src/gromacs/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecEwSh_VdwLJEwSh_GeomP1P1_c.c

   1 /*
   2  * This file is part of the GROMACS molecular simulation package.
   3  *
   4  * Copyright (c) 2012,2013,2014, by the GROMACS development team, led by
   5  * Mark Abraham, David van der Spoel, Berk Hess, and Erik Lindahl,
   6  * and including many others, as listed in the AUTHORS file in the
   7  * top-level source directory and at http://www.gromacs.org.
   8  *
   9  * GROMACS is free software; you can redistribute it and/or
  10  * modify it under the terms of the GNU Lesser General Public License
  11  * as published by the Free Software Foundation; either version 2.1
  12  * of the License, or (at your option) any later version.
  13  *
  14  * GROMACS is distributed in the hope that it will be useful,
  15  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  16  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  17  * Lesser General Public License for more details.
  18  *
  19  * You should have received a copy of the GNU Lesser General Public
  20  * License along with GROMACS; if not, see
  21  * http://www.gnu.org/licenses, or write to the Free Software Foundation,
  22  * Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA.
  23  *
  24  * If you want to redistribute modifications to GROMACS, please
  25  * consider that scientific software is very special. Version
  26  * control is crucial - bugs must be traceable. We will be happy to
  27  * consider code for inclusion in the official distribution, but
  28  * derived work must not be called official GROMACS. Details are found
  29  * in the README & COPYING files - if they are missing, get the
  30  * official version at http://www.gromacs.org.
  31  *
  32  * To help us fund GROMACS development, we humbly ask that you cite
  33  * the research papers on the package. Check out http://www.gromacs.org.
  34  */
  35 /*
  36  * Note: this file was generated by the GROMACS c kernel generator.
  37  */
  38 #include "config.h"
  39
  40 #include <math.h>
  41
  42 #include "../nb_kernel.h"
  43 #include "gromacs/legacyheaders/types/simple.h"
  44 #include "gromacs/math/vec.h"
  45 #include "gromacs/legacyheaders/nrnb.h"
  46
  47 /*
  48  * Gromacs nonbonded kernel:   nb_kernel_ElecEwSh_VdwLJEwSh_GeomP1P1_VF_c
  49  * Electrostatics interaction: Ewald
  50  * VdW interaction:            LJEwald
  51  * Geometry:                   Particle-Particle
  52  * Calculate force/pot:        PotentialAndForce
  53  */
  54 void
  55 nb_kernel_ElecEwSh_VdwLJEwSh_GeomP1P1_VF_c
  56                     (t_nblist                    * gmx_restrict       nlist,
  57                      rvec                        * gmx_restrict          xx,
  58                      rvec                        * gmx_restrict          ff,
  59                      t_forcerec                  * gmx_restrict          fr,
  60                      t_mdatoms                   * gmx_restrict     mdatoms,
  61                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
  62                      t_nrnb                      * gmx_restrict        nrnb)
  63 {
  64     int              i_shift_offset,i_coord_offset,j_coord_offset;
  65     int              j_index_start,j_index_end;
  66     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
  67     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
  68     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  69     real             *shiftvec,*fshift,*x,*f;
  70     int              vdwioffset0;
  71     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  72     int              vdwjidx0;
  73     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  74     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
  75     real             velec,felec,velecsum,facel,crf,krf,krf2;
  76     real             *charge;
  77     int              nvdwtype;
  78     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
  79     int              *vdwtype;
  80     real             *vdwparam;
  81     real             c6grid_00;
  82     real             ewclj,ewclj2,ewclj6,ewcljrsq,poly,exponent,sh_lj_ewald;
  83     real             *vdwgridparam;
  84     int              ewitab;
  85     real             ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
  86     real             *ewtab;
  87
  88     x                = xx[0];
  89     f                = ff[0];
  90
  91     nri              = nlist->nri;
  92     iinr             = nlist->iinr;
  93     jindex           = nlist->jindex;
  94     jjnr             = nlist->jjnr;
  95     shiftidx         = nlist->shift;
  96     gid              = nlist->gid;
  97     shiftvec         = fr->shift_vec[0];
  98     fshift           = fr->fshift[0];
  99     facel            = fr->epsfac;
 100     charge           = mdatoms->chargeA;
 101     nvdwtype         = fr->ntype;
 102     vdwparam         = fr->nbfp;
 103     vdwtype          = mdatoms->typeA;
 104     vdwgridparam     = fr->ljpme_c6grid;
 105     ewclj            = fr->ewaldcoeff_lj;
 106     sh_lj_ewald      = fr->ic->sh_lj_ewald;
 107     ewclj2           = ewclj*ewclj;
 108     ewclj6           = ewclj2*ewclj2*ewclj2;
