src/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecCSTab_VdwLJ_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_ElecCSTab_VdwLJ_GeomW4P1_VF_c
  35  * Electrostatics interaction: CubicSplineTable
  36  * VdW interaction:            LennardJones
  37  * Geometry:                   Water4-Particle
  38  * Calculate force/pot:        PotentialAndForce
  39  */
  40 void
  41 nb_kernel_ElecCSTab_VdwLJ_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     nvdwtype         = fr->ntype;
  94     vdwparam         = fr->nbfp;
  95     vdwtype          = mdatoms->typeA;
  96
  97     vftab            = kernel_data->table_elec->data;
  98     vftabscale       = kernel_data->table_elec->scale;
  99
 100     /* Setup water-specific parameters */
 101     inr              = nlist->iinr[0];
 102     iq1              = facel*charge[inr+1];
 103     iq2              = facel*charge[inr+2];
 104     iq3              = facel*charge[inr+3];
 105     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 106
 107     outeriter        = 0;
 108     inneriter        = 0;
 109
 110     /* Start outer loop over neighborlists */
 111     for(iidx=0; iidx<nri; iidx++)
 112     {
 113         /* Load shift vector for this list */
 114         i_shift_offset   = DIM*shiftidx[iidx];
 115         shX              = shiftvec[i_shift_offset+XX];
 116         shY              = shiftvec[i_shift_offset+YY];
 117         shZ              = shiftvec[i_shift_offset+ZZ];
 118
 119         /* Load limits for loop over neighbors */
 120         j_index_start    = jindex[iidx];
 121         j_index_end      = jindex[iidx+1];
 122
 123         /* Get outer coordinate index */
 124         inr              = iinr[iidx];
 125         i_coord_offset   = DIM*inr;
 126
 127         /* Load i particle coords and add shift vector */
 128         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 129         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 130         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 131         ix1              = shX + x[i_coord_offset+DIM*1+XX];
 132         iy1              = shY + x[i_coord_offset+DIM*1+YY];
 133         iz1              = shZ + x[i_coord_offset+DIM*1+ZZ];
 134         ix2              = shX + x[i_coord_offset+DIM*2+XX];
 135         iy2              = shY + x[i_coord_offset+DIM*2+YY];
 136         iz2              = shZ + x[i_coord_offset+DIM*2+ZZ];
 137         ix3              = shX + x[i_coord_offset+DIM*3+XX];
 138         iy3              = shY + x[i_coord_offset+DIM*3+YY];
 139         iz3              = shZ + x[i_coord_offset+DIM*3+ZZ];
 140
 141         fix0             = 0.0;
 142         fiy0             = 0.0;
 143         fiz0             = 0.0;
 144         fix1             = 0.0;
 145         fiy1             = 0.0;
 146         fiz1             = 0.0;
 147         fix2             = 0.0;
 148         fiy2             = 0.0;
 149         fiz2             = 0.0;
 150         fix3             = 0.0;
 151         fiy3             = 0.0;
 152         fiz3             = 0.0;
 153
 154         /* Reset potential sums */
 155         velecsum         = 0.0;
 156         vvdwsum          = 0.0;
 157
 158         /* Start inner kernel loop */
 159         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 160         {
 161             /* Get j neighbor index, and coordinate index */
 162             jnr              = jjnr[jidx];
 163             j_coord_offset   = DIM*jnr;
 164
 165             /* load j atom coordinates */
 166             jx0              = x[j_coord_offset+DIM*0+XX];
 167             jy0              = x[j_coord_offset+DIM*0+YY];
 168             jz0              = x[j_coord_offset+DIM*0+ZZ];
 169
 170             /* Calculate displacement vector */
 171             dx00             = ix0 - jx0;
 172             dy00             = iy0 - jy0;
 173             dz00             = iz0 - jz0;
 174             dx10             = ix1 - jx0;
 175             dy10             = iy1 - jy0;
 176             dz10             = iz1 - jz0;
 177             dx20             = ix2 - jx0;
 178             dy20             = iy2 - jy0;
 179             dz20             = iz2 - jz0;
 180             dx30             = ix3 - jx0;
 181             dy30             = iy3 - jy0;
 182             dz30             = iz3 - jz0;
 183
 184             /* Calculate squared distance and things based on it */
 185             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 186             rsq10            = dx10*dx10+dy10*dy10+dz10*dz10;
