src/gromacs/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecRFCut_VdwBhamSw_GeomW3P1_c.c

   1 /*
   2  * This file is part of the GROMACS molecular simulation package.
   3  *
   4  * Copyright (c) 2012,2013,2014, by the GROMACS development team, led by
   5  * Mark Abraham, David van der Spoel, Berk Hess, and Erik Lindahl,
   6  * and including many others, as listed in the AUTHORS file in the
   7  * top-level source directory and at http://www.gromacs.org.
   8  *
   9  * GROMACS is free software; you can redistribute it and/or
  10  * modify it under the terms of the GNU Lesser General Public License
  11  * as published by the Free Software Foundation; either version 2.1
  12  * of the License, or (at your option) any later version.
  13  *
  14  * GROMACS is distributed in the hope that it will be useful,
  15  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  16  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  17  * Lesser General Public License for more details.
  18  *
  19  * You should have received a copy of the GNU Lesser General Public
  20  * License along with GROMACS; if not, see
  21  * http://www.gnu.org/licenses, or write to the Free Software Foundation,
  22  * Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA.
  23  *
  24  * If you want to redistribute modifications to GROMACS, please
  25  * consider that scientific software is very special. Version
  26  * control is crucial - bugs must be traceable. We will be happy to
  27  * consider code for inclusion in the official distribution, but
  28  * derived work must not be called official GROMACS. Details are found
  29  * in the README & COPYING files - if they are missing, get the
  30  * official version at http://www.gromacs.org.
  31  *
  32  * To help us fund GROMACS development, we humbly ask that you cite
  33  * the research papers on the package. Check out http://www.gromacs.org.
  34  */
  35 /*
  36  * Note: this file was generated by the GROMACS c kernel generator.
  37  */
  38 #include "config.h"
  39
  40 #include <math.h>
  41
  42 #include "../nb_kernel.h"
  43 #include "gromacs/legacyheaders/types/simple.h"
  44 #include "gromacs/math/vec.h"
  45 #include "gromacs/legacyheaders/nrnb.h"
  46
  47 /*
  48  * Gromacs nonbonded kernel:   nb_kernel_ElecRFCut_VdwBhamSw_GeomW3P1_VF_c
  49  * Electrostatics interaction: ReactionField
  50  * VdW interaction:            Buckingham
  51  * Geometry:                   Water3-Particle
  52  * Calculate force/pot:        PotentialAndForce
  53  */
  54 void
  55 nb_kernel_ElecRFCut_VdwBhamSw_GeomW3P1_VF_c
  56                     (t_nblist                    * gmx_restrict       nlist,
  57                      rvec                        * gmx_restrict          xx,
  58                      rvec                        * gmx_restrict          ff,
  59                      t_forcerec                  * gmx_restrict          fr,
  60                      t_mdatoms                   * gmx_restrict     mdatoms,
  61                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
  62                      t_nrnb                      * gmx_restrict        nrnb)
  63 {
  64     int              i_shift_offset,i_coord_offset,j_coord_offset;
  65     int              j_index_start,j_index_end;
  66     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
  67     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
  68     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  69     real             *shiftvec,*fshift,*x,*f;
  70     int              vdwioffset0;
  71     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  72     int              vdwioffset1;
  73     real             ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
  74     int              vdwioffset2;
  75     real             ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
  76     int              vdwjidx0;
  77     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  78     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
  79     real             dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10;
  80     real             dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20;
  81     real             velec,felec,velecsum,facel,crf,krf,krf2;
  82     real             *charge;
  83     int              nvdwtype;
  84     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
  85     int              *vdwtype;
  86     real             *vdwparam;
  87     real             rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
  88
  89     x                = xx[0];
  90     f                = ff[0];
  91
  92     nri              = nlist->nri;
  93     iinr             = nlist->iinr;
  94     jindex           = nlist->jindex;