 109
 110     sh_ewald         = fr->ic->sh_ewald;
 111     ewtab            = fr->ic->tabq_coul_FDV0;
 112     ewtabscale       = fr->ic->tabq_scale;
 113     ewtabhalfspace   = 0.5/ewtabscale;
 114
 115     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 116     rcutoff          = fr->rcoulomb;
 117     rcutoff2         = rcutoff*rcutoff;
 118
 119     sh_vdw_invrcut6  = fr->ic->sh_invrc6;
 120     rvdw             = fr->rvdw;
 121
 122     outeriter        = 0;
 123     inneriter        = 0;
 124
 125     /* Start outer loop over neighborlists */
 126     for(iidx=0; iidx<nri; iidx++)
 127     {
 128         /* Load shift vector for this list */
 129         i_shift_offset   = DIM*shiftidx[iidx];
 130         shX              = shiftvec[i_shift_offset+XX];
 131         shY              = shiftvec[i_shift_offset+YY];
 132         shZ              = shiftvec[i_shift_offset+ZZ];
 133
 134         /* Load limits for loop over neighbors */
 135         j_index_start    = jindex[iidx];
 136         j_index_end      = jindex[iidx+1];
 137
 138         /* Get outer coordinate index */
 139         inr              = iinr[iidx];
 140         i_coord_offset   = DIM*inr;
 141
 142         /* Load i particle coords and add shift vector */
 143         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 144         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 145         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 146
 147         fix0             = 0.0;
 148         fiy0             = 0.0;
 149         fiz0             = 0.0;
 150
 151         /* Load parameters for i particles */
 152         iq0              = facel*charge[inr+0];
 153         vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 154
 155         /* Reset potential sums */
 156         velecsum         = 0.0;
 157         vvdwsum          = 0.0;
 158
 159         /* Start inner kernel loop */
 160         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 161         {
 162             /* Get j neighbor index, and coordinate index */
 163             jnr              = jjnr[jidx];
 164             j_coord_offset   = DIM*jnr;
 165
 166             /* load j atom coordinates */
 167             jx0              = x[j_coord_offset+DIM*0+XX];
 168             jy0              = x[j_coord_offset+DIM*0+YY];
 169             jz0              = x[j_coord_offset+DIM*0+ZZ];
 170
 171             /* Calculate displacement vector */
 172             dx00             = ix0 - jx0;
 173             dy00             = iy0 - jy0;
 174             dz00             = iz0 - jz0;
 175
 176             /* Calculate squared distance and things based on it */
 177             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 178
 179             rinv00           = gmx_invsqrt(rsq00);
 180
 181             rinvsq00         = rinv00*rinv00;
 182
 183             /* Load parameters for j particles */
 184             jq0              = charge[jnr+0];
 185             vdwjidx0         = 2*vdwtype[jnr+0];
 186
 187             /**************************
 188              * CALCULATE INTERACTIONS *
 189              **************************/
 190
 191             if (rsq00<rcutoff2)
 192             {
 193
 194             r00              = rsq00*rinv00;
 195
 196             qq00             = iq0*jq0;
 197             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 198             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 199             c6grid_00        = vdwgridparam[vdwioffset0+vdwjidx0];
 200
 201             /* EWALD ELECTROSTATICS */
 202
 203             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 204             ewrt             = r00*ewtabscale;
 205             ewitab           = ewrt;
 206             eweps            = ewrt-ewitab;
 207             ewitab           = 4*ewitab;
 208             felec            = ewtab[ewitab]+eweps*ewtab[ewitab+1];
 209             velec            = qq00*((rinv00-sh_ewald)-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
 210             felec            = qq00*rinv00*(rinvsq00-felec);
 211
 212             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 213             ewcljrsq         = ewclj2*rsq00;
 214             exponent         = exp(-ewcljrsq);
 215             poly             = exponent*(1.0 + ewcljrsq + ewcljrsq*ewcljrsq*0.5);
 216             vvdw6            = (c6_00-c6grid_00*(1.0-poly))*rinvsix;
 217             vvdw12           = c12_00*rinvsix*rinvsix;
 218             vvdw             = (vvdw12 - c12_00*sh_vdw_invrcut6*sh_vdw_invrcut6)*(1.0/12.0) - (vvdw6 - c6_00*sh_vdw_invrcut6 - c6grid_00*sh_lj_ewald)*(1.0/6.0);
 219             fvdw             = (vvdw12 - vvdw6 - c6grid_00*(1.0/6.0)*exponent*ewclj6)*rinvsq00;
 220
 221             /* Update potential sums from outer loop */
 222             velecsum        += velec;
 223             vvdwsum         += vvdw;
 224
 225             fscal            = felec+fvdw;
 226
 227             /* Calculate temporary vectorial force */
 228             tx               = fscal*dx00;
 229             ty               = fscal*dy00;
 230             tz               = fscal*dz00;
 231
 232             /* Update vectorial force */
