Fix component for libcudart
[alexxy/gromacs.git] / src / gmxlib / nonbonded / nb_kernel_c / nb_kernel_ElecCoul_VdwBham_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_VdwBham_GeomP1P1_VF_c
35  * Electrostatics interaction: Coulomb
36  * VdW interaction:            Buckingham
37  * Geometry:                   Particle-Particle
38  * Calculate force/pot:        PotentialAndForce
39  */
40 void
41 nb_kernel_ElecCoul_VdwBham_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
68     x                = xx[0];
69     f                = ff[0];
70
71     nri              = nlist->nri;
72     iinr             = nlist->iinr;
73     jindex           = nlist->jindex;
74     jjnr             = nlist->jjnr;
75     shiftidx         = nlist->shift;
76     gid              = nlist->gid;
77     shiftvec         = fr->shift_vec[0];
78     fshift           = fr->fshift[0];
79     facel            = fr->epsfac;
80     charge           = mdatoms->chargeA;
81     nvdwtype         = fr->ntype;
82     vdwparam         = fr->nbfp;
83     vdwtype          = mdatoms->typeA;
84
85     outeriter        = 0;
86     inneriter        = 0;
87
88     /* Start outer loop over neighborlists */
89     for(iidx=0; iidx<nri; iidx++)
90     {
91         /* Load shift vector for this list */
92         i_shift_offset   = DIM*shiftidx[iidx];
93         shX              = shiftvec[i_shift_offset+XX];
94         shY              = shiftvec[i_shift_offset+YY];
95         shZ              = shiftvec[i_shift_offset+ZZ];
96
97         /* Load limits for loop over neighbors */
98         j_index_start    = jindex[iidx];
99         j_index_end      = jindex[iidx+1];
100
101         /* Get outer coordinate index */
102         inr              = iinr[iidx];
103         i_coord_offset   = DIM*inr;
104
105         /* Load i particle coords and add shift vector */
106         ix0              = shX + x[i_coord_offset+DIM*0+XX];
107         iy0              = shY + x[i_coord_offset+DIM*0+YY];
108         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
109
110         fix0             = 0.0;
111         fiy0             = 0.0;
112         fiz0             = 0.0;
113
114         /* Load parameters for i particles */
115         iq0              = facel*charge[inr+0];
116         vdwioffset0      = 3*nvdwtype*vdwtype[inr+0];
117
118         /* Reset potential sums */
119         velecsum         = 0.0;
120         vvdwsum          = 0.0;
121
122         /* Start inner kernel loop */
123         for(jidx=j_index_start; jidx<j_index_end; jidx++)
124         {
125             /* Get j neighbor index, and coordinate index */
126             jnr              = jjnr[jidx];
127             j_coord_offset   = DIM*jnr;
128
129             /* load j atom coordinates */
130             jx0              = x[j_coord_offset+DIM*0+XX];
131             jy0              = x[j_coord_offset+DIM*0+YY];
132             jz0              = x[j_coord_offset+DIM*0+ZZ];
133
134             /* Calculate displacement vector */
135             dx00             = ix0 - jx0;
136             dy00             = iy0 - jy0;
137             dz00             = iz0 - jz0;
138
139             /* Calculate squared distance and things based on it */
140             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
141
142             rinv00           = gmx_invsqrt(rsq00);
143
144             rinvsq00         = rinv00*rinv00;
145
146             /* Load parameters for j particles */
147             jq0              = charge[jnr+0];
148             vdwjidx0         = 3*vdwtype[jnr+0];
149
150             /**************************
151              * CALCULATE INTERACTIONS *
152              **************************/
153
154             r00              = rsq00*rinv00;
155
156             qq00             = iq0*jq0;
157             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
158             cexp1_00         = vdwparam[vdwioffset0+vdwjidx0+1];
159             cexp2_00         = vdwparam[vdwioffset0+vdwjidx0+2];
160
161             /* COULOMB ELECTROSTATICS */
162             velec            = qq00*rinv00;
163             felec            = velec*rinvsq00;
164
165             /* BUCKINGHAM DISPERSION/REPULSION */
166             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
167             vvdw6            = c6_00*rinvsix;
168             br               = cexp2_00*r00;
169             vvdwexp          = cexp1_00*exp(-br);
170             vvdw             = vvdwexp - vvdw6*(1.0/6.0);
171             fvdw             = (br*vvdwexp-vvdw6)*rinvsq00;
172
173             /* Update potential sums from outer loop */
