src/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecRFCut_VdwCSTab_GeomW3W3_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_ElecRFCut_VdwCSTab_GeomW3W3_VF_c
  35  * Electrostatics interaction: ReactionField
  36  * VdW interaction:            CubicSplineTable
  37  * Geometry:                   Water3-Water3
  38  * Calculate force/pot:        PotentialAndForce
  39  */
  40 void
  41 nb_kernel_ElecRFCut_VdwCSTab_GeomW3W3_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     int              vdwjidx1;
  65     real             jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
  66     int              vdwjidx2;
  67     real             jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
  68     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
  69     real             dx01,dy01,dz01,rsq01,rinv01,rinvsq01,r01,qq01,c6_01,c12_01,cexp1_01,cexp2_01;
  70     real             dx02,dy02,dz02,rsq02,rinv02,rinvsq02,r02,qq02,c6_02,c12_02,cexp1_02,cexp2_02;
  71     real             dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10;
  72     real             dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11,cexp1_11,cexp2_11;
  73     real             dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12,cexp1_12,cexp2_12;
  74     real             dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20;
  75     real             dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21,cexp1_21,cexp2_21;
  76     real             dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22,cexp1_22,cexp2_22;
  77     real             velec,felec,velecsum,facel,crf,krf,krf2;
  78     real             *charge;
  79     int              nvdwtype;
  80     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
  81     int              *vdwtype;
  82     real             *vdwparam;
  83     int              vfitab;
  84     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
  85     real             *vftab;
  86
  87     x                = xx[0];
  88     f                = ff[0];
  89
  90     nri              = nlist->nri;
  91     iinr             = nlist->iinr;
  92     jindex           = nlist->jindex;
  93     jjnr             = nlist->jjnr;
  94     shiftidx         = nlist->shift;
  95     gid              = nlist->gid;
  96     shiftvec         = fr->shift_vec[0];
  97     fshift           = fr->fshift[0];
  98     facel            = fr->epsfac;
  99     charge           = mdatoms->chargeA;
 100     krf              = fr->ic->k_rf;
 101     krf2             = krf*2.0;
 102     crf              = fr->ic->c_rf;
 103     nvdwtype         = fr->ntype;
 104     vdwparam         = fr->nbfp;
 105     vdwtype          = mdatoms->typeA;
 106
 107     vftab            = kernel_data->table_vdw->data;
 108     vftabscale       = kernel_data->table_vdw->scale;
 109
 110     /* Setup water-specific parameters */
 111     inr              = nlist->iinr[0];
 112     iq0              = facel*charge[inr+0];
 113     iq1              = facel*charge[inr+1];
 114     iq2              = facel*charge[inr+2];
 115     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 116
 117     jq0              = charge[inr+0];
 118     jq1              = charge[inr+1];
 119     jq2              = charge[inr+2];
 120     vdwjidx0         = 2*vdwtype[inr+0];
 121     qq00             = iq0*jq0;
 122     c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 123     c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 124     qq01             = iq0*jq1;
 125     qq02             = iq0*jq2;
 126     qq10             = iq1*jq0;
 127     qq11             = iq1*jq1;
 128     qq12             = iq1*jq2;
 129     qq20             = iq2*jq0;
 130     qq21             = iq2*jq1;
 131     qq22             = iq2*jq2;
 132
 133     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 134     rcutoff          = fr->rcoulomb;
 135     rcutoff2         = rcutoff*rcutoff;
 136
 137     outeriter        = 0;
 138     inneriter        = 0;
 139
 140     /* Start outer loop over neighborlists */
 141     for(iidx=0; iidx<nri; iidx++)
 142     {
 143         /* Load shift vector for this list */
 144         i_shift_offset   = DIM*shiftidx[iidx];
 145         shX              = shiftvec[i_shift_offset+XX];
 146         shY              = shiftvec[i_shift_offset+YY];
 147         shZ              = shiftvec[i_shift_offset+ZZ];
 148
 149         /* Load limits for loop over neighbors */
 150         j_index_start    = jindex[iidx];
 151         j_index_end      = jindex[iidx+1];
 152
 153         /* Get outer coordinate index */
 154         inr              = iinr[iidx];
 155         i_coord_offset   = DIM*inr;
 156
 157         /* Load i particle coords and add shift vector */
 158         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 159         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 160         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 161         ix1              = shX + x[i_coord_offset+DIM*1+XX];
 162         iy1              = shY + x[i_coord_offset+DIM*1+YY];
 163         iz1              = shZ + x[i_coord_offset+DIM*1+ZZ];
 164         ix2              = shX + x[i_coord_offset+DIM*2+XX];
 165         iy2              = shY + x[i_coord_offset+DIM*2+YY];
 166         iz2              = shZ + x[i_coord_offset+DIM*2+ZZ];
 167
 168         fix0             = 0.0;
 169         fiy0             = 0.0;
 170         fiz0             = 0.0;
 171         fix1             = 0.0;
 172         fiy1             = 0.0;
 173         fiz1             = 0.0;
 174         fix2             = 0.0;
 175         fiy2             = 0.0;
