src/gromacs/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecRFCut_VdwBhamSh_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_ElecRFCut_VdwBhamSh_GeomW4P1_VF_c
  51  * Electrostatics interaction: ReactionField
  52  * VdW interaction:            Buckingham
  53  * Geometry:                   Water4-Particle
  54  * Calculate force/pot:        PotentialAndForce
  55  */
  56 void
  57 nb_kernel_ElecRFCut_VdwBhamSh_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      = 3*nvdwtype*vdwtype[inr+0];
 119
 120     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 121     rcutoff          = fr->rcoulomb;
 122     rcutoff2         = rcutoff*rcutoff;
 123
 124     sh_vdw_invrcut6  = fr->ic->sh_invrc6;
 125     rvdw             = fr->rvdw;
 126
 127     outeriter        = 0;
 128     inneriter        = 0;
 129
 130     /* Start outer loop over neighborlists */
 131     for(iidx=0; iidx<nri; iidx++)
 132     {
 133         /* Load shift vector for this list */
 134         i_shift_offset   = DIM*shiftidx[iidx];
 135         shX              = shiftvec[i_shift_offset+XX];
 136         shY              = shiftvec[i_shift_offset+YY];
 137         shZ              = shiftvec[i_shift_offset+ZZ];
 138
 139         /* Load limits for loop over neighbors */
 140         j_index_start    = jindex[iidx];
 141         j_index_end      = jindex[iidx+1];
 142
 143         /* Get outer coordinate index */
 144         inr              = iinr[iidx];
 145         i_coord_offset   = DIM*inr;
 146
 147         /* Load i particle coords and add shift vector */
 148         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 149         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 150         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 151         ix1              = shX + x[i_coord_offset+DIM*1+XX];
 152         iy1              = shY + x[i_coord_offset+DIM*1+YY];
 153         iz1              = shZ + x[i_coord_offset+DIM*1+ZZ];
 154         ix2              = shX + x[i_coord_offset+DIM*2+XX];
 155         iy2              = shY + x[i_coord_offset+DIM*2+YY];
 156         iz2              = shZ + x[i_coord_offset+DIM*2+ZZ];
 157         ix3              = shX + x[i_coord_offset+DIM*3+XX];
 158         iy3              = shY + x[i_coord_offset+DIM*3+YY];
 159         iz3              = shZ + x[i_coord_offset+DIM*3+ZZ];
 160
 161         fix0             = 0.0;
 162         fiy0             = 0.0;
 163         fiz0             = 0.0;
 164         fix1             = 0.0;
 165         fiy1             = 0.0;
 166         fiz1             = 0.0;
 167         fix2             = 0.0;
 168         fiy2             = 0.0;
 169         fiz2             = 0.0;
 170         fix3             = 0.0;
 171         fiy3             = 0.0;
 172         fiz3             = 0.0;
 173
 174         /* Reset potential sums */
 175         velecsum         = 0.0;
 176         vvdwsum          = 0.0;
 177
 178         /* Start inner kernel loop */
 179         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 180         {
 181             /* Get j neighbor index, and coordinate index */
 182             jnr              = jjnr[jidx];
 183             j_coord_offset   = DIM*jnr;
 184
 185             /* load j atom coordinates */
 186             jx0              = x[j_coord_offset+DIM*0+XX];
 187             jy0              = x[j_coord_offset+DIM*0+YY];
 188             jz0              = x[j_coord_offset+DIM*0+ZZ];
 189
 190             /* Calculate displacement vector */
 191             dx00             = ix0 - jx0;
 192             dy00             = iy0 - jy0;
 193             dz00             = iz0 - jz0;
 194             dx10             = ix1 - jx0;
 195             dy10             = iy1 - jy0;
 196             dz10             = iz1 - jz0;
 197             dx20             = ix2 - jx0;
 198             dy20             = iy2 - jy0;
 199             dz20             = iz2 - jz0;
 200             dx30             = ix3 - jx0;
 201             dy30             = iy3 - jy0;
 202             dz30             = iz3 - jz0;
 203
 204             /* Calculate squared distance and things based on it */
 205             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 206             rsq10            = dx10*dx10+dy10*dy10+dz10*dz10;
