src/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecRF_VdwCSTab_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_ElecRF_VdwCSTab_GeomW4P1_VF_c
  35  * Electrostatics interaction: ReactionField
  36  * VdW interaction:            CubicSplineTable
  37  * Geometry:                   Water4-Particle
  38  * Calculate force/pot:        PotentialAndForce
  39  */
  40 void
  41 nb_kernel_ElecRF_VdwCSTab_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     int              vfitab;
  77     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
  78     real             *vftab;
  79
  80     x                = xx[0];
  81     f                = ff[0];
  82
  83     nri              = nlist->nri;
  84     iinr             = nlist->iinr;
  85     jindex           = nlist->jindex;
  86     jjnr             = nlist->jjnr;
  87     shiftidx         = nlist->shift;
  88     gid              = nlist->gid;
  89     shiftvec         = fr->shift_vec[0];
  90     fshift           = fr->fshift[0];
  91     facel            = fr->epsfac;
  92     charge           = mdatoms->chargeA;
  93     krf              = fr->ic->k_rf;
  94     krf2             = krf*2.0;
  95     crf              = fr->ic->c_rf;
  96     nvdwtype         = fr->ntype;
  97     vdwparam         = fr->nbfp;
  98     vdwtype          = mdatoms->typeA;
  99
 100     vftab            = kernel_data->table_vdw->data;
 101     vftabscale       = kernel_data->table_vdw->scale;
 102
 103     /* Setup water-specific parameters */
 104     inr              = nlist->iinr[0];
 105     iq1              = facel*charge[inr+1];
 106     iq2              = facel*charge[inr+2];
 107     iq3              = facel*charge[inr+3];
 108     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 109
 110     outeriter        = 0;
 111     inneriter        = 0;
 112
 113     /* Start outer loop over neighborlists */
 114     for(iidx=0; iidx<nri; iidx++)
 115     {
 116         /* Load shift vector for this list */
 117         i_shift_offset   = DIM*shiftidx[iidx];
 118         shX              = shiftvec[i_shift_offset+XX];
 119         shY              = shiftvec[i_shift_offset+YY];
 120         shZ              = shiftvec[i_shift_offset+ZZ];
 121
 122         /* Load limits for loop over neighbors */
 123         j_index_start    = jindex[iidx];
 124         j_index_end      = jindex[iidx+1];
 125
 126         /* Get outer coordinate index */
 127         inr              = iinr[iidx];
 128         i_coord_offset   = DIM*inr;
 129
 130         /* Load i particle coords and add shift vector */
 131         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 132         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 133         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 134         ix1              = shX + x[i_coord_offset+DIM*1+XX];
 135         iy1              = shY + x[i_coord_offset+DIM*1+YY];
 136         iz1              = shZ + x[i_coord_offset+DIM*1+ZZ];
 137         ix2              = shX + x[i_coord_offset+DIM*2+XX];
 138         iy2              = shY + x[i_coord_offset+DIM*2+YY];
 139         iz2              = shZ + x[i_coord_offset+DIM*2+ZZ];
 140         ix3              = shX + x[i_coord_offset+DIM*3+XX];
 141         iy3              = shY + x[i_coord_offset+DIM*3+YY];
 142         iz3              = shZ + x[i_coord_offset+DIM*3+ZZ];
 143
 144         fix0             = 0.0;
 145         fiy0             = 0.0;
 146         fiz0             = 0.0;
 147         fix1             = 0.0;
 148         fiy1             = 0.0;
 149         fiz1             = 0.0;
 150         fix2             = 0.0;
 151         fiy2             = 0.0;
 152         fiz2             = 0.0;
 153         fix3             = 0.0;
 154         fiy3             = 0.0;
 155         fiz3             = 0.0;
 156
 157         /* Reset potential sums */
 158         velecsum         = 0.0;
 159         vvdwsum          = 0.0;
 160
 161         /* Start inner kernel loop */
 162         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 163         {
 164             /* Get j neighbor index, and coordinate index */
 165             jnr              = jjnr[jidx];
 166             j_coord_offset   = DIM*jnr;
 167
 168             /* load j atom coordinates */
 169             jx0              = x[j_coord_offset+DIM*0+XX];
 170             jy0              = x[j_coord_offset+DIM*0+YY];
 171             jz0              = x[j_coord_offset+DIM*0+ZZ];
 172
 173             /* Calculate displacement vector */
 174             dx00             = ix0 - jx0;
 175             dy00             = iy0 - jy0;
 176             dz00             = iz0 - jz0;
 177             dx10             = ix1 - jx0;
 178             dy10             = iy1 - jy0;
