src/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecRFCut_VdwLJSh_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_VdwLJSh_GeomW3W3_VF_c
  35  * Electrostatics interaction: ReactionField
  36  * VdW interaction:            LennardJones
  37  * Geometry:                   Water3-Water3
  38  * Calculate force/pot:        PotentialAndForce
  39  */
  40 void
  41 nb_kernel_ElecRFCut_VdwLJSh_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
  84     x                = xx[0];
  85     f                = ff[0];
  86
  87     nri              = nlist->nri;
  88     iinr             = nlist->iinr;
  89     jindex           = nlist->jindex;
  90     jjnr             = nlist->jjnr;
  91     shiftidx         = nlist->shift;
  92     gid              = nlist->gid;
  93     shiftvec         = fr->shift_vec[0];
  94     fshift           = fr->fshift[0];
  95     facel            = fr->epsfac;
  96     charge           = mdatoms->chargeA;
  97     krf              = fr->ic->k_rf;
  98     krf2             = krf*2.0;
  99     crf              = fr->ic->c_rf;
 100     nvdwtype         = fr->ntype;
 101     vdwparam         = fr->nbfp;
 102     vdwtype          = mdatoms->typeA;
 103
 104     /* Setup water-specific parameters */
 105     inr              = nlist->iinr[0];
 106     iq0              = facel*charge[inr+0];
 107     iq1              = facel*charge[inr+1];
 108     iq2              = facel*charge[inr+2];
 109     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 110
 111     jq0              = charge[inr+0];
 112     jq1              = charge[inr+1];
 113     jq2              = charge[inr+2];
 114     vdwjidx0         = 2*vdwtype[inr+0];
 115     qq00             = iq0*jq0;
 116     c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 117     c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 118     qq01             = iq0*jq1;
 119     qq02             = iq0*jq2;
 120     qq10             = iq1*jq0;
 121     qq11             = iq1*jq1;
 122     qq12             = iq1*jq2;
 123     qq20             = iq2*jq0;
 124     qq21             = iq2*jq1;
 125     qq22             = iq2*jq2;
 126
 127     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 128     rcutoff          = fr->rcoulomb;
 129     rcutoff2         = rcutoff*rcutoff;
 130
 131     sh_vdw_invrcut6  = fr->ic->sh_invrc6;
 132     rvdw             = fr->rvdw;
 133
 134     outeriter        = 0;
 135     inneriter        = 0;
 136
 137     /* Start outer loop over neighborlists */
 138     for(iidx=0; iidx<nri; iidx++)
 139     {
 140         /* Load shift vector for this list */
 141         i_shift_offset   = DIM*shiftidx[iidx];
 142         shX              = shiftvec[i_shift_offset+XX];
 143         shY              = shiftvec[i_shift_offset+YY];
 144         shZ              = shiftvec[i_shift_offset+ZZ];
 145
 146         /* Load limits for loop over neighbors */
 147         j_index_start    = jindex[iidx];
 148         j_index_end      = jindex[iidx+1];
 149
 150         /* Get outer coordinate index */
 151         inr              = iinr[iidx];
 152         i_coord_offset   = DIM*inr;
 153
 154         /* Load i particle coords and add shift vector */
 155         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 156         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 157         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 158         ix1              = shX + x[i_coord_offset+DIM*1+XX];
 159         iy1              = shY + x[i_coord_offset+DIM*1+YY];
 160         iz1              = shZ + x[i_coord_offset+DIM*1+ZZ];
 161         ix2              = shX + x[i_coord_offset+DIM*2+XX];
 162         iy2              = shY + x[i_coord_offset+DIM*2+YY];
 163         iz2              = shZ + x[i_coord_offset+DIM*2+ZZ];
 164
 165         fix0             = 0.0;
 166         fiy0             = 0.0;
 167         fiz0             = 0.0;
 168         fix1             = 0.0;
 169         fiy1             = 0.0;
 170         fiz1             = 0.0;
 171         fix2             = 0.0;
 172         fiy2             = 0.0;
 173         fiz2             = 0.0;
 174
 175         /* Reset potential sums */
 176         velecsum         = 0.0;
 177         vvdwsum          = 0.0;
 178
 179         /* Start inner kernel loop */
 180         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 181         {
 182             /* Get j neighbor index, and coordinate index */
 183             jnr              = jjnr[jidx];
 184             j_coord_offset   = DIM*jnr;
 185
 186             /* load j atom coordinates */
 187             jx0              = x[j_coord_offset+DIM*0+XX];
 188             jy0              = x[j_coord_offset+DIM*0+YY];
 189             jz0              = x[j_coord_offset+DIM*0+ZZ];
 190             jx1              = x[j_coord_offset+DIM*1+XX];
 191             jy1              = x[j_coord_offset+DIM*1+YY];
 192             jz1              = x[j_coord_offset+DIM*1+ZZ];
 193             jx2              = x[j_coord_offset+DIM*2+XX];
 194             jy2              = x[j_coord_offset+DIM*2+YY];
 195             jz2              = x[j_coord_offset+DIM*2+ZZ];
