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 #ifdef HAVE_CONFIG_H
  39 #include <config.h>
  40 #endif
  41
  42 #include <math.h>
  43
  44 #include "../nb_kernel.h"
  45 #include "types/simple.h"
  46 #include "vec.h"
  47 #include "nrnb.h"
  48
  49 /*
  50  * Gromacs nonbonded kernel:   nb_kernel_ElecEwSh_VdwLJEwSh_GeomP1P1_VF_c
  51  * Electrostatics interaction: Ewald
  52  * VdW interaction:            LJEwald
  53  * Geometry:                   Particle-Particle
  54  * Calculate force/pot:        PotentialAndForce
  55  */
  56 void
  57 nb_kernel_ElecEwSh_VdwLJEwSh_GeomP1P1_VF_c
  58                     (t_nblist                    * gmx_restrict       nlist,
  59                      rvec                        * gmx_restrict          xx,
  60                      rvec                        * gmx_restrict          ff,
  61                      t_forcerec                  * gmx_restrict          fr,
  62                      t_mdatoms                   * gmx_restrict     mdatoms,
  63                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
  64                      t_nrnb                      * gmx_restrict        nrnb)
  65 {
  66     int              i_shift_offset,i_coord_offset,j_coord_offset;
  67     int              j_index_start,j_index_end;
  68     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
  69     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
  70     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  71     real             *shiftvec,*fshift,*x,*f;
  72     int              vdwioffset0;
  73     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  74     int              vdwjidx0;
  75     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  76     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
  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     real             c6grid_00;
  84     real             ewclj,ewclj2,ewclj6,ewcljrsq,poly,exponent,sh_lj_ewald;
  85     real             *vdwgridparam;
  86     int              ewitab;
  87     real             ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
  88     real             *ewtab;
  89
  90     x                = xx[0];
  91     f                = ff[0];
  92
  93     nri              = nlist->nri;
  94     iinr             = nlist->iinr;
  95     jindex           = nlist->jindex;
  96     jjnr             = nlist->jjnr;
  97     shiftidx         = nlist->shift;
  98     gid              = nlist->gid;
  99     shiftvec         = fr->shift_vec[0];
 100     fshift           = fr->fshift[0];
 101     facel            = fr->epsfac;
 102     charge           = mdatoms->chargeA;
 103     nvdwtype         = fr->ntype;
 104     vdwparam         = fr->nbfp;
 105     vdwtype          = mdatoms->typeA;
 106     vdwgridparam     = fr->ljpme_c6grid;
 107     ewclj            = fr->ewaldcoeff_lj;
 108     sh_lj_ewald      = fr->ic->sh_lj_ewald;
 109     ewclj2           = ewclj*ewclj;
 110     ewclj6           = ewclj2*ewclj2*ewclj2;
 111
 112     sh_ewald         = fr->ic->sh_ewald;
 113     ewtab            = fr->ic->tabq_coul_FDV0;
 114     ewtabscale       = fr->ic->tabq_scale;
 115     ewtabhalfspace   = 0.5/ewtabscale;
 116
 117     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 118     rcutoff          = fr->rcoulomb;
 119     rcutoff2         = rcutoff*rcutoff;
 120
 121     sh_vdw_invrcut6  = fr->ic->sh_invrc6;
 122     rvdw             = fr->rvdw;
 123
 124     outeriter        = 0;
 125     inneriter        = 0;
 126
 127     /* Start outer loop over neighborlists */
 128     for(iidx=0; iidx<nri; iidx++)
 129     {
 130         /* Load shift vector for this list */
 131         i_shift_offset   = DIM*shiftidx[iidx];
 132         shX              = shiftvec[i_shift_offset+XX];
 133         shY              = shiftvec[i_shift_offset+YY];
 134         shZ              = shiftvec[i_shift_offset+ZZ];
 135
 136         /* Load limits for loop over neighbors */
 137         j_index_start    = jindex[iidx];
 138         j_index_end      = jindex[iidx+1];
 139
 140         /* Get outer coordinate index */
 141         inr              = iinr[iidx];
 142         i_coord_offset   = DIM*inr;
 143
 144         /* Load i particle coords and add shift vector */
 145         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 146         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 147         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 148
 149         fix0             = 0.0;
 150         fiy0             = 0.0;
 151         fiz0             = 0.0;
 152
 153         /* Load parameters for i particles */
