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 "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_ElecEwSh_VdwBhamSh_GeomP1P1_VF_c
  49  * Electrostatics interaction: Ewald
  50  * VdW interaction:            Buckingham
  51  * Geometry:                   Particle-Particle
  52  * Calculate force/pot:        PotentialAndForce
  53  */
  54 void
  55 nb_kernel_ElecEwSh_VdwBhamSh_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              nvdwtype;
  78     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
  79     int              *vdwtype;
  80     real             *vdwparam;
  81     int              ewitab;
  82     real             ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
  83     real             *ewtab;
  84
  85     x                = xx[0];
  86     f                = ff[0];
  87
  88     nri              = nlist->nri;
  89     iinr             = nlist->iinr;
  90     jindex           = nlist->jindex;
  91     jjnr             = nlist->jjnr;
  92     shiftidx         = nlist->shift;
  93     gid              = nlist->gid;
  94     shiftvec         = fr->shift_vec[0];
  95     fshift           = fr->fshift[0];
  96     facel            = fr->epsfac;
  97     charge           = mdatoms->chargeA;
  98     nvdwtype         = fr->ntype;
  99     vdwparam         = fr->nbfp;
 100     vdwtype          = mdatoms->typeA;
 101
 102     sh_ewald         = fr->ic->sh_ewald;
 103     ewtab            = fr->ic->tabq_coul_FDV0;
 104     ewtabscale       = fr->ic->tabq_scale;
 105     ewtabhalfspace   = 0.5/ewtabscale;
 106
 107     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 108     rcutoff          = fr->rcoulomb;
 109     rcutoff2         = rcutoff*rcutoff;
 110
 111     sh_vdw_invrcut6  = fr->ic->sh_invrc6;
 112     rvdw             = fr->rvdw;
 113
 114     outeriter        = 0;
 115     inneriter        = 0;
 116
 117     /* Start outer loop over neighborlists */
 118     for(iidx=0; iidx<nri; iidx++)
 119     {
 120         /* Load shift vector for this list */
 121         i_shift_offset   = DIM*shiftidx[iidx];
 122         shX              = shiftvec[i_shift_offset+XX];
 123         shY              = shiftvec[i_shift_offset+YY];
 124         shZ              = shiftvec[i_shift_offset+ZZ];
 125
 126         /* Load limits for loop over neighbors */
 127         j_index_start    = jindex[iidx];
 128         j_index_end      = jindex[iidx+1];
 129
 130         /* Get outer coordinate index */
 131         inr              = iinr[iidx];
 132         i_coord_offset   = DIM*inr;
 133
 134         /* Load i particle coords and add shift vector */
 135         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 136         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 137         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 138
 139         fix0             = 0.0;
 140         fiy0             = 0.0;
 141         fiz0             = 0.0;
 142
 143         /* Load parameters for i particles */
 144         iq0              = facel*charge[inr+0];
 145         vdwioffset0      = 3*nvdwtype*vdwtype[inr+0];
 146
 147         /* Reset potential sums */
 148         velecsum         = 0.0;
 149         vvdwsum          = 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             vdwjidx0         = 3*vdwtype[jnr+0];
 178
 179             /**************************
 180              * CALCULATE INTERACTIONS *
 181              **************************/
 182
 183             if (rsq00<rcutoff2)
 184             {
 185
 186             r00              = rsq00*rinv00;
 187
 188             qq00             = iq0*jq0;
 189             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 190             cexp1_00         = vdwparam[vdwioffset0+vdwjidx0+1];
 191             cexp2_00         = vdwparam[vdwioffset0+vdwjidx0+2];
 192
 193             /* EWALD ELECTROSTATICS */
 194
 195             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 196             ewrt             = r00*ewtabscale;
 197             ewitab           = ewrt;
 198             eweps            = ewrt-ewitab;
 199             ewitab           = 4*ewitab;
 200             felec            = ewtab[ewitab]+eweps*ewtab[ewitab+1];
 201             velec            = qq00*((rinv00-sh_ewald)-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
 202             felec            = qq00*rinv00*(rinvsq00-felec);
 203
 204             /* BUCKINGHAM DISPERSION/REPULSION */
 205             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 206             vvdw6            = c6_00*rinvsix;
 207             br               = cexp2_00*r00;
 208             vvdwexp          = cexp1_00*exp(-br);
 209             vvdw             = (vvdwexp-cexp1_00*exp(-cexp2_00*rvdw)) - (vvdw6 - c6_00*sh_vdw_invrcut6)*(1.0/6.0);
 210             fvdw             = (br*vvdwexp-vvdw6)*rinvsq00;
 211
 212             /* Update potential sums from outer loop */
 213             velecsum        += velec;
 214             vvdwsum         += vvdw;
 215
 216             fscal            = felec+fvdw;
 217