174             velecsum        += velec;
175             vvdwsum         += vvdw;
176
177             fscal            = felec+fvdw;
178
179             /* Calculate temporary vectorial force */
180             tx               = fscal*dx00;
181             ty               = fscal*dy00;
182             tz               = fscal*dz00;
183
184             /* Update vectorial force */
185             fix0            += tx;
186             fiy0            += ty;
187             fiz0            += tz;
188             f[j_coord_offset+DIM*0+XX] -= tx;
189             f[j_coord_offset+DIM*0+YY] -= ty;
190             f[j_coord_offset+DIM*0+ZZ] -= tz;
191
192             /* Inner loop uses 67 flops */
193         }
194         /* End of innermost loop */
195
196         tx = ty = tz = 0;
197         f[i_coord_offset+DIM*0+XX] += fix0;
198         f[i_coord_offset+DIM*0+YY] += fiy0;
199         f[i_coord_offset+DIM*0+ZZ] += fiz0;
200         tx                         += fix0;
201         ty                         += fiy0;
202         tz                         += fiz0;
203         fshift[i_shift_offset+XX]  += tx;
204         fshift[i_shift_offset+YY]  += ty;
205         fshift[i_shift_offset+ZZ]  += tz;
206
207         ggid                        = gid[iidx];
208         /* Update potential energies */
209         kernel_data->energygrp_elec[ggid] += velecsum;
210         kernel_data->energygrp_vdw[ggid] += vvdwsum;
211
212         /* Increment number of inner iterations */
213         inneriter                  += j_index_end - j_index_start;
214
215         /* Outer loop uses 15 flops */
216     }
217
218     /* Increment number of outer iterations */
219     outeriter        += nri;
220
221     /* Update outer/inner flops */
222
223     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_VF,outeriter*15 + inneriter*67);
224 }
225 /*
226  * Gromacs nonbonded kernel:   nb_kernel_ElecCoul_VdwBham_GeomP1P1_F_c
227  * Electrostatics interaction: Coulomb
228  * VdW interaction:            Buckingham
229  * Geometry:                   Particle-Particle
230  * Calculate force/pot:        Force
231  */
232 void
233 nb_kernel_ElecCoul_VdwBham_GeomP1P1_F_c
234                     (t_nblist * gmx_restrict                nlist,
235                      rvec * gmx_restrict                    xx,
236                      rvec * gmx_restrict                    ff,
237                      t_forcerec * gmx_restrict              fr,
238                      t_mdatoms * gmx_restrict               mdatoms,
239                      nb_kernel_data_t * gmx_restrict        kernel_data,
240                      t_nrnb * gmx_restrict                  nrnb)
241 {
242     int              i_shift_offset,i_coord_offset,j_coord_offset;
243     int              j_index_start,j_index_end;
244     int              nri,inr,ggid,iidx,jidx,jnr,outeriter,inneriter;
245     real             shX,shY,shZ,tx,ty,tz,fscal,rcutoff,rcutoff2;
246     int              *iinr,*jindex,*jjnr,*shiftidx,*gid;
247     real             *shiftvec,*fshift,*x,*f;
248     int              vdwioffset0;
249     real             ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
250     int              vdwjidx0;
251     real             jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
252     real             dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00,cexp1_00,cexp2_00;
253     real             velec,felec,velecsum,facel,crf,krf,krf2;
254     real             *charge;
255     int              nvdwtype;
256     real             rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,br,vvdwexp,sh_vdw_invrcut6;
257     int              *vdwtype;
258     real             *vdwparam;
259
260     x                = xx[0];
261     f                = ff[0];
262
263     nri              = nlist->nri;
264     iinr             = nlist->iinr;
265     jindex           = nlist->jindex;
266     jjnr             = nlist->jjnr;
267     shiftidx         = nlist->shift;
268     gid              = nlist->gid;
269     shiftvec         = fr->shift_vec[0];
270     fshift           = fr->fshift[0];
271     facel            = fr->epsfac;
272     charge           = mdatoms->chargeA;
273     nvdwtype         = fr->ntype;
274     vdwparam         = fr->nbfp;
275     vdwtype          = mdatoms->typeA;
276
277     outeriter        = 0;
278     inneriter        = 0;
279
280     /* Start outer loop over neighborlists */
281     for(iidx=0; iidx<nri; iidx++)
282     {
283         /* Load shift vector for this list */
284         i_shift_offset   = DIM*shiftidx[iidx];
