src/gromacs/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecRFCut_VdwBhamSw_GeomP1P1_c.c

   1 /*
   2  * This file is part of the GROMACS molecular simulation package.
   3  *
   4  * Copyright (c) 2012,2013,2014.2015,2017, 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/gmxlib/nrnb.h"
  46
  47 /*
  48  * Gromacs nonbonded kernel:   nb_kernel_ElecRFCut_VdwBhamSw_GeomP1P1_VF_c
  49  * Electrostatics interaction: ReactionField
  50  * VdW interaction:            Buckingham
  51  * Geometry:                   Particle-Particle
  52  * Calculate force/pot:        PotentialAndForce
  53  */
  54 void
  55 nb_kernel_ElecRFCut_VdwBhamSw_GeomP1P1_VF_c
  56                     (t_nblist                    * gmx_restrict       nlist,
  57                      rvec                        * gmx_restrict          xx,
  58                      rvec                        * gmx_restrict          ff,
  59                      struct 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             rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
  82
  83     x                = xx[0];
  84     f                = ff[0];
  85
  86     nri              = nlist->nri;
  87     iinr             = nlist->iinr;
  88     jindex           = nlist->jindex;
  89     jjnr             = nlist->jjnr;
  90     shiftidx         = nlist->shift;
  91     gid              = nlist->gid;
  92     shiftvec         = fr->shift_vec[0];
  93     fshift           = fr->fshift[0];
  94     facel            = fr->ic->epsfac;
  95     charge           = mdatoms->chargeA;
  96     krf              = fr->ic->k_rf;
  97     krf2             = krf*2.0;
  98     crf              = fr->ic->c_rf;
  99     nvdwtype         = fr->ntype;
 100     vdwparam         = fr->nbfp;
 101     vdwtype          = mdatoms->typeA;
 102
 103     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 104     rcutoff          = fr->ic->rcoulomb;
 105     rcutoff2         = rcutoff*rcutoff;
 106
 107     rswitch          = fr->ic->rvdw_switch;
 108     /* Setup switch parameters */
 109     d                = rcutoff-rswitch;
 110     swV3             = -10.0/(d*d*d);
 111     swV4             =  15.0/(d*d*d*d);
 112     swV5             =  -6.0/(d*d*d*d*d);
 113     swF2             = -30.0/(d*d*d);
 114     swF3             =  60.0/(d*d*d*d);
 115     swF4             = -30.0/(d*d*d*d*d);
 116
 117     outeriter        = 0;
 118     inneriter        = 0;
 119
 120     /* Start outer loop over neighborlists */
 121     for(iidx=0; iidx<nri; iidx++)
 122     {
 123         /* Load shift vector for this list */
 124         i_shift_offset   = DIM*shiftidx[iidx];
 125         shX              = shiftvec[i_shift_offset+XX];
 126         shY              = shiftvec[i_shift_offset+YY];
 127         shZ              = shiftvec[i_shift_offset+ZZ];
 128
 129         /* Load limits for loop over neighbors */
 130         j_index_start    = jindex[iidx];
 131         j_index_end      = jindex[iidx+1];
 132
 133         /* Get outer coordinate index */
 134         inr              = iinr[iidx];
 135         i_coord_offset   = DIM*inr;
 136
 137         /* Load i particle coords and add shift vector */
 138         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 139         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 140         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 141
 142         fix0             = 0.0;
 143         fiy0             = 0.0;
 144         fiz0             = 0.0;
 145
 146         /* Load parameters for i particles */
 147         iq0              = facel*charge[inr+0];
 148         vdwioffset0      = 3*nvdwtype*vdwtype[inr+0];
 149
 150         /* Reset potential sums */
 151         velecsum         = 0.0;
 152         vvdwsum          = 0.0;
 153
 154         /* Start inner kernel loop */
 155         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 156         {
 157             /* Get j neighbor index, and coordinate index */
 158             jnr              = jjnr[jidx];
 159             j_coord_offset   = DIM*jnr;
 160
 161             /* load j atom coordinates */
 162             jx0              = x[j_coord_offset+DIM*0+XX];
 163             jy0              = x[j_coord_offset+DIM*0+YY];
 164             jz0              = x[j_coord_offset+DIM*0+ZZ];
 165
 166             /* Calculate displacement vector */