 196
 197             /* Calculate displacement vector */
 198             dx00             = ix0 - jx0;
 199             dy00             = iy0 - jy0;
 200             dz00             = iz0 - jz0;
 201             dx01             = ix0 - jx1;
 202             dy01             = iy0 - jy1;
 203             dz01             = iz0 - jz1;
 204             dx02             = ix0 - jx2;
 205             dy02             = iy0 - jy2;
 206             dz02             = iz0 - jz2;
 207             dx10             = ix1 - jx0;
 208             dy10             = iy1 - jy0;
 209             dz10             = iz1 - jz0;
 210             dx11             = ix1 - jx1;
 211             dy11             = iy1 - jy1;
 212             dz11             = iz1 - jz1;
 213             dx12             = ix1 - jx2;
 214             dy12             = iy1 - jy2;
 215             dz12             = iz1 - jz2;
 216             dx20             = ix2 - jx0;
 217             dy20             = iy2 - jy0;
 218             dz20             = iz2 - jz0;
 219             dx21             = ix2 - jx1;
 220             dy21             = iy2 - jy1;
 221             dz21             = iz2 - jz1;
 222             dx22             = ix2 - jx2;
 223             dy22             = iy2 - jy2;
 224             dz22             = iz2 - jz2;
 225
 226             /* Calculate squared distance and things based on it */
 227             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 228             rsq01            = dx01*dx01+dy01*dy01+dz01*dz01;
 229             rsq02            = dx02*dx02+dy02*dy02+dz02*dz02;
 230             rsq10            = dx10*dx10+dy10*dy10+dz10*dz10;
 231             rsq11            = dx11*dx11+dy11*dy11+dz11*dz11;
 232             rsq12            = dx12*dx12+dy12*dy12+dz12*dz12;
 233             rsq20            = dx20*dx20+dy20*dy20+dz20*dz20;
 234             rsq21            = dx21*dx21+dy21*dy21+dz21*dz21;
 235             rsq22            = dx22*dx22+dy22*dy22+dz22*dz22;
 236
 237             rinv00           = gmx_invsqrt(rsq00);
 238             rinv01           = gmx_invsqrt(rsq01);
 239             rinv02           = gmx_invsqrt(rsq02);
 240             rinv10           = gmx_invsqrt(rsq10);
 241             rinv11           = gmx_invsqrt(rsq11);
 242             rinv12           = gmx_invsqrt(rsq12);
 243             rinv20           = gmx_invsqrt(rsq20);
 244             rinv21           = gmx_invsqrt(rsq21);
 245             rinv22           = gmx_invsqrt(rsq22);
 246
 247             rinvsq00         = rinv00*rinv00;
 248             rinvsq01         = rinv01*rinv01;
 249             rinvsq02         = rinv02*rinv02;
 250             rinvsq10         = rinv10*rinv10;
 251             rinvsq11         = rinv11*rinv11;
 252             rinvsq12         = rinv12*rinv12;
 253             rinvsq20         = rinv20*rinv20;
 254             rinvsq21         = rinv21*rinv21;
 255             rinvsq22         = rinv22*rinv22;
 256
 257             /**************************
 258              * CALCULATE INTERACTIONS *
 259              **************************/
 260
 261             if (rsq00<rcutoff2)
 262             {
 263
 264             /* REACTION-FIELD ELECTROSTATICS */
 265             velec            = qq00*(rinv00+krf*rsq00-crf);
 266             felec            = qq00*(rinv00*rinvsq00-krf2);
 267
 268             /* LENNARD-JONES DISPERSION/REPULSION */
 269
 270             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 271             vvdw6            = c6_00*rinvsix;
 272             vvdw12           = c12_00*rinvsix*rinvsix;
 273             vvdw             = (vvdw12 - c12_00*sh_vdw_invrcut6*sh_vdw_invrcut6)*(1.0/12.0) - (vvdw6 - c6_00*sh_vdw_invrcut6)*(1.0/6.0);
 274             fvdw             = (vvdw12-vvdw6)*rinvsq00;
 275
 276             /* Update potential sums from outer loop */
 277             velecsum        += velec;
 278             vvdwsum         += vvdw;
 279
 280             fscal            = felec+fvdw;
 281
 282             /* Calculate temporary vectorial force */
 283             tx               = fscal*dx00;
 284             ty               = fscal*dy00;
 285             tz               = fscal*dz00;
 286
 287             /* Update vectorial force */
 288             fix0            += tx;
 289             fiy0            += ty;
 290             fiz0            += tz;
 291             f[j_coord_offset+DIM*0+XX] -= tx;
 292             f[j_coord_offset+DIM*0+YY] -= ty;
 293             f[j_coord_offset+DIM*0+ZZ] -= tz;
 294
 295             }
 296
 297             /**************************
 298              * CALCULATE INTERACTIONS *
 299              **************************/
 300
 301             if (rsq01<rcutoff2)
 302             {
 303
 304             /* REACTION-FIELD ELECTROSTATICS */
 305             velec            = qq01*(rinv01+krf*rsq01-crf);
 306             felec            = qq01*(rinv01*rinvsq01-krf2);
 307
 308             /* Update potential sums from outer loop */
 309             velecsum        += velec;
 310
 311             fscal            = felec;
 312
