src/gromacs/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecEwSh_VdwBhamSh_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_ElecEwSh_VdwBhamSh_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_ElecEwSh_VdwBhamSh_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
  87     x                = xx[0];
  88     f                = ff[0];
  89
  90     nri              = nlist->nri;
  91     iinr             = nlist->iinr;
  92     jindex           = nlist->jindex;
  93     jjnr             = nlist->jjnr;
  94     shiftidx         = nlist->shift;
  95     gid              = nlist->gid;
  96     shiftvec         = fr->shift_vec[0];
  97     fshift           = fr->fshift[0];
  98     facel            = fr->epsfac;
  99     charge           = mdatoms->chargeA;
 100     nvdwtype         = fr->ntype;
 101     vdwparam         = fr->nbfp;
 102     vdwtype          = mdatoms->typeA;
 103
 104     sh_ewald         = fr->ic->sh_ewald;
 105     ewtab            = fr->ic->tabq_coul_FDV0;
 106     ewtabscale       = fr->ic->tabq_scale;
 107     ewtabhalfspace   = 0.5/ewtabscale;
 108
 109     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 110     rcutoff          = fr->rcoulomb;
 111     rcutoff2         = rcutoff*rcutoff;
 112
 113     sh_vdw_invrcut6  = fr->ic->sh_invrc6;
 114     rvdw             = fr->rvdw;
 115
 116     outeriter        = 0;
 117     inneriter        = 0;
 118
 119     /* Start outer loop over neighborlists */
 120     for(iidx=0; iidx<nri; iidx++)
 121     {
 122         /* Load shift vector for this list */
 123         i_shift_offset   = DIM*shiftidx[iidx];
 124         shX              = shiftvec[i_shift_offset+XX];
 125         shY              = shiftvec[i_shift_offset+YY];
 126         shZ              = shiftvec[i_shift_offset+ZZ];
 127
 128         /* Load limits for loop over neighbors */
 129         j_index_start    = jindex[iidx];
 130         j_index_end      = jindex[iidx+1];
 131
 132         /* Get outer coordinate index */
 133         inr              = iinr[iidx];
 134         i_coord_offset   = DIM*inr;
 135
 136         /* Load i particle coords and add shift vector */
 137         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 138         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 139         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 140
 141         fix0             = 0.0;
 142         fiy0             = 0.0;
 143         fiz0             = 0.0;
 144
 145         /* Load parameters for i particles */
 146         iq0              = facel*charge[inr+0];
 147         vdwioffset0      = 3*nvdwtype*vdwtype[inr+0];
 148
 149         /* Reset potential sums */
 150         velecsum         = 0.0;
 151         vvdwsum          = 0.0;
 152
 153         /* Start inner kernel loop */
 154         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 155         {
 156             /* Get j neighbor index, and coordinate index */
 157             jnr              = jjnr[jidx];
 158             j_coord_offset   = DIM*jnr;
 159
 160             /* load j atom coordinates */
 161             jx0              = x[j_coord_offset+DIM*0+XX];
 162             jy0              = x[j_coord_offset+DIM*0+YY];
 163             jz0              = x[j_coord_offset+DIM*0+ZZ];
 164
 165             /* Calculate displacement vector */
 166             dx00             = ix0 - jx0;
 167             dy00             = iy0 - jy0;
 168             dz00             = iz0 - jz0;
 169
 170             /* Calculate squared distance and things based on it */
 171             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 172
 173             rinv00           = gmx_invsqrt(rsq00);
 174
 175             rinvsq00         = rinv00*rinv00;
 176
 177             /* Load parameters for j particles */
 178             jq0              = charge[jnr+0];
 179             vdwjidx0         = 3*vdwtype[jnr+0];
 180
 181             /**************************
 182              * CALCULATE INTERACTIONS *
 183              **************************/
 184
 185             if (rsq00<rcutoff2)
 186             {
 187
 188             r00              = rsq00*rinv00;
 189
 190             qq00             = iq0*jq0;
 191             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 192             cexp1_00         = vdwparam[vdwioffset0+vdwjidx0+1];
 193             cexp2_00         = vdwparam[vdwioffset0+vdwjidx0+2];
 194
 195             /* EWALD ELECTROSTATICS */
 196
 197             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 198             ewrt             = r00*ewtabscale;
 199             ewitab           = ewrt;
 200             eweps            = ewrt-ewitab;
 201             ewitab           = 4*ewitab;
 202             felec            = ewtab[ewitab]+eweps*ewtab[ewitab+1];
 203             velec            = qq00*((rinv00-sh_ewald)-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
 204             felec            = qq00*rinv00*(rinvsq00-felec);
 205
 206             /* BUCKINGHAM DISPERSION/REPULSION */
 207             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 208             vvdw6            = c6_00*rinvsix;
 209             br               = cexp2_00*r00;
 210             vvdwexp          = cexp1_00*exp(-br);
