src/gromacs/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecGB_VdwBham_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 "config.h"
  39
  40 #include <math.h>
  41
  42 #include "../nb_kernel.h"
  43 #include "gromacs/legacyheaders/types/simple.h"
  44 #include "gromacs/math/vec.h"
  45 #include "gromacs/legacyheaders/nrnb.h"
  46
  47 /*
  48  * Gromacs nonbonded kernel:   nb_kernel_ElecGB_VdwBham_GeomP1P1_VF_c
  49  * Electrostatics interaction: GeneralizedBorn
  50  * VdW interaction:            Buckingham
  51  * Geometry:                   Particle-Particle
  52  * Calculate force/pot:        PotentialAndForce
  53  */
  54 void
  55 nb_kernel_ElecGB_VdwBham_GeomP1P1_VF_c
  56                     (t_nblist                    * gmx_restrict       nlist,
  57                      rvec                        * gmx_restrict          xx,
  58                      rvec                        * gmx_restrict          ff,
  59                      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->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->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      = 3*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           = gmx_invsqrt(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         = 3*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             cexp1_00         = vdwparam[vdwioffset0+vdwjidx0+1];
 189             cexp2_00         = vdwparam[vdwioffset0+vdwjidx0+2];
 190
 191             /* GENERALIZED BORN AND COULOMB ELECTROSTATICS */
 192             isaprod          = isai0*isaj0;
 193             gbqqfactor       = isaprod*(-qq00)*gbinvepsdiff;
 194             gbscale          = isaprod*gbtabscale;
 195             dvdaj            = dvda[jnr+0];
 196
 197             /* Calculate generalized born table index - this is a separate table from the normal one,
 198              * but we use the same procedure by multiplying r with scale and truncating to integer.
 199              */
 200             rt               = r00*gbscale;
 201             gbitab           = rt;
 202             gbeps            = rt-gbitab;
 203             gbitab           = 4*gbitab;
 204
 205             Y                = gbtab[gbitab];
 206             F                = gbtab[gbitab+1];
 207             Geps             = gbeps*gbtab[gbitab+2];
 208             Heps2            = gbeps*gbeps*gbtab[gbitab+3];
 209             Fp               = F+Geps+Heps2;
 210             VV               = Y+gbeps*Fp;
 211             vgb              = gbqqfactor*VV;
 212
 213             FF               = Fp+Geps+2.0*Heps2;
 214             fgb              = gbqqfactor*FF*gbscale;
 215             dvdatmp          = -0.5*(vgb+fgb*r00);
 216             dvdasum          = dvdasum + dvdatmp;
 217             dvda[jnr]        = dvdaj+dvdatmp*isaj0*isaj0;
 218             velec            = qq00*rinv00;
 219             felec            = (velec*rinv00-fgb)*rinv00;
 220
 221             /* BUCKINGHAM DISPERSION/REPULSION */
 222             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 223             vvdw6            = c6_00*rinvsix;
 224             br               = cexp2_00*r00;
 225             vvdwexp          = cexp1_00*exp(-br);
 226             vvdw             = vvdwexp - vvdw6*(1.0/6.0);
 227             fvdw             = (br*vvdwexp-vvdw6)*rinvsq00;
 228
 229             /* Update potential sums from outer loop */
 230             velecsum        += velec;
 231             vgbsum          += vgb;
 232             vvdwsum         += vvdw;
 233
 234             fscal            = felec+fvdw;
 235
 236             /* Calculate temporary vectorial force */
 237             tx               = fscal*dx00;
 238             ty               = fscal*dy00;
 239             tz               = fscal*dz00;
 240
 241             /* Update vectorial force */
 242             fix0            += tx;
 243             fiy0            += ty;
 244             fiz0            += tz;
 245             f[j_coord_offset+DIM*0+XX] -= tx;
 246             f[j_coord_offset+DIM*0+YY] -= ty;
 247             f[j_coord_offset+DIM*0+ZZ] -= tz;
 248
 249             /* Inner loop uses 97 flops */
 250         }
 251         /* End of innermost loop */
 252
 253         tx = ty = tz = 0;
 254         f[i_coord_offset+DIM*0+XX] += fix0;
 255         f[i_coord_offset+DIM*0+YY] += fiy0;
 256         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 257         tx                         += fix0;
 258         ty                         += fiy0;
 259         tz                         += fiz0;
 260         fshift[i_shift_offset+XX]  += tx;
 261         fshift[i_shift_offset+YY]  += ty;
 262         fshift[i_shift_offset+ZZ]  += tz;
 263
 264         ggid                        = gid[iidx];
 265         /* Update potential energies */
 266         kernel_data->energygrp_elec[ggid] += velecsum;
 267         kernel_data->energygrp_polarization[ggid] += vgbsum;
 268         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 269         dvda[inr]                   = dvda[inr] + dvdasum*isai0*isai0;