 176         fiz2             = 0.0;
 177
 178         /* Reset potential sums */
 179         velecsum         = 0.0;
 180         vvdwsum          = 0.0;
 181
 182         /* Start inner kernel loop */
 183         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 184         {
 185             /* Get j neighbor index, and coordinate index */
 186             jnr              = jjnr[jidx];
 187             j_coord_offset   = DIM*jnr;
 188
 189             /* load j atom coordinates */
 190             jx0              = x[j_coord_offset+DIM*0+XX];
 191             jy0              = x[j_coord_offset+DIM*0+YY];
 192             jz0              = x[j_coord_offset+DIM*0+ZZ];
 193             jx1              = x[j_coord_offset+DIM*1+XX];
 194             jy1              = x[j_coord_offset+DIM*1+YY];
 195             jz1              = x[j_coord_offset+DIM*1+ZZ];
 196             jx2              = x[j_coord_offset+DIM*2+XX];
 197             jy2              = x[j_coord_offset+DIM*2+YY];
 198             jz2              = x[j_coord_offset+DIM*2+ZZ];
 199
 200             /* Calculate displacement vector */
 201             dx00             = ix0 - jx0;
 202             dy00             = iy0 - jy0;
 203             dz00             = iz0 - jz0;
 204             dx01             = ix0 - jx1;
 205             dy01             = iy0 - jy1;
 206             dz01             = iz0 - jz1;
 207             dx02             = ix0 - jx2;
 208             dy02             = iy0 - jy2;
 209             dz02             = iz0 - jz2;
 210             dx10             = ix1 - jx0;
 211             dy10             = iy1 - jy0;
 212             dz10             = iz1 - jz0;
 213             dx11             = ix1 - jx1;
 214             dy11             = iy1 - jy1;
 215             dz11             = iz1 - jz1;
 216             dx12             = ix1 - jx2;
 217             dy12             = iy1 - jy2;
 218             dz12             = iz1 - jz2;
 219             dx20             = ix2 - jx0;
 220             dy20             = iy2 - jy0;
 221             dz20             = iz2 - jz0;
 222             dx21             = ix2 - jx1;
 223             dy21             = iy2 - jy1;
 224             dz21             = iz2 - jz1;
 225             dx22             = ix2 - jx2;
 226             dy22             = iy2 - jy2;
 227             dz22             = iz2 - jz2;
 228
 229             /* Calculate squared distance and things based on it */
 230             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 231             rsq01            = dx01*dx01+dy01*dy01+dz01*dz01;
 232             rsq02            = dx02*dx02+dy02*dy02+dz02*dz02;
 233             rsq10            = dx10*dx10+dy10*dy10+dz10*dz10;
 234             rsq11            = dx11*dx11+dy11*dy11+dz11*dz11;
 235             rsq12            = dx12*dx12+dy12*dy12+dz12*dz12;
 236             rsq20            = dx20*dx20+dy20*dy20+dz20*dz20;
 237             rsq21            = dx21*dx21+dy21*dy21+dz21*dz21;
 238             rsq22            = dx22*dx22+dy22*dy22+dz22*dz22;
 239
 240             rinv00           = gmx_invsqrt(rsq00);
 241             rinv01           = gmx_invsqrt(rsq01);
 242             rinv02           = gmx_invsqrt(rsq02);
 243             rinv10           = gmx_invsqrt(rsq10);
 244             rinv11           = gmx_invsqrt(rsq11);
 245             rinv12           = gmx_invsqrt(rsq12);
 246             rinv20           = gmx_invsqrt(rsq20);
 247             rinv21           = gmx_invsqrt(rsq21);
 248             rinv22           = gmx_invsqrt(rsq22);
 249
 250             rinvsq00         = rinv00*rinv00;
 251             rinvsq01         = rinv01*rinv01;
 252             rinvsq02         = rinv02*rinv02;
 253             rinvsq10         = rinv10*rinv10;
 254             rinvsq11         = rinv11*rinv11;
 255             rinvsq12         = rinv12*rinv12;
 256             rinvsq20         = rinv20*rinv20;
 257             rinvsq21         = rinv21*rinv21;
 258             rinvsq22         = rinv22*rinv22;
 259
 260             /**************************
 261              * CALCULATE INTERACTIONS *
 262              **************************/
 263
 264             if (rsq00<rcutoff2)
 265             {
 266
 267             r00              = rsq00*rinv00;
 268
 269             /* Calculate table index by multiplying r with table scale and truncate to integer */
 270             rt               = r00*vftabscale;
 271             vfitab           = rt;
 272             vfeps            = rt-vfitab;
 273             vfitab           = 2*4*vfitab;
 274
 275             /* REACTION-FIELD ELECTROSTATICS */
 276             velec            = qq00*(rinv00+krf*rsq00-crf);
 277             felec            = qq00*(rinv00*rinvsq00-krf2);
 278
 279             /* CUBIC SPLINE TABLE DISPERSION */
 280             vfitab          += 0;
 281             Y                = vftab[vfitab];
 282             F                = vftab[vfitab+1];
 283             Geps             = vfeps*vftab[vfitab+2];
 284             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 285             Fp               = F+Geps+Heps2;
 286             VV               = Y+vfeps*Fp;
 287             vvdw6            = c6_00*VV;
 288             FF               = Fp+Geps+2.0*Heps2;
 289             fvdw6            = c6_00*FF;
 290
 291             /* CUBIC SPLINE TABLE REPULSION */
 292             Y                = vftab[vfitab+4];
 293             F                = vftab[vfitab+5];
 294             Geps             = vfeps*vftab[vfitab+6];