 179             dz10             = iz1 - jz0;
 180             dx20             = ix2 - jx0;
 181             dy20             = iy2 - jy0;
 182             dz20             = iz2 - jz0;
 183             dx30             = ix3 - jx0;
 184             dy30             = iy3 - jy0;
 185             dz30             = iz3 - jz0;
 186
 187             /* Calculate squared distance and things based on it */
 188             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 189             rsq10            = dx10*dx10+dy10*dy10+dz10*dz10;
 190             rsq20            = dx20*dx20+dy20*dy20+dz20*dz20;
 191             rsq30            = dx30*dx30+dy30*dy30+dz30*dz30;
 192
 193             rinv00           = gmx_invsqrt(rsq00);
 194             rinv10           = gmx_invsqrt(rsq10);
 195             rinv20           = gmx_invsqrt(rsq20);
 196             rinv30           = gmx_invsqrt(rsq30);
 197
 198             rinvsq10         = rinv10*rinv10;
 199             rinvsq20         = rinv20*rinv20;
 200             rinvsq30         = rinv30*rinv30;
 201
 202             /* Load parameters for j particles */
 203             jq0              = charge[jnr+0];
 204             vdwjidx0         = 2*vdwtype[jnr+0];
 205
 206             /**************************
 207              * CALCULATE INTERACTIONS *
 208              **************************/
 209
 210             r00              = rsq00*rinv00;
 211
 212             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 213             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 214
 215             /* Calculate table index by multiplying r with table scale and truncate to integer */
 216             rt               = r00*vftabscale;
 217             vfitab           = rt;
 218             vfeps            = rt-vfitab;
 219             vfitab           = 2*4*vfitab;
 220
 221             /* CUBIC SPLINE TABLE DISPERSION */
 222             vfitab          += 0;
 223             Y                = vftab[vfitab];
 224             F                = vftab[vfitab+1];
 225             Geps             = vfeps*vftab[vfitab+2];
 226             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 227             Fp               = F+Geps+Heps2;
 228             VV               = Y+vfeps*Fp;
 229             vvdw6            = c6_00*VV;
 230             FF               = Fp+Geps+2.0*Heps2;
 231             fvdw6            = c6_00*FF;
 232
 233             /* CUBIC SPLINE TABLE REPULSION */
 234             Y                = vftab[vfitab+4];
 235             F                = vftab[vfitab+5];
 236             Geps             = vfeps*vftab[vfitab+6];
 237             Heps2            = vfeps*vfeps*vftab[vfitab+7];
 238             Fp               = F+Geps+Heps2;
 239             VV               = Y+vfeps*Fp;
 240             vvdw12           = c12_00*VV;
 241             FF               = Fp+Geps+2.0*Heps2;
 242             fvdw12           = c12_00*FF;
 243             vvdw             = vvdw12+vvdw6;
 244             fvdw             = -(fvdw6+fvdw12)*vftabscale*rinv00;
 245
 246             /* Update potential sums from outer loop */
 247             vvdwsum         += vvdw;
 248
 249             fscal            = fvdw;
 250
 251             /* Calculate temporary vectorial force */
 252             tx               = fscal*dx00;
 253             ty               = fscal*dy00;
 254             tz               = fscal*dz00;
 255
 256             /* Update vectorial force */
 257             fix0            += tx;
 258             fiy0            += ty;
 259             fiz0            += tz;
 260             f[j_coord_offset+DIM*0+XX] -= tx;
 261             f[j_coord_offset+DIM*0+YY] -= ty;
 262             f[j_coord_offset+DIM*0+ZZ] -= tz;
 263
 264             /**************************
 265              * CALCULATE INTERACTIONS *
 266              **************************/
 267
 268             qq10             = iq1*jq0;
 269
 270             /* REACTION-FIELD ELECTROSTATICS */
 271             velec            = qq10*(rinv10+krf*rsq10-crf);
 272             felec            = qq10*(rinv10*rinvsq10-krf2);
 273
 274             /* Update potential sums from outer loop */
 275             velecsum        += velec;
 276
 277             fscal            = felec;
 278
 279             /* Calculate temporary vectorial force */
 280             tx               = fscal*dx10;
 281             ty               = fscal*dy10;
 282             tz               = fscal*dz10;
 283
 284             /* Update vectorial force */
 285             fix1            += tx;
 286             fiy1            += ty;
 287             fiz1            += tz;
 288             f[j_coord_offset+DIM*0+XX] -= tx;
