src/gromacs/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecEw_VdwLJ_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 "gmxpre.h"
  39
  40 #include "config.h"
  41
  42 #include <math.h>
  43
  44 #include "../nb_kernel.h"
  45 #include "gromacs/legacyheaders/types/simple.h"
  46 #include "gromacs/math/vec.h"
  47 #include "gromacs/legacyheaders/nrnb.h"
  48
  49 /*
  50  * Gromacs nonbonded kernel:   nb_kernel_ElecEw_VdwLJ_GeomP1P1_VF_c
  51  * Electrostatics interaction: Ewald
  52  * VdW interaction:            LennardJones
  53  * Geometry:                   Particle-Particle
  54  * Calculate force/pot:        PotentialAndForce
  55  */
  56 void
  57 nb_kernel_ElecEw_VdwLJ_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     int              ewitab;
  84     real             ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
  85     real             *ewtab;
  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     nvdwtype         = fr->ntype;
 101     vdwparam         = fr->nbfp;
 102     vdwtype          = mdatoms->typeA;
 103
 104     sh_ewald         = fr->ic->sh_ewald;
 105     ewtab            = fr->ic->tabq_coul_FDV0;
 106     ewtabscale       = fr->ic->tabq_scale;
 107     ewtabhalfspace   = 0.5/ewtabscale;
 108
 109     outeriter        = 0;
 110     inneriter        = 0;
 111
 112     /* Start outer loop over neighborlists */
 113     for(iidx=0; iidx<nri; iidx++)
 114     {
 115         /* Load shift vector for this list */
 116         i_shift_offset   = DIM*shiftidx[iidx];
 117         shX              = shiftvec[i_shift_offset+XX];
 118         shY              = shiftvec[i_shift_offset+YY];
 119         shZ              = shiftvec[i_shift_offset+ZZ];
 120
 121         /* Load limits for loop over neighbors */
 122         j_index_start    = jindex[iidx];
 123         j_index_end      = jindex[iidx+1];
 124
 125         /* Get outer coordinate index */
 126         inr              = iinr[iidx];
 127         i_coord_offset   = DIM*inr;
 128
 129         /* Load i particle coords and add shift vector */
 130         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 131         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 132         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 133
 134         fix0             = 0.0;
 135         fiy0             = 0.0;
 136         fiz0             = 0.0;
 137
 138         /* Load parameters for i particles */
 139         iq0              = facel*charge[inr+0];
 140         vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 141
 142         /* Reset potential sums */
 143         velecsum         = 0.0;
 144         vvdwsum          = 0.0;
 145
 146         /* Start inner kernel loop */
 147         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 148         {
 149             /* Get j neighbor index, and coordinate index */
 150             jnr              = jjnr[jidx];
 151             j_coord_offset   = DIM*jnr;
 152
 153             /* load j atom coordinates */
 154             jx0              = x[j_coord_offset+DIM*0+XX];
 155             jy0              = x[j_coord_offset+DIM*0+YY];
 156             jz0              = x[j_coord_offset+DIM*0+ZZ];
 157
 158             /* Calculate displacement vector */
 159             dx00             = ix0 - jx0;
 160             dy00             = iy0 - jy0;
 161             dz00             = iz0 - jz0;
 162
 163             /* Calculate squared distance and things based on it */
 164             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 165
 166             rinv00           = gmx_invsqrt(rsq00);
 167
 168             rinvsq00         = rinv00*rinv00;
 169
 170             /* Load parameters for j particles */
 171             jq0              = charge[jnr+0];
 172             vdwjidx0         = 2*vdwtype[jnr+0];
 173
 174             /**************************
 175              * CALCULATE INTERACTIONS *
 176              **************************/
 177
 178             r00              = rsq00*rinv00;
 179
 180             qq00             = iq0*jq0;
 181             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 182             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 183
 184             /* EWALD ELECTROSTATICS */
 185
 186             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 187             ewrt             = r00*ewtabscale;
 188             ewitab           = ewrt;
 189             eweps            = ewrt-ewitab;
 190             ewitab           = 4*ewitab;
 191             felec            = ewtab[ewitab]+eweps*ewtab[ewitab+1];
 192             velec            = qq00*(rinv00-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
 193             felec            = qq00*rinv00*(rinvsq00-felec);
 194
 195             /* LENNARD-JONES DISPERSION/REPULSION */
 196
 197             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 198             vvdw6            = c6_00*rinvsix;
 199             vvdw12           = c12_00*rinvsix*rinvsix;