 233             fix0            += tx;
 234             fiy0            += ty;
 235             fiz0            += tz;
 236             f[j_coord_offset+DIM*0+XX] -= tx;
 237             f[j_coord_offset+DIM*0+YY] -= ty;
 238             f[j_coord_offset+DIM*0+ZZ] -= tz;
 239
 240             }
 241
 242             /* Inner loop uses 74 flops */
 243         }
 244         /* End of innermost loop */
 245
 246         tx = ty = tz = 0;
 247         f[i_coord_offset+DIM*0+XX] += fix0;
 248         f[i_coord_offset+DIM*0+YY] += fiy0;
 249         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 250         tx                         += fix0;
 251         ty                         += fiy0;
 252         tz                         += fiz0;
 253         fshift[i_shift_offset+XX]  += tx;
 254         fshift[i_shift_offset+YY]  += ty;
 255         fshift[i_shift_offset+ZZ]  += tz;
 256
 257         ggid                        = gid[iidx];
 258         /* Update potential energies */
 259         kernel_data->energygrp_elec[ggid] += velecsum;
 260         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 261
 262         /* Increment number of inner iterations */
 263         inneriter                  += j_index_end - j_index_start;
 264
 265         /* Outer loop uses 15 flops */
 266     }
 267
 268     /* Increment number of outer iterations */
 269     outeriter        += nri;
 270
 271     /* Update outer/inner flops */
 272
 273     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_VF,outeriter*15 + inneriter*74);
 274 }
 275 /*
 276  * Gromacs nonbonded kernel:   nb_kernel_ElecEwSh_VdwLJEwSh_GeomP1P1_F_c
 277  * Electrostatics interaction: Ewald
 278  * VdW interaction:            LJEwald
 279  * Geometry:                   Particle-Particle
 280  * Calculate force/pot:        Force
 281  */
 282 void
 283 nb_kernel_ElecEwSh_VdwLJEwSh_GeomP1P1_F_c
 284                     (t_nblist                    * gmx_restrict       nlist,
 285                      rvec                        * gmx_restrict          xx,
 286                      rvec                        * gmx_restrict          ff,
 287                      t_forcerec                  * gmx_restrict          fr,
 288                      t_mdatoms                   * gmx_restrict     mdatoms,
 289                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 290                      t_nrnb                      * gmx_restrict        nrnb)
 291 {
 292     int              i_shift_offset,i_coord_offset,j_coord_offset;
 293     int              j_index_start,j_index_end;
 294     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 295     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 296     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 297     real             *shiftvec,*fshift,*x,*f;
 298     int              vdwioffset0;
 299     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 300     int              vdwjidx0;
 301     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 302     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 303     real             velec,felec,velecsum,facel,crf,krf,krf2;
 304     real             *charge;
 305     int              nvdwtype;
 306     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 307     int              *vdwtype;
 308     real             *vdwparam;
 309     real             c6grid_00;
 310     real             ewclj,ewclj2,ewclj6,ewcljrsq,poly,exponent,sh_lj_ewald;
 311     real             *vdwgridparam;
 312     int              ewitab;
 313     real             ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
 314     real             *ewtab;
 315
 316     x                = xx[0];
 317     f                = ff[0];
 318
 319     nri              = nlist->nri;
 320     iinr             = nlist->iinr;
 321     jindex           = nlist->jindex;
 322     jjnr             = nlist->jjnr;
 323     shiftidx         = nlist->shift;
 324     gid              = nlist->gid;
 325     shiftvec         = fr->shift_vec[0];
 326     fshift           = fr->fshift[0];
 327     facel            = fr->epsfac;
 328     charge           = mdatoms->chargeA;
 329     nvdwtype         = fr->ntype;
 330     vdwparam         = fr->nbfp;
 331     vdwtype          = mdatoms->typeA;
 332     vdwgridparam     = fr->ljpme_c6grid;
 333     ewclj            = fr->ewaldcoeff_lj;
 334     sh_lj_ewald      = fr->ic->sh_lj_ewald;
 335     ewclj2           = ewclj*ewclj;
 336     ewclj6           = ewclj2*ewclj2*ewclj2;
 337
 338     sh_ewald         = fr->ic->sh_ewald;
 339     ewtab            = fr->ic->tabq_coul_F;
 340     ewtabscale       = fr->ic->tabq_scale;
 341     ewtabhalfspace   = 0.5/ewtabscale;
 342
 343     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 344     rcutoff          = fr->rcoulomb;
 345     rcutoff2         = rcutoff*rcutoff;
 346
 347     sh_vdw_invrcut6  = fr->ic->sh_invrc6;
 348     rvdw             = fr->rvdw;
 349
 350     outeriter        = 0;
 351     inneriter        = 0;
 352
 353     /* Start outer loop over neighborlists */
 354     for(iidx=0; iidx<nri; iidx++)
 355     {
 356         /* Load shift vector for this list */
