src/gromacs/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecCSTab_VdwCSTab_GeomW4P1_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_ElecCSTab_VdwCSTab_GeomW4P1_VF_c
  35  * Electrostatics interaction: CubicSplineTable
  36  * VdW interaction:            CubicSplineTable
  37  * Geometry:                   Water4-Particle
  38  * Calculate force/pot:        PotentialAndForce
  39  */
  40 void
  41 nb_kernel_ElecCSTab_VdwCSTab_GeomW4P1_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              vdwioffset3;
  63     real             ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
  64     int              vdwjidx0;
  65     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  66     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
  67     real             dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10;
  68     real             dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20;
  69     real             dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30,cexp1_30,cexp2_30;
  70     real             velec,felec,velecsum,facel,crf,krf,krf2;
  71     real             *charge;
  72     int              nvdwtype;
  73     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
  74     int              *vdwtype;
  75     real             *vdwparam;
  76     int              vfitab;
  77     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
  78     real             *vftab;
  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_elec_vdw->data;
  98     vftabscale       = kernel_data->table_elec_vdw->scale;
  99
 100     /* Setup water-specific parameters */
 101     inr              = nlist->iinr[0];
 102     iq1              = facel*charge[inr+1];
 103     iq2              = facel*charge[inr+2];
 104     iq3              = facel*charge[inr+3];
 105     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 106
 107     outeriter        = 0;
 108     inneriter        = 0;
 109
 110     /* Start outer loop over neighborlists */
 111     for(iidx=0; iidx<nri; iidx++)
 112     {
 113         /* Load shift vector for this list */
 114         i_shift_offset   = DIM*shiftidx[iidx];
 115         shX              = shiftvec[i_shift_offset+XX];
 116         shY              = shiftvec[i_shift_offset+YY];
 117         shZ              = shiftvec[i_shift_offset+ZZ];
 118
 119         /* Load limits for loop over neighbors */
 120         j_index_start    = jindex[iidx];
 121         j_index_end      = jindex[iidx+1];
 122
 123         /* Get outer coordinate index */
 124         inr              = iinr[iidx];
 125         i_coord_offset   = DIM*inr;
 126
 127         /* Load i particle coords and add shift vector */
 128         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 129         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 130         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 131         ix1              = shX + x[i_coord_offset+DIM*1+XX];
 132         iy1              = shY + x[i_coord_offset+DIM*1+YY];
 133         iz1              = shZ + x[i_coord_offset+DIM*1+ZZ];
 134         ix2              = shX + x[i_coord_offset+DIM*2+XX];
 135         iy2              = shY + x[i_coord_offset+DIM*2+YY];
 136         iz2              = shZ + x[i_coord_offset+DIM*2+ZZ];
 137         ix3              = shX + x[i_coord_offset+DIM*3+XX];
 138         iy3              = shY + x[i_coord_offset+DIM*3+YY];
 139         iz3              = shZ + x[i_coord_offset+DIM*3+ZZ];
 140
 141         fix0             = 0.0;
 142         fiy0             = 0.0;
 143         fiz0             = 0.0;
 144         fix1             = 0.0;
 145         fiy1             = 0.0;
 146         fiz1             = 0.0;
 147         fix2             = 0.0;
 148         fiy2             = 0.0;
 149         fiz2             = 0.0;
 150         fix3             = 0.0;
 151         fiy3             = 0.0;
 152         fiz3             = 0.0;
 153
 154         /* Reset potential sums */
 155         velecsum         = 0.0;
 156         vvdwsum          = 0.0;
 157
 158         /* Start inner kernel loop */
 159         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 160         {
 161             /* Get j neighbor index, and coordinate index */
 162             jnr              = jjnr[jidx];
 163             j_coord_offset   = DIM*jnr;
 164
 165             /* load j atom coordinates */
 166             jx0              = x[j_coord_offset+DIM*0+XX];
 167             jy0              = x[j_coord_offset+DIM*0+YY];
 168             jz0              = x[j_coord_offset+DIM*0+ZZ];
 169