 200             vvdw             = vvdw12*(1.0/12.0) - vvdw6*(1.0/6.0);
 201             fvdw             = (vvdw12-vvdw6)*rinvsq00;
 202
 203             /* Update potential sums from outer loop */
 204             velecsum        += velec;
 205             vvdwsum         += vvdw;
 206
 207             fscal            = felec+fvdw;
 208
 209             /* Calculate temporary vectorial force */
 210             tx               = fscal*dx00;
 211             ty               = fscal*dy00;
 212             tz               = fscal*dz00;
 213
 214             /* Update vectorial force */
 215             fix0            += tx;
 216             fiy0            += ty;
 217             fiz0            += tz;
 218             f[j_coord_offset+DIM*0+XX] -= tx;
 219             f[j_coord_offset+DIM*0+YY] -= ty;
 220             f[j_coord_offset+DIM*0+ZZ] -= tz;
 221
 222             /* Inner loop uses 53 flops */
 223         }
 224         /* End of innermost loop */
 225
 226         tx = ty = tz = 0;
 227         f[i_coord_offset+DIM*0+XX] += fix0;
 228         f[i_coord_offset+DIM*0+YY] += fiy0;
 229         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 230         tx                         += fix0;
 231         ty                         += fiy0;
 232         tz                         += fiz0;
 233         fshift[i_shift_offset+XX]  += tx;
 234         fshift[i_shift_offset+YY]  += ty;
 235         fshift[i_shift_offset+ZZ]  += tz;
 236
 237         ggid                        = gid[iidx];
 238         /* Update potential energies */
 239         kernel_data->energygrp_elec[ggid] += velecsum;
 240         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 241
 242         /* Increment number of inner iterations */
 243         inneriter                  += j_index_end - j_index_start;
 244
 245         /* Outer loop uses 15 flops */
 246     }
 247
 248     /* Increment number of outer iterations */
 249     outeriter        += nri;
 250
 251     /* Update outer/inner flops */
 252
 253     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_VF,outeriter*15 + inneriter*53);
 254 }
 255 /*
 256  * Gromacs nonbonded kernel:   nb_kernel_ElecEw_VdwLJ_GeomP1P1_F_c
 257  * Electrostatics interaction: Ewald
 258  * VdW interaction:            LennardJones
 259  * Geometry:                   Particle-Particle
 260  * Calculate force/pot:        Force
 261  */
 262 void
 263 nb_kernel_ElecEw_VdwLJ_GeomP1P1_F_c
 264                     (t_nblist                    * gmx_restrict       nlist,
 265                      rvec                        * gmx_restrict          xx,
 266                      rvec                        * gmx_restrict          ff,
 267                      t_forcerec                  * gmx_restrict          fr,
 268                      t_mdatoms                   * gmx_restrict     mdatoms,
 269                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 270                      t_nrnb                      * gmx_restrict        nrnb)
 271 {
 272     int              i_shift_offset,i_coord_offset,j_coord_offset;
 273     int              j_index_start,j_index_end;
 274     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 275     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 276     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 277     real             *shiftvec,*fshift,*x,*f;
 278     int              vdwioffset0;
 279     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 280     int              vdwjidx0;
 281     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 282     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 283     real             velec,felec,velecsum,facel,crf,krf,krf2;
 284     real             *charge;
 285     int              nvdwtype;
 286     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 287     int              *vdwtype;
 288     real             *vdwparam;
 289     int              ewitab;
 290     real             ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
 291     real             *ewtab;
 292
 293     x                = xx[0];
 294     f                = ff[0];
 295
 296     nri              = nlist->nri;
 297     iinr             = nlist->iinr;
 298     jindex           = nlist->jindex;
 299     jjnr             = nlist->jjnr;
 300     shiftidx         = nlist->shift;
 301     gid              = nlist->gid;
 302     shiftvec         = fr->shift_vec[0];
 303     fshift           = fr->fshift[0];
 304     facel            = fr->epsfac;
 305     charge           = mdatoms->chargeA;
 306     nvdwtype         = fr->ntype;
 307     vdwparam         = fr->nbfp;
 308     vdwtype          = mdatoms->typeA;
 309
 310     sh_ewald         = fr->ic->sh_ewald;
 311     ewtab            = fr->ic->tabq_coul_F;
 312     ewtabscale       = fr->ic->tabq_scale;
 313     ewtabhalfspace   = 0.5/ewtabscale;
 314
 315     outeriter        = 0;
 316     inneriter        = 0;
 317
