src/gromacs/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecEw_VdwBham_GeomW4P1_c.c

   1 /*
   2  * This file is part of the GROMACS molecular simulation package.
   3  *
   4  * Copyright (c) 2012,2013, 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 #ifdef HAVE_CONFIG_H
  39 #include <config.h>
  40 #endif
  41
  42 #include <math.h>
  43
  44 #include "../nb_kernel.h"
  45 #include "types/simple.h"
  46 #include "vec.h"
  47 #include "nrnb.h"
  48
  49 /*
  50  * Gromacs nonbonded kernel:   nb_kernel_ElecEw_VdwBham_GeomW4P1_VF_c
  51  * Electrostatics interaction: Ewald
  52  * VdW interaction:            Buckingham
  53  * Geometry:                   Water4-Particle
  54  * Calculate force/pot:        PotentialAndForce
  55  */
  56 void
  57 nb_kernel_ElecEw_VdwBham_GeomW4P1_VF_c
  58                     (t_nblist                    * gmx_restrict       nlist,
  59                      rvec                        * gmx_restrict          xx,
  60                      rvec                        * gmx_restrict          ff,
  61                      t_forcerec                  * gmx_restrict          fr,
  62                      t_mdatoms                   * gmx_restrict     mdatoms,
  63                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
  64                      t_nrnb                      * gmx_restrict        nrnb)
  65 {
  66     int              i_shift_offset,i_coord_offset,j_coord_offset;
  67     int              j_index_start,j_index_end;
  68     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
  69     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
  70     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  71     real             *shiftvec,*fshift,*x,*f;
  72     int              vdwioffset0;
  73     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  74     int              vdwioffset1;
  75     real             ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
  76     int              vdwioffset2;
  77     real             ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
  78     int              vdwioffset3;
  79     real             ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
  80     int              vdwjidx0;
  81     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  82     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
  83     real             dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10;
  84     real             dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20;
  85     real             dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30,cexp1_30,cexp2_30;
  86     real             velec,felec,velecsum,facel,crf,krf,krf2;
  87     real             *charge;
  88     int              nvdwtype;
  89     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
  90     int              *vdwtype;
  91     real             *vdwparam;
  92     int              ewitab;
  93     real             ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
  94     real             *ewtab;
  95
  96     x                = xx[0];
  97     f                = ff[0];
  98
  99     nri              = nlist->nri;
 100     iinr             = nlist->iinr;
 101     jindex           = nlist->jindex;
 102     jjnr             = nlist->jjnr;
 103     shiftidx         = nlist->shift;
 104     gid              = nlist->gid;
 105     shiftvec         = fr->shift_vec[0];
 106     fshift           = fr->fshift[0];
 107     facel            = fr->epsfac;
 108     charge           = mdatoms->chargeA;
 109     nvdwtype         = fr->ntype;
 110     vdwparam         = fr->nbfp;
 111     vdwtype          = mdatoms->typeA;
 112
 113     sh_ewald         = fr->ic->sh_ewald;
 114     ewtab            = fr->ic->tabq_coul_FDV0;
 115     ewtabscale       = fr->ic->tabq_scale;
 116     ewtabhalfspace   = 0.5/ewtabscale;
 117
 118     /* Setup water-specific parameters */
 119     inr              = nlist->iinr[0];
 120     iq1              = facel*charge[inr+1];
 121     iq2              = facel*charge[inr+2];
 122     iq3              = facel*charge[inr+3];
 123     vdwioffset0      = 3*nvdwtype*vdwtype[inr+0];
 124
 125     outeriter        = 0;
 126     inneriter        = 0;
 127
 128     /* Start outer loop over neighborlists */
 129     for(iidx=0; iidx<nri; iidx++)
 130     {
 131         /* Load shift vector for this list */
 132         i_shift_offset   = DIM*shiftidx[iidx];
 133         shX              = shiftvec[i_shift_offset+XX];
 134         shY              = shiftvec[i_shift_offset+YY];
 135         shZ              = shiftvec[i_shift_offset+ZZ];
 136
 137         /* Load limits for loop over neighbors */
 138         j_index_start    = jindex[iidx];
 139         j_index_end      = jindex[iidx+1];
 140