 187             rsq20            = dx20*dx20+dy20*dy20+dz20*dz20;
 188             rsq30            = dx30*dx30+dy30*dy30+dz30*dz30;
 189
 190             rinv10           = gmx_invsqrt(rsq10);
 191             rinv20           = gmx_invsqrt(rsq20);
 192             rinv30           = gmx_invsqrt(rsq30);
 193
 194             rinvsq00         = 1.0/rsq00;
 195
 196             /* Load parameters for j particles */
 197             jq0              = charge[jnr+0];
 198             vdwjidx0         = 2*vdwtype[jnr+0];
 199
 200             /**************************
 201              * CALCULATE INTERACTIONS *
 202              **************************/
 203
 204             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 205             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 206
 207             /* LENNARD-JONES DISPERSION/REPULSION */
 208
 209             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 210             vvdw6            = c6_00*rinvsix;
 211             vvdw12           = c12_00*rinvsix*rinvsix;
 212             vvdw             = vvdw12*(1.0/12.0) - vvdw6*(1.0/6.0);
 213             fvdw             = (vvdw12-vvdw6)*rinvsq00;
 214
 215             /* Update potential sums from outer loop */
 216             vvdwsum         += vvdw;
 217
 218             fscal            = fvdw;
 219
 220             /* Calculate temporary vectorial force */
 221             tx               = fscal*dx00;
 222             ty               = fscal*dy00;
 223             tz               = fscal*dz00;
 224
 225             /* Update vectorial force */
 226             fix0            += tx;
 227             fiy0            += ty;
 228             fiz0            += tz;
 229             f[j_coord_offset+DIM*0+XX] -= tx;
 230             f[j_coord_offset+DIM*0+YY] -= ty;
 231             f[j_coord_offset+DIM*0+ZZ] -= tz;
 232
 233             /**************************
 234              * CALCULATE INTERACTIONS *
 235              **************************/
 236
 237             r10              = rsq10*rinv10;
 238
 239             qq10             = iq1*jq0;
 240
 241             /* Calculate table index by multiplying r with table scale and truncate to integer */
 242             rt               = r10*vftabscale;
 243             vfitab           = rt;
 244             vfeps            = rt-vfitab;
 245             vfitab           = 1*4*vfitab;
 246
 247             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 248             Y                = vftab[vfitab];
 249             F                = vftab[vfitab+1];
 250             Geps             = vfeps*vftab[vfitab+2];
 251             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 252             Fp               = F+Geps+Heps2;
 253             VV               = Y+vfeps*Fp;
 254             velec            = qq10*VV;
 255             FF               = Fp+Geps+2.0*Heps2;
 256             felec            = -qq10*FF*vftabscale*rinv10;
 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              * CALCULATE INTERACTIONS *
 278              **************************/
 279
 280             r20              = rsq20*rinv20;
 281
 282             qq20             = iq2*jq0;
 283
 284             /* Calculate table index by multiplying r with table scale and truncate to integer */
 285             rt               = r20*vftabscale;
 286             vfitab           = rt;
 287             vfeps            = rt-vfitab;
 288             vfitab           = 1*4*vfitab;
 289
 290             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 291             Y                = vftab[vfitab];
 292             F                = vftab[vfitab+1];
 293             Geps             = vfeps*vftab[vfitab+2];
 294             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 295             Fp               = F+Geps+Heps2;
 296             VV               = Y+vfeps*Fp;
 297             velec            = qq20*VV;
 298             FF               = Fp+Geps+2.0*Heps2;
 299             felec            = -qq20*FF*vftabscale*rinv20;
 300
 301             /* Update potential sums from outer loop */
 302             velecsum        += velec;
 303
 304             fscal            = felec;
 305
 306             /* Calculate temporary vectorial force */
 307             tx               = fscal*dx20;
 308             ty               = fscal*dy20;
 309             tz               = fscal*dz20;
 310
 311             /* Update vectorial force */
 312             fix2            += tx;