 154         iq0              = facel*charge[inr+0];
 155         vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 156
 157         /* Reset potential sums */
 158         velecsum         = 0.0;
 159         vvdwsum          = 0.0;
 160
 161         /* Start inner kernel loop */
 162         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 163         {
 164             /* Get j neighbor index, and coordinate index */
 165             jnr              = jjnr[jidx];
 166             j_coord_offset   = DIM*jnr;
 167
 168             /* load j atom coordinates */
 169             jx0              = x[j_coord_offset+DIM*0+XX];
 170             jy0              = x[j_coord_offset+DIM*0+YY];
 171             jz0              = x[j_coord_offset+DIM*0+ZZ];
 172
 173             /* Calculate displacement vector */
 174             dx00             = ix0 - jx0;
 175             dy00             = iy0 - jy0;
 176             dz00             = iz0 - jz0;
 177
 178             /* Calculate squared distance and things based on it */
 179             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 180
 181             rinv00           = gmx_invsqrt(rsq00);
 182
 183             rinvsq00         = rinv00*rinv00;
 184
 185             /* Load parameters for j particles */
 186             jq0              = charge[jnr+0];
 187             vdwjidx0         = 2*vdwtype[jnr+0];
 188
 189             /**************************
 190              * CALCULATE INTERACTIONS *
 191              **************************/
 192
 193             if (rsq00<rcutoff2)
 194             {
 195
 196             r00              = rsq00*rinv00;
 197
 198             qq00             = iq0*jq0;
 199             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 200             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 201             c6grid_00        = vdwgridparam[vdwioffset0+vdwjidx0];
 202
 203             /* EWALD ELECTROSTATICS */
 204
 205             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 206             ewrt             = r00*ewtabscale;
 207             ewitab           = ewrt;
 208             eweps            = ewrt-ewitab;
 209             ewitab           = 4*ewitab;
 210             felec            = ewtab[ewitab]+eweps*ewtab[ewitab+1];
 211             velec            = qq00*((rinv00-sh_ewald)-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
 212             felec            = qq00*rinv00*(rinvsq00-felec);
 213
 214             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 215             ewcljrsq         = ewclj2*rsq00;
 216             exponent         = exp(-ewcljrsq);
 217             poly             = exponent*(1.0 + ewcljrsq + ewcljrsq*ewcljrsq*0.5);
 218             vvdw6            = (c6_00-c6grid_00*(1.0-poly))*rinvsix;
 219             vvdw12           = c12_00*rinvsix*rinvsix;
 220             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);
 221             fvdw             = (vvdw12 - vvdw6 - c6grid_00*(1.0/6.0)*exponent*ewclj6)*rinvsq00;
 222
 223             /* Update potential sums from outer loop */
 224             velecsum        += velec;
 225             vvdwsum         += vvdw;
 226
 227             fscal            = felec+fvdw;
 228
 229             /* Calculate temporary vectorial force */
 230             tx               = fscal*dx00;
 231             ty               = fscal*dy00;
 232             tz               = fscal*dz00;
 233
 234             /* Update vectorial force */
 235             fix0            += tx;
 236             fiy0            += ty;
 237             fiz0            += tz;
 238             f[j_coord_offset+DIM*0+XX] -= tx;
 239             f[j_coord_offset+DIM*0+YY] -= ty;
 240             f[j_coord_offset+DIM*0+ZZ] -= tz;
 241
 242             }
 243
 244             /* Inner loop uses 74 flops */
 245         }
 246         /* End of innermost loop */
 247
 248         tx = ty = tz = 0;
 249         f[i_coord_offset+DIM*0+XX] += fix0;
 250         f[i_coord_offset+DIM*0+YY] += fiy0;
 251         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 252         tx                         += fix0;
 253         ty                         += fiy0;
 254         tz                         += fiz0;
 255         fshift[i_shift_offset+XX]  += tx;
 256         fshift[i_shift_offset+YY]  += ty;
 257         fshift[i_shift_offset+ZZ]  += tz;
 258
 259         ggid                        = gid[iidx];
 260         /* Update potential energies */
 261         kernel_data->energygrp_elec[ggid] += velecsum;
 262         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 263
