src/gromacs/gmxlib/nonbonded/nb_kernel_c/nb_kernel_ElecCoul_VdwCSTab_GeomP1P1_c.c

   1 /*
   2  * Note: this file was generated by the Gromacs c kernel generator.
   3  *
   4  *                This source code is part of
   5  *
   6  *                 G   R   O   M   A   C   S
   7  *
   8  * Copyright (c) 2001-2012, The GROMACS Development Team
   9  *
  10  * Gromacs is a library for molecular simulation and trajectory analysis,
  11  * written by Erik Lindahl, David van der Spoel, Berk Hess, and others - for
  12  * a full list of developers and information, check out http://www.gromacs.org
  13  *
  14  * This program is free software; you can redistribute it and/or modify it under
  15  * the terms of the GNU Lesser General Public License as published by the Free
  16  * Software Foundation; either version 2 of the License, or (at your option) any
  17  * later version.
  18  *
  19  * To help fund GROMACS development, we humbly ask that you cite
  20  * the papers people have written on it - you can find them on the website.
  21  */
  22 #ifdef HAVE_CONFIG_H
  23 #include <config.h>
  24 #endif
  25
  26 #include <math.h>
  27
  28 #include "../nb_kernel.h"
  29 #include "types/simple.h"
  30 #include "vec.h"
  31 #include "nrnb.h"
  32
  33 /*
  34  * Gromacs nonbonded kernel:   nb_kernel_ElecCoul_VdwCSTab_GeomP1P1_VF_c
  35  * Electrostatics interaction: Coulomb
  36  * VdW interaction:            CubicSplineTable
  37  * Geometry:                   Particle-Particle
  38  * Calculate force/pot:        PotentialAndForce
  39  */
  40 void
  41 nb_kernel_ElecCoul_VdwCSTab_GeomP1P1_VF_c
  42                     (t_nblist * gmx_restrict                nlist,
  43                      rvec * gmx_restrict                    xx,
  44                      rvec * gmx_restrict                    ff,
  45                      t_forcerec * gmx_restrict              fr,
  46                      t_mdatoms * gmx_restrict               mdatoms,
  47                      nb_kernel_data_t * gmx_restrict        kernel_data,
  48                      t_nrnb * gmx_restrict                  nrnb)
  49 {
  50     int              i_shift_offset,i_coord_offset,j_coord_offset;
  51     int              j_index_start,j_index_end;
  52     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
  53     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
  54     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
  55     real             *shiftvec,*fshift,*x,*f;
  56     int              vdwioffset0;
  57     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
  58     int              vdwjidx0;
  59     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
  60     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
  61     real             velec,felec,velecsum,facel,crf,krf,krf2;
  62     real             *charge;
  63     int              nvdwtype;
  64     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
  65     int              *vdwtype;
  66     real             *vdwparam;
  67     int              vfitab;
  68     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
  69     real             *vftab;
  70
  71     x                = xx[0];
  72     f                = ff[0];
  73
  74     nri              = nlist->nri;
  75     iinr             = nlist->iinr;
  76     jindex           = nlist->jindex;
  77     jjnr             = nlist->jjnr;
  78     shiftidx         = nlist->shift;
  79     gid              = nlist->gid;
  80     shiftvec         = fr->shift_vec[0];
  81     fshift           = fr->fshift[0];
  82     facel            = fr->epsfac;
  83     charge           = mdatoms->chargeA;
  84     nvdwtype         = fr->ntype;
  85     vdwparam         = fr->nbfp;
  86     vdwtype          = mdatoms->typeA;
  87
  88     vftab            = kernel_data->table_vdw->data;
  89     vftabscale       = kernel_data->table_vdw->scale;
  90
  91     outeriter        = 0;
  92     inneriter        = 0;
  93
  94     /* Start outer loop over neighborlists */
  95     for(iidx=0; iidx<nri; iidx++)
  96     {
  97         /* Load shift vector for this list */
  98         i_shift_offset   = DIM*shiftidx[iidx];
  99         shX              = shiftvec[i_shift_offset+XX];
 100         shY              = shiftvec[i_shift_offset+YY];
 101         shZ              = shiftvec[i_shift_offset+ZZ];
 102
 103         /* Load limits for loop over neighbors */
 104         j_index_start    = jindex[iidx];
 105         j_index_end      = jindex[iidx+1];
 106
 107         /* Get outer coordinate index */
 108         inr              = iinr[iidx];
 109         i_coord_offset   = DIM*inr;
 110
 111         /* Load i particle coords and add shift vector */