 313             /* Calculate temporary vectorial force */
 314             tx               = fscal*dx01;
 315             ty               = fscal*dy01;
 316             tz               = fscal*dz01;
 317
 318             /* Update vectorial force */
 319             fix0            += tx;
 320             fiy0            += ty;
 321             fiz0            += tz;
 322             f[j_coord_offset+DIM*1+XX] -= tx;
 323             f[j_coord_offset+DIM*1+YY] -= ty;
 324             f[j_coord_offset+DIM*1+ZZ] -= tz;
 325
 326             }
 327
 328             /**************************
 329              * CALCULATE INTERACTIONS *
 330              **************************/
 331
 332             if (rsq02<rcutoff2)
 333             {
 334
 335             /* REACTION-FIELD ELECTROSTATICS */
 336             velec            = qq02*(rinv02+krf*rsq02-crf);
 337             felec            = qq02*(rinv02*rinvsq02-krf2);
 338
 339             /* Update potential sums from outer loop */
 340             velecsum        += velec;
 341
 342             fscal            = felec;
 343
 344             /* Calculate temporary vectorial force */
 345             tx               = fscal*dx02;
 346             ty               = fscal*dy02;
 347             tz               = fscal*dz02;
 348
 349             /* Update vectorial force */
 350             fix0            += tx;
 351             fiy0            += ty;
 352             fiz0            += tz;
 353             f[j_coord_offset+DIM*2+XX] -= tx;
 354             f[j_coord_offset+DIM*2+YY] -= ty;
 355             f[j_coord_offset+DIM*2+ZZ] -= tz;
 356
 357             }
 358
 359             /**************************
 360              * CALCULATE INTERACTIONS *
 361              **************************/
 362
 363             if (rsq10<rcutoff2)
 364             {
 365
 366             /* REACTION-FIELD ELECTROSTATICS */
 367             velec            = qq10*(rinv10+krf*rsq10-crf);
 368             felec            = qq10*(rinv10*rinvsq10-krf2);
 369
 370             /* Update potential sums from outer loop */
 371             velecsum        += velec;
 372
 373             fscal            = felec;
 374
 375             /* Calculate temporary vectorial force */
 376             tx               = fscal*dx10;
 377             ty               = fscal*dy10;
 378             tz               = fscal*dz10;
 379
 380             /* Update vectorial force */
 381             fix1            += tx;
 382             fiy1            += ty;
 383             fiz1            += tz;
 384             f[j_coord_offset+DIM*0+XX] -= tx;
 385             f[j_coord_offset+DIM*0+YY] -= ty;
 386             f[j_coord_offset+DIM*0+ZZ] -= tz;
 387
 388             }
 389
 390             /**************************
 391              * CALCULATE INTERACTIONS *
 392              **************************/
 393
 394             if (rsq11<rcutoff2)
 395             {
 396
 397             /* REACTION-FIELD ELECTROSTATICS */
 398             velec            = qq11*(rinv11+krf*rsq11-crf);
 399             felec            = qq11*(rinv11*rinvsq11-krf2);
 400
 401             /* Update potential sums from outer loop */
 402             velecsum        += velec;
 403
 404             fscal            = felec;
 405
 406             /* Calculate temporary vectorial force */
 407             tx               = fscal*dx11;
 408             ty               = fscal*dy11;
 409             tz               = fscal*dz11;
 410
 411             /* Update vectorial force */
 412             fix1            += tx;
 413             fiy1            += ty;
 414             fiz1            += tz;
 415             f[j_coord_offset+DIM*1+XX] -= tx;
 416             f[j_coord_offset+DIM*1+YY] -= ty;
 417             f[j_coord_offset+DIM*1+ZZ] -= tz;
 418
 419             }
 420
 421             /**************************
 422              * CALCULATE INTERACTIONS *
 423              **************************/
 424
 425             if (rsq12<rcutoff2)
 426             {
 427
 428             /* REACTION-FIELD ELECTROSTATICS */
 429             velec            = qq12*(rinv12+krf*rsq12-crf);
 430             felec            = qq12*(rinv12*rinvsq12-krf2);
 431
 432             /* Update potential sums from outer loop */
 433             velecsum        += velec;
 434
 435             fscal            = felec;
 436
 437             /* Calculate temporary vectorial force */
 438             tx               = fscal*dx12;
 439             ty               = fscal*dy12;
 440             tz               = fscal*dz12;
 441
 442             /* Update vectorial force */
 443             fix1            += tx;
 444             fiy1            += ty;
 445             fiz1            += tz;
 446             f[j_coord_offset+DIM*2+XX] -= tx;
 447             f[j_coord_offset+DIM*2+YY] -= ty;
 448             f[j_coord_offset+DIM*2+ZZ] -= tz;
 449
 450             }
 451
 452             /**************************
 453              * CALCULATE INTERACTIONS *
 454              **************************/
 455
 456             if (rsq20<rcutoff2)
 457             {
