src/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecRFCut_VdwLJSh_GeomW4P1_c.c

   1 /*
   2  * Note: this file was generated by the Gromacs c kernel generator.
   3  *
   4  *                This source code is part of
   5  *
   6  *                 G   R   O   M   A   C   S
   7  *
   8  * Copyright (c) 2001-2012, The GROMACS Development Team
   9  *
  10  * Gromacs is a library for molecular simulation and trajectory analysis,
  11  * written by Erik Lindahl, David van der Spoel, Berk Hess, and others - for
  12  * a full list of developers and information, check out http://www.gromacs.org
  13  *
  14  * This program is free software; you can redistribute it and/or modify it under
  15  * the terms of the GNU Lesser General Public License as published by the Free
  16  * Software Foundation; either version 2 of the License, or (at your option) any
  17  * later version.
  18  *
  19  * To help fund GROMACS development, we humbly ask that you cite
  20  * the papers people have written on it - you can find them on the website.
  21  */
  22 #ifdef HAVE_CONFIG_H
  23 #include <config.h>
  24 #endif
  25
  26 #include <math.h>
  27
  28 #include "../nb_kernel.h"
  29 #include "types/simple.h"
  30 #include "vec.h"
  31 #include "nrnb.h"
  32
  33 /*
  34  * Gromacs nonbonded kernel:   nb_kernel_ElecRFCut_VdwLJSh_GeomW4P1_VF_c
  35  * Electrostatics interaction: ReactionField
  36  * VdW interaction:            LennardJones
  37  * Geometry:                   Water4-Particle
  38  * Calculate force/pot:        PotentialAndForce
  39  */
  40 void
  41 nb_kernel_ElecRFCut_VdwLJSh_GeomW4P1_VF_c
  42                     (t_nblist * gmx_restrict                nlist,
  43                      rvec * gmx_restrict                    xx,
  44                      rvec * gmx_restrict                    ff,
  45                      t_forcerec * gmx_restrict              fr,
  46                      t_mdatoms * gmx_restrict               mdatoms,
  47                      nb_kernel_data_t * gmx_restrict        kernel_data,
  48                      t_nrnb * gmx_restrict                  nrnb)
  49 {
  50     int              i_shift_offset,i_coord_offset,j_coord_offset;
  51     int              j_index_start,j_index_end;
  52     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
  53     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
  54     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  55     real             *shiftvec,*fshift,*x,*f;
  56     int              vdwioffset0;
  57     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  58     int              vdwioffset1;
  59     real             ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
  60     int              vdwioffset2;
  61     real             ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
  62     int              vdwioffset3;
  63     real             ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
  64     int              vdwjidx0;
  65     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  66     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
  67     real             dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10;
  68     real             dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20;
  69     real             dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30,cexp1_30,cexp2_30;
  70     real             velec,felec,velecsum,facel,crf,krf,krf2;
  71     real             *charge;
  72     int              nvdwtype;
  73     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
  74     int              *vdwtype;
  75     real             *vdwparam;
  76
  77     x                = xx[0];
  78     f                = ff[0];
  79
  80     nri              = nlist->nri;
  81     iinr             = nlist->iinr;
  82     jindex           = nlist->jindex;
  83     jjnr             = nlist->jjnr;
  84     shiftidx         = nlist->shift;
  85     gid              = nlist->gid;
  86     shiftvec         = fr->shift_vec[0];
  87     fshift           = fr->fshift[0];
  88     facel            = fr->epsfac;
  89     charge           = mdatoms->chargeA;
  90     krf              = fr->ic->k_rf;
  91     krf2             = krf*2.0;
  92     crf              = fr->ic->c_rf;