 141         /* Get outer coordinate index */
 142         inr              = iinr[iidx];
 143         i_coord_offset   = DIM*inr;
 144
 145         /* Load i particle coords and add shift vector */
 146         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 147         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 148         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 149         ix1              = shX + x[i_coord_offset+DIM*1+XX];
 150         iy1              = shY + x[i_coord_offset+DIM*1+YY];
 151         iz1              = shZ + x[i_coord_offset+DIM*1+ZZ];
 152         ix2              = shX + x[i_coord_offset+DIM*2+XX];
 153         iy2              = shY + x[i_coord_offset+DIM*2+YY];
 154         iz2              = shZ + x[i_coord_offset+DIM*2+ZZ];
 155         ix3              = shX + x[i_coord_offset+DIM*3+XX];
 156         iy3              = shY + x[i_coord_offset+DIM*3+YY];
 157         iz3              = shZ + x[i_coord_offset+DIM*3+ZZ];
 158
 159         fix0             = 0.0;
 160         fiy0             = 0.0;
 161         fiz0             = 0.0;
 162         fix1             = 0.0;
 163         fiy1             = 0.0;
 164         fiz1             = 0.0;
 165         fix2             = 0.0;
 166         fiy2             = 0.0;
 167         fiz2             = 0.0;
 168         fix3             = 0.0;
 169         fiy3             = 0.0;
 170         fiz3             = 0.0;
 171
 172         /* Reset potential sums */
 173         velecsum         = 0.0;
 174         vvdwsum          = 0.0;
 175
 176         /* Start inner kernel loop */
 177         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 178         {
 179             /* Get j neighbor index, and coordinate index */
 180             jnr              = jjnr[jidx];
 181             j_coord_offset   = DIM*jnr;
 182
 183             /* load j atom coordinates */
 184             jx0              = x[j_coord_offset+DIM*0+XX];
 185             jy0              = x[j_coord_offset+DIM*0+YY];
 186             jz0              = x[j_coord_offset+DIM*0+ZZ];
 187
 188             /* Calculate displacement vector */
 189             dx00             = ix0 - jx0;
 190             dy00             = iy0 - jy0;
 191             dz00             = iz0 - jz0;
 192             dx10             = ix1 - jx0;
 193             dy10             = iy1 - jy0;
 194             dz10             = iz1 - jz0;
 195             dx20             = ix2 - jx0;
 196             dy20             = iy2 - jy0;
 197             dz20             = iz2 - jz0;
 198             dx30             = ix3 - jx0;
 199             dy30             = iy3 - jy0;
 200             dz30             = iz3 - jz0;
 201
 202             /* Calculate squared distance and things based on it */
 203             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 204             rsq10            = dx10*dx10+dy10*dy10+dz10*dz10;
 205             rsq20            = dx20*dx20+dy20*dy20+dz20*dz20;
 206             rsq30            = dx30*dx30+dy30*dy30+dz30*dz30;
 207
 208             rinv00           = gmx_invsqrt(rsq00);
 209             rinv10           = gmx_invsqrt(rsq10);
 210             rinv20           = gmx_invsqrt(rsq20);
 211             rinv30           = gmx_invsqrt(rsq30);
 212
 213             rinvsq00         = rinv00*rinv00;
 214             rinvsq10         = rinv10*rinv10;
 215             rinvsq20         = rinv20*rinv20;
 216             rinvsq30         = rinv30*rinv30;
 217
 218             /* Load parameters for j particles */
 219             jq0              = charge[jnr+0];
 220             vdwjidx0         = 3*vdwtype[jnr+0];
 221
 222             /**************************
 223              * CALCULATE INTERACTIONS *
 224              **************************/
 225
 226             r00              = rsq00*rinv00;
 227
 228             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 229             cexp1_00         = vdwparam[vdwioffset0+vdwjidx0+1];
 230             cexp2_00         = vdwparam[vdwioffset0+vdwjidx0+2];
 231
 232             /* BUCKINGHAM DISPERSION/REPULSION */
 233             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 234             vvdw6            = c6_00*rinvsix;
 235             br               = cexp2_00*r00;
 236             vvdwexp          = cexp1_00*exp(-br);
 237             vvdw             = vvdwexp - vvdw6*(1.0/6.0);
 238             fvdw             = (br*vvdwexp-vvdw6)*rinvsq00;
 239
 240             /* Update potential sums from outer loop */
 241             vvdwsum         += vvdw;
 242
 243             fscal            = fvdw;
 244
 245             /* Calculate temporary vectorial force */
 246             tx               = fscal*dx00;
 247             ty               = fscal*dy00;