 218             /* Calculate temporary vectorial force */
 219             tx               = fscal*dx00;
 220             ty               = fscal*dy00;
 221             tz               = fscal*dz00;
 222
 223             /* Update vectorial force */
 224             fix0            += tx;
 225             fiy0            += ty;
 226             fiz0            += tz;
 227             f[j_coord_offset+DIM*0+XX] -= tx;
 228             f[j_coord_offset+DIM*0+YY] -= ty;
 229             f[j_coord_offset+DIM*0+ZZ] -= tz;
 230
 231             }
 232
 233             /* Inner loop uses 111 flops */
 234         }
 235         /* End of innermost loop */
 236
 237         tx = ty = tz = 0;
 238         f[i_coord_offset+DIM*0+XX] += fix0;
 239         f[i_coord_offset+DIM*0+YY] += fiy0;
 240         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 241         tx                         += fix0;
 242         ty                         += fiy0;
 243         tz                         += fiz0;
 244         fshift[i_shift_offset+XX]  += tx;
 245         fshift[i_shift_offset+YY]  += ty;
 246         fshift[i_shift_offset+ZZ]  += tz;
 247
 248         ggid                        = gid[iidx];
 249         /* Update potential energies */
 250         kernel_data->energygrp_elec[ggid] += velecsum;
 251         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 252
 253         /* Increment number of inner iterations */
 254         inneriter                  += j_index_end - j_index_start;
 255
 256         /* Outer loop uses 15 flops */
 257     }
 258
 259     /* Increment number of outer iterations */
 260     outeriter        += nri;
 261
 262     /* Update outer/inner flops */
 263
 264     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_VF,outeriter*15 + inneriter*111);
 265 }
 266 /*
 267  * Gromacs nonbonded kernel:   nb_kernel_ElecEwSh_VdwBhamSh_GeomP1P1_F_c
 268  * Electrostatics interaction: Ewald
 269  * VdW interaction:            Buckingham
 270  * Geometry:                   Particle-Particle
 271  * Calculate force/pot:        Force
 272  */
 273 void
 274 nb_kernel_ElecEwSh_VdwBhamSh_GeomP1P1_F_c
 275                     (t_nblist                    * gmx_restrict       nlist,
 276                      rvec                        * gmx_restrict          xx,
 277                      rvec                        * gmx_restrict          ff,
 278                      t_forcerec                  * gmx_restrict          fr,
 279                      t_mdatoms                   * gmx_restrict     mdatoms,
 280                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 281                      t_nrnb                      * gmx_restrict        nrnb)
 282 {
 283     int              i_shift_offset,i_coord_offset,j_coord_offset;
 284     int              j_index_start,j_index_end;
 285     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 286     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 287     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 288     real             *shiftvec,*fshift,*x,*f;
 289     int              vdwioffset0;
 290     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 291     int              vdwjidx0;
 292     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 293     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 294     real             velec,felec,velecsum,facel,crf,krf,krf2;
 295     real             *charge;
 296     int              nvdwtype;
 297     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 298     int              *vdwtype;
 299     real             *vdwparam;
 300     int              ewitab;
 301     real             ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
 302     real             *ewtab;
 303
 304     x                = xx[0];
 305     f                = ff[0];
 306
 307     nri              = nlist->nri;
 308     iinr             = nlist->iinr;
 309     jindex           = nlist->jindex;
 310     jjnr             = nlist->jjnr;
 311     shiftidx         = nlist->shift;
 312     gid              = nlist->gid;
 313     shiftvec         = fr->shift_vec[0];
 314     fshift           = fr->fshift[0];
 315     facel            = fr->epsfac;
 316     charge           = mdatoms->chargeA;
 317     nvdwtype         = fr->ntype;
 318     vdwparam         = fr->nbfp;
 319     vdwtype          = mdatoms->typeA;
 320
 321     sh_ewald         = fr->ic->sh_ewald;
 322     ewtab            = fr->ic->tabq_coul_F;
 323     ewtabscale       = fr->ic->tabq_scale;
 324     ewtabhalfspace   = 0.5/ewtabscale;
 325
 326     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 327     rcutoff          = fr->rcoulomb;
 328     rcutoff2         = rcutoff*rcutoff;
 329
 330     sh_vdw_invrcut6  = fr->ic->sh_invrc6;
 331     rvdw             = fr->rvdw;
 332
 333     outeriter        = 0;
 334     inneriter        = 0;
 335
 336     /* Start outer loop over neighborlists */
 337     for(iidx=0; iidx<nri; iidx++)
 338     {
 339         /* Load shift vector for this list */
 340         i_shift_offset   = DIM*shiftidx[iidx];
