src/gromacs/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecEw_VdwCSTab_GeomW3P1_c.c

   1 /*
   2  * Note: this file was generated by the Gromacs c kernel generator.
   3  *
   4  *                This source code is part of
   5  *
   6  *                 G   R   O   M   A   C   S
   7  *
   8  * Copyright (c) 2001-2012, The GROMACS Development Team
   9  *
  10  * Gromacs is a library for molecular simulation and trajectory analysis,
  11  * written by Erik Lindahl, David van der Spoel, Berk Hess, and others - for
  12  * a full list of developers and information, check out http://www.gromacs.org
  13  *
  14  * This program is free software; you can redistribute it and/or modify it under
  15  * the terms of the GNU Lesser General Public License as published by the Free
  16  * Software Foundation; either version 2 of the License, or (at your option) any
  17  * later version.
  18  *
  19  * To help fund GROMACS development, we humbly ask that you cite
  20  * the papers people have written on it - you can find them on the website.
  21  */
  22 #ifdef HAVE_CONFIG_H
  23 #include <config.h>
  24 #endif
  25
  26 #include <math.h>
  27
  28 #include "../nb_kernel.h"
  29 #include "types/simple.h"
  30 #include "vec.h"
  31 #include "nrnb.h"
  32
  33 /*
  34  * Gromacs nonbonded kernel:   nb_kernel_ElecEw_VdwCSTab_GeomW3P1_VF_c
  35  * Electrostatics interaction: Ewald
  36  * VdW interaction:            CubicSplineTable
  37  * Geometry:                   Water3-Particle
  38  * Calculate force/pot:        PotentialAndForce
  39  */
  40 void
  41 nb_kernel_ElecEw_VdwCSTab_GeomW3P1_VF_c
  42                     (t_nblist * gmx_restrict                nlist,
  43                      rvec * gmx_restrict                    xx,
  44                      rvec * gmx_restrict                    ff,
  45                      t_forcerec * gmx_restrict              fr,
  46                      t_mdatoms * gmx_restrict               mdatoms,
  47                      nb_kernel_data_t * gmx_restrict        kernel_data,
  48                      t_nrnb * gmx_restrict                  nrnb)
  49 {
  50     int              i_shift_offset,i_coord_offset,j_coord_offset;
  51     int              j_index_start,j_index_end;
  52     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
  53     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
  54     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  55     real             *shiftvec,*fshift,*x,*f;
  56     int              vdwioffset0;
  57     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  58     int              vdwioffset1;
  59     real             ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
  60     int              vdwioffset2;
  61     real             ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
  62     int              vdwjidx0;
  63     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  64     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
  65     real             dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10;
  66     real             dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20;
  67     real             velec,felec,velecsum,facel,crf,krf,krf2;
  68     real             *charge;
  69     int              nvdwtype;
  70     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
  71     int              *vdwtype;
  72     real             *vdwparam;
  73     int              vfitab;
  74     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
  75     real             *vftab;
  76     int              ewitab;
  77     real             ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
  78     real             *ewtab;
  79
  80     x                = xx[0];
  81     f                = ff[0];
  82
  83     nri              = nlist->nri;
  84     iinr             = nlist->iinr;
  85     jindex           = nlist->jindex;
  86     jjnr             = nlist->jjnr;
  87     shiftidx         = nlist->shift;
  88     gid              = nlist->gid;
  89     shiftvec         = fr->shift_vec[0];
  90     fshift           = fr->fshift[0];
  91     facel            = fr->epsfac;
  92     charge           = mdatoms->chargeA;
  93     nvdwtype         = fr->ntype;
  94     vdwparam         = fr->nbfp;
  95     vdwtype          = mdatoms->typeA;
  96
  97     vftab            = kernel_data->table_vdw->data;