  95     jjnr             = nlist->jjnr;
  96     shiftidx         = nlist->shift;
  97     gid              = nlist->gid;
  98     shiftvec         = fr->shift_vec[0];
  99     fshift           = fr->fshift[0];
 100     facel            = fr->epsfac;
 101     charge           = mdatoms->chargeA;
 102     krf              = fr->ic->k_rf;
 103     krf2             = krf*2.0;
 104     crf              = fr->ic->c_rf;
 105     nvdwtype         = fr->ntype;
 106     vdwparam         = fr->nbfp;
 107     vdwtype          = mdatoms->typeA;
 108
 109     /* Setup water-specific parameters */
 110     inr              = nlist->iinr[0];
 111     iq0              = facel*charge[inr+0];
 112     iq1              = facel*charge[inr+1];
 113     iq2              = facel*charge[inr+2];
 114     vdwioffset0      = 3*nvdwtype*vdwtype[inr+0];
 115
 116     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 117     rcutoff          = fr->rcoulomb;
 118     rcutoff2         = rcutoff*rcutoff;
 119
 120     rswitch          = fr->rvdw_switch;
 121     /* Setup switch parameters */
 122     d                = rcutoff-rswitch;
 123     swV3             = -10.0/(d*d*d);
 124     swV4             =  15.0/(d*d*d*d);
 125     swV5             =  -6.0/(d*d*d*d*d);
 126     swF2             = -30.0/(d*d*d);
 127     swF3             =  60.0/(d*d*d*d);
 128     swF4             = -30.0/(d*d*d*d*d);
 129
 130     outeriter        = 0;
 131     inneriter        = 0;
 132
 133     /* Start outer loop over neighborlists */
 134     for(iidx=0; iidx<nri; iidx++)
 135     {
 136         /* Load shift vector for this list */
 137         i_shift_offset   = DIM*shiftidx[iidx];
 138         shX              = shiftvec[i_shift_offset+XX];
 139         shY              = shiftvec[i_shift_offset+YY];
 140         shZ              = shiftvec[i_shift_offset+ZZ];
 141
 142         /* Load limits for loop over neighbors */
 143         j_index_start    = jindex[iidx];
 144         j_index_end      = jindex[iidx+1];
 145
 146         /* Get outer coordinate index */
 147         inr              = iinr[iidx];
 148         i_coord_offset   = DIM*inr;
 149
 150         /* Load i particle coords and add shift vector */
 151         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 152         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 153         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 154         ix1              = shX + x[i_coord_offset+DIM*1+XX];
 155         iy1              = shY + x[i_coord_offset+DIM*1+YY];
 156         iz1              = shZ + x[i_coord_offset+DIM*1+ZZ];
 157         ix2              = shX + x[i_coord_offset+DIM*2+XX];
 158         iy2              = shY + x[i_coord_offset+DIM*2+YY];
 159         iz2              = shZ + x[i_coord_offset+DIM*2+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
 171         /* Reset potential sums */
 172         velecsum         = 0.0;
 173         vvdwsum          = 0.0;
 174
 175         /* Start inner kernel loop */
 176         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 177         {
 178             /* Get j neighbor index, and coordinate index */
 179             jnr              = jjnr[jidx];
 180             j_coord_offset   = DIM*jnr;
 181
 182             /* load j atom coordinates */
 183             jx0              = x[j_coord_offset+DIM*0+XX];
 184             jy0              = x[j_coord_offset+DIM*0+YY];
 185             jz0              = x[j_coord_offset+DIM*0+ZZ];
 186
 187             /* Calculate displacement vector */
 188             dx00             = ix0 - jx0;
 189             dy00             = iy0 - jy0;
 190             dz00             = iz0 - jz0;
 191             dx10             = ix1 - jx0;
 192             dy10             = iy1 - jy0;
 193             dz10             = iz1 - jz0;
 194             dx20             = ix2 - jx0;
 195             dy20             = iy2 - jy0;
 196             dz20             = iz2 - jz0;
 197
 198             /* Calculate squared distance and things based on it */
 199             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 200             rsq10            = dx10*dx10+dy10*dy10+dz10*dz10;
 201             rsq20            = dx20*dx20+dy20*dy20+dz20*dz20;
 202
 203             rinv00           = gmx_invsqrt(rsq00);
 204             rinv10           = gmx_invsqrt(rsq10);
 205             rinv20           = gmx_invsqrt(rsq20);
 206
 207             rinvsq00         = rinv00*rinv00;
 208             rinvsq10         = rinv10*rinv10;
 209             rinvsq20         = rinv20*rinv20;
 210
 211             /* Load parameters for j particles */
 212             jq0              = charge[jnr+0];
 213             vdwjidx0         = 3*vdwtype[jnr+0];
 214
