];
 180             jy0              = x[j_coord_offset+DIM*0+YY];
 181             jz0              = x[j_coord_offset+DIM*0+ZZ];
 182
 183             /* Calculate displacement vector */
 184             dx00             = ix0 - jx0;
 185             dy00             = iy0 - jy0;
 186             dz00             = iz0 - jz0;
 187             dx10             = ix1 - jx0;
 188             dy10             = iy1 - jy0;
 189             dz10             = iz1 - jz0;
 190             dx20             = ix2 - jx0;
 191             dy20             = iy2 - jy0;
 192             dz20             = iz2 - jz0;
 193             dx30             = ix3 - jx0;
 194             dy30             = iy3 - jy0;
 195             dz30             = iz3 - jz0;
 196
 197             /* Calculate squared distance and things based on it */
 198             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 199             rsq10            = dx10*dx10+dy10*dy10+dz10*dz10;
 200             rsq20            = dx20*dx20+dy20*dy20+dz20*dz20;
 201             rsq30            = dx30*dx30+dy30*dy30+dz30*dz30;
 202
 203             rinv10           = gmx_invsqrt(rsq10);
 204             rinv20           = gmx_invsqrt(rsq20);
 205             rinv30           = gmx_invsqrt(rsq30);
 206
 207             rinvsq00         = 1.0/rsq00;
 208             rinvsq10         = rinv10*rinv10;
 209             rinvsq20         = rinv20*rinv20;
 210             rinvsq30         = rinv30*rinv30;
 211
 212             /* Load parameters for j particles */
 213             jq0              = charge[jnr+0];
 214             vdwjidx0         = 2*vdwtype[jnr+0];
 215
 216             /**************************
 217              * CALCULATE INTERACTIONS *
 218              **************************/
 219
 220             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 221             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 222
 223             /* LENNARD-JONES DISPERSION/REPULSION */
 224
 225             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 226             vvdw6            = c6_00*rinvsix;
 227             vvdw12           = c12_00*rinvsix*rinvsix;
 228             vvdw             = vvdw12*(1.0/12.0) - vvdw6*(1.0/6.0);
 229             fvdw             = (vvdw12-vvdw6)*rinvsq00;
 230
 231             /* Update potential sums from outer loop */
 232             vvdwsum         += vvdw;
 233
 234             fscal            = fvdw;
 235
 236             /* Calculate temporary vectorial force */
 237             tx               = fscal*dx00;
 238             ty               = fscal*dy00;
 239             tz               = fscal*dz00;
 240
 241             /* Update vectorial force */
 242             fix0            += tx;
 243             fiy0            += ty;
 244             fiz0            += tz;
 245             f[j_coord_offset+DIM*0+XX] -= tx;
 246             f[j_coord_offset+DIM*0+YY] -= ty;
 247             f[j_coord_offset+DIM*0+ZZ] -= tz;
 248
 249             /**************************
 250              * CALCULATE INTERACTIONS *
 251              **************************/
 252
 253             qq10             = iq1*jq0;
 254
 255             /* REACTION-FIELD ELECTROSTATICS */
 256             velec            = qq10*(rinv10+krf*rsq10-crf);
 257             felec            = qq10*(rinv10*rinvsq10-krf2);
 258
 259             /* Update potential sums from outer loop */
 260             velecsum        += velec;
 261
 262             fscal            = felec;
 263
 264             /* Calculate temporary vectorial force */
 265             tx               = fscal*dx10;
 266             ty               = fscal*dy10;
 267             tz               = fscal*dz10;
 268
 269             /* Update vectorial force */
 270             fix1            += tx;
 271             fiy1            += ty;
 272             fiz1            += 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             qq20             = iq2*jq0;
 282
 283             /* REACTION-FIELD ELECTROSTATICS */
 284             velec            = qq20*(rinv20+krf*rsq20-crf);
 285             felec            = qq20*(rinv20*rinvsq20-krf2);
 286
