src/gromacs/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecEwSw_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_ElecEwSw_VdwBhamSw_GeomP1P1_VF_c
  51  * Electrostatics interaction: Ewald
  52  * VdW interaction:            Buckingham
  53  * Geometry:                   Particle-Particle
  54  * Calculate force/pot:        PotentialAndForce
  55  */
  56 void
  57 nb_kernel_ElecEwSw_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     int              ewitab;
  84     real             ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
  85     real             *ewtab;
  86     real             rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
  87
  88     x                = xx[0];
  89     f                = ff[0];
  90
  91     nri              = nlist->nri;
  92     iinr             = nlist->iinr;
  93     jindex           = nlist->jindex;
  94     jjnr             = nlist->jjnr;
  95     shiftidx         = nlist->shift;
  96     gid              = nlist->gid;
  97     shiftvec         = fr->shift_vec[0];
  98     fshift           = fr->fshift[0];
  99     facel            = fr->epsfac;
 100     charge           = mdatoms->chargeA;
 101     nvdwtype         = fr->ntype;
 102     vdwparam         = fr->nbfp;
 103     vdwtype          = mdatoms->typeA;
 104
 105     sh_ewald         = fr->ic->sh_ewald;
 106     ewtab            = fr->ic->tabq_coul_FDV0;
 107     ewtabscale       = fr->ic->tabq_scale;
 108     ewtabhalfspace   = 0.5/ewtabscale;
 109
 110     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 111     rcutoff          = fr->rcoulomb;
 112     rcutoff2         = rcutoff*rcutoff;
 113
 114     rswitch          = fr->rcoulomb_switch;
 115     /* Setup switch parameters */
 116     d                = rcutoff-rswitch;
 117     swV3             = -10.0/(d*d*d);
 118     swV4             =  15.0/(d*d*d*d);
 119     swV5             =  -6.0/(d*d*d*d*d);
 120     swF2             = -30.0/(d*d*d);
 121     swF3             =  60.0/(d*d*d*d);
 122     swF4             = -30.0/(d*d*d*d*d);
 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      = 3*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         = 3*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             cexp1_00         = vdwparam[vdwioffset0+vdwjidx0+1];
 201             cexp2_00         = vdwparam[vdwioffset0+vdwjidx0+2];
 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-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
 212             felec            = qq00*rinv00*(rinvsq00-felec);
 213
 214             /* BUCKINGHAM DISPERSION/REPULSION */
 215             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 216             vvdw6            = c6_00*rinvsix;
 217             br               = cexp2_00*r00;
 218             vvdwexp          = cexp1_00*exp(-br);
 219             vvdw             = vvdwexp - vvdw6*(1.0/6.0);
 220             fvdw             = (br*vvdwexp-vvdw6)*rinvsq00;
 221
 222             d                = r00-rswitch;
 223             d                = (d>0.0) ? d : 0.0;
 224             d2               = d*d;
 225             sw               = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
 226
 227             dsw              = d2*(swF2+d*(swF3+d*swF4));
 228
 229             /* Evaluate switch function */
 230             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 231             felec            = felec*sw - rinv00*velec*dsw;
 232             fvdw             = fvdw*sw - rinv00*vvdw*dsw;
 233             velec           *= sw;
 234             vvdw            *= sw;
 235
 236             /* Update potential sums from outer loop */
 237             velecsum        += velec;
 238             vvdwsum         += vvdw;
 239
 240             fscal            = felec+fvdw;
 241
 242             /* Calculate temporary vectorial force */
 243             tx               = fscal*dx00;
 244             ty               = fscal*dy00;
 245             tz               = fscal*dz00;
 246
 247             /* Update vectorial force */
 248             fix0            += tx;
 249             fiy0            += ty;
 250             fiz0            += tz;
 251             f[j_coord_offset+DIM*0+XX] -= tx;
 252             f[j_coord_offset+DIM*0+YY] -= ty;
 253             f[j_coord_offset+DIM*0+ZZ] -= tz;
 254
 255             }
 256
 257             /* Inner loop uses 101 flops */
 258         }
 259         /* End of innermost loop */
 260
 261         tx = ty = tz = 0;
 262         f[i_coord_offset+DIM*0+XX] += fix0;
 263         f[i_coord_offset+DIM*0+YY] += fiy0;
 264         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 265         tx                         += fix0;
 266         ty                         += fiy0;
 267         tz                         += fiz0;
 268         fshift[i_shift_offset+XX]  += tx;
 269         fshift[i_shift_offset+YY]  += ty;
 270         fshift[i_shift_offset+ZZ]  += tz;
 271
 272         ggid                        = gid[iidx];
 273         /* Update potential energies */