 295             Heps2            = vfeps*vfeps*vftab[vfitab+7];
 296             Fp               = F+Geps+Heps2;
 297             VV               = Y+vfeps*Fp;
 298             vvdw12           = c12_00*VV;
 299             FF               = Fp+Geps+2.0*Heps2;
 300             fvdw12           = c12_00*FF;
 301             vvdw             = vvdw12+vvdw6;
 302             fvdw             = -(fvdw6+fvdw12)*vftabscale*rinv00;
 303
 304             /* Update potential sums from outer loop */
 305             velecsum        += velec;
 306             vvdwsum         += vvdw;
 307
 308             fscal            = felec+fvdw;
 309
 310             /* Calculate temporary vectorial force */
 311             tx               = fscal*dx00;
 312             ty               = fscal*dy00;
 313             tz               = fscal*dz00;
 314
 315             /* Update vectorial force */
 316             fix0            += tx;
 317             fiy0            += ty;
 318             fiz0            += tz;
 319             f[j_coord_offset+DIM*0+XX] -= tx;
 320             f[j_coord_offset+DIM*0+YY] -= ty;
 321             f[j_coord_offset+DIM*0+ZZ] -= tz;
 322
 323             }
 324
 325             /**************************
 326              * CALCULATE INTERACTIONS *
 327              **************************/
 328
 329             if (rsq01<rcutoff2)
 330             {
 331
 332             /* REACTION-FIELD ELECTROSTATICS */
 333             velec            = qq01*(rinv01+krf*rsq01-crf);
 334             felec            = qq01*(rinv01*rinvsq01-krf2);
 335
 336             /* Update potential sums from outer loop */
 337             velecsum        += velec;
 338
 339             fscal            = felec;
 340
 341             /* Calculate temporary vectorial force */
 342             tx               = fscal*dx01;
 343             ty               = fscal*dy01;
 344             tz               = fscal*dz01;
 345
 346             /* Update vectorial force */
 347             fix0            += tx;
 348             fiy0            += ty;
 349             fiz0            += tz;
 350             f[j_coord_offset+DIM*1+XX] -= tx;
 351             f[j_coord_offset+DIM*1+YY] -= ty;
 352             f[j_coord_offset+DIM*1+ZZ] -= tz;
 353
 354             }
 355
 356             /**************************
 357              * CALCULATE INTERACTIONS *
 358              **************************/
 359
 360             if (rsq02<rcutoff2)
 361             {
 362
 363             /* REACTION-FIELD ELECTROSTATICS */
 364             velec            = qq02*(rinv02+krf*rsq02-crf);
 365             felec            = qq02*(rinv02*rinvsq02-krf2);
 366
 367             /* Update potential sums from outer loop */
 368             velecsum        += velec;
 369
 370             fscal            = felec;
 371
 372             /* Calculate temporary vectorial force */
 373             tx               = fscal*dx02;
 374             ty               = fscal*dy02;
 375             tz               = fscal*dz02;
 376
 377             /* Update vectorial force */
 378             fix0            += tx;
 379             fiy0            += ty;
 380             fiz0            += tz;
 381             f[j_coord_offset+DIM*2+XX] -= tx;
 382             f[j_coord_offset+DIM*2+YY] -= ty;
 383             f[j_coord_offset+DIM*2+ZZ] -= tz;
 384
 385             }
 386
 387             /**************************
 388              * CALCULATE INTERACTIONS *
 389              **************************/
 390
 391             if (rsq10<rcutoff2)
 392             {
 393
 394             /* REACTION-FIELD ELECTROSTATICS */
 395             velec            = qq10*(rinv10+krf*rsq10-crf);
 396             felec            = qq10*(rinv10*rinvsq10-krf2);
 397
 398             /* Update potential sums from outer loop */
 399             velecsum        += velec;
 400
 401             fscal            = felec;
 402
 403             /* Calculate temporary vectorial force */
 404             tx               = fscal*dx10;
 405             ty               = fscal*dy10;
 406             tz               = fscal*dz10;
 407
 408             /* Update vectorial force */
 409             fix1            += tx;
 410             fiy1            += ty;
 411             fiz1            += tz;
 412             f[j_coord_offset+DIM*0+XX] -= tx;
 413             f[j_coord_offset+DIM*0+YY] -= ty;
 414             f[j_coord_offset+DIM*0+ZZ] -= tz;
 415
 416             }
 417
 418             /**************************
 419              * CALCULATE INTERACTIONS *
 420              **************************/
 421
 422             if (rsq11<rcutoff2)
 423             {
 424
 425             /* REACTION-FIELD ELECTROSTATICS */
 426             velec            = qq11*(rinv11+krf*rsq11-crf);
 427             felec            = qq11*(rinv11*rinvsq11-krf2);
 428
 429             /* Update potential sums from outer loop */
 430             velecsum        += velec;
 431
 432             fscal            = felec;
 433
 434             /* Calculate temporary vectorial force */
 435             tx               = fscal*dx11;
 436             ty               = fscal*dy11;
 437             tz               = fscal*dz11;
 438
 439             /* Update vectorial force */
 440             fix1            += tx;
 441             fiy1            += ty;
 442             fiz1            += tz;
 443             f[j_coord_offset+DIM*1+XX] -= tx;
 444             f[j_coord_offset+DIM*1+YY] -= ty;
 445             f[j_coord_offset+DIM*1+ZZ] -= tz;
 446
 447             }
 448
 449             /**************************