 170             /* Calculate displacement vector */
 171             dx00             = ix0 - jx0;
 172             dy00             = iy0 - jy0;
 173             dz00             = iz0 - jz0;
 174             dx10             = ix1 - jx0;
 175             dy10             = iy1 - jy0;
 176             dz10             = iz1 - jz0;
 177             dx20             = ix2 - jx0;
 178             dy20             = iy2 - jy0;
 179             dz20             = iz2 - jz0;
 180             dx30             = ix3 - jx0;
 181             dy30             = iy3 - jy0;
 182             dz30             = iz3 - jz0;
 183
 184             /* Calculate squared distance and things based on it */
 185             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 186             rsq10            = dx10*dx10+dy10*dy10+dz10*dz10;
 187             rsq20            = dx20*dx20+dy20*dy20+dz20*dz20;
 188             rsq30            = dx30*dx30+dy30*dy30+dz30*dz30;
 189
 190             rinv00           = gmx_invsqrt(rsq00);
 191             rinv10           = gmx_invsqrt(rsq10);
 192             rinv20           = gmx_invsqrt(rsq20);
 193             rinv30           = gmx_invsqrt(rsq30);
 194
 195             /* Load parameters for j particles */
 196             jq0              = charge[jnr+0];
 197             vdwjidx0         = 2*vdwtype[jnr+0];
 198
 199             /**************************
 200              * CALCULATE INTERACTIONS *
 201              **************************/
 202
 203             r00              = rsq00*rinv00;
 204
 205             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 206             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 207
 208             /* Calculate table index by multiplying r with table scale and truncate to integer */
 209             rt               = r00*vftabscale;
 210             vfitab           = rt;
 211             vfeps            = rt-vfitab;
 212             vfitab           = 3*4*vfitab;
 213
 214             /* CUBIC SPLINE TABLE DISPERSION */
 215             vfitab          += 4;
 216             Y                = vftab[vfitab];
 217             F                = vftab[vfitab+1];
 218             Geps             = vfeps*vftab[vfitab+2];
 219             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 220             Fp               = F+Geps+Heps2;
 221             VV               = Y+vfeps*Fp;
 222             vvdw6            = c6_00*VV;
 223             FF               = Fp+Geps+2.0*Heps2;
 224             fvdw6            = c6_00*FF;
 225
 226             /* CUBIC SPLINE TABLE REPULSION */
 227             Y                = vftab[vfitab+4];
 228             F                = vftab[vfitab+5];
 229             Geps             = vfeps*vftab[vfitab+6];
 230             Heps2            = vfeps*vfeps*vftab[vfitab+7];
 231             Fp               = F+Geps+Heps2;
 232             VV               = Y+vfeps*Fp;
 233             vvdw12           = c12_00*VV;
 234             FF               = Fp+Geps+2.0*Heps2;
 235             fvdw12           = c12_00*FF;
 236             vvdw             = vvdw12+vvdw6;
 237             fvdw             = -(fvdw6+fvdw12)*vftabscale*rinv00;
 238
 239             /* Update potential sums from outer loop */
 240             vvdwsum         += vvdw;
 241
 242             fscal            = fvdw;
 243
 244             /* Calculate temporary vectorial force */
 245             tx               = fscal*dx00;
 246             ty               = fscal*dy00;
 247             tz               = fscal*dz00;
 248
 249             /* Update vectorial force */
 250             fix0            += tx;
 251             fiy0            += ty;
 252             fiz0            += tz;
 253             f[j_coord_offset+DIM*0+XX] -= tx;
 254             f[j_coord_offset+DIM*0+YY] -= ty;
 255             f[j_coord_offset+DIM*0+ZZ] -= tz;
 256
 257             /**************************
 258              * CALCULATE INTERACTIONS *
 259              **************************/
 260
 261             r10              = rsq10*rinv10;
 262
 263             qq10             = iq1*jq0;
 264
 265             /* Calculate table index by multiplying r with table scale and truncate to integer */
 266             rt               = r10*vftabscale;
 267             vfitab           = rt;
 268             vfeps            = rt-vfitab;
 269             vfitab           = 3*4*vfitab;
 270
 271             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 272             Y                = vftab[vfitab];
 273             F                = vftab[vfitab+1];