  93     nvdwtype         = fr->ntype;
  94     vdwparam         = fr->nbfp;
  95     vdwtype          = mdatoms->typeA;
  96
  97     /* Setup water-specific parameters */
  98     inr              = nlist->iinr[0];
  99     iq1              = facel*charge[inr+1];
 100     iq2              = facel*charge[inr+2];
 101     iq3              = facel*charge[inr+3];
 102     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 103
 104     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 105     rcutoff          = fr->rcoulomb;
 106     rcutoff2         = rcutoff*rcutoff;
 107
 108     sh_vdw_invrcut6  = fr->ic->sh_invrc6;
 109     rvdw             = fr->rvdw;
 110
 111     outeriter        = 0;
 112     inneriter        = 0;
 113
 114     /* Start outer loop over neighborlists */
 115     for(iidx=0; iidx<nri; iidx++)
 116     {
 117         /* Load shift vector for this list */
 118         i_shift_offset   = DIM*shiftidx[iidx];
 119         shX              = shiftvec[i_shift_offset+XX];
 120         shY              = shiftvec[i_shift_offset+YY];
 121         shZ              = shiftvec[i_shift_offset+ZZ];
 122
 123         /* Load limits for loop over neighbors */
 124         j_index_start    = jindex[iidx];
 125         j_index_end      = jindex[iidx+1];
 126
 127         /* Get outer coordinate index */
 128         inr              = iinr[iidx];
 129         i_coord_offset   = DIM*inr;
 130
 131         /* Load i particle coords and add shift vector */
 132         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 133         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 134         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 135         ix1              = shX + x[i_coord_offset+DIM*1+XX];
 136         iy1              = shY + x[i_coord_offset+DIM*1+YY];
 137         iz1              = shZ + x[i_coord_offset+DIM*1+ZZ];
 138         ix2              = shX + x[i_coord_offset+DIM*2+XX];
 139         iy2              = shY + x[i_coord_offset+DIM*2+YY];
 140         iz2              = shZ + x[i_coord_offset+DIM*2+ZZ];
 141         ix3              = shX + x[i_coord_offset+DIM*3+XX];
 142         iy3              = shY + x[i_coord_offset+DIM*3+YY];
 143         iz3              = shZ + x[i_coord_offset+DIM*3+ZZ];
 144
 145         fix0             = 0.0;
 146         fiy0             = 0.0;
 147         fiz0             = 0.0;
 148         fix1             = 0.0;
 149         fiy1             = 0.0;
 150         fiz1             = 0.0;
 151         fix2             = 0.0;
 152         fiy2             = 0.0;
 153         fiz2             = 0.0;
 154         fix3             = 0.0;
 155         fiy3             = 0.0;
 156         fiz3             = 0.0;
 157
 158         /* Reset potential sums */
 159         velecsum         = 0.0;
 160         vvdwsum          = 0.0;
 161
 162         /* Start inner kernel loop */
 163         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 164         {
 165             /* Get j neighbor index, and coordinate index */
 166             jnr              = jjnr[jidx];
 167             j_coord_offset   = DIM*jnr;
 168
 169             /* load j atom coordinates */
 170             jx0              = x[j_coord_offset+DIM*0+XX];
 171             jy0              = x[j_coord_offset+DIM*0+YY];
 172             jz0              = x[j_coord_offset+DIM*0+ZZ];
 173
 174             /* Calculate displacement vector */
 175             dx00             = ix0 - jx0;
 176             dy00             = iy0 - jy0;
 177             dz00             = iz0 - jz0;
 178             dx10             = ix1 - jx0;
 179             dy10             = iy1 - jy0;
 180             dz10             = iz1 - jz0;
 181             dx20             = ix2 - jx0;
 182             dy20             = iy2 - jy0;
 183             dz20             = iz2 - jz0;
 184             dx30             = ix3 - jx0;
 185             dy30             = iy3 - jy0;
 186             dz30             = iz3 - jz0;
 187
 188             /* Calculate squared distance and things based on it */
 189             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 190             rsq10            = dx10*dx10+dy10*dy10+dz10*dz10;
 191             rsq20            = dx20*dx20+dy20*dy20+dz20*dz20;
