src/gromacs/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecGB_VdwNone_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 "types/simple.h"
  44 #include "gromacs/math/vec.h"
  45 #include "nrnb.h"
  46
  47 /*
  48  * Gromacs nonbonded kernel:   nb_kernel_ElecGB_VdwNone_GeomP1P1_VF_c
  49  * Electrostatics interaction: GeneralizedBorn
  50  * VdW interaction:            None
  51  * Geometry:                   Particle-Particle
  52  * Calculate force/pot:        PotentialAndForce
  53  */
  54 void
  55 nb_kernel_ElecGB_VdwNone_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              vfitab;
  81     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
  82     real             *vftab;
  83
  84     x                = xx[0];
  85     f                = ff[0];
  86
  87     nri              = nlist->nri;
  88     iinr             = nlist->iinr;
  89     jindex           = nlist->jindex;
  90     jjnr             = nlist->jjnr;
  91     shiftidx         = nlist->shift;
  92     gid              = nlist->gid;
  93     shiftvec         = fr->shift_vec[0];
  94     fshift           = fr->fshift[0];
  95     facel            = fr->epsfac;
  96     charge           = mdatoms->chargeA;
  97
  98     invsqrta         = fr->invsqrta;
  99     dvda             = fr->dvda;
 100     gbtabscale       = fr->gbtab.scale;
 101     gbtab            = fr->gbtab.data;
 102     gbinvepsdiff     = (1.0/fr->epsilon_r) - (1.0/fr->gb_epsilon_solvent);
 103
 104     outeriter        = 0;
 105     inneriter        = 0;
 106
 107     /* Start outer loop over neighborlists */
 108     for(iidx=0; iidx<nri; iidx++)
 109     {
 110         /* Load shift vector for this list */
 111         i_shift_offset   = DIM*shiftidx[iidx];
 112         shX              = shiftvec[i_shift_offset+XX];
 113         shY              = shiftvec[i_shift_offset+YY];
 114         shZ              = shiftvec[i_shift_offset+ZZ];
 115
 116         /* Load limits for loop over neighbors */
 117         j_index_start    = jindex[iidx];
 118         j_index_end      = jindex[iidx+1];
 119
 120         /* Get outer coordinate index */
 121         inr              = iinr[iidx];
 122         i_coord_offset   = DIM*inr;
 123
 124         /* Load i particle coords and add shift vector */
 125         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 126         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 127         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 128
 129         fix0             = 0.0;
 130         fiy0             = 0.0;
 131         fiz0             = 0.0;
 132
 133         /* Load parameters for i particles */
 134         iq0              = facel*charge[inr+0];
 135         isai0            = invsqrta[inr+0];
 136
 137         /* Reset potential sums */
 138         velecsum         = 0.0;
 139         vgbsum           = 0.0;
 140         dvdasum          = 0.0;
 141
 142         /* Start inner kernel loop */
 143         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 144         {
 145             /* Get j neighbor index, and coordinate index */
 146             jnr              = jjnr[jidx];
 147             j_coord_offset   = DIM*jnr;
 148
 149             /* load j atom coordinates */
 150             jx0              = x[j_coord_offset+DIM*0+XX];
 151             jy0              = x[j_coord_offset+DIM*0+YY];
 152             jz0              = x[j_coord_offset+DIM*0+ZZ];
 153
 154             /* Calculate displacement vector */
 155             dx00             = ix0 - jx0;
 156             dy00             = iy0 - jy0;
 157             dz00             = iz0 - jz0;
 158
 159             /* Calculate squared distance and things based on it */
 160             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 161
 162             rinv00           = gmx_invsqrt(rsq00);
 163
 164             /* Load parameters for j particles */
 165             jq0              = charge[jnr+0];
 166             isaj0           = invsqrta[jnr+0];
 167
 168             /**************************
 169              * CALCULATE INTERACTIONS *
 170              **************************/
 171
 172             r00              = rsq00*rinv00;
 173
 174             qq00             = iq0*jq0;
 175
 176             /* GENERALIZED BORN AND COULOMB ELECTROSTATICS */
 177             isaprod          = isai0*isaj0;
 178             gbqqfactor       = isaprod*(-qq00)*gbinvepsdiff;
 179             gbscale          = isaprod*gbtabscale;
 180             dvdaj            = dvda[jnr+0];
 181
 182             /* Calculate generalized born table index - this is a separate table from the normal one,
 183              * but we use the same procedure by multiplying r with scale and truncating to integer.
