src/gromacs/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecCSTab_VdwCSTab_GeomW4P1_c.c

   1 /*
   2  * This file is part of the GROMACS molecular simulation package.
   3  *
   4  * Copyright (c) 2012,2013,2014, by the GROMACS development team, led by
   5  * Mark Abraham, David van der Spoel, Berk Hess, and Erik Lindahl,
   6  * and including many others, as listed in the AUTHORS file in the
   7  * top-level source directory and at http://www.gromacs.org.
   8  *
   9  * GROMACS is free software; you can redistribute it and/or
  10  * modify it under the terms of the GNU Lesser General Public License
  11  * as published by the Free Software Foundation; either version 2.1
  12  * of the License, or (at your option) any later version.
  13  *
  14  * GROMACS is distributed in the hope that it will be useful,
  15  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  16  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  17  * Lesser General Public License for more details.
  18  *
  19  * You should have received a copy of the GNU Lesser General Public
  20  * License along with GROMACS; if not, see
  21  * http://www.gnu.org/licenses, or write to the Free Software Foundation,
  22  * Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA.
  23  *
  24  * If you want to redistribute modifications to GROMACS, please
  25  * consider that scientific software is very special. Version
  26  * control is crucial - bugs must be traceable. We will be happy to
  27  * consider code for inclusion in the official distribution, but
  28  * derived work must not be called official GROMACS. Details are found
  29  * in the README & COPYING files - if they are missing, get the
  30  * official version at http://www.gromacs.org.
  31  *
  32  * To help us fund GROMACS development, we humbly ask that you cite
  33  * the research papers on the package. Check out http://www.gromacs.org.
  34  */
  35 /*
  36  * Note: this file was generated by the GROMACS c kernel generator.
  37  */
  38 #include "config.h"
  39
  40 #include <math.h>
  41
  42 #include "../nb_kernel.h"
  43 #include "types/simple.h"
  44 #include "gromacs/math/vec.h"
  45 #include "nrnb.h"
  46
  47 /*
  48  * Gromacs nonbonded kernel:   nb_kernel_ElecCSTab_VdwCSTab_GeomW4P1_VF_c
  49  * Electrostatics interaction: CubicSplineTable
  50  * VdW interaction:            CubicSplineTable
  51  * Geometry:                   Water4-Particle
  52  * Calculate force/pot:        PotentialAndForce
  53  */
  54 void
  55 nb_kernel_ElecCSTab_VdwCSTab_GeomW4P1_VF_c
  56                     (t_nblist                    * gmx_restrict       nlist,
  57                      rvec                        * gmx_restrict          xx,
  58                      rvec                        * gmx_restrict          ff,
  59                      t_forcerec                  * gmx_restrict          fr,
  60                      t_mdatoms                   * gmx_restrict     mdatoms,
  61                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
  62                      t_nrnb                      * gmx_restrict        nrnb)
  63 {
  64     int              i_shift_offset,i_coord_offset,j_coord_offset;
  65     int              j_index_start,j_index_end;
  66     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
  67     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
  68     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  69     real             *shiftvec,*fshift,*x,*f;
  70     int              vdwioffset0;
  71     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  72     int              vdwioffset1;
  73     real             ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
  74     int              vdwioffset2;
  75     real             ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
  76     int              vdwioffset3;
  77     real             ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
  78     int              vdwjidx0;
  79     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  80     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
  81     real             dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10;
  82     real             dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20;
  83     real             dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30,cexp1_30,cexp2_30;
  84     real             velec,felec,velecsum,facel,crf,krf,krf2;
  85     real             *charge;
  86     int              nvdwtype;
