src/gromacs/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecEw_VdwCSTab_GeomW3P1_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_VdwCSTab_GeomW3P1_VF_c
  51  * Electrostatics interaction: Ewald
  52  * VdW interaction:            CubicSplineTable
  53  * Geometry:                   Water3-Particle
  54  * Calculate force/pot:        PotentialAndForce
  55  */
  56 void
  57 nb_kernel_ElecEw_VdwCSTab_GeomW3P1_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              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             velec,felec,velecsum,facel,crf,krf,krf2;
  84     real             *charge;
  85     int              nvdwtype;
  86     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
  87     int              *vdwtype;
  88     real             *vdwparam;
  89     int              vfitab;
  90     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
  91     real             *vftab;
  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     vftab            = kernel_data->table_vdw->data;
 114     vftabscale       = kernel_data->table_vdw->scale;
 115
 116     sh_ewald         = fr->ic->sh_ewald;
 117     ewtab            = fr->ic->tabq_coul_FDV0;
 118     ewtabscale       = fr->ic->tabq_scale;
 119     ewtabhalfspace   = 0.5/ewtabscale;
 120
 121     /* Setup water-specific parameters */
 122     inr              = nlist->iinr[0];
 123     iq0              = facel*charge[inr+0];
 124     iq1              = facel*charge[inr+1];
 125     iq2              = facel*charge[inr+2];
 126     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 127
 128     outeriter        = 0;
 129     inneriter        = 0;
 130
 131     /* Start outer loop over neighborlists */
 132     for(iidx=0; iidx<nri; iidx++)
 133     {
 134         /* Load shift vector for this list */
 135         i_shift_offset   = DIM*shiftidx[iidx];
 136         shX              = shiftvec[i_shift_offset+XX];
 137         shY              = shiftvec[i_shift_offset+YY];
 138         shZ              = shiftvec[i_shift_offset+ZZ];
 139
 140         /* Load limits for loop over neighbors */
 141         j_index_start    = jindex[iidx];
 142         j_index_end      = jindex[iidx+1];
 143
 144         /* Get outer coordinate index */
 145         inr              = iinr[iidx];
 146         i_coord_offset   = DIM*inr;
 147
 148         /* Load i particle coords and add shift vector */
 149         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 150         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 151         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 152         ix1              = shX + x[i_coord_offset+DIM*1+XX];
 153         iy1              = shY + x[i_coord_offset+DIM*1+YY];
 154         iz1              = shZ + x[i_coord_offset+DIM*1+ZZ];
 155         ix2              = shX + x[i_coord_offset+DIM*2+XX];
 156         iy2              = shY + x[i_coord_offset+DIM*2+YY];
 157         iz2              = shZ + x[i_coord_offset+DIM*2+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
 169         /* Reset potential sums */
 170         velecsum         = 0.0;
 171         vvdwsum          = 0.0;
 172
 173         /* Start inner kernel loop */
 174         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 175         {
 176             /* Get j neighbor index, and coordinate index */
 177             jnr              = jjnr[jidx];
 178             j_coord_offset   = DIM*jnr;
 179
 180             /* load j atom coordinates */
 181             jx0              = x[j_coord_offset+DIM*0+XX];
 182             jy0              = x[j_coord_offset+DIM*0+YY];
 183             jz0              = x[j_coord_offset+DIM*0+ZZ];
 184
 185             /* Calculate displacement vector */
 186             dx00             = ix0 - jx0;
 187             dy00             = iy0 - jy0;
 188             dz00             = iz0 - jz0;
 189             dx10             = ix1 - jx0;
 190             dy10             = iy1 - jy0;
 191             dz10             = iz1 - jz0;
 192             dx20             = ix2 - jx0;
 193             dy20             = iy2 - jy0;
 194             dz20             = iz2 - jz0;
 195
 196             /* Calculate squared distance and things based on it */
 197             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 198             rsq10            = dx10*dx10+dy10*dy10+dz10*dz10;
 199             rsq20            = dx20*dx20+dy20*dy20+dz20*dz20;
 200
 201             rinv00           = gmx_invsqrt(rsq00);
 202             rinv10           = gmx_invsqrt(rsq10);
 203             rinv20           = gmx_invsqrt(rsq20);
 204
 205             rinvsq00         = rinv00*rinv00;
 206             rinvsq10         = rinv10*rinv10;