 192             rsq30            = dx30*dx30+dy30*dy30+dz30*dz30;
 193
 194             rinv10           = gmx_invsqrt(rsq10);
 195             rinv20           = gmx_invsqrt(rsq20);
 196             rinv30           = gmx_invsqrt(rsq30);
 197
 198             rinvsq00         = 1.0/rsq00;
 199             rinvsq10         = rinv10*rinv10;
 200             rinvsq20         = rinv20*rinv20;
 201             rinvsq30         = rinv30*rinv30;
 202
 203             /* Load parameters for j particles */
 204             jq0              = charge[jnr+0];
 205             vdwjidx0         = 2*vdwtype[jnr+0];
 206
 207             /**************************
 208              * CALCULATE INTERACTIONS *
 209              **************************/
 210
 211             if (rsq00<rcutoff2)
 212             {
 213
 214             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 215             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 216
 217             /* LENNARD-JONES DISPERSION/REPULSION */
 218
 219             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 220             vvdw6            = c6_00*rinvsix;
 221             vvdw12           = c12_00*rinvsix*rinvsix;
 222             vvdw             = (vvdw12 - c12_00*sh_vdw_invrcut6*sh_vdw_invrcut6)*(1.0/12.0) - (vvdw6 - c6_00*sh_vdw_invrcut6)*(1.0/6.0);
 223             fvdw             = (vvdw12-vvdw6)*rinvsq00;
 224
 225             /* Update potential sums from outer loop */
 226             vvdwsum         += vvdw;
 227
 228             fscal            = fvdw;
 229
 230             /* Calculate temporary vectorial force */
 231             tx               = fscal*dx00;
 232             ty               = fscal*dy00;
 233             tz               = fscal*dz00;
 234
 235             /* Update vectorial force */
 236             fix0            += tx;
 237             fiy0            += ty;
 238             fiz0            += tz;
 239             f[j_coord_offset+DIM*0+XX] -= tx;
 240             f[j_coord_offset+DIM*0+YY] -= ty;
 241             f[j_coord_offset+DIM*0+ZZ] -= tz;
 242
 243             }
 244
 245             /**************************
 246              * CALCULATE INTERACTIONS *
 247              **************************/
 248
 249             if (rsq10<rcutoff2)
 250             {
 251
 252             qq10             = iq1*jq0;
 253
 254             /* REACTION-FIELD ELECTROSTATICS */
 255             velec            = qq10*(rinv10+krf*rsq10-crf);
 256             felec            = qq10*(rinv10*rinvsq10-krf2);
 257
 258             /* Update potential sums from outer loop */
 259             velecsum        += velec;
 260
 261             fscal            = felec;
 262
 263             /* Calculate temporary vectorial force */
 264             tx               = fscal*dx10;
 265             ty               = fscal*dy10;
 266             tz               = fscal*dz10;
 267
 268             /* Update vectorial force */
 269             fix1            += tx;
 270             fiy1            += ty;
 271             fiz1            += tz;
 272             f[j_coord_offset+DIM*0+XX] -= tx;
 273             f[j_coord_offset+DIM*0+YY] -= ty;
 274             f[j_coord_offset+DIM*0+ZZ] -= tz;
 275
 276             }
 277
 278             /**************************
 279              * CALCULATE INTERACTIONS *
 280              **************************/
 281
 282             if (rsq20<rcutoff2)
 283             {
 284
 285             qq20             = iq2*jq0;
 286
 287             /* REACTION-FIELD ELECTROSTATICS */
 288             velec            = qq20*(rinv20+krf*rsq20-crf);
 289             felec            = qq20*(rinv20*rinvsq20-krf2);
 290
 291             /* Update potential sums from outer loop */
 292             velecsum        += velec;
 293
 294             fscal            = felec;
 295
 296             /* Calculate temporary vectorial force */
 297             tx               = fscal*dx20;
 298             ty               = fscal*dy20;
 299             tz               = fscal*dz20;
 300
 301             /* Update vectorial force */
 302             fix2            += tx;
 303             fiy2            += ty;
 304             fiz2            += tz;
 305             f[j_coord_offset+DIM*0+XX] -= tx;