 458
 459             /* REACTION-FIELD ELECTROSTATICS */
 460             velec            = qq20*(rinv20+krf*rsq20-crf);
 461             felec            = qq20*(rinv20*rinvsq20-krf2);
 462
 463             /* Update potential sums from outer loop */
 464             velecsum        += velec;
 465
 466             fscal            = felec;
 467
 468             /* Calculate temporary vectorial force */
 469             tx               = fscal*dx20;
 470             ty               = fscal*dy20;
 471             tz               = fscal*dz20;
 472
 473             /* Update vectorial force */
 474             fix2            += tx;
 475             fiy2            += ty;
 476             fiz2            += tz;
 477             f[j_coord_offset+DIM*0+XX] -= tx;
 478             f[j_coord_offset+DIM*0+YY] -= ty;
 479             f[j_coord_offset+DIM*0+ZZ] -= tz;
 480
 481             }
 482
 483             /**************************
 484              * CALCULATE INTERACTIONS *
 485              **************************/
 486
 487             if (rsq21<rcutoff2)
 488             {
 489
 490             /* REACTION-FIELD ELECTROSTATICS */
 491             velec            = qq21*(rinv21+krf*rsq21-crf);
 492             felec            = qq21*(rinv21*rinvsq21-krf2);
 493
 494             /* Update potential sums from outer loop */
 495             velecsum        += velec;
 496
 497             fscal            = felec;
 498
 499             /* Calculate temporary vectorial force */
 500             tx               = fscal*dx21;
 501             ty               = fscal*dy21;
 502             tz               = fscal*dz21;
 503
 504             /* Update vectorial force */
 505             fix2            += tx;
 506             fiy2            += ty;
 507             fiz2            += tz;
 508             f[j_coord_offset+DIM*1+XX] -= tx;
 509             f[j_coord_offset+DIM*1+YY] -= ty;
 510             f[j_coord_offset+DIM*1+ZZ] -= tz;
 511
 512             }
 513
 514             /**************************
 515              * CALCULATE INTERACTIONS *
 516              **************************/
 517
 518             if (rsq22<rcutoff2)
 519             {
 520
 521             /* REACTION-FIELD ELECTROSTATICS */
 522             velec            = qq22*(rinv22+krf*rsq22-crf);
 523             felec            = qq22*(rinv22*rinvsq22-krf2);
 524
 525             /* Update potential sums from outer loop */
 526             velecsum        += velec;
 527
 528             fscal            = felec;
 529
 530             /* Calculate temporary vectorial force */
 531             tx               = fscal*dx22;
 532             ty               = fscal*dy22;
 533             tz               = fscal*dz22;
 534
 535             /* Update vectorial force */
 536             fix2            += tx;
 537             fiy2            += ty;
 538             fiz2            += tz;
 539             f[j_coord_offset+DIM*2+XX] -= tx;
 540             f[j_coord_offset+DIM*2+YY] -= ty;
 541             f[j_coord_offset+DIM*2+ZZ] -= tz;
 542
 543             }
 544
 545             /* Inner loop uses 296 flops */
 546         }
 547         /* End of innermost loop */
 548
 549         tx = ty = tz = 0;
 550         f[i_coord_offset+DIM*0+XX] += fix0;
 551         f[i_coord_offset+DIM*0+YY] += fiy0;
 552         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 553         tx                         += fix0;
 554         ty                         += fiy0;
 555         tz                         += fiz0;
 556         f[i_coord_offset+DIM*1+XX] += fix1;
 557         f[i_coord_offset+DIM*1+YY] += fiy1;
 558         f[i_coord_offset+DIM*1+ZZ] += fiz1;
 559         tx                         += fix1;
 560         ty                         += fiy1;
 561         tz                         += fiz1;
 562         f[i_coord_offset+DIM*2+XX] += fix2;
 563         f[i_coord_offset+DIM*2+YY] += fiy2;
 564         f[i_coord_offset+DIM*2+ZZ] += fiz2;
 565         tx                         += fix2;
 566         ty                         += fiy2;
 567         tz                         += fiz2;
 568         fshift[i_shift_offset+XX]  += tx;
 569         fshift[i_shift_offset+YY]  += ty;
 570         fshift[i_shift_offset+ZZ]  += tz;
 571
 572         ggid                        = gid[iidx];
 573         /* Update potential energies */
 574         kernel_data->energygrp_elec[ggid] += velecsum;
 575         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 576
 577         /* Increment number of inner iterations */
 578         inneriter                  += j_index_end - j_index_start;
 579
 580         /* Outer loop uses 32 flops */
 581     }
 582
 583     /* Increment number of outer iterations */
 584     outeriter        += nri;
 585
 586     /* Update outer/inner flops */
 587
 588     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3W3_VF,outeriter*32 + inneriter*296);
 589 }
 590 /*
 591  * Gromacs nonbonded kernel:   nb_kernel_ElecRFCut_VdwLJSh_GeomW3W3_F_c
 592  * Electrostatics interaction: ReactionField
 593  * VdW interaction:            LennardJones