 207             rinvsq20         = rinv20*rinv20;
 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             qq00             = iq0*jq0;
 220             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 221             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 222
 223             /* Calculate table index by multiplying r with table scale and truncate to integer */
 224             rt               = r00*vftabscale;
 225             vfitab           = rt;
 226             vfeps            = rt-vfitab;
 227             vfitab           = 2*4*vfitab;
 228
 229             /* EWALD ELECTROSTATICS */
 230
 231             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 232             ewrt             = r00*ewtabscale;
 233             ewitab           = ewrt;
 234             eweps            = ewrt-ewitab;
 235             ewitab           = 4*ewitab;
 236             felec            = ewtab[ewitab]+eweps*ewtab[ewitab+1];
 237             velec            = qq00*(rinv00-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
 238             felec            = qq00*rinv00*(rinvsq00-felec);
 239
 240             /* CUBIC SPLINE TABLE DISPERSION */
 241             vfitab          += 0;
 242             Y                = vftab[vfitab];
 243             F                = vftab[vfitab+1];
 244             Geps             = vfeps*vftab[vfitab+2];
 245             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 246             Fp               = F+Geps+Heps2;
 247             VV               = Y+vfeps*Fp;
 248             vvdw6            = c6_00*VV;
 249             FF               = Fp+Geps+2.0*Heps2;
 250             fvdw6            = c6_00*FF;
 251
 252             /* CUBIC SPLINE TABLE REPULSION */
 253             Y                = vftab[vfitab+4];
 254             F                = vftab[vfitab+5];
 255             Geps             = vfeps*vftab[vfitab+6];
 256             Heps2            = vfeps*vfeps*vftab[vfitab+7];
 257             Fp               = F+Geps+Heps2;
 258             VV               = Y+vfeps*Fp;
 259             vvdw12           = c12_00*VV;
 260             FF               = Fp+Geps+2.0*Heps2;
 261             fvdw12           = c12_00*FF;
 262             vvdw             = vvdw12+vvdw6;
 263             fvdw             = -(fvdw6+fvdw12)*vftabscale*rinv00;
 264
 265             /* Update potential sums from outer loop */
 266             velecsum        += velec;
 267             vvdwsum         += vvdw;
 268
 269             fscal            = felec+fvdw;
 270
 271             /* Calculate temporary vectorial force */
 272             tx               = fscal*dx00;
 273             ty               = fscal*dy00;
 274             tz               = fscal*dz00;
 275
 276             /* Update vectorial force */
 277             fix0            += tx;
 278             fiy0            += ty;
 279             fiz0            += tz;
 280             f[j_coord_offset+DIM*0+XX] -= tx;
 281             f[j_coord_offset+DIM*0+YY] -= ty;
 282             f[j_coord_offset+DIM*0+ZZ] -= tz;
 283
 284             /**************************
 285              * CALCULATE INTERACTIONS *
 286              **************************/
 287
 288             r10              = rsq10*rinv10;
 289
 290             qq10             = iq1*jq0;
 291
 292             /* EWALD ELECTROSTATICS */
 293
 294             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 295             ewrt             = r10*ewtabscale;
 296             ewitab           = ewrt;
 297             eweps            = ewrt-ewitab;
 298             ewitab           = 4*ewitab;
 299             felec            = ewtab[ewitab]+eweps*ewtab[ewitab+1];
 300             velec            = qq10*(rinv10-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
 301             felec            = qq10*rinv10*(rinvsq10-felec);
 302
 303             /* Update potential sums from outer loop */
 304             velecsum        += velec;
 305
 306             fscal            = felec;
 307
 308             /* Calculate temporary vectorial force */
 309             tx               = fscal*dx10;
 310             ty               = fscal*dy10;
 311             tz               = fscal*dz10;
 312
 313             /* Update vectorial force */
 314             fix1            += tx;
 315             fiy1            += ty;
 316             fiz1            += tz;
 317             f[j_coord_offset+DIM*0+XX] -= tx;
 318             f[j_coord_offset+DIM*0+YY] -= ty;
 319             f[j_coord_offset+DIM*0+ZZ] -= tz;
 320
 321             /**************************
 322              * CALCULATE INTERACTIONS *
 323              **************************/