 306             f[j_coord_offset+DIM*0+YY] -= ty;
 307             f[j_coord_offset+DIM*0+ZZ] -= tz;
 308
 309             }
 310
 311             /**************************
 312              * CALCULATE INTERACTIONS *
 313              **************************/
 314
 315             if (rsq30<rcutoff2)
 316             {
 317
 318             qq30             = iq3*jq0;
 319
 320             /* REACTION-FIELD ELECTROSTATICS */
 321             velec            = qq30*(rinv30+krf*rsq30-crf);
 322             felec            = qq30*(rinv30*rinvsq30-krf2);
 323
 324             /* Update potential sums from outer loop */
 325             velecsum        += velec;
 326
 327             fscal            = felec;
 328
 329             /* Calculate temporary vectorial force */
 330             tx               = fscal*dx30;
 331             ty               = fscal*dy30;
 332             tz               = fscal*dz30;
 333
 334             /* Update vectorial force */
 335             fix3            += tx;
 336             fiy3            += ty;
 337             fiz3            += tz;
 338             f[j_coord_offset+DIM*0+XX] -= tx;
 339             f[j_coord_offset+DIM*0+YY] -= ty;
 340             f[j_coord_offset+DIM*0+ZZ] -= tz;
 341
 342             }
 343
 344             /* Inner loop uses 133 flops */
 345         }
 346         /* End of innermost loop */
 347
 348         tx = ty = tz = 0;
 349         f[i_coord_offset+DIM*0+XX] += fix0;
 350         f[i_coord_offset+DIM*0+YY] += fiy0;
 351         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 352         tx                         += fix0;
 353         ty                         += fiy0;
 354         tz                         += fiz0;
 355         f[i_coord_offset+DIM*1+XX] += fix1;
 356         f[i_coord_offset+DIM*1+YY] += fiy1;
 357         f[i_coord_offset+DIM*1+ZZ] += fiz1;
 358         tx                         += fix1;
 359         ty                         += fiy1;
 360         tz                         += fiz1;
 361         f[i_coord_offset+DIM*2+XX] += fix2;
 362         f[i_coord_offset+DIM*2+YY] += fiy2;
 363         f[i_coord_offset+DIM*2+ZZ] += fiz2;
 364         tx                         += fix2;
 365         ty                         += fiy2;
 366         tz                         += fiz2;
 367         f[i_coord_offset+DIM*3+XX] += fix3;
 368         f[i_coord_offset+DIM*3+YY] += fiy3;
 369         f[i_coord_offset+DIM*3+ZZ] += fiz3;
 370         tx                         += fix3;
 371         ty                         += fiy3;
 372         tz                         += fiz3;
 373         fshift[i_shift_offset+XX]  += tx;
 374         fshift[i_shift_offset+YY]  += ty;
 375         fshift[i_shift_offset+ZZ]  += tz;
 376
 377         ggid                        = gid[iidx];
 378         /* Update potential energies */
 379         kernel_data->energygrp_elec[ggid] += velecsum;
 380         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 381
 382         /* Increment number of inner iterations */
 383         inneriter                  += j_index_end - j_index_start;
 384
 385         /* Outer loop uses 41 flops */
 386     }
 387
 388     /* Increment number of outer iterations */
 389     outeriter        += nri;
 390
 391     /* Update outer/inner flops */
 392
 393     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_VF,outeriter*41 + inneriter*133);
 394 }
 395 /*
 396  * Gromacs nonbonded kernel:   nb_kernel_ElecRFCut_VdwLJSh_GeomW4P1_F_c
 397  * Electrostatics interaction: ReactionField
 398  * VdW interaction:            LennardJones
 399  * Geometry:                   Water4-Particle
 400  * Calculate force/pot:        Force
 401  */
 402 void
 403 nb_kernel_ElecRFCut_VdwLJSh_GeomW4P1_F_c
 404                     (t_nblist * gmx_restrict                nlist,
 405                      rvec * gmx_restrict                    xx,
 406                      rvec * gmx_restrict                    ff,
 407                      t_forcerec * gmx_restrict              fr,
 408                      t_mdatoms * gmx_restrict               mdatoms,
 409                      nb_kernel_data_t * gmx_restrict        kernel_data,
 410                      t_nrnb * gmx_restrict                  nrnb)
 411 {
 412     int              i_shift_offset,i_coord_offset,j_coord_offset;