 184              */
 185             rt               = r00*gbscale;
 186             gbitab           = rt;
 187             gbeps            = rt-gbitab;
 188             gbitab           = 4*gbitab;
 189
 190             Y                = gbtab[gbitab];
 191             F                = gbtab[gbitab+1];
 192             Geps             = gbeps*gbtab[gbitab+2];
 193             Heps2            = gbeps*gbeps*gbtab[gbitab+3];
 194             Fp               = F+Geps+Heps2;
 195             VV               = Y+gbeps*Fp;
 196             vgb              = gbqqfactor*VV;
 197
 198             FF               = Fp+Geps+2.0*Heps2;
 199             fgb              = gbqqfactor*FF*gbscale;
 200             dvdatmp          = -0.5*(vgb+fgb*r00);
 201             dvdasum          = dvdasum + dvdatmp;
 202             dvda[jnr]        = dvdaj+dvdatmp*isaj0*isaj0;
 203             velec            = qq00*rinv00;
 204             felec            = (velec*rinv00-fgb)*rinv00;
 205
 206             /* Update potential sums from outer loop */
 207             velecsum        += velec;
 208             vgbsum          += vgb;
 209
 210             fscal            = felec;
 211
 212             /* Calculate temporary vectorial force */
 213             tx               = fscal*dx00;
 214             ty               = fscal*dy00;
 215             tz               = fscal*dz00;
 216
 217             /* Update vectorial force */
 218             fix0            += tx;
 219             fiy0            += ty;
 220             fiz0            += tz;
 221             f[j_coord_offset+DIM*0+XX] -= tx;
 222             f[j_coord_offset+DIM*0+YY] -= ty;
 223             f[j_coord_offset+DIM*0+ZZ] -= tz;
 224
 225             /* Inner loop uses 58 flops */
 226         }
 227         /* End of innermost loop */
 228
 229         tx = ty = tz = 0;
 230         f[i_coord_offset+DIM*0+XX] += fix0;
 231         f[i_coord_offset+DIM*0+YY] += fiy0;
 232         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 233         tx                         += fix0;
 234         ty                         += fiy0;
 235         tz                         += fiz0;
 236         fshift[i_shift_offset+XX]  += tx;
 237         fshift[i_shift_offset+YY]  += ty;
 238         fshift[i_shift_offset+ZZ]  += tz;
 239
 240         ggid                        = gid[iidx];
 241         /* Update potential energies */
 242         kernel_data->energygrp_elec[ggid] += velecsum;
 243         kernel_data->energygrp_polarization[ggid] += vgbsum;
 244         dvda[inr]                   = dvda[inr] + dvdasum*isai0*isai0;
 245
 246         /* Increment number of inner iterations */
 247         inneriter                  += j_index_end - j_index_start;
 248
 249         /* Outer loop uses 15 flops */
 250     }
 251
 252     /* Increment number of outer iterations */
 253     outeriter        += nri;
 254
 255     /* Update outer/inner flops */
 256
 257     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VF,outeriter*15 + inneriter*58);
 258 }
 259 /*
 260  * Gromacs nonbonded kernel:   nb_kernel_ElecGB_VdwNone_GeomP1P1_F_c
 261  * Electrostatics interaction: GeneralizedBorn
 262  * VdW interaction:            None
 263  * Geometry:                   Particle-Particle
 264  * Calculate force/pot:        Force
 265  */
 266 void
 267 nb_kernel_ElecGB_VdwNone_GeomP1P1_F_c
 268                     (t_nblist                    * gmx_restrict       nlist,
 269                      rvec                        * gmx_restrict          xx,
 270                      rvec                        * gmx_restrict          ff,
 271                      t_forcerec                  * gmx_restrict          fr,
 272                      t_mdatoms                   * gmx_restrict     mdatoms,
 273                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 274                      t_nrnb                      * gmx_restrict        nrnb)