  98     vftabscale       = kernel_data->table_vdw->scale;
  99
 100     sh_ewald         = fr->ic->sh_ewald;
 101     ewtab            = fr->ic->tabq_coul_FDV0;
 102     ewtabscale       = fr->ic->tabq_scale;
 103     ewtabhalfspace   = 0.5/ewtabscale;
 104
 105     /* Setup water-specific parameters */
 106     inr              = nlist->iinr[0];
 107     iq0              = facel*charge[inr+0];
 108     iq1              = facel*charge[inr+1];
 109     iq2              = facel*charge[inr+2];
 110     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 111
 112     outeriter        = 0;
 113     inneriter        = 0;
 114
 115     /* Start outer loop over neighborlists */
 116     for(iidx=0; iidx<nri; iidx++)
 117     {
 118         /* Load shift vector for this list */
 119         i_shift_offset   = DIM*shiftidx[iidx];
 120         shX              = shiftvec[i_shift_offset+XX];
 121         shY              = shiftvec[i_shift_offset+YY];
 122         shZ              = shiftvec[i_shift_offset+ZZ];
 123
 124         /* Load limits for loop over neighbors */
 125         j_index_start    = jindex[iidx];
 126         j_index_end      = jindex[iidx+1];
 127
 128         /* Get outer coordinate index */
 129         inr              = iinr[iidx];
 130         i_coord_offset   = DIM*inr;
 131
 132         /* Load i particle coords and add shift vector */
 133         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 134         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 135         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 136         ix1              = shX + x[i_coord_offset+DIM*1+XX];
 137         iy1              = shY + x[i_coord_offset+DIM*1+YY];
 138         iz1              = shZ + x[i_coord_offset+DIM*1+ZZ];
 139         ix2              = shX + x[i_coord_offset+DIM*2+XX];
 140         iy2              = shY + x[i_coord_offset+DIM*2+YY];
 141         iz2              = shZ + x[i_coord_offset+DIM*2+ZZ];
 142
 143         fix0             = 0.0;
 144         fiy0             = 0.0;
 145         fiz0             = 0.0;
 146         fix1             = 0.0;
 147         fiy1             = 0.0;
 148         fiz1             = 0.0;
 149         fix2             = 0.0;
 150         fiy2             = 0.0;
 151         fiz2             = 0.0;
 152
 153         /* Reset potential sums */
 154         velecsum         = 0.0;
 155         vvdwsum          = 0.0;
 156
 157         /* Start inner kernel loop */
 158         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 159         {
 160             /* Get j neighbor index, and coordinate index */
 161             jnr              = jjnr[jidx];
 162             j_coord_offset   = DIM*jnr;
 163
 164             /* load j atom coordinates */
 165             jx0              = x[j_coord_offset+DIM*0+XX];
 166             jy0              = x[j_coord_offset+DIM*0+YY];
 167             jz0              = x[j_coord_offset+DIM*0+ZZ];
 168
 169             /* Calculate displacement vector */
 170             dx00             = ix0 - jx0;
 171             dy00             = iy0 - jy0;
 172             dz00             = iz0 - jz0;
 173             dx10             = ix1 - jx0;
 174             dy10             = iy1 - jy0;
 175             dz10             = iz1 - jz0;
 176             dx20             = ix2 - jx0;
 177             dy20             = iy2 - jy0;
 178             dz20             = iz2 - jz0;
 179
 180             /* Calculate squared distance and things based on it */
 181             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 182             rsq10            = dx10*dx10+dy10*dy10+dz10*dz10;
 183             rsq20            = dx20*dx20+dy20*dy20+dz20*dz20;
 184
 185             rinv00           = gmx_invsqrt(rsq00);
 186             rinv10           = gmx_invsqrt(rsq10);
 187             rinv20           = gmx_invsqrt(rsq20);
 188
 189             rinvsq00         = rinv00*rinv00;
 190             rinvsq10         = rinv10*rinv10;
 191             rinvsq20         = rinv20*rinv20;
 192
 193             /* Load parameters for j particles */
 194             jq0              = charge[jnr+0];
 195             vdwjidx0         = 2*vdwtype[jnr+0];
 196
 197             /**************************
 198              * CALCULATE INTERACTIONS *
 199              **************************/
