src/gromacs/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecEw_VdwCSTab_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_ElecEw_VdwCSTab_GeomP1P1_VF_c
  49  * Electrostatics interaction: Ewald
  50  * VdW interaction:            CubicSplineTable
  51  * Geometry:                   Particle-Particle
  52  * Calculate force/pot:        PotentialAndForce
  53  */
  54 void
  55 nb_kernel_ElecEw_VdwCSTab_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              vfitab;
  82     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
  83     real             *vftab;
  84     int              ewitab;
  85     real             ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
  86     real             *ewtab;
  87
  88     x                = xx[0];
  89     f                = ff[0];
  90
  91     nri              = nlist->nri;
  92     iinr             = nlist->iinr;
  93     jindex           = nlist->jindex;
  94     jjnr             = nlist->jjnr;
  95     shiftidx         = nlist->shift;
  96     gid              = nlist->gid;
  97     shiftvec         = fr->shift_vec[0];
  98     fshift           = fr->fshift[0];
  99     facel            = fr->epsfac;
 100     charge           = mdatoms->chargeA;
 101     nvdwtype         = fr->ntype;
 102     vdwparam         = fr->nbfp;
 103     vdwtype          = mdatoms->typeA;
 104
 105     vftab            = kernel_data->table_vdw->data;
 106     vftabscale       = kernel_data->table_vdw->scale;
 107
 108     sh_ewald         = fr->ic->sh_ewald;
 109     ewtab            = fr->ic->tabq_coul_FDV0;
 110     ewtabscale       = fr->ic->tabq_scale;
 111     ewtabhalfspace   = 0.5/ewtabscale;
 112
 113     outeriter        = 0;
 114     inneriter        = 0;
 115
 116     /* Start outer loop over neighborlists */
 117     for(iidx=0; iidx<nri; iidx++)
 118     {
 119         /* Load shift vector for this list */
 120         i_shift_offset   = DIM*shiftidx[iidx];
 121         shX              = shiftvec[i_shift_offset+XX];
 122         shY              = shiftvec[i_shift_offset+YY];
 123         shZ              = shiftvec[i_shift_offset+ZZ];
 124
 125         /* Load limits for loop over neighbors */
 126         j_index_start    = jindex[iidx];
 127         j_index_end      = jindex[iidx+1];
 128
 129         /* Get outer coordinate index */
 130         inr              = iinr[iidx];
 131         i_coord_offset   = DIM*inr;
 132
 133         /* Load i particle coords and add shift vector */
 134         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 135         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 136         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 137
 138         fix0             = 0.0;
 139         fiy0             = 0.0;
 140         fiz0             = 0.0;
 141
 142         /* Load parameters for i particles */
 143         iq0              = facel*charge[inr+0];
 144         vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 145
 146         /* Reset potential sums */
 147         velecsum         = 0.0;
 148         vvdwsum          = 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             vdwjidx0         = 2*vdwtype[jnr+0];
 177
 178             /**************************
 179              * CALCULATE INTERACTIONS *
 180              **************************/
 181
 182             r00              = rsq00*rinv00;
 183
 184             qq00             = iq0*jq0;
 185             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 186             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 187
 188             /* Calculate table index by multiplying r with table scale and truncate to integer */
 189             rt               = r00*vftabscale;
 190             vfitab           = rt;
 191             vfeps            = rt-vfitab;
 192             vfitab           = 2*4*vfitab;
 193
 194             /* EWALD ELECTROSTATICS */
 195
 196             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 197             ewrt             = r00*ewtabscale;
 198             ewitab           = ewrt;
 199             eweps            = ewrt-ewitab;
 200             ewitab           = 4*ewitab;
 201             felec            = ewtab[ewitab]+eweps*ewtab[ewitab+1];
 202             velec            = qq00*(rinv00-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
 203             felec            = qq00*rinv00*(rinvsq00-felec);
 204
 205             /* CUBIC SPLINE TABLE DISPERSION */
 206             vfitab          += 0;
 207             Y                = vftab[vfitab];
 208             F                = vftab[vfitab+1];
 209             Geps             = vfeps*vftab[vfitab+2];
 210             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 211             Fp               = F+Geps+Heps2;
 212             VV               = Y+vfeps*Fp;
 213             vvdw6            = c6_00*VV;
 214             FF               = Fp+Geps+2.0*Heps2;
 215             fvdw6            = c6_00*FF;
 216
 217             /* CUBIC SPLINE TABLE REPULSION */
 218             Y                = vftab[vfitab+4];
 219             F                = vftab[vfitab+5];
 220             Geps             = vfeps*vftab[vfitab+6];
 221             Heps2            = vfeps*vfeps*vftab[vfitab+7];
 222             Fp               = F+Geps+Heps2;
 223             VV               = Y+vfeps*Fp;
 224             vvdw12           = c12_00*VV;
 225             FF               = Fp+Geps+2.0*Heps2;
 226             fvdw12           = c12_00*FF;
 227             vvdw             = vvdw12+vvdw6;
