2 * This file is part of the GROMACS molecular simulation package.
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.
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.
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.
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.
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.
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.
36 #define UNROLLI NBNXN_CPU_CLUSTER_I_SIZE
37 #define UNROLLJ NBNXN_CPU_CLUSTER_I_SIZE
39 /* We could use nbat->xstride and nbat->fstride, but macros might be faster */
42 /* Local i-atom buffer strides */
47 /* All functionality defines are set here, except for:
48 * CALC_ENERGIES, ENERGY_GROUPS which are defined before.
49 * CHECK_EXCLS, which is set just before including the inner loop contents.
52 /* We always calculate shift forces, because it's cheap anyhow */
53 #define CALC_SHIFTFORCES
56 #define NBK_FUNC_NAME2(ljt, feg) nbnxn_kernel ## _ElecRF ## ljt ## feg ## _ref
59 #ifndef VDW_CUTOFF_CHECK
60 #define NBK_FUNC_NAME2(ljt, feg) nbnxn_kernel ## _ElecQSTab ## ljt ## feg ## _ref
62 #define NBK_FUNC_NAME2(ljt, feg) nbnxn_kernel ## _ElecQSTabTwinCut ## ljt ## feg ## _ref
66 #ifdef VDW_FORCE_SWITCH
67 #define NBK_FUNC_NAME(feg) NBK_FUNC_NAME2(_VdwLJFsw, feg)
70 #define NBK_FUNC_NAME(feg) NBK_FUNC_NAME2(_VdwLJPsw, feg)
72 #define NBK_FUNC_NAME(feg) NBK_FUNC_NAME2(_VdwLJ, feg)
88 (const nbnxn_pairlist_t *nbl,
89 const nbnxn_atomdata_t *nbat,
90 const interaction_const_t *ic,
93 #ifdef CALC_SHIFTFORCES
104 const nbnxn_ci_t *nbln;
105 const nbnxn_cj_t *l_cj;
108 const real *shiftvec;
112 #ifdef VDW_CUTOFF_CHECK
120 gmx_bool do_LJ, half_LJ, do_coul;
121 int cjind0, cjind1, cjind;
124 real xi[UNROLLI*XI_STRIDE];
125 real fi[UNROLLI*FI_STRIDE];
129 #ifndef ENERGY_GROUPS
134 int egp_sh_i[UNROLLI];
137 #ifdef VDW_POT_SWITCH
138 real swV3, swV4, swV5;
139 real swF2, swF3, swF4;
154 const real *tab_coul_FDV0;
156 const real *tab_coul_F;
157 const real *tab_coul_V;
167 #ifdef VDW_POT_SWITCH
168 swV3 = ic->vdw_switch.c3;
169 swV4 = ic->vdw_switch.c4;
170 swV5 = ic->vdw_switch.c5;
171 swF2 = 3*ic->vdw_switch.c3;
172 swF3 = 4*ic->vdw_switch.c4;
173 swF4 = 5*ic->vdw_switch.c5;
184 tabscale = ic->tabq_scale;
186 halfsp = 0.5/ic->tabq_scale;
190 tab_coul_FDV0 = ic->tabq_coul_FDV0;
192 tab_coul_F = ic->tabq_coul_F;
193 tab_coul_V = ic->tabq_coul_V;
198 egp_mask = (1<<nbat->neg_2log) - 1;
202 rcut2 = ic->rcoulomb*ic->rcoulomb;
203 #ifdef VDW_CUTOFF_CHECK
204 rvdw2 = ic->rvdw*ic->rvdw;
207 ntype2 = nbat->ntype*2;
212 shiftvec = shift_vec[0];
218 for (n = 0; n < nbl->nci; n++)
224 ish = (nbln->shift & NBNXN_CI_SHIFT);
225 /* x, f and fshift are assumed to be stored with stride 3 */
227 cjind0 = nbln->cj_ind_start;
228 cjind1 = nbln->cj_ind_end;
229 /* Currently only works super-cells equal to sub-cells */
231 ci_sh = (ish == CENTRAL ? ci : -1);
233 /* We have 5 LJ/C combinations, but use only three inner loops,
234 * as the other combinations are unlikely and/or not much faster:
235 * inner half-LJ + C for half-LJ + C / no-LJ + C
236 * inner LJ + C for full-LJ + C
237 * inner LJ for full-LJ + no-C / half-LJ + no-C
239 do_LJ = (nbln->shift & NBNXN_CI_DO_LJ(0));
240 do_coul = (nbln->shift & NBNXN_CI_DO_COUL(0));
241 half_LJ = ((nbln->shift & NBNXN_CI_HALF_LJ(0)) || !do_LJ) && do_coul;
244 #ifndef ENERGY_GROUPS
248 for (i = 0; i < UNROLLI; i++)
250 egp_sh_i[i] = ((nbat->energrp[ci]>>(i*nbat->neg_2log)) & egp_mask)*nbat->nenergrp;
255 for (i = 0; i < UNROLLI; i++)
257 for (d = 0; d < DIM; d++)
259 xi[i*XI_STRIDE+d] = x[(ci*UNROLLI+i)*X_STRIDE+d] + shiftvec[ishf+d];
260 fi[i*FI_STRIDE+d] = 0;
270 Vc_sub_self = 0.5*c_rf;
274 Vc_sub_self = 0.5*tab_coul_V[0];
276 Vc_sub_self = 0.5*tab_coul_FDV0[2];
281 for (i = 0; i < UNROLLI; i++)
283 qi[i] = facel*q[ci*UNROLLI+i];
286 if (l_cj[nbln->cj_ind_start].cj == ci_sh)
289 Vc[egp_sh_i[i]+((nbat->energrp[ci]>>(i*nbat->neg_2log)) & egp_mask)]
293 -= qi[i]*q[ci*UNROLLI+i]*Vc_sub_self;
300 while (cjind < cjind1 && nbl->cj[cjind].excl != 0xffff)
307 #include "nbnxn_kernel_ref_inner.h"
311 /* cppcheck-suppress duplicateBranch */
315 #include "nbnxn_kernel_ref_inner.h"
320 #include "nbnxn_kernel_ref_inner.h"
326 for (; (cjind < cjind1); cjind++)
332 #include "nbnxn_kernel_ref_inner.h"
336 /* cppcheck-suppress duplicateBranch */
340 #include "nbnxn_kernel_ref_inner.h"
345 #include "nbnxn_kernel_ref_inner.h"
348 ninner += cjind1 - cjind0;
350 /* Add accumulated i-forces to the force array */
351 for (i = 0; i < UNROLLI; i++)
353 for (d = 0; d < DIM; d++)
355 f[(ci*UNROLLI+i)*F_STRIDE+d] += fi[i*FI_STRIDE+d];
358 #ifdef CALC_SHIFTFORCES
361 /* Add i forces to shifted force list */
362 for (i = 0; i < UNROLLI; i++)
364 for (d = 0; d < DIM; d++)
366 fshift[ishf+d] += fi[i*FI_STRIDE+d];
373 #ifndef ENERGY_GROUPS
381 printf("atom pairs %d\n", npair);
385 #undef CALC_SHIFTFORCES