src/gromacs/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecRFCut_VdwLJSw_GeomP1P1_c.c

   1 /*
   2  * This file is part of the GROMACS molecular simulation package.
   3  *
   4  * Copyright (c) 2012,2013, 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_ElecRFCut_VdwLJSw_GeomP1P1_VF_c
  51  * Electrostatics interaction: ReactionField
  52  * VdW interaction:            LennardJones
  53  * Geometry:                   Particle-Particle
  54  * Calculate force/pot:        PotentialAndForce
  55  */
  56 void
  57 nb_kernel_ElecRFCut_VdwLJSw_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      = 2*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         = 2*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             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 196
 197             /* REACTION-FIELD ELECTROSTATICS */
 198             velec            = qq00*(rinv00+krf*rsq00-crf);
 199             felec            = qq00*(rinv00*rinvsq00-krf2);
 200
 201             /* LENNARD-JONES DISPERSION/REPULSION */
 202
 203             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 204             vvdw6            = c6_00*rinvsix;
 205             vvdw12           = c12_00*rinvsix*rinvsix;
 206             vvdw             = vvdw12*(1.0/12.0) - vvdw6*(1.0/6.0);
 207             fvdw             = (vvdw12-vvdw6)*rinvsq00;
 208
 209             d                = r00-rswitch;
 210             d                = (d>0.0) ? d : 0.0;
 211             d2               = d*d;
 212             sw               = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
 213
 214             dsw              = d2*(swF2+d*(swF3+d*swF4));
 215
 216             /* Evaluate switch function */
 217             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 218             fvdw             = fvdw*sw - rinv00*vvdw*dsw;
 219             vvdw            *= sw;
 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 63 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*63);
 274 }
 275 /*
 276  * Gromacs nonbonded kernel:   nb_kernel_ElecRFCut_VdwLJSw_GeomP1P1_F_c
 277  * Electrostatics interaction: ReactionField
 278  * VdW interaction:            LennardJones
 279  * Geometry:                   Particle-Particle
 280  * Calculate force/pot:        Force
 281  */
 282 void
 283 nb_kernel_ElecRFCut_VdwLJSw_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             rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
 310
 311     x                = xx[0];
 312     f                = ff[0];
 313
 314     nri              = nlist->nri;
 315     iinr             = nlist->iinr;
 316     jindex           = nlist->jindex;
 317     jjnr             = nlist->jjnr;
 318     shiftidx         = nlist->shift;
 319     gid              = nlist->gid;
 320     shiftvec         = fr->shift_vec[0];
 321     fshift           = fr->fshift[0];
 322     facel            = fr->epsfac;
 323     charge           = mdatoms->chargeA;
 324     krf              = fr->ic->k_rf;
 325     krf2             = krf*2.0;
 326     crf              = fr->ic->c_rf;
 327     nvdwtype         = fr->ntype;
 328     vdwparam         = fr->nbfp;
 329     vdwtype          = mdatoms->typeA;
 330
 331     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 332     rcutoff          = fr->rcoulomb;
 333     rcutoff2         = rcutoff*rcutoff;
 334
 335     rswitch          = fr->rvdw_switch;
 336     /* Setup switch parameters */
 337     d                = rcutoff-rswitch;
 338     swV3             = -10.0/(d*d*d);
 339     swV4             =  15.0/(d*d*d*d);
 340     swV5             =  -6.0/(d*d*d*d*d);
 341     swF2             = -30.0/(d*d*d);
 342     swF3             =  60.0/(d*d*d*d);
 343     swF4             = -30.0/(d*d*d*d*d);
 344
 345     outeriter        = 0;
 346     inneriter        = 0;
 347
 348     /* Start outer loop over neighborlists */
 349     for(iidx=0; iidx<nri; iidx++)
 350     {
 351         /* Load shift vector for this list */
 352         i_shift_offset   = DIM*shiftidx[iidx];
 353         shX              = shiftvec[i_shift_offset+XX];
 354         shY              = shiftvec[i_shift_offset+YY];
 355         shZ              = shiftvec[i_shift_offset+ZZ];
 356
 357         /* Load limits for loop over neighbors */
 358         j_index_start    = jindex[iidx];
 359         j_index_end      = jindex[iidx+1];
 360
 361         /* Get outer coordinate index */
 362         inr              = iinr[iidx];
 363         i_coord_offset   = DIM*inr;
 364
 365         /* Load i particle coords and add shift vector */
 366         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 367         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 368         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 369
 370         fix0             = 0.0;
 371         fiy0             = 0.0;
 372         fiz0             = 0.0;
 373
 374         /* Load parameters for i particles */
 375         iq0              = facel*charge[inr+0];
 376         vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 377
 378         /* Start inner kernel loop */
 379         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 380         {
 381             /* Get j neighbor index, and coordinate index */
 382             jnr              = jjnr[jidx];
 383             j_coord_offset   = DIM*jnr;
 384
 385             /* load j atom coordinates */
 386             jx0              = x[j_coord_offset+DIM*0+XX];
 387             jy0              = x[j_coord_offset+DIM*0+YY];
 388             jz0              = x[j_coord_offset+DIM*0+ZZ];
 389
 390             /* Calculate displacement vector */
 391             dx00             = ix0 - jx0;
 392             dy00             = iy0 - jy0;
 393             dz00             = iz0 - jz0;
 394
 395             /* Calculate squared distance and things based on it */
 396             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 397
 398             rinv00           = gmx_invsqrt(rsq00);
 399
 400             rinvsq00         = rinv00*rinv00;
 401
 402             /* Load parameters for j particles */
 403             jq0              = charge[jnr+0];
 404             vdwjidx0         = 2*vdwtype[jnr+0];
 405
 406             /**************************
 407              * CALCULATE INTERACTIONS *
 408              **************************/
 409
 410             if (rsq00<rcutoff2)
 411             {
 412
 413             r00              = rsq00*rinv00;
 414
 415             qq00             = iq0*jq0;
 416             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 417             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 418
 419             /* REACTION-FIELD ELECTROSTATICS */
 420             felec            = qq00*(rinv00*rinvsq00-krf2);
 421
 422             /* LENNARD-JONES DISPERSION/REPULSION */
 423
 424             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 425             vvdw6            = c6_00*rinvsix;
 426             vvdw12           = c12_00*rinvsix*rinvsix;
 427             vvdw             = vvdw12*(1.0/12.0) - vvdw6*(1.0/6.0);
 428             fvdw             = (vvdw12-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 56 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*56);
 485 }