 287             /* Update potential sums from outer loop */
 288             velecsum        += velec;
 289
 290             fscal            = felec;
 291
 292             /* Calculate temporary vectorial force */
 293             tx               = fscal*dx20;
 294             ty               = fscal*dy20;
 295             tz               = fscal*dz20;
 296
 297             /* Update vectorial force */
 298             fix2            += tx;
 299             fiy2            += ty;
 300             fiz2            += 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             qq30             = iq3*jq0;
 310
 311             /* REACTION-FIELD ELECTROSTATICS */
 312             velec            = qq30*(rinv30+krf*rsq30-crf);
 313             felec            = qq30*(rinv30*rinvsq30-krf2);
 314
 315             /* Update potential sums from outer loop */
 316             velecsum        += velec;
 317
 318             fscal            = felec;
 319
 320             /* Calculate temporary vectorial force */
 321             tx               = fscal*dx30;
 322             ty               = fscal*dy30;
 323             tz               = fscal*dz30;
 324
 325             /* Update vectorial force */
 326             fix3            += tx;
 327             fiy3            += ty;
 328             fiz3            += 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             /* Inner loop uses 128 flops */
 334         }
 335         /* End of innermost loop */
 336
 337         tx = ty = tz = 0;
 338         f[i_coord_offset+DIM*0+XX] += fix0;
 339         f[i_coord_offset+DIM*0+YY] += fiy0;
 340         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 341         tx                         += fix0;
 342         ty                         += fiy0;
 343         tz                         += fiz0;
 344         f[i_coord_offset+DIM*1+XX] += fix1;
 345         f[i_coord_offset+DIM*1+YY] += fiy1;
 346         f[i_coord_offset+DIM*1+ZZ] += fiz1;
 347         tx                         += fix1;
 348         ty                         += fiy1;
 349         tz                         += fiz1;
 350         f[i_coord_offset+DIM*2+XX] += fix2;
 351         f[i_coord_offset+DIM*2+YY] += fiy2;
 352         f[i_coord_offset+DIM*2+ZZ] += fiz2;
 353         tx                         += fix2;
 354         ty                         += fiy2;
 355         tz                         += fiz2;
 356         f[i_coord_offset+DIM*3+XX] += fix3;
 357         f[i_coord_offset+DIM*3+YY] += fiy3;
 358         f[i_coord_offset+DIM*3+ZZ] += fiz3;
 359         tx                         += fix3;
 360         ty                         += fiy3;
 361         tz                         += fiz3;
 362         fshift[i_shift_offset+XX]  += tx;
 363         fshift[i_shift_offset+YY]  += ty;
 364         fshift[i_shift_offset+ZZ]  += tz;
 365
 366         ggid                        = gid[iidx];
 367         /* Update potential energies */
 368         kernel_data->energygrp_elec[ggid] += velecsum;
 369         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 370
 371         /* Increment number of inner iterations */
 372         inneriter                  += j_index_end - j_index_start;
 373
 374         /* Outer loop uses 41 flops */
 375     }
 376
 377     /* Increment number of outer iterations */
 378     outeriter        += nri;
 379
 380     /* Update outer/inner flops */
 381
 382     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_VF,outeriter*41 + inneriter*128);
 383 }
 384 /*
 385  * Gromacs nonbonded kernel:   nb_kernel_ElecRF_VdwLJ_GeomW4P1_F_c
 386  * Electrostatics interaction: ReactionField
 387  * VdW interaction:            LennardJones
 388  * Geometry:                   Water4-Particle
 389  * Calculate force/pot:        Force
 390  */
 391 void
 392 nb_kernel_ElecRF_VdwLJ_GeomW4P1_F_c
 393                     (t_nblist                    * gmx_restrict       nlist,
 394                      rvec                        * gmx_restrict          xx,
 395                      rvec                        * gmx_restrict          ff,
 396                      t_forcerec                  * gmx_restrict          fr,
