src/gromacs/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecGB_VdwLJ_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_ElecGB_VdwLJ_GeomP1P1_VF_c
  49  * Electrostatics interaction: GeneralizedBorn
  50  * VdW interaction:            LennardJones
  51  * Geometry:                   Particle-Particle
  52  * Calculate force/pot:        PotentialAndForce
  53  */
  54 void
  55 nb_kernel_ElecGB_VdwLJ_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              gbitab;
  78     real             vgb,fgb,vgbsum,dvdasum,gbscale,gbtabscale,isaprod,gbqqfactor,gbinvepsdiff,dvdaj,gbeps,dvdatmp;
  79     real             *invsqrta,*dvda,*gbtab;
  80     int              nvdwtype;
  81     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
  82     int              *vdwtype;
  83     real             *vdwparam;
  84     int              vfitab;
  85     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
  86     real             *vftab;
  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->ic->epsfac;
 100     charge           = mdatoms->chargeA;
 101     nvdwtype         = fr->ntype;
 102     vdwparam         = fr->nbfp;
 103     vdwtype          = mdatoms->typeA;
 104
 105     invsqrta         = fr->invsqrta;
 106     dvda             = fr->dvda;
 107     gbtabscale       = fr->gbtab->scale;
 108     gbtab            = fr->gbtab->data;
 109     gbinvepsdiff     = (1.0/fr->ic->epsilon_r) - (1.0/fr->gb_epsilon_solvent);
 110
 111     outeriter        = 0;
 112     inneriter        = 0;
 113
 114     /* Start outer loop over neighborlists */
 115     for(iidx=0; iidx<nri; iidx++)
 116     {
 117         /* Load shift vector for this list */
 118         i_shift_offset   = DIM*shiftidx[iidx];
 119         shX              = shiftvec[i_shift_offset+XX];
 120         shY              = shiftvec[i_shift_offset+YY];
 121         shZ              = shiftvec[i_shift_offset+ZZ];
 122
 123         /* Load limits for loop over neighbors */
 124         j_index_start    = jindex[iidx];
 125         j_index_end      = jindex[iidx+1];
 126
 127         /* Get outer coordinate index */
 128         inr              = iinr[iidx];
 129         i_coord_offset   = DIM*inr;
 130
 131         /* Load i particle coords and add shift vector */
 132         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 133         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 134         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 135
 136         fix0             = 0.0;
 137         fiy0             = 0.0;
 138         fiz0             = 0.0;
 139
 140         /* Load parameters for i particles */
 141         iq0              = facel*charge[inr+0];
 142         isai0            = invsqrta[inr+0];
 143         vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 144
 145         /* Reset potential sums */
 146         velecsum         = 0.0;
 147         vgbsum           = 0.0;
 148         vvdwsum          = 0.0;
 149         dvdasum          = 0.0;
 150
 151         /* Start inner kernel loop */
 152         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 153         {
 154             /* Get j neighbor index, and coordinate index */
 155             jnr              = jjnr[jidx];
 156             j_coord_offset   = DIM*jnr;
 157
 158             /* load j atom coordinates */
 159             jx0              = x[j_coord_offset+DIM*0+XX];
 160             jy0              = x[j_coord_offset+DIM*0+YY];
 161             jz0              = x[j_coord_offset+DIM*0+ZZ];
 162
 163             /* Calculate displacement vector */
 164             dx00             = ix0 - jx0;
 165             dy00             = iy0 - jy0;
 166             dz00             = iz0 - jz0;
 167
 168             /* Calculate squared distance and things based on it */
 169             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 170
 171             rinv00           = 1.0/sqrt(rsq00);
 172
 173             rinvsq00         = rinv00*rinv00;
 174
 175             /* Load parameters for j particles */
 176             jq0              = charge[jnr+0];
 177             isaj0           = invsqrta[jnr+0];
 178             vdwjidx0         = 2*vdwtype[jnr+0];
 179
 180             /**************************
 181              * CALCULATE INTERACTIONS *
 182              **************************/
 183
 184             r00              = rsq00*rinv00;
 185
 186             qq00             = iq0*jq0;
 187             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 188             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 189
 190             /* GENERALIZED BORN AND COULOMB ELECTROSTATICS */
 191             isaprod          = isai0*isaj0;
 192             gbqqfactor       = isaprod*(-qq00)*gbinvepsdiff;
 193             gbscale          = isaprod*gbtabscale;
 194             dvdaj            = dvda[jnr+0];
 195
 196             /* Calculate generalized born table index - this is a separate table from the normal one,
 197              * but we use the same procedure by multiplying r with scale and truncating to integer.
