src/gromacs/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecCSTab_VdwBham_GeomW3P1_c.c

   1 /*
   2  * This file is part of the GROMACS molecular simulation package.
   3  *
   4  * Copyright (c) 2012,2013,2014, by the GROMACS development team, led by
   5  * Mark Abraham, David van der Spoel, Berk Hess, and Erik Lindahl,
   6  * and including many others, as listed in the AUTHORS file in the
   7  * top-level source directory and at http://www.gromacs.org.
   8  *
   9  * GROMACS is free software; you can redistribute it and/or
  10  * modify it under the terms of the GNU Lesser General Public License
  11  * as published by the Free Software Foundation; either version 2.1
  12  * of the License, or (at your option) any later version.
  13  *
  14  * GROMACS is distributed in the hope that it will be useful,
  15  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  16  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  17  * Lesser General Public License for more details.
  18  *
  19  * You should have received a copy of the GNU Lesser General Public
  20  * License along with GROMACS; if not, see
  21  * http://www.gnu.org/licenses, or write to the Free Software Foundation,
  22  * Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA.
  23  *
  24  * If you want to redistribute modifications to GROMACS, please
  25  * consider that scientific software is very special. Version
  26  * control is crucial - bugs must be traceable. We will be happy to
  27  * consider code for inclusion in the official distribution, but
  28  * derived work must not be called official GROMACS. Details are found
  29  * in the README & COPYING files - if they are missing, get the
  30  * official version at http://www.gromacs.org.
  31  *
  32  * To help us fund GROMACS development, we humbly ask that you cite
  33  * the research papers on the package. Check out http://www.gromacs.org.
  34  */
  35 /*
  36  * Note: this file was generated by the GROMACS c kernel generator.
  37  */
  38 #include "config.h"
  39
  40 #include <math.h>
  41
  42 #include "../nb_kernel.h"
  43 #include "gromacs/legacyheaders/types/simple.h"
  44 #include "gromacs/math/vec.h"
  45 #include "gromacs/legacyheaders/nrnb.h"
  46
  47 /*
  48  * Gromacs nonbonded kernel:   nb_kernel_ElecCSTab_VdwBham_GeomW3P1_VF_c
  49  * Electrostatics interaction: CubicSplineTable
  50  * VdW interaction:            Buckingham
  51  * Geometry:                   Water3-Particle
  52  * Calculate force/pot:        PotentialAndForce
  53  */
  54 void
  55 nb_kernel_ElecCSTab_VdwBham_GeomW3P1_VF_c
  56                     (t_nblist                    * gmx_restrict       nlist,
  57                      rvec                        * gmx_restrict          xx,
  58                      rvec                        * gmx_restrict          ff,
  59                      t_forcerec                  * gmx_restrict          fr,
  60                      t_mdatoms                   * gmx_restrict     mdatoms,
  61                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
  62                      t_nrnb                      * gmx_restrict        nrnb)
  63 {
  64     int              i_shift_offset,i_coord_offset,j_coord_offset;
  65     int              j_index_start,j_index_end;
  66     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
  67     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
  68     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  69     real             *shiftvec,*fshift,*x,*f;
  70     int              vdwioffset0;
  71     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  72     int              vdwioffset1;
  73     real             ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
  74     int              vdwioffset2;
  75     real             ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
  76     int              vdwjidx0;
  77     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  78     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
  79     real             dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10;
  80     real             dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20;
  81     real             velec,felec,velecsum,facel,crf,krf,krf2;
  82     real             *charge;
  83     int              nvdwtype;
  84     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