 357         i_shift_offset   = DIM*shiftidx[iidx];
 358         shX              = shiftvec[i_shift_offset+XX];
 359         shY              = shiftvec[i_shift_offset+YY];
 360         shZ              = shiftvec[i_shift_offset+ZZ];
 361
 362         /* Load limits for loop over neighbors */
 363         j_index_start    = jindex[iidx];
 364         j_index_end      = jindex[iidx+1];
 365
 366         /* Get outer coordinate index */
 367         inr              = iinr[iidx];
 368         i_coord_offset   = DIM*inr;
 369
 370         /* Load i particle coords and add shift vector */
 371         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 372         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 373         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 374
 375         fix0             = 0.0;
 376         fiy0             = 0.0;
 377         fiz0             = 0.0;
 378
 379         /* Load parameters for i particles */
 380         iq0              = facel*charge[inr+0];
 381         vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 382
 383         /* Start inner kernel loop */
 384         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 385         {
 386             /* Get j neighbor index, and coordinate index */
 387             jnr              = jjnr[jidx];
 388             j_coord_offset   = DIM*jnr;
 389
 390             /* load j atom coordinates */
 391             jx0              = x[j_coord_offset+DIM*0+XX];
 392             jy0              = x[j_coord_offset+DIM*0+YY];
 393             jz0              = x[j_coord_offset+DIM*0+ZZ];
 394
 395             /* Calculate displacement vector */
 396             dx00             = ix0 - jx0;
 397             dy00             = iy0 - jy0;
 398             dz00             = iz0 - jz0;
 399
 400             /* Calculate squared distance and things based on it */
 401             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 402
 403             rinv00           = gmx_invsqrt(rsq00);
 404
 405             rinvsq00         = rinv00*rinv00;
 406
 407             /* Load parameters for j particles */
 408             jq0              = charge[jnr+0];
 409             vdwjidx0         = 2*vdwtype[jnr+0];
 410
 411             /**************************
 412              * CALCULATE INTERACTIONS *
 413              **************************/
 414
 415             if (rsq00<rcutoff2)
 416             {
 417
 418             r00              = rsq00*rinv00;
 419
 420             qq00             = iq0*jq0;
 421             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 422             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 423             c6grid_00        = vdwgridparam[vdwioffset0+vdwjidx0];
 424
 425             /* EWALD ELECTROSTATICS */
 426
 427             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 428             ewrt             = r00*ewtabscale;
 429             ewitab           = ewrt;
 430             eweps            = ewrt-ewitab;
 431             felec            = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1];
 432             felec            = qq00*rinv00*(rinvsq00-felec);
 433
 434             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 435             ewcljrsq         = ewclj2*rsq00;
 436             exponent         = exp(-ewcljrsq);
 437             poly             = exponent*(1.0 + ewcljrsq + ewcljrsq*ewcljrsq*0.5);
 438             fvdw             = (((c12_00*rinvsix - c6_00 + c6grid_00*(1.0-poly))*rinvsix) - c6grid_00*(1.0/6.0)*exponent*ewclj6)*rinvsq00;
 439
 440             fscal            = felec+fvdw;
 441
 442             /* Calculate temporary vectorial force */
 443             tx               = fscal*dx00;
 444             ty               = fscal*dy00;
 445             tz               = fscal*dz00;
 446
 447             /* Update vectorial force */
 448             fix0            += tx;
 449             fiy0            += ty;
 450             fiz0            += tz;
 451             f[j_coord_offset+DIM*0+XX] -= tx;
 452             f[j_coord_offset+DIM*0+YY] -= ty;
 453             f[j_coord_offset+DIM*0+ZZ] -= tz;
 454
 455             }
 456
 457             /* Inner loop uses 55 flops */
 458         }
 459         /* End of innermost loop */
 460
 461         tx = ty = tz = 0;
 462         f[i_coord_offset+DIM*0+XX] += fix0;
 463         f[i_coord_offset+DIM*0+YY] += fiy0;
 464         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 465         tx                         += fix0;
 466         ty                         += fiy0;
 467         tz                         += fiz0;
 468         fshift[i_shift_offset+XX]  += tx;
 469         fshift[i_shift_offset+YY]  += ty;
 470         fshift[i_shift_offset+ZZ]  += tz;
 471
 472         /* Increment number of inner iterations */
 473         inneriter                  += j_index_end - j_index_start;
 474
 475         /* Outer loop uses 13 flops */
 476     }
 477
 478     /* Increment number of outer iterations */
 479     outeriter        += nri;
 480
 481     /* Update outer/inner flops */
 482
 483     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_F,outeriter*13 + inneriter*55);
 484 }