 264         /* Increment number of inner iterations */
 265         inneriter                  += j_index_end - j_index_start;
 266
 267         /* Outer loop uses 15 flops */
 268     }
 269
 270     /* Increment number of outer iterations */
 271     outeriter        += nri;
 272
 273     /* Update outer/inner flops */
 274
 275     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_VF,outeriter*15 + inneriter*74);
 276 }
 277 /*
 278  * Gromacs nonbonded kernel:   nb_kernel_ElecEwSh_VdwLJEwSh_GeomP1P1_F_c
 279  * Electrostatics interaction: Ewald
 280  * VdW interaction:            LJEwald
 281  * Geometry:                   Particle-Particle
 282  * Calculate force/pot:        Force
 283  */
 284 void
 285 nb_kernel_ElecEwSh_VdwLJEwSh_GeomP1P1_F_c
 286                     (t_nblist                    * gmx_restrict       nlist,
 287                      rvec                        * gmx_restrict          xx,
 288                      rvec                        * gmx_restrict          ff,
 289                      t_forcerec                  * gmx_restrict          fr,
 290                      t_mdatoms                   * gmx_restrict     mdatoms,
 291                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 292                      t_nrnb                      * gmx_restrict        nrnb)
 293 {
 294     int              i_shift_offset,i_coord_offset,j_coord_offset;
 295     int              j_index_start,j_index_end;
 296     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 297     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 298     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 299     real             *shiftvec,*fshift,*x,*f;
 300     int              vdwioffset0;
 301     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 302     int              vdwjidx0;
 303     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 304     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 305     real             velec,felec,velecsum,facel,crf,krf,krf2;
 306     real             *charge;
 307     int              nvdwtype;
 308     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 309     int              *vdwtype;
 310     real             *vdwparam;
 311     real             c6grid_00;
 312     real             ewclj,ewclj2,ewclj6,ewcljrsq,poly,exponent,sh_lj_ewald;
 313     real             *vdwgridparam;
 314     int              ewitab;
 315     real             ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
 316     real             *ewtab;
 317
 318     x                = xx[0];
 319     f                = ff[0];
 320
 321     nri              = nlist->nri;
 322     iinr             = nlist->iinr;
 323     jindex           = nlist->jindex;
 324     jjnr             = nlist->jjnr;
 325     shiftidx         = nlist->shift;
 326     gid              = nlist->gid;
 327     shiftvec         = fr->shift_vec[0];
 328     fshift           = fr->fshift[0];
 329     facel            = fr->epsfac;
 330     charge           = mdatoms->chargeA;
 331     nvdwtype         = fr->ntype;
 332     vdwparam         = fr->nbfp;
 333     vdwtype          = mdatoms->typeA;
 334     vdwgridparam     = fr->ljpme_c6grid;
 335     ewclj            = fr->ewaldcoeff_lj;
 336     sh_lj_ewald      = fr->ic->sh_lj_ewald;
 337     ewclj2           = ewclj*ewclj;
 338     ewclj6           = ewclj2*ewclj2*ewclj2;
 339
 340     sh_ewald         = fr->ic->sh_ewald;
 341     ewtab            = fr->ic->tabq_coul_F;
 342     ewtabscale       = fr->ic->tabq_scale;
 343     ewtabhalfspace   = 0.5/ewtabscale;
 344
 345     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 346     rcutoff          = fr->rcoulomb;
 347     rcutoff2         = rcutoff*rcutoff;
 348
 349     sh_vdw_invrcut6  = fr->ic->sh_invrc6;
 350     rvdw             = fr->rvdw;
 351
 352     outeriter        = 0;
 353     inneriter        = 0;
 354
 355     /* Start outer loop over neighborlists */
 356     for(iidx=0; iidx<nri; iidx++)
 357     {
 358         /* Load shift vector for this list */
 359         i_shift_offset   = DIM*shiftidx[iidx];
 360         shX              = shiftvec[i_shift_offset+XX];
 361         shY              = shiftvec[i_shift_offset+YY];
 362         shZ              = shiftvec[i_shift_offset+ZZ];
 363
 364         /* Load limits for loop over neighbors */
 365         j_index_start    = jindex[iidx];
 366         j_index_end      = jindex[iidx+1];
 367
 368         /* Get outer coordinate index */
 369         inr              = iinr[iidx];
 370         i_coord_offset   = DIM*inr;
 371
 372         /* Load i particle coords and add shift vector */