 594  * Geometry:                   Water3-Water3
 595  * Calculate force/pot:        Force
 596  */
 597 void
 598 nb_kernel_ElecRFCut_VdwLJSh_GeomW3W3_F_c
 599                     (t_nblist * gmx_restrict                nlist,
 600                      rvec * gmx_restrict                    xx,
 601                      rvec * gmx_restrict                    ff,
 602                      t_forcerec * gmx_restrict              fr,
 603                      t_mdatoms * gmx_restrict               mdatoms,
 604                      nb_kernel_data_t * gmx_restrict        kernel_data,
 605                      t_nrnb * gmx_restrict                  nrnb)
 606 {
 607     int              i_shift_offset,i_coord_offset,j_coord_offset;
 608     int              j_index_start,j_index_end;
 609     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 610     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 611     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 612     real             *shiftvec,*fshift,*x,*f;
 613     int              vdwioffset0;
 614     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 615     int              vdwioffset1;
 616     real             ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
 617     int              vdwioffset2;
 618     real             ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
 619     int              vdwjidx0;
 620     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 621     int              vdwjidx1;
 622     real             jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
 623     int              vdwjidx2;
 624     real             jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
 625     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 626     real             dx01,dy01,dz01,rsq01,rinv01,rinvsq01,r01,qq01,c6_01,c12_01,cexp1_01,cexp2_01;
 627     real             dx02,dy02,dz02,rsq02,rinv02,rinvsq02,r02,qq02,c6_02,c12_02,cexp1_02,cexp2_02;
 628     real             dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10,cexp1_10,cexp2_10;
 629     real             dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11,cexp1_11,cexp2_11;
 630     real             dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12,cexp1_12,cexp2_12;
 631     real             dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20,cexp1_20,cexp2_20;
 632     real             dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21,cexp1_21,cexp2_21;
 633     real             dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22,cexp1_22,cexp2_22;
 634     real             velec,felec,velecsum,facel,crf,krf,krf2;
 635     real             *charge;
 636     int              nvdwtype;
 637     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 638     int              *vdwtype;
 639     real             *vdwparam;
 640
 641     x                = xx[0];
 642     f                = ff[0];
 643
 644     nri              = nlist->nri;
 645     iinr             = nlist->iinr;
 646     jindex           = nlist->jindex;
 647     jjnr             = nlist->jjnr;
 648     shiftidx         = nlist->shift;
 649     gid              = nlist->gid;
 650     shiftvec         = fr->shift_vec[0];
 651     fshift           = fr->fshift[0];
 652     facel            = fr->epsfac;
 653     charge           = mdatoms->chargeA;
 654     krf              = fr->ic->k_rf;
 655     krf2             = krf*2.0;
 656     crf              = fr->ic->c_rf;
 657     nvdwtype         = fr->ntype;
 658     vdwparam         = fr->nbfp;
 659     vdwtype          = mdatoms->typeA;
 660
 661     /* Setup water-specific parameters */
 662     inr              = nlist->iinr[0];
 663     iq0              = facel*charge[inr+0];
 664     iq1              = facel*charge[inr+1];
 665     iq2              = facel*charge[inr+2];
 666     vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 667
 668     jq0              = charge[inr+0];
 669     jq1              = charge[inr+1];
 670     jq2              = charge[inr+2];
 671     vdwjidx0         = 2*vdwtype[inr+0];
 672     qq00             = iq0*jq0;
 673     c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 674     c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 675     qq01             = iq0*jq1;
 676     qq02             = iq0*jq2;
 677     qq10             = iq1*jq0;
 678     qq11             = iq1*jq1;
 679     qq12             = iq1*jq2;
 680     qq20             = iq2*jq0;
 681     qq21             = iq2*jq1;
 682     qq22             = iq2*jq2;
 683
 684     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 685     rcutoff          = fr->rcoulomb;
 686     rcutoff2         = rcutoff*rcutoff;
 687
 688     sh_vdw_invrcut6  = fr->ic->sh_invrc6;
 689     rvdw             = fr->rvdw;
 690
 691     outeriter        = 0;
 692     inneriter        = 0;
 693
 694     /* Start outer loop over neighborlists */
 695     for(iidx=0; iidx<nri; iidx++)
 696     {
 697         /* Load shift vector for this list */
 698         i_shift_offset   = DIM*shiftidx[iidx];
 699         shX              = shiftvec[i_shift_offset+XX];
 700         shY              = shiftvec[i_shift_offset+YY];