 198              */
 199             rt               = r00*gbscale;
 200             gbitab           = rt;
 201             gbeps            = rt-gbitab;
 202             gbitab           = 4*gbitab;
 203
 204             Y                = gbtab[gbitab];
 205             F                = gbtab[gbitab+1];
 206             Geps             = gbeps*gbtab[gbitab+2];
 207             Heps2            = gbeps*gbeps*gbtab[gbitab+3];
 208             Fp               = F+Geps+Heps2;
 209             VV               = Y+gbeps*Fp;
 210             vgb              = gbqqfactor*VV;
 211
 212             FF               = Fp+Geps+2.0*Heps2;
 213             fgb              = gbqqfactor*FF*gbscale;
 214             dvdatmp          = -0.5*(vgb+fgb*r00);
 215             dvdasum          = dvdasum + dvdatmp;
 216             dvda[jnr]        = dvdaj+dvdatmp*isaj0*isaj0;
 217             velec            = qq00*rinv00;
 218             felec            = (velec*rinv00-fgb)*rinv00;
 219
 220             /* LENNARD-JONES DISPERSION/REPULSION */
 221
 222             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 223             vvdw6            = c6_00*rinvsix;
 224             vvdw12           = c12_00*rinvsix*rinvsix;
 225             vvdw             = vvdw12*(1.0/12.0) - vvdw6*(1.0/6.0);
 226             fvdw             = (vvdw12-vvdw6)*rinvsq00;
 227
 228             /* Update potential sums from outer loop */
 229             velecsum        += velec;
 230             vgbsum          += vgb;
 231             vvdwsum         += vvdw;
 232
 233             fscal            = felec+fvdw;
 234
 235             /* Calculate temporary vectorial force */
 236             tx               = fscal*dx00;
 237             ty               = fscal*dy00;
 238             tz               = fscal*dz00;
 239
 240             /* Update vectorial force */
 241             fix0            += tx;
 242             fiy0            += ty;
 243             fiz0            += tz;
 244             f[j_coord_offset+DIM*0+XX] -= tx;
 245             f[j_coord_offset+DIM*0+YY] -= ty;
 246             f[j_coord_offset+DIM*0+ZZ] -= tz;
 247
 248             /* Inner loop uses 71 flops */
 249         }
 250         /* End of innermost loop */
 251
 252         tx = ty = tz = 0;
 253         f[i_coord_offset+DIM*0+XX] += fix0;
 254         f[i_coord_offset+DIM*0+YY] += fiy0;
 255         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 256         tx                         += fix0;
 257         ty                         += fiy0;
 258         tz                         += fiz0;
 259         fshift[i_shift_offset+XX]  += tx;
 260         fshift[i_shift_offset+YY]  += ty;
 261         fshift[i_shift_offset+ZZ]  += tz;
 262
 263         ggid                        = gid[iidx];
 264         /* Update potential energies */
 265         kernel_data->energygrp_elec[ggid] += velecsum;
 266         kernel_data->energygrp_polarization[ggid] += vgbsum;
 267         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 268         dvda[inr]                   = dvda[inr] + dvdasum*isai0*isai0;
 269
 270         /* Increment number of inner iterations */
 271         inneriter                  += j_index_end - j_index_start;
 272
 273         /* Outer loop uses 16 flops */
 274     }
 275
 276     /* Increment number of outer iterations */
 277     outeriter        += nri;
 278
 279     /* Update outer/inner flops */
 280
 281     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_VF,outeriter*16 + inneriter*71);
 282 }
 283 /*
 284  * Gromacs nonbonded kernel:   nb_kernel_ElecGB_VdwLJ_GeomP1P1_F_c
 285  * Electrostatics interaction: GeneralizedBorn
 286  * VdW interaction:            LennardJones
 287  * Geometry:                   Particle-Particle
 288  * Calculate force/pot:        Force
 289  */
 290 void
 291 nb_kernel_ElecGB_VdwLJ_GeomP1P1_F_c
 292                     (t_nblist                    * gmx_restrict       nlist,
 293                      rvec                        * gmx_restrict          xx,
 294                      rvec                        * gmx_restrict          ff,
 295                      struct t_forcerec           * gmx_restrict          fr,
 296                      t_mdatoms                   * gmx_restrict     mdatoms,