  85     int              *vdwtype;
  86     real             *vdwparam;
  87     int              vfitab;
  88     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
  89     real             *vftab;
  90
  91     x                = xx[0];
  92     f                = ff[0];
  93
  94     nri              = nlist->nri;
  95     iinr             = nlist->iinr;
  96     jindex           = nlist->jindex;
  97     jjnr             = nlist->jjnr;
  98     shiftidx         = nlist->shift;
  99     gid              = nlist->gid;
 100     shiftvec         = fr->shift_vec[0];
 101     fshift           = fr->fshift[0];
 102     facel            = fr->epsfac;
 103     charge           = mdatoms->chargeA;
 104     nvdwtype         = fr->ntype;
 105     vdwparam         = fr->nbfp;
 106     vdwtype          = mdatoms->typeA;
 107
 108     vftab            = kernel_data->table_elec->data;
 109     vftabscale       = kernel_data->table_elec->scale;
 110
 111     /* Setup water-specific parameters */
 112     inr              = nlist->iinr[0];
 113     iq0              = facel*charge[inr+0];
 114     iq1              = facel*charge[inr+1];
 115     iq2              = facel*charge[inr+2];
 116     vdwioffset0      = 3*nvdwtype*vdwtype[inr+0];
 117
 118     outeriter        = 0;
 119     inneriter        = 0;
 120
 121     /* Start outer loop over neighborlists */
 122     for(iidx=0; iidx<nri; iidx++)
 123     {
 124         /* Load shift vector for this list */
 125         i_shift_offset   = DIM*shiftidx[iidx];
 126         shX              = shiftvec[i_shift_offset+XX];
 127         shY              = shiftvec[i_shift_offset+YY];
 128         shZ              = shiftvec[i_shift_offset+ZZ];
 129
 130         /* Load limits for loop over neighbors */
 131         j_index_start    = jindex[iidx];
 132         j_index_end      = jindex[iidx+1];
 133
 134         /* Get outer coordinate index */
 135         inr              = iinr[iidx];
 136         i_coord_offset   = DIM*inr;
 137
 138         /* Load i particle coords and add shift vector */
 139         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 140         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 141         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 142         ix1              = shX + x[i_coord_offset+DIM*1+XX];
 143         iy1              = shY + x[i_coord_offset+DIM*1+YY];
 144         iz1              = shZ + x[i_coord_offset+DIM*1+ZZ];
 145         ix2              = shX + x[i_coord_offset+DIM*2+XX];
 146         iy2              = shY + x[i_coord_offset+DIM*2+YY];
 147         iz2              = shZ + x[i_coord_offset+DIM*2+ZZ];
 148
 149         fix0             = 0.0;
 150         fiy0             = 0.0;
 151         fiz0             = 0.0;
 152         fix1             = 0.0;
 153         fiy1             = 0.0;
 154         fiz1             = 0.0;
 155         fix2             = 0.0;
 156         fiy2             = 0.0;
 157         fiz2             = 0.0;
 158
 159         /* Reset potential sums */
 160         velecsum         = 0.0;
 161         vvdwsum          = 0.0;
 162
 163         /* Start inner kernel loop */
 164         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 165         {
 166             /* Get j neighbor index, and coordinate index */
 167             jnr              = jjnr[jidx];
 168             j_coord_offset   = DIM*jnr;
 169
 170             /* load j atom coordinates */
 171             jx0              = x[j_coord_offset+DIM*0+XX];
 172             jy0              = x[j_coord_offset+DIM*0+YY];
 173             jz0              = x[j_coord_offset+DIM*0+ZZ];
 174
 175             /* Calculate displacement vector */
 176             dx00             = ix0 - jx0;
 177             dy00             = iy0 - jy0;
 178             dz00             = iz0 - jz0;
 179             dx10             = ix1 - jx0;
 180             dy10             = iy1 - jy0;
 181             dz10             = iz1 - jz0;
 182             dx20             = ix2 - jx0;
 183             dy20             = iy2 - jy0;
 184             dz20             = iz2 - jz0;
 185
 186             /* Calculate squared distance and things based on it */
 187             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 188             rsq10            = dx10*dx10+dy10*dy10+dz10*dz10;
 189             rsq20            = dx20*dx20+dy20*dy20+dz20*dz20;