 701         shZ              = shiftvec[i_shift_offset+ZZ];
 702
 703         /* Load limits for loop over neighbors */
 704         j_index_start    = jindex[iidx];
 705         j_index_end      = jindex[iidx+1];
 706
 707         /* Get outer coordinate index */
 708         inr              = iinr[iidx];
 709         i_coord_offset   = DIM*inr;
 710
 711         /* Load i particle coords and add shift vector */
 712         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 713         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 714         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 715         ix1              = shX + x[i_coord_offset+DIM*1+XX];
 716         iy1              = shY + x[i_coord_offset+DIM*1+YY];
 717         iz1              = shZ + x[i_coord_offset+DIM*1+ZZ];
 718         ix2              = shX + x[i_coord_offset+DIM*2+XX];
 719         iy2              = shY + x[i_coord_offset+DIM*2+YY];
 720         iz2              = shZ + x[i_coord_offset+DIM*2+ZZ];
 721
 722         fix0             = 0.0;
 723         fiy0             = 0.0;
 724         fiz0             = 0.0;
 725         fix1             = 0.0;
 726         fiy1             = 0.0;
 727         fiz1             = 0.0;
 728         fix2             = 0.0;
 729         fiy2             = 0.0;
 730         fiz2             = 0.0;
 731
 732         /* Start inner kernel loop */
 733         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 734         {
 735             /* Get j neighbor index, and coordinate index */
 736             jnr              = jjnr[jidx];
 737             j_coord_offset   = DIM*jnr;
 738
 739             /* load j atom coordinates */
 740             jx0              = x[j_coord_offset+DIM*0+XX];
 741             jy0              = x[j_coord_offset+DIM*0+YY];
 742             jz0              = x[j_coord_offset+DIM*0+ZZ];
 743             jx1              = x[j_coord_offset+DIM*1+XX];
 744             jy1              = x[j_coord_offset+DIM*1+YY];
 745             jz1              = x[j_coord_offset+DIM*1+ZZ];
 746             jx2              = x[j_coord_offset+DIM*2+XX];
 747             jy2              = x[j_coord_offset+DIM*2+YY];
 748             jz2              = x[j_coord_offset+DIM*2+ZZ];
 749
 750             /* Calculate displacement vector */
 751             dx00             = ix0 - jx0;
 752             dy00             = iy0 - jy0;
 753             dz00             = iz0 - jz0;
 754             dx01             = ix0 - jx1;
 755             dy01             = iy0 - jy1;
 756             dz01             = iz0 - jz1;
 757             dx02             = ix0 - jx2;
 758             dy02             = iy0 - jy2;
 759             dz02             = iz0 - jz2;
 760             dx10             = ix1 - jx0;
 761             dy10             = iy1 - jy0;
 762             dz10             = iz1 - jz0;
 763             dx11             = ix1 - jx1;
 764             dy11             = iy1 - jy1;
 765             dz11             = iz1 - jz1;
 766             dx12             = ix1 - jx2;
 767             dy12             = iy1 - jy2;
 768             dz12             = iz1 - jz2;
 769             dx20             = ix2 - jx0;
 770             dy20             = iy2 - jy0;
 771             dz20             = iz2 - jz0;
 772             dx21             = ix2 - jx1;
 773             dy21             = iy2 - jy1;
 774             dz21             = iz2 - jz1;
 775             dx22             = ix2 - jx2;
 776             dy22             = iy2 - jy2;
 777             dz22             = iz2 - jz2;
 778
 779             /* Calculate squared distance and things based on it */
 780             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 781             rsq01            = dx01*dx01+dy01*dy01+dz01*dz01;
 782             rsq02            = dx02*dx02+dy02*dy02+dz02*dz02;
 783             rsq10            = dx10*dx10+dy10*dy10+dz10*dz10;
 784             rsq11            = dx11*dx11+dy11*dy11+dz11*dz11;
 785             rsq12            = dx12*dx12+dy12*dy12+dz12*dz12;
 786             rsq20            = dx20*dx20+dy20*dy20+dz20*dz20;
 787             rsq21            = dx21*dx21+dy21*dy21+dz21*dz21;
 788             rsq22            = dx22*dx22+dy22*dy22+dz22*dz22;
 789
 790             rinv00           = gmx_invsqrt(rsq00);
 791             rinv01           = gmx_invsqrt(rsq01);
 792             rinv02           = gmx_invsqrt(rsq02);
 793             rinv10           = gmx_invsqrt(rsq10);
 794             rinv11           = gmx_invsqrt(rsq11);
 795             rinv12           = gmx_invsqrt(rsq12);
 796             rinv20           = gmx_invsqrt(rsq20);
 797             rinv21           = gmx_invsqrt(rsq21);
 798             rinv22           = gmx_invsqrt(rsq22);
 799
 800             rinvsq00         = rinv00*rinv00;
 801             rinvsq01         = rinv01*rinv01;
 802             rinvsq02         = rinv02*rinv02;
 803             rinvsq10         = rinv10*rinv10;
 804             rinvsq11         = rinv11*rinv11;
 805             rinvsq12         = rinv12*rinv12;
 806             rinvsq20         = rinv20*rinv20;