 270
 271         /* Increment number of inner iterations */
 272         inneriter                  += j_index_end - j_index_start;
 273
 274         /* Outer loop uses 16 flops */
 275     }
 276
 277     /* Increment number of outer iterations */
 278     outeriter        += nri;
 279
 280     /* Update outer/inner flops */
 281
 282     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_VF,outeriter*16 + inneriter*97);
 283 }
 284 /*
 285  * Gromacs nonbonded kernel:   nb_kernel_ElecGB_VdwBham_GeomP1P1_F_c
 286  * Electrostatics interaction: GeneralizedBorn
 287  * VdW interaction:            Buckingham
 288  * Geometry:                   Particle-Particle
 289  * Calculate force/pot:        Force
 290  */
 291 void
 292 nb_kernel_ElecGB_VdwBham_GeomP1P1_F_c
 293                     (t_nblist                    * gmx_restrict       nlist,
 294                      rvec                        * gmx_restrict          xx,
 295                      rvec                        * gmx_restrict          ff,
 296                      t_forcerec                  * gmx_restrict          fr,
 297                      t_mdatoms                   * gmx_restrict     mdatoms,
 298                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 299                      t_nrnb                      * gmx_restrict        nrnb)
 300 {
 301     int              i_shift_offset,i_coord_offset,j_coord_offset;
 302     int              j_index_start,j_index_end;
 303     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 304     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 305     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 306     real             *shiftvec,*fshift,*x,*f;
 307     int              vdwioffset0;
 308     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 309     int              vdwjidx0;
 310     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 311     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 312     real             velec,felec,velecsum,facel,crf,krf,krf2;
 313     real             *charge;
 314     int              gbitab;
 315     real             vgb,fgb,vgbsum,dvdasum,gbscale,gbtabscale,isaprod,gbqqfactor,gbinvepsdiff,dvdaj,gbeps,dvdatmp;
 316     real             *invsqrta,*dvda,*gbtab;
 317     int              nvdwtype;
 318     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 319     int              *vdwtype;
 320     real             *vdwparam;
 321     int              vfitab;
 322     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
 323     real             *vftab;
 324
 325     x                = xx[0];
 326     f                = ff[0];
 327
 328     nri              = nlist->nri;
 329     iinr             = nlist->iinr;
 330     jindex           = nlist->jindex;
 331     jjnr             = nlist->jjnr;
 332     shiftidx         = nlist->shift;
 333     gid              = nlist->gid;
 334     shiftvec         = fr->shift_vec[0];
 335     fshift           = fr->fshift[0];
 336     facel            = fr->epsfac;
 337     charge           = mdatoms->chargeA;
 338     nvdwtype         = fr->ntype;
 339     vdwparam         = fr->nbfp;
 340     vdwtype          = mdatoms->typeA;
 341
 342     invsqrta         = fr->invsqrta;
 343     dvda             = fr->dvda;
 344     gbtabscale       = fr->gbtab.scale;
 345     gbtab            = fr->gbtab.data;
 346     gbinvepsdiff     = (1.0/fr->epsilon_r) - (1.0/fr->gb_epsilon_solvent);
 347
 348     outeriter        = 0;
 349     inneriter        = 0;
 350
 351     /* Start outer loop over neighborlists */
 352     for(iidx=0; iidx<nri; iidx++)
 353     {
 354         /* Load shift vector for this list */
 355         i_shift_offset   = DIM*shiftidx[iidx];
 356         shX              = shiftvec[i_shift_offset+XX];
 357         shY              = shiftvec[i_shift_offset+YY];
 358         shZ              = shiftvec[i_shift_offset+ZZ];
 359
 360         /* Load limits for loop over neighbors */
 361         j_index_start    = jindex[iidx];
 362         j_index_end      = jindex[iidx+1];
 363
 364         /* Get outer coordinate index */
 365         inr              = iinr[iidx];
 366         i_coord_offset   = DIM*inr;
 367
 368         /* Load i particle coords and add shift vector */
 369         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 370         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 371         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 372
 373         fix0             = 0.0;
 374         fiy0             = 0.0;
 375         fiz0             = 0.0;
 376
 377         /* Load parameters for i particles */
 378         iq0              = facel*charge[inr+0];
 379         isai0            = invsqrta[inr+0];
 380         vdwioffset0      = 3*nvdwtype*vdwtype[inr+0];
 381
 382         dvdasum          = 0.0;
 383
 384         /* Start inner kernel loop */
 385         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 386         {