 313             fiy2            += ty;
 314             fiz2            += tz;
 315             f[j_coord_offset+DIM*0+XX] -= tx;
 316             f[j_coord_offset+DIM*0+YY] -= ty;
 317             f[j_coord_offset+DIM*0+ZZ] -= tz;
 318
 319             /**************************
 320              * CALCULATE INTERACTIONS *
 321              **************************/
 322
 323             r30              = rsq30*rinv30;
 324
 325             qq30             = iq3*jq0;
 326
 327             /* Calculate table index by multiplying r with table scale and truncate to integer */
 328             rt               = r30*vftabscale;
 329             vfitab           = rt;
 330             vfeps            = rt-vfitab;
 331             vfitab           = 1*4*vfitab;
 332
 333             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 334             Y                = vftab[vfitab];
 335             F                = vftab[vfitab+1];
 336             Geps             = vfeps*vftab[vfitab+2];
 337             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 338             Fp               = F+Geps+Heps2;
 339             VV               = Y+vfeps*Fp;
 340             velec            = qq30*VV;
 341             FF               = Fp+Geps+2.0*Heps2;
 342             felec            = -qq30*FF*vftabscale*rinv30;
 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             /* Inner loop uses 158 flops */
 363         }
 364         /* End of innermost loop */
 365
 366         tx = ty = tz = 0;
 367         f[i_coord_offset+DIM*0+XX] += fix0;
 368         f[i_coord_offset+DIM*0+YY] += fiy0;
 369         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 370         tx                         += fix0;
 371         ty                         += fiy0;
 372         tz                         += fiz0;
 373         f[i_coord_offset+DIM*1+XX] += fix1;
 374         f[i_coord_offset+DIM*1+YY] += fiy1;
 375         f[i_coord_offset+DIM*1+ZZ] += fiz1;
 376         tx                         += fix1;
 377         ty                         += fiy1;
 378         tz                         += fiz1;
 379         f[i_coord_offset+DIM*2+XX] += fix2;
 380         f[i_coord_offset+DIM*2+YY] += fiy2;
 381         f[i_coord_offset+DIM*2+ZZ] += fiz2;
 382         tx                         += fix2;
 383         ty                         += fiy2;
 384         tz                         += fiz2;
 385         f[i_coord_offset+DIM*3+XX] += fix3;
 386         f[i_coord_offset+DIM*3+YY] += fiy3;
 387         f[i_coord_offset+DIM*3+ZZ] += fiz3;
 388         tx                         += fix3;
 389         ty                         += fiy3;
 390         tz                         += fiz3;
 391         fshift[i_shift_offset+XX]  += tx;
 392         fshift[i_shift_offset+YY]  += ty;
 393         fshift[i_shift_offset+ZZ]  += tz;
 394
 395         ggid                        = gid[iidx];
 396         /* Update potential energies */
 397         kernel_data->energygrp_elec[ggid] += velecsum;
 398         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 399
 400         /* Increment number of inner iterations */
 401         inneriter                  += j_index_end - j_index_start;
 402
 403         /* Outer loop uses 41 flops */
 404     }
 405
 406     /* Increment number of outer iterations */
 407     outeriter        += nri;
 408
 409     /* Update outer/inner flops */
 410
 411     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_VF,outeriter*41 + inneriter*158);
 412 }
 413 /*
 414  * Gromacs nonbonded kernel:   nb_kernel_ElecCSTab_VdwLJ_GeomW4P1_F_c
 415  * Electrostatics interaction: CubicSplineTable
 416  * VdW interaction:            LennardJones
 417  * Geometry:                   Water4-Particle
 418  * Calculate force/pot:        Force
 419  */
 420 void
 421 nb_kernel_ElecCSTab_VdwLJ_GeomW4P1_F_c
 422                     (t_nblist * gmx_restrict                nlist,
 423                      rvec * gmx_restrict                    xx,
 424                      rvec * gmx_restrict                    ff,
 425                      t_forcerec * gmx_restrict              fr,
 426                      t_mdatoms * gmx_restrict               mdatoms,
 427                      nb_kernel_data_t * gmx_restrict        kernel_data,
 428                      t_nrnb * gmx_restrict                  nrnb)
 429 {
 430     int              i_shift_offset,i_coord_offset,j_coord_offset;
 431     int              j_index_start,j_index_end;