  87     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
  88     int              *vdwtype;
  89     real             *vdwparam;
  90     int              vfitab;
  91     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
  92     real             *vftab;
  93
  94     x                = xx[0];
  95     f                = ff[0];
  96
  97     nri              = nlist->nri;
  98     iinr             = nlist->iinr;
  99     jindex           = nlist->jindex;
 100     jjnr             = nlist->jjnr;
 101     shiftidx         = nlist->shift;
 102     gid              = nlist->gid;
 103     shiftvec         = fr->shift_vec[0];
 104     fshift           = fr->fshift[0];
 105     facel            = fr->epsfac;
 106     charge           = mdatoms->chargeA;
 107     nvdwtype         = fr->ntype;
 108     vdwparam         = fr->nbfp;
 109     vdwtype          = mdatoms->typeA;
 110
 111     vftab            = kernel_data->table_elec_vdw->data;
 112     vftabscale       = kernel_data->table_elec_vdw->scale;
 113
 114     /* Setup water-specific parameters */
 115     inr              = nlist->iinr[0];
 116     iq1              = facel*charge[inr+1];
 117     iq2              = facel*charge[inr+2];
 118     iq3              = facel*charge[inr+3];
 119     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 120
 121     outeriter        = 0;
 122     inneriter        = 0;
 123
 124     /* Start outer loop over neighborlists */
 125     for(iidx=0; iidx<nri; iidx++)
 126     {
 127         /* Load shift vector for this list */
 128         i_shift_offset   = DIM*shiftidx[iidx];
 129         shX              = shiftvec[i_shift_offset+XX];
 130         shY              = shiftvec[i_shift_offset+YY];
 131         shZ              = shiftvec[i_shift_offset+ZZ];
 132
 133         /* Load limits for loop over neighbors */
 134         j_index_start    = jindex[iidx];
 135         j_index_end      = jindex[iidx+1];
 136
 137         /* Get outer coordinate index */
 138         inr              = iinr[iidx];
 139         i_coord_offset   = DIM*inr;
 140
 141         /* Load i particle coords and add shift vector */
 142         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 143         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 144         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 145         ix1              = shX + x[i_coord_offset+DIM*1+XX];
 146         iy1              = shY + x[i_coord_offset+DIM*1+YY];
 147         iz1              = shZ + x[i_coord_offset+DIM*1+ZZ];
 148         ix2              = shX + x[i_coord_offset+DIM*2+XX];
 149         iy2              = shY + x[i_coord_offset+DIM*2+YY];
 150         iz2              = shZ + x[i_coord_offset+DIM*2+ZZ];
 151         ix3              = shX + x[i_coord_offset+DIM*3+XX];
 152         iy3              = shY + x[i_coord_offset+DIM*3+YY];
 153         iz3              = shZ + x[i_coord_offset+DIM*3+ZZ];
 154
 155         fix0             = 0.0;
 156         fiy0             = 0.0;
 157         fiz0             = 0.0;
 158         fix1             = 0.0;
 159         fiy1             = 0.0;
 160         fiz1             = 0.0;
 161         fix2             = 0.0;
 162         fiy2             = 0.0;
 163         fiz2             = 0.0;
 164         fix3             = 0.0;
 165         fiy3             = 0.0;
 166         fiz3             = 0.0;
 167
 168         /* Reset potential sums */
 169         velecsum         = 0.0;
 170         vvdwsum          = 0.0;
 171
 172         /* Start inner kernel loop */
 173         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 174         {
 175             /* Get j neighbor index, and coordinate index */
 176             jnr              = jjnr[jidx];
 177             j_coord_offset   = DIM*jnr;
 178
 179             /* load j atom coordinates */
 180             jx0              = x[j_coord_offset+DIM*0+XX];
 181             jy0              = x[j_coord_offset+DIM*0+YY];
 182             jz0              = x[j_coord_offset+DIM*0+ZZ];
 183
 184             /* Calculate displacement vector */
 185             dx00             = ix0 - jx0;
 186             dy00             = iy0 - jy0;
 187             dz00             = iz0 - jz0;
 188             dx10             = ix1 - jx0;
 189             dy10             = iy1 - jy0;
 190             dz10             = iz1 - jz0;
 191             dx20             = ix2 - jx0;
 192             dy20             = iy2 - jy0;
 193             dz20             = iz2 - jz0;
 194             dx30             = ix3 - jx0;
