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 * Note: this file was generated by the GROMACS sparc64_hpc_ace_double kernel generator.
42 #include "../nb_kernel.h"
43 #include "types/simple.h"
44 #include "gromacs/math/vec.h"
47 #include "kernelutil_sparc64_hpc_ace_double.h"
50 * Gromacs nonbonded kernel: nb_kernel_ElecNone_VdwLJ_GeomP1P1_VF_sparc64_hpc_ace_double
51 * Electrostatics interaction: None
52 * VdW interaction: LennardJones
53 * Geometry: Particle-Particle
54 * Calculate force/pot: PotentialAndForce
57 nb_kernel_ElecNone_VdwLJ_GeomP1P1_VF_sparc64_hpc_ace_double
58 (t_nblist * gmx_restrict nlist,
59 rvec * gmx_restrict xx,
60 rvec * gmx_restrict ff,
61 t_forcerec * gmx_restrict fr,
62 t_mdatoms * gmx_restrict mdatoms,
63 nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
64 t_nrnb * gmx_restrict nrnb)
66 /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
67 * just 0 for non-waters.
68 * Suffixes A,B refer to j loop unrolling done with double precision SIMD, e.g. for the two different
69 * jnr indices corresponding to data put in the four positions in the SIMD register.
71 int i_shift_offset,i_coord_offset,outeriter,inneriter;
72 int j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
74 int j_coord_offsetA,j_coord_offsetB;
75 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
77 real *shiftvec,*fshift,*x,*f;
78 _fjsp_v2r8 tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
80 _fjsp_v2r8 ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
81 int vdwjidx0A,vdwjidx0B;
82 _fjsp_v2r8 jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
83 _fjsp_v2r8 dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
85 _fjsp_v2r8 rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
88 _fjsp_v2r8 one_sixth = gmx_fjsp_set1_v2r8(1.0/6.0);
89 _fjsp_v2r8 one_twelfth = gmx_fjsp_set1_v2r8(1.0/12.0);
91 _fjsp_v2r8 dummy_mask,cutoff_mask;
92 _fjsp_v2r8 one = gmx_fjsp_set1_v2r8(1.0);
93 _fjsp_v2r8 two = gmx_fjsp_set1_v2r8(2.0);
94 union { _fjsp_v2r8 simd; long long int i[2]; } vfconv,gbconv,ewconv;
101 jindex = nlist->jindex;
103 shiftidx = nlist->shift;
105 shiftvec = fr->shift_vec[0];
106 fshift = fr->fshift[0];
107 nvdwtype = fr->ntype;
109 vdwtype = mdatoms->typeA;
111 /* Avoid stupid compiler warnings */
119 /* Start outer loop over neighborlists */
120 for(iidx=0; iidx<nri; iidx++)
122 /* Load shift vector for this list */
123 i_shift_offset = DIM*shiftidx[iidx];
125 /* Load limits for loop over neighbors */
126 j_index_start = jindex[iidx];
127 j_index_end = jindex[iidx+1];
129 /* Get outer coordinate index */
131 i_coord_offset = DIM*inr;
133 /* Load i particle coords and add shift vector */
134 gmx_fjsp_load_shift_and_1rvec_broadcast_v2r8(shiftvec+i_shift_offset,x+i_coord_offset,&ix0,&iy0,&iz0);
136 fix0 = _fjsp_setzero_v2r8();
137 fiy0 = _fjsp_setzero_v2r8();
138 fiz0 = _fjsp_setzero_v2r8();
140 /* Load parameters for i particles */
141 vdwioffset0 = 2*nvdwtype*vdwtype[inr+0];
143 /* Reset potential sums */
144 vvdwsum = _fjsp_setzero_v2r8();
146 /* Start inner kernel loop */
147 for(jidx=j_index_start; jidx<j_index_end-1; jidx+=2)