 432     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 433     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 434     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 435     real             *shiftvec,*fshift,*x,*f;
 436     int              vdwioffset0;
 437     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 438     int              vdwioffset1;
 439     real             ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
 440     int              vdwioffset2;
 441     real             ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
 442     int              vdwioffset3;
 443     real             ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
 444     int              vdwjidx0;
 445     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 446     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 447     real             dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10;
 448     real             dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20;
 449     real             dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30,cexp1_30,cexp2_30;
 450     real             velec,felec,velecsum,facel,crf,krf,krf2;
 451     real             *charge;
 452     int              nvdwtype;
 453     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 454     int              *vdwtype;
 455     real             *vdwparam;
 456     int              vfitab;
 457     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
 458     real             *vftab;
 459
 460     x                = xx[0];
 461     f                = ff[0];
 462
 463     nri              = nlist->nri;
 464     iinr             = nlist->iinr;
 465     jindex           = nlist->jindex;
 466     jjnr             = nlist->jjnr;
 467     shiftidx         = nlist->shift;
 468     gid              = nlist->gid;
 469     shiftvec         = fr->shift_vec[0];
 470     fshift           = fr->fshift[0];
 471     facel            = fr->epsfac;
 472     charge           = mdatoms->chargeA;
 473     nvdwtype         = fr->ntype;
 474     vdwparam         = fr->nbfp;
 475     vdwtype          = mdatoms->typeA;
 476
 477     vftab            = kernel_data->table_elec->data;
 478     vftabscale       = kernel_data->table_elec->scale;
 479
 480     /* Setup water-specific parameters */
 481     inr              = nlist->iinr[0];
 482     iq1              = facel*charge[inr+1];
 483     iq2              = facel*charge[inr+2];
 484     iq3              = facel*charge[inr+3];
 485     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 486
 487     outeriter        = 0;
 488     inneriter        = 0;
 489
 490     /* Start outer loop over neighborlists */
 491     for(iidx=0; iidx<nri; iidx++)
 492     {
 493         /* Load shift vector for this list */
 494         i_shift_offset   = DIM*shiftidx[iidx];
 495         shX              = shiftvec[i_shift_offset+XX];
 496         shY              = shiftvec[i_shift_offset+YY];
 497         shZ              = shiftvec[i_shift_offset+ZZ];
 498
 499         /* Load limits for loop over neighbors */
 500         j_index_start    = jindex[iidx];
 501         j_index_end      = jindex[iidx+1];
 502
 503         /* Get outer coordinate index */
 504         inr              = iinr[iidx];
 505         i_coord_offset   = DIM*inr;
 506
 507         /* Load i particle coords and add shift vector */
 508         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 509         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 510         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 511         ix1              = shX + x[i_coord_offset+DIM*1+XX];
 512         iy1              = shY + x[i_coord_offset+DIM*1+YY];
 513         iz1              = shZ + x[i_coord_offset+DIM*1+ZZ];
 514         ix2              = shX + x[i_coord_offset+DIM*2+XX];
 515         iy2              = shY + x[i_coord_offset+DIM*2+YY];
 516         iz2              = shZ + x[i_coord_offset+DIM*2+ZZ];
 517         ix3              = shX + x[i_coord_offset+DIM*3+XX];
 518         iy3              = shY + x[i_coord_offset+DIM*3+YY];
 519         iz3              = shZ + x[i_coord_offset+DIM*3+ZZ];
 520
 521         fix0             = 0.0;
 522         fiy0             = 0.0;
 523         fiz0             = 0.0;
 524         fix1             = 0.0;
 525         fiy1             = 0.0;
 526         fiz1             = 0.0;
 527         fix2             = 0.0;
 528         fiy2             = 0.0;
 529         fiz2             = 0.0;
 530         fix3             = 0.0;
 531         fiy3             = 0.0;