 413     int              j_index_start,j_index_end;
 414     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 415     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 416     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 417     real             *shiftvec,*fshift,*x,*f;
 418     int              vdwioffset0;
 419     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 420     int              vdwioffset1;
 421     real             ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
 422     int              vdwioffset2;
 423     real             ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
 424     int              vdwioffset3;
 425     real             ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
 426     int              vdwjidx0;
 427     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 428     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 429     real             dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10;
 430     real             dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20;
 431     real             dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30,cexp1_30,cexp2_30;
 432     real             velec,felec,velecsum,facel,crf,krf,krf2;
 433     real             *charge;
 434     int              nvdwtype;
 435     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 436     int              *vdwtype;
 437     real             *vdwparam;
 438
 439     x                = xx[0];
 440     f                = ff[0];
 441
 442     nri              = nlist->nri;
 443     iinr             = nlist->iinr;
 444     jindex           = nlist->jindex;
 445     jjnr             = nlist->jjnr;
 446     shiftidx         = nlist->shift;
 447     gid              = nlist->gid;
 448     shiftvec         = fr->shift_vec[0];
 449     fshift           = fr->fshift[0];
 450     facel            = fr->epsfac;
 451     charge           = mdatoms->chargeA;
 452     krf              = fr->ic->k_rf;
 453     krf2             = krf*2.0;
 454     crf              = fr->ic->c_rf;
 455     nvdwtype         = fr->ntype;
 456     vdwparam         = fr->nbfp;
 457     vdwtype          = mdatoms->typeA;
 458
 459     /* Setup water-specific parameters */
 460     inr              = nlist->iinr[0];
 461     iq1              = facel*charge[inr+1];
 462     iq2              = facel*charge[inr+2];
 463     iq3              = facel*charge[inr+3];
 464     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 465
 466     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 467     rcutoff          = fr->rcoulomb;
 468     rcutoff2         = rcutoff*rcutoff;
 469
 470     sh_vdw_invrcut6  = fr->ic->sh_invrc6;
 471     rvdw             = fr->rvdw;
 472
 473     outeriter        = 0;
 474     inneriter        = 0;
 475
 476     /* Start outer loop over neighborlists */
 477     for(iidx=0; iidx<nri; iidx++)
 478     {
 479         /* Load shift vector for this list */
 480         i_shift_offset   = DIM*shiftidx[iidx];
 481         shX              = shiftvec[i_shift_offset+XX];
 482         shY              = shiftvec[i_shift_offset+YY];
 483         shZ              = shiftvec[i_shift_offset+ZZ];
 484
 485         /* Load limits for loop over neighbors */
 486         j_index_start    = jindex[iidx];
 487         j_index_end      = jindex[iidx+1];
 488
 489         /* Get outer coordinate index */
 490         inr              = iinr[iidx];
 491         i_coord_offset   = DIM*inr;
 492
 493         /* Load i particle coords and add shift vector */
 494         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 495         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 496         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 497         ix1              = shX + x[i_coord_offset+DIM*1+XX];
 498         iy1              = shY + x[i_coord_offset+DIM*1+YY];
 499         iz1              = shZ + x[i_coord_offset+DIM*1+ZZ];
 500         ix2              = shX + x[i_coord_offset+DIM*2+XX];
 501         iy2              = shY + x[i_coord_offset+DIM*2+YY];
 502         iz2              = shZ + x[i_coord_offset+DIM*2+ZZ];
 503         ix3              = shX + x[i_coord_offset+DIM*3+XX];
 504         iy3              = shY + x[i_coord_offset+DIM*3+YY];