 397                      t_mdatoms                   * gmx_restrict     mdatoms,
 398                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 399                      t_nrnb                      * gmx_restrict        nrnb)
 400 {
 401     int              i_shift_offset,i_coord_offset,j_coord_offset;
 402     int              j_index_start,j_index_end;
 403     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 404     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 405     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 406     real             *shiftvec,*fshift,*x,*f;
 407     int              vdwioffset0;
 408     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 409     int              vdwioffset1;
 410     real             ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
 411     int              vdwioffset2;
 412     real             ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
 413     int              vdwioffset3;
 414     real             ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
 415     int              vdwjidx0;
 416     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 417     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 418     real             dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10;
 419     real             dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20;
 420     real             dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30,cexp1_30,cexp2_30;
 421     real             velec,felec,velecsum,facel,crf,krf,krf2;
 422     real             *charge;
 423     int              nvdwtype;
 424     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 425     int              *vdwtype;
 426     real             *vdwparam;
 427
 428     x                = xx[0];
 429     f                = ff[0];
 430
 431     nri              = nlist->nri;
 432     iinr             = nlist->iinr;
 433     jindex           = nlist->jindex;
 434     jjnr             = nlist->jjnr;
 435     shiftidx         = nlist->shift;
 436     gid              = nlist->gid;
 437     shiftvec         = fr->shift_vec[0];
 438     fshift           = fr->fshift[0];
 439     facel            = fr->epsfac;
 440     charge           = mdatoms->chargeA;
 441     krf              = fr->ic->k_rf;
 442     krf2             = krf*2.0;
 443     crf              = fr->ic->c_rf;
 444     nvdwtype         = fr->ntype;
 445     vdwparam         = fr->nbfp;
 446     vdwtype          = mdatoms->typeA;
 447
 448     /* Setup water-specific parameters */
 449     inr              = nlist->iinr[0];
 450     iq1              = facel*charge[inr+1];
 451     iq2              = facel*charge[inr+2];
 452     iq3              = facel*charge[inr+3];
 453     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 454
 455     outeriter        = 0;
 456     inneriter        = 0;
 457
 458     /* Start outer loop over neighborlists */
 459     for(iidx=0; iidx<nri; iidx++)
 460     {
 461         /* Load shift vector for this list */
 462         i_shift_offset   = DIM*shiftidx[iidx];
 463         shX              = shiftvec[i_shift_offset+XX];
 464         shY              = shiftvec[i_shift_offset+YY];
 465         shZ              = shiftvec[i_shift_offset+ZZ];
 466
 467         /* Load limits for loop over neighbors */
 468         j_index_start    = jindex[iidx];
 469         j_index_end      = jindex[iidx+1];
 470
 471         /* Get outer coordinate index */
 472         inr              = iinr[iidx];
 473         i_coord_offset   = DIM*inr;
 474
 475         /* Load i particle coords and add shift vector */
 476         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 477         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 478         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 479         ix1              = shX + x[i_coord_offset+DIM*1+XX];
 480         iy1              = shY + x[i_coord_offset+DIM*1+YY];
 481         iz1              = shZ + x[i_coord_offset+DIM*1+ZZ];
 482         ix2              = shX + x[i_coord_offset+DIM*2+XX];
 483         iy2              = shY + x[i_coord_offset+DIM*2+YY];
 484         iz2              = shZ + x[i_coord_offset+DIM*2+ZZ];