 190
 191             rinv00           = gmx_invsqrt(rsq00);
 192             rinv10           = gmx_invsqrt(rsq10);
 193             rinv20           = gmx_invsqrt(rsq20);
 194
 195             rinvsq00         = rinv00*rinv00;
 196
 197             /* Load parameters for j particles */
 198             jq0              = charge[jnr+0];
 199             vdwjidx0         = 3*vdwtype[jnr+0];
 200
 201             /**************************
 202              * CALCULATE INTERACTIONS *
 203              **************************/
 204
 205             r00              = rsq00*rinv00;
 206
 207             qq00             = iq0*jq0;
 208             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 209             cexp1_00         = vdwparam[vdwioffset0+vdwjidx0+1];
 210             cexp2_00         = vdwparam[vdwioffset0+vdwjidx0+2];
 211
 212             /* Calculate table index by multiplying r with table scale and truncate to integer */
 213             rt               = r00*vftabscale;
 214             vfitab           = rt;
 215             vfeps            = rt-vfitab;
 216             vfitab           = 1*4*vfitab;
 217
 218             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 219             Y                = vftab[vfitab];
 220             F                = vftab[vfitab+1];
 221             Geps             = vfeps*vftab[vfitab+2];
 222             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 223             Fp               = F+Geps+Heps2;
 224             VV               = Y+vfeps*Fp;
 225             velec            = qq00*VV;
 226             FF               = Fp+Geps+2.0*Heps2;
 227             felec            = -qq00*FF*vftabscale*rinv00;
 228
 229             /* BUCKINGHAM DISPERSION/REPULSION */
 230             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 231             vvdw6            = c6_00*rinvsix;
 232             br               = cexp2_00*r00;
 233             vvdwexp          = cexp1_00*exp(-br);
 234             vvdw             = vvdwexp - vvdw6*(1.0/6.0);
 235             fvdw             = (br*vvdwexp-vvdw6)*rinvsq00;
 236
 237             /* Update potential sums from outer loop */
 238             velecsum        += velec;
 239             vvdwsum         += vvdw;
 240
 241             fscal            = felec+fvdw;
 242
 243             /* Calculate temporary vectorial force */
 244             tx               = fscal*dx00;
 245             ty               = fscal*dy00;
 246             tz               = fscal*dz00;
 247
 248             /* Update vectorial force */
 249             fix0            += tx;
 250             fiy0            += ty;
 251             fiz0            += tz;
 252             f[j_coord_offset+DIM*0+XX] -= tx;
 253             f[j_coord_offset+DIM*0+YY] -= ty;
 254             f[j_coord_offset+DIM*0+ZZ] -= tz;
 255
 256             /**************************
 257              * CALCULATE INTERACTIONS *
 258              **************************/
 259
 260             r10              = rsq10*rinv10;
 261
 262             qq10             = iq1*jq0;
 263
 264             /* Calculate table index by multiplying r with table scale and truncate to integer */
 265             rt               = r10*vftabscale;
 266             vfitab           = rt;
 267             vfeps            = rt-vfitab;
 268             vfitab           = 1*4*vfitab;
 269
 270             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 271             Y                = vftab[vfitab];
 272             F                = vftab[vfitab+1];
 273             Geps             = vfeps*vftab[vfitab+2];
 274             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 275             Fp               = F+Geps+Heps2;
 276             VV               = Y+vfeps*Fp;
 277             velec            = qq10*VV;
 278             FF               = Fp+Geps+2.0*Heps2;
 279             felec            = -qq10*FF*vftabscale*rinv10;
 280
 281             /* Update potential sums from outer loop */
 282             velecsum        += velec;
 283
 284             fscal            = felec;
 285
 286             /* Calculate temporary vectorial force */
 287             tx               = fscal*dx10;
 288             ty               = fscal*dy10;
 289             tz               = fscal*dz10;
 290
 291             /* Update vectorial force */