 341         shX              = shiftvec[i_shift_offset+XX];
 342         shY              = shiftvec[i_shift_offset+YY];
 343         shZ              = shiftvec[i_shift_offset+ZZ];
 344
 345         /* Load limits for loop over neighbors */
 346         j_index_start    = jindex[iidx];
 347         j_index_end      = jindex[iidx+1];
 348
 349         /* Get outer coordinate index */
 350         inr              = iinr[iidx];
 351         i_coord_offset   = DIM*inr;
 352
 353         /* Load i particle coords and add shift vector */
 354         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 355         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 356         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 357
 358         fix0             = 0.0;
 359         fiy0             = 0.0;
 360         fiz0             = 0.0;
 361
 362         /* Load parameters for i particles */
 363         iq0              = facel*charge[inr+0];
 364         vdwioffset0      = 3*nvdwtype*vdwtype[inr+0];
 365
 366         /* Start inner kernel loop */
 367         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 368         {
 369             /* Get j neighbor index, and coordinate index */
 370             jnr              = jjnr[jidx];
 371             j_coord_offset   = DIM*jnr;
 372
 373             /* load j atom coordinates */
 374             jx0              = x[j_coord_offset+DIM*0+XX];
 375             jy0              = x[j_coord_offset+DIM*0+YY];
 376             jz0              = x[j_coord_offset+DIM*0+ZZ];
 377
 378             /* Calculate displacement vector */
 379             dx00             = ix0 - jx0;
 380             dy00             = iy0 - jy0;
 381             dz00             = iz0 - jz0;
 382
 383             /* Calculate squared distance and things based on it */
 384             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 385
 386             rinv00           = gmx_invsqrt(rsq00);
 387
 388             rinvsq00         = rinv00*rinv00;
 389
 390             /* Load parameters for j particles */
 391             jq0              = charge[jnr+0];
 392             vdwjidx0         = 3*vdwtype[jnr+0];
 393
 394             /**************************
 395              * CALCULATE INTERACTIONS *
 396              **************************/
 397
 398             if (rsq00<rcutoff2)
 399             {
 400
 401             r00              = rsq00*rinv00;
 402
 403             qq00             = iq0*jq0;
 404             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 405             cexp1_00         = vdwparam[vdwioffset0+vdwjidx0+1];
 406             cexp2_00         = vdwparam[vdwioffset0+vdwjidx0+2];
 407
 408             /* EWALD ELECTROSTATICS */
 409
 410             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 411             ewrt             = r00*ewtabscale;
 412             ewitab           = ewrt;
 413             eweps            = ewrt-ewitab;
 414             felec            = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1];
 415             felec            = qq00*rinv00*(rinvsq00-felec);
 416
 417             /* BUCKINGHAM DISPERSION/REPULSION */
 418             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
 419             vvdw6            = c6_00*rinvsix;
 420             br               = cexp2_00*r00;
 421             vvdwexp          = cexp1_00*exp(-br);
 422             fvdw             = (br*vvdwexp-vvdw6)*rinvsq00;
 423
 424             fscal            = felec+fvdw;
 425
 426             /* Calculate temporary vectorial force */
 427             tx               = fscal*dx00;
 428             ty               = fscal*dy00;
 429             tz               = fscal*dz00;
 430
 431             /* Update vectorial force */
 432             fix0            += tx;
 433             fiy0            += ty;
 434             fiz0            += tz;
 435             f[j_coord_offset+DIM*0+XX] -= tx;
 436             f[j_coord_offset+DIM*0+YY] -= ty;
 437             f[j_coord_offset+DIM*0+ZZ] -= tz;
 438
 439             }
 440
 441             /* Inner loop uses 69 flops */
 442         }
 443         /* End of innermost loop */
 444
 445         tx = ty = tz = 0;
 446         f[i_coord_offset+DIM*0+XX] += fix0;
 447         f[i_coord_offset+DIM*0+YY] += fiy0;
 448         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 449         tx                         += fix0;
 450         ty                         += fiy0;
 451         tz                         += fiz0;
 452         fshift[i_shift_offset+XX]  += tx;
 453         fshift[i_shift_offset+YY]  += ty;
 454         fshift[i_shift_offset+ZZ]  += tz;
 455
 456         /* Increment number of inner iterations */
 457         inneriter                  += j_index_end - j_index_start;
 458
 459         /* Outer loop uses 13 flops */
 460     }
 461
 462     /* Increment number of outer iterations */
 463     outeriter        += nri;
 464
 465     /* Update outer/inner flops */
 466
 467     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_F,outeriter*13 + inneriter*69);
 468 }