 112         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 113         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 114         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 115
 116         fix0             = 0.0;
 117         fiy0             = 0.0;
 118         fiz0             = 0.0;
 119
 120         /* Load parameters for i particles */
 121         iq0              = facel*charge[inr+0];
 122         vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 123
 124         /* Reset potential sums */
 125         velecsum         = 0.0;
 126         vvdwsum          = 0.0;
 127
 128         /* Start inner kernel loop */
 129         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 130         {
 131             /* Get j neighbor index, and coordinate index */
 132             jnr              = jjnr[jidx];
 133             j_coord_offset   = DIM*jnr;
 134
 135             /* load j atom coordinates */
 136             jx0              = x[j_coord_offset+DIM*0+XX];
 137             jy0              = x[j_coord_offset+DIM*0+YY];
 138             jz0              = x[j_coord_offset+DIM*0+ZZ];
 139
 140             /* Calculate displacement vector */
 141             dx00             = ix0 - jx0;
 142             dy00             = iy0 - jy0;
 143             dz00             = iz0 - jz0;
 144
 145             /* Calculate squared distance and things based on it */
 146             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 147
 148             rinv00           = gmx_invsqrt(rsq00);
 149
 150             rinvsq00         = rinv00*rinv00;
 151
 152             /* Load parameters for j particles */
 153             jq0              = charge[jnr+0];
 154             vdwjidx0         = 2*vdwtype[jnr+0];
 155
 156             /**************************
 157              * CALCULATE INTERACTIONS *
 158              **************************/
 159
 160             r00              = rsq00*rinv00;
 161
 162             qq00             = iq0*jq0;
 163             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 164             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 165
 166             /* Calculate table index by multiplying r with table scale and truncate to integer */
 167             rt               = r00*vftabscale;
 168             vfitab           = rt;
 169             vfeps            = rt-vfitab;
 170             vfitab           = 2*4*vfitab;
 171
 172             /* COULOMB ELECTROSTATICS */
 173             velec            = qq00*rinv00;
 174             felec            = velec*rinvsq00;
 175
 176             /* CUBIC SPLINE TABLE DISPERSION */
 177             vfitab          += 0;
 178             Y                = vftab[vfitab];
 179             F                = vftab[vfitab+1];
 180             Geps             = vfeps*vftab[vfitab+2];
 181             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 182             Fp               = F+Geps+Heps2;
 183             VV               = Y+vfeps*Fp;
 184             vvdw6            = c6_00*VV;
 185             FF               = Fp+Geps+2.0*Heps2;
 186             fvdw6            = c6_00*FF;
 187
 188             /* CUBIC SPLINE TABLE REPULSION */
 189             Y                = vftab[vfitab+4];
 190             F                = vftab[vfitab+5];
 191             Geps             = vfeps*vftab[vfitab+6];
 192             Heps2            = vfeps*vfeps*vftab[vfitab+7];
 193             Fp               = F+Geps+Heps2;
 194             VV               = Y+vfeps*Fp;
 195             vvdw12           = c12_00*VV;
 196             FF               = Fp+Geps+2.0*Heps2;
 197             fvdw12           = c12_00*FF;
 198             vvdw             = vvdw12+vvdw6;
 199             fvdw             = -(fvdw6+fvdw12)*vftabscale*rinv00;
 200
 201             /* Update potential sums from outer loop */
 202             velecsum        += velec;
 203             vvdwsum         += vvdw;
 204
 205             fscal            = felec+fvdw;
 206
 207             /* Calculate temporary vectorial force */
 208             tx               = fscal*dx00;
 209             ty               = fscal*dy00;
 210             tz               = fscal*dz00;
 211
 212             /* Update vectorial force */
 213             fix0            += tx;
 214             fiy0            += ty;
 215             fiz0            += tz;
 216             f[j_coord_offset+DIM*0+XX] -= tx;
 217             f[j_coord_offset+DIM*0+YY] -= ty;
 218             f[j_coord_offset+DIM*0+ZZ] -= tz;
 219
 220             /* Inner loop uses 62 flops */
 221         }
 222         /* End of innermost loop */
 223
 224         tx = ty = tz = 0;
 225         f[i_coord_offset+DIM*0+XX] += fix0;
