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, 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_ElecRFCut_VdwBhamSw_GeomP1P1_VF_c
  51  * Electrostatics interaction: ReactionField
  52  * VdW interaction:            Buckingham
  53  * Geometry:                   Particle-Particle
  54  * Calculate force/pot:        PotentialAndForce
  55  */
  56 void
  57 nb_kernel_ElecRFCut_VdwBhamSw_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             rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
  84
  85     x                = xx[0];
  86     f                = ff[0];
  87
  88     nri              = nlist->nri;
  89     iinr             = nlist->iinr;
  90     jindex           = nlist->jindex;
  91     jjnr             = nlist->jjnr;
  92     shiftidx         = nlist->shift;
  93     gid              = nlist->gid;
  94     shiftvec         = fr->shift_vec[0];
  95     fshift           = fr->fshift[0];
  96     facel            = fr->epsfac;
  97     charge           = mdatoms->chargeA;
  98     krf              = fr->ic->k_rf;
  99     krf2             = krf*2.0;
 100     crf              = fr->ic->c_rf;
 101     nvdwtype         = fr->ntype;
 102     vdwparam         = fr->nbfp;
 103     vdwtype          = mdatoms->typeA;
 104
 105     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 106     rcutoff          = fr->rcoulomb;
 107     rcutoff2         = rcutoff*rcutoff;
 108
 109     rswitch          = fr->rvdw_switch;
 110     /* Setup switch parameters */
 111     d                = rcutoff-rswitch;
 112     swV3             = -10.0/(d*d*d);
 113     swV4             =  15.0/(d*d*d*d);
 114     swV5             =  -6.0/(d*d*d*d*d);
 115     swF2             = -30.0/(d*d*d);
 116     swF3             =  60.0/(d*d*d*d);
 117     swF4             = -30.0/(d*d*d*d*d);
 118
 119     outeriter        = 0;
 120     inneriter        = 0;
 121
 122     /* Start outer loop over neighborlists */
 123     for(iidx=0; iidx<nri; iidx++)
 124     {
 125         /* Load shift vector for this list */
 126         i_shift_offset   = DIM*shiftidx[iidx];
 127         shX              = shiftvec[i_shift_offset+XX];
 128         shY              = shiftvec[i_shift_offset+YY];
 129         shZ              = shiftvec[i_shift_offset+ZZ];
 130
 131         /* Load limits for loop over neighbors */
 132         j_index_start    = jindex[iidx];
 133         j_index_end      = jindex[iidx+1];
 134
 135         /* Get outer coordinate index */
 136         inr              = iinr[iidx];
 137         i_coord_offset   = DIM*inr;
 138
 139         /* Load i particle coords and add shift vector */
 140         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 141         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 142         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 143
 144         fix0             = 0.0;
 145         fiy0             = 0.0;
 146         fiz0             = 0.0;
 147
 148         /* Load parameters for i particles */
 149         iq0              = facel*charge[inr+0];
 150         vdwioffset0      = 3*nvdwtype*vdwtype[inr+0];
 151
 152         /* Reset potential sums */
 153         velecsum         = 0.0;
 154         vvdwsum          = 0.0;
 155
 156         /* Start inner kernel loop */
 157         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 158         {
 159             /* Get j neighbor index, and coordinate index */
 160             jnr              = jjnr[jidx];
 161             j_coord_offset   = DIM*jnr;
 162
 163             /* load j atom coordinates */
 164             jx0              = x[j_coord_offset+DIM*0+XX];
 165             jy0              = x[j_coord_offset+DIM*0+YY];
 166             jz0              = x[j_coord_offset+DIM*0+ZZ];
 167
 168             /* Calculate displacement vector */
 169             dx00             = ix0 - jx0;
 170             dy00             = iy0 - jy0;
 171             dz00             = iz0 - jz0;
 172
 173             /* Calculate squared distance and things based on it */
 174             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 175
 176             rinv00           = gmx_invsqrt(rsq00);
 177
 178             rinvsq00         = rinv00*rinv00;