 807             rinvsq21         = rinv21*rinv21;
 808             rinvsq22         = rinv22*rinv22;
 809
 810             /**************************
 811              * CALCULATE INTERACTIONS *
 812              **************************/
 813
 814             if (rsq00<rcutoff2)
 815             {
 816
 817             /* REACTION-FIELD ELECTROSTATICS */
 818             felec            = qq00*(rinv00*rinvsq00-krf2);
 819
 820             /* LENNARD-JONES DISPERSION/REPULSION */
 821
 822             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 823             fvdw             = (c12_00*rinvsix-c6_00)*rinvsix*rinvsq00;
 824
 825             fscal            = felec+fvdw;
 826
 827             /* Calculate temporary vectorial force */
 828             tx               = fscal*dx00;
 829             ty               = fscal*dy00;
 830             tz               = fscal*dz00;
 831
 832             /* Update vectorial force */
 833             fix0            += tx;
 834             fiy0            += ty;
 835             fiz0            += tz;
 836             f[j_coord_offset+DIM*0+XX] -= tx;
 837             f[j_coord_offset+DIM*0+YY] -= ty;
 838             f[j_coord_offset+DIM*0+ZZ] -= tz;
 839
 840             }
 841
 842             /**************************
 843              * CALCULATE INTERACTIONS *
 844              **************************/
 845
 846             if (rsq01<rcutoff2)
 847             {
 848
 849             /* REACTION-FIELD ELECTROSTATICS */
 850             felec            = qq01*(rinv01*rinvsq01-krf2);
 851
 852             fscal            = felec;
 853
 854             /* Calculate temporary vectorial force */
 855             tx               = fscal*dx01;
 856             ty               = fscal*dy01;
 857             tz               = fscal*dz01;
 858
 859             /* Update vectorial force */
 860             fix0            += tx;
 861             fiy0            += ty;
 862             fiz0            += tz;
 863             f[j_coord_offset+DIM*1+XX] -= tx;
 864             f[j_coord_offset+DIM*1+YY] -= ty;
 865             f[j_coord_offset+DIM*1+ZZ] -= tz;
 866
 867             }
 868
 869             /**************************
 870              * CALCULATE INTERACTIONS *
 871              **************************/
 872
 873             if (rsq02<rcutoff2)
 874             {
 875
 876             /* REACTION-FIELD ELECTROSTATICS */
 877             felec            = qq02*(rinv02*rinvsq02-krf2);
 878
 879             fscal            = felec;
 880
 881             /* Calculate temporary vectorial force */
 882             tx               = fscal*dx02;
 883             ty               = fscal*dy02;
 884             tz               = fscal*dz02;
 885
 886             /* Update vectorial force */
 887             fix0            += tx;
 888             fiy0            += ty;
 889             fiz0            += tz;
 890             f[j_coord_offset+DIM*2+XX] -= tx;
 891             f[j_coord_offset+DIM*2+YY] -= ty;
 892             f[j_coord_offset+DIM*2+ZZ] -= tz;
 893
 894             }
 895
 896             /**************************
 897              * CALCULATE INTERACTIONS *
 898              **************************/
 899
 900             if (rsq10<rcutoff2)
 901             {
 902
 903             /* REACTION-FIELD ELECTROSTATICS */
 904             felec            = qq10*(rinv10*rinvsq10-krf2);
 905
 906             fscal            = felec;
 907
 908             /* Calculate temporary vectorial force */
 909             tx               = fscal*dx10;
 910             ty               = fscal*dy10;
 911             tz               = fscal*dz10;
 912
 913             /* Update vectorial force */
 914             fix1            += tx;
 915             fiy1            += ty;
 916             fiz1            += tz;
 917             f[j_coord_offset+DIM*0+XX] -= tx;
 918             f[j_coord_offset+DIM*0+YY] -= ty;
 919             f[j_coord_offset+DIM*0+ZZ] -= tz;
 920
 921             }
 922
 923             /**************************
 924              * CALCULATE INTERACTIONS *
 925              **************************/
 926
 927             if (rsq11<rcutoff2)
 928             {
 929
 930             /* REACTION-FIELD ELECTROSTATICS */
 931             felec            = qq11*(rinv11*rinvsq11-krf2);
 932
 933             fscal            = felec;
 934
 935             /* Calculate temporary vectorial force */
 936             tx               = fscal*dx11;
 937             ty               = fscal*dy11;
 938             tz               = fscal*dz11;
 939
 940             /* Update vectorial force */
 941             fix1            += tx;
 942             fiy1            += ty;
 943             fiz1            += tz;
 944             f[j_coord_offset+DIM*1+XX] -= tx;
 945             f[j_coord_offset+DIM*1+YY] -= ty;
 946             f[j_coord_offset+DIM*1+ZZ] -= tz;
 947
 948             }
 949
 950             /**************************
 951              * CALCULATE INTERACTIONS *
 952              **************************/
 953
 954             if (rsq12<rcutoff2)
