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 "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_ElecEwSw_VdwNone_GeomP1P1_VF_c
  49  * Electrostatics interaction: Ewald
  50  * VdW interaction:            None
  51  * Geometry:                   Particle-Particle
  52  * Calculate force/pot:        PotentialAndForce
  53  */
  54 void
  55 nb_kernel_ElecEwSw_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              ewitab;
  78     real             ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
  79     real             *ewtab;
  80     real             rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
  81
  82     x                = xx[0];
  83     f                = ff[0];
  84
  85     nri              = nlist->nri;
  86     iinr             = nlist->iinr;
  87     jindex           = nlist->jindex;
  88     jjnr             = nlist->jjnr;
  89     shiftidx         = nlist->shift;
  90     gid              = nlist->gid;
  91     shiftvec         = fr->shift_vec[0];
  92     fshift           = fr->fshift[0];
  93     facel            = fr->epsfac;
  94     charge           = mdatoms->chargeA;
  95
  96     sh_ewald         = fr->ic->sh_ewald;
  97     ewtab            = fr->ic->tabq_coul_FDV0;
  98     ewtabscale       = fr->ic->tabq_scale;
  99     ewtabhalfspace   = 0.5/ewtabscale;
 100
 101     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 102     rcutoff          = fr->rcoulomb;
 103     rcutoff2         = rcutoff*rcutoff;
 104
 105     rswitch          = fr->rcoulomb_switch;
 106     /* Setup switch parameters */
 107     d                = rcutoff-rswitch;
 108     swV3             = -10.0/(d*d*d);
 109     swV4             =  15.0/(d*d*d*d);
 110     swV5             =  -6.0/(d*d*d*d*d);
 111     swF2             = -30.0/(d*d*d);
 112     swF3             =  60.0/(d*d*d*d);
 113     swF4             = -30.0/(d*d*d*d*d);
 114
 115     outeriter        = 0;
 116     inneriter        = 0;
 117
 118     /* Start outer loop over neighborlists */
 119     for(iidx=0; iidx<nri; iidx++)
 120     {
 121         /* Load shift vector for this list */
 122         i_shift_offset   = DIM*shiftidx[iidx];
 123         shX              = shiftvec[i_shift_offset+XX];
 124         shY              = shiftvec[i_shift_offset+YY];
 125         shZ              = shiftvec[i_shift_offset+ZZ];
 126
 127         /* Load limits for loop over neighbors */
 128         j_index_start    = jindex[iidx];
 129         j_index_end      = jindex[iidx+1];
 130
 131         /* Get outer coordinate index */
 132         inr              = iinr[iidx];
 133         i_coord_offset   = DIM*inr;
 134
 135         /* Load i particle coords and add shift vector */
 136         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 137         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 138         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 139
 140         fix0             = 0.0;
 141         fiy0             = 0.0;
 142         fiz0             = 0.0;
 143
 144         /* Load parameters for i particles */
 145         iq0              = facel*charge[inr+0];
 146
 147         /* Reset potential sums */
 148         velecsum         = 0.0;
 149
 150         /* Start inner kernel loop */
 151         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 152         {
 153             /* Get j neighbor index, and coordinate index */
 154             jnr              = jjnr[jidx];
 155             j_coord_offset   = DIM*jnr;
 156
 157             /* load j atom coordinates */
 158             jx0              = x[j_coord_offset+DIM*0+XX];
 159             jy0              = x[j_coord_offset+DIM*0+YY];
 160             jz0              = x[j_coord_offset+DIM*0+ZZ];
 161
 162             /* Calculate displacement vector */
 163             dx00             = ix0 - jx0;
 164             dy00             = iy0 - jy0;
 165             dz00             = iz0 - jz0;
 166
 167             /* Calculate squared distance and things based on it */
 168             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 169
 170             rinv00           = gmx_invsqrt(rsq00);
 171
 172             rinvsq00         = rinv00*rinv00;
 173
 174             /* Load parameters for j particles */
 175             jq0              = charge[jnr+0];
 176
 177             /**************************
 178              * CALCULATE INTERACTIONS *
 179              **************************/
 180
 181             if (rsq00<rcutoff2)
 182             {
 183
 184             r00              = rsq00*rinv00;
 185
 186             qq00             = iq0*jq0;
 187
 188             /* EWALD ELECTROSTATICS */
 189
 190             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 191             ewrt             = r00*ewtabscale;
 192             ewitab           = ewrt;
 193             eweps            = ewrt-ewitab;
 194             ewitab           = 4*ewitab;
 195             felec            = ewtab[ewitab]+eweps*ewtab[ewitab+1];
 196             velec            = qq00*(rinv00-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
 197             felec            = qq00*rinv00*(rinvsq00-felec);
 198
 199             d                = r00-rswitch;
 200             d                = (d>0.0) ? d : 0.0;
 201             d2               = d*d;