 179
 180             /* Load parameters for j particles */
 181             jq0              = charge[jnr+0];
 182             vdwjidx0         = 3*vdwtype[jnr+0];
 183
 184             /**************************
 185              * CALCULATE INTERACTIONS *
 186              **************************/
 187
 188             if (rsq00<rcutoff2)
 189             {
 190
 191             r00              = rsq00*rinv00;
 192
 193             qq00             = iq0*jq0;
 194             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 195             cexp1_00         = vdwparam[vdwioffset0+vdwjidx0+1];
 196             cexp2_00         = vdwparam[vdwioffset0+vdwjidx0+2];
 197
 198             /* REACTION-FIELD ELECTROSTATICS */
 199             velec            = qq00*(rinv00+krf*rsq00-crf);
 200             felec            = qq00*(rinv00*rinvsq00-krf2);
 201
 202             /* BUCKINGHAM DISPERSION/REPULSION */
 203             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 204             vvdw6            = c6_00*rinvsix;
 205             br               = cexp2_00*r00;
 206             vvdwexp          = cexp1_00*exp(-br);
 207             vvdw             = vvdwexp - vvdw6*(1.0/6.0);
 208             fvdw             = (br*vvdwexp-vvdw6)*rinvsq00;
 209
 210             d                = r00-rswitch;
 211             d                = (d>0.0) ? d : 0.0;
 212             d2               = d*d;
 213             sw               = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
 214
 215             dsw              = d2*(swF2+d*(swF3+d*swF4));
 216
 217             /* Evaluate switch function */
 218             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 219             fvdw             = fvdw*sw - rinv00*vvdw*dsw;
 220             vvdw            *= sw;
 221
 222             /* Update potential sums from outer loop */
 223             velecsum        += velec;
 224             vvdwsum         += vvdw;
 225
 226             fscal            = felec+fvdw;
 227
 228             /* Calculate temporary vectorial force */
 229             tx               = fscal*dx00;
 230             ty               = fscal*dy00;
 231             tz               = fscal*dz00;
 232
 233             /* Update vectorial force */
 234             fix0            += tx;
 235             fiy0            += ty;
 236             fiz0            += tz;
 237             f[j_coord_offset+DIM*0+XX] -= tx;
 238             f[j_coord_offset+DIM*0+YY] -= ty;
 239             f[j_coord_offset+DIM*0+ZZ] -= tz;
 240
 241             }
 242
 243             /* Inner loop uses 89 flops */
 244         }
 245         /* End of innermost loop */
 246
 247         tx = ty = tz = 0;
 248         f[i_coord_offset+DIM*0+XX] += fix0;
 249         f[i_coord_offset+DIM*0+YY] += fiy0;
 250         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 251         tx                         += fix0;
 252         ty                         += fiy0;
 253         tz                         += fiz0;
 254         fshift[i_shift_offset+XX]  += tx;
 255         fshift[i_shift_offset+YY]  += ty;
 256         fshift[i_shift_offset+ZZ]  += tz;
 257
 258         ggid                        = gid[iidx];
 259         /* Update potential energies */
 260         kernel_data->energygrp_elec[ggid] += velecsum;
 261         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 262
 263         /* Increment number of inner iterations */
 264         inneriter                  += j_index_end - j_index_start;
 265
 266         /* Outer loop uses 15 flops */
 267     }
 268
 269     /* Increment number of outer iterations */
 270     outeriter        += nri;
 271
 272     /* Update outer/inner flops */
 273
 274     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_VF,outeriter*15 + inneriter*89);
 275 }
 276 /*
 277  * Gromacs nonbonded kernel:   nb_kernel_ElecRFCut_VdwBhamSw_GeomP1P1_F_c
 278  * Electrostatics interaction: ReactionField
 279  * VdW interaction:            Buckingham
 280  * Geometry:                   Particle-Particle
 281  * Calculate force/pot:        Force
 282  */
 283 void
 284 nb_kernel_ElecRFCut_VdwBhamSw_GeomP1P1_F_c