 226         f[i_coord_offset+DIM*0+YY] += fiy0;
 227         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 228         tx                         += fix0;
 229         ty                         += fiy0;
 230         tz                         += fiz0;
 231         fshift[i_shift_offset+XX]  += tx;
 232         fshift[i_shift_offset+YY]  += ty;
 233         fshift[i_shift_offset+ZZ]  += tz;
 234
 235         ggid                        = gid[iidx];
 236         /* Update potential energies */
 237         kernel_data->energygrp_elec[ggid] += velecsum;
 238         kernel_data->energygrp_vdw[ggid] += vvdwsum;
 239
 240         /* Increment number of inner iterations */
 241         inneriter                  += j_index_end - j_index_start;
 242
 243         /* Outer loop uses 15 flops */
 244     }
 245
 246     /* Increment number of outer iterations */
 247     outeriter        += nri;
 248
 249     /* Update outer/inner flops */
 250
 251     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_VF,outeriter*15 + inneriter*62);
 252 }
 253 /*
 254  * Gromacs nonbonded kernel:   nb_kernel_ElecCoul_VdwCSTab_GeomP1P1_F_c
 255  * Electrostatics interaction: Coulomb
 256  * VdW interaction:            CubicSplineTable
 257  * Geometry:                   Particle-Particle
 258  * Calculate force/pot:        Force
 259  */
 260 void
 261 nb_kernel_ElecCoul_VdwCSTab_GeomP1P1_F_c
 262                     (t_nblist * gmx_restrict                nlist,
 263                      rvec * gmx_restrict                    xx,
 264                      rvec * gmx_restrict                    ff,
 265                      t_forcerec * gmx_restrict              fr,
 266                      t_mdatoms * gmx_restrict               mdatoms,
 267                      nb_kernel_data_t * gmx_restrict        kernel_data,
 268                      t_nrnb * gmx_restrict                  nrnb)
 269 {
 270     int              i_shift_offset,i_coord_offset,j_coord_offset;
 271     int              j_index_start,j_index_end;
 272     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
 273     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
 274     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
 275     real             *shiftvec,*fshift,*x,*f;
 276     int              vdwioffset0;
 277     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
 278     int              vdwjidx0;
 279     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
 280     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
 281     real             velec,felec,velecsum,facel,crf,krf,krf2;
 282     real             *charge;
 283     int              nvdwtype;
 284     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
 285     int              *vdwtype;
 286     real             *vdwparam;
 287     int              vfitab;
 288     real             rt,vfeps,vftabscale,Y,F,Geps,Heps2,Fp,VV,FF;
 289     real             *vftab;
 290
 291     x                = xx[0];
 292     f                = ff[0];
 293
 294     nri              = nlist->nri;
 295     iinr             = nlist->iinr;
 296     jindex           = nlist->jindex;
 297     jjnr             = nlist->jjnr;
 298     shiftidx         = nlist->shift;
 299     gid              = nlist->gid;
 300     shiftvec         = fr->shift_vec[0];
 301     fshift           = fr->fshift[0];
 302     facel            = fr->epsfac;
 303     charge           = mdatoms->chargeA;
 304     nvdwtype         = fr->ntype;
 305     vdwparam         = fr->nbfp;
 306     vdwtype          = mdatoms->typeA;
 307
 308     vftab            = kernel_data->table_vdw->data;
 309     vftabscale       = kernel_data->table_vdw->scale;
 310
 311     outeriter        = 0;
 312     inneriter        = 0;
 313
 314     /* Start outer loop over neighborlists */
 315     for(iidx=0; iidx<nri; iidx++)
 316     {
 317         /* Load shift vector for this list */
 318         i_shift_offset   = DIM*shiftidx[iidx];
 319         shX              = shiftvec[i_shift_offset+XX];
 320         shY              = shiftvec[i_shift_offset+YY];
 321         shZ              = shiftvec[i_shift_offset+ZZ];
 322
 323         /* Load limits for loop over neighbors */
 324         j_index_start    = jindex[iidx];
 325         j_index_end      = jindex[iidx+1];
 326
 327         /* Get outer coordinate index */
 328         inr              = iinr[iidx];
 329         i_coord_offset   = DIM*inr;
 330
 331         /* Load i particle coords and add shift vector */
 332         ix0              = shX + x[i_coord_offset+DIM*0+XX];
 333         iy0              = shY + x[i_coord_offset+DIM*0+YY];
 334         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
 335