 200
 201             r00              = rsq00*rinv00;
 202
 203             qq00             = iq0*jq0;
 204             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 205             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 206
 207             /* Calculate table index by multiplying r with table scale and truncate to integer */
 208             rt               = r00*vftabscale;
 209             vfitab           = rt;
 210             vfeps            = rt-vfitab;
 211             vfitab           = 2*4*vfitab;
 212
 213             /* EWALD ELECTROSTATICS */
 214
 215             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 216             ewrt             = r00*ewtabscale;
 217             ewitab           = ewrt;
 218             eweps            = ewrt-ewitab;
 219             ewitab           = 4*ewitab;
 220             felec            = ewtab[ewitab]+eweps*ewtab[ewitab+1];
 221             velec            = qq00*(rinv00-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
 222             felec            = qq00*rinv00*(rinvsq00-felec);
 223
 224             /* CUBIC SPLINE TABLE DISPERSION */
 225             vfitab          += 0;
 226             Y                = vftab[vfitab];
 227             F                = vftab[vfitab+1];
 228             Geps             = vfeps*vftab[vfitab+2];
 229             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 230             Fp               = F+Geps+Heps2;
 231             VV               = Y+vfeps*Fp;
 232             vvdw6            = c6_00*VV;
 233             FF               = Fp+Geps+2.0*Heps2;
 234             fvdw6            = c6_00*FF;
 235
 236             /* CUBIC SPLINE TABLE REPULSION */
 237             Y                = vftab[vfitab+4];
 238             F                = vftab[vfitab+5];
 239             Geps             = vfeps*vftab[vfitab+6];
 240             Heps2            = vfeps*vfeps*vftab[vfitab+7];
 241             Fp               = F+Geps+Heps2;
 242             VV               = Y+vfeps*Fp;
 243             vvdw12           = c12_00*VV;
 244             FF               = Fp+Geps+2.0*Heps2;
 245             fvdw12           = c12_00*FF;
 246             vvdw             = vvdw12+vvdw6;
 247             fvdw             = -(fvdw6+fvdw12)*vftabscale*rinv00;
 248
 249             /* Update potential sums from outer loop */
 250             velecsum        += velec;
 251             vvdwsum         += vvdw;
 252
 253             fscal            = felec+fvdw;
 254
 255             /* Calculate temporary vectorial force */
 256             tx               = fscal*dx00;
 257             ty               = fscal*dy00;
 258             tz               = fscal*dz00;
 259
 260             /* Update vectorial force */
 261             fix0            += tx;
 262             fiy0            += ty;
 263             fiz0            += tz;
 264             f[j_coord_offset+DIM*0+XX] -= tx;
 265             f[j_coord_offset+DIM*0+YY] -= ty;
 266             f[j_coord_offset+DIM*0+ZZ] -= tz;
 267
 268             /**************************
 269              * CALCULATE INTERACTIONS *
 270              **************************/
 271
 272             r10              = rsq10*rinv10;
 273
 274             qq10             = iq1*jq0;
 275
 276             /* EWALD ELECTROSTATICS */
 277
 278             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 279             ewrt             = r10*ewtabscale;
 280             ewitab           = ewrt;
 281             eweps            = ewrt-ewitab;
 282             ewitab           = 4*ewitab;
 283             felec            = ewtab[ewitab]+eweps*ewtab[ewitab+1];
 284             velec            = qq10*(rinv10-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
 285             felec            = qq10*rinv10*(rinvsq10-felec);
 286
 287             /* Update potential sums from outer loop */
 288             velecsum        += velec;
 289
 290             fscal            = felec;
 291
 292             /* Calculate temporary vectorial force */
 293             tx               = fscal*dx10;
 294             ty               = fscal*dy10;
 295             tz               = fscal*dz10;
 296
 297             /* Update vectorial force */
 298             fix1            += tx;
 299             fiy1            += ty;
 300             fiz1            += tz;