 202             sw               = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
 203
 204             dsw              = d2*(swF2+d*(swF3+d*swF4));
 205
 206             /* Evaluate switch function */
 207             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 208             felec            = felec*sw - rinv00*velec*dsw;
 209             velec           *= sw;
 210
 211             /* Update potential sums from outer loop */
 212             velecsum        += velec;
 213
 214             fscal            = felec;
 215
 216             /* Calculate temporary vectorial force */
 217             tx               = fscal*dx00;
 218             ty               = fscal*dy00;
 219             tz               = fscal*dz00;
 220
 221             /* Update vectorial force */
 222             fix0            += tx;
 223             fiy0            += ty;
 224             fiz0            += tz;
 225             f[j_coord_offset+DIM*0+XX] -= tx;
 226             f[j_coord_offset+DIM*0+YY] -= ty;
 227             f[j_coord_offset+DIM*0+ZZ] -= tz;
 228
 229             }
 230
 231             /* Inner loop uses 59 flops */
 232         }
 233         /* End of innermost loop */
 234
 235         tx = ty = tz = 0;
 236         f[i_coord_offset+DIM*0+XX] += fix0;
 237         f[i_coord_offset+DIM*0+YY] += fiy0;
 238         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 239         tx                         += fix0;
 240         ty                         += fiy0;
 241         tz                         += fiz0;
 242         fshift[i_shift_offset+XX]  += tx;
 243         fshift[i_shift_offset+YY]  += ty;
 244         fshift[i_shift_offset+ZZ]  += tz;
 245
 246         ggid                        = gid[iidx];
 247         /* Update potential energies */
 248         kernel_data->energygrp_elec[ggid] += velecsum;
 249
 250         /* Increment number of inner iterations */
 251         inneriter                  += j_index_end - j_index_start;
 252
 253         /* Outer loop uses 14 flops */
 254     }
 255
 256     /* Increment number of outer iterations */
 257     outeriter        += nri;
 258
 259     /* Update outer/inner flops */
 260
 261     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VF,outeriter*14 + inneriter*59);
 262 }
 263 /*
 264  * Gromacs nonbonded kernel:   nb_kernel_ElecEwSw_VdwNone_GeomP1P1_F_c
 265  * Electrostatics interaction: Ewald
 266  * VdW interaction:            None
 267  * Geometry:                   Particle-Particle
 268  * Calculate force/pot:        Force
 269  */
 270 void
 271 nb_kernel_ElecEwSw_VdwNone_GeomP1P1_F_c
 272                     (t_nblist                    * gmx_restrict       nlist,
 273                      rvec                        * gmx_restrict          xx,
 274                      rvec                        * gmx_restrict          ff,
 275                      t_forcerec                  * gmx_restrict          fr,
 276                      t_mdatoms                   * gmx_restrict     mdatoms,
 277                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 278                      t_nrnb                      * gmx_restrict        nrnb)
 279 {
 280     int              i_shift_offset,i_coord_offset,j_coord_offset;
 281     int              j_index_start,j_index_end;
 282     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 283     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 284     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 285     real             *shiftvec,*fshift,*x,*f;
 286     int              vdwioffset0;
 287     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 288     int              vdwjidx0;
 289     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 290     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 291     real             velec,felec,velecsum,facel,crf,krf,krf2;
 292     real             *charge;
 293     int              ewitab;
 294     real             ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
 295     real             *ewtab;
 296     real             rswitch,swV3,swV4,swV5,swF2,swF3,swF4,d,d2,sw,dsw;
 297
 298     x                = xx[0];
 299     f                = ff[0];
 300
 301     nri              = nlist->nri;
 302     iinr             = nlist->iinr;
 303     jindex           = nlist->jindex;
 304     jjnr             = nlist->jjnr;
 305     shiftidx         = nlist->shift;
 306     gid              = nlist->gid;
 307     shiftvec         = fr->shift_vec[0];
 308     fshift           = fr->fshift[0];
 309     facel            = fr->epsfac;
 310     charge           = mdatoms->chargeA;
 311
 312     sh_ewald         = fr->ic->sh_ewald;
 313     ewtab            = fr->ic->tabq_coul_FDV0;
 314     ewtabscale       = fr->ic->tabq_scale;
 315     ewtabhalfspace   = 0.5/ewtabscale;
 316
 317     /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
 318     rcutoff          = fr->rcoulomb;
 319     rcutoff2         = rcutoff*rcutoff;
 320
 321     rswitch          = fr->rcoulomb_switch;
 322     /* Setup switch parameters */
 323     d                = rcutoff-rswitch;
 324     swV3             = -10.0/(d*d*d);
 325     swV4             =  15.0/(d*d*d*d);
 326     swV5             =  -6.0/(d*d*d*d*d);
 327     swF2             = -30.0/(d*d*d);
 328     swF3             =  60.0/(d*d*d*d);
 329     swF4             = -30.0/(d*d*d*d*d);