 289             f[j_coord_offset+DIM*0+YY] -= ty;
 290             f[j_coord_offset+DIM*0+ZZ] -= tz;
 291
 292             /**************************
 293              * CALCULATE INTERACTIONS *
 294              **************************/
 295
 296             qq20             = iq2*jq0;
 297
 298             /* REACTION-FIELD ELECTROSTATICS */
 299             velec            = qq20*(rinv20+krf*rsq20-crf);
 300             felec            = qq20*(rinv20*rinvsq20-krf2);
 301
 302             /* Update potential sums from outer loop */
 303             velecsum        += velec;
 304
 305             fscal            = felec;
 306
 307             /* Calculate temporary vectorial force */
 308             tx               = fscal*dx20;
 309             ty               = fscal*dy20;
 310             tz               = fscal*dz20;
 311
 312             /* Update vectorial force */
 313             fix2            += tx;
 314             fiy2            += ty;
 315             fiz2            += tz;
 316             f[j_coord_offset+DIM*0+XX] -= tx;
 317             f[j_coord_offset+DIM*0+YY] -= ty;
 318             f[j_coord_offset+DIM*0+ZZ] -= tz;
 319
 320             /**************************
 321              * CALCULATE INTERACTIONS *
 322              **************************/
 323
 324             qq30             = iq3*jq0;
 325
 326             /* REACTION-FIELD ELECTROSTATICS */
 327             velec            = qq30*(rinv30+krf*rsq30-crf);
 328             felec            = qq30*(rinv30*rinvsq30-krf2);
 329
 330             /* Update potential sums from outer loop */
 331             velecsum        += velec;
 332
 333             fscal            = felec;
 334
 335             /* Calculate temporary vectorial force */
 336             tx               = fscal*dx30;
 337             ty               = fscal*dy30;
 338             tz               = fscal*dz30;
 339
 340             /* Update vectorial force */
 341             fix3            += tx;
 342             fiy3            += ty;
 343             fiz3            += tz;
 344             f[j_coord_offset+DIM*0+XX] -= tx;
 345             f[j_coord_offset+DIM*0+YY] -= ty;
 346             f[j_coord_offset+DIM*0+ZZ] -= tz;
 347
 348             /* Inner loop uses 151 flops */
 349         }
 350         /* End of innermost loop */
 351
 352         tx = ty = tz = 0;
 353         f[i_coord_offset+DIM*0+XX] += fix0;
 354         f[i_coord_offset+DIM*0+YY] += fiy0;
 355         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 356         tx                         += fix0;
 357         ty                         += fiy0;
 358         tz                         += fiz0;
 359         f[i_coord_offset+DIM*1+XX] += fix1;
 360         f[i_coord_offset+DIM*1+YY] += fiy1;
 361         f[i_coord_offset+DIM*1+ZZ] += fiz1;
 362         tx                         += fix1;
 363         ty                         += fiy1;
 364         tz                         += fiz1;
 365         f[i_coord_offset+DIM*2+XX] += fix2;
 366         f[i_coord_offset+DIM*2+YY] += fiy2;
 367         f[i_coord_offset+DIM*2+ZZ] += fiz2;
 368         tx                         += fix2;
 369         ty                         += fiy2;
 370         tz                         += fiz2;
 371         f[i_coord_offset+DIM*3+XX] += fix3;
 372         f[i_coord_offset+DIM*3+YY] += fiy3;
 373         f[i_coord_offset+DIM*3+ZZ] += fiz3;
 374         tx                         += fix3;
 375         ty                         += fiy3;
 376         tz                         += fiz3;
 377         fshift[i_shift_offset+XX]  += tx;
 378         fshift[i_shift_offset+YY]  += ty;
 379         fshift[i_shift_offset+ZZ]  += tz;
 380
 381         ggid                        = gid[iidx];
 382         /* Update potential energies */
 383         kernel_data->energygrp_elec[ggid] += velecsum;
 384         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 385
 386         /* Increment number of inner iterations */
 387         inneriter                  += j_index_end - j_index_start;
 388
 389         /* Outer loop uses 41 flops */
 390     }
 391
 392     /* Increment number of outer iterations */
 393     outeriter        += nri;
 394
 395     /* Update outer/inner flops */
 396
 397     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_VF,outeriter*41 + inneriter*151);
 398 }
 399 /*
 400  * Gromacs nonbonded kernel:   nb_kernel_ElecRF_VdwCSTab_GeomW4P1_F_c
 401  * Electrostatics interaction: ReactionField
 402  * VdW interaction:            CubicSplineTable
 403  * Geometry:                   Water4-Particle
 404  * Calculate force/pot:        Force
 405  */
 406 void
 407 nb_kernel_ElecRF_VdwCSTab_GeomW4P1_F_c