 195             dy30             = iy3 - jy0;
 196             dz30             = iz3 - jz0;
 197
 198             /* Calculate squared distance and things based on it */
 199             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 200             rsq10            = dx10*dx10+dy10*dy10+dz10*dz10;
 201             rsq20            = dx20*dx20+dy20*dy20+dz20*dz20;
 202             rsq30            = dx30*dx30+dy30*dy30+dz30*dz30;
 203
 204             rinv00           = gmx_invsqrt(rsq00);
 205             rinv10           = gmx_invsqrt(rsq10);
 206             rinv20           = gmx_invsqrt(rsq20);
 207             rinv30           = gmx_invsqrt(rsq30);
 208
 209             /* Load parameters for j particles */
 210             jq0              = charge[jnr+0];
 211             vdwjidx0         = 2*vdwtype[jnr+0];
 212
 213             /**************************
 214              * CALCULATE INTERACTIONS *
 215              **************************/
 216
 217             r00              = rsq00*rinv00;
 218
 219             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 220             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 221
 222             /* Calculate table index by multiplying r with table scale and truncate to integer */
 223             rt               = r00*vftabscale;
 224             vfitab           = rt;
 225             vfeps            = rt-vfitab;
 226             vfitab           = 3*4*vfitab;
 227
 228             /* CUBIC SPLINE TABLE DISPERSION */
 229             vfitab          += 4;
 230             Y                = vftab[vfitab];
 231             F                = vftab[vfitab+1];
 232             Geps             = vfeps*vftab[vfitab+2];
 233             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 234             Fp               = F+Geps+Heps2;
 235             VV               = Y+vfeps*Fp;
 236             vvdw6            = c6_00*VV;
 237             FF               = Fp+Geps+2.0*Heps2;
 238             fvdw6            = c6_00*FF;
 239
 240             /* CUBIC SPLINE TABLE REPULSION */
 241             Y                = vftab[vfitab+4];
 242             F                = vftab[vfitab+5];
 243             Geps             = vfeps*vftab[vfitab+6];
 244             Heps2            = vfeps*vfeps*vftab[vfitab+7];
 245             Fp               = F+Geps+Heps2;
 246             VV               = Y+vfeps*Fp;
 247             vvdw12           = c12_00*VV;
 248             FF               = Fp+Geps+2.0*Heps2;
 249             fvdw12           = c12_00*FF;
 250             vvdw             = vvdw12+vvdw6;
 251             fvdw             = -(fvdw6+fvdw12)*vftabscale*rinv00;
 252
 253             /* Update potential sums from outer loop */
 254             vvdwsum         += vvdw;
 255
 256             fscal            = fvdw;
 257
 258             /* Calculate temporary vectorial force */
 259             tx               = fscal*dx00;
 260             ty               = fscal*dy00;
 261             tz               = fscal*dz00;
 262
 263             /* Update vectorial force */
 264             fix0            += tx;
 265             fiy0            += ty;
 266             fiz0            += tz;
 267             f[j_coord_offset+DIM*0+XX] -= tx;
 268             f[j_coord_offset+DIM*0+YY] -= ty;
 269             f[j_coord_offset+DIM*0+ZZ] -= tz;
 270
 271             /**************************
 272              * CALCULATE INTERACTIONS *
 273              **************************/
 274
 275             r10              = rsq10*rinv10;
 276
 277             qq10             = iq1*jq0;
 278
 279             /* Calculate table index by multiplying r with table scale and truncate to integer */
 280             rt               = r10*vftabscale;
 281             vfitab           = rt;
 282             vfeps            = rt-vfitab;
 283             vfitab           = 3*4*vfitab;
 284
 285             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 286             Y                = vftab[vfitab];
 287             F                = vftab[vfitab+1];
 288             Geps             = vfeps*vftab[vfitab+2];
 289             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 290             Fp               = F+Geps+Heps2;
 291             VV               = Y+vfeps*Fp;
 292             velec            = qq10*VV;
 293             FF               = Fp+Geps+2.0*Heps2;
 294             felec            = -qq10*FF*vftabscale*rinv10;
 295