 450              * CALCULATE INTERACTIONS *
 451              **************************/
 452
 453             if (rsq12<rcutoff2)
 454             {
 455
 456             /* REACTION-FIELD ELECTROSTATICS */
 457             velec            = qq12*(rinv12+krf*rsq12-crf);
 458             felec            = qq12*(rinv12*rinvsq12-krf2);
 459
 460             /* Update potential sums from outer loop */
 461             velecsum        += velec;
 462
 463             fscal            = felec;
 464
 465             /* Calculate temporary vectorial force */
 466             tx               = fscal*dx12;
 467             ty               = fscal*dy12;
 468             tz               = fscal*dz12;
 469
 470             /* Update vectorial force */
 471             fix1            += tx;
 472             fiy1            += ty;
 473             fiz1            += tz;
 474             f[j_coord_offset+DIM*2+XX] -= tx;
 475             f[j_coord_offset+DIM*2+YY] -= ty;
 476             f[j_coord_offset+DIM*2+ZZ] -= tz;
 477
 478             }
 479
 480             /**************************
 481              * CALCULATE INTERACTIONS *
 482              **************************/
 483
 484             if (rsq20<rcutoff2)
 485             {
 486
 487             /* REACTION-FIELD ELECTROSTATICS */
 488             velec            = qq20*(rinv20+krf*rsq20-crf);
 489             felec            = qq20*(rinv20*rinvsq20-krf2);
 490
 491             /* Update potential sums from outer loop */
 492             velecsum        += velec;
 493
 494             fscal            = felec;
 495
 496             /* Calculate temporary vectorial force */
 497             tx               = fscal*dx20;
 498             ty               = fscal*dy20;
 499             tz               = fscal*dz20;
 500
 501             /* Update vectorial force */
 502             fix2            += tx;
 503             fiy2            += ty;
 504             fiz2            += tz;
 505             f[j_coord_offset+DIM*0+XX] -= tx;
 506             f[j_coord_offset+DIM*0+YY] -= ty;
 507             f[j_coord_offset+DIM*0+ZZ] -= tz;
 508
 509             }
 510
 511             /**************************
 512              * CALCULATE INTERACTIONS *
 513              **************************/
 514
 515             if (rsq21<rcutoff2)
 516             {
 517
 518             /* REACTION-FIELD ELECTROSTATICS */
 519             velec            = qq21*(rinv21+krf*rsq21-crf);
 520             felec            = qq21*(rinv21*rinvsq21-krf2);
 521
 522             /* Update potential sums from outer loop */
 523             velecsum        += velec;
 524
 525             fscal            = felec;
 526
 527             /* Calculate temporary vectorial force */
 528             tx               = fscal*dx21;
 529             ty               = fscal*dy21;
 530             tz               = fscal*dz21;
 531
 532             /* Update vectorial force */
 533             fix2            += tx;
 534             fiy2            += ty;
 535             fiz2            += tz;
 536             f[j_coord_offset+DIM*1+XX] -= tx;
 537             f[j_coord_offset+DIM*1+YY] -= ty;
 538             f[j_coord_offset+DIM*1+ZZ] -= tz;
 539
 540             }
 541
 542             /**************************
 543              * CALCULATE INTERACTIONS *
 544              **************************/
 545
 546             if (rsq22<rcutoff2)
 547             {
 548
 549             /* REACTION-FIELD ELECTROSTATICS */
 550             velec            = qq22*(rinv22+krf*rsq22-crf);
 551             felec            = qq22*(rinv22*rinvsq22-krf2);
 552
 553             /* Update potential sums from outer loop */
 554             velecsum        += velec;
 555
 556             fscal            = felec;
 557
 558             /* Calculate temporary vectorial force */
 559             tx               = fscal*dx22;
 560             ty               = fscal*dy22;
 561             tz               = fscal*dz22;
 562
 563             /* Update vectorial force */
 564             fix2            += tx;
 565             fiy2            += ty;
 566             fiz2            += tz;
 567             f[j_coord_offset+DIM*2+XX] -= tx;
 568             f[j_coord_offset+DIM*2+YY] -= ty;
 569             f[j_coord_offset+DIM*2+ZZ] -= tz;
 570
 571             }
 572
 573             /* Inner loop uses 313 flops */
 574         }
 575         /* End of innermost loop */
 576
 577         tx = ty = tz = 0;
 578         f[i_coord_offset+DIM*0+XX] += fix0;
 579         f[i_coord_offset+DIM*0+YY] += fiy0;
 580         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 581         tx                         += fix0;
 582         ty                         += fiy0;
 583         tz                         += fiz0;
 584         f[i_coord_offset+DIM*1+XX] += fix1;
 585         f[i_coord_offset+DIM*1+YY] += fiy1;
 586         f[i_coord_offset+DIM*1+ZZ] += fiz1;
 587         tx                         += fix1;
 588         ty                         += fiy1;
 589         tz                         += fiz1;
 590         f[i_coord_offset+DIM*2+XX] += fix2;
 591         f[i_coord_offset+DIM*2+YY] += fiy2;
 592         f[i_coord_offset+DIM*2+ZZ] += fiz2;
 593         tx                         += fix2;
 594         ty                         += fiy2;
 595         tz                         += fiz2;
 596         fshift[i_shift_offset+XX]  += tx;
 597         fshift[i_shift_offset+YY]  += ty;
 598         fshift[i_shift_offset+ZZ]  += tz;
 599
 600         ggid                        = gid[iidx];
 601         /* Update potential energies */