285         shX              = shiftvec[i_shift_offset+XX];
286         shY              = shiftvec[i_shift_offset+YY];
287         shZ              = shiftvec[i_shift_offset+ZZ];
288
289         /* Load limits for loop over neighbors */
290         j_index_start    = jindex[iidx];
291         j_index_end      = jindex[iidx+1];
292
293         /* Get outer coordinate index */
294         inr              = iinr[iidx];
295         i_coord_offset   = DIM*inr;
296
297         /* Load i particle coords and add shift vector */
298         ix0              = shX + x[i_coord_offset+DIM*0+XX];
299         iy0              = shY + x[i_coord_offset+DIM*0+YY];
300         iz0              = shZ + x[i_coord_offset+DIM*0+ZZ];
301
302         fix0             = 0.0;
303         fiy0             = 0.0;
304         fiz0             = 0.0;
305
306         /* Load parameters for i particles */
307         iq0              = facel*charge[inr+0];
308         vdwioffset0      = 3*nvdwtype*vdwtype[inr+0];
309
310         /* Start inner kernel loop */
311         for(jidx=j_index_start; jidx<j_index_end; jidx++)
312         {
313             /* Get j neighbor index, and coordinate index */
314             jnr              = jjnr[jidx];
315             j_coord_offset   = DIM*jnr;
316
317             /* load j atom coordinates */
318             jx0              = x[j_coord_offset+DIM*0+XX];
319             jy0              = x[j_coord_offset+DIM*0+YY];
320             jz0              = x[j_coord_offset+DIM*0+ZZ];
321
322             /* Calculate displacement vector */
323             dx00             = ix0 - jx0;
324             dy00             = iy0 - jy0;
325             dz00             = iz0 - jz0;
326
327             /* Calculate squared distance and things based on it */
328             rsq00            = dx00*dx00+dy00*dy00+dz00*dz00;
329
330             rinv00           = gmx_invsqrt(rsq00);
331
332             rinvsq00         = rinv00*rinv00;
333
334             /* Load parameters for j particles */
335             jq0              = charge[jnr+0];
336             vdwjidx0         = 3*vdwtype[jnr+0];
337
338             /**************************
339              * CALCULATE INTERACTIONS *
340              **************************/
341
342             r00              = rsq00*rinv00;
343
344             qq00             = iq0*jq0;
345             c6_00            = vdwparam[vdwioffset0+vdwjidx0];
346             cexp1_00         = vdwparam[vdwioffset0+vdwjidx0+1];
347             cexp2_00         = vdwparam[vdwioffset0+vdwjidx0+2];
348
349             /* COULOMB ELECTROSTATICS */
350             velec            = qq00*rinv00;
351             felec            = velec*rinvsq00;
352
353             /* BUCKINGHAM DISPERSION/REPULSION */
354             rinvsix          = rinvsq00*rinvsq00*rinvsq00;
355             vvdw6            = c6_00*rinvsix;
356             br               = cexp2_00*r00;
357             vvdwexp          = cexp1_00*exp(-br);
358             fvdw             = (br*vvdwexp-vvdw6)*rinvsq00;
359
360             fscal            = felec+fvdw;
361
362             /* Calculate temporary vectorial force */
363             tx               = fscal*dx00;
364             ty               = fscal*dy00;
365             tz               = fscal*dz00;
366
367             /* Update vectorial force */
368             fix0            += tx;
369             fiy0            += ty;
370             fiz0            += tz;
371             f[j_coord_offset+DIM*0+XX] -= tx;
372             f[j_coord_offset+DIM*0+YY] -= ty;
373             f[j_coord_offset+DIM*0+ZZ] -= tz;
374
375             /* Inner loop uses 63 flops */
376         }
377         /* End of innermost loop */
378
379         tx = ty = tz = 0;
380         f[i_coord_offset+DIM*0+XX] += fix0;
381         f[i_coord_offset+DIM*0+YY] += fiy0;
382         f[i_coord_offset+DIM*0+ZZ] += fiz0;
383         tx                         += fix0;
384         ty                         += fiy0;
385         tz                         += fiz0;
386         fshift[i_shift_offset+XX]  += tx;
387         fshift[i_shift_offset+YY]  += ty;
388         fshift[i_shift_offset+ZZ]  += tz;
389
390         /* Increment number of inner iterations */
391         inneriter                  += j_index_end - j_index_start;
392
393         /* Outer loop uses 13 flops */
394     }
395
396     /* Increment number of outer iterations */
397     outeriter        += nri;
398
399     /* Update outer/inner flops */
400
401     inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_F,outeriter*13 + inneriter*63);
402 }