 336         fix0             = 0.0;
 337         fiy0             = 0.0;
 338         fiz0             = 0.0;
 339
 340         /* Load parameters for i particles */
 341         iq0              = facel*charge[inr+0];
 342         vdwioffset0      = 2*nvdwtype*vdwtype[inr+0];
 343
 344         /* Start inner kernel loop */
 345         for(jidx=j_index_start; jidx<j_index_end; jidx++)
 346         {
 347             /* Get j neighbor index, and coordinate index */
 348             jnr              = jjnr[jidx];
 349             j_coord_offset   = DIM*jnr;
 350
 351             /* load j atom coordinates */
 352             jx0              = x[j_coord_offset+DIM*0+XX];
 353             jy0              = x[j_coord_offset+DIM*0+YY];
 354             jz0              = x[j_coord_offset+DIM*0+ZZ];
 355
 356             /* Calculate displacement vector */
 357             dx00             = ix0 - jx0;
 358             dy00             = iy0 - jy0;
 359             dz00             = iz0 - jz0;
 360
 361             /* Calculate squared distance and things based on it */
 362             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
 363
 364             rinv00           = gmx_invsqrt(rsq00);
 365
 366             rinvsq00         = rinv00*rinv00;
 367
 368             /* Load parameters for j particles */
 369             jq0              = charge[jnr+0];
 370             vdwjidx0         = 2*vdwtype[jnr+0];
 371
 372             /**************************
 373              * CALCULATE INTERACTIONS *
 374              **************************/
 375
 376             r00              = rsq00*rinv00;
 377
 378             qq00             = iq0*jq0;
 379             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
 380             c12_00           = vdwparam[vdwioffset0+vdwjidx0+1];
 381
 382             /* Calculate table index by multiplying r with table scale and truncate to integer */
 383             rt               = r00*vftabscale;
 384             vfitab           = rt;
 385             vfeps            = rt-vfitab;
 386             vfitab           = 2*4*vfitab;
 387
 388             /* COULOMB ELECTROSTATICS */
 389             velec            = qq00*rinv00;
 390             felec            = velec*rinvsq00;
 391
 392             /* CUBIC SPLINE TABLE DISPERSION */
 393             vfitab          += 0;
 394             F                = vftab[vfitab+1];
 395             Geps             = vfeps*vftab[vfitab+2];
 396             Heps2            = vfeps*vfeps*vftab[vfitab+3];
 397             Fp               = F+Geps+Heps2;
 398             FF               = Fp+Geps+2.0*Heps2;
 399             fvdw6            = c6_00*FF;
 400
 401             /* CUBIC SPLINE TABLE REPULSION */
 402             F                = vftab[vfitab+5];
 403             Geps             = vfeps*vftab[vfitab+6];
 404             Heps2            = vfeps*vfeps*vftab[vfitab+7];
 405             Fp               = F+Geps+Heps2;
 406             FF               = Fp+Geps+2.0*Heps2;
 407             fvdw12           = c12_00*FF;
 408             fvdw             = -(fvdw6+fvdw12)*vftabscale*rinv00;
 409
 410             fscal            = felec+fvdw;
 411
 412             /* Calculate temporary vectorial force */
 413             tx               = fscal*dx00;
 414             ty               = fscal*dy00;
 415             tz               = fscal*dz00;
 416
 417             /* Update vectorial force */
 418             fix0            += tx;
 419             fiy0            += ty;
 420             fiz0            += tz;
 421             f[j_coord_offset+DIM*0+XX] -= tx;
 422             f[j_coord_offset+DIM*0+YY] -= ty;
 423             f[j_coord_offset+DIM*0+ZZ] -= tz;
 424
 425             /* Inner loop uses 53 flops */
 426         }
 427         /* End of innermost loop */
 428
 429         tx = ty = tz = 0;
 430         f[i_coord_offset+DIM*0+XX] += fix0;
 431         f[i_coord_offset+DIM*0+YY] += fiy0;
 432         f[i_coord_offset+DIM*0+ZZ] += fiz0;
 433         tx                         += fix0;
 434         ty                         += fiy0;
 435         tz                         += fiz0;
 436         fshift[i_shift_offset+XX]  += tx;
 437         fshift[i_shift_offset+YY]  += ty;
 438         fshift[i_shift_offset+ZZ]  += tz;
 439
 440         /* Increment number of inner iterations */
 441         inneriter                  += j_index_end - j_index_start;
 442
 443         /* Outer loop uses 13 flops */
 444     }
 445
 446     /* Increment number of outer iterations */
 447     outeriter        += nri;
 448
 449     /* Update outer/inner flops */
 450
 451     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_F,outeriter*13 + inneriter*53);
 452 }