 207             rsq20            = dx20*dx20+dy20*dy20+dz20*dz20;
 208             rsq30            = dx30*dx30+dy30*dy30+dz30*dz30;
 209
 210             rinv00           = gmx_invsqrt(rsq00);
 211             rinv10           = gmx_invsqrt(rsq10);
 212             rinv20           = gmx_invsqrt(rsq20);
 213             rinv30           = gmx_invsqrt(rsq30);
 214
 215             rinvsq00         = rinv00*rinv00;
 216             rinvsq10         = rinv10*rinv10;
 217             rinvsq20         = rinv20*rinv20;
 218             rinvsq30         = rinv30*rinv30;
 219
 220             /* Load parameters for j particles */
 221             jq0              = charge[jnr+0];
 222             vdwjidx0         = 3*vdwtype[jnr+0];
 223
 224             /**************************
 225              * CALCULATE INTERACTIONS *
 226              **************************/
 227
 228             if (rsq00<rcutoff2)
 229             {
 230
 231             r00              = rsq00*rinv00;
 232
 233             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 234             cexp1_00         = vdwparam[vdwioffset0+vdwjidx0+1];
 235             cexp2_00         = vdwparam[vdwioffset0+vdwjidx0+2];
 236
 237             /* BUCKINGHAM DISPERSION/REPULSION */
 238             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 239             vvdw6            = c6_00*rinvsix;
 240             br               = cexp2_00*r00;
 241             vvdwexp          = cexp1_00*exp(-br);
 242             vvdw             = (vvdwexp-cexp1_00*exp(-cexp2_00*rvdw)) - (vvdw6 - c6_00*sh_vdw_invrcut6)*(1.0/6.0);
 243             fvdw             = (br*vvdwexp-vvdw6)*rinvsq00;
 244
 245             /* Update potential sums from outer loop */
 246             vvdwsum         += vvdw;
 247
 248             fscal            = fvdw;
 249
 250             /* Calculate temporary vectorial force */
 251             tx               = fscal*dx00;
 252             ty               = fscal*dy00;
 253             tz               = fscal*dz00;
 254
 255             /* Update vectorial force */
 256             fix0            += tx;
 257             fiy0            += ty;
 258             fiz0            += tz;
 259             f[j_coord_offset+DIM*0+XX] -= tx;
 260             f[j_coord_offset+DIM*0+YY] -= ty;
 261             f[j_coord_offset+DIM*0+ZZ] -= tz;
 262
 263             }
 264
 265             /**************************
 266              * CALCULATE INTERACTIONS *
 267              **************************/
 268
 269             if (rsq10<rcutoff2)
 270             {
 271
 272             qq10             = iq1*jq0;
 273
 274             /* REACTION-FIELD ELECTROSTATICS */
 275             velec            = qq10*(rinv10+krf*rsq10-crf);
 276             felec            = qq10*(rinv10*rinvsq10-krf2);
 277
 278             /* Update potential sums from outer loop */
 279             velecsum        += velec;
 280
 281             fscal            = felec;
 282
 283             /* Calculate temporary vectorial force */
 284             tx               = fscal*dx10;
 285             ty               = fscal*dy10;
 286             tz               = fscal*dz10;
 287
 288             /* Update vectorial force */
 289             fix1            += tx;
 290             fiy1            += ty;
 291             fiz1            += tz;
 292             f[j_coord_offset+DIM*0+XX] -= tx;
 293             f[j_coord_offset+DIM*0+YY] -= ty;
 294             f[j_coord_offset+DIM*0+ZZ] -= tz;
 295
 296             }
 297
 298             /**************************
 299              * CALCULATE INTERACTIONS *
 300              **************************/
 301
 302             if (rsq20<rcutoff2)
 303             {
 304
 305             qq20             = iq2*jq0;
 306
 307             /* REACTION-FIELD ELECTROSTATICS */
 308             velec            = qq20*(rinv20+krf*rsq20-crf);
 309             felec            = qq20*(rinv20*rinvsq20-krf2);
 310
 311             /* Update potential sums from outer loop */
 312             velecsum        += velec;
 313
 314             fscal            = felec;
 315
 316             /* Calculate temporary vectorial force */
 317             tx               = fscal*dx20;
 318             ty               = fscal*dy20;