 301             f[j_coord_offset+DIM*0+XX] -= tx;
 302             f[j_coord_offset+DIM*0+YY] -= ty;
 303             f[j_coord_offset+DIM*0+ZZ] -= tz;
 304
 305             /**************************
 306              * CALCULATE INTERACTIONS *
 307              **************************/
 308
 309             r20              = rsq20*rinv20;
 310
 311             qq20             = iq2*jq0;
 312
 313             /* EWALD ELECTROSTATICS */
 314
 315             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 316             ewrt             = r20*ewtabscale;
 317             ewitab           = ewrt;
 318             eweps            = ewrt-ewitab;
 319             ewitab           = 4*ewitab;
 320             felec            = ewtab[ewitab]+eweps*ewtab[ewitab+1];
 321             velec            = qq20*(rinv20-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
 322             felec            = qq20*rinv20*(rinvsq20-felec);
 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 156 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*156);
 386 }
 387 /*
 388  * Gromacs nonbonded kernel:   nb_kernel_ElecEw_VdwCSTab_GeomW3P1_F_c
 389  * Electrostatics interaction: Ewald
 390  * VdW interaction:            CubicSplineTable
 391  * Geometry:                   Water3-Particle
 392  * Calculate force/pot:        Force
 393  */
 394 void
 395 nb_kernel_ElecEw_VdwCSTab_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_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     int              ewitab;
 431     real             ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
 432     real             *ewtab;
 433
 434     x                = xx[0];
 435     f                = ff[0];
 436
 437     nri              = nlist->nri;
 438     iinr             = nlist->iinr;
 439     jindex           = nlist->jindex;
 440     jjnr             = nlist->jjnr;
 441     shiftidx         = nlist->shift;
 442     gid              = nlist->gid;
 443     shiftvec         = fr->shift_vec[0];
 444     fshift           = fr->fshift[0];
 445     facel            = fr->epsfac;
 446     charge           = mdatoms->chargeA;
 447     nvdwtype         = fr->ntype;
 448     vdwparam         = fr->nbfp;
 449     vdwtype          = mdatoms->typeA;
 450
 451     vftab            = kernel_data->table_vdw->data;
 452     vftabscale       = kernel_data->table_vdw->scale;
 453
 454     sh_ewald         = fr->ic->sh_ewald;
 455     ewtab            = fr->ic->tabq_coul_F;
 456     ewtabscale       = fr->ic->tabq_scale;
 457     ewtabhalfspace   = 0.5/ewtabscale;
 458
 459     /* Setup water-specific parameters */
 460     inr              = nlist->iinr[0];
 461     iq0              = facel*charge[inr+0];
 462     iq1              = facel*charge[inr+1];
 463     iq2              = facel*charge[inr+2];
 464     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 465
 466     outeriter        = 0;
 467     inneriter        = 0;
 468
 469     /* Start outer loop over neighborlists */
 470     for(iidx=0; iidx<nri; iidx++)
 471     {
 472         /* Load shift vector for this list */
 473         i_shift_offset   = DIM*shiftidx[iidx];
 474         shX              = shiftvec[i_shift_offset+XX];
 475         shY              = shiftvec[i_shift_offset+YY];
 476         shZ              = shiftvec[i_shift_offset+ZZ];
 477
 478         /* Load limits for loop over neighbors */
 479         j_index_start    = jindex[iidx];
 480         j_index_end      = jindex[iidx+1];
 481
 482         /* Get outer coordinate index */
 483         inr              = iinr[iidx];
 484         i_coord_offset   = DIM*inr;
 485
 486         /* Load i particle coords and add shift vector */
 487         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 488         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 489         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 490         ix1              = shX + x[i_coord_offset+DIM*1+XX];
 491         iy1              = shY + x[i_coord_offset+DIM*1+YY];
 492         iz1              = shZ + x[i_coord_offset+DIM*1+ZZ];
 493         ix2              = shX + x[i_coord_offset+DIM*2+XX];
 494         iy2              = shY + x[i_coord_offset+DIM*2+YY];
 495         iz2              = shZ + x[i_coord_offset+DIM*2+ZZ];
 496
 497         fix0             = 0.0;
 498         fiy0             = 0.0;