 324
 325             r20              = rsq20*rinv20;
 326
 327             qq20             = iq2*jq0;
 328
 329             /* EWALD ELECTROSTATICS */
 330
 331             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 332             ewrt             = r20*ewtabscale;
 333             ewitab           = ewrt;
 334             eweps            = ewrt-ewitab;
 335             ewitab           = 4*ewitab;
 336             felec            = ewtab[ewitab]+eweps*ewtab[ewitab+1];
 337             velec            = qq20*(rinv20-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
 338             felec            = qq20*rinv20*(rinvsq20-felec);
 339
 340             /* Update potential sums from outer loop */
 341             velecsum        += velec;
 342
 343             fscal            = felec;
 344
 345             /* Calculate temporary vectorial force */
 346             tx               = fscal*dx20;
 347             ty               = fscal*dy20;
 348             tz               = fscal*dz20;
 349
 350             /* Update vectorial force */
 351             fix2            += tx;
 352             fiy2            += ty;
 353             fiz2            += tz;
 354             f[j_coord_offset+DIM*0+XX] -= tx;
 355             f[j_coord_offset+DIM*0+YY] -= ty;
 356             f[j_coord_offset+DIM*0+ZZ] -= tz;
 357
 358             /* Inner loop uses 156 flops */
 359         }
 360         /* End of innermost loop */
 361
 362         tx = ty = tz = 0;
 363         f[i_coord_offset+DIM*0+XX] += fix0;
 364         f[i_coord_offset+DIM*0+YY] += fiy0;
 365         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 366         tx                         += fix0;
 367         ty                         += fiy0;
 368         tz                         += fiz0;
 369         f[i_coord_offset+DIM*1+XX] += fix1;
 370         f[i_coord_offset+DIM*1+YY] += fiy1;
 371         f[i_coord_offset+DIM*1+ZZ] += fiz1;
 372         tx                         += fix1;
 373         ty                         += fiy1;
 374         tz                         += fiz1;
 375         f[i_coord_offset+DIM*2+XX] += fix2;
 376         f[i_coord_offset+DIM*2+YY] += fiy2;
 377         f[i_coord_offset+DIM*2+ZZ] += fiz2;
 378         tx                         += fix2;
 379         ty                         += fiy2;
 380         tz                         += fiz2;
 381         fshift[i_shift_offset+XX]  += tx;
 382         fshift[i_shift_offset+YY]  += ty;
 383         fshift[i_shift_offset+ZZ]  += tz;
 384
 385         ggid                        = gid[iidx];
 386         /* Update potential energies */
 387         kernel_data->energygrp_elec[ggid] += velecsum;
 388         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 389
 390         /* Increment number of inner iterations */
 391         inneriter                  += j_index_end - j_index_start;
 392
 393         /* Outer loop uses 32 flops */
 394     }
 395
 396     /* Increment number of outer iterations */
 397     outeriter        += nri;
 398
 399     /* Update outer/inner flops */
 400
 401     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3_VF,outeriter*32 + inneriter*156);
 402 }
 403 /*
 404  * Gromacs nonbonded kernel:   nb_kernel_ElecEw_VdwCSTab_GeomW3P1_F_c
 405  * Electrostatics interaction: Ewald
 406  * VdW interaction:            CubicSplineTable
 407  * Geometry:                   Water3-Particle
 408  * Calculate force/pot:        Force
 409  */
 410 void
 411 nb_kernel_ElecEw_VdwCSTab_GeomW3P1_F_c
 412                     (t_nblist                    * gmx_restrict       nlist,
 413                      rvec                        * gmx_restrict          xx,
 414                      rvec                        * gmx_restrict          ff,
 415                      t_forcerec                  * gmx_restrict          fr,
 416                      t_mdatoms                   * gmx_restrict     mdatoms,
 417                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 418                      t_nrnb                      * gmx_restrict        nrnb)
 419 {
 420     int              i_shift_offset,i_coord_offset,j_coord_offset;
 421     int              j_index_start,j_index_end;
 422     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 423     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 424     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 425     real             *shiftvec,*fshift,*x,*f;
 426     int              vdwioffset0;
 427     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 428     int              vdwioffset1;
 429     real             ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
 430     int              vdwioffset2;