 297                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 298                      t_nrnb                      * gmx_restrict        nrnb)
 299 {
 300     int              i_shift_offset,i_coord_offset,j_coord_offset;
 301     int              j_index_start,j_index_end;
 302     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 303     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 304     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 305     real             *shiftvec,*fshift,*x,*f;
 306     int              vdwioffset0;
 307     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 308     int              vdwjidx0;
 309     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 310     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 311     real             velec,felec,velecsum,facel,crf,krf,krf2;
 312     real             *charge;
 313     int              gbitab;
 314     real             vgb,fgb,vgbsum,dvdasum,gbscale,gbtabscale,isaprod,gbqqfactor,gbinvepsdiff,dvdaj,gbeps,dvdatmp;
 315     real             *invsqrta,*dvda,*gbtab;
 316     int              nvdwtype;
 317     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 318     int              *vdwtype;
 319     real             *vdwparam;
 320     int              vfitab;
 321     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
 322     real             *vftab;
 323
 324     x                = xx[0];
 325     f                = ff[0];
 326
 327     nri              = nlist->nri;
 328     iinr             = nlist->iinr;
 329     jindex           = nlist->jindex;
 330     jjnr             = nlist->jjnr;
 331     shiftidx         = nlist->shift;
 332     gid              = nlist->gid;
 333     shiftvec         = fr->shift_vec[0];
 334     fshift           = fr->fshift[0];
 335     facel            = fr->ic->epsfac;
 336     charge           = mdatoms->chargeA;
 337     nvdwtype         = fr->ntype;
 338     vdwparam         = fr->nbfp;
 339     vdwtype          = mdatoms->typeA;
 340
 341     invsqrta         = fr->invsqrta;
 342     dvda             = fr->dvda;
 343     gbtabscale       = fr->gbtab->scale;
 344     gbtab            = fr->gbtab->data;
 345     gbinvepsdiff     = (1.0/fr->ic->epsilon_r) - (1.0/fr->gb_epsilon_solvent);
 346
 347     outeriter        = 0;
 348     inneriter        = 0;
 349
 350     /* Start outer loop over neighborlists */
 351     for(iidx=0; iidx<nri; iidx++)
 352     {
 353         /* Load shift vector for this list */
 354         i_shift_offset   = DIM*shiftidx[iidx];
 355         shX              = shiftvec[i_shift_offset+XX];
 356         shY              = shiftvec[i_shift_offset+YY];
 357         shZ              = shiftvec[i_shift_offset+ZZ];
 358
 359         /* Load limits for loop over neighbors */
 360         j_index_start    = jindex[iidx];
 361         j_index_end      = jindex[iidx+1];
 362
 363         /* Get outer coordinate index */
 364         inr              = iinr[iidx];
 365         i_coord_offset   = DIM*inr;
 366
 367         /* Load i particle coords and add shift vector */
 368         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 369         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 370         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 371
 372         fix0             = 0.0;
 373         fiy0             = 0.0;
 374         fiz0             = 0.0;
 375
 376         /* Load parameters for i particles */
 377         iq0              = facel*charge[inr+0];
 378         isai0            = invsqrta[inr+0];
 379         vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 380
 381         dvdasum          = 0.0;
 382
 383         /* Start inner kernel loop */
 384         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 385         {
 386             /* Get j neighbor index, and coordinate index */
 387             jnr              = jjnr[jidx];
 388             j_coord_offset   = DIM*jnr;
 389
 390             /* load j atom coordinates */
 391             jx0              = x[j_coord_offset+DIM*0+XX];
 392             jy0              = x[j_coord_offset+DIM*0+YY];
 393             jz0              = x[j_coord_offset+DIM*0+ZZ];
 394
 395             /* Calculate displacement vector */
 396             dx00             = ix0 - jx0;
 397             dy00             = iy0 - jy0;