 532         fiz3             = 0.0;
 533
 534         /* Start inner kernel loop */
 535         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 536         {
 537             /* Get j neighbor index, and coordinate index */
 538             jnr              = jjnr[jidx];
 539             j_coord_offset   = DIM*jnr;
 540
 541             /* load j atom coordinates */
 542             jx0              = x[j_coord_offset+DIM*0+XX];
 543             jy0              = x[j_coord_offset+DIM*0+YY];
 544             jz0              = x[j_coord_offset+DIM*0+ZZ];
 545
 546             /* Calculate displacement vector */
 547             dx00             = ix0 - jx0;
 548             dy00             = iy0 - jy0;
 549             dz00             = iz0 - jz0;
 550             dx10             = ix1 - jx0;
 551             dy10             = iy1 - jy0;
 552             dz10             = iz1 - jz0;
 553             dx20             = ix2 - jx0;
 554             dy20             = iy2 - jy0;
 555             dz20             = iz2 - jz0;
 556             dx30             = ix3 - jx0;
 557             dy30             = iy3 - jy0;
 558             dz30             = iz3 - jz0;
 559
 560             /* Calculate squared distance and things based on it */
 561             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 562             rsq10            = dx10*dx10+dy10*dy10+dz10*dz10;
 563             rsq20            = dx20*dx20+dy20*dy20+dz20*dz20;
 564             rsq30            = dx30*dx30+dy30*dy30+dz30*dz30;
 565
 566             rinv10           = gmx_invsqrt(rsq10);
 567             rinv20           = gmx_invsqrt(rsq20);
 568             rinv30           = gmx_invsqrt(rsq30);
 569
 570             rinvsq00         = 1.0/rsq00;
 571
 572             /* Load parameters for j particles */
 573             jq0              = charge[jnr+0];
 574             vdwjidx0         = 2*vdwtype[jnr+0];
 575
 576             /**************************
 577              * CALCULATE INTERACTIONS *
 578              **************************/
 579
 580             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 581             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 582
 583             /* LENNARD-JONES DISPERSION/REPULSION */
 584
 585             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 586             fvdw             = (c12_00*rinvsix-c6_00)*rinvsix*rinvsq00;
 587
 588             fscal            = fvdw;
 589
 590             /* Calculate temporary vectorial force */
 591             tx               = fscal*dx00;
 592             ty               = fscal*dy00;
 593             tz               = fscal*dz00;
 594
 595             /* Update vectorial force */
 596             fix0            += tx;
 597             fiy0            += ty;
 598             fiz0            += tz;
 599             f[j_coord_offset+DIM*0+XX] -= tx;
 600             f[j_coord_offset+DIM*0+YY] -= ty;
 601             f[j_coord_offset+DIM*0+ZZ] -= tz;
 602
 603             /**************************
 604              * CALCULATE INTERACTIONS *
 605              **************************/
 606
 607             r10              = rsq10*rinv10;
 608
 609             qq10             = iq1*jq0;
 610
 611             /* Calculate table index by multiplying r with table scale and truncate to integer */
 612             rt               = r10*vftabscale;
 613             vfitab           = rt;
 614             vfeps            = rt-vfitab;
 615             vfitab           = 1*4*vfitab;
 616
 617             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 618             Y                = vftab[vfitab];
 619             F                = vftab[vfitab+1];
 620             Geps             = vfeps*vftab[vfitab+2];
 621             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 622             Fp               = F+Geps+Heps2;
 623             FF               = Fp+Geps+2.0*Heps2;
 624             felec            = -qq10*FF*vftabscale*rinv10;
 625
 626             fscal            = felec;
 627
 628             /* Calculate temporary vectorial force */
 629             tx               = fscal*dx10;
 630             ty               = fscal*dy10;
 631             tz               = fscal*dz10;
 632
 633             /* Update vectorial force */
 634             fix1            += tx;
 635             fiy1            += ty;
 636             fiz1            += tz;
 637             f[j_coord_offset+DIM*0+XX] -= tx;
 638             f[j_coord_offset+DIM*0+YY] -= ty;