 408                     (t_nblist * gmx_restrict                nlist,
 409                      rvec * gmx_restrict                    xx,
 410                      rvec * gmx_restrict                    ff,
 411                      t_forcerec * gmx_restrict              fr,
 412                      t_mdatoms * gmx_restrict               mdatoms,
 413                      nb_kernel_data_t * gmx_restrict        kernel_data,
 414                      t_nrnb * gmx_restrict                  nrnb)
 415 {
 416     int              i_shift_offset,i_coord_offset,j_coord_offset;
 417     int              j_index_start,j_index_end;
 418     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 419     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 420     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 421     real             *shiftvec,*fshift,*x,*f;
 422     int              vdwioffset0;
 423     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 424     int              vdwioffset1;
 425     real             ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
 426     int              vdwioffset2;
 427     real             ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
 428     int              vdwioffset3;
 429     real             ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
 430     int              vdwjidx0;
 431     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 432     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 433     real             dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10;
 434     real             dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20;
 435     real             dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30,cexp1_30,cexp2_30;
 436     real             velec,felec,velecsum,facel,crf,krf,krf2;
 437     real             *charge;
 438     int              nvdwtype;
 439     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 440     int              *vdwtype;
 441     real             *vdwparam;
 442     int              vfitab;
 443     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
 444     real             *vftab;
 445
 446     x                = xx[0];
 447     f                = ff[0];
 448
 449     nri              = nlist->nri;
 450     iinr             = nlist->iinr;
 451     jindex           = nlist->jindex;
 452     jjnr             = nlist->jjnr;
 453     shiftidx         = nlist->shift;
 454     gid              = nlist->gid;
 455     shiftvec         = fr->shift_vec[0];
 456     fshift           = fr->fshift[0];
 457     facel            = fr->epsfac;
 458     charge           = mdatoms->chargeA;
 459     krf              = fr->ic->k_rf;
 460     krf2             = krf*2.0;
 461     crf              = fr->ic->c_rf;
 462     nvdwtype         = fr->ntype;
 463     vdwparam         = fr->nbfp;
 464     vdwtype          = mdatoms->typeA;
 465
 466     vftab            = kernel_data->table_vdw->data;
 467     vftabscale       = kernel_data->table_vdw->scale;
 468
 469     /* Setup water-specific parameters */
 470     inr              = nlist->iinr[0];
 471     iq1              = facel*charge[inr+1];
 472     iq2              = facel*charge[inr+2];
 473     iq3              = facel*charge[inr+3];
 474     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 475
 476     outeriter        = 0;
 477     inneriter        = 0;
 478
 479     /* Start outer loop over neighborlists */
 480     for(iidx=0; iidx<nri; iidx++)
 481     {
 482         /* Load shift vector for this list */
 483         i_shift_offset   = DIM*shiftidx[iidx];
 484         shX              = shiftvec[i_shift_offset+XX];
 485         shY              = shiftvec[i_shift_offset+YY];
 486         shZ              = shiftvec[i_shift_offset+ZZ];
 487
 488         /* Load limits for loop over neighbors */
 489         j_index_start    = jindex[iidx];
 490         j_index_end      = jindex[iidx+1];
 491
 492         /* Get outer coordinate index */
 493         inr              = iinr[iidx];
 494         i_coord_offset   = DIM*inr;
 495
 496         /* Load i particle coords and add shift vector */
 497         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 498         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 499         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 500         ix1              = shX + x[i_coord_offset+DIM*1+XX];
 501         iy1              = shY + x[i_coord_offset+DIM*1+YY];
 502         iz1              = shZ + x[i_coord_offset+DIM*1+ZZ];
 503         ix2              = shX + x[i_coord_offset+DIM*2+XX];
 504         iy2              = shY + x[i_coord_offset+DIM*2+YY];
 505         iz2              = shZ + x[i_coord_offset+DIM*2+ZZ];