 274             Geps             = vfeps*vftab[vfitab+2];
 275             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 276             Fp               = F+Geps+Heps2;
 277             VV               = Y+vfeps*Fp;
 278             velec            = qq10*VV;
 279             FF               = Fp+Geps+2.0*Heps2;
 280             felec            = -qq10*FF*vftabscale*rinv10;
 281
 282             /* Update potential sums from outer loop */
 283             velecsum        += velec;
 284
 285             fscal            = felec;
 286
 287             /* Calculate temporary vectorial force */
 288             tx               = fscal*dx10;
 289             ty               = fscal*dy10;
 290             tz               = fscal*dz10;
 291
 292             /* Update vectorial force */
 293             fix1            += tx;
 294             fiy1            += ty;
 295             fiz1            += tz;
 296             f[j_coord_offset+DIM*0+XX] -= tx;
 297             f[j_coord_offset+DIM*0+YY] -= ty;
 298             f[j_coord_offset+DIM*0+ZZ] -= tz;
 299
 300             /**************************
 301              * CALCULATE INTERACTIONS *
 302              **************************/
 303
 304             r20              = rsq20*rinv20;
 305
 306             qq20             = iq2*jq0;
 307
 308             /* Calculate table index by multiplying r with table scale and truncate to integer */
 309             rt               = r20*vftabscale;
 310             vfitab           = rt;
 311             vfeps            = rt-vfitab;
 312             vfitab           = 3*4*vfitab;
 313
 314             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 315             Y                = vftab[vfitab];
 316             F                = vftab[vfitab+1];
 317             Geps             = vfeps*vftab[vfitab+2];
 318             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 319             Fp               = F+Geps+Heps2;
 320             VV               = Y+vfeps*Fp;
 321             velec            = qq20*VV;
 322             FF               = Fp+Geps+2.0*Heps2;
 323             felec            = -qq20*FF*vftabscale*rinv20;
 324
 325             /* Update potential sums from outer loop */
 326             velecsum        += velec;
 327
 328             fscal            = felec;
 329
 330             /* Calculate temporary vectorial force */
 331             tx               = fscal*dx20;
 332             ty               = fscal*dy20;
 333             tz               = fscal*dz20;
 334
 335             /* Update vectorial force */
 336             fix2            += tx;
 337             fiy2            += ty;
 338             fiz2            += tz;
 339             f[j_coord_offset+DIM*0+XX] -= tx;
 340             f[j_coord_offset+DIM*0+YY] -= ty;
 341             f[j_coord_offset+DIM*0+ZZ] -= tz;
 342
 343             /**************************
 344              * CALCULATE INTERACTIONS *
 345              **************************/
 346
 347             r30              = rsq30*rinv30;
 348
 349             qq30             = iq3*jq0;
 350
 351             /* Calculate table index by multiplying r with table scale and truncate to integer */
 352             rt               = r30*vftabscale;
 353             vfitab           = rt;
 354             vfeps            = rt-vfitab;
 355             vfitab           = 3*4*vfitab;
 356
 357             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 358             Y                = vftab[vfitab];
 359             F                = vftab[vfitab+1];
 360             Geps             = vfeps*vftab[vfitab+2];
 361             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 362             Fp               = F+Geps+Heps2;
 363             VV               = Y+vfeps*Fp;
 364             velec            = qq30*VV;
 365             FF               = Fp+Geps+2.0*Heps2;
 366             felec            = -qq30*FF*vftabscale*rinv30;
 367
 368             /* Update potential sums from outer loop */
 369             velecsum        += velec;
 370
 371             fscal            = felec;
 372
 373             /* Calculate temporary vectorial force */
 374             tx               = fscal*dx30;
 375             ty               = fscal*dy30;
 376             tz               = fscal*dz30;
 377
 378             /* Update vectorial force */
 379             fix3            += tx;
 380             fiy3            += ty;
 381             fiz3            += tz;
 382             f[j_coord_offset+DIM*0+XX] -= tx;