 248             tz               = fscal*dz00;
 249
 250             /* Update vectorial force */
 251             fix0            += tx;
 252             fiy0            += ty;
 253             fiz0            += tz;
 254             f[j_coord_offset+DIM*0+XX] -= tx;
 255             f[j_coord_offset+DIM*0+YY] -= ty;
 256             f[j_coord_offset+DIM*0+ZZ] -= tz;
 257
 258             /**************************
 259              * CALCULATE INTERACTIONS *
 260              **************************/
 261
 262             r10              = rsq10*rinv10;
 263
 264             qq10             = iq1*jq0;
 265
 266             /* EWALD ELECTROSTATICS */
 267
 268             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 269             ewrt             = r10*ewtabscale;
 270             ewitab           = ewrt;
 271             eweps            = ewrt-ewitab;
 272             ewitab           = 4*ewitab;
 273             felec            = ewtab[ewitab]+eweps*ewtab[ewitab+1];
 274             velec            = qq10*(rinv10-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
 275             felec            = qq10*rinv10*(rinvsq10-felec);
 276
 277             /* Update potential sums from outer loop */
 278             velecsum        += velec;
 279
 280             fscal            = felec;
 281
 282             /* Calculate temporary vectorial force */
 283             tx               = fscal*dx10;
 284             ty               = fscal*dy10;
 285             tz               = fscal*dz10;
 286
 287             /* Update vectorial force */
 288             fix1            += tx;
 289             fiy1            += ty;
 290             fiz1            += tz;
 291             f[j_coord_offset+DIM*0+XX] -= tx;
 292             f[j_coord_offset+DIM*0+YY] -= ty;
 293             f[j_coord_offset+DIM*0+ZZ] -= tz;
 294
 295             /**************************
 296              * CALCULATE INTERACTIONS *
 297              **************************/
 298
 299             r20              = rsq20*rinv20;
 300
 301             qq20             = iq2*jq0;
 302
 303             /* EWALD ELECTROSTATICS */
 304
 305             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 306             ewrt             = r20*ewtabscale;
 307             ewitab           = ewrt;
 308             eweps            = ewrt-ewitab;
 309             ewitab           = 4*ewitab;
 310             felec            = ewtab[ewitab]+eweps*ewtab[ewitab+1];
 311             velec            = qq20*(rinv20-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
 312             felec            = qq20*rinv20*(rinvsq20-felec);
 313
 314             /* Update potential sums from outer loop */
 315             velecsum        += velec;
 316
 317             fscal            = felec;
 318
 319             /* Calculate temporary vectorial force */
 320             tx               = fscal*dx20;
 321             ty               = fscal*dy20;
 322             tz               = fscal*dz20;
 323
 324             /* Update vectorial force */
 325             fix2            += tx;
 326             fiy2            += ty;
 327             fiz2            += tz;
 328             f[j_coord_offset+DIM*0+XX] -= tx;
 329             f[j_coord_offset+DIM*0+YY] -= ty;
 330             f[j_coord_offset+DIM*0+ZZ] -= tz;
 331
 332             /**************************
 333              * CALCULATE INTERACTIONS *
 334              **************************/
 335
 336             r30              = rsq30*rinv30;
 337
 338             qq30             = iq3*jq0;
 339
 340             /* EWALD ELECTROSTATICS */
 341
 342             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 343             ewrt             = r30*ewtabscale;
 344             ewitab           = ewrt;
 345             eweps            = ewrt-ewitab;
 346             ewitab           = 4*ewitab;
 347             felec            = ewtab[ewitab]+eweps*ewtab[ewitab+1];
 348             velec            = qq30*(rinv30-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
 349             felec            = qq30*rinv30*(rinvsq30-felec);
 350
 351             /* Update potential sums from outer loop */
 352             velecsum        += velec;
 353
 354             fscal            = felec;
 355
 356             /* Calculate temporary vectorial force */
 357             tx               = fscal*dx30;
 358             ty               = fscal*dy30;
 359             tz               = fscal*dz30;
 360
 361             /* Update vectorial force */
 362             fix3            += tx;
 363             fiy3            += ty;
 364             fiz3            += tz;
 365             f[j_coord_offset+DIM*0+XX] -= tx;
 366             f[j_coord_offset+DIM*0+YY] -= ty;