 319             tz               = fscal*dz20;
 320
 321             /* Update vectorial force */
 322             fix2            += tx;
 323             fiy2            += ty;
 324             fiz2            += tz;
 325             f[j_coord_offset+DIM*0+XX] -= tx;
 326             f[j_coord_offset+DIM*0+YY] -= ty;
 327             f[j_coord_offset+DIM*0+ZZ] -= tz;
 328
 329             }
 330
 331             /**************************
 332              * CALCULATE INTERACTIONS *
 333              **************************/
 334
 335             if (rsq30<rcutoff2)
 336             {
 337
 338             qq30             = iq3*jq0;
 339
 340             /* REACTION-FIELD ELECTROSTATICS */
 341             velec            = qq30*(rinv30+krf*rsq30-crf);
 342             felec            = qq30*(rinv30*rinvsq30-krf2);
 343
 344             /* Update potential sums from outer loop */
 345             velecsum        += velec;
 346
 347             fscal            = felec;
 348
 349             /* Calculate temporary vectorial force */
 350             tx               = fscal*dx30;
 351             ty               = fscal*dy30;
 352             tz               = fscal*dz30;
 353
 354             /* Update vectorial force */
 355             fix3            += tx;
 356             fiy3            += ty;
 357             fiz3            += tz;
 358             f[j_coord_offset+DIM*0+XX] -= tx;
 359             f[j_coord_offset+DIM*0+YY] -= ty;
 360             f[j_coord_offset+DIM*0+ZZ] -= tz;
 361
 362             }
 363
 364             /* Inner loop uses 188 flops */
 365         }
 366         /* End of innermost loop */
 367
 368         tx = ty = tz = 0;
 369         f[i_coord_offset+DIM*0+XX] += fix0;
 370         f[i_coord_offset+DIM*0+YY] += fiy0;
 371         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 372         tx                         += fix0;
 373         ty                         += fiy0;
 374         tz                         += fiz0;
 375         f[i_coord_offset+DIM*1+XX] += fix1;
 376         f[i_coord_offset+DIM*1+YY] += fiy1;
 377         f[i_coord_offset+DIM*1+ZZ] += fiz1;
 378         tx                         += fix1;
 379         ty                         += fiy1;
 380         tz                         += fiz1;
 381         f[i_coord_offset+DIM*2+XX] += fix2;
 382         f[i_coord_offset+DIM*2+YY] += fiy2;
 383         f[i_coord_offset+DIM*2+ZZ] += fiz2;
 384         tx                         += fix2;
 385         ty                         += fiy2;
 386         tz                         += fiz2;
 387         f[i_coord_offset+DIM*3+XX] += fix3;
 388         f[i_coord_offset+DIM*3+YY] += fiy3;
 389         f[i_coord_offset+DIM*3+ZZ] += fiz3;
 390         tx                         += fix3;
 391         ty                         += fiy3;
 392         tz                         += fiz3;
 393         fshift[i_shift_offset+XX]  += tx;
 394         fshift[i_shift_offset+YY]  += ty;
 395         fshift[i_shift_offset+ZZ]  += tz;
 396
 397         ggid                        = gid[iidx];
 398         /* Update potential energies */
 399         kernel_data->energygrp_elec[ggid] += velecsum;
 400         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 401
 402         /* Increment number of inner iterations */
 403         inneriter                  += j_index_end - j_index_start;
 404
 405         /* Outer loop uses 41 flops */
 406     }
 407
 408     /* Increment number of outer iterations */
 409     outeriter        += nri;
 410
 411     /* Update outer/inner flops */
 412
 413     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_VF,outeriter*41 + inneriter*188);
 414 }
 415 /*
 416  * Gromacs nonbonded kernel:   nb_kernel_ElecRFCut_VdwBhamSh_GeomW4P1_F_c
 417  * Electrostatics interaction: ReactionField
 418  * VdW interaction:            Buckingham
 419  * Geometry:                   Water4-Particle
 420  * Calculate force/pot:        Force
 421  */
 422 void
 423 nb_kernel_ElecRFCut_VdwBhamSh_GeomW4P1_F_c
 424                     (t_nblist                    * gmx_restrict       nlist,
 425                      rvec                        * gmx_restrict          xx,
 426                      rvec                        * gmx_restrict          ff,