 296             /* Update potential sums from outer loop */
 297             velecsum        += velec;
 298
 299             fscal            = felec;
 300
 301             /* Calculate temporary vectorial force */
 302             tx               = fscal*dx10;
 303             ty               = fscal*dy10;
 304             tz               = fscal*dz10;
 305
 306             /* Update vectorial force */
 307             fix1            += tx;
 308             fiy1            += ty;
 309             fiz1            += tz;
 310             f[j_coord_offset+DIM*0+XX] -= tx;
 311             f[j_coord_offset+DIM*0+YY] -= ty;
 312             f[j_coord_offset+DIM*0+ZZ] -= tz;
 313
 314             /**************************
 315              * CALCULATE INTERACTIONS *
 316              **************************/
 317
 318             r20              = rsq20*rinv20;
 319
 320             qq20             = iq2*jq0;
 321
 322             /* Calculate table index by multiplying r with table scale and truncate to integer */
 323             rt               = r20*vftabscale;
 324             vfitab           = rt;
 325             vfeps            = rt-vfitab;
 326             vfitab           = 3*4*vfitab;
 327
 328             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 329             Y                = vftab[vfitab];
 330             F                = vftab[vfitab+1];
 331             Geps             = vfeps*vftab[vfitab+2];
 332             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 333             Fp               = F+Geps+Heps2;
 334             VV               = Y+vfeps*Fp;
 335             velec            = qq20*VV;
 336             FF               = Fp+Geps+2.0*Heps2;
 337             felec            = -qq20*FF*vftabscale*rinv20;
 338
 339             /* Update potential sums from outer loop */
 340             velecsum        += velec;
 341
 342             fscal            = felec;
 343
 344             /* Calculate temporary vectorial force */
 345             tx               = fscal*dx20;
 346             ty               = fscal*dy20;
 347             tz               = fscal*dz20;
 348
 349             /* Update vectorial force */
 350             fix2            += tx;
 351             fiy2            += ty;
 352             fiz2            += tz;
 353             f[j_coord_offset+DIM*0+XX] -= tx;
 354             f[j_coord_offset+DIM*0+YY] -= ty;
 355             f[j_coord_offset+DIM*0+ZZ] -= tz;
 356
 357             /**************************
 358              * CALCULATE INTERACTIONS *
 359              **************************/
 360
 361             r30              = rsq30*rinv30;
 362
 363             qq30             = iq3*jq0;
 364
 365             /* Calculate table index by multiplying r with table scale and truncate to integer */
 366             rt               = r30*vftabscale;
 367             vfitab           = rt;
 368             vfeps            = rt-vfitab;
 369             vfitab           = 3*4*vfitab;
 370
 371             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 372             Y                = vftab[vfitab];
 373             F                = vftab[vfitab+1];
 374             Geps             = vfeps*vftab[vfitab+2];
 375             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 376             Fp               = F+Geps+Heps2;
 377             VV               = Y+vfeps*Fp;
 378             velec            = qq30*VV;
 379             FF               = Fp+Geps+2.0*Heps2;
 380             felec            = -qq30*FF*vftabscale*rinv30;
 381
 382             /* Update potential sums from outer loop */
 383             velecsum        += velec;
 384
 385             fscal            = felec;
 386
 387             /* Calculate temporary vectorial force */
 388             tx               = fscal*dx30;
 389             ty               = fscal*dy30;
 390             tz               = fscal*dz30;
 391
 392             /* Update vectorial force */
 393             fix3            += tx;
 394             fiy3            += ty;
 395             fiz3            += tz;
 396             f[j_coord_offset+DIM*0+XX] -= tx;
 397             f[j_coord_offset+DIM*0+YY] -= ty;
 398             f[j_coord_offset+DIM*0+ZZ] -= tz;
 399
 400             /* Inner loop uses 181 flops */
 401         }
 402         /* End of innermost loop */
 403
 404         tx = ty = tz = 0;
 405         f[i_coord_offset+DIM*0+XX] += fix0;
 406         f[i_coord_offset+DIM*0+YY] += fiy0;