 955             {
 956
 957             /* REACTION-FIELD ELECTROSTATICS */
 958             felec            = qq12*(rinv12*rinvsq12-krf2);
 959
 960             fscal            = felec;
 961
 962             /* Calculate temporary vectorial force */
 963             tx               = fscal*dx12;
 964             ty               = fscal*dy12;
 965             tz               = fscal*dz12;
 966
 967             /* Update vectorial force */
 968             fix1            += tx;
 969             fiy1            += ty;
 970             fiz1            += tz;
 971             f[j_coord_offset+DIM*2+XX] -= tx;
 972             f[j_coord_offset+DIM*2+YY] -= ty;
 973             f[j_coord_offset+DIM*2+ZZ] -= tz;
 974
 975             }
 976
 977             /**************************
 978              * CALCULATE INTERACTIONS *
 979              **************************/
 980
 981             if (rsq20<rcutoff2)
 982             {
 983
 984             /* REACTION-FIELD ELECTROSTATICS */
 985             felec            = qq20*(rinv20*rinvsq20-krf2);
 986
 987             fscal            = felec;
 988
 989             /* Calculate temporary vectorial force */
 990             tx               = fscal*dx20;
 991             ty               = fscal*dy20;
 992             tz               = fscal*dz20;
 993
 994             /* Update vectorial force */
 995             fix2            += tx;
 996             fiy2            += ty;
 997             fiz2            += tz;
 998             f[j_coord_offset+DIM*0+XX] -= tx;
 999             f[j_coord_offset+DIM*0+YY] -= ty;
1000             f[j_coord_offset+DIM*0+ZZ] -= tz;
1001
1002             }
1003
1004             /**************************
1005              * CALCULATE INTERACTIONS *
1006              **************************/
1007
1008             if (rsq21<rcutoff2)
1009             {
1010
1011             /* REACTION-FIELD ELECTROSTATICS */
1012             felec            = qq21*(rinv21*rinvsq21-krf2);
1013
1014             fscal            = felec;
1015
1016             /* Calculate temporary vectorial force */
1017             tx               = fscal*dx21;
1018             ty               = fscal*dy21;
1019             tz               = fscal*dz21;
1020
1021             /* Update vectorial force */
1022             fix2            += tx;
1023             fiy2            += ty;
1024             fiz2            += tz;
1025             f[j_coord_offset+DIM*1+XX] -= tx;
1026             f[j_coord_offset+DIM*1+YY] -= ty;
1027             f[j_coord_offset+DIM*1+ZZ] -= tz;
1028
1029             }
1030
1031             /**************************
1032              * CALCULATE INTERACTIONS *
1033              **************************/
1034
1035             if (rsq22<rcutoff2)
1036             {
1037
1038             /* REACTION-FIELD ELECTROSTATICS */
1039             felec            = qq22*(rinv22*rinvsq22-krf2);
1040
1041             fscal            = felec;
1042
1043             /* Calculate temporary vectorial force */
1044             tx               = fscal*dx22;
1045             ty               = fscal*dy22;
1046             tz               = fscal*dz22;
1047
1048             /* Update vectorial force */
1049             fix2            += tx;
1050             fiy2            += ty;
1051             fiz2            += tz;
1052             f[j_coord_offset+DIM*2+XX] -= tx;
1053             f[j_coord_offset+DIM*2+YY] -= ty;
1054             f[j_coord_offset+DIM*2+ZZ] -= tz;
1055
1056             }
1057
1058             /* Inner loop uses 241 flops */
1059         }
1060         /* End of innermost loop */
1061
1062         tx = ty = tz = 0;
1063         f[i_coord_offset+DIM*0+XX] += fix0;
1064         f[i_coord_offset+DIM*0+YY] += fiy0;
1065         f[i_coord_offset+DIM*0+ZZ] += fiz0;
1066         tx                         += fix0;
1067         ty                         += fiy0;
1068         tz                         += fiz0;
1069         f[i_coord_offset+DIM*1+XX] += fix1;
1070         f[i_coord_offset+DIM*1+YY] += fiy1;
1071         f[i_coord_offset+DIM*1+ZZ] += fiz1;
1072         tx                         += fix1;
1073         ty                         += fiy1;
1074         tz                         += fiz1;
1075         f[i_coord_offset+DIM*2+XX] += fix2;
1076         f[i_coord_offset+DIM*2+YY] += fiy2;
1077         f[i_coord_offset+DIM*2+ZZ] += fiz2;
1078         tx                         += fix2;
1079         ty                         += fiy2;
1080         tz                         += fiz2;
1081         fshift[i_shift_offset+XX]  += tx;
1082         fshift[i_shift_offset+YY]  += ty;
1083         fshift[i_shift_offset+ZZ]  += tz;
1084
1085         /* Increment number of inner iterations */
1086         inneriter                  += j_index_end - j_index_start;
1087
1088         /* Outer loop uses 30 flops */
1089     }
1090
1091     /* Increment number of outer iterations */
1092     outeriter        += nri;
1093
1094     /* Update outer/inner flops */
1095
1096     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3W3_F,outeriter*30 + inneriter*241);
1097 }