 505         iz3              = shZ + x[i_coord_offset+DIM*3+ZZ];
 506
 507         fix0             = 0.0;
 508         fiy0             = 0.0;
 509         fiz0             = 0.0;
 510         fix1             = 0.0;
 511         fiy1             = 0.0;
 512         fiz1             = 0.0;
 513         fix2             = 0.0;
 514         fiy2             = 0.0;
 515         fiz2             = 0.0;
 516         fix3             = 0.0;
 517         fiy3             = 0.0;
 518         fiz3             = 0.0;
 519
 520         /* Start inner kernel loop */
 521         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 522         {
 523             /* Get j neighbor index, and coordinate index */
 524             jnr              = jjnr[jidx];
 525             j_coord_offset   = DIM*jnr;
 526
 527             /* load j atom coordinates */
 528             jx0              = x[j_coord_offset+DIM*0+XX];
 529             jy0              = x[j_coord_offset+DIM*0+YY];
 530             jz0              = x[j_coord_offset+DIM*0+ZZ];
 531
 532             /* Calculate displacement vector */
 533             dx00             = ix0 - jx0;
 534             dy00             = iy0 - jy0;
 535             dz00             = iz0 - jz0;
 536             dx10             = ix1 - jx0;
 537             dy10             = iy1 - jy0;
 538             dz10             = iz1 - jz0;
 539             dx20             = ix2 - jx0;
 540             dy20             = iy2 - jy0;
 541             dz20             = iz2 - jz0;
 542             dx30             = ix3 - jx0;
 543             dy30             = iy3 - jy0;
 544             dz30             = iz3 - jz0;
 545
 546             /* Calculate squared distance and things based on it */
 547             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 548             rsq10            = dx10*dx10+dy10*dy10+dz10*dz10;
 549             rsq20            = dx20*dx20+dy20*dy20+dz20*dz20;
 550             rsq30            = dx30*dx30+dy30*dy30+dz30*dz30;
 551
 552             rinv10           = gmx_invsqrt(rsq10);
 553             rinv20           = gmx_invsqrt(rsq20);
 554             rinv30           = gmx_invsqrt(rsq30);
 555
 556             rinvsq00         = 1.0/rsq00;
 557             rinvsq10         = rinv10*rinv10;
 558             rinvsq20         = rinv20*rinv20;
 559             rinvsq30         = rinv30*rinv30;
 560
 561             /* Load parameters for j particles */
 562             jq0              = charge[jnr+0];
 563             vdwjidx0         = 2*vdwtype[jnr+0];
 564
 565             /**************************
 566              * CALCULATE INTERACTIONS *
 567              **************************/
 568
 569             if (rsq00<rcutoff2)
 570             {
 571
 572             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 573             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 574
 575             /* LENNARD-JONES DISPERSION/REPULSION */
 576
 577             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 578             fvdw             = (c12_00*rinvsix-c6_00)*rinvsix*rinvsq00;
 579
 580             fscal            = fvdw;
 581
 582             /* Calculate temporary vectorial force */
 583             tx               = fscal*dx00;
 584             ty               = fscal*dy00;
 585             tz               = fscal*dz00;
 586
 587             /* Update vectorial force */
 588             fix0            += tx;
 589             fiy0            += ty;
 590             fiz0            += tz;
 591             f[j_coord_offset+DIM*0+XX] -= tx;
 592             f[j_coord_offset+DIM*0+YY] -= ty;
 593             f[j_coord_offset+DIM*0+ZZ] -= tz;
 594
 595             }
 596
 597             /**************************
 598              * CALCULATE INTERACTIONS *
 599              **************************/
 600
 601             if (rsq10<rcutoff2)
 602             {
 603
 604             qq10             = iq1*jq0;
 605
 606             /* REACTION-FIELD ELECTROSTATICS */
 607             felec            = qq10*(rinv10*rinvsq10-krf2);
 608
 609             fscal            = felec;
 610
 611             /* Calculate temporary vectorial force */
 612             tx               = fscal*dx10;