 367             f[j_coord_offset+DIM*0+ZZ] -= tz;
 368
 369             /* Inner loop uses 184 flops */
 370         }
 371         /* End of innermost loop */
 372
 373         tx = ty = tz = 0;
 374         f[i_coord_offset+DIM*0+XX] += fix0;
 375         f[i_coord_offset+DIM*0+YY] += fiy0;
 376         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 377         tx                         += fix0;
 378         ty                         += fiy0;
 379         tz                         += fiz0;
 380         f[i_coord_offset+DIM*1+XX] += fix1;
 381         f[i_coord_offset+DIM*1+YY] += fiy1;
 382         f[i_coord_offset+DIM*1+ZZ] += fiz1;
 383         tx                         += fix1;
 384         ty                         += fiy1;
 385         tz                         += fiz1;
 386         f[i_coord_offset+DIM*2+XX] += fix2;
 387         f[i_coord_offset+DIM*2+YY] += fiy2;
 388         f[i_coord_offset+DIM*2+ZZ] += fiz2;
 389         tx                         += fix2;
 390         ty                         += fiy2;
 391         tz                         += fiz2;
 392         f[i_coord_offset+DIM*3+XX] += fix3;
 393         f[i_coord_offset+DIM*3+YY] += fiy3;
 394         f[i_coord_offset+DIM*3+ZZ] += fiz3;
 395         tx                         += fix3;
 396         ty                         += fiy3;
 397         tz                         += fiz3;
 398         fshift[i_shift_offset+XX]  += tx;
 399         fshift[i_shift_offset+YY]  += ty;
 400         fshift[i_shift_offset+ZZ]  += tz;
 401
 402         ggid                        = gid[iidx];
 403         /* Update potential energies */
 404         kernel_data->energygrp_elec[ggid] += velecsum;
 405         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 406
 407         /* Increment number of inner iterations */
 408         inneriter                  += j_index_end - j_index_start;
 409
 410         /* Outer loop uses 41 flops */
 411     }
 412
 413     /* Increment number of outer iterations */
 414     outeriter        += nri;
 415
 416     /* Update outer/inner flops */
 417
 418     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_VF,outeriter*41 + inneriter*184);
 419 }
 420 /*
 421  * Gromacs nonbonded kernel:   nb_kernel_ElecEw_VdwBham_GeomW4P1_F_c
 422  * Electrostatics interaction: Ewald
 423  * VdW interaction:            Buckingham
 424  * Geometry:                   Water4-Particle
 425  * Calculate force/pot:        Force
 426  */
 427 void
 428 nb_kernel_ElecEw_VdwBham_GeomW4P1_F_c
 429                     (t_nblist                    * gmx_restrict       nlist,
 430                      rvec                        * gmx_restrict          xx,
 431                      rvec                        * gmx_restrict          ff,
 432                      t_forcerec                  * gmx_restrict          fr,
 433                      t_mdatoms                   * gmx_restrict     mdatoms,
 434                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 435                      t_nrnb                      * gmx_restrict        nrnb)
 436 {
 437     int              i_shift_offset,i_coord_offset,j_coord_offset;
 438     int              j_index_start,j_index_end;
 439     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 440     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 441     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 442     real             *shiftvec,*fshift,*x,*f;
 443     int              vdwioffset0;
 444     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 445     int              vdwioffset1;
 446     real             ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
 447     int              vdwioffset2;
 448     real             ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
 449     int              vdwioffset3;
 450     real             ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
 451     int              vdwjidx0;
 452     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 453     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 454     real             dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10;
 455     real             dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20;
 456     real             dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30,cexp1_30,cexp2_30;
 457     real             velec,felec,velecsum,facel,crf,krf,krf2;
 458     real             *charge;
 459     int              nvdwtype;
 460     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 461     int              *vdwtype;
 462     real             *vdwparam;
 463     int              ewitab;
 464     real             ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
 465     real             *ewtab;
 466
 467     x                = xx[0];