 506         ix3              = shX + x[i_coord_offset+DIM*3+XX];
 507         iy3              = shY + x[i_coord_offset+DIM*3+YY];
 508         iz3              = shZ + x[i_coord_offset+DIM*3+ZZ];
 509
 510         fix0             = 0.0;
 511         fiy0             = 0.0;
 512         fiz0             = 0.0;
 513         fix1             = 0.0;
 514         fiy1             = 0.0;
 515         fiz1             = 0.0;
 516         fix2             = 0.0;
 517         fiy2             = 0.0;
 518         fiz2             = 0.0;
 519         fix3             = 0.0;
 520         fiy3             = 0.0;
 521         fiz3             = 0.0;
 522
 523         /* Start inner kernel loop */
 524         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 525         {
 526             /* Get j neighbor index, and coordinate index */
 527             jnr              = jjnr[jidx];
 528             j_coord_offset   = DIM*jnr;
 529
 530             /* load j atom coordinates */
 531             jx0              = x[j_coord_offset+DIM*0+XX];
 532             jy0              = x[j_coord_offset+DIM*0+YY];
 533             jz0              = x[j_coord_offset+DIM*0+ZZ];
 534
 535             /* Calculate displacement vector */
 536             dx00             = ix0 - jx0;
 537             dy00             = iy0 - jy0;
 538             dz00             = iz0 - jz0;
 539             dx10             = ix1 - jx0;
 540             dy10             = iy1 - jy0;
 541             dz10             = iz1 - jz0;
 542             dx20             = ix2 - jx0;
 543             dy20             = iy2 - jy0;
 544             dz20             = iz2 - jz0;
 545             dx30             = ix3 - jx0;
 546             dy30             = iy3 - jy0;
 547             dz30             = iz3 - jz0;
 548
 549             /* Calculate squared distance and things based on it */
 550             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 551             rsq10            = dx10*dx10+dy10*dy10+dz10*dz10;
 552             rsq20            = dx20*dx20+dy20*dy20+dz20*dz20;
 553             rsq30            = dx30*dx30+dy30*dy30+dz30*dz30;
 554
 555             rinv00           = gmx_invsqrt(rsq00);
 556             rinv10           = gmx_invsqrt(rsq10);
 557             rinv20           = gmx_invsqrt(rsq20);
 558             rinv30           = gmx_invsqrt(rsq30);
 559
 560             rinvsq10         = rinv10*rinv10;
 561             rinvsq20         = rinv20*rinv20;
 562             rinvsq30         = rinv30*rinv30;
 563
 564             /* Load parameters for j particles */
 565             jq0              = charge[jnr+0];
 566             vdwjidx0         = 2*vdwtype[jnr+0];
 567
 568             /**************************
 569              * CALCULATE INTERACTIONS *
 570              **************************/
 571
 572             r00              = rsq00*rinv00;
 573
 574             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 575             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 576
 577             /* Calculate table index by multiplying r with table scale and truncate to integer */
 578             rt               = r00*vftabscale;
 579             vfitab           = rt;
 580             vfeps            = rt-vfitab;
 581             vfitab           = 2*4*vfitab;
 582
 583             /* CUBIC SPLINE TABLE DISPERSION */
 584             vfitab          += 0;
 585             Y                = vftab[vfitab];
 586             F                = vftab[vfitab+1];
 587             Geps             = vfeps*vftab[vfitab+2];
 588             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 589             Fp               = F+Geps+Heps2;
 590             FF               = Fp+Geps+2.0*Heps2;
 591             fvdw6            = c6_00*FF;
 592
 593             /* CUBIC SPLINE TABLE REPULSION */
 594             Y                = vftab[vfitab+4];
 595             F                = vftab[vfitab+5];
 596             Geps             = vfeps*vftab[vfitab+6];
 597             Heps2            = vfeps*vfeps*vftab[vfitab+7];
 598             Fp               = F+Geps+Heps2;
 599             FF               = Fp+Geps+2.0*Heps2;
 600             fvdw12           = c12_00*FF;
 601             fvdw             = -(fvdw6+fvdw12)*vftabscale*rinv00;
 602
 603             fscal            = fvdw;
 604
 605             /* Calculate temporary vectorial force */
 606             tx               = fscal*dx00;
 607             ty               = fscal*dy00;
 608             tz               = fscal*dz00;
 609
 610             /* Update vectorial force */
 611             fix0            += tx;
 612             fiy0            += ty;
 613             fiz0            += tz;