 602         kernel_data->energygrp_elec[ggid] += velecsum;
 603         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 604
 605         /* Increment number of inner iterations */
 606         inneriter                  += j_index_end - j_index_start;
 607
 608         /* Outer loop uses 32 flops */
 609     }
 610
 611     /* Increment number of outer iterations */
 612     outeriter        += nri;
 613
 614     /* Update outer/inner flops */
 615
 616     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3W3_VF,outeriter*32 + inneriter*313);
 617 }
 618 /*
 619  * Gromacs nonbonded kernel:   nb_kernel_ElecRFCut_VdwCSTab_GeomW3W3_F_c
 620  * Electrostatics interaction: ReactionField
 621  * VdW interaction:            CubicSplineTable
 622  * Geometry:                   Water3-Water3
 623  * Calculate force/pot:        Force
 624  */
 625 void
 626 nb_kernel_ElecRFCut_VdwCSTab_GeomW3W3_F_c
 627                     (t_nblist * gmx_restrict                nlist,
 628                      rvec * gmx_restrict                    xx,
 629                      rvec * gmx_restrict                    ff,
 630                      t_forcerec * gmx_restrict              fr,
 631                      t_mdatoms * gmx_restrict               mdatoms,
 632                      nb_kernel_data_t * gmx_restrict        kernel_data,
 633                      t_nrnb * gmx_restrict                  nrnb)
 634 {
 635     int              i_shift_offset,i_coord_offset,j_coord_offset;
 636     int              j_index_start,j_index_end;
 637     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 638     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 639     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 640     real             *shiftvec,*fshift,*x,*f;
 641     int              vdwioffset0;
 642     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 643     int              vdwioffset1;
 644     real             ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
 645     int              vdwioffset2;
 646     real             ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
 647     int              vdwjidx0;
 648     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 649     int              vdwjidx1;
 650     real             jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
 651     int              vdwjidx2;
 652     real             jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
 653     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 654     real             dx01,dy01,dz01,rsq01,rinv01,rinvsq01,r01,qq01,c6_01,c12_01,cexp1_01,cexp2_01;
 655     real             dx02,dy02,dz02,rsq02,rinv02,rinvsq02,r02,qq02,c6_02,c12_02,cexp1_02,cexp2_02;
 656     real             dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10;
 657     real             dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11,cexp1_11,cexp2_11;
 658     real             dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12,cexp1_12,cexp2_12;
 659     real             dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20;
 660     real             dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21,cexp1_21,cexp2_21;
 661     real             dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22,cexp1_22,cexp2_22;
 662     real             velec,felec,velecsum,facel,crf,krf,krf2;
 663     real             *charge;
 664     int              nvdwtype;
 665     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 666     int              *vdwtype;
 667     real             *vdwparam;
 668     int              vfitab;
 669     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
 670     real             *vftab;
 671
 672     x                = xx[0];
 673     f                = ff[0];
 674
 675     nri              = nlist->nri;
 676     iinr             = nlist->iinr;
 677     jindex           = nlist->jindex;
 678     jjnr             = nlist->jjnr;
 679     shiftidx         = nlist->shift;
 680     gid              = nlist->gid;
 681     shiftvec         = fr->shift_vec[0];
 682     fshift           = fr->fshift[0];
 683     facel            = fr->epsfac;
 684     charge           = mdatoms->chargeA;
 685     krf              = fr->ic->k_rf;
 686     krf2             = krf*2.0;
 687     crf              = fr->ic->c_rf;
 688     nvdwtype         = fr->ntype;
 689     vdwparam         = fr->nbfp;
 690     vdwtype          = mdatoms->typeA;
 691
 692     vftab            = kernel_data->table_vdw->data;
 693     vftabscale       = kernel_data->table_vdw->scale;
 694
 695     /* Setup water-specific parameters */
 696     inr              = nlist->iinr[0];
 697     iq0              = facel*charge[inr+0];
 698     iq1              = facel*charge[inr+1];
 699     iq2              = facel*charge[inr+2];
 700     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 701
 702     jq0              = charge[inr+0];
 703     jq1              = charge[inr+1];
 704     jq2              = charge[inr+2];
 705     vdwjidx0         = 2*vdwtype[inr+0];
 706     qq00             = iq0*jq0;
 707     c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 708     c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 709     qq01             = iq0*jq1;
 710     qq02             = iq0*jq2;
 711     qq10             = iq1*jq0;
 712     qq11             = iq1*jq1;
 713     qq12             = iq1*jq2;
 714     qq20             = iq2*jq0;
 715     qq21             = iq2*jq1;
 716     qq22             = iq2*jq2;
 717
 718     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 719     rcutoff          = fr->rcoulomb;