 427                      t_forcerec                  * gmx_restrict          fr,
 428                      t_mdatoms                   * gmx_restrict     mdatoms,
 429                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 430                      t_nrnb                      * gmx_restrict        nrnb)
 431 {
 432     int              i_shift_offset,i_coord_offset,j_coord_offset;
 433     int              j_index_start,j_index_end;
 434     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 435     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 436     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 437     real             *shiftvec,*fshift,*x,*f;
 438     int              vdwioffset0;
 439     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 440     int              vdwioffset1;
 441     real             ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
 442     int              vdwioffset2;
 443     real             ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
 444     int              vdwioffset3;
 445     real             ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
 446     int              vdwjidx0;
 447     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 448     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 449     real             dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10;
 450     real             dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20;
 451     real             dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30,cexp1_30,cexp2_30;
 452     real             velec,felec,velecsum,facel,crf,krf,krf2;
 453     real             *charge;
 454     int              nvdwtype;
 455     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 456     int              *vdwtype;
 457     real             *vdwparam;
 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     krf              = fr->ic->k_rf;
 473     krf2             = krf*2.0;
 474     crf              = fr->ic->c_rf;
 475     nvdwtype         = fr->ntype;
 476     vdwparam         = fr->nbfp;
 477     vdwtype          = mdatoms->typeA;
 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      = 3*nvdwtype*vdwtype[inr+0];
 485
 486     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 487     rcutoff          = fr->rcoulomb;
 488     rcutoff2         = rcutoff*rcutoff;
 489
 490     sh_vdw_invrcut6  = fr->ic->sh_invrc6;
 491     rvdw             = fr->rvdw;
 492
 493     outeriter        = 0;
 494     inneriter        = 0;
 495
 496     /* Start outer loop over neighborlists */
 497     for(iidx=0; iidx<nri; iidx++)
 498     {
 499         /* Load shift vector for this list */
 500         i_shift_offset   = DIM*shiftidx[iidx];
 501         shX              = shiftvec[i_shift_offset+XX];
 502         shY              = shiftvec[i_shift_offset+YY];
 503         shZ              = shiftvec[i_shift_offset+ZZ];
 504
 505         /* Load limits for loop over neighbors */
 506         j_index_start    = jindex[iidx];
 507         j_index_end      = jindex[iidx+1];
 508
 509         /* Get outer coordinate index */
 510         inr              = iinr[iidx];
 511         i_coord_offset   = DIM*inr;
 512
 513         /* Load i particle coords and add shift vector */
 514         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 515         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 516         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 517         ix1              = shX + x[i_coord_offset+DIM*1+XX];
 518         iy1              = shY + x[i_coord_offset+DIM*1+YY];
 519         iz1              = shZ + x[i_coord_offset+DIM*1+ZZ];
 520         ix2              = shX + x[i_coord_offset+DIM*2+XX];
 521         iy2              = shY + x[i_coord_offset+DIM*2+YY];
 522         iz2              = shZ + x[i_coord_offset+DIM*2+ZZ];
 523         ix3              = shX + x[i_coord_offset+DIM*3+XX];
 524         iy3              = shY + x[i_coord_offset+DIM*3+YY];
 525         iz3              = shZ + x[i_coord_offset+DIM*3+ZZ];
 526
 527         fix0             = 0.0;
 528         fiy0             = 0.0;
 529         fiz0             = 0.0;
 530         fix1             = 0.0;
 531         fiy1             = 0.0;
 532         fiz1             = 0.0;