 292             fix1            += tx;
 293             fiy1            += ty;
 294             fiz1            += tz;
 295             f[j_coord_offset+DIM*0+XX] -= tx;
 296             f[j_coord_offset+DIM*0+YY] -= ty;
 297             f[j_coord_offset+DIM*0+ZZ] -= tz;
 298
 299             /**************************
 300              * CALCULATE INTERACTIONS *
 301              **************************/
 302
 303             r20              = rsq20*rinv20;
 304
 305             qq20             = iq2*jq0;
 306
 307             /* Calculate table index by multiplying r with table scale and truncate to integer */
 308             rt               = r20*vftabscale;
 309             vfitab           = rt;
 310             vfeps            = rt-vfitab;
 311             vfitab           = 1*4*vfitab;
 312
 313             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 314             Y                = vftab[vfitab];
 315             F                = vftab[vfitab+1];
 316             Geps             = vfeps*vftab[vfitab+2];
 317             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 318             Fp               = F+Geps+Heps2;
 319             VV               = Y+vfeps*Fp;
 320             velec            = qq20*VV;
 321             FF               = Fp+Geps+2.0*Heps2;
 322             felec            = -qq20*FF*vftabscale*rinv20;
 323
 324             /* Update potential sums from outer loop */
 325             velecsum        += velec;
 326
 327             fscal            = felec;
 328
 329             /* Calculate temporary vectorial force */
 330             tx               = fscal*dx20;
 331             ty               = fscal*dy20;
 332             tz               = fscal*dz20;
 333
 334             /* Update vectorial force */
 335             fix2            += tx;
 336             fiy2            += ty;
 337             fiz2            += tz;
 338             f[j_coord_offset+DIM*0+XX] -= tx;
 339             f[j_coord_offset+DIM*0+YY] -= ty;
 340             f[j_coord_offset+DIM*0+ZZ] -= tz;
 341
 342             /* Inner loop uses 165 flops */
 343         }
 344         /* End of innermost loop */
 345
 346         tx = ty = tz = 0;
 347         f[i_coord_offset+DIM*0+XX] += fix0;
 348         f[i_coord_offset+DIM*0+YY] += fiy0;
 349         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 350         tx                         += fix0;
 351         ty                         += fiy0;
 352         tz                         += fiz0;
 353         f[i_coord_offset+DIM*1+XX] += fix1;
 354         f[i_coord_offset+DIM*1+YY] += fiy1;
 355         f[i_coord_offset+DIM*1+ZZ] += fiz1;
 356         tx                         += fix1;
 357         ty                         += fiy1;
 358         tz                         += fiz1;
 359         f[i_coord_offset+DIM*2+XX] += fix2;
 360         f[i_coord_offset+DIM*2+YY] += fiy2;
 361         f[i_coord_offset+DIM*2+ZZ] += fiz2;
 362         tx                         += fix2;
 363         ty                         += fiy2;
 364         tz                         += fiz2;
 365         fshift[i_shift_offset+XX]  += tx;
 366         fshift[i_shift_offset+YY]  += ty;
 367         fshift[i_shift_offset+ZZ]  += tz;
 368
 369         ggid                        = gid[iidx];
 370         /* Update potential energies */
 371         kernel_data->energygrp_elec[ggid] += velecsum;
 372         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 373
 374         /* Increment number of inner iterations */
 375         inneriter                  += j_index_end - j_index_start;
 376
 377         /* Outer loop uses 32 flops */
 378     }
 379
 380     /* Increment number of outer iterations */
 381     outeriter        += nri;
 382
 383     /* Update outer/inner flops */
 384
 385     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3_VF,outeriter*32 + inneriter*165);
 386 }
 387 /*
 388  * Gromacs nonbonded kernel:   nb_kernel_ElecCSTab_VdwBham_GeomW3P1_F_c
 389  * Electrostatics interaction: CubicSplineTable
 390  * VdW interaction:            Buckingham
 391  * Geometry:                   Water3-Particle
 392  * Calculate force/pot:        Force
 393  */
 394 void
 395 nb_kernel_ElecCSTab_VdwBham_GeomW3P1_F_c
 396                     (t_nblist                    * gmx_restrict       nlist,