 720     rcutoff2         = rcutoff*rcutoff;
 721
 722     outeriter        = 0;
 723     inneriter        = 0;
 724
 725     /* Start outer loop over neighborlists */
 726     for(iidx=0; iidx<nri; iidx++)
 727     {
 728         /* Load shift vector for this list */
 729         i_shift_offset   = DIM*shiftidx[iidx];
 730         shX              = shiftvec[i_shift_offset+XX];
 731         shY              = shiftvec[i_shift_offset+YY];
 732         shZ              = shiftvec[i_shift_offset+ZZ];
 733
 734         /* Load limits for loop over neighbors */
 735         j_index_start    = jindex[iidx];
 736         j_index_end      = jindex[iidx+1];
 737
 738         /* Get outer coordinate index */
 739         inr              = iinr[iidx];
 740         i_coord_offset   = DIM*inr;
 741
 742         /* Load i particle coords and add shift vector */
 743         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 744         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 745         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 746         ix1              = shX + x[i_coord_offset+DIM*1+XX];
 747         iy1              = shY + x[i_coord_offset+DIM*1+YY];
 748         iz1              = shZ + x[i_coord_offset+DIM*1+ZZ];
 749         ix2              = shX + x[i_coord_offset+DIM*2+XX];
 750         iy2              = shY + x[i_coord_offset+DIM*2+YY];
 751         iz2              = shZ + x[i_coord_offset+DIM*2+ZZ];
 752
 753         fix0             = 0.0;
 754         fiy0             = 0.0;
 755         fiz0             = 0.0;
 756         fix1             = 0.0;
 757         fiy1             = 0.0;
 758         fiz1             = 0.0;
 759         fix2             = 0.0;
 760         fiy2             = 0.0;
 761         fiz2             = 0.0;
 762
 763         /* Start inner kernel loop */
 764         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 765         {
 766             /* Get j neighbor index, and coordinate index */
 767             jnr              = jjnr[jidx];
 768             j_coord_offset   = DIM*jnr;
 769
 770             /* load j atom coordinates */
 771             jx0              = x[j_coord_offset+DIM*0+XX];
 772             jy0              = x[j_coord_offset+DIM*0+YY];
 773             jz0              = x[j_coord_offset+DIM*0+ZZ];
 774             jx1              = x[j_coord_offset+DIM*1+XX];
 775             jy1              = x[j_coord_offset+DIM*1+YY];
 776             jz1              = x[j_coord_offset+DIM*1+ZZ];
 777             jx2              = x[j_coord_offset+DIM*2+XX];
 778             jy2              = x[j_coord_offset+DIM*2+YY];
 779             jz2              = x[j_coord_offset+DIM*2+ZZ];
 780
 781             /* Calculate displacement vector */
 782             dx00             = ix0 - jx0;
 783             dy00             = iy0 - jy0;
 784             dz00             = iz0 - jz0;
 785             dx01             = ix0 - jx1;
 786             dy01             = iy0 - jy1;
 787             dz01             = iz0 - jz1;
 788             dx02             = ix0 - jx2;
 789             dy02             = iy0 - jy2;
 790             dz02             = iz0 - jz2;
 791             dx10             = ix1 - jx0;
 792             dy10             = iy1 - jy0;
 793             dz10             = iz1 - jz0;
 794             dx11             = ix1 - jx1;
 795             dy11             = iy1 - jy1;
 796             dz11             = iz1 - jz1;
 797             dx12             = ix1 - jx2;
 798             dy12             = iy1 - jy2;
 799             dz12             = iz1 - jz2;
 800             dx20             = ix2 - jx0;
 801             dy20             = iy2 - jy0;
 802             dz20             = iz2 - jz0;
 803             dx21             = ix2 - jx1;
 804             dy21             = iy2 - jy1;
 805             dz21             = iz2 - jz1;
 806             dx22             = ix2 - jx2;
 807             dy22             = iy2 - jy2;
 808             dz22             = iz2 - jz2;
 809
 810             /* Calculate squared distance and things based on it */
 811             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 812             rsq01            = dx01*dx01+dy01*dy01+dz01*dz01;
 813             rsq02            = dx02*dx02+dy02*dy02+dz02*dz02;
 814             rsq10            = dx10*dx10+dy10*dy10+dz10*dz10;
 815             rsq11            = dx11*dx11+dy11*dy11+dz11*dz11;
 816             rsq12            = dx12*dx12+dy12*dy12+dz12*dz12;
 817             rsq20            = dx20*dx20+dy20*dy20+dz20*dz20;
 818             rsq21            = dx21*dx21+dy21*dy21+dz21*dz21;
 819             rsq22            = dx22*dx22+dy22*dy22+dz22*dz22;
 820
 821             rinv00           = gmx_invsqrt(rsq00);
 822             rinv01           = gmx_invsqrt(rsq01);
 823             rinv02           = gmx_invsqrt(rsq02);
 824             rinv10           = gmx_invsqrt(rsq10);
 825             rinv11           = gmx_invsqrt(rsq11);
 826             rinv12           = gmx_invsqrt(rsq12);
 827             rinv20           = gmx_invsqrt(rsq20);
 828             rinv21           = gmx_invsqrt(rsq21);
 829             rinv22           = gmx_invsqrt(rsq22);
 830
 831             rinvsq00         = rinv00*rinv00;
 832             rinvsq01         = rinv01*rinv01;
 833             rinvsq02         = rinv02*rinv02;
 834             rinvsq10         = rinv10*rinv10;
 835             rinvsq11         = rinv11*rinv11;
 836             rinvsq12         = rinv12*rinv12;
 837             rinvsq20         = rinv20*rinv20;
 838             rinvsq21         = rinv21*rinv21;