 285                     (t_nblist                    * gmx_restrict       nlist,
 286                      rvec                        * gmx_restrict          xx,
 287                      rvec                        * gmx_restrict          ff,
 288                      t_forcerec                  * gmx_restrict          fr,
 289                      t_mdatoms                   * gmx_restrict     mdatoms,
 290                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 291                      t_nrnb                      * gmx_restrict        nrnb)
 292 {
 293     int              i_shift_offset,i_coord_offset,j_coord_offset;
 294     int              j_index_start,j_index_end;
 295     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 296     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 297     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 298     real             *shiftvec,*fshift,*x,*f;
 299     int              vdwioffset0;
 300     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 301     int              vdwjidx0;
 302     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 303     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 304     real             velec,felec,velecsum,facel,crf,krf,krf2;
 305     real             *charge;
 306     int              nvdwtype;
 307     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 308     int              *vdwtype;
 309     real             *vdwparam;
 310     real             rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
 311
 312     x                = xx[0];
 313     f                = ff[0];
 314
 315     nri              = nlist->nri;
 316     iinr             = nlist->iinr;
 317     jindex           = nlist->jindex;
 318     jjnr             = nlist->jjnr;
 319     shiftidx         = nlist->shift;
 320     gid              = nlist->gid;
 321     shiftvec         = fr->shift_vec[0];
 322     fshift           = fr->fshift[0];
 323     facel            = fr->epsfac;
 324     charge           = mdatoms->chargeA;
 325     krf              = fr->ic->k_rf;
 326     krf2             = krf*2.0;
 327     crf              = fr->ic->c_rf;
 328     nvdwtype         = fr->ntype;
 329     vdwparam         = fr->nbfp;
 330     vdwtype          = mdatoms->typeA;
 331
 332     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 333     rcutoff          = fr->rcoulomb;
 334     rcutoff2         = rcutoff*rcutoff;
 335
 336     rswitch          = fr->rvdw_switch;
 337     /* Setup switch parameters */
 338     d                = rcutoff-rswitch;
 339     swV3             = -10.0/(d*d*d);
 340     swV4             =  15.0/(d*d*d*d);
 341     swV5             =  -6.0/(d*d*d*d*d);
 342     swF2             = -30.0/(d*d*d);
 343     swF3             =  60.0/(d*d*d*d);
 344     swF4             = -30.0/(d*d*d*d*d);
 345
 346     outeriter        = 0;
 347     inneriter        = 0;
 348
 349     /* Start outer loop over neighborlists */
 350     for(iidx=0; iidx<nri; iidx++)
 351     {
 352         /* Load shift vector for this list */
 353         i_shift_offset   = DIM*shiftidx[iidx];
 354         shX              = shiftvec[i_shift_offset+XX];
 355         shY              = shiftvec[i_shift_offset+YY];
 356         shZ              = shiftvec[i_shift_offset+ZZ];
 357
 358         /* Load limits for loop over neighbors */
 359         j_index_start    = jindex[iidx];
 360         j_index_end      = jindex[iidx+1];
 361
 362         /* Get outer coordinate index */
 363         inr              = iinr[iidx];
 364         i_coord_offset   = DIM*inr;
 365
 366         /* Load i particle coords and add shift vector */
 367         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 368         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 369         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 370
 371         fix0             = 0.0;
 372         fiy0             = 0.0;
 373         fiz0             = 0.0;
 374
 375         /* Load parameters for i particles */
 376         iq0              = facel*charge[inr+0];
 377         vdwioffset0      = 3*nvdwtype*vdwtype[inr+0];
 378
 379         /* Start inner kernel loop */
 380         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 381         {