 211             vvdw             = (vvdwexp-cexp1_00*exp(-cexp2_00*rvdw)) - (vvdw6 - c6_00*sh_vdw_invrcut6)*(1.0/6.0);
 212             fvdw             = (br*vvdwexp-vvdw6)*rinvsq00;
 213
 214             /* Update potential sums from outer loop */
 215             velecsum        += velec;
 216             vvdwsum         += vvdw;
 217
 218             fscal            = felec+fvdw;
 219
 220             /* Calculate temporary vectorial force */
 221             tx               = fscal*dx00;
 222             ty               = fscal*dy00;
 223             tz               = fscal*dz00;
 224
 225             /* Update vectorial force */
 226             fix0            += tx;
 227             fiy0            += ty;
 228             fiz0            += tz;
 229             f[j_coord_offset+DIM*0+XX] -= tx;
 230             f[j_coord_offset+DIM*0+YY] -= ty;
 231             f[j_coord_offset+DIM*0+ZZ] -= tz;
 232
 233             }
 234
 235             /* Inner loop uses 111 flops */
 236         }
 237         /* End of innermost loop */
 238
 239         tx = ty = tz = 0;
 240         f[i_coord_offset+DIM*0+XX] += fix0;
 241         f[i_coord_offset+DIM*0+YY] += fiy0;
 242         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 243         tx                         += fix0;
 244         ty                         += fiy0;
 245         tz                         += fiz0;
 246         fshift[i_shift_offset+XX]  += tx;
 247         fshift[i_shift_offset+YY]  += ty;
 248         fshift[i_shift_offset+ZZ]  += tz;
 249
 250         ggid                        = gid[iidx];
 251         /* Update potential energies */
 252         kernel_data->energygrp_elec[ggid] += velecsum;
 253         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 254
 255         /* Increment number of inner iterations */
 256         inneriter                  += j_index_end - j_index_start;
 257
 258         /* Outer loop uses 15 flops */
 259     }
 260
 261     /* Increment number of outer iterations */
 262     outeriter        += nri;
 263
 264     /* Update outer/inner flops */
 265
 266     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_VF,outeriter*15 + inneriter*111);
 267 }
 268 /*
 269  * Gromacs nonbonded kernel:   nb_kernel_ElecEwSh_VdwBhamSh_GeomP1P1_F_c
 270  * Electrostatics interaction: Ewald
 271  * VdW interaction:            Buckingham
 272  * Geometry:                   Particle-Particle
 273  * Calculate force/pot:        Force
 274  */
 275 void
 276 nb_kernel_ElecEwSh_VdwBhamSh_GeomP1P1_F_c
 277                     (t_nblist                    * gmx_restrict       nlist,
 278                      rvec                        * gmx_restrict          xx,
 279                      rvec                        * gmx_restrict          ff,
 280                      t_forcerec                  * gmx_restrict          fr,
 281                      t_mdatoms                   * gmx_restrict     mdatoms,
 282                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 283                      t_nrnb                      * gmx_restrict        nrnb)
 284 {
 285     int              i_shift_offset,i_coord_offset,j_coord_offset;
 286     int              j_index_start,j_index_end;
 287     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 288     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 289     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 290     real             *shiftvec,*fshift,*x,*f;
 291     int              vdwioffset0;
 292     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 293     int              vdwjidx0;
 294     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 295     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 296     real             velec,felec,velecsum,facel,crf,krf,krf2;
 297     real             *charge;
 298     int              nvdwtype;
 299     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 300     int              *vdwtype;
 301     real             *vdwparam;
 302     int              ewitab;
 303     real             ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
 304     real             *ewtab;
 305
 306     x                = xx[0];
 307     f                = ff[0];
 308
 309     nri              = nlist->nri;
 310     iinr             = nlist->iinr;
 311     jindex           = nlist->jindex;
 312     jjnr             = nlist->jjnr;
 313     shiftidx         = nlist->shift;
 314     gid              = nlist->gid;
 315     shiftvec         = fr->shift_vec[0];
 316     fshift           = fr->fshift[0];
 317     facel            = fr->epsfac;
 318     charge           = mdatoms->chargeA;
 319     nvdwtype         = fr->ntype;
 320     vdwparam         = fr->nbfp;
 321     vdwtype          = mdatoms->typeA;
 322
 323     sh_ewald         = fr->ic->sh_ewald;
 324     ewtab            = fr->ic->tabq_coul_F;
 325     ewtabscale       = fr->ic->tabq_scale;
 326     ewtabhalfspace   = 0.5/ewtabscale;
 327
 328     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 329     rcutoff          = fr->rcoulomb;
 330     rcutoff2         = rcutoff*rcutoff;
 331
 332     sh_vdw_invrcut6  = fr->ic->sh_invrc6;
 333     rvdw             = fr->rvdw;
 334
 335     outeriter        = 0;
 336     inneriter        = 0;
 337