 318     /* Start outer loop over neighborlists */
 319     for(iidx=0; iidx<nri; iidx++)
 320     {
 321         /* Load shift vector for this list */
 322         i_shift_offset   = DIM*shiftidx[iidx];
 323         shX              = shiftvec[i_shift_offset+XX];
 324         shY              = shiftvec[i_shift_offset+YY];
 325         shZ              = shiftvec[i_shift_offset+ZZ];
 326
 327         /* Load limits for loop over neighbors */
 328         j_index_start    = jindex[iidx];
 329         j_index_end      = jindex[iidx+1];
 330
 331         /* Get outer coordinate index */
 332         inr              = iinr[iidx];
 333         i_coord_offset   = DIM*inr;
 334
 335         /* Load i particle coords and add shift vector */
 336         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 337         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 338         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 339
 340         fix0             = 0.0;
 341         fiy0             = 0.0;
 342         fiz0             = 0.0;
 343
 344         /* Load parameters for i particles */
 345         iq0              = facel*charge[inr+0];
 346         vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 347
 348         /* Start inner kernel loop */
 349         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 350         {
 351             /* Get j neighbor index, and coordinate index */
 352             jnr              = jjnr[jidx];
 353             j_coord_offset   = DIM*jnr;
 354
 355             /* load j atom coordinates */
 356             jx0              = x[j_coord_offset+DIM*0+XX];
 357             jy0              = x[j_coord_offset+DIM*0+YY];
 358             jz0              = x[j_coord_offset+DIM*0+ZZ];
 359
 360             /* Calculate displacement vector */
 361             dx00             = ix0 - jx0;
 362             dy00             = iy0 - jy0;
 363             dz00             = iz0 - jz0;
 364
 365             /* Calculate squared distance and things based on it */
 366             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 367
 368             rinv00           = gmx_invsqrt(rsq00);
 369
 370             rinvsq00         = rinv00*rinv00;
 371
 372             /* Load parameters for j particles */
 373             jq0              = charge[jnr+0];
 374             vdwjidx0         = 2*vdwtype[jnr+0];
 375
 376             /**************************
 377              * CALCULATE INTERACTIONS *
 378              **************************/
 379
 380             r00              = rsq00*rinv00;
 381
 382             qq00             = iq0*jq0;
 383             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 384             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 385
 386             /* EWALD ELECTROSTATICS */
 387
 388             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 389             ewrt             = r00*ewtabscale;
 390             ewitab           = ewrt;
 391             eweps            = ewrt-ewitab;
 392             felec            = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1];
 393             felec            = qq00*rinv00*(rinvsq00-felec);
 394
 395             /* LENNARD-JONES DISPERSION/REPULSION */
 396
 397             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 398             fvdw             = (c12_00*rinvsix-c6_00)*rinvsix*rinvsq00;
 399
 400             fscal            = felec+fvdw;
 401
 402             /* Calculate temporary vectorial force */
 403             tx               = fscal*dx00;
 404             ty               = fscal*dy00;
 405             tz               = fscal*dz00;
 406
 407             /* Update vectorial force */
 408             fix0            += tx;
 409             fiy0            += ty;
 410             fiz0            += tz;
 411             f[j_coord_offset+DIM*0+XX] -= tx;
 412             f[j_coord_offset+DIM*0+YY] -= ty;
 413             f[j_coord_offset+DIM*0+ZZ] -= tz;
 414
 415             /* Inner loop uses 41 flops */
 416         }
 417         /* End of innermost loop */
 418
 419         tx = ty = tz = 0;
 420         f[i_coord_offset+DIM*0+XX] += fix0;
 421         f[i_coord_offset+DIM*0+YY] += fiy0;
 422         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 423         tx                         += fix0;
 424         ty                         += fiy0;
 425         tz                         += fiz0;
 426         fshift[i_shift_offset+XX]  += tx;
 427         fshift[i_shift_offset+YY]  += ty;
 428         fshift[i_shift_offset+ZZ]  += tz;
 429
 430         /* Increment number of inner iterations */
 431         inneriter                  += j_index_end - j_index_start;
 432
 433         /* Outer loop uses 13 flops */
 434     }
 435
 436     /* Increment number of outer iterations */
 437     outeriter        += nri;
 438
 439     /* Update outer/inner flops */
 440
 441     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_F,outeriter*13 + inneriter*41);
 442 }