 382             /* Get j neighbor index, and coordinate index */
 383             jnr              = jjnr[jidx];
 384             j_coord_offset   = DIM*jnr;
 385
 386             /* load j atom coordinates */
 387             jx0              = x[j_coord_offset+DIM*0+XX];
 388             jy0              = x[j_coord_offset+DIM*0+YY];
 389             jz0              = x[j_coord_offset+DIM*0+ZZ];
 390
 391             /* Calculate displacement vector */
 392             dx00             = ix0 - jx0;
 393             dy00             = iy0 - jy0;
 394             dz00             = iz0 - jz0;
 395
 396             /* Calculate squared distance and things based on it */
 397             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 398
 399             rinv00           = gmx_invsqrt(rsq00);
 400
 401             rinvsq00         = rinv00*rinv00;
 402
 403             /* Load parameters for j particles */
 404             jq0              = charge[jnr+0];
 405             vdwjidx0         = 3*vdwtype[jnr+0];
 406
 407             /**************************
 408              * CALCULATE INTERACTIONS *
 409              **************************/
 410
 411             if (rsq00<rcutoff2)
 412             {
 413
 414             r00              = rsq00*rinv00;
 415
 416             qq00             = iq0*jq0;
 417             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 418             cexp1_00         = vdwparam[vdwioffset0+vdwjidx0+1];
 419             cexp2_00         = vdwparam[vdwioffset0+vdwjidx0+2];
 420
 421             /* REACTION-FIELD ELECTROSTATICS */
 422             felec            = qq00*(rinv00*rinvsq00-krf2);
 423
 424             /* BUCKINGHAM DISPERSION/REPULSION */
 425             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 426             vvdw6            = c6_00*rinvsix;
 427             br               = cexp2_00*r00;
 428             vvdwexp          = cexp1_00*exp(-br);
 429             vvdw             = vvdwexp - vvdw6*(1.0/6.0);
 430             fvdw             = (br*vvdwexp-vvdw6)*rinvsq00;
 431
 432             d                = r00-rswitch;
 433             d                = (d>0.0) ? d : 0.0;
 434             d2               = d*d;
 435             sw               = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
 436
 437             dsw              = d2*(swF2+d*(swF3+d*swF4));
 438
 439             /* Evaluate switch function */
 440             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 441             fvdw             = fvdw*sw - rinv00*vvdw*dsw;
 442
 443             fscal            = felec+fvdw;
 444
 445             /* Calculate temporary vectorial force */
 446             tx               = fscal*dx00;
 447             ty               = fscal*dy00;
 448             tz               = fscal*dz00;
 449
 450             /* Update vectorial force */
 451             fix0            += tx;
 452             fiy0            += ty;
 453             fiz0            += tz;
 454             f[j_coord_offset+DIM*0+XX] -= tx;
 455             f[j_coord_offset+DIM*0+YY] -= ty;
 456             f[j_coord_offset+DIM*0+ZZ] -= tz;
 457
 458             }
 459
 460             /* Inner loop uses 82 flops */
 461         }
 462         /* End of innermost loop */
 463
 464         tx = ty = tz = 0;
 465         f[i_coord_offset+DIM*0+XX] += fix0;
 466         f[i_coord_offset+DIM*0+YY] += fiy0;
 467         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 468         tx                         += fix0;
 469         ty                         += fiy0;
 470         tz                         += fiz0;
 471         fshift[i_shift_offset+XX]  += tx;
 472         fshift[i_shift_offset+YY]  += ty;
 473         fshift[i_shift_offset+ZZ]  += tz;
 474
 475         /* Increment number of inner iterations */
 476         inneriter                  += j_index_end - j_index_start;
 477
 478         /* Outer loop uses 13 flops */
 479     }
 480
 481     /* Increment number of outer iterations */
 482     outeriter        += nri;
 483
 484     /* Update outer/inner flops */
 485
 486     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_F,outeriter*13 + inneriter*82);
 487 }