 383             f[j_coord_offset+DIM*0+YY] -= ty;
 384             f[j_coord_offset+DIM*0+ZZ] -= tz;
 385
 386             /* Inner loop uses 181 flops */
 387         }
 388         /* End of innermost loop */
 389
 390         tx = ty = tz = 0;
 391         f[i_coord_offset+DIM*0+XX] += fix0;
 392         f[i_coord_offset+DIM*0+YY] += fiy0;
 393         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 394         tx                         += fix0;
 395         ty                         += fiy0;
 396         tz                         += fiz0;
 397         f[i_coord_offset+DIM*1+XX] += fix1;
 398         f[i_coord_offset+DIM*1+YY] += fiy1;
 399         f[i_coord_offset+DIM*1+ZZ] += fiz1;
 400         tx                         += fix1;
 401         ty                         += fiy1;
 402         tz                         += fiz1;
 403         f[i_coord_offset+DIM*2+XX] += fix2;
 404         f[i_coord_offset+DIM*2+YY] += fiy2;
 405         f[i_coord_offset+DIM*2+ZZ] += fiz2;
 406         tx                         += fix2;
 407         ty                         += fiy2;
 408         tz                         += fiz2;
 409         f[i_coord_offset+DIM*3+XX] += fix3;
 410         f[i_coord_offset+DIM*3+YY] += fiy3;
 411         f[i_coord_offset+DIM*3+ZZ] += fiz3;
 412         tx                         += fix3;
 413         ty                         += fiy3;
 414         tz                         += fiz3;
 415         fshift[i_shift_offset+XX]  += tx;
 416         fshift[i_shift_offset+YY]  += ty;
 417         fshift[i_shift_offset+ZZ]  += tz;
 418
 419         ggid                        = gid[iidx];
 420         /* Update potential energies */
 421         kernel_data->energygrp_elec[ggid] += velecsum;
 422         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 423
 424         /* Increment number of inner iterations */
 425         inneriter                  += j_index_end - j_index_start;
 426
 427         /* Outer loop uses 41 flops */
 428     }
 429
 430     /* Increment number of outer iterations */
 431     outeriter        += nri;
 432
 433     /* Update outer/inner flops */
 434
 435     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_VF,outeriter*41 + inneriter*181);
 436 }
 437 /*
 438  * Gromacs nonbonded kernel:   nb_kernel_ElecCSTab_VdwCSTab_GeomW4P1_F_c
 439  * Electrostatics interaction: CubicSplineTable
 440  * VdW interaction:            CubicSplineTable
 441  * Geometry:                   Water4-Particle
 442  * Calculate force/pot:        Force
 443  */
 444 void
 445 nb_kernel_ElecCSTab_VdwCSTab_GeomW4P1_F_c
 446                     (t_nblist * gmx_restrict                nlist,
 447                      rvec * gmx_restrict                    xx,
 448                      rvec * gmx_restrict                    ff,
 449                      t_forcerec * gmx_restrict              fr,
 450                      t_mdatoms * gmx_restrict               mdatoms,
 451                      nb_kernel_data_t * gmx_restrict        kernel_data,
 452                      t_nrnb * gmx_restrict                  nrnb)
 453 {
 454     int              i_shift_offset,i_coord_offset,j_coord_offset;
 455     int              j_index_start,j_index_end;
 456     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 457     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 458     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 459     real             *shiftvec,*fshift,*x,*f;
 460     int              vdwioffset0;
 461     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 462     int              vdwioffset1;
 463     real             ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
 464     int              vdwioffset2;
 465     real             ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
 466     int              vdwioffset3;
 467     real             ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
 468     int              vdwjidx0;
 469     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 470     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 471     real             dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10;
 472     real             dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20;
 473     real             dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30,cexp1_30,cexp2_30;
 474     real             velec,felec,velecsum,facel,crf,krf,krf2;
 475     real             *charge;
 476     int              nvdwtype;
 477     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 478     int              *vdwtype;