 215             /**************************
 216              * CALCULATE INTERACTIONS *
 217              **************************/
 218
 219             if (rsq00<rcutoff2)
 220             {
 221
 222             r00              = rsq00*rinv00;
 223
 224             qq00             = iq0*jq0;
 225             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 226             cexp1_00         = vdwparam[vdwioffset0+vdwjidx0+1];
 227             cexp2_00         = vdwparam[vdwioffset0+vdwjidx0+2];
 228
 229             /* REACTION-FIELD ELECTROSTATICS */
 230             velec            = qq00*(rinv00+krf*rsq00-crf);
 231             felec            = qq00*(rinv00*rinvsq00-krf2);
 232
 233             /* BUCKINGHAM DISPERSION/REPULSION */
 234             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 235             vvdw6            = c6_00*rinvsix;
 236             br               = cexp2_00*r00;
 237             vvdwexp          = cexp1_00*exp(-br);
 238             vvdw             = vvdwexp - vvdw6*(1.0/6.0);
 239             fvdw             = (br*vvdwexp-vvdw6)*rinvsq00;
 240
 241             d                = r00-rswitch;
 242             d                = (d>0.0) ? d : 0.0;
 243             d2               = d*d;
 244             sw               = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
 245
 246             dsw              = d2*(swF2+d*(swF3+d*swF4));
 247
 248             /* Evaluate switch function */
 249             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 250             fvdw             = fvdw*sw - rinv00*vvdw*dsw;
 251             vvdw            *= sw;
 252
 253             /* Update potential sums from outer loop */
 254             velecsum        += velec;
 255             vvdwsum         += vvdw;
 256
 257             fscal            = felec+fvdw;
 258
 259             /* Calculate temporary vectorial force */
 260             tx               = fscal*dx00;
 261             ty               = fscal*dy00;
 262             tz               = fscal*dz00;
 263
 264             /* Update vectorial force */
 265             fix0            += tx;
 266             fiy0            += ty;
 267             fiz0            += tz;
 268             f[j_coord_offset+DIM*0+XX] -= tx;
 269             f[j_coord_offset+DIM*0+YY] -= ty;
 270             f[j_coord_offset+DIM*0+ZZ] -= tz;
 271
 272             }
 273
 274             /**************************
 275              * CALCULATE INTERACTIONS *
 276              **************************/
 277
 278             if (rsq10<rcutoff2)
 279             {
 280
 281             qq10             = iq1*jq0;
 282
 283             /* REACTION-FIELD ELECTROSTATICS */
 284             velec            = qq10*(rinv10+krf*rsq10-crf);
 285             felec            = qq10*(rinv10*rinvsq10-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*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
 307             /**************************
 308              * CALCULATE INTERACTIONS *
 309              **************************/
 310
 311             if (rsq20<rcutoff2)
 312             {
 313
 314             qq20             = iq2*jq0;
 315
 316             /* REACTION-FIELD ELECTROSTATICS */
 317             velec            = qq20*(rinv20+krf*rsq20-crf);
 318             felec            = qq20*(rinv20*rinvsq20-krf2);
 319
 320             /* Update potential sums from outer loop */
 321             velecsum        += velec;
 322
 323             fscal            = felec;
 324
 325             /* Calculate temporary vectorial force */
 326             tx               = fscal*dx20;
 327             ty               = fscal*dy20;
 328             tz               = fscal*dz20;
 329
 330             /* Update vectorial force */
 331             fix2            += tx;
 332             fiy2            += ty;
 333             fiz2            += tz;
 334             f[j_coord_offset+DIM*0+XX] -= tx;
 335             f[j_coord_offset+DIM*0+YY] -= ty;
 336             f[j_coord_offset+DIM*0+ZZ] -= tz;
 337
 338             }
 339
 340             /* Inner loop uses 153 flops */
 341         }
 342         /* End of innermost loop */
 343
 344         tx = ty = tz = 0;
 345         f[i_coord_offset+DIM*0+XX] += fix0;
 346         f[i_coord_offset+DIM*0+YY] += fiy0;
 347         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 348         tx                         += fix0;
 349         ty                         += fiy0;
 350         tz                         += fiz0;
 351         f[i_coord_offset+DIM*1+XX] += fix1;
 352         f[i_coord_offset+DIM*1+YY] += fiy1;
 353         f[i_coord_offset+DIM*1+ZZ] += fiz1;
 354         tx                         += fix1;
 355         ty                         += fiy1;