 167             dx00             = ix0 - jx0;
 168             dy00             = iy0 - jy0;
 169             dz00             = iz0 - jz0;
 170
 171             /* Calculate squared distance and things based on it */
 172             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 173
 174             rinv00           = 1.0/sqrt(rsq00);
 175
 176             rinvsq00         = rinv00*rinv00;
 177
 178             /* Load parameters for j particles */
 179             jq0              = charge[jnr+0];
 180             vdwjidx0         = 3*vdwtype[jnr+0];
 181
 182             /**************************
 183              * CALCULATE INTERACTIONS *
 184              **************************/
 185
 186             if (rsq00<rcutoff2)
 187             {
 188
 189             r00              = rsq00*rinv00;
 190
 191             qq00             = iq0*jq0;
 192             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 193             cexp1_00         = vdwparam[vdwioffset0+vdwjidx0+1];
 194             cexp2_00         = vdwparam[vdwioffset0+vdwjidx0+2];
 195
 196             /* REACTION-FIELD ELECTROSTATICS */
 197             velec            = qq00*(rinv00+krf*rsq00-crf);
 198             felec            = qq00*(rinv00*rinvsq00-krf2);
 199
 200             /* BUCKINGHAM DISPERSION/REPULSION */
 201             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 202             vvdw6            = c6_00*rinvsix;
 203             br               = cexp2_00*r00;
 204             vvdwexp          = cexp1_00*exp(-br);
 205             vvdw             = vvdwexp - vvdw6*(1.0/6.0);
 206             fvdw             = (br*vvdwexp-vvdw6)*rinvsq00;
 207
 208             d                = r00-rswitch;
 209             d                = (d>0.0) ? d : 0.0;
 210             d2               = d*d;
 211             sw               = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
 212
 213             dsw              = d2*(swF2+d*(swF3+d*swF4));
 214
 215             /* Evaluate switch function */
 216             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 217             fvdw             = fvdw*sw - rinv00*vvdw*dsw;
 218             vvdw            *= sw;
 219
 220             /* Update potential sums from outer loop */
 221             velecsum        += velec;
 222             vvdwsum         += vvdw;
 223
 224             fscal            = felec+fvdw;
 225
 226             /* Calculate temporary vectorial force */
 227             tx               = fscal*dx00;
 228             ty               = fscal*dy00;
 229             tz               = fscal*dz00;
 230
 231             /* Update vectorial force */
 232             fix0            += tx;
 233             fiy0            += ty;
 234             fiz0            += tz;
 235             f[j_coord_offset+DIM*0+XX] -= tx;
 236             f[j_coord_offset+DIM*0+YY] -= ty;
 237             f[j_coord_offset+DIM*0+ZZ] -= tz;
 238
 239             }
 240
 241             /* Inner loop uses 89 flops */
 242         }
 243         /* End of innermost loop */
 244
 245         tx = ty = tz = 0;
 246         f[i_coord_offset+DIM*0+XX] += fix0;
 247         f[i_coord_offset+DIM*0+YY] += fiy0;
 248         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 249         tx                         += fix0;
 250         ty                         += fiy0;
 251         tz                         += fiz0;
 252         fshift[i_shift_offset+XX]  += tx;
 253         fshift[i_shift_offset+YY]  += ty;
 254         fshift[i_shift_offset+ZZ]  += tz;
 255
 256         ggid                        = gid[iidx];
 257         /* Update potential energies */
 258         kernel_data->energygrp_elec[ggid] += velecsum;
 259         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 260
 261         /* Increment number of inner iterations */
 262         inneriter                  += j_index_end - j_index_start;
 263
 264         /* Outer loop uses 15 flops */
 265     }
 266
 267     /* Increment number of outer iterations */
 268     outeriter        += nri;
 269
 270     /* Update outer/inner flops */
 271
 272     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_VF,outeriter*15 + inneriter*89);
 273 }
 274 /*
 275  * Gromacs nonbonded kernel:   nb_kernel_ElecRFCut_VdwBhamSw_GeomP1P1_F_c