 479     real             *vdwparam;
 480     int              vfitab;
 481     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
 482     real             *vftab;
 483
 484     x                = xx[0];
 485     f                = ff[0];
 486
 487     nri              = nlist->nri;
 488     iinr             = nlist->iinr;
 489     jindex           = nlist->jindex;
 490     jjnr             = nlist->jjnr;
 491     shiftidx         = nlist->shift;
 492     gid              = nlist->gid;
 493     shiftvec         = fr->shift_vec[0];
 494     fshift           = fr->fshift[0];
 495     facel            = fr->epsfac;
 496     charge           = mdatoms->chargeA;
 497     nvdwtype         = fr->ntype;
 498     vdwparam         = fr->nbfp;
 499     vdwtype          = mdatoms->typeA;
 500
 501     vftab            = kernel_data->table_elec_vdw->data;
 502     vftabscale       = kernel_data->table_elec_vdw->scale;
 503
 504     /* Setup water-specific parameters */
 505     inr              = nlist->iinr[0];
 506     iq1              = facel*charge[inr+1];
 507     iq2              = facel*charge[inr+2];
 508     iq3              = facel*charge[inr+3];
 509     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 510
 511     outeriter        = 0;
 512     inneriter        = 0;
 513
 514     /* Start outer loop over neighborlists */
 515     for(iidx=0; iidx<nri; iidx++)
 516     {
 517         /* Load shift vector for this list */
 518         i_shift_offset   = DIM*shiftidx[iidx];
 519         shX              = shiftvec[i_shift_offset+XX];
 520         shY              = shiftvec[i_shift_offset+YY];
 521         shZ              = shiftvec[i_shift_offset+ZZ];
 522
 523         /* Load limits for loop over neighbors */
 524         j_index_start    = jindex[iidx];
 525         j_index_end      = jindex[iidx+1];
 526
 527         /* Get outer coordinate index */
 528         inr              = iinr[iidx];
 529         i_coord_offset   = DIM*inr;
 530
 531         /* Load i particle coords and add shift vector */
 532         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 533         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 534         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 535         ix1              = shX + x[i_coord_offset+DIM*1+XX];
 536         iy1              = shY + x[i_coord_offset+DIM*1+YY];
 537         iz1              = shZ + x[i_coord_offset+DIM*1+ZZ];
 538         ix2              = shX + x[i_coord_offset+DIM*2+XX];
 539         iy2              = shY + x[i_coord_offset+DIM*2+YY];
 540         iz2              = shZ + x[i_coord_offset+DIM*2+ZZ];
 541         ix3              = shX + x[i_coord_offset+DIM*3+XX];
 542         iy3              = shY + x[i_coord_offset+DIM*3+YY];
 543         iz3              = shZ + x[i_coord_offset+DIM*3+ZZ];
 544
 545         fix0             = 0.0;
 546         fiy0             = 0.0;
 547         fiz0             = 0.0;
 548         fix1             = 0.0;
 549         fiy1             = 0.0;
 550         fiz1             = 0.0;
 551         fix2             = 0.0;
 552         fiy2             = 0.0;
 553         fiz2             = 0.0;
 554         fix3             = 0.0;
 555         fiy3             = 0.0;
 556         fiz3             = 0.0;
 557
 558         /* Start inner kernel loop */
 559         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 560         {
 561             /* Get j neighbor index, and coordinate index */
 562             jnr              = jjnr[jidx];
 563             j_coord_offset   = DIM*jnr;
 564
 565             /* load j atom coordinates */
 566             jx0              = x[j_coord_offset+DIM*0+XX];
 567             jy0              = x[j_coord_offset+DIM*0+YY];
 568             jz0              = x[j_coord_offset+DIM*0+ZZ];
 569
 570             /* Calculate displacement vector */
 571             dx00             = ix0 - jx0;
 572             dy00             = iy0 - jy0;
 573             dz00             = iz0 - jz0;
 574             dx10             = ix1 - jx0;
 575             dy10             = iy1 - jy0;
 576             dz10             = iz1 - jz0;
 577             dx20             = ix2 - jx0;
 578             dy20             = iy2 - jy0;
 579             dz20             = iz2 - jz0;
 580             dx30             = ix3 - jx0;
 581             dy30             = iy3 - jy0;
 582             dz30             = iz3 - jz0;
 583
 584             /* Calculate squared distance and things based on it */
 585             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 586             rsq10            = dx10*dx10+dy10*dy10+dz10*dz10;