 330
 331     outeriter        = 0;
 332     inneriter        = 0;
 333
 334     /* Start outer loop over neighborlists */
 335     for(iidx=0; iidx<nri; iidx++)
 336     {
 337         /* Load shift vector for this list */
 338         i_shift_offset   = DIM*shiftidx[iidx];
 339         shX              = shiftvec[i_shift_offset+XX];
 340         shY              = shiftvec[i_shift_offset+YY];
 341         shZ              = shiftvec[i_shift_offset+ZZ];
 342
 343         /* Load limits for loop over neighbors */
 344         j_index_start    = jindex[iidx];
 345         j_index_end      = jindex[iidx+1];
 346
 347         /* Get outer coordinate index */
 348         inr              = iinr[iidx];
 349         i_coord_offset   = DIM*inr;
 350
 351         /* Load i particle coords and add shift vector */
 352         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 353         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 354         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 355
 356         fix0             = 0.0;
 357         fiy0             = 0.0;
 358         fiz0             = 0.0;
 359
 360         /* Load parameters for i particles */
 361         iq0              = facel*charge[inr+0];
 362
 363         /* Start inner kernel loop */
 364         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 365         {
 366             /* Get j neighbor index, and coordinate index */
 367             jnr              = jjnr[jidx];
 368             j_coord_offset   = DIM*jnr;
 369
 370             /* load j atom coordinates */
 371             jx0              = x[j_coord_offset+DIM*0+XX];
 372             jy0              = x[j_coord_offset+DIM*0+YY];
 373             jz0              = x[j_coord_offset+DIM*0+ZZ];
 374
 375             /* Calculate displacement vector */
 376             dx00             = ix0 - jx0;
 377             dy00             = iy0 - jy0;
 378             dz00             = iz0 - jz0;
 379
 380             /* Calculate squared distance and things based on it */
 381             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 382
 383             rinv00           = gmx_invsqrt(rsq00);
 384
 385             rinvsq00         = rinv00*rinv00;
 386
 387             /* Load parameters for j particles */
 388             jq0              = charge[jnr+0];
 389
 390             /**************************
 391              * CALCULATE INTERACTIONS *
 392              **************************/
 393
 394             if (rsq00<rcutoff2)
 395             {
 396
 397             r00              = rsq00*rinv00;
 398
 399             qq00             = iq0*jq0;
 400
 401             /* EWALD ELECTROSTATICS */
 402
 403             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 404             ewrt             = r00*ewtabscale;
 405             ewitab           = ewrt;
 406             eweps            = ewrt-ewitab;
 407             ewitab           = 4*ewitab;
 408             felec            = ewtab[ewitab]+eweps*ewtab[ewitab+1];
 409             velec            = qq00*(rinv00-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
 410             felec            = qq00*rinv00*(rinvsq00-felec);
 411
 412             d                = r00-rswitch;
 413             d                = (d>0.0) ? d : 0.0;
 414             d2               = d*d;
 415             sw               = 1.0+d2*d*(swV3+d*(swV4+d*swV5));
 416
 417             dsw              = d2*(swF2+d*(swF3+d*swF4));
 418
 419             /* Evaluate switch function */
 420             /* fscal'=f'/r=-(v*sw)'/r=-(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=fscal*sw-v*dsw/r */
 421             felec            = felec*sw - rinv00*velec*dsw;
 422
 423             fscal            = felec;
 424
 425             /* Calculate temporary vectorial force */
 426             tx               = fscal*dx00;
 427             ty               = fscal*dy00;
 428             tz               = fscal*dz00;
 429
 430             /* Update vectorial force */
 431             fix0            += tx;
 432             fiy0            += ty;
 433             fiz0            += tz;
 434             f[j_coord_offset+DIM*0+XX] -= tx;
 435             f[j_coord_offset+DIM*0+YY] -= ty;
 436             f[j_coord_offset+DIM*0+ZZ] -= tz;
 437
 438             }
 439
 440             /* Inner loop uses 57 flops */
 441         }
 442         /* End of innermost loop */
 443
 444         tx = ty = tz = 0;
 445         f[i_coord_offset+DIM*0+XX] += fix0;
 446         f[i_coord_offset+DIM*0+YY] += fiy0;
 447         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 448         tx                         += fix0;
 449         ty                         += fiy0;
 450         tz                         += fiz0;
 451         fshift[i_shift_offset+XX]  += tx;
 452         fshift[i_shift_offset+YY]  += ty;
 453         fshift[i_shift_offset+ZZ]  += tz;
 454
 455         /* Increment number of inner iterations */
 456         inneriter                  += j_index_end - j_index_start;
 457
 458         /* Outer loop uses 13 flops */
 459     }
 460
 461     /* Increment number of outer iterations */
 462     outeriter        += nri;
 463
 464     /* Update outer/inner flops */
 465
 466     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_F,outeriter*13 + inneriter*57);
 467 }