 614             f[j_coord_offset+DIM*0+XX] -= tx;
 615             f[j_coord_offset+DIM*0+YY] -= ty;
 616             f[j_coord_offset+DIM*0+ZZ] -= tz;
 617
 618             /**************************
 619              * CALCULATE INTERACTIONS *
 620              **************************/
 621
 622             qq10             = iq1*jq0;
 623
 624             /* REACTION-FIELD ELECTROSTATICS */
 625             felec            = qq10*(rinv10*rinvsq10-krf2);
 626
 627             fscal            = felec;
 628
 629             /* Calculate temporary vectorial force */
 630             tx               = fscal*dx10;
 631             ty               = fscal*dy10;
 632             tz               = fscal*dz10;
 633
 634             /* Update vectorial force */
 635             fix1            += tx;
 636             fiy1            += ty;
 637             fiz1            += tz;
 638             f[j_coord_offset+DIM*0+XX] -= tx;
 639             f[j_coord_offset+DIM*0+YY] -= ty;
 640             f[j_coord_offset+DIM*0+ZZ] -= tz;
 641
 642             /**************************
 643              * CALCULATE INTERACTIONS *
 644              **************************/
 645
 646             qq20             = iq2*jq0;
 647
 648             /* REACTION-FIELD ELECTROSTATICS */
 649             felec            = qq20*(rinv20*rinvsq20-krf2);
 650
 651             fscal            = felec;
 652
 653             /* Calculate temporary vectorial force */
 654             tx               = fscal*dx20;
 655             ty               = fscal*dy20;
 656             tz               = fscal*dz20;
 657
 658             /* Update vectorial force */
 659             fix2            += tx;
 660             fiy2            += ty;
 661             fiz2            += tz;
 662             f[j_coord_offset+DIM*0+XX] -= tx;
 663             f[j_coord_offset+DIM*0+YY] -= ty;
 664             f[j_coord_offset+DIM*0+ZZ] -= tz;
 665
 666             /**************************
 667              * CALCULATE INTERACTIONS *
 668              **************************/
 669
 670             qq30             = iq3*jq0;
 671
 672             /* REACTION-FIELD ELECTROSTATICS */
 673             felec            = qq30*(rinv30*rinvsq30-krf2);
 674
 675             fscal            = felec;
 676
 677             /* Calculate temporary vectorial force */
 678             tx               = fscal*dx30;
 679             ty               = fscal*dy30;
 680             tz               = fscal*dz30;
 681
 682             /* Update vectorial force */
 683             fix3            += tx;
 684             fiy3            += ty;
 685             fiz3            += tz;
 686             f[j_coord_offset+DIM*0+XX] -= tx;
 687             f[j_coord_offset+DIM*0+YY] -= ty;
 688             f[j_coord_offset+DIM*0+ZZ] -= tz;
 689
 690             /* Inner loop uses 128 flops */
 691         }
 692         /* End of innermost loop */
 693
 694         tx = ty = tz = 0;
 695         f[i_coord_offset+DIM*0+XX] += fix0;
 696         f[i_coord_offset+DIM*0+YY] += fiy0;
 697         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 698         tx                         += fix0;
 699         ty                         += fiy0;
 700         tz                         += fiz0;
 701         f[i_coord_offset+DIM*1+XX] += fix1;
 702         f[i_coord_offset+DIM*1+YY] += fiy1;
 703         f[i_coord_offset+DIM*1+ZZ] += fiz1;
 704         tx                         += fix1;
 705         ty                         += fiy1;
 706         tz                         += fiz1;
 707         f[i_coord_offset+DIM*2+XX] += fix2;
 708         f[i_coord_offset+DIM*2+YY] += fiy2;
 709         f[i_coord_offset+DIM*2+ZZ] += fiz2;
 710         tx                         += fix2;
 711         ty                         += fiy2;
 712         tz                         += fiz2;
 713         f[i_coord_offset+DIM*3+XX] += fix3;
 714         f[i_coord_offset+DIM*3+YY] += fiy3;
 715         f[i_coord_offset+DIM*3+ZZ] += fiz3;
 716         tx                         += fix3;
 717         ty                         += fiy3;
 718         tz                         += fiz3;
 719         fshift[i_shift_offset+XX]  += tx;
 720         fshift[i_shift_offset+YY]  += ty;
 721         fshift[i_shift_offset+ZZ]  += tz;
 722
 723         /* Increment number of inner iterations */
 724         inneriter                  += j_index_end - j_index_start;
 725
 726         /* Outer loop uses 39 flops */
 727     }
 728
 729     /* Increment number of outer iterations */
 730     outeriter        += nri;
 731
 732     /* Update outer/inner flops */
 733
 734     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_F,outeriter*39 + inneriter*128);
 735 }