 499         fiz0             = 0.0;
 500         fix1             = 0.0;
 501         fiy1             = 0.0;
 502         fiz1             = 0.0;
 503         fix2             = 0.0;
 504         fiy2             = 0.0;
 505         fiz2             = 0.0;
 506
 507         /* Start inner kernel loop */
 508         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 509         {
 510             /* Get j neighbor index, and coordinate index */
 511             jnr              = jjnr[jidx];
 512             j_coord_offset   = DIM*jnr;
 513
 514             /* load j atom coordinates */
 515             jx0              = x[j_coord_offset+DIM*0+XX];
 516             jy0              = x[j_coord_offset+DIM*0+YY];
 517             jz0              = x[j_coord_offset+DIM*0+ZZ];
 518
 519             /* Calculate displacement vector */
 520             dx00             = ix0 - jx0;
 521             dy00             = iy0 - jy0;
 522             dz00             = iz0 - jz0;
 523             dx10             = ix1 - jx0;
 524             dy10             = iy1 - jy0;
 525             dz10             = iz1 - jz0;
 526             dx20             = ix2 - jx0;
 527             dy20             = iy2 - jy0;
 528             dz20             = iz2 - jz0;
 529
 530             /* Calculate squared distance and things based on it */
 531             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 532             rsq10            = dx10*dx10+dy10*dy10+dz10*dz10;
 533             rsq20            = dx20*dx20+dy20*dy20+dz20*dz20;
 534
 535             rinv00           = gmx_invsqrt(rsq00);
 536             rinv10           = gmx_invsqrt(rsq10);
 537             rinv20           = gmx_invsqrt(rsq20);
 538
 539             rinvsq00         = rinv00*rinv00;
 540             rinvsq10         = rinv10*rinv10;
 541             rinvsq20         = rinv20*rinv20;
 542
 543             /* Load parameters for j particles */
 544             jq0              = charge[jnr+0];
 545             vdwjidx0         = 2*vdwtype[jnr+0];
 546
 547             /**************************
 548              * CALCULATE INTERACTIONS *
 549              **************************/
 550
 551             r00              = rsq00*rinv00;
 552
 553             qq00             = iq0*jq0;
 554             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 555             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 556
 557             /* Calculate table index by multiplying r with table scale and truncate to integer */
 558             rt               = r00*vftabscale;
 559             vfitab           = rt;
 560             vfeps            = rt-vfitab;
 561             vfitab           = 2*4*vfitab;
 562
 563             /* EWALD ELECTROSTATICS */
 564
 565             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 566             ewrt             = r00*ewtabscale;
 567             ewitab           = ewrt;
 568             eweps            = ewrt-ewitab;
 569             felec            = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1];
 570             felec            = qq00*rinv00*(rinvsq00-felec);
 571
 572             /* CUBIC SPLINE TABLE DISPERSION */
 573             vfitab          += 0;
 574             F                = vftab[vfitab+1];
 575             Geps             = vfeps*vftab[vfitab+2];
 576             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 577             Fp               = F+Geps+Heps2;
 578             FF               = Fp+Geps+2.0*Heps2;
 579             fvdw6            = c6_00*FF;
 580
 581             /* CUBIC SPLINE TABLE REPULSION */
 582             F                = vftab[vfitab+5];
 583             Geps             = vfeps*vftab[vfitab+6];
 584             Heps2            = vfeps*vfeps*vftab[vfitab+7];
 585             Fp               = F+Geps+Heps2;
 586             FF               = Fp+Geps+2.0*Heps2;
 587             fvdw12           = c12_00*FF;
 588             fvdw             = -(fvdw6+fvdw12)*vftabscale*rinv00;
 589
 590             fscal            = felec+fvdw;
 591
 592             /* Calculate temporary vectorial force */
 593             tx               = fscal*dx00;
 594             ty               = fscal*dy00;
 595             tz               = fscal*dz00;
 596
 597             /* Update vectorial force */