 356         tz                         += fiz1;
 357         f[i_coord_offset+DIM*2+XX] += fix2;
 358         f[i_coord_offset+DIM*2+YY] += fiy2;
 359         f[i_coord_offset+DIM*2+ZZ] += fiz2;
 360         tx                         += fix2;
 361         ty                         += fiy2;
 362         tz                         += fiz2;
 363         fshift[i_shift_offset+XX]  += tx;
 364         fshift[i_shift_offset+YY]  += ty;
 365         fshift[i_shift_offset+ZZ]  += tz;
 366
 367         ggid                        = gid[iidx];
 368         /* Update potential energies */
 369         kernel_data->energygrp_elec[ggid] += velecsum;
 370         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 371
 372         /* Increment number of inner iterations */
 373         inneriter                  += j_index_end - j_index_start;
 374
 375         /* Outer loop uses 32 flops */
 376     }
 377
 378     /* Increment number of outer iterations */
 379     outeriter        += nri;
 380
 381     /* Update outer/inner flops */
 382
 383     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3_VF,outeriter*32 + inneriter*153);
 384 }
 385 /*
 386  * Gromacs nonbonded kernel:   nb_kernel_ElecRFCut_VdwBhamSw_GeomW3P1_F_c
 387  * Electrostatics interaction: ReactionField
 388  * VdW interaction:            Buckingham
 389  * Geometry:                   Water3-Particle
 390  * Calculate force/pot:        Force
 391  */
 392 void
 393 nb_kernel_ElecRFCut_VdwBhamSw_GeomW3P1_F_c
 394                     (t_nblist                    * gmx_restrict       nlist,
 395                      rvec                        * gmx_restrict          xx,
 396                      rvec                        * gmx_restrict          ff,
 397                      t_forcerec                  * gmx_restrict          fr,
 398                      t_mdatoms                   * gmx_restrict     mdatoms,
 399                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 400                      t_nrnb                      * gmx_restrict        nrnb)
 401 {
 402     int              i_shift_offset,i_coord_offset,j_coord_offset;
 403     int              j_index_start,j_index_end;
 404     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 405     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 406     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 407     real             *shiftvec,*fshift,*x,*f;
 408     int              vdwioffset0;
 409     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 410     int              vdwioffset1;
 411     real             ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
 412     int              vdwioffset2;
 413     real             ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
 414     int              vdwjidx0;
 415     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 416     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 417     real             dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10;
 418     real             dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20;
 419     real             velec,felec,velecsum,facel,crf,krf,krf2;
 420     real             *charge;
 421     int              nvdwtype;
 422     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 423     int              *vdwtype;
 424     real             *vdwparam;
 425     real             rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
 426
 427     x                = xx[0];
 428     f                = ff[0];
 429
 430     nri              = nlist->nri;
 431     iinr             = nlist->iinr;
 432     jindex           = nlist->jindex;
 433     jjnr             = nlist->jjnr;
 434     shiftidx         = nlist->shift;
 435     gid              = nlist->gid;
 436     shiftvec         = fr->shift_vec[0];
 437     fshift           = fr->fshift[0];
 438     facel            = fr->epsfac;
 439     charge           = mdatoms->chargeA;
 440     krf              = fr->ic->k_rf;
 441     krf2             = krf*2.0;
 442     crf              = fr->ic->c_rf;
 443     nvdwtype         = fr->ntype;
 444     vdwparam         = fr->nbfp;
 445     vdwtype          = mdatoms->typeA;
 446
 447     /* Setup water-specific parameters */
 448     inr              = nlist->iinr[0];
 449     iq0              = facel*charge[inr+0];
 450     iq1              = facel*charge[inr+1];
 451     iq2              = facel*charge[inr+2];
 452     vdwioffset0      = 3*nvdwtype*vdwtype[inr+0];
 453
 454     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 455     rcutoff          = fr->rcoulomb;
 456     rcutoff2         = rcutoff*rcutoff;
 457
 458     rswitch          = fr->rvdw_switch;
 459     /* Setup switch parameters */
 460     d                = rcutoff-rswitch;