 407         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 408         tx                         += fix0;
 409         ty                         += fiy0;
 410         tz                         += fiz0;
 411         f[i_coord_offset+DIM*1+XX] += fix1;
 412         f[i_coord_offset+DIM*1+YY] += fiy1;
 413         f[i_coord_offset+DIM*1+ZZ] += fiz1;
 414         tx                         += fix1;
 415         ty                         += fiy1;
 416         tz                         += fiz1;
 417         f[i_coord_offset+DIM*2+XX] += fix2;
 418         f[i_coord_offset+DIM*2+YY] += fiy2;
 419         f[i_coord_offset+DIM*2+ZZ] += fiz2;
 420         tx                         += fix2;
 421         ty                         += fiy2;
 422         tz                         += fiz2;
 423         f[i_coord_offset+DIM*3+XX] += fix3;
 424         f[i_coord_offset+DIM*3+YY] += fiy3;
 425         f[i_coord_offset+DIM*3+ZZ] += fiz3;
 426         tx                         += fix3;
 427         ty                         += fiy3;
 428         tz                         += fiz3;
 429         fshift[i_shift_offset+XX]  += tx;
 430         fshift[i_shift_offset+YY]  += ty;
 431         fshift[i_shift_offset+ZZ]  += tz;
 432
 433         ggid                        = gid[iidx];
 434         /* Update potential energies */
 435         kernel_data->energygrp_elec[ggid] += velecsum;
 436         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 437
 438         /* Increment number of inner iterations */
 439         inneriter                  += j_index_end - j_index_start;
 440
 441         /* Outer loop uses 41 flops */
 442     }
 443
 444     /* Increment number of outer iterations */
 445     outeriter        += nri;
 446
 447     /* Update outer/inner flops */
 448
 449     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_VF,outeriter*41 + inneriter*181);
 450 }
 451 /*
 452  * Gromacs nonbonded kernel:   nb_kernel_ElecCSTab_VdwCSTab_GeomW4P1_F_c
 453  * Electrostatics interaction: CubicSplineTable
 454  * VdW interaction:            CubicSplineTable
 455  * Geometry:                   Water4-Particle
 456  * Calculate force/pot:        Force
 457  */
 458 void
 459 nb_kernel_ElecCSTab_VdwCSTab_GeomW4P1_F_c
 460                     (t_nblist                    * gmx_restrict       nlist,
 461                      rvec                        * gmx_restrict          xx,
 462                      rvec                        * gmx_restrict          ff,
 463                      t_forcerec                  * gmx_restrict          fr,
 464                      t_mdatoms                   * gmx_restrict     mdatoms,
 465                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 466                      t_nrnb                      * gmx_restrict        nrnb)
 467 {
 468     int              i_shift_offset,i_coord_offset,j_coord_offset;
 469     int              j_index_start,j_index_end;
 470     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 471     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 472     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 473     real             *shiftvec,*fshift,*x,*f;
 474     int              vdwioffset0;
 475     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 476     int              vdwioffset1;
 477     real             ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
 478     int              vdwioffset2;
 479     real             ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
 480     int              vdwioffset3;
 481     real             ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
 482     int              vdwjidx0;
 483     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 484     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 485     real             dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10;
 486     real             dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20;
 487     real             dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30,cexp1_30,cexp2_30;
 488     real             velec,felec,velecsum,facel,crf,krf,krf2;
 489     real             *charge;
 490     int              nvdwtype;
 491     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 492     int              *vdwtype;
 493     real             *vdwparam;
 494     int              vfitab;
 495     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
 496     real             *vftab;
 497
 498     x                = xx[0];
 499     f                = ff[0];