 274         kernel_data->energygrp_elec[ggid] += velecsum;
 275         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 276
 277         /* Increment number of inner iterations */
 278         inneriter                  += j_index_end - j_index_start;
 279
 280         /* Outer loop uses 15 flops */
 281     }
 282
 283     /* Increment number of outer iterations */
 284     outeriter        += nri;
 285
 286     /* Update outer/inner flops */
 287
 288     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_VF,outeriter*15 + inneriter*101);
 289 }
 290 /*
 291  * Gromacs nonbonded kernel:   nb_kernel_ElecEwSw_VdwBhamSw_GeomP1P1_F_c
 292  * Electrostatics interaction: Ewald
 293  * VdW interaction:            Buckingham
 294  * Geometry:                   Particle-Particle
 295  * Calculate force/pot:        Force
 296  */
 297 void
 298 nb_kernel_ElecEwSw_VdwBhamSw_GeomP1P1_F_c
 299                     (t_nblist                    * gmx_restrict       nlist,
 300                      rvec                        * gmx_restrict          xx,
 301                      rvec                        * gmx_restrict          ff,
 302                      t_forcerec                  * gmx_restrict          fr,
 303                      t_mdatoms                   * gmx_restrict     mdatoms,
 304                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 305                      t_nrnb                      * gmx_restrict        nrnb)
 306 {
 307     int              i_shift_offset,i_coord_offset,j_coord_offset;
 308     int              j_index_start,j_index_end;
 309     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 310     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 311     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 312     real             *shiftvec,*fshift,*x,*f;
 313     int              vdwioffset0;
 314     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 315     int              vdwjidx0;
 316     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 317     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 318     real             velec,felec,velecsum,facel,crf,krf,krf2;
 319     real             *charge;
 320     int              nvdwtype;
 321     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 322     int              *vdwtype;
 323     real             *vdwparam;
 324     int              ewitab;
 325     real             ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
 326     real             *ewtab;
 327     real             rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
 328
 329     x                = xx[0];
 330     f                = ff[0];
 331
 332     nri              = nlist->nri;
 333     iinr             = nlist->iinr;
 334     jindex           = nlist->jindex;
 335     jjnr             = nlist->jjnr;
 336     shiftidx         = nlist->shift;
 337     gid              = nlist->gid;
 338     shiftvec         = fr->shift_vec[0];
 339     fshift           = fr->fshift[0];
 340     facel            = fr->epsfac;
 341     charge           = mdatoms->chargeA;
 342     nvdwtype         = fr->ntype;
 343     vdwparam         = fr->nbfp;
 344     vdwtype          = mdatoms->typeA;
 345
 346     sh_ewald         = fr->ic->sh_ewald;
 347     ewtab            = fr->ic->tabq_coul_FDV0;
 348     ewtabscale       = fr->ic->tabq_scale;
 349     ewtabhalfspace   = 0.5/ewtabscale;
 350
 351     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 352     rcutoff          = fr->rcoulomb;
 353     rcutoff2         = rcutoff*rcutoff;
 354
 355     rswitch          = fr->rcoulomb_switch;
 356     /* Setup switch parameters */
 357     d                = rcutoff-rswitch;
 358     swV3             = -10.0/(d*d*d);
 359     swV4             =  15.0/(d*d*d*d);
 360     swV5             =  -6.0/(d*d*d*d*d);
 361     swF2             = -30.0/(d*d*d);
 362     swF3             =  60.0/(d*d*d*d);
 363     swF4             = -30.0/(d*d*d*d*d);
 364
 365     outeriter        = 0;
 366     inneriter        = 0;
 367
 368     /* Start outer loop over neighborlists */
 369     for(iidx=0; iidx<nri; iidx++)
 370     {
 371         /* Load shift vector for this list */
 372         i_shift_offset   = DIM*shiftidx[iidx];
 373         shX              = shiftvec[i_shift_offset+XX];
 374         shY              = shiftvec[i_shift_offset+YY];
 375         shZ              = shiftvec[i_shift_offset+ZZ];
 376
 377         /* Load limits for loop over neighbors */
 378         j_index_start    = jindex[iidx];
 379         j_index_end      = jindex[iidx+1];
 380
 381         /* Get outer coordinate index */
 382         inr              = iinr[iidx];
 383         i_coord_offset   = DIM*inr;
 384
 385         /* Load i particle coords and add shift vector */
 386         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 387         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 388         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 389
 390         fix0             = 0.0;
 391         fiy0             = 0.0;