 338     /* Start outer loop over neighborlists */
 339     for(iidx=0; iidx<nri; iidx++)
 340     {
 341         /* Load shift vector for this list */
 342         i_shift_offset   = DIM*shiftidx[iidx];
 343         shX              = shiftvec[i_shift_offset+XX];
 344         shY              = shiftvec[i_shift_offset+YY];
 345         shZ              = shiftvec[i_shift_offset+ZZ];
 346
 347         /* Load limits for loop over neighbors */
 348         j_index_start    = jindex[iidx];
 349         j_index_end      = jindex[iidx+1];
 350
 351         /* Get outer coordinate index */
 352         inr              = iinr[iidx];
 353         i_coord_offset   = DIM*inr;
 354
 355         /* Load i particle coords and add shift vector */
 356         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 357         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 358         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 359
 360         fix0             = 0.0;
 361         fiy0             = 0.0;
 362         fiz0             = 0.0;
 363
 364         /* Load parameters for i particles */
 365         iq0              = facel*charge[inr+0];
 366         vdwioffset0      = 3*nvdwtype*vdwtype[inr+0];
 367
 368         /* Start inner kernel loop */
 369         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 370         {
 371             /* Get j neighbor index, and coordinate index */
 372             jnr              = jjnr[jidx];
 373             j_coord_offset   = DIM*jnr;
 374
 375             /* load j atom coordinates */
 376             jx0              = x[j_coord_offset+DIM*0+XX];
 377             jy0              = x[j_coord_offset+DIM*0+YY];
 378             jz0              = x[j_coord_offset+DIM*0+ZZ];
 379
 380             /* Calculate displacement vector */
 381             dx00             = ix0 - jx0;
 382             dy00             = iy0 - jy0;
 383             dz00             = iz0 - jz0;
 384
 385             /* Calculate squared distance and things based on it */
 386             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 387
 388             rinv00           = gmx_invsqrt(rsq00);
 389
 390             rinvsq00         = rinv00*rinv00;
 391
 392             /* Load parameters for j particles */
 393             jq0              = charge[jnr+0];
 394             vdwjidx0         = 3*vdwtype[jnr+0];
 395
 396             /**************************
 397              * CALCULATE INTERACTIONS *
 398              **************************/
 399
 400             if (rsq00<rcutoff2)
 401             {
 402
 403             r00              = rsq00*rinv00;
 404
 405             qq00             = iq0*jq0;
 406             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 407             cexp1_00         = vdwparam[vdwioffset0+vdwjidx0+1];
 408             cexp2_00         = vdwparam[vdwioffset0+vdwjidx0+2];
 409
 410             /* EWALD ELECTROSTATICS */
 411
 412             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 413             ewrt             = r00*ewtabscale;
 414             ewitab           = ewrt;
 415             eweps            = ewrt-ewitab;
 416             felec            = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1];
 417             felec            = qq00*rinv00*(rinvsq00-felec);
 418
 419             /* BUCKINGHAM DISPERSION/REPULSION */
 420             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 421             vvdw6            = c6_00*rinvsix;
 422             br               = cexp2_00*r00;
 423             vvdwexp          = cexp1_00*exp(-br);
 424             fvdw             = (br*vvdwexp-vvdw6)*rinvsq00;
 425
 426             fscal            = felec+fvdw;
 427
 428             /* Calculate temporary vectorial force */
 429             tx               = fscal*dx00;
 430             ty               = fscal*dy00;
 431             tz               = fscal*dz00;
 432
 433             /* Update vectorial force */
 434             fix0            += tx;
 435             fiy0            += ty;
 436             fiz0            += tz;
 437             f[j_coord_offset+DIM*0+XX] -= tx;
 438             f[j_coord_offset+DIM*0+YY] -= ty;
 439             f[j_coord_offset+DIM*0+ZZ] -= tz;
 440
 441             }
 442
 443             /* Inner loop uses 69 flops */
 444         }
 445         /* End of innermost loop */
 446
 447         tx = ty = tz = 0;
 448         f[i_coord_offset+DIM*0+XX] += fix0;
 449         f[i_coord_offset+DIM*0+YY] += fiy0;
 450         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 451         tx                         += fix0;
 452         ty                         += fiy0;
 453         tz                         += fiz0;
 454         fshift[i_shift_offset+XX]  += tx;
 455         fshift[i_shift_offset+YY]  += ty;
 456         fshift[i_shift_offset+ZZ]  += tz;
 457
 458         /* Increment number of inner iterations */
 459         inneriter                  += j_index_end - j_index_start;
 460
 461         /* Outer loop uses 13 flops */
 462     }
 463
 464     /* Increment number of outer iterations */
 465     outeriter        += nri;
 466
 467     /* Update outer/inner flops */
 468
 469     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_F,outeriter*13 + inneriter*69);
 470 }