 276  * Electrostatics interaction: ReactionField
 277  * VdW interaction:            Buckingham
 278  * Geometry:                   Particle-Particle
 279  * Calculate force/pot:        Force
 280  */
 281 void
 282 nb_kernel_ElecRFCut_VdwBhamSw_GeomP1P1_F_c
 283                     (t_nblist                    * gmx_restrict       nlist,
 284                      rvec                        * gmx_restrict          xx,
 285                      rvec                        * gmx_restrict          ff,
 286                      struct t_forcerec           * gmx_restrict          fr,
 287                      t_mdatoms                   * gmx_restrict     mdatoms,
 288                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 289                      t_nrnb                      * gmx_restrict        nrnb)
 290 {
 291     int              i_shift_offset,i_coord_offset,j_coord_offset;
 292     int              j_index_start,j_index_end;
 293     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 294     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 295     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 296     real             *shiftvec,*fshift,*x,*f;
 297     int              vdwioffset0;
 298     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 299     int              vdwjidx0;
 300     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 301     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 302     real             velec,felec,velecsum,facel,crf,krf,krf2;
 303     real             *charge;
 304     int              nvdwtype;
 305     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 306     int              *vdwtype;
 307     real             *vdwparam;
 308     real             rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
 309
 310     x                = xx[0];
 311     f                = ff[0];
 312
 313     nri              = nlist->nri;
 314     iinr             = nlist->iinr;
 315     jindex           = nlist->jindex;
 316     jjnr             = nlist->jjnr;
 317     shiftidx         = nlist->shift;
 318     gid              = nlist->gid;
 319     shiftvec         = fr->shift_vec[0];
 320     fshift           = fr->fshift[0];
 321     facel            = fr->ic->epsfac;
 322     charge           = mdatoms->chargeA;
 323     krf              = fr->ic->k_rf;
 324     krf2             = krf*2.0;
 325     crf              = fr->ic->c_rf;
 326     nvdwtype         = fr->ntype;
 327     vdwparam         = fr->nbfp;
 328     vdwtype          = mdatoms->typeA;
 329
 330     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 331     rcutoff          = fr->ic->rcoulomb;
 332     rcutoff2         = rcutoff*rcutoff;
 333
 334     rswitch          = fr->ic->rvdw_switch;
 335     /* Setup switch parameters */
 336     d                = rcutoff-rswitch;
 337     swV3             = -10.0/(d*d*d);
 338     swV4             =  15.0/(d*d*d*d);
 339     swV5             =  -6.0/(d*d*d*d*d);
 340     swF2             = -30.0/(d*d*d);
 341     swF3             =  60.0/(d*d*d*d);
 342     swF4             = -30.0/(d*d*d*d*d);
 343
 344     outeriter        = 0;
 345     inneriter        = 0;
 346
 347     /* Start outer loop over neighborlists */
 348     for(iidx=0; iidx<nri; iidx++)
 349     {
 350         /* Load shift vector for this list */
 351         i_shift_offset   = DIM*shiftidx[iidx];
 352         shX              = shiftvec[i_shift_offset+XX];
 353         shY              = shiftvec[i_shift_offset+YY];
 354         shZ              = shiftvec[i_shift_offset+ZZ];
 355
 356         /* Load limits for loop over neighbors */
 357         j_index_start    = jindex[iidx];
 358         j_index_end      = jindex[iidx+1];
 359
 360         /* Get outer coordinate index */
 361         inr              = iinr[iidx];
 362         i_coord_offset   = DIM*inr;
 363
 364         /* Load i particle coords and add shift vector */
 365         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 366         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 367         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 368
 369         fix0             = 0.0;
 370         fiy0             = 0.0;
 371         fiz0             = 0.0;
 372
 373         /* Load parameters for i particles */
 374         iq0              = facel*charge[inr+0];
 375         vdwioffset0      = 3*nvdwtype*vdwtype[inr+0];