 533         fix2             = 0.0;
 534         fiy2             = 0.0;
 535         fiz2             = 0.0;
 536         fix3             = 0.0;
 537         fiy3             = 0.0;
 538         fiz3             = 0.0;
 539
 540         /* Start inner kernel loop */
 541         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 542         {
 543             /* Get j neighbor index, and coordinate index */
 544             jnr              = jjnr[jidx];
 545             j_coord_offset   = DIM*jnr;
 546
 547             /* load j atom coordinates */
 548             jx0              = x[j_coord_offset+DIM*0+XX];
 549             jy0              = x[j_coord_offset+DIM*0+YY];
 550             jz0              = x[j_coord_offset+DIM*0+ZZ];
 551
 552             /* Calculate displacement vector */
 553             dx00             = ix0 - jx0;
 554             dy00             = iy0 - jy0;
 555             dz00             = iz0 - jz0;
 556             dx10             = ix1 - jx0;
 557             dy10             = iy1 - jy0;
 558             dz10             = iz1 - jz0;
 559             dx20             = ix2 - jx0;
 560             dy20             = iy2 - jy0;
 561             dz20             = iz2 - jz0;
 562             dx30             = ix3 - jx0;
 563             dy30             = iy3 - jy0;
 564             dz30             = iz3 - jz0;
 565
 566             /* Calculate squared distance and things based on it */
 567             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 568             rsq10            = dx10*dx10+dy10*dy10+dz10*dz10;
 569             rsq20            = dx20*dx20+dy20*dy20+dz20*dz20;
 570             rsq30            = dx30*dx30+dy30*dy30+dz30*dz30;
 571
 572             rinv00           = gmx_invsqrt(rsq00);
 573             rinv10           = gmx_invsqrt(rsq10);
 574             rinv20           = gmx_invsqrt(rsq20);
 575             rinv30           = gmx_invsqrt(rsq30);
 576
 577             rinvsq00         = rinv00*rinv00;
 578             rinvsq10         = rinv10*rinv10;
 579             rinvsq20         = rinv20*rinv20;
 580             rinvsq30         = rinv30*rinv30;
 581
 582             /* Load parameters for j particles */
 583             jq0              = charge[jnr+0];
 584             vdwjidx0         = 3*vdwtype[jnr+0];
 585
 586             /**************************
 587              * CALCULATE INTERACTIONS *
 588              **************************/
 589
 590             if (rsq00<rcutoff2)
 591             {
 592
 593             r00              = rsq00*rinv00;
 594
 595             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 596             cexp1_00         = vdwparam[vdwioffset0+vdwjidx0+1];
 597             cexp2_00         = vdwparam[vdwioffset0+vdwjidx0+2];
 598
 599             /* BUCKINGHAM DISPERSION/REPULSION */
 600             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 601             vvdw6            = c6_00*rinvsix;
 602             br               = cexp2_00*r00;
 603             vvdwexp          = cexp1_00*exp(-br);
 604             fvdw             = (br*vvdwexp-vvdw6)*rinvsq00;
 605
 606             fscal            = fvdw;
 607
 608             /* Calculate temporary vectorial force */
 609             tx               = fscal*dx00;
 610             ty               = fscal*dy00;
 611             tz               = fscal*dz00;
 612
 613             /* Update vectorial force */
 614             fix0            += tx;
 615             fiy0            += ty;
 616             fiz0            += tz;
 617             f[j_coord_offset+DIM*0+XX] -= tx;
 618             f[j_coord_offset+DIM*0+YY] -= ty;
 619             f[j_coord_offset+DIM*0+ZZ] -= tz;
 620
 621             }
 622
 623             /**************************
 624              * CALCULATE INTERACTIONS *
 625              **************************/
 626
 627             if (rsq10<rcutoff2)
 628             {
 629
 630             qq10             = iq1*jq0;
 631
 632             /* REACTION-FIELD ELECTROSTATICS */
 633             felec            = qq10*(rinv10*rinvsq10-krf2);
 634
 635             fscal            = felec;
 636
 637             /* Calculate temporary vectorial force */
 638             tx               = fscal*dx10;