 468     f                = ff[0];
 469
 470     nri              = nlist->nri;
 471     iinr             = nlist->iinr;
 472     jindex           = nlist->jindex;
 473     jjnr             = nlist->jjnr;
 474     shiftidx         = nlist->shift;
 475     gid              = nlist->gid;
 476     shiftvec         = fr->shift_vec[0];
 477     fshift           = fr->fshift[0];
 478     facel            = fr->epsfac;
 479     charge           = mdatoms->chargeA;
 480     nvdwtype         = fr->ntype;
 481     vdwparam         = fr->nbfp;
 482     vdwtype          = mdatoms->typeA;
 483
 484     sh_ewald         = fr->ic->sh_ewald;
 485     ewtab            = fr->ic->tabq_coul_F;
 486     ewtabscale       = fr->ic->tabq_scale;
 487     ewtabhalfspace   = 0.5/ewtabscale;
 488
 489     /* Setup water-specific parameters */
 490     inr              = nlist->iinr[0];
 491     iq1              = facel*charge[inr+1];
 492     iq2              = facel*charge[inr+2];
 493     iq3              = facel*charge[inr+3];
 494     vdwioffset0      = 3*nvdwtype*vdwtype[inr+0];
 495
 496     outeriter        = 0;
 497     inneriter        = 0;
 498
 499     /* Start outer loop over neighborlists */
 500     for(iidx=0; iidx<nri; iidx++)
 501     {
 502         /* Load shift vector for this list */
 503         i_shift_offset   = DIM*shiftidx[iidx];
 504         shX              = shiftvec[i_shift_offset+XX];
 505         shY              = shiftvec[i_shift_offset+YY];
 506         shZ              = shiftvec[i_shift_offset+ZZ];
 507
 508         /* Load limits for loop over neighbors */
 509         j_index_start    = jindex[iidx];
 510         j_index_end      = jindex[iidx+1];
 511
 512         /* Get outer coordinate index */
 513         inr              = iinr[iidx];
 514         i_coord_offset   = DIM*inr;
 515
 516         /* Load i particle coords and add shift vector */
 517         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 518         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 519         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 520         ix1              = shX + x[i_coord_offset+DIM*1+XX];
 521         iy1              = shY + x[i_coord_offset+DIM*1+YY];
 522         iz1              = shZ + x[i_coord_offset+DIM*1+ZZ];
 523         ix2              = shX + x[i_coord_offset+DIM*2+XX];
 524         iy2              = shY + x[i_coord_offset+DIM*2+YY];
 525         iz2              = shZ + x[i_coord_offset+DIM*2+ZZ];
 526         ix3              = shX + x[i_coord_offset+DIM*3+XX];
 527         iy3              = shY + x[i_coord_offset+DIM*3+YY];
 528         iz3              = shZ + x[i_coord_offset+DIM*3+ZZ];
 529
 530         fix0             = 0.0;
 531         fiy0             = 0.0;
 532         fiz0             = 0.0;
 533         fix1             = 0.0;
 534         fiy1             = 0.0;
 535         fiz1             = 0.0;
 536         fix2             = 0.0;
 537         fiy2             = 0.0;
 538         fiz2             = 0.0;
 539         fix3             = 0.0;
 540         fiy3             = 0.0;
 541         fiz3             = 0.0;
 542
 543         /* Start inner kernel loop */
 544         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 545         {
 546             /* Get j neighbor index, and coordinate index */
 547             jnr              = jjnr[jidx];
 548             j_coord_offset   = DIM*jnr;
 549
 550             /* load j atom coordinates */
 551             jx0              = x[j_coord_offset+DIM*0+XX];
 552             jy0              = x[j_coord_offset+DIM*0+YY];
 553             jz0              = x[j_coord_offset+DIM*0+ZZ];
 554
 555             /* Calculate displacement vector */
 556             dx00             = ix0 - jx0;
 557             dy00             = iy0 - jy0;
 558             dz00             = iz0 - jz0;
 559             dx10             = ix1 - jx0;
 560             dy10             = iy1 - jy0;
 561             dz10             = iz1 - jz0;
 562             dx20             = ix2 - jx0;
 563             dy20             = iy2 - jy0;
 564             dz20             = iz2 - jz0;
 565             dx30             = ix3 - jx0;
 566             dy30             = iy3 - jy0;
 567             dz30             = iz3 - jz0;
 568
 569             /* Calculate squared distance and things based on it */
 570             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 571             rsq10            = dx10*dx10+dy10*dy10+dz10*dz10;
 572             rsq20            = dx20*dx20+dy20*dy20+dz20*dz20;
 573             rsq30            = dx30*dx30+dy30*dy30+dz30*dz30;
 574
 575             rinv00           = gmx_invsqrt(rsq00);
 576             rinv10           = gmx_invsqrt(rsq10);
 577             rinv20           = gmx_invsqrt(rsq20);