 500
 501     nri              = nlist->nri;
 502     iinr             = nlist->iinr;
 503     jindex           = nlist->jindex;
 504     jjnr             = nlist->jjnr;
 505     shiftidx         = nlist->shift;
 506     gid              = nlist->gid;
 507     shiftvec         = fr->shift_vec[0];
 508     fshift           = fr->fshift[0];
 509     facel            = fr->epsfac;
 510     charge           = mdatoms->chargeA;
 511     nvdwtype         = fr->ntype;
 512     vdwparam         = fr->nbfp;
 513     vdwtype          = mdatoms->typeA;
 514
 515     vftab            = kernel_data->table_elec_vdw->data;
 516     vftabscale       = kernel_data->table_elec_vdw->scale;
 517
 518     /* Setup water-specific parameters */
 519     inr              = nlist->iinr[0];
 520     iq1              = facel*charge[inr+1];
 521     iq2              = facel*charge[inr+2];
 522     iq3              = facel*charge[inr+3];
 523     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 524
 525     outeriter        = 0;
 526     inneriter        = 0;
 527
 528     /* Start outer loop over neighborlists */
 529     for(iidx=0; iidx<nri; iidx++)
 530     {
 531         /* Load shift vector for this list */
 532         i_shift_offset   = DIM*shiftidx[iidx];
 533         shX              = shiftvec[i_shift_offset+XX];
 534         shY              = shiftvec[i_shift_offset+YY];
 535         shZ              = shiftvec[i_shift_offset+ZZ];
 536
 537         /* Load limits for loop over neighbors */
 538         j_index_start    = jindex[iidx];
 539         j_index_end      = jindex[iidx+1];
 540
 541         /* Get outer coordinate index */
 542         inr              = iinr[iidx];
 543         i_coord_offset   = DIM*inr;
 544
 545         /* Load i particle coords and add shift vector */
 546         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 547         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 548         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 549         ix1              = shX + x[i_coord_offset+DIM*1+XX];
 550         iy1              = shY + x[i_coord_offset+DIM*1+YY];
 551         iz1              = shZ + x[i_coord_offset+DIM*1+ZZ];
 552         ix2              = shX + x[i_coord_offset+DIM*2+XX];
 553         iy2              = shY + x[i_coord_offset+DIM*2+YY];
 554         iz2              = shZ + x[i_coord_offset+DIM*2+ZZ];
 555         ix3              = shX + x[i_coord_offset+DIM*3+XX];
 556         iy3              = shY + x[i_coord_offset+DIM*3+YY];
 557         iz3              = shZ + x[i_coord_offset+DIM*3+ZZ];
 558
 559         fix0             = 0.0;
 560         fiy0             = 0.0;
 561         fiz0             = 0.0;
 562         fix1             = 0.0;
 563         fiy1             = 0.0;
 564         fiz1             = 0.0;
 565         fix2             = 0.0;
 566         fiy2             = 0.0;
 567         fiz2             = 0.0;
 568         fix3             = 0.0;
 569         fiy3             = 0.0;
 570         fiz3             = 0.0;
 571
 572         /* Start inner kernel loop */
 573         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 574         {
 575             /* Get j neighbor index, and coordinate index */
 576             jnr              = jjnr[jidx];
 577             j_coord_offset   = DIM*jnr;
 578
 579             /* load j atom coordinates */
 580             jx0              = x[j_coord_offset+DIM*0+XX];
 581             jy0              = x[j_coord_offset+DIM*0+YY];
 582             jz0              = x[j_coord_offset+DIM*0+ZZ];
 583
 584             /* Calculate displacement vector */
 585             dx00             = ix0 - jx0;
 586             dy00             = iy0 - jy0;
 587             dz00             = iz0 - jz0;
 588             dx10             = ix1 - jx0;
 589             dy10             = iy1 - jy0;
 590             dz10             = iz1 - jz0;
 591             dx20             = ix2 - jx0;
 592             dy20             = iy2 - jy0;
 593             dz20             = iz2 - jz0;
 594             dx30             = ix3 - jx0;
 595             dy30             = iy3 - jy0;
 596             dz30             = iz3 - jz0;
 597
 598             /* Calculate squared distance and things based on it */
 599             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 600             rsq10            = dx10*dx10+dy10*dy10+dz10*dz10;
 601             rsq20            = dx20*dx20+dy20*dy20+dz20*dz20;
 602             rsq30            = dx30*dx30+dy30*dy30+dz30*dz30;