 587             rsq20            = dx20*dx20+dy20*dy20+dz20*dz20;
 588             rsq30            = dx30*dx30+dy30*dy30+dz30*dz30;
 589
 590             rinv00           = gmx_invsqrt(rsq00);
 591             rinv10           = gmx_invsqrt(rsq10);
 592             rinv20           = gmx_invsqrt(rsq20);
 593             rinv30           = gmx_invsqrt(rsq30);
 594
 595             /* Load parameters for j particles */
 596             jq0              = charge[jnr+0];
 597             vdwjidx0         = 2*vdwtype[jnr+0];
 598
 599             /**************************
 600              * CALCULATE INTERACTIONS *
 601              **************************/
 602
 603             r00              = rsq00*rinv00;
 604
 605             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 606             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 607
 608             /* Calculate table index by multiplying r with table scale and truncate to integer */
 609             rt               = r00*vftabscale;
 610             vfitab           = rt;
 611             vfeps            = rt-vfitab;
 612             vfitab           = 3*4*vfitab;
 613
 614             /* CUBIC SPLINE TABLE DISPERSION */
 615             vfitab          += 4;
 616             F                = vftab[vfitab+1];
 617             Geps             = vfeps*vftab[vfitab+2];
 618             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 619             Fp               = F+Geps+Heps2;
 620             FF               = Fp+Geps+2.0*Heps2;
 621             fvdw6            = c6_00*FF;
 622
 623             /* CUBIC SPLINE TABLE REPULSION */
 624             F                = vftab[vfitab+5];
 625             Geps             = vfeps*vftab[vfitab+6];
 626             Heps2            = vfeps*vfeps*vftab[vfitab+7];
 627             Fp               = F+Geps+Heps2;
 628             FF               = Fp+Geps+2.0*Heps2;
 629             fvdw12           = c12_00*FF;
 630             fvdw             = -(fvdw6+fvdw12)*vftabscale*rinv00;
 631
 632             fscal            = fvdw;
 633
 634             /* Calculate temporary vectorial force */
 635             tx               = fscal*dx00;
 636             ty               = fscal*dy00;
 637             tz               = fscal*dz00;
 638
 639             /* Update vectorial force */
 640             fix0            += tx;
 641             fiy0            += ty;
 642             fiz0            += tz;
 643             f[j_coord_offset+DIM*0+XX] -= tx;
 644             f[j_coord_offset+DIM*0+YY] -= ty;
 645             f[j_coord_offset+DIM*0+ZZ] -= tz;
 646
 647             /**************************
 648              * CALCULATE INTERACTIONS *
 649              **************************/
 650
 651             r10              = rsq10*rinv10;
 652
 653             qq10             = iq1*jq0;
 654
 655             /* Calculate table index by multiplying r with table scale and truncate to integer */
 656             rt               = r10*vftabscale;
 657             vfitab           = rt;
 658             vfeps            = rt-vfitab;
 659             vfitab           = 3*4*vfitab;
 660
 661             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 662             F                = vftab[vfitab+1];
 663             Geps             = vfeps*vftab[vfitab+2];
 664             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 665             Fp               = F+Geps+Heps2;
 666             FF               = Fp+Geps+2.0*Heps2;
 667             felec            = -qq10*FF*vftabscale*rinv10;
 668
 669             fscal            = felec;
 670
 671             /* Calculate temporary vectorial force */
 672             tx               = fscal*dx10;
 673             ty               = fscal*dy10;
 674             tz               = fscal*dz10;
 675
 676             /* Update vectorial force */
 677             fix1            += tx;
 678             fiy1            += ty;
 679             fiz1            += tz;
 680             f[j_coord_offset+DIM*0+XX] -= tx;
 681             f[j_coord_offset+DIM*0+YY] -= ty;
 682             f[j_coord_offset+DIM*0+ZZ] -= tz;
 683
 684             /**************************
 685              * CALCULATE INTERACTIONS *
 686              **************************/
 687
 688             r20              = rsq20*rinv20;
 689
 690             qq20             = iq2*jq0;
 691
 692             /* Calculate table index by multiplying r with table scale and truncate to integer */
 693             rt               = r20*vftabscale;