 228             fvdw             = -(fvdw6+fvdw12)*vftabscale*rinv00;
 229
 230             /* Update potential sums from outer loop */
 231             velecsum        += velec;
 232             vvdwsum         += vvdw;
 233
 234             fscal            = felec+fvdw;
 235
 236             /* Calculate temporary vectorial force */
 237             tx               = fscal*dx00;
 238             ty               = fscal*dy00;
 239             tz               = fscal*dz00;
 240
 241             /* Update vectorial force */
 242             fix0            += tx;
 243             fiy0            += ty;
 244             fiz0            += tz;
 245             f[j_coord_offset+DIM*0+XX] -= tx;
 246             f[j_coord_offset+DIM*0+YY] -= ty;
 247             f[j_coord_offset+DIM*0+ZZ] -= tz;
 248
 249             /* Inner loop uses 74 flops */
 250         }
 251         /* End of innermost loop */
 252
 253         tx = ty = tz = 0;
 254         f[i_coord_offset+DIM*0+XX] += fix0;
 255         f[i_coord_offset+DIM*0+YY] += fiy0;
 256         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 257         tx                         += fix0;
 258         ty                         += fiy0;
 259         tz                         += fiz0;
 260         fshift[i_shift_offset+XX]  += tx;
 261         fshift[i_shift_offset+YY]  += ty;
 262         fshift[i_shift_offset+ZZ]  += tz;
 263
 264         ggid                        = gid[iidx];
 265         /* Update potential energies */
 266         kernel_data->energygrp_elec[ggid] += velecsum;
 267         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 268
 269         /* Increment number of inner iterations */
 270         inneriter                  += j_index_end - j_index_start;
 271
 272         /* Outer loop uses 15 flops */
 273     }
 274
 275     /* Increment number of outer iterations */
 276     outeriter        += nri;
 277
 278     /* Update outer/inner flops */
 279
 280     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_VF,outeriter*15 + inneriter*74);
 281 }
 282 /*
 283  * Gromacs nonbonded kernel:   nb_kernel_ElecEw_VdwCSTab_GeomP1P1_F_c
 284  * Electrostatics interaction: Ewald
 285  * VdW interaction:            CubicSplineTable
 286  * Geometry:                   Particle-Particle
 287  * Calculate force/pot:        Force
 288  */
 289 void
 290 nb_kernel_ElecEw_VdwCSTab_GeomP1P1_F_c
 291                     (t_nblist                    * gmx_restrict       nlist,
 292                      rvec                        * gmx_restrict          xx,
 293                      rvec                        * gmx_restrict          ff,
 294                      t_forcerec                  * gmx_restrict          fr,
 295                      t_mdatoms                   * gmx_restrict     mdatoms,
 296                      nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
 297                      t_nrnb                      * gmx_restrict        nrnb)
 298 {
 299     int              i_shift_offset,i_coord_offset,j_coord_offset;
 300     int              j_index_start,j_index_end;
 301     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 302     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 303     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 304     real             *shiftvec,*fshift,*x,*f;
 305     int              vdwioffset0;
 306     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 307     int              vdwjidx0;
 308     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 309     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 310     real             velec,felec,velecsum,facel,crf,krf,krf2;
 311     real             *charge;
 312     int              nvdwtype;
 313     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 314     int              *vdwtype;
 315     real             *vdwparam;
 316     int              vfitab;
 317     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
 318     real             *vftab;
 319     int              ewitab;
 320     real             ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace;
 321     real             *ewtab;
 322
 323     x                = xx[0];
 324     f                = ff[0];
 325
 326     nri              = nlist->nri;
 327     iinr             = nlist->iinr;
 328     jindex           = nlist->jindex;
 329     jjnr             = nlist->jjnr;
 330     shiftidx         = nlist->shift;
 331     gid              = nlist->gid;
 332     shiftvec         = fr->shift_vec[0];
 333     fshift           = fr->fshift[0];
 334     facel            = fr->epsfac;
 335     charge           = mdatoms->chargeA;
 336     nvdwtype         = fr->ntype;
 337     vdwparam         = fr->nbfp;
 338     vdwtype          = mdatoms->typeA;
 339
 340     vftab            = kernel_data->table_vdw->data;
 341     vftabscale       = kernel_data->table_vdw->scale;
 342
 343     sh_ewald         = fr->ic->sh_ewald;
 344     ewtab            = fr->ic->tabq_coul_F;
 345     ewtabscale       = fr->ic->tabq_scale;
 346     ewtabhalfspace   = 0.5/ewtabscale;
 347
 348     outeriter        = 0;
 349     inneriter        = 0;
 350
 351     /* Start outer loop over neighborlists */
 352     for(iidx=0; iidx<nri; iidx++)
 353     {
 354         /* Load shift vector for this list */
 355         i_shift_offset   = DIM*shiftidx[iidx];
 356         shX              = shiftvec[i_shift_offset+XX];
 357         shY              = shiftvec[i_shift_offset+YY];
 358         shZ              = shiftvec[i_shift_offset+ZZ];
 359