 275 {
 276     int              i_shift_offset,i_coord_offset,j_coord_offset;
 277     int              j_index_start,j_index_end;
 278     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 279     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 280     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 281     real             *shiftvec,*fshift,*x,*f;
 282     int              vdwioffset0;
 283     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 284     int              vdwjidx0;
 285     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 286     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 287     real             velec,felec,velecsum,facel,crf,krf,krf2;
 288     real             *charge;
 289     int              gbitab;
 290     real             vgb,fgb,vgbsum,dvdasum,gbscale,gbtabscale,isaprod,gbqqfactor,gbinvepsdiff,dvdaj,gbeps,dvdatmp;
 291     real             *invsqrta,*dvda,*gbtab;
 292     int              vfitab;
 293     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
 294     real             *vftab;
 295
 296     x                = xx[0];
 297     f                = ff[0];
 298
 299     nri              = nlist->nri;
 300     iinr             = nlist->iinr;
 301     jindex           = nlist->jindex;
 302     jjnr             = nlist->jjnr;
 303     shiftidx         = nlist->shift;
 304     gid              = nlist->gid;
 305     shiftvec         = fr->shift_vec[0];
 306     fshift           = fr->fshift[0];
 307     facel            = fr->epsfac;
 308     charge           = mdatoms->chargeA;
 309
 310     invsqrta         = fr->invsqrta;
 311     dvda             = fr->dvda;
 312     gbtabscale       = fr->gbtab.scale;
 313     gbtab            = fr->gbtab.data;
 314     gbinvepsdiff     = (1.0/fr->epsilon_r) - (1.0/fr->gb_epsilon_solvent);
 315
 316     outeriter        = 0;
 317     inneriter        = 0;
 318
 319     /* Start outer loop over neighborlists */
 320     for(iidx=0; iidx<nri; iidx++)
 321     {
 322         /* Load shift vector for this list */
 323         i_shift_offset   = DIM*shiftidx[iidx];
 324         shX              = shiftvec[i_shift_offset+XX];
 325         shY              = shiftvec[i_shift_offset+YY];
 326         shZ              = shiftvec[i_shift_offset+ZZ];
 327
 328         /* Load limits for loop over neighbors */
 329         j_index_start    = jindex[iidx];
 330         j_index_end      = jindex[iidx+1];
 331
 332         /* Get outer coordinate index */
 333         inr              = iinr[iidx];
 334         i_coord_offset   = DIM*inr;
 335
 336         /* Load i particle coords and add shift vector */
 337         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 338         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 339         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 340
 341         fix0             = 0.0;
 342         fiy0             = 0.0;
 343         fiz0             = 0.0;
 344
 345         /* Load parameters for i particles */
 346         iq0              = facel*charge[inr+0];
 347         isai0            = invsqrta[inr+0];
 348
 349         dvdasum          = 0.0;
 350
 351         /* Start inner kernel loop */
 352         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 353         {
 354             /* Get j neighbor index, and coordinate index */
 355             jnr              = jjnr[jidx];
 356             j_coord_offset   = DIM*jnr;
 357
 358             /* load j atom coordinates */
 359             jx0              = x[j_coord_offset+DIM*0+XX];
 360             jy0              = x[j_coord_offset+DIM*0+YY];
 361             jz0              = x[j_coord_offset+DIM*0+ZZ];
 362
 363             /* Calculate displacement vector */
 364             dx00             = ix0 - jx0;