 373         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 374         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 375         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 376
 377         fix0             = 0.0;
 378         fiy0             = 0.0;
 379         fiz0             = 0.0;
 380
 381         /* Load parameters for i particles */
 382         iq0              = facel*charge[inr+0];
 383         vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 384
 385         /* Start inner kernel loop */
 386         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 387         {
 388             /* Get j neighbor index, and coordinate index */
 389             jnr              = jjnr[jidx];
 390             j_coord_offset   = DIM*jnr;
 391
 392             /* load j atom coordinates */
 393             jx0              = x[j_coord_offset+DIM*0+XX];
 394             jy0              = x[j_coord_offset+DIM*0+YY];
 395             jz0              = x[j_coord_offset+DIM*0+ZZ];
 396
 397             /* Calculate displacement vector */
 398             dx00             = ix0 - jx0;
 399             dy00             = iy0 - jy0;
 400             dz00             = iz0 - jz0;
 401
 402             /* Calculate squared distance and things based on it */
 403             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 404
 405             rinv00           = gmx_invsqrt(rsq00);
 406
 407             rinvsq00         = rinv00*rinv00;
 408
 409             /* Load parameters for j particles */
 410             jq0              = charge[jnr+0];
 411             vdwjidx0         = 2*vdwtype[jnr+0];
 412
 413             /**************************
 414              * CALCULATE INTERACTIONS *
 415              **************************/
 416
 417             if (rsq00<rcutoff2)
 418             {
 419
 420             r00              = rsq00*rinv00;
 421
 422             qq00             = iq0*jq0;
 423             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 424             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 425             c6grid_00        = vdwgridparam[vdwioffset0+vdwjidx0];
 426
 427             /* EWALD ELECTROSTATICS */
 428
 429             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 430             ewrt             = r00*ewtabscale;
 431             ewitab           = ewrt;
 432             eweps            = ewrt-ewitab;
 433             felec            = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1];
 434             felec            = qq00*rinv00*(rinvsq00-felec);
 435
 436             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 437             ewcljrsq         = ewclj2*rsq00;
 438             exponent         = exp(-ewcljrsq);
 439             poly             = exponent*(1.0 + ewcljrsq + ewcljrsq*ewcljrsq*0.5);
 440             fvdw             = (((c12_00*rinvsix - c6_00 + c6grid_00*(1.0-poly))*rinvsix) - c6grid_00*(1.0/6.0)*exponent*ewclj6)*rinvsq00;
 441
 442             fscal            = felec+fvdw;
 443
 444             /* Calculate temporary vectorial force */
 445             tx               = fscal*dx00;
 446             ty               = fscal*dy00;
 447             tz               = fscal*dz00;
 448
 449             /* Update vectorial force */
 450             fix0            += tx;
 451             fiy0            += ty;
 452             fiz0            += tz;
 453             f[j_coord_offset+DIM*0+XX] -= tx;
 454             f[j_coord_offset+DIM*0+YY] -= ty;
 455             f[j_coord_offset+DIM*0+ZZ] -= tz;
 456
 457             }
 458
 459             /* Inner loop uses 55 flops */
 460         }
 461         /* End of innermost loop */
 462
 463         tx = ty = tz = 0;
 464         f[i_coord_offset+DIM*0+XX] += fix0;
 465         f[i_coord_offset+DIM*0+YY] += fiy0;
 466         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 467         tx                         += fix0;
 468         ty                         += fiy0;
 469         tz                         += fiz0;
 470         fshift[i_shift_offset+XX]  += tx;
 471         fshift[i_shift_offset+YY]  += ty;
 472         fshift[i_shift_offset+ZZ]  += tz;
 473
 474         /* Increment number of inner iterations */
 475         inneriter                  += j_index_end - j_index_start;
 476
 477         /* Outer loop uses 13 flops */
 478     }
 479
 480     /* Increment number of outer iterations */
 481     outeriter        += nri;
 482
 483     /* Update outer/inner flops */
 484
 485     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_F,outeriter*13 + inneriter*55);
 486 }