 839             rinvsq22         = rinv22*rinv22;
 840
 841             /**************************
 842              * CALCULATE INTERACTIONS *
 843              **************************/
 844
 845             if (rsq00<rcutoff2)
 846             {
 847
 848             r00              = rsq00*rinv00;
 849
 850             /* Calculate table index by multiplying r with table scale and truncate to integer */
 851             rt               = r00*vftabscale;
 852             vfitab           = rt;
 853             vfeps            = rt-vfitab;
 854             vfitab           = 2*4*vfitab;
 855
 856             /* REACTION-FIELD ELECTROSTATICS */
 857             felec            = qq00*(rinv00*rinvsq00-krf2);
 858
 859             /* CUBIC SPLINE TABLE DISPERSION */
 860             vfitab          += 0;
 861             F                = vftab[vfitab+1];
 862             Geps             = vfeps*vftab[vfitab+2];
 863             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 864             Fp               = F+Geps+Heps2;
 865             FF               = Fp+Geps+2.0*Heps2;
 866             fvdw6            = c6_00*FF;
 867
 868             /* CUBIC SPLINE TABLE REPULSION */
 869             F                = vftab[vfitab+5];
 870             Geps             = vfeps*vftab[vfitab+6];
 871             Heps2            = vfeps*vfeps*vftab[vfitab+7];
 872             Fp               = F+Geps+Heps2;
 873             FF               = Fp+Geps+2.0*Heps2;
 874             fvdw12           = c12_00*FF;
 875             fvdw             = -(fvdw6+fvdw12)*vftabscale*rinv00;
 876
 877             fscal            = felec+fvdw;
 878
 879             /* Calculate temporary vectorial force */
 880             tx               = fscal*dx00;
 881             ty               = fscal*dy00;
 882             tz               = fscal*dz00;
 883
 884             /* Update vectorial force */
 885             fix0            += tx;
 886             fiy0            += ty;
 887             fiz0            += tz;
 888             f[j_coord_offset+DIM*0+XX] -= tx;
 889             f[j_coord_offset+DIM*0+YY] -= ty;
 890             f[j_coord_offset+DIM*0+ZZ] -= tz;
 891
 892             }
 893
 894             /**************************
 895              * CALCULATE INTERACTIONS *
 896              **************************/
 897
 898             if (rsq01<rcutoff2)
 899             {
 900
 901             /* REACTION-FIELD ELECTROSTATICS */
 902             felec            = qq01*(rinv01*rinvsq01-krf2);
 903
 904             fscal            = felec;
 905
 906             /* Calculate temporary vectorial force */
 907             tx               = fscal*dx01;
 908             ty               = fscal*dy01;
 909             tz               = fscal*dz01;
 910
 911             /* Update vectorial force */
 912             fix0            += tx;
 913             fiy0            += ty;
 914             fiz0            += tz;
 915             f[j_coord_offset+DIM*1+XX] -= tx;
 916             f[j_coord_offset+DIM*1+YY] -= ty;
 917             f[j_coord_offset+DIM*1+ZZ] -= tz;
 918
 919             }
 920
 921             /**************************
 922              * CALCULATE INTERACTIONS *
 923              **************************/
 924
 925             if (rsq02<rcutoff2)
 926             {
 927
 928             /* REACTION-FIELD ELECTROSTATICS */
 929             felec            = qq02*(rinv02*rinvsq02-krf2);
 930
 931             fscal            = felec;
 932
 933             /* Calculate temporary vectorial force */
 934             tx               = fscal*dx02;
 935             ty               = fscal*dy02;
 936             tz               = fscal*dz02;
 937
 938             /* Update vectorial force */
 939             fix0            += tx;
 940             fiy0            += ty;
 941             fiz0            += tz;
 942             f[j_coord_offset+DIM*2+XX] -= tx;
 943             f[j_coord_offset+DIM*2+YY] -= ty;
 944             f[j_coord_offset+DIM*2+ZZ] -= tz;
 945
 946             }
 947
 948             /**************************
 949              * CALCULATE INTERACTIONS *
 950              **************************/
 951
 952             if (rsq10<rcutoff2)
 953             {
 954
 955             /* REACTION-FIELD ELECTROSTATICS */
 956             felec            = qq10*(rinv10*rinvsq10-krf2);
 957
 958             fscal            = felec;
 959
 960             /* Calculate temporary vectorial force */
 961             tx               = fscal*dx10;
 962             ty               = fscal*dy10;
 963             tz               = fscal*dz10;
 964
 965             /* Update vectorial force */
 966             fix1            += tx;
 967             fiy1            += ty;
 968             fiz1            += tz;
 969             f[j_coord_offset+DIM*0+XX] -= tx;
 970             f[j_coord_offset+DIM*0+YY] -= ty;
 971             f[j_coord_offset+DIM*0+ZZ] -= tz;
 972
 973             }
 974
 975             /**************************
 976              * CALCULATE INTERACTIONS *
 977              **************************/
 978
 979             if (rsq11<rcutoff2)
 980             {
 981
 982             /* REACTION-FIELD ELECTROSTATICS */
 983             felec            = qq11*(rinv11*rinvsq11-krf2);
 984
 985             fscal            = felec;
 986
 987             /* Calculate temporary vectorial force */
 988             tx               = fscal*dx11;
 989             ty               = fscal*dy11;
 990             tz               = fscal*dz11;
 991
 992             /* Update vectorial force */