 639             ty               = fscal*dy10;
 640             tz               = fscal*dz10;
 641
 642             /* Update vectorial force */
 643             fix1            += tx;
 644             fiy1            += ty;
 645             fiz1            += tz;
 646             f[j_coord_offset+DIM*0+XX] -= tx;
 647             f[j_coord_offset+DIM*0+YY] -= ty;
 648             f[j_coord_offset+DIM*0+ZZ] -= tz;
 649
 650             }
 651
 652             /**************************
 653              * CALCULATE INTERACTIONS *
 654              **************************/
 655
 656             if (rsq20<rcutoff2)
 657             {
 658
 659             qq20             = iq2*jq0;
 660
 661             /* REACTION-FIELD ELECTROSTATICS */
 662             felec            = qq20*(rinv20*rinvsq20-krf2);
 663
 664             fscal            = felec;
 665
 666             /* Calculate temporary vectorial force */
 667             tx               = fscal*dx20;
 668             ty               = fscal*dy20;
 669             tz               = fscal*dz20;
 670
 671             /* Update vectorial force */
 672             fix2            += tx;
 673             fiy2            += ty;
 674             fiz2            += tz;
 675             f[j_coord_offset+DIM*0+XX] -= tx;
 676             f[j_coord_offset+DIM*0+YY] -= ty;
 677             f[j_coord_offset+DIM*0+ZZ] -= tz;
 678
 679             }
 680
 681             /**************************
 682              * CALCULATE INTERACTIONS *
 683              **************************/
 684
 685             if (rsq30<rcutoff2)
 686             {
 687
 688             qq30             = iq3*jq0;
 689
 690             /* REACTION-FIELD ELECTROSTATICS */
 691             felec            = qq30*(rinv30*rinvsq30-krf2);
 692
 693             fscal            = felec;
 694
 695             /* Calculate temporary vectorial force */
 696             tx               = fscal*dx30;
 697             ty               = fscal*dy30;
 698             tz               = fscal*dz30;
 699
 700             /* Update vectorial force */
 701             fix3            += tx;
 702             fiy3            += ty;
 703             fiz3            += tz;
 704             f[j_coord_offset+DIM*0+XX] -= tx;
 705             f[j_coord_offset+DIM*0+YY] -= ty;
 706             f[j_coord_offset+DIM*0+ZZ] -= tz;
 707
 708             }
 709
 710             /* Inner loop uses 139 flops */
 711         }
 712         /* End of innermost loop */
 713
 714         tx = ty = tz = 0;
 715         f[i_coord_offset+DIM*0+XX] += fix0;
 716         f[i_coord_offset+DIM*0+YY] += fiy0;
 717         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 718         tx                         += fix0;
 719         ty                         += fiy0;
 720         tz                         += fiz0;
 721         f[i_coord_offset+DIM*1+XX] += fix1;
 722         f[i_coord_offset+DIM*1+YY] += fiy1;
 723         f[i_coord_offset+DIM*1+ZZ] += fiz1;
 724         tx                         += fix1;
 725         ty                         += fiy1;
 726         tz                         += fiz1;
 727         f[i_coord_offset+DIM*2+XX] += fix2;
 728         f[i_coord_offset+DIM*2+YY] += fiy2;
 729         f[i_coord_offset+DIM*2+ZZ] += fiz2;
 730         tx                         += fix2;
 731         ty                         += fiy2;
 732         tz                         += fiz2;
 733         f[i_coord_offset+DIM*3+XX] += fix3;
 734         f[i_coord_offset+DIM*3+YY] += fiy3;
 735         f[i_coord_offset+DIM*3+ZZ] += fiz3;
 736         tx                         += fix3;
 737         ty                         += fiy3;
 738         tz                         += fiz3;
 739         fshift[i_shift_offset+XX]  += tx;
 740         fshift[i_shift_offset+YY]  += ty;
 741         fshift[i_shift_offset+ZZ]  += tz;
 742
 743         /* Increment number of inner iterations */
 744         inneriter                  += j_index_end - j_index_start;
 745
 746         /* Outer loop uses 39 flops */
 747     }
 748
 749     /* Increment number of outer iterations */
 750     outeriter        += nri;
 751
 752     /* Update outer/inner flops */
 753
 754     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_F,outeriter*39 + inneriter*139);
 755 }