 397                      rvec                        * gmx_restrict          xx,
 398                      rvec                        * gmx_restrict          ff,
 399                      t_forcerec                  * gmx_restrict          fr,
 400                      t_mdatoms                   * gmx_restrict     mdatoms,
 401                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 402                      t_nrnb                      * gmx_restrict        nrnb)
 403 {
 404     int              i_shift_offset,i_coord_offset,j_coord_offset;
 405     int              j_index_start,j_index_end;
 406     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 407     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 408     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 409     real             *shiftvec,*fshift,*x,*f;
 410     int              vdwioffset0;
 411     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 412     int              vdwioffset1;
 413     real             ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
 414     int              vdwioffset2;
 415     real             ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
 416     int              vdwjidx0;
 417     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 418     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 419     real             dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10;
 420     real             dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20;
 421     real             velec,felec,velecsum,facel,crf,krf,krf2;
 422     real             *charge;
 423     int              nvdwtype;
 424     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 425     int              *vdwtype;
 426     real             *vdwparam;
 427     int              vfitab;
 428     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
 429     real             *vftab;
 430
 431     x                = xx[0];
 432     f                = ff[0];
 433
 434     nri              = nlist->nri;
 435     iinr             = nlist->iinr;
 436     jindex           = nlist->jindex;
 437     jjnr             = nlist->jjnr;
 438     shiftidx         = nlist->shift;
 439     gid              = nlist->gid;
 440     shiftvec         = fr->shift_vec[0];
 441     fshift           = fr->fshift[0];
 442     facel            = fr->epsfac;
 443     charge           = mdatoms->chargeA;
 444     nvdwtype         = fr->ntype;
 445     vdwparam         = fr->nbfp;
 446     vdwtype          = mdatoms->typeA;
 447
 448     vftab            = kernel_data->table_elec->data;
 449     vftabscale       = kernel_data->table_elec->scale;
 450
 451     /* Setup water-specific parameters */
 452     inr              = nlist->iinr[0];
 453     iq0              = facel*charge[inr+0];
 454     iq1              = facel*charge[inr+1];
 455     iq2              = facel*charge[inr+2];
 456     vdwioffset0      = 3*nvdwtype*vdwtype[inr+0];
 457
 458     outeriter        = 0;
 459     inneriter        = 0;
 460
 461     /* Start outer loop over neighborlists */
 462     for(iidx=0; iidx<nri; iidx++)
 463     {
 464         /* Load shift vector for this list */
 465         i_shift_offset   = DIM*shiftidx[iidx];
 466         shX              = shiftvec[i_shift_offset+XX];
 467         shY              = shiftvec[i_shift_offset+YY];
 468         shZ              = shiftvec[i_shift_offset+ZZ];
 469
 470         /* Load limits for loop over neighbors */
 471         j_index_start    = jindex[iidx];
 472         j_index_end      = jindex[iidx+1];
 473
 474         /* Get outer coordinate index */
 475         inr              = iinr[iidx];
 476         i_coord_offset   = DIM*inr;
 477
 478         /* Load i particle coords and add shift vector */
 479         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 480         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 481         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 482         ix1              = shX + x[i_coord_offset+DIM*1+XX];
 483         iy1              = shY + x[i_coord_offset+DIM*1+YY];
 484         iz1              = shZ + x[i_coord_offset+DIM*1+ZZ];
 485         ix2              = shX + x[i_coord_offset+DIM*2+XX];
 486         iy2              = shY + x[i_coord_offset+DIM*2+YY];
 487         iz2              = shZ + x[i_coord_offset+DIM*2+ZZ];
 488
 489         fix0             = 0.0;