 598             fix0            += tx;
 599             fiy0            += ty;
 600             fiz0            += tz;
 601             f[j_coord_offset+DIM*0+XX] -= tx;
 602             f[j_coord_offset+DIM*0+YY] -= ty;
 603             f[j_coord_offset+DIM*0+ZZ] -= tz;
 604
 605             /**************************
 606              * CALCULATE INTERACTIONS *
 607              **************************/
 608
 609             r10              = rsq10*rinv10;
 610
 611             qq10             = iq1*jq0;
 612
 613             /* EWALD ELECTROSTATICS */
 614
 615             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 616             ewrt             = r10*ewtabscale;
 617             ewitab           = ewrt;
 618             eweps            = ewrt-ewitab;
 619             felec            = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1];
 620             felec            = qq10*rinv10*(rinvsq10-felec);
 621
 622             fscal            = felec;
 623
 624             /* Calculate temporary vectorial force */
 625             tx               = fscal*dx10;
 626             ty               = fscal*dy10;
 627             tz               = fscal*dz10;
 628
 629             /* Update vectorial force */
 630             fix1            += tx;
 631             fiy1            += ty;
 632             fiz1            += tz;
 633             f[j_coord_offset+DIM*0+XX] -= tx;
 634             f[j_coord_offset+DIM*0+YY] -= ty;
 635             f[j_coord_offset+DIM*0+ZZ] -= tz;
 636
 637             /**************************
 638              * CALCULATE INTERACTIONS *
 639              **************************/
 640
 641             r20              = rsq20*rinv20;
 642
 643             qq20             = iq2*jq0;
 644
 645             /* EWALD ELECTROSTATICS */
 646
 647             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 648             ewrt             = r20*ewtabscale;
 649             ewitab           = ewrt;
 650             eweps            = ewrt-ewitab;
 651             felec            = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1];
 652             felec            = qq20*rinv20*(rinvsq20-felec);
 653
 654             fscal            = felec;
 655
 656             /* Calculate temporary vectorial force */
 657             tx               = fscal*dx20;
 658             ty               = fscal*dy20;
 659             tz               = fscal*dz20;
 660
 661             /* Update vectorial force */
 662             fix2            += tx;
 663             fiy2            += ty;
 664             fiz2            += tz;
 665             f[j_coord_offset+DIM*0+XX] -= tx;
 666             f[j_coord_offset+DIM*0+YY] -= ty;
 667             f[j_coord_offset+DIM*0+ZZ] -= tz;
 668
 669             /* Inner loop uses 127 flops */
 670         }
 671         /* End of innermost loop */
 672
 673         tx = ty = tz = 0;
 674         f[i_coord_offset+DIM*0+XX] += fix0;
 675         f[i_coord_offset+DIM*0+YY] += fiy0;
 676         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 677         tx                         += fix0;
 678         ty                         += fiy0;
 679         tz                         += fiz0;
 680         f[i_coord_offset+DIM*1+XX] += fix1;
 681         f[i_coord_offset+DIM*1+YY] += fiy1;
 682         f[i_coord_offset+DIM*1+ZZ] += fiz1;
 683         tx                         += fix1;
 684         ty                         += fiy1;
 685         tz                         += fiz1;
 686         f[i_coord_offset+DIM*2+XX] += fix2;
 687         f[i_coord_offset+DIM*2+YY] += fiy2;
 688         f[i_coord_offset+DIM*2+ZZ] += fiz2;
 689         tx                         += fix2;
 690         ty                         += fiy2;
 691         tz                         += fiz2;
 692         fshift[i_shift_offset+XX]  += tx;
 693         fshift[i_shift_offset+YY]  += ty;
 694         fshift[i_shift_offset+ZZ]  += tz;
 695
 696         /* Increment number of inner iterations */
 697         inneriter                  += j_index_end - j_index_start;
 698
 699         /* Outer loop uses 30 flops */
 700     }
 701
 702     /* Increment number of outer iterations */
 703     outeriter        += nri;
 704
 705     /* Update outer/inner flops */
 706
 707     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3_F,outeriter*30 + inneriter*127);
 708 }