 431     real             ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
 432     int              vdwjidx0;
 433     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 434     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 435     real             dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10;
 436     real             dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20;
 437     real             velec,felec,velecsum,facel,crf,krf,krf2;
 438     real             *charge;
 439     int              nvdwtype;
 440     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 441     int              *vdwtype;
 442     real             *vdwparam;
 443     int              vfitab;
 444     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
 445     real             *vftab;
 446     int              ewitab;
 447     real             ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
 448     real             *ewtab;
 449
 450     x                = xx[0];
 451     f                = ff[0];
 452
 453     nri              = nlist->nri;
 454     iinr             = nlist->iinr;
 455     jindex           = nlist->jindex;
 456     jjnr             = nlist->jjnr;
 457     shiftidx         = nlist->shift;
 458     gid              = nlist->gid;
 459     shiftvec         = fr->shift_vec[0];
 460     fshift           = fr->fshift[0];
 461     facel            = fr->epsfac;
 462     charge           = mdatoms->chargeA;
 463     nvdwtype         = fr->ntype;
 464     vdwparam         = fr->nbfp;
 465     vdwtype          = mdatoms->typeA;
 466
 467     vftab            = kernel_data->table_vdw->data;
 468     vftabscale       = kernel_data->table_vdw->scale;
 469
 470     sh_ewald         = fr->ic->sh_ewald;
 471     ewtab            = fr->ic->tabq_coul_F;
 472     ewtabscale       = fr->ic->tabq_scale;
 473     ewtabhalfspace   = 0.5/ewtabscale;
 474
 475     /* Setup water-specific parameters */
 476     inr              = nlist->iinr[0];
 477     iq0              = facel*charge[inr+0];
 478     iq1              = facel*charge[inr+1];
 479     iq2              = facel*charge[inr+2];
 480     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 481
 482     outeriter        = 0;
 483     inneriter        = 0;
 484
 485     /* Start outer loop over neighborlists */
 486     for(iidx=0; iidx<nri; iidx++)
 487     {
 488         /* Load shift vector for this list */
 489         i_shift_offset   = DIM*shiftidx[iidx];
 490         shX              = shiftvec[i_shift_offset+XX];
 491         shY              = shiftvec[i_shift_offset+YY];
 492         shZ              = shiftvec[i_shift_offset+ZZ];
 493
 494         /* Load limits for loop over neighbors */
 495         j_index_start    = jindex[iidx];
 496         j_index_end      = jindex[iidx+1];
 497
 498         /* Get outer coordinate index */
 499         inr              = iinr[iidx];
 500         i_coord_offset   = DIM*inr;
 501
 502         /* Load i particle coords and add shift vector */
 503         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 504         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 505         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 506         ix1              = shX + x[i_coord_offset+DIM*1+XX];
 507         iy1              = shY + x[i_coord_offset+DIM*1+YY];
 508         iz1              = shZ + x[i_coord_offset+DIM*1+ZZ];
 509         ix2              = shX + x[i_coord_offset+DIM*2+XX];
 510         iy2              = shY + x[i_coord_offset+DIM*2+YY];
 511         iz2              = shZ + x[i_coord_offset+DIM*2+ZZ];
 512
 513         fix0             = 0.0;
 514         fiy0             = 0.0;
 515         fiz0             = 0.0;
 516         fix1             = 0.0;
 517         fiy1             = 0.0;
 518         fiz1             = 0.0;
 519         fix2             = 0.0;
 520         fiy2             = 0.0;
 521         fiz2             = 0.0;
 522
 523         /* Start inner kernel loop */
 524         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 525         {
 526             /* Get j neighbor index, and coordinate index */
 527             jnr              = jjnr[jidx];
 528             j_coord_offset   = DIM*jnr;
 529
 530             /* load j atom coordinates */
 531             jx0              = x[j_coord_offset+DIM*0+XX];
 532             jy0              = x[j_coord_offset+DIM*0+YY];
 533             jz0              = x[j_coord_offset+DIM*0+ZZ];
 534
 535             /* Calculate displacement vector */
 536             dx00             = ix0 - jx0;
 537             dy00             = iy0 - jy0;
 538             dz00             = iz0 - jz0;
 539             dx10             = ix1 - jx0;