 387             /* Get j neighbor index, and coordinate index */
 388             jnr              = jjnr[jidx];
 389             j_coord_offset   = DIM*jnr;
 390
 391             /* load j atom coordinates */
 392             jx0              = x[j_coord_offset+DIM*0+XX];
 393             jy0              = x[j_coord_offset+DIM*0+YY];
 394             jz0              = x[j_coord_offset+DIM*0+ZZ];
 395
 396             /* Calculate displacement vector */
 397             dx00             = ix0 - jx0;
 398             dy00             = iy0 - jy0;
 399             dz00             = iz0 - jz0;
 400
 401             /* Calculate squared distance and things based on it */
 402             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 403
 404             rinv00           = gmx_invsqrt(rsq00);
 405
 406             rinvsq00         = rinv00*rinv00;
 407
 408             /* Load parameters for j particles */
 409             jq0              = charge[jnr+0];
 410             isaj0           = invsqrta[jnr+0];
 411             vdwjidx0         = 3*vdwtype[jnr+0];
 412
 413             /**************************
 414              * CALCULATE INTERACTIONS *
 415              **************************/
 416
 417             r00              = rsq00*rinv00;
 418
 419             qq00             = iq0*jq0;
 420             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 421             cexp1_00         = vdwparam[vdwioffset0+vdwjidx0+1];
 422             cexp2_00         = vdwparam[vdwioffset0+vdwjidx0+2];
 423
 424             /* GENERALIZED BORN AND COULOMB ELECTROSTATICS */
 425             isaprod          = isai0*isaj0;
 426             gbqqfactor       = isaprod*(-qq00)*gbinvepsdiff;
 427             gbscale          = isaprod*gbtabscale;
 428             dvdaj            = dvda[jnr+0];
 429
 430             /* Calculate generalized born table index - this is a separate table from the normal one,
 431              * but we use the same procedure by multiplying r with scale and truncating to integer.
 432              */
 433             rt               = r00*gbscale;
 434             gbitab           = rt;
 435             gbeps            = rt-gbitab;
 436             gbitab           = 4*gbitab;
 437
 438             Y                = gbtab[gbitab];
 439             F                = gbtab[gbitab+1];
 440             Geps             = gbeps*gbtab[gbitab+2];
 441             Heps2            = gbeps*gbeps*gbtab[gbitab+3];
 442             Fp               = F+Geps+Heps2;
 443             VV               = Y+gbeps*Fp;
 444             vgb              = gbqqfactor*VV;
 445
 446             FF               = Fp+Geps+2.0*Heps2;
 447             fgb              = gbqqfactor*FF*gbscale;
 448             dvdatmp          = -0.5*(vgb+fgb*r00);
 449             dvdasum          = dvdasum + dvdatmp;
 450             dvda[jnr]        = dvdaj+dvdatmp*isaj0*isaj0;
 451             velec            = qq00*rinv00;
 452             felec            = (velec*rinv00-fgb)*rinv00;
 453
 454             /* BUCKINGHAM DISPERSION/REPULSION */
 455             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 456             vvdw6            = c6_00*rinvsix;
 457             br               = cexp2_00*r00;
 458             vvdwexp          = cexp1_00*exp(-br);
 459             fvdw             = (br*vvdwexp-vvdw6)*rinvsq00;
 460
 461             fscal            = felec+fvdw;
 462
 463             /* Calculate temporary vectorial force */
 464             tx               = fscal*dx00;
 465             ty               = fscal*dy00;
 466             tz               = fscal*dz00;
 467
 468             /* Update vectorial force */
 469             fix0            += tx;
 470             fiy0            += ty;
 471             fiz0            += tz;
 472             f[j_coord_offset+DIM*0+XX] -= tx;
 473             f[j_coord_offset+DIM*0+YY] -= ty;
 474             f[j_coord_offset+DIM*0+ZZ] -= tz;
 475
 476             /* Inner loop uses 92 flops */
 477         }
 478         /* End of innermost loop */
 479
 480         tx = ty = tz = 0;
 481         f[i_coord_offset+DIM*0+XX] += fix0;
 482         f[i_coord_offset+DIM*0+YY] += fiy0;
 483         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 484         tx                         += fix0;
 485         ty                         += fiy0;
 486         tz                         += fiz0;
 487         fshift[i_shift_offset+XX]  += tx;
 488         fshift[i_shift_offset+YY]  += ty;
 489         fshift[i_shift_offset+ZZ]  += tz;
 490
 491         dvda[inr]                   = dvda[inr] + dvdasum*isai0*isai0;
 492
 493         /* Increment number of inner iterations */
 494         inneriter                  += j_index_end - j_index_start;
 495
 496         /* Outer loop uses 13 flops */
 497     }
 498
 499     /* Increment number of outer iterations */
 500     outeriter        += nri;
 501
 502     /* Update outer/inner flops */
 503
 504     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_F,outeriter*13 + inneriter*92);
 505 }