 376
 377         /* Start inner kernel loop */
 378         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 379         {
 380             /* Get j neighbor index, and coordinate index */
 381             jnr              = jjnr[jidx];
 382             j_coord_offset   = DIM*jnr;
 383
 384             /* load j atom coordinates */
 385             jx0              = x[j_coord_offset+DIM*0+XX];
 386             jy0              = x[j_coord_offset+DIM*0+YY];
 387             jz0              = x[j_coord_offset+DIM*0+ZZ];
 388
 389             /* Calculate displacement vector */
 390             dx00             = ix0 - jx0;
 391             dy00             = iy0 - jy0;
 392             dz00             = iz0 - jz0;
 393
 394             /* Calculate squared distance and things based on it */
 395             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 396
 397             rinv00           = 1.0/sqrt(rsq00);
 398
 399             rinvsq00         = rinv00*rinv00;
 400
 401             /* Load parameters for j particles */
 402             jq0              = charge[jnr+0];
 403             vdwjidx0         = 3*vdwtype[jnr+0];
 404
 405             /**************************
 406              * CALCULATE INTERACTIONS *
 407              **************************/
 408
 409             if (rsq00<rcutoff2)
 410             {
 411
 412             r00              = rsq00*rinv00;
 413
 414             qq00             = iq0*jq0;
 415             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 416             cexp1_00         = vdwparam[vdwioffset0+vdwjidx0+1];
 417             cexp2_00         = vdwparam[vdwioffset0+vdwjidx0+2];
 418
 419             /* REACTION-FIELD ELECTROSTATICS */
 420             felec            = qq00*(rinv00*rinvsq00-krf2);
 421
 422             /* BUCKINGHAM DISPERSION/REPULSION */
 423             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 424             vvdw6            = c6_00*rinvsix;
 425             br               = cexp2_00*r00;
 426             vvdwexp          = cexp1_00*exp(-br);
 427             vvdw             = vvdwexp - vvdw6*(1.0/6.0);
 428             fvdw             = (br*vvdwexp-vvdw6)*rinvsq00;
 429
 430             d                = r00-rswitch;
 431             d                = (d>0.0) ? d : 0.0;
 432             d2               = d*d;
 433             sw               = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
 434
 435             dsw              = d2*(swF2+d*(swF3+d*swF4));
 436
 437             /* Evaluate switch function */
 438             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 439             fvdw             = fvdw*sw - rinv00*vvdw*dsw;
 440
 441             fscal            = felec+fvdw;
 442
 443             /* Calculate temporary vectorial force */
 444             tx               = fscal*dx00;
 445             ty               = fscal*dy00;
 446             tz               = fscal*dz00;
 447
 448             /* Update vectorial force */
 449             fix0            += tx;
 450             fiy0            += ty;
 451             fiz0            += tz;
 452             f[j_coord_offset+DIM*0+XX] -= tx;
 453             f[j_coord_offset+DIM*0+YY] -= ty;
 454             f[j_coord_offset+DIM*0+ZZ] -= tz;
 455
 456             }
 457
 458             /* Inner loop uses 82 flops */
 459         }
 460         /* End of innermost loop */
 461
 462         tx = ty = tz = 0;
 463         f[i_coord_offset+DIM*0+XX] += fix0;
 464         f[i_coord_offset+DIM*0+YY] += fiy0;
 465         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 466         tx                         += fix0;
 467         ty                         += fiy0;
 468         tz                         += fiz0;
 469         fshift[i_shift_offset+XX]  += tx;
 470         fshift[i_shift_offset+YY]  += ty;
 471         fshift[i_shift_offset+ZZ]  += tz;
 472
 473         /* Increment number of inner iterations */
 474         inneriter                  += j_index_end - j_index_start;
 475
 476         /* Outer loop uses 13 flops */
 477     }
 478
 479     /* Increment number of outer iterations */
 480     outeriter        += nri;
 481
 482     /* Update outer/inner flops */
 483
 484     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_F,outeriter*13 + inneriter*82);
 485 }