 398             dz00             = iz0 - jz0;
 399
 400             /* Calculate squared distance and things based on it */
 401             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 402
 403             rinv00           = 1.0/sqrt(rsq00);
 404
 405             rinvsq00         = rinv00*rinv00;
 406
 407             /* Load parameters for j particles */
 408             jq0              = charge[jnr+0];
 409             isaj0           = invsqrta[jnr+0];
 410             vdwjidx0         = 2*vdwtype[jnr+0];
 411
 412             /**************************
 413              * CALCULATE INTERACTIONS *
 414              **************************/
 415
 416             r00              = rsq00*rinv00;
 417
 418             qq00             = iq0*jq0;
 419             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 420             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 421
 422             /* GENERALIZED BORN AND COULOMB ELECTROSTATICS */
 423             isaprod          = isai0*isaj0;
 424             gbqqfactor       = isaprod*(-qq00)*gbinvepsdiff;
 425             gbscale          = isaprod*gbtabscale;
 426             dvdaj            = dvda[jnr+0];
 427
 428             /* Calculate generalized born table index - this is a separate table from the normal one,
 429              * but we use the same procedure by multiplying r with scale and truncating to integer.
 430              */
 431             rt               = r00*gbscale;
 432             gbitab           = rt;
 433             gbeps            = rt-gbitab;
 434             gbitab           = 4*gbitab;
 435
 436             Y                = gbtab[gbitab];
 437             F                = gbtab[gbitab+1];
 438             Geps             = gbeps*gbtab[gbitab+2];
 439             Heps2            = gbeps*gbeps*gbtab[gbitab+3];
 440             Fp               = F+Geps+Heps2;
 441             VV               = Y+gbeps*Fp;
 442             vgb              = gbqqfactor*VV;
 443
 444             FF               = Fp+Geps+2.0*Heps2;
 445             fgb              = gbqqfactor*FF*gbscale;
 446             dvdatmp          = -0.5*(vgb+fgb*r00);
 447             dvdasum          = dvdasum + dvdatmp;
 448             dvda[jnr]        = dvdaj+dvdatmp*isaj0*isaj0;
 449             velec            = qq00*rinv00;
 450             felec            = (velec*rinv00-fgb)*rinv00;
 451
 452             /* LENNARD-JONES DISPERSION/REPULSION */
 453
 454             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 455             fvdw             = (c12_00*rinvsix-c6_00)*rinvsix*rinvsq00;
 456
 457             fscal            = felec+fvdw;
 458
 459             /* Calculate temporary vectorial force */
 460             tx               = fscal*dx00;
 461             ty               = fscal*dy00;
 462             tz               = fscal*dz00;
 463
 464             /* Update vectorial force */
 465             fix0            += tx;
 466             fiy0            += ty;
 467             fiz0            += tz;
 468             f[j_coord_offset+DIM*0+XX] -= tx;
 469             f[j_coord_offset+DIM*0+YY] -= ty;
 470             f[j_coord_offset+DIM*0+ZZ] -= tz;
 471
 472             /* Inner loop uses 64 flops */
 473         }
 474         /* End of innermost loop */
 475
 476         tx = ty = tz = 0;
 477         f[i_coord_offset+DIM*0+XX] += fix0;
 478         f[i_coord_offset+DIM*0+YY] += fiy0;
 479         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 480         tx                         += fix0;
 481         ty                         += fiy0;
 482         tz                         += fiz0;
 483         fshift[i_shift_offset+XX]  += tx;
 484         fshift[i_shift_offset+YY]  += ty;
 485         fshift[i_shift_offset+ZZ]  += tz;
 486
 487         dvda[inr]                   = dvda[inr] + dvdasum*isai0*isai0;
 488
 489         /* Increment number of inner iterations */
 490         inneriter                  += j_index_end - j_index_start;
 491
 492         /* Outer loop uses 13 flops */
 493     }
 494
 495     /* Increment number of outer iterations */
 496     outeriter        += nri;
 497
 498     /* Update outer/inner flops */
 499
 500     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_F,outeriter*13 + inneriter*64);
 501 }