 993             fix1            += tx;
 994             fiy1            += ty;
 995             fiz1            += tz;
 996             f[j_coord_offset+DIM*1+XX] -= tx;
 997             f[j_coord_offset+DIM*1+YY] -= ty;
 998             f[j_coord_offset+DIM*1+ZZ] -= tz;
 999
1000             }
1001
1002             /**************************
1003              * CALCULATE INTERACTIONS *
1004              **************************/
1005
1006             if (rsq12<rcutoff2)
1007             {
1008
1009             /* REACTION-FIELD ELECTROSTATICS */
1010             felec            = qq12*(rinv12*rinvsq12-krf2);
1011
1012             fscal            = felec;
1013
1014             /* Calculate temporary vectorial force */
1015             tx               = fscal*dx12;
1016             ty               = fscal*dy12;
1017             tz               = fscal*dz12;
1018
1019             /* Update vectorial force */
1020             fix1            += tx;
1021             fiy1            += ty;
1022             fiz1            += tz;
1023             f[j_coord_offset+DIM*2+XX] -= tx;
1024             f[j_coord_offset+DIM*2+YY] -= ty;
1025             f[j_coord_offset+DIM*2+ZZ] -= tz;
1026
1027             }
1028
1029             /**************************
1030              * CALCULATE INTERACTIONS *
1031              **************************/
1032
1033             if (rsq20<rcutoff2)
1034             {
1035
1036             /* REACTION-FIELD ELECTROSTATICS */
1037             felec            = qq20*(rinv20*rinvsq20-krf2);
1038
1039             fscal            = felec;
1040
1041             /* Calculate temporary vectorial force */
1042             tx               = fscal*dx20;
1043             ty               = fscal*dy20;
1044             tz               = fscal*dz20;
1045
1046             /* Update vectorial force */
1047             fix2            += tx;
1048             fiy2            += ty;
1049             fiz2            += tz;
1050             f[j_coord_offset+DIM*0+XX] -= tx;
1051             f[j_coord_offset+DIM*0+YY] -= ty;
1052             f[j_coord_offset+DIM*0+ZZ] -= tz;
1053
1054             }
1055
1056             /**************************
1057              * CALCULATE INTERACTIONS *
1058              **************************/
1059
1060             if (rsq21<rcutoff2)
1061             {
1062
1063             /* REACTION-FIELD ELECTROSTATICS */
1064             felec            = qq21*(rinv21*rinvsq21-krf2);
1065
1066             fscal            = felec;
1067
1068             /* Calculate temporary vectorial force */
1069             tx               = fscal*dx21;
1070             ty               = fscal*dy21;
1071             tz               = fscal*dz21;
1072
1073             /* Update vectorial force */
1074             fix2            += tx;
1075             fiy2            += ty;
1076             fiz2            += tz;
1077             f[j_coord_offset+DIM*1+XX] -= tx;
1078             f[j_coord_offset+DIM*1+YY] -= ty;
1079             f[j_coord_offset+DIM*1+ZZ] -= tz;
1080
1081             }
1082
1083             /**************************
1084              * CALCULATE INTERACTIONS *
1085              **************************/
1086
1087             if (rsq22<rcutoff2)
1088             {
1089
1090             /* REACTION-FIELD ELECTROSTATICS */
1091             felec            = qq22*(rinv22*rinvsq22-krf2);
1092
1093             fscal            = felec;
1094
1095             /* Calculate temporary vectorial force */
1096             tx               = fscal*dx22;
1097             ty               = fscal*dy22;
1098             tz               = fscal*dz22;
1099
1100             /* Update vectorial force */
1101             fix2            += tx;
1102             fiy2            += ty;
1103             fiz2            += tz;
1104             f[j_coord_offset+DIM*2+XX] -= tx;
1105             f[j_coord_offset+DIM*2+YY] -= ty;
1106             f[j_coord_offset+DIM*2+ZZ] -= tz;
1107
1108             }
1109
1110             /* Inner loop uses 260 flops */
1111         }
1112         /* End of innermost loop */
1113
1114         tx = ty = tz = 0;
1115         f[i_coord_offset+DIM*0+XX] += fix0;
1116         f[i_coord_offset+DIM*0+YY] += fiy0;
1117         f[i_coord_offset+DIM*0+ZZ] += fiz0;
1118         tx                         += fix0;
1119         ty                         += fiy0;
1120         tz                         += fiz0;
1121         f[i_coord_offset+DIM*1+XX] += fix1;
1122         f[i_coord_offset+DIM*1+YY] += fiy1;
1123         f[i_coord_offset+DIM*1+ZZ] += fiz1;
1124         tx                         += fix1;
1125         ty                         += fiy1;
1126         tz                         += fiz1;
1127         f[i_coord_offset+DIM*2+XX] += fix2;
1128         f[i_coord_offset+DIM*2+YY] += fiy2;
1129         f[i_coord_offset+DIM*2+ZZ] += fiz2;
1130         tx                         += fix2;
1131         ty                         += fiy2;
1132         tz                         += fiz2;
1133         fshift[i_shift_offset+XX]  += tx;
1134         fshift[i_shift_offset+YY]  += ty;
1135         fshift[i_shift_offset+ZZ]  += tz;
1136
1137         /* Increment number of inner iterations */
1138         inneriter                  += j_index_end - j_index_start;
1139
1140         /* Outer loop uses 30 flops */
1141     }
1142
1143     /* Increment number of outer iterations */
1144     outeriter        += nri;
1145
1146     /* Update outer/inner flops */
1147
1148     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3W3_F,outeriter*30 + inneriter*260);
1149 }