src/gromacs/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecEwSw_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 "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_ElecEwSw_VdwNone_GeomP1P1_VF_c
  51  * Electrostatics interaction: Ewald
  52  * VdW interaction:            None
  53  * Geometry:                   Particle-Particle
  54  * Calculate force/pot:        PotentialAndForce
  55  */
  56 void
  57 nb_kernel_ElecEwSw_VdwNone_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              ewitab;
  80     real             ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
  81     real             *ewtab;
  82     real             rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
  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     sh_ewald         = fr->ic->sh_ewald;
  99     ewtab            = fr->ic->tabq_coul_FDV0;
 100     ewtabscale       = fr->ic->tabq_scale;
 101     ewtabhalfspace   = 0.5/ewtabscale;
 102
 103     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 104     rcutoff          = fr->rcoulomb;
 105     rcutoff2         = rcutoff*rcutoff;
 106
 107     rswitch          = fr->rcoulomb_switch;
 108     /* Setup switch parameters */
 109     d                = rcutoff-rswitch;
 110     swV3             = -10.0/(d*d*d);
 111     swV4             =  15.0/(d*d*d*d);
 112     swV5             =  -6.0/(d*d*d*d*d);
 113     swF2             = -30.0/(d*d*d);
 114     swF3             =  60.0/(d*d*d*d);
 115     swF4             = -30.0/(d*d*d*d*d);
 116
 117     outeriter        = 0;
 118     inneriter        = 0;
 119
 120     /* Start outer loop over neighborlists */
 121     for(iidx=0; iidx<nri; iidx++)
 122     {
 123         /* Load shift vector for this list */
 124         i_shift_offset   = DIM*shiftidx[iidx];
 125         shX              = shiftvec[i_shift_offset+XX];
 126         shY              = shiftvec[i_shift_offset+YY];
 127         shZ              = shiftvec[i_shift_offset+ZZ];
 128
 129         /* Load limits for loop over neighbors */
 130         j_index_start    = jindex[iidx];
 131         j_index_end      = jindex[iidx+1];
 132
 133         /* Get outer coordinate index */
 134         inr              = iinr[iidx];
 135         i_coord_offset   = DIM*inr;
 136
 137         /* Load i particle coords and add shift vector */
 138         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 139         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 140         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 141
 142         fix0             = 0.0;
 143         fiy0             = 0.0;
 144         fiz0             = 0.0;
 145
 146         /* Load parameters for i particles */
 147         iq0              = facel*charge[inr+0];
 148
 149         /* Reset potential sums */
 150         velecsum         = 0.0;
 151
 152         /* Start inner kernel loop */
 153         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 154         {
 155             /* Get j neighbor index, and coordinate index */
 156             jnr              = jjnr[jidx];
 157             j_coord_offset   = DIM*jnr;
 158
 159             /* load j atom coordinates */
 160             jx0              = x[j_coord_offset+DIM*0+XX];
 161             jy0              = x[j_coord_offset+DIM*0+YY];
 162             jz0              = x[j_coord_offset+DIM*0+ZZ];
 163
 164             /* Calculate displacement vector */
 165             dx00             = ix0 - jx0;
 166             dy00             = iy0 - jy0;
 167             dz00             = iz0 - jz0;
 168
 169             /* Calculate squared distance and things based on it */
 170             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 171
 172             rinv00           = gmx_invsqrt(rsq00);
 173
 174             rinvsq00         = rinv00*rinv00;
 175
 176             /* Load parameters for j particles */
 177             jq0              = charge[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
 190             /* EWALD ELECTROSTATICS */
 191
 192             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 193             ewrt             = r00*ewtabscale;
 194             ewitab           = ewrt;
 195             eweps            = ewrt-ewitab;
 196             ewitab           = 4*ewitab;
 197             felec            = ewtab[ewitab]+eweps*ewtab[ewitab+1];
 198             velec            = qq00*(rinv00-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
 199             felec            = qq00*rinv00*(rinvsq00-felec);
 200
 201             d                = r00-rswitch;
 202             d                = (d>0.0) ? d : 0.0;
 203             d2               = d*d;
 204             sw               = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
 205
 206             dsw              = d2*(swF2+d*(swF3+d*swF4));
 207
 208             /* Evaluate switch function */
 209             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 210             felec            = felec*sw - rinv00*velec*dsw;
 211             velec           *= sw;
 212
 213             /* Update potential sums from outer loop */
 214             velecsum        += velec;
 215
 216             fscal            = felec;
 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 59 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
 252         /* Increment number of inner iterations */
 253         inneriter                  += j_index_end - j_index_start;
 254
 255         /* Outer loop uses 14 flops */
 256     }
 257
 258     /* Increment number of outer iterations */
 259     outeriter        += nri;
 260
 261     /* Update outer/inner flops */
 262
 263     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VF,outeriter*14 + inneriter*59);
 264 }
 265 /*
 266  * Gromacs nonbonded kernel:   nb_kernel_ElecEwSw_VdwNone_GeomP1P1_F_c
 267  * Electrostatics interaction: Ewald
 268  * VdW interaction:            None
 269  * Geometry:                   Particle-Particle
 270  * Calculate force/pot:        Force
 271  */
 272 void
 273 nb_kernel_ElecEwSw_VdwNone_GeomP1P1_F_c
 274                     (t_nblist                    * gmx_restrict       nlist,
 275                      rvec                        * gmx_restrict          xx,
 276                      rvec                        * gmx_restrict          ff,
 277                      t_forcerec                  * gmx_restrict          fr,
 278                      t_mdatoms                   * gmx_restrict     mdatoms,