 603
 604             rinv00           = gmx_invsqrt(rsq00);
 605             rinv10           = gmx_invsqrt(rsq10);
 606             rinv20           = gmx_invsqrt(rsq20);
 607             rinv30           = gmx_invsqrt(rsq30);
 608
 609             /* Load parameters for j particles */
 610             jq0              = charge[jnr+0];
 611             vdwjidx0         = 2*vdwtype[jnr+0];
 612
 613             /**************************
 614              * CALCULATE INTERACTIONS *
 615              **************************/
 616
 617             r00              = rsq00*rinv00;
 618
 619             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 620             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 621
 622             /* Calculate table index by multiplying r with table scale and truncate to integer */
 623             rt               = r00*vftabscale;
 624             vfitab           = rt;
 625             vfeps            = rt-vfitab;
 626             vfitab           = 3*4*vfitab;
 627
 628             /* CUBIC SPLINE TABLE DISPERSION */
 629             vfitab          += 4;
 630             F                = vftab[vfitab+1];
 631             Geps             = vfeps*vftab[vfitab+2];
 632             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 633             Fp               = F+Geps+Heps2;
 634             FF               = Fp+Geps+2.0*Heps2;
 635             fvdw6            = c6_00*FF;
 636
 637             /* CUBIC SPLINE TABLE REPULSION */
 638             F                = vftab[vfitab+5];
 639             Geps             = vfeps*vftab[vfitab+6];
 640             Heps2            = vfeps*vfeps*vftab[vfitab+7];
 641             Fp               = F+Geps+Heps2;
 642             FF               = Fp+Geps+2.0*Heps2;
 643             fvdw12           = c12_00*FF;
 644             fvdw             = -(fvdw6+fvdw12)*vftabscale*rinv00;
 645
 646             fscal            = fvdw;
 647
 648             /* Calculate temporary vectorial force */
 649             tx               = fscal*dx00;
 650             ty               = fscal*dy00;
 651             tz               = fscal*dz00;
 652
 653             /* Update vectorial force */
 654             fix0            += tx;
 655             fiy0            += ty;
 656             fiz0            += tz;
 657             f[j_coord_offset+DIM*0+XX] -= tx;
 658             f[j_coord_offset+DIM*0+YY] -= ty;
 659             f[j_coord_offset+DIM*0+ZZ] -= tz;
 660
 661             /**************************
 662              * CALCULATE INTERACTIONS *
 663              **************************/
 664
 665             r10              = rsq10*rinv10;
 666
 667             qq10             = iq1*jq0;
 668
 669             /* Calculate table index by multiplying r with table scale and truncate to integer */
 670             rt               = r10*vftabscale;
 671             vfitab           = rt;
 672             vfeps            = rt-vfitab;
 673             vfitab           = 3*4*vfitab;
 674
 675             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 676             F                = vftab[vfitab+1];
 677             Geps             = vfeps*vftab[vfitab+2];
 678             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 679             Fp               = F+Geps+Heps2;
 680             FF               = Fp+Geps+2.0*Heps2;
 681             felec            = -qq10*FF*vftabscale*rinv10;
 682
 683             fscal            = felec;
 684
 685             /* Calculate temporary vectorial force */
 686             tx               = fscal*dx10;
 687             ty               = fscal*dy10;
 688             tz               = fscal*dz10;
 689
 690             /* Update vectorial force */
 691             fix1            += tx;
 692             fiy1            += ty;
 693             fiz1            += tz;
 694             f[j_coord_offset+DIM*0+XX] -= tx;
 695             f[j_coord_offset+DIM*0+YY] -= ty;
 696             f[j_coord_offset+DIM*0+ZZ] -= tz;
 697
 698             /**************************
 699              * CALCULATE INTERACTIONS *
 700              **************************/
 701
 702             r20              = rsq20*rinv20;
 703
 704             qq20             = iq2*jq0;
 705
 706             /* Calculate table index by multiplying r with table scale and truncate to integer */
 707             rt               = r20*vftabscale;
 708             vfitab           = rt;
 709             vfeps            = rt-vfitab;