 578             rinv30           = gmx_invsqrt(rsq30);
 579
 580             rinvsq00         = rinv00*rinv00;
 581             rinvsq10         = rinv10*rinv10;
 582             rinvsq20         = rinv20*rinv20;
 583             rinvsq30         = rinv30*rinv30;
 584
 585             /* Load parameters for j particles */
 586             jq0              = charge[jnr+0];
 587             vdwjidx0         = 3*vdwtype[jnr+0];
 588
 589             /**************************
 590              * CALCULATE INTERACTIONS *
 591              **************************/
 592
 593             r00              = rsq00*rinv00;
 594
 595             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 596             cexp1_00         = vdwparam[vdwioffset0+vdwjidx0+1];
 597             cexp2_00         = vdwparam[vdwioffset0+vdwjidx0+2];
 598
 599             /* BUCKINGHAM DISPERSION/REPULSION */
 600             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 601             vvdw6            = c6_00*rinvsix;
 602             br               = cexp2_00*r00;
 603             vvdwexp          = cexp1_00*exp(-br);
 604             fvdw             = (br*vvdwexp-vvdw6)*rinvsq00;
 605
 606             fscal            = fvdw;
 607
 608             /* Calculate temporary vectorial force */
 609             tx               = fscal*dx00;
 610             ty               = fscal*dy00;
 611             tz               = fscal*dz00;
 612
 613             /* Update vectorial force */
 614             fix0            += tx;
 615             fiy0            += ty;
 616             fiz0            += 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             r10              = rsq10*rinv10;
 626
 627             qq10             = iq1*jq0;
 628
 629             /* EWALD ELECTROSTATICS */
 630
 631             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 632             ewrt             = r10*ewtabscale;
 633             ewitab           = ewrt;
 634             eweps            = ewrt-ewitab;
 635             felec            = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1];
 636             felec            = qq10*rinv10*(rinvsq10-felec);
 637
 638             fscal            = felec;
 639
 640             /* Calculate temporary vectorial force */
 641             tx               = fscal*dx10;
 642             ty               = fscal*dy10;
 643             tz               = fscal*dz10;
 644
 645             /* Update vectorial force */
 646             fix1            += tx;
 647             fiy1            += ty;
 648             fiz1            += tz;
 649             f[j_coord_offset+DIM*0+XX] -= tx;
 650             f[j_coord_offset+DIM*0+YY] -= ty;
 651             f[j_coord_offset+DIM*0+ZZ] -= tz;
 652
 653             /**************************
 654              * CALCULATE INTERACTIONS *
 655              **************************/
 656
 657             r20              = rsq20*rinv20;
 658
 659             qq20             = iq2*jq0;
 660
 661             /* EWALD ELECTROSTATICS */
 662
 663             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 664             ewrt             = r20*ewtabscale;
 665             ewitab           = ewrt;
 666             eweps            = ewrt-ewitab;
 667             felec            = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1];
 668             felec            = qq20*rinv20*(rinvsq20-felec);
 669
 670             fscal            = felec;
 671
 672             /* Calculate temporary vectorial force */
 673             tx               = fscal*dx20;
 674             ty               = fscal*dy20;
 675             tz               = fscal*dz20;
 676
 677             /* Update vectorial force */
 678             fix2            += tx;
 679             fiy2            += ty;
 680             fiz2            += tz;
 681             f[j_coord_offset+DIM*0+XX] -= tx;
 682             f[j_coord_offset+DIM*0+YY] -= ty;
 683             f[j_coord_offset+DIM*0+ZZ] -= tz;
 684
 685             /**************************
 686              * CALCULATE INTERACTIONS *
 687              **************************/
 688
 689             r30              = rsq30*rinv30;
 690
 691             qq30             = iq3*jq0;
 692
 693             /* EWALD ELECTROSTATICS */
 694
 695             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 696             ewrt             = r30*ewtabscale;
 697             ewitab           = ewrt;
 698             eweps            = ewrt-ewitab;
 699             felec            = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1];
 700             felec            = qq30*rinv30*(rinvsq30-felec);
 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 160 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*160);
 762 }