 485         ix3              = shX + x[i_coord_offset+DIM*3+XX];
 486         iy3              = shY + x[i_coord_offset+DIM*3+YY];
 487         iz3              = shZ + x[i_coord_offset+DIM*3+ZZ];
 488
 489         fix0             = 0.0;
 490         fiy0             = 0.0;
 491         fiz0             = 0.0;
 492         fix1             = 0.0;
 493         fiy1             = 0.0;
 494         fiz1             = 0.0;
 495         fix2             = 0.0;
 496         fiy2             = 0.0;
 497         fiz2             = 0.0;
 498         fix3             = 0.0;
 499         fiy3             = 0.0;
 500         fiz3             = 0.0;
 501
 502         /* Start inner kernel loop */
 503         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 504         {
 505             /* Get j neighbor index, and coordinate index */
 506             jnr              = jjnr[jidx];
 507             j_coord_offset   = DIM*jnr;
 508
 509             /* load j atom coordinates */
 510             jx0              = x[j_coord_offset+DIM*0+XX];
 511             jy0              = x[j_coord_offset+DIM*0+YY];
 512             jz0              = x[j_coord_offset+DIM*0+ZZ];
 513
 514             /* Calculate displacement vector */
 515             dx00             = ix0 - jx0;
 516             dy00             = iy0 - jy0;
 517             dz00             = iz0 - jz0;
 518             dx10             = ix1 - jx0;
 519             dy10             = iy1 - jy0;
 520             dz10             = iz1 - jz0;
 521             dx20             = ix2 - jx0;
 522             dy20             = iy2 - jy0;
 523             dz20             = iz2 - jz0;
 524             dx30             = ix3 - jx0;
 525             dy30             = iy3 - jy0;
 526             dz30             = iz3 - jz0;
 527
 528             /* Calculate squared distance and things based on it */
 529             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 530             rsq10            = dx10*dx10+dy10*dy10+dz10*dz10;
 531             rsq20            = dx20*dx20+dy20*dy20+dz20*dz20;
 532             rsq30            = dx30*dx30+dy30*dy30+dz30*dz30;
 533
 534             rinv10           = gmx_invsqrt(rsq10);
 535             rinv20           = gmx_invsqrt(rsq20);
 536             rinv30           = gmx_invsqrt(rsq30);
 537
 538             rinvsq00         = 1.0/rsq00;
 539             rinvsq10         = rinv10*rinv10;
 540             rinvsq20         = rinv20*rinv20;
 541             rinvsq30         = rinv30*rinv30;
 542
 543             /* Load parameters for j particles */
 544             jq0              = charge[jnr+0];
 545             vdwjidx0         = 2*vdwtype[jnr+0];
 546
 547             /**************************
 548              * CALCULATE INTERACTIONS *
 549              **************************/
 550
 551             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 552             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 553
 554             /* LENNARD-JONES DISPERSION/REPULSION */
 555
 556             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 557             fvdw             = (c12_00*rinvsix-c6_00)*rinvsix*rinvsq00;
 558
 559             fscal            = fvdw;
 560
 561             /* Calculate temporary vectorial force */
 562             tx               = fscal*dx00;
 563             ty               = fscal*dy00;
 564             tz               = fscal*dz00;
 565
 566             /* Update vectorial force */
 567             fix0            += tx;
 568             fiy0            += ty;
 569             fiz0            += tz;
 570             f[j_coord_offset+DIM*0+XX] -= tx;
 571             f[j_coord_offset+DIM*0+YY] -= ty;
 572             f[j_coord_offset+DIM*0+ZZ] -= tz;
 573
 574             /**************************
 575              * CALCULATE INTERACTIONS *
 576              **************************/
 577
 578             qq10             = iq1*jq0;
 579
 580             /* REACTION-FIELD ELECTROSTATICS */
 581             felec            = qq10*(rinv10*rinvsq10-krf2);
 582
 583             fscal            = felec;
 584
 585             /* Calculate temporary vectorial force */