 365             dy00             = iy0 - jy0;
 366             dz00             = iz0 - jz0;
 367
 368             /* Calculate squared distance and things based on it */
 369             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 370
 371             rinv00           = gmx_invsqrt(rsq00);
 372
 373             /* Load parameters for j particles */
 374             jq0              = charge[jnr+0];
 375             isaj0           = invsqrta[jnr+0];
 376
 377             /**************************
 378              * CALCULATE INTERACTIONS *
 379              **************************/
 380
 381             r00              = rsq00*rinv00;
 382
 383             qq00             = iq0*jq0;
 384
 385             /* GENERALIZED BORN AND COULOMB ELECTROSTATICS */
 386             isaprod          = isai0*isaj0;
 387             gbqqfactor       = isaprod*(-qq00)*gbinvepsdiff;
 388             gbscale          = isaprod*gbtabscale;
 389             dvdaj            = dvda[jnr+0];
 390
 391             /* Calculate generalized born table index - this is a separate table from the normal one,
 392              * but we use the same procedure by multiplying r with scale and truncating to integer.
 393              */
 394             rt               = r00*gbscale;
 395             gbitab           = rt;
 396             gbeps            = rt-gbitab;
 397             gbitab           = 4*gbitab;
 398
 399             Y                = gbtab[gbitab];
 400             F                = gbtab[gbitab+1];
 401             Geps             = gbeps*gbtab[gbitab+2];
 402             Heps2            = gbeps*gbeps*gbtab[gbitab+3];
 403             Fp               = F+Geps+Heps2;
 404             VV               = Y+gbeps*Fp;
 405             vgb              = gbqqfactor*VV;
 406
 407             FF               = Fp+Geps+2.0*Heps2;
 408             fgb              = gbqqfactor*FF*gbscale;
 409             dvdatmp          = -0.5*(vgb+fgb*r00);
 410             dvdasum          = dvdasum + dvdatmp;
 411             dvda[jnr]        = dvdaj+dvdatmp*isaj0*isaj0;
 412             velec            = qq00*rinv00;
 413             felec            = (velec*rinv00-fgb)*rinv00;
 414
 415             fscal            = felec;
 416
 417             /* Calculate temporary vectorial force */
 418             tx               = fscal*dx00;
 419             ty               = fscal*dy00;
 420             tz               = fscal*dz00;
 421
 422             /* Update vectorial force */
 423             fix0            += tx;
 424             fiy0            += ty;
 425             fiz0            += tz;
 426             f[j_coord_offset+DIM*0+XX] -= tx;
 427             f[j_coord_offset+DIM*0+YY] -= ty;
 428             f[j_coord_offset+DIM*0+ZZ] -= tz;
 429
 430             /* Inner loop uses 56 flops */
 431         }
 432         /* End of innermost loop */
 433
 434         tx = ty = tz = 0;
 435         f[i_coord_offset+DIM*0+XX] += fix0;
 436         f[i_coord_offset+DIM*0+YY] += fiy0;
 437         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 438         tx                         += fix0;
 439         ty                         += fiy0;
 440         tz                         += fiz0;
 441         fshift[i_shift_offset+XX]  += tx;
 442         fshift[i_shift_offset+YY]  += ty;
 443         fshift[i_shift_offset+ZZ]  += tz;
 444
 445         dvda[inr]                   = dvda[inr] + dvdasum*isai0*isai0;
 446
 447         /* Increment number of inner iterations */
 448         inneriter                  += j_index_end - j_index_start;
 449
 450         /* Outer loop uses 13 flops */
 451     }
 452
 453     /* Increment number of outer iterations */
 454     outeriter        += nri;
 455
 456     /* Update outer/inner flops */
 457
 458     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_F,outeriter*13 + inneriter*56);
 459 }