 490         fiy0             = 0.0;
 491         fiz0             = 0.0;
 492         fix1             = 0.0;
 493         fiy1             = 0.0;
 494         fiz1             = 0.0;
 495         fix2             = 0.0;
 496         fiy2             = 0.0;
 497         fiz2             = 0.0;
 498
 499         /* Start inner kernel loop */
 500         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 501         {
 502             /* Get j neighbor index, and coordinate index */
 503             jnr              = jjnr[jidx];
 504             j_coord_offset   = DIM*jnr;
 505
 506             /* load j atom coordinates */
 507             jx0              = x[j_coord_offset+DIM*0+XX];
 508             jy0              = x[j_coord_offset+DIM*0+YY];
 509             jz0              = x[j_coord_offset+DIM*0+ZZ];
 510
 511             /* Calculate displacement vector */
 512             dx00             = ix0 - jx0;
 513             dy00             = iy0 - jy0;
 514             dz00             = iz0 - jz0;
 515             dx10             = ix1 - jx0;
 516             dy10             = iy1 - jy0;
 517             dz10             = iz1 - jz0;
 518             dx20             = ix2 - jx0;
 519             dy20             = iy2 - jy0;
 520             dz20             = iz2 - jz0;
 521
 522             /* Calculate squared distance and things based on it */
 523             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 524             rsq10            = dx10*dx10+dy10*dy10+dz10*dz10;
 525             rsq20            = dx20*dx20+dy20*dy20+dz20*dz20;
 526
 527             rinv00           = gmx_invsqrt(rsq00);
 528             rinv10           = gmx_invsqrt(rsq10);
 529             rinv20           = gmx_invsqrt(rsq20);
 530
 531             rinvsq00         = rinv00*rinv00;
 532
 533             /* Load parameters for j particles */
 534             jq0              = charge[jnr+0];
 535             vdwjidx0         = 3*vdwtype[jnr+0];
 536
 537             /**************************
 538              * CALCULATE INTERACTIONS *
 539              **************************/
 540
 541             r00              = rsq00*rinv00;
 542
 543             qq00             = iq0*jq0;
 544             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 545             cexp1_00         = vdwparam[vdwioffset0+vdwjidx0+1];
 546             cexp2_00         = vdwparam[vdwioffset0+vdwjidx0+2];
 547
 548             /* Calculate table index by multiplying r with table scale and truncate to integer */
 549             rt               = r00*vftabscale;
 550             vfitab           = rt;
 551             vfeps            = rt-vfitab;
 552             vfitab           = 1*4*vfitab;
 553
 554             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 555             F                = vftab[vfitab+1];
 556             Geps             = vfeps*vftab[vfitab+2];
 557             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 558             Fp               = F+Geps+Heps2;
 559             FF               = Fp+Geps+2.0*Heps2;
 560             felec            = -qq00*FF*vftabscale*rinv00;
 561
 562             /* BUCKINGHAM DISPERSION/REPULSION */
 563             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 564             vvdw6            = c6_00*rinvsix;
 565             br               = cexp2_00*r00;
 566             vvdwexp          = cexp1_00*exp(-br);
 567             fvdw             = (br*vvdwexp-vvdw6)*rinvsq00;
 568
 569             fscal            = felec+fvdw;
 570
 571             /* Calculate temporary vectorial force */
 572             tx               = fscal*dx00;
 573             ty               = fscal*dy00;
 574             tz               = fscal*dz00;
 575
 576             /* Update vectorial force */
 577             fix0            += tx;
 578             fiy0            += ty;
 579             fiz0            += tz;
 580             f[j_coord_offset+DIM*0+XX] -= tx;
 581             f[j_coord_offset+DIM*0+YY] -= ty;
 582             f[j_coord_offset+DIM*0+ZZ] -= tz;
 583
 584             /**************************
 585              * CALCULATE INTERACTIONS *
 586              **************************/
 587
 588             r10              = rsq10*rinv10;
 589
 590             qq10             = iq1*jq0;
 591
 592             /* Calculate table index by multiplying r with table scale and truncate to integer */