 710             vfitab           = 3*4*vfitab;
 711
 712             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 713             F                = vftab[vfitab+1];
 714             Geps             = vfeps*vftab[vfitab+2];
 715             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 716             Fp               = F+Geps+Heps2;
 717             FF               = Fp+Geps+2.0*Heps2;
 718             felec            = -qq20*FF*vftabscale*rinv20;
 719
 720             fscal            = felec;
 721
 722             /* Calculate temporary vectorial force */
 723             tx               = fscal*dx20;
 724             ty               = fscal*dy20;
 725             tz               = fscal*dz20;
 726
 727             /* Update vectorial force */
 728             fix2            += tx;
 729             fiy2            += ty;
 730             fiz2            += tz;
 731             f[j_coord_offset+DIM*0+XX] -= tx;
 732             f[j_coord_offset+DIM*0+YY] -= ty;
 733             f[j_coord_offset+DIM*0+ZZ] -= tz;
 734
 735             /**************************
 736              * CALCULATE INTERACTIONS *
 737              **************************/
 738
 739             r30              = rsq30*rinv30;
 740
 741             qq30             = iq3*jq0;
 742
 743             /* Calculate table index by multiplying r with table scale and truncate to integer */
 744             rt               = r30*vftabscale;
 745             vfitab           = rt;
 746             vfeps            = rt-vfitab;
 747             vfitab           = 3*4*vfitab;
 748
 749             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 750             F                = vftab[vfitab+1];
 751             Geps             = vfeps*vftab[vfitab+2];
 752             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 753             Fp               = F+Geps+Heps2;
 754             FF               = Fp+Geps+2.0*Heps2;
 755             felec            = -qq30*FF*vftabscale*rinv30;
 756
 757             fscal            = felec;
 758
 759             /* Calculate temporary vectorial force */
 760             tx               = fscal*dx30;
 761             ty               = fscal*dy30;
 762             tz               = fscal*dz30;
 763
 764             /* Update vectorial force */
 765             fix3            += tx;
 766             fiy3            += ty;
 767             fiz3            += tz;
 768             f[j_coord_offset+DIM*0+XX] -= tx;
 769             f[j_coord_offset+DIM*0+YY] -= ty;
 770             f[j_coord_offset+DIM*0+ZZ] -= tz;
 771
 772             /* Inner loop uses 161 flops */
 773         }
 774         /* End of innermost loop */
 775
 776         tx = ty = tz = 0;
 777         f[i_coord_offset+DIM*0+XX] += fix0;
 778         f[i_coord_offset+DIM*0+YY] += fiy0;
 779         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 780         tx                         += fix0;
 781         ty                         += fiy0;
 782         tz                         += fiz0;
 783         f[i_coord_offset+DIM*1+XX] += fix1;
 784         f[i_coord_offset+DIM*1+YY] += fiy1;
 785         f[i_coord_offset+DIM*1+ZZ] += fiz1;
 786         tx                         += fix1;
 787         ty                         += fiy1;
 788         tz                         += fiz1;
 789         f[i_coord_offset+DIM*2+XX] += fix2;
 790         f[i_coord_offset+DIM*2+YY] += fiy2;
 791         f[i_coord_offset+DIM*2+ZZ] += fiz2;
 792         tx                         += fix2;
 793         ty                         += fiy2;
 794         tz                         += fiz2;
 795         f[i_coord_offset+DIM*3+XX] += fix3;
 796         f[i_coord_offset+DIM*3+YY] += fiy3;
 797         f[i_coord_offset+DIM*3+ZZ] += fiz3;
 798         tx                         += fix3;
 799         ty                         += fiy3;
 800         tz                         += fiz3;
 801         fshift[i_shift_offset+XX]  += tx;
 802         fshift[i_shift_offset+YY]  += ty;
 803         fshift[i_shift_offset+ZZ]  += tz;
 804
 805         /* Increment number of inner iterations */
 806         inneriter                  += j_index_end - j_index_start;
 807
 808         /* Outer loop uses 39 flops */
 809     }
 810
 811     /* Increment number of outer iterations */
 812     outeriter        += nri;
 813
 814     /* Update outer/inner flops */
 815
 816     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_F,outeriter*39 + inneriter*161);
 817 }