 360         /* Load limits for loop over neighbors */
 361         j_index_start    = jindex[iidx];
 362         j_index_end      = jindex[iidx+1];
 363
 364         /* Get outer coordinate index */
 365         inr              = iinr[iidx];
 366         i_coord_offset   = DIM*inr;
 367
 368         /* Load i particle coords and add shift vector */
 369         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 370         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 371         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 372
 373         fix0             = 0.0;
 374         fiy0             = 0.0;
 375         fiz0             = 0.0;
 376
 377         /* Load parameters for i particles */
 378         iq0              = facel*charge[inr+0];
 379         vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 380
 381         /* Start inner kernel loop */
 382         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 383         {
 384             /* Get j neighbor index, and coordinate index */
 385             jnr              = jjnr[jidx];
 386             j_coord_offset   = DIM*jnr;
 387
 388             /* load j atom coordinates */
 389             jx0              = x[j_coord_offset+DIM*0+XX];
 390             jy0              = x[j_coord_offset+DIM*0+YY];
 391             jz0              = x[j_coord_offset+DIM*0+ZZ];
 392
 393             /* Calculate displacement vector */
 394             dx00             = ix0 - jx0;
 395             dy00             = iy0 - jy0;
 396             dz00             = iz0 - jz0;
 397
 398             /* Calculate squared distance and things based on it */
 399             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 400
 401             rinv00           = gmx_invsqrt(rsq00);
 402
 403             rinvsq00         = rinv00*rinv00;
 404
 405             /* Load parameters for j particles */
 406             jq0              = charge[jnr+0];
 407             vdwjidx0         = 2*vdwtype[jnr+0];
 408
 409             /**************************
 410              * CALCULATE INTERACTIONS *
 411              **************************/
 412
 413             r00              = rsq00*rinv00;
 414
 415             qq00             = iq0*jq0;
 416             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 417             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 418
 419             /* Calculate table index by multiplying r with table scale and truncate to integer */
 420             rt               = r00*vftabscale;
 421             vfitab           = rt;
 422             vfeps            = rt-vfitab;
 423             vfitab           = 2*4*vfitab;
 424
 425             /* EWALD ELECTROSTATICS */
 426
 427             /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
 428             ewrt             = r00*ewtabscale;
 429             ewitab           = ewrt;
 430             eweps            = ewrt-ewitab;
 431             felec            = (1.0-eweps)*ewtab[ewitab]+eweps*ewtab[ewitab+1];
 432             felec            = qq00*rinv00*(rinvsq00-felec);
 433
 434             /* CUBIC SPLINE TABLE DISPERSION */
 435             vfitab          += 0;
 436             F                = vftab[vfitab+1];
 437             Geps             = vfeps*vftab[vfitab+2];
 438             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 439             Fp               = F+Geps+Heps2;
 440             FF               = Fp+Geps+2.0*Heps2;
 441             fvdw6            = c6_00*FF;
 442
 443             /* CUBIC SPLINE TABLE REPULSION */
 444             F                = vftab[vfitab+5];
 445             Geps             = vfeps*vftab[vfitab+6];
 446             Heps2            = vfeps*vfeps*vftab[vfitab+7];
 447             Fp               = F+Geps+Heps2;
 448             FF               = Fp+Geps+2.0*Heps2;
 449             fvdw12           = c12_00*FF;
 450             fvdw             = -(fvdw6+fvdw12)*vftabscale*rinv00;
 451
 452             fscal            = felec+fvdw;
 453
 454             /* Calculate temporary vectorial force */
 455             tx               = fscal*dx00;
 456             ty               = fscal*dy00;
 457             tz               = fscal*dz00;
 458
 459             /* Update vectorial force */
 460             fix0            += tx;
 461             fiy0            += ty;
 462             fiz0            += tz;
 463             f[j_coord_offset+DIM*0+XX] -= tx;
 464             f[j_coord_offset+DIM*0+YY] -= ty;
 465             f[j_coord_offset+DIM*0+ZZ] -= tz;
 466
 467             /* Inner loop uses 59 flops */
 468         }
 469         /* End of innermost loop */
 470
 471         tx = ty = tz = 0;
 472         f[i_coord_offset+DIM*0+XX] += fix0;
 473         f[i_coord_offset+DIM*0+YY] += fiy0;
 474         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 475         tx                         += fix0;
 476         ty                         += fiy0;
 477         tz                         += fiz0;
 478         fshift[i_shift_offset+XX]  += tx;
 479         fshift[i_shift_offset+YY]  += ty;
 480         fshift[i_shift_offset+ZZ]  += tz;
 481
 482         /* Increment number of inner iterations */
 483         inneriter                  += j_index_end - j_index_start;
 484
 485         /* Outer loop uses 13 flops */
 486     }
 487
 488     /* Increment number of outer iterations */
 489     outeriter        += nri;
 490
 491     /* Update outer/inner flops */
 492
 493     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_F,outeriter*13 + inneriter*59);
 494 }