 461     swV3             = -10.0/(d*d*d);
 462     swV4             =  15.0/(d*d*d*d);
 463     swV5             =  -6.0/(d*d*d*d*d);
 464     swF2             = -30.0/(d*d*d);
 465     swF3             =  60.0/(d*d*d*d);
 466     swF4             = -30.0/(d*d*d*d*d);
 467
 468     outeriter        = 0;
 469     inneriter        = 0;
 470
 471     /* Start outer loop over neighborlists */
 472     for(iidx=0; iidx<nri; iidx++)
 473     {
 474         /* Load shift vector for this list */
 475         i_shift_offset   = DIM*shiftidx[iidx];
 476         shX              = shiftvec[i_shift_offset+XX];
 477         shY              = shiftvec[i_shift_offset+YY];
 478         shZ              = shiftvec[i_shift_offset+ZZ];
 479
 480         /* Load limits for loop over neighbors */
 481         j_index_start    = jindex[iidx];
 482         j_index_end      = jindex[iidx+1];
 483
 484         /* Get outer coordinate index */
 485         inr              = iinr[iidx];
 486         i_coord_offset   = DIM*inr;
 487
 488         /* Load i particle coords and add shift vector */
 489         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 490         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 491         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 492         ix1              = shX + x[i_coord_offset+DIM*1+XX];
 493         iy1              = shY + x[i_coord_offset+DIM*1+YY];
 494         iz1              = shZ + x[i_coord_offset+DIM*1+ZZ];
 495         ix2              = shX + x[i_coord_offset+DIM*2+XX];
 496         iy2              = shY + x[i_coord_offset+DIM*2+YY];
 497         iz2              = shZ + x[i_coord_offset+DIM*2+ZZ];
 498
 499         fix0             = 0.0;
 500         fiy0             = 0.0;
 501         fiz0             = 0.0;
 502         fix1             = 0.0;
 503         fiy1             = 0.0;
 504         fiz1             = 0.0;
 505         fix2             = 0.0;
 506         fiy2             = 0.0;
 507         fiz2             = 0.0;
 508
 509         /* Start inner kernel loop */
 510         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 511         {
 512             /* Get j neighbor index, and coordinate index */
 513             jnr              = jjnr[jidx];
 514             j_coord_offset   = DIM*jnr;
 515
 516             /* load j atom coordinates */
 517             jx0              = x[j_coord_offset+DIM*0+XX];
 518             jy0              = x[j_coord_offset+DIM*0+YY];
 519             jz0              = x[j_coord_offset+DIM*0+ZZ];
 520
 521             /* Calculate displacement vector */
 522             dx00             = ix0 - jx0;
 523             dy00             = iy0 - jy0;
 524             dz00             = iz0 - jz0;
 525             dx10             = ix1 - jx0;
 526             dy10             = iy1 - jy0;
 527             dz10             = iz1 - jz0;
 528             dx20             = ix2 - jx0;
 529             dy20             = iy2 - jy0;
 530             dz20             = iz2 - jz0;
 531
 532             /* Calculate squared distance and things based on it */
 533             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 534             rsq10            = dx10*dx10+dy10*dy10+dz10*dz10;
 535             rsq20            = dx20*dx20+dy20*dy20+dz20*dz20;
 536
 537             rinv00           = gmx_invsqrt(rsq00);
 538             rinv10           = gmx_invsqrt(rsq10);
 539             rinv20           = gmx_invsqrt(rsq20);
 540
 541             rinvsq00         = rinv00*rinv00;
 542             rinvsq10         = rinv10*rinv10;
 543             rinvsq20         = rinv20*rinv20;
 544
 545             /* Load parameters for j particles */
 546             jq0              = charge[jnr+0];
 547             vdwjidx0         = 3*vdwtype[jnr+0];
 548
 549             /**************************
 550              * CALCULATE INTERACTIONS *
 551              **************************/
 552
 553             if (rsq00<rcutoff2)
 554             {
 555
 556             r00              = rsq00*rinv00;
 557
 558             qq00             = iq0*jq0;
 559             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 560             cexp1_00         = vdwparam[vdwioffset0+vdwjidx0+1];
 561             cexp2_00         = vdwparam[vdwioffset0+vdwjidx0+2];
 562
 563             /* REACTION-FIELD ELECTROSTATICS */
 564             felec            = qq00*(rinv00*rinvsq00-krf2);
 565
 566             /* BUCKINGHAM DISPERSION/REPULSION */
 567             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 568             vvdw6            = c6_00*rinvsix;
 569             br               = cexp2_00*r00;
 570             vvdwexp          = cexp1_00*exp(-br);
 571             vvdw             = vvdwexp - vvdw6*(1.0/6.0);