 639             f[j_coord_offset+DIM*0+ZZ] -= tz;
 640
 641             /**************************
 642              * CALCULATE INTERACTIONS *
 643              **************************/
 644
 645             r20              = rsq20*rinv20;
 646
 647             qq20             = iq2*jq0;
 648
 649             /* Calculate table index by multiplying r with table scale and truncate to integer */
 650             rt               = r20*vftabscale;
 651             vfitab           = rt;
 652             vfeps            = rt-vfitab;
 653             vfitab           = 1*4*vfitab;
 654
 655             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 656             Y                = vftab[vfitab];
 657             F                = vftab[vfitab+1];
 658             Geps             = vfeps*vftab[vfitab+2];
 659             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 660             Fp               = F+Geps+Heps2;
 661             FF               = Fp+Geps+2.0*Heps2;
 662             felec            = -qq20*FF*vftabscale*rinv20;
 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              * CALCULATE INTERACTIONS *
 681              **************************/
 682
 683             r30              = rsq30*rinv30;
 684
 685             qq30             = iq3*jq0;
 686
 687             /* Calculate table index by multiplying r with table scale and truncate to integer */
 688             rt               = r30*vftabscale;
 689             vfitab           = rt;
 690             vfeps            = rt-vfitab;
 691             vfitab           = 1*4*vfitab;
 692
 693             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 694             Y                = vftab[vfitab];
 695             F                = vftab[vfitab+1];
 696             Geps             = vfeps*vftab[vfitab+2];
 697             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 698             Fp               = F+Geps+Heps2;
 699             FF               = Fp+Geps+2.0*Heps2;
 700             felec            = -qq30*FF*vftabscale*rinv30;
 701
 702             fscal            = felec;
 703
 704             /* Calculate temporary vectorial force */
 705             tx               = fscal*dx30;
 706             ty               = fscal*dy30;
 707             tz               = fscal*dz30;
 708
 709             /* Update vectorial force */
 710             fix3            += tx;
 711             fiy3            += ty;
 712             fiz3            += tz;
 713             f[j_coord_offset+DIM*0+XX] -= tx;
 714             f[j_coord_offset+DIM*0+YY] -= ty;
 715             f[j_coord_offset+DIM*0+ZZ] -= tz;
 716
 717             /* Inner loop uses 141 flops */
 718         }
 719         /* End of innermost loop */
 720
 721         tx = ty = tz = 0;
 722         f[i_coord_offset+DIM*0+XX] += fix0;
 723         f[i_coord_offset+DIM*0+YY] += fiy0;
 724         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 725         tx                         += fix0;
 726         ty                         += fiy0;
 727         tz                         += fiz0;
 728         f[i_coord_offset+DIM*1+XX] += fix1;
 729         f[i_coord_offset+DIM*1+YY] += fiy1;
 730         f[i_coord_offset+DIM*1+ZZ] += fiz1;
 731         tx                         += fix1;
 732         ty                         += fiy1;
 733         tz                         += fiz1;
 734         f[i_coord_offset+DIM*2+XX] += fix2;
 735         f[i_coord_offset+DIM*2+YY] += fiy2;
 736         f[i_coord_offset+DIM*2+ZZ] += fiz2;
 737         tx                         += fix2;
 738         ty                         += fiy2;
 739         tz                         += fiz2;
 740         f[i_coord_offset+DIM*3+XX] += fix3;
 741         f[i_coord_offset+DIM*3+YY] += fiy3;
 742         f[i_coord_offset+DIM*3+ZZ] += fiz3;
 743         tx                         += fix3;
 744         ty                         += fiy3;
 745         tz                         += fiz3;
 746         fshift[i_shift_offset+XX]  += tx;
 747         fshift[i_shift_offset+YY]  += ty;
 748         fshift[i_shift_offset+ZZ]  += tz;
 749
 750         /* Increment number of inner iterations */
 751         inneriter                  += j_index_end - j_index_start;
 752
 753         /* Outer loop uses 39 flops */
 754     }
 755
 756     /* Increment number of outer iterations */
 757     outeriter        += nri;
 758
 759     /* Update outer/inner flops */
 760
 761     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_F,outeriter*39 + inneriter*141);
 762 }