 586             tx               = fscal*dx10;
 587             ty               = fscal*dy10;
 588             tz               = fscal*dz10;
 589
 590             /* Update vectorial force */
 591             fix1            += tx;
 592             fiy1            += ty;
 593             fiz1            += tz;
 594             f[j_coord_offset+DIM*0+XX] -= tx;
 595             f[j_coord_offset+DIM*0+YY] -= ty;
 596             f[j_coord_offset+DIM*0+ZZ] -= tz;
 597
 598             /**************************
 599              * CALCULATE INTERACTIONS *
 600              **************************/
 601
 602             qq20             = iq2*jq0;
 603
 604             /* REACTION-FIELD ELECTROSTATICS */
 605             felec            = qq20*(rinv20*rinvsq20-krf2);
 606
 607             fscal            = felec;
 608
 609             /* Calculate temporary vectorial force */
 610             tx               = fscal*dx20;
 611             ty               = fscal*dy20;
 612             tz               = fscal*dz20;
 613
 614             /* Update vectorial force */
 615             fix2            += tx;
 616             fiy2            += ty;
 617             fiz2            += tz;
 618             f[j_coord_offset+DIM*0+XX] -= tx;
 619             f[j_coord_offset+DIM*0+YY] -= ty;
 620             f[j_coord_offset+DIM*0+ZZ] -= tz;
 621
 622             /**************************
 623              * CALCULATE INTERACTIONS *
 624              **************************/
 625
 626             qq30             = iq3*jq0;
 627
 628             /* REACTION-FIELD ELECTROSTATICS */
 629             felec            = qq30*(rinv30*rinvsq30-krf2);
 630
 631             fscal            = felec;
 632
 633             /* Calculate temporary vectorial force */
 634             tx               = fscal*dx30;
 635             ty               = fscal*dy30;
 636             tz               = fscal*dz30;
 637
 638             /* Update vectorial force */
 639             fix3            += tx;
 640             fiy3            += ty;
 641             fiz3            += tz;
 642             f[j_coord_offset+DIM*0+XX] -= tx;
 643             f[j_coord_offset+DIM*0+YY] -= ty;
 644             f[j_coord_offset+DIM*0+ZZ] -= tz;
 645
 646             /* Inner loop uses 108 flops */
 647         }
 648         /* End of innermost loop */
 649
 650         tx = ty = tz = 0;
 651         f[i_coord_offset+DIM*0+XX] += fix0;
 652         f[i_coord_offset+DIM*0+YY] += fiy0;
 653         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 654         tx                         += fix0;
 655         ty                         += fiy0;
 656         tz                         += fiz0;
 657         f[i_coord_offset+DIM*1+XX] += fix1;
 658         f[i_coord_offset+DIM*1+YY] += fiy1;
 659         f[i_coord_offset+DIM*1+ZZ] += fiz1;
 660         tx                         += fix1;
 661         ty                         += fiy1;
 662         tz                         += fiz1;
 663         f[i_coord_offset+DIM*2+XX] += fix2;
 664         f[i_coord_offset+DIM*2+YY] += fiy2;
 665         f[i_coord_offset+DIM*2+ZZ] += fiz2;
 666         tx                         += fix2;
 667         ty                         += fiy2;
 668         tz                         += fiz2;
 669         f[i_coord_offset+DIM*3+XX] += fix3;
 670         f[i_coord_offset+DIM*3+YY] += fiy3;
 671         f[i_coord_offset+DIM*3+ZZ] += fiz3;
 672         tx                         += fix3;
 673         ty                         += fiy3;
 674         tz                         += fiz3;
 675         fshift[i_shift_offset+XX]  += tx;
 676         fshift[i_shift_offset+YY]  += ty;
 677         fshift[i_shift_offset+ZZ]  += tz;
 678
 679         /* Increment number of inner iterations */
 680         inneriter                  += j_index_end - j_index_start;
 681
 682         /* Outer loop uses 39 flops */
 683     }
 684
 685     /* Increment number of outer iterations */
 686     outeriter        += nri;
 687
 688     /* Update outer/inner flops */
 689
 690     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_F,outeriter*39 + inneriter*108);
 691 }