 572             fvdw             = (br*vvdwexp-vvdw6)*rinvsq00;
 573
 574             d                = r00-rswitch;
 575             d                = (d>0.0) ? d : 0.0;
 576             d2               = d*d;
 577             sw               = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
 578
 579             dsw              = d2*(swF2+d*(swF3+d*swF4));
 580
 581             /* Evaluate switch function */
 582             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 583             fvdw             = fvdw*sw - rinv00*vvdw*dsw;
 584
 585             fscal            = felec+fvdw;
 586
 587             /* Calculate temporary vectorial force */
 588             tx               = fscal*dx00;
 589             ty               = fscal*dy00;
 590             tz               = fscal*dz00;
 591
 592             /* Update vectorial force */
 593             fix0            += tx;
 594             fiy0            += ty;
 595             fiz0            += tz;
 596             f[j_coord_offset+DIM*0+XX] -= tx;
 597             f[j_coord_offset+DIM*0+YY] -= ty;
 598             f[j_coord_offset+DIM*0+ZZ] -= tz;
 599
 600             }
 601
 602             /**************************
 603              * CALCULATE INTERACTIONS *
 604              **************************/
 605
 606             if (rsq10<rcutoff2)
 607             {
 608
 609             qq10             = iq1*jq0;
 610
 611             /* REACTION-FIELD ELECTROSTATICS */
 612             felec            = qq10*(rinv10*rinvsq10-krf2);
 613
 614             fscal            = felec;
 615
 616             /* Calculate temporary vectorial force */
 617             tx               = fscal*dx10;
 618             ty               = fscal*dy10;
 619             tz               = fscal*dz10;
 620
 621             /* Update vectorial force */
 622             fix1            += tx;
 623             fiy1            += ty;
 624             fiz1            += tz;
 625             f[j_coord_offset+DIM*0+XX] -= tx;
 626             f[j_coord_offset+DIM*0+YY] -= ty;
 627             f[j_coord_offset+DIM*0+ZZ] -= tz;
 628
 629             }
 630
 631             /**************************
 632              * CALCULATE INTERACTIONS *
 633              **************************/
 634
 635             if (rsq20<rcutoff2)
 636             {
 637
 638             qq20             = iq2*jq0;
 639
 640             /* REACTION-FIELD ELECTROSTATICS */
 641             felec            = qq20*(rinv20*rinvsq20-krf2);
 642
 643             fscal            = felec;
 644
 645             /* Calculate temporary vectorial force */
 646             tx               = fscal*dx20;
 647             ty               = fscal*dy20;
 648             tz               = fscal*dz20;
 649
 650             /* Update vectorial force */
 651             fix2            += tx;
 652             fiy2            += ty;
 653             fiz2            += tz;
 654             f[j_coord_offset+DIM*0+XX] -= tx;
 655             f[j_coord_offset+DIM*0+YY] -= ty;
 656             f[j_coord_offset+DIM*0+ZZ] -= tz;
 657
 658             }
 659
 660             /* Inner loop uses 136 flops */
 661         }
 662         /* End of innermost loop */
 663
 664         tx = ty = tz = 0;
 665         f[i_coord_offset+DIM*0+XX] += fix0;
 666         f[i_coord_offset+DIM*0+YY] += fiy0;
 667         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 668         tx                         += fix0;
 669         ty                         += fiy0;
 670         tz                         += fiz0;
 671         f[i_coord_offset+DIM*1+XX] += fix1;
 672         f[i_coord_offset+DIM*1+YY] += fiy1;
 673         f[i_coord_offset+DIM*1+ZZ] += fiz1;
 674         tx                         += fix1;
 675         ty                         += fiy1;
 676         tz                         += fiz1;
 677         f[i_coord_offset+DIM*2+XX] += fix2;
 678         f[i_coord_offset+DIM*2+YY] += fiy2;
 679         f[i_coord_offset+DIM*2+ZZ] += fiz2;
 680         tx                         += fix2;
 681         ty                         += fiy2;
 682         tz                         += fiz2;
 683         fshift[i_shift_offset+XX]  += tx;
 684         fshift[i_shift_offset+YY]  += ty;
 685         fshift[i_shift_offset+ZZ]  += tz;
 686
 687         /* Increment number of inner iterations */
 688         inneriter                  += j_index_end - j_index_start;
 689
 690         /* Outer loop uses 30 flops */
 691     }
 692
 693     /* Increment number of outer iterations */
 694     outeriter        += nri;
 695
 696     /* Update outer/inner flops */
 697
 698     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3_F,outeriter*30 + inneriter*136);
 699 }