 593             rt               = r10*vftabscale;
 594             vfitab           = rt;
 595             vfeps            = rt-vfitab;
 596             vfitab           = 1*4*vfitab;
 597
 598             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 599             F                = vftab[vfitab+1];
 600             Geps             = vfeps*vftab[vfitab+2];
 601             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 602             Fp               = F+Geps+Heps2;
 603             FF               = Fp+Geps+2.0*Heps2;
 604             felec            = -qq10*FF*vftabscale*rinv10;
 605
 606             fscal            = felec;
 607
 608             /* Calculate temporary vectorial force */
 609             tx               = fscal*dx10;
 610             ty               = fscal*dy10;
 611             tz               = fscal*dz10;
 612
 613             /* Update vectorial force */
 614             fix1            += tx;
 615             fiy1            += ty;
 616             fiz1            += tz;
 617             f[j_coord_offset+DIM*0+XX] -= tx;
 618             f[j_coord_offset+DIM*0+YY] -= ty;
 619             f[j_coord_offset+DIM*0+ZZ] -= tz;
 620
 621             /**************************
 622              * CALCULATE INTERACTIONS *
 623              **************************/
 624
 625             r20              = rsq20*rinv20;
 626
 627             qq20             = iq2*jq0;
 628
 629             /* Calculate table index by multiplying r with table scale and truncate to integer */
 630             rt               = r20*vftabscale;
 631             vfitab           = rt;
 632             vfeps            = rt-vfitab;
 633             vfitab           = 1*4*vfitab;
 634
 635             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 636             F                = vftab[vfitab+1];
 637             Geps             = vfeps*vftab[vfitab+2];
 638             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 639             Fp               = F+Geps+Heps2;
 640             FF               = Fp+Geps+2.0*Heps2;
 641             felec            = -qq20*FF*vftabscale*rinv20;
 642
 643             fscal            = felec;
 644
 645             /* Calculate temporary vectorial force */
 646             tx               = fscal*dx20;
 647             ty               = fscal*dy20;
 648             tz               = fscal*dz20;
 649
 650             /* Update vectorial force */
 651             fix2            += tx;
 652             fiy2            += ty;
 653             fiz2            += tz;
 654             f[j_coord_offset+DIM*0+XX] -= tx;
 655             f[j_coord_offset+DIM*0+YY] -= ty;
 656             f[j_coord_offset+DIM*0+ZZ] -= tz;
 657
 658             /* Inner loop uses 150 flops */
 659         }
 660         /* End of innermost loop */
 661
 662         tx = ty = tz = 0;
 663         f[i_coord_offset+DIM*0+XX] += fix0;
 664         f[i_coord_offset+DIM*0+YY] += fiy0;
 665         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 666         tx                         += fix0;
 667         ty                         += fiy0;
 668         tz                         += fiz0;
 669         f[i_coord_offset+DIM*1+XX] += fix1;
 670         f[i_coord_offset+DIM*1+YY] += fiy1;
 671         f[i_coord_offset+DIM*1+ZZ] += fiz1;
 672         tx                         += fix1;
 673         ty                         += fiy1;
 674         tz                         += fiz1;
 675         f[i_coord_offset+DIM*2+XX] += fix2;
 676         f[i_coord_offset+DIM*2+YY] += fiy2;
 677         f[i_coord_offset+DIM*2+ZZ] += fiz2;
 678         tx                         += fix2;
 679         ty                         += fiy2;
 680         tz                         += fiz2;
 681         fshift[i_shift_offset+XX]  += tx;
 682         fshift[i_shift_offset+YY]  += ty;
 683         fshift[i_shift_offset+ZZ]  += tz;
 684
 685         /* Increment number of inner iterations */
 686         inneriter                  += j_index_end - j_index_start;
 687
 688         /* Outer loop uses 30 flops */
 689     }
 690
 691     /* Increment number of outer iterations */
 692     outeriter        += nri;
 693
 694     /* Update outer/inner flops */
 695
 696     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3_F,outeriter*30 + inneriter*150);
 697 }