 613             ty               = fscal*dy10;
 614             tz               = fscal*dz10;
 615
 616             /* Update vectorial force */
 617             fix1            += tx;
 618             fiy1            += ty;
 619             fiz1            += tz;
 620             f[j_coord_offset+DIM*0+XX] -= tx;
 621             f[j_coord_offset+DIM*0+YY] -= ty;
 622             f[j_coord_offset+DIM*0+ZZ] -= tz;
 623
 624             }
 625
 626             /**************************
 627              * CALCULATE INTERACTIONS *
 628              **************************/
 629
 630             if (rsq20<rcutoff2)
 631             {
 632
 633             qq20             = iq2*jq0;
 634
 635             /* REACTION-FIELD ELECTROSTATICS */
 636             felec            = qq20*(rinv20*rinvsq20-krf2);
 637
 638             fscal            = felec;
 639
 640             /* Calculate temporary vectorial force */
 641             tx               = fscal*dx20;
 642             ty               = fscal*dy20;
 643             tz               = fscal*dz20;
 644
 645             /* Update vectorial force */
 646             fix2            += tx;
 647             fiy2            += ty;
 648             fiz2            += tz;
 649             f[j_coord_offset+DIM*0+XX] -= tx;
 650             f[j_coord_offset+DIM*0+YY] -= ty;
 651             f[j_coord_offset+DIM*0+ZZ] -= tz;
 652
 653             }
 654
 655             /**************************
 656              * CALCULATE INTERACTIONS *
 657              **************************/
 658
 659             if (rsq30<rcutoff2)
 660             {
 661
 662             qq30             = iq3*jq0;
 663
 664             /* REACTION-FIELD ELECTROSTATICS */
 665             felec            = qq30*(rinv30*rinvsq30-krf2);
 666
 667             fscal            = felec;
 668
 669             /* Calculate temporary vectorial force */
 670             tx               = fscal*dx30;
 671             ty               = fscal*dy30;
 672             tz               = fscal*dz30;
 673
 674             /* Update vectorial force */
 675             fix3            += tx;
 676             fiy3            += ty;
 677             fiz3            += tz;
 678             f[j_coord_offset+DIM*0+XX] -= tx;
 679             f[j_coord_offset+DIM*0+YY] -= ty;
 680             f[j_coord_offset+DIM*0+ZZ] -= tz;
 681
 682             }
 683
 684             /* Inner loop uses 108 flops */
 685         }
 686         /* End of innermost loop */
 687
 688         tx = ty = tz = 0;
 689         f[i_coord_offset+DIM*0+XX] += fix0;
 690         f[i_coord_offset+DIM*0+YY] += fiy0;
 691         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 692         tx                         += fix0;
 693         ty                         += fiy0;
 694         tz                         += fiz0;
 695         f[i_coord_offset+DIM*1+XX] += fix1;
 696         f[i_coord_offset+DIM*1+YY] += fiy1;
 697         f[i_coord_offset+DIM*1+ZZ] += fiz1;
 698         tx                         += fix1;
 699         ty                         += fiy1;
 700         tz                         += fiz1;
 701         f[i_coord_offset+DIM*2+XX] += fix2;
 702         f[i_coord_offset+DIM*2+YY] += fiy2;
 703         f[i_coord_offset+DIM*2+ZZ] += fiz2;
 704         tx                         += fix2;
 705         ty                         += fiy2;
 706         tz                         += fiz2;
 707         f[i_coord_offset+DIM*3+XX] += fix3;
 708         f[i_coord_offset+DIM*3+YY] += fiy3;
 709         f[i_coord_offset+DIM*3+ZZ] += fiz3;
 710         tx                         += fix3;
 711         ty                         += fiy3;
 712         tz                         += fiz3;
 713         fshift[i_shift_offset+XX]  += tx;
 714         fshift[i_shift_offset+YY]  += ty;
 715         fshift[i_shift_offset+ZZ]  += tz;
 716
 717         /* Increment number of inner iterations */
 718         inneriter                  += j_index_end - j_index_start;
 719
 720         /* Outer loop uses 39 flops */
 721     }
 722
 723     /* Increment number of outer iterations */
 724     outeriter        += nri;
 725
 726     /* Update outer/inner flops */
 727
 728     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_F,outeriter*39 + inneriter*108);
 729 }