 540             dy10             = iy1 - jy0;
 541             dz10             = iz1 - jz0;
 542             dx20             = ix2 - jx0;
 543             dy20             = iy2 - jy0;
 544             dz20             = iz2 - jz0;
 545
 546             /* Calculate squared distance and things based on it */
 547             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 548             rsq10            = dx10*dx10+dy10*dy10+dz10*dz10;
 549             rsq20            = dx20*dx20+dy20*dy20+dz20*dz20;
 550
 551             rinv00           = gmx_invsqrt(rsq00);
 552             rinv10           = gmx_invsqrt(rsq10);
 553             rinv20           = gmx_invsqrt(rsq20);
 554
 555             rinvsq00         = rinv00*rinv00;
 556             rinvsq10         = rinv10*rinv10;
 557             rinvsq20         = rinv20*rinv20;
 558
 559             /* Load parameters for j particles */
 560             jq0              = charge[jnr+0];
 561             vdwjidx0         = 2*vdwtype[jnr+0];
 562
 563             /**************************
 564              * CALCULATE INTERACTIONS *
 565              **************************/
 566
 567             r00              = rsq00*rinv00;
 568
 569             qq00             = iq0*jq0;
 570             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 571             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 572
 573             /* Calculate table index by multiplying r with table scale and truncate to integer */
 574             rt               = r00*vftabscale;
 575             vfitab           = rt;
 576             vfeps            = rt-vfitab;
 577             vfitab           = 2*4*vfitab;
 578
 579             /* EWALD ELECTROSTATICS */
 580
 581             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 582             ewrt             = r00*ewtabscale;
 583             ewitab           = ewrt;
 584             eweps            = ewrt-ewitab;
 585             felec            = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1];
 586             felec            = qq00*rinv00*(rinvsq00-felec);
 587
 588             /* CUBIC SPLINE TABLE DISPERSION */
 589             vfitab          += 0;
 590             F                = vftab[vfitab+1];
 591             Geps             = vfeps*vftab[vfitab+2];
 592             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 593             Fp               = F+Geps+Heps2;
 594             FF               = Fp+Geps+2.0*Heps2;
 595             fvdw6            = c6_00*FF;
 596
 597             /* CUBIC SPLINE TABLE REPULSION */
 598             F                = vftab[vfitab+5];
 599             Geps             = vfeps*vftab[vfitab+6];
 600             Heps2            = vfeps*vfeps*vftab[vfitab+7];
 601             Fp               = F+Geps+Heps2;
 602             FF               = Fp+Geps+2.0*Heps2;
 603             fvdw12           = c12_00*FF;
 604             fvdw             = -(fvdw6+fvdw12)*vftabscale*rinv00;
 605
 606             fscal            = felec+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             /* Inner loop uses 127 flops */
 686         }
 687         /* End of innermost loop */
 688
 689         tx = ty = tz = 0;
 690         f[i_coord_offset+DIM*0+XX] += fix0;
 691         f[i_coord_offset+DIM*0+YY] += fiy0;
 692         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 693         tx                         += fix0;
 694         ty                         += fiy0;
 695         tz                         += fiz0;
 696         f[i_coord_offset+DIM*1+XX] += fix1;
 697         f[i_coord_offset+DIM*1+YY] += fiy1;
 698         f[i_coord_offset+DIM*1+ZZ] += fiz1;
 699         tx                         += fix1;
 700         ty                         += fiy1;
 701         tz                         += fiz1;
 702         f[i_coord_offset+DIM*2+XX] += fix2;
 703         f[i_coord_offset+DIM*2+YY] += fiy2;
 704         f[i_coord_offset+DIM*2+ZZ] += fiz2;
 705         tx                         += fix2;
 706         ty                         += fiy2;
 707         tz                         += fiz2;
 708         fshift[i_shift_offset+XX]  += tx;
 709         fshift[i_shift_offset+YY]  += ty;
 710         fshift[i_shift_offset+ZZ]  += tz;
 711
 712         /* Increment number of inner iterations */
 713         inneriter                  += j_index_end - j_index_start;
 714
 715         /* Outer loop uses 30 flops */
 716     }
 717
 718     /* Increment number of outer iterations */
 719     outeriter        += nri;
 720
 721     /* Update outer/inner flops */
 722
 723     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3_F,outeriter*30 + inneriter*127);
 724 }