 279                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 280                      t_nrnb                      * gmx_restrict        nrnb)
 281 {
 282     int              i_shift_offset,i_coord_offset,j_coord_offset;
 283     int              j_index_start,j_index_end;
 284     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 285     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 286     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 287     real             *shiftvec,*fshift,*x,*f;
 288     int              vdwioffset0;
 289     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 290     int              vdwjidx0;
 291     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 292     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 293     real             velec,felec,velecsum,facel,crf,krf,krf2;
 294     real             *charge;
 295     int              ewitab;
 296     real             ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
 297     real             *ewtab;
 298     real             rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
 299
 300     x                = xx[0];
 301     f                = ff[0];
 302
 303     nri              = nlist->nri;
 304     iinr             = nlist->iinr;
 305     jindex           = nlist->jindex;
 306     jjnr             = nlist->jjnr;
 307     shiftidx         = nlist->shift;
 308     gid              = nlist->gid;
 309     shiftvec         = fr->shift_vec[0];
 310     fshift           = fr->fshift[0];
 311     facel            = fr->epsfac;
 312     charge           = mdatoms->chargeA;
 313
 314     sh_ewald         = fr->ic->sh_ewald;
 315     ewtab            = fr->ic->tabq_coul_FDV0;
 316     ewtabscale       = fr->ic->tabq_scale;
 317     ewtabhalfspace   = 0.5/ewtabscale;
 318
 319     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 320     rcutoff          = fr->rcoulomb;
 321     rcutoff2         = rcutoff*rcutoff;
 322
 323     rswitch          = fr->rcoulomb_switch;
 324     /* Setup switch parameters */
 325     d                = rcutoff-rswitch;
 326     swV3             = -10.0/(d*d*d);
 327     swV4             =  15.0/(d*d*d*d);
 328     swV5             =  -6.0/(d*d*d*d*d);
 329     swF2             = -30.0/(d*d*d);
 330     swF3             =  60.0/(d*d*d*d);
 331     swF4             = -30.0/(d*d*d*d*d);
 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
 365         /* Start inner kernel loop */
 366         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 367         {
 368             /* Get j neighbor index, and coordinate index */
 369             jnr              = jjnr[jidx];
 370             j_coord_offset   = DIM*jnr;
 371
 372             /* load j atom coordinates */
 373             jx0              = x[j_coord_offset+DIM*0+XX];
 374             jy0              = x[j_coord_offset+DIM*0+YY];
 375             jz0              = x[j_coord_offset+DIM*0+ZZ];
 376
 377             /* Calculate displacement vector */
 378             dx00             = ix0 - jx0;
 379             dy00             = iy0 - jy0;
 380             dz00             = iz0 - jz0;
 381
 382             /* Calculate squared distance and things based on it */
 383             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 384
 385             rinv00           = gmx_invsqrt(rsq00);
 386
 387             rinvsq00         = rinv00*rinv00;
 388
 389             /* Load parameters for j particles */
 390             jq0              = charge[jnr+0];
 391
 392             /**************************
 393              * CALCULATE INTERACTIONS *
 394              **************************/
 395
 396             if (rsq00<rcutoff2)
 397             {
 398
 399             r00              = rsq00*rinv00;
 400
 401             qq00             = iq0*jq0;
 402
 403             /* EWALD ELECTROSTATICS */
 404
 405             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 406             ewrt             = r00*ewtabscale;
 407             ewitab           = ewrt;
 408             eweps            = ewrt-ewitab;
 409             ewitab           = 4*ewitab;
 410             felec            = ewtab[ewitab]+eweps*ewtab[ewitab+1];
 411             velec            = qq00*(rinv00-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
 412             felec            = qq00*rinv00*(rinvsq00-felec);
 413
 414             d                = r00-rswitch;
 415             d                = (d>0.0) ? d : 0.0;
 416             d2               = d*d;
 417             sw               = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
 418
 419             dsw              = d2*(swF2+d*(swF3+d*swF4));
 420
 421             /* Evaluate switch function */
 422             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 423             felec            = felec*sw - rinv00*velec*dsw;
 424
 425             fscal            = felec;
 426
 427             /* Calculate temporary vectorial force */
 428             tx               = fscal*dx00;
 429             ty               = fscal*dy00;
 430             tz               = fscal*dz00;
 431
 432             /* Update vectorial force */
 433             fix0            += tx;
 434             fiy0            += ty;
 435             fiz0            += tz;
 436             f[j_coord_offset+DIM*0+XX] -= tx;
 437             f[j_coord_offset+DIM*0+YY] -= ty;
 438             f[j_coord_offset+DIM*0+ZZ] -= tz;
 439
 440             }
 441
 442             /* Inner loop uses 57 flops */
 443         }
 444         /* End of innermost loop */
 445
 446         tx = ty = tz = 0;
 447         f[i_coord_offset+DIM*0+XX] += fix0;
 448         f[i_coord_offset+DIM*0+YY] += fiy0;
 449         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 450         tx                         += fix0;
 451         ty                         += fiy0;
 452         tz                         += fiz0;
 453         fshift[i_shift_offset+XX]  += tx;
 454         fshift[i_shift_offset+YY]  += ty;
 455         fshift[i_shift_offset+ZZ]  += tz;
 456
 457         /* Increment number of inner iterations */
 458         inneriter                  += j_index_end - j_index_start;
 459
 460         /* Outer loop uses 13 flops */
 461     }
 462
 463     /* Increment number of outer iterations */
 464     outeriter        += nri;
 465
 466     /* Update outer/inner flops */
 467
 468     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_F,outeriter*13 + inneriter*57);
 469 }