 694             vfitab           = rt;
 695             vfeps            = rt-vfitab;
 696             vfitab           = 3*4*vfitab;
 697
 698             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 699             F                = vftab[vfitab+1];
 700             Geps             = vfeps*vftab[vfitab+2];
 701             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 702             Fp               = F+Geps+Heps2;
 703             FF               = Fp+Geps+2.0*Heps2;
 704             felec            = -qq20*FF*vftabscale*rinv20;
 705
 706             fscal            = felec;
 707
 708             /* Calculate temporary vectorial force */
 709             tx               = fscal*dx20;
 710             ty               = fscal*dy20;
 711             tz               = fscal*dz20;
 712
 713             /* Update vectorial force */
 714             fix2            += tx;
 715             fiy2            += ty;
 716             fiz2            += tz;
 717             f[j_coord_offset+DIM*0+XX] -= tx;
 718             f[j_coord_offset+DIM*0+YY] -= ty;
 719             f[j_coord_offset+DIM*0+ZZ] -= tz;
 720
 721             /**************************
 722              * CALCULATE INTERACTIONS *
 723              **************************/
 724
 725             r30              = rsq30*rinv30;
 726
 727             qq30             = iq3*jq0;
 728
 729             /* Calculate table index by multiplying r with table scale and truncate to integer */
 730             rt               = r30*vftabscale;
 731             vfitab           = rt;
 732             vfeps            = rt-vfitab;
 733             vfitab           = 3*4*vfitab;
 734
 735             /* CUBIC SPLINE TABLE ELECTROSTATICS */
 736             F                = vftab[vfitab+1];
 737             Geps             = vfeps*vftab[vfitab+2];
 738             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 739             Fp               = F+Geps+Heps2;
 740             FF               = Fp+Geps+2.0*Heps2;
 741             felec            = -qq30*FF*vftabscale*rinv30;
 742
 743             fscal            = felec;
 744
 745             /* Calculate temporary vectorial force */
 746             tx               = fscal*dx30;
 747             ty               = fscal*dy30;
 748             tz               = fscal*dz30;
 749
 750             /* Update vectorial force */
 751             fix3            += tx;
 752             fiy3            += ty;
 753             fiz3            += tz;
 754             f[j_coord_offset+DIM*0+XX] -= tx;
 755             f[j_coord_offset+DIM*0+YY] -= ty;
 756             f[j_coord_offset+DIM*0+ZZ] -= tz;
 757
 758             /* Inner loop uses 161 flops */
 759         }
 760         /* End of innermost loop */
 761
 762         tx = ty = tz = 0;
 763         f[i_coord_offset+DIM*0+XX] += fix0;
 764         f[i_coord_offset+DIM*0+YY] += fiy0;
 765         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 766         tx                         += fix0;
 767         ty                         += fiy0;
 768         tz                         += fiz0;
 769         f[i_coord_offset+DIM*1+XX] += fix1;
 770         f[i_coord_offset+DIM*1+YY] += fiy1;
 771         f[i_coord_offset+DIM*1+ZZ] += fiz1;
 772         tx                         += fix1;
 773         ty                         += fiy1;
 774         tz                         += fiz1;
 775         f[i_coord_offset+DIM*2+XX] += fix2;
 776         f[i_coord_offset+DIM*2+YY] += fiy2;
 777         f[i_coord_offset+DIM*2+ZZ] += fiz2;
 778         tx                         += fix2;
 779         ty                         += fiy2;
 780         tz                         += fiz2;
 781         f[i_coord_offset+DIM*3+XX] += fix3;
 782         f[i_coord_offset+DIM*3+YY] += fiy3;
 783         f[i_coord_offset+DIM*3+ZZ] += fiz3;
 784         tx                         += fix3;
 785         ty                         += fiy3;
 786         tz                         += fiz3;
 787         fshift[i_shift_offset+XX]  += tx;
 788         fshift[i_shift_offset+YY]  += ty;
 789         fshift[i_shift_offset+ZZ]  += tz;
 790
 791         /* Increment number of inner iterations */
 792         inneriter                  += j_index_end - j_index_start;
 793
 794         /* Outer loop uses 39 flops */
 795     }
 796
 797     /* Increment number of outer iterations */
 798     outeriter        += nri;
 799
 800     /* Update outer/inner flops */
 801
 802     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_F,outeriter*39 + inneriter*161);
 803 }