 392         fiz0             = 0.0;
 393
 394         /* Load parameters for i particles */
 395         iq0              = facel*charge[inr+0];
 396         vdwioffset0      = 3*nvdwtype*vdwtype[inr+0];
 397
 398         /* Start inner kernel loop */
 399         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 400         {
 401             /* Get j neighbor index, and coordinate index */
 402             jnr              = jjnr[jidx];
 403             j_coord_offset   = DIM*jnr;
 404
 405             /* load j atom coordinates */
 406             jx0              = x[j_coord_offset+DIM*0+XX];
 407             jy0              = x[j_coord_offset+DIM*0+YY];
 408             jz0              = x[j_coord_offset+DIM*0+ZZ];
 409
 410             /* Calculate displacement vector */
 411             dx00             = ix0 - jx0;
 412             dy00             = iy0 - jy0;
 413             dz00             = iz0 - jz0;
 414
 415             /* Calculate squared distance and things based on it */
 416             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 417
 418             rinv00           = gmx_invsqrt(rsq00);
 419
 420             rinvsq00         = rinv00*rinv00;
 421
 422             /* Load parameters for j particles */
 423             jq0              = charge[jnr+0];
 424             vdwjidx0         = 3*vdwtype[jnr+0];
 425
 426             /**************************
 427              * CALCULATE INTERACTIONS *
 428              **************************/
 429
 430             if (rsq00<rcutoff2)
 431             {
 432
 433             r00              = rsq00*rinv00;
 434
 435             qq00             = iq0*jq0;
 436             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 437             cexp1_00         = vdwparam[vdwioffset0+vdwjidx0+1];
 438             cexp2_00         = vdwparam[vdwioffset0+vdwjidx0+2];
 439
 440             /* EWALD ELECTROSTATICS */
 441
 442             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 443             ewrt             = r00*ewtabscale;
 444             ewitab           = ewrt;
 445             eweps            = ewrt-ewitab;
 446             ewitab           = 4*ewitab;
 447             felec            = ewtab[ewitab]+eweps*ewtab[ewitab+1];
 448             velec            = qq00*(rinv00-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
 449             felec            = qq00*rinv00*(rinvsq00-felec);
 450
 451             /* BUCKINGHAM DISPERSION/REPULSION */
 452             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 453             vvdw6            = c6_00*rinvsix;
 454             br               = cexp2_00*r00;
 455             vvdwexp          = cexp1_00*exp(-br);
 456             vvdw             = vvdwexp - vvdw6*(1.0/6.0);
 457             fvdw             = (br*vvdwexp-vvdw6)*rinvsq00;
 458
 459             d                = r00-rswitch;
 460             d                = (d>0.0) ? d : 0.0;
 461             d2               = d*d;
 462             sw               = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
 463
 464             dsw              = d2*(swF2+d*(swF3+d*swF4));
 465
 466             /* Evaluate switch function */
 467             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 468             felec            = felec*sw - rinv00*velec*dsw;
 469             fvdw             = fvdw*sw - rinv00*vvdw*dsw;
 470
 471             fscal            = felec+fvdw;
 472
 473             /* Calculate temporary vectorial force */
 474             tx               = fscal*dx00;
 475             ty               = fscal*dy00;
 476             tz               = fscal*dz00;
 477
 478             /* Update vectorial force */
 479             fix0            += tx;
 480             fiy0            += ty;
 481             fiz0            += tz;
 482             f[j_coord_offset+DIM*0+XX] -= tx;
 483             f[j_coord_offset+DIM*0+YY] -= ty;
 484             f[j_coord_offset+DIM*0+ZZ] -= tz;
 485
 486             }
 487
 488             /* Inner loop uses 97 flops */
 489         }
 490         /* End of innermost loop */
 491
 492         tx = ty = tz = 0;
 493         f[i_coord_offset+DIM*0+XX] += fix0;
 494         f[i_coord_offset+DIM*0+YY] += fiy0;
 495         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 496         tx                         += fix0;
 497         ty                         += fiy0;
 498         tz                         += fiz0;
 499         fshift[i_shift_offset+XX]  += tx;
 500         fshift[i_shift_offset+YY]  += ty;
 501         fshift[i_shift_offset+ZZ]  += tz;
 502
 503         /* Increment number of inner iterations */
 504         inneriter                  += j_index_end - j_index_start;
 505
 506         /* Outer loop uses 13 flops */
 507     }
 508
 509     /* Increment number of outer iterations */
 510     outeriter        += nri;
 511
 512     /* Update outer/inner flops */
 513
 514     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_F,outeriter*13 + inneriter*97);
 515 }