150 /* Get j neighbor index, and coordinate index */
153 j_coord_offsetA = DIM*jnrA;
154 j_coord_offsetB = DIM*jnrB;
156 /* load j atom coordinates */
157 gmx_fjsp_load_1rvec_2ptr_swizzle_v2r8(x+j_coord_offsetA,x+j_coord_offsetB,
160 /* Calculate displacement vector */
161 dx00 = _fjsp_sub_v2r8(ix0,jx0);
162 dy00 = _fjsp_sub_v2r8(iy0,jy0);
163 dz00 = _fjsp_sub_v2r8(iz0,jz0);
165 /* Calculate squared distance and things based on it */
166 rsq00 = gmx_fjsp_calc_rsq_v2r8(dx00,dy00,dz00);
168 rinvsq00 = gmx_fjsp_inv_v2r8(rsq00);
170 /* Load parameters for j particles */
171 vdwjidx0A = 2*vdwtype[jnrA+0];
172 vdwjidx0B = 2*vdwtype[jnrB+0];
174 /**************************
175 * CALCULATE INTERACTIONS *
176 **************************/
178 /* Compute parameters for interactions between i and j atoms */
179 gmx_fjsp_load_2pair_swizzle_v2r8(vdwparam+vdwioffset0+vdwjidx0A,
180 vdwparam+vdwioffset0+vdwjidx0B,&c6_00,&c12_00);
182 /* LENNARD-JONES DISPERSION/REPULSION */
184 rinvsix = _fjsp_mul_v2r8(_fjsp_mul_v2r8(rinvsq00,rinvsq00),rinvsq00);
185 vvdw6 = _fjsp_mul_v2r8(c6_00,rinvsix);
186 vvdw12 = _fjsp_mul_v2r8(c12_00,_fjsp_mul_v2r8(rinvsix,rinvsix));
187 vvdw = _fjsp_msub_v2r8( vvdw12,one_twelfth, _fjsp_mul_v2r8(vvdw6,one_sixth) );
188 fvdw = _fjsp_mul_v2r8(_fjsp_sub_v2r8(vvdw12,vvdw6),rinvsq00);
190 /* Update potential sum for this i atom from the interaction with this j atom. */
191 vvdwsum = _fjsp_add_v2r8(vvdwsum,vvdw);
195 /* Update vectorial force */
196 fix0 = _fjsp_madd_v2r8(dx00,fscal,fix0);
197 fiy0 = _fjsp_madd_v2r8(dy00,fscal,fiy0);
198 fiz0 = _fjsp_madd_v2r8(dz00,fscal,fiz0);
200 gmx_fjsp_decrement_fma_1rvec_2ptr_swizzle_v2r8(f+j_coord_offsetA,f+j_coord_offsetB,fscal,dx00,dy00,dz00);
202 /* Inner loop uses 35 flops */
209 j_coord_offsetA = DIM*jnrA;
211 /* load j atom coordinates */
212 gmx_fjsp_load_1rvec_1ptr_swizzle_v2r8(x+j_coord_offsetA,
215 /* Calculate displacement vector */
216 dx00 = _fjsp_sub_v2r8(ix0,jx0);
217 dy00 = _fjsp_sub_v2r8(iy0,jy0);
218 dz00 = _fjsp_sub_v2r8(iz0,jz0);
220 /* Calculate squared distance and things based on it */
221 rsq00 = gmx_fjsp_calc_rsq_v2r8(dx00,dy00,dz00);
223 rinvsq00 = gmx_fjsp_inv_v2r8(rsq00);
225 /* Load parameters for j particles */
226 vdwjidx0A = 2*vdwtype[jnrA+0];
228 /**************************
229 * CALCULATE INTERACTIONS *
230 **************************/
232 /* Compute parameters for interactions between i and j atoms */
233 gmx_fjsp_load_1pair_swizzle_v2r8(vdwparam+vdwioffset0+vdwjidx0A,&c6_00,&c12_00);
235 /* LENNARD-JONES DISPERSION/REPULSION */
237 rinvsix = _fjsp_mul_v2r8(_fjsp_mul_v2r8(rinvsq00,rinvsq00),rinvsq00);
238 vvdw6 = _fjsp_mul_v2r8(c6_00,rinvsix);
239 vvdw12 = _fjsp_mul_v2r8(c12_00,_fjsp_mul_v2r8(rinvsix,rinvsix));
240 vvdw = _fjsp_msub_v2r8( vvdw12,one_twelfth, _fjsp_mul_v2r8(vvdw6,one_sixth) );
241 fvdw = _fjsp_mul_v2r8(_fjsp_sub_v2r8(vvdw12,vvdw6),rinvsq00);
243 /* Update potential sum for this i atom from the interaction with this j atom. */
244 vvdw = _fjsp_unpacklo_v2r8(vvdw,_fjsp_setzero_v2r8());
245 vvdwsum = _fjsp_add_v2r8(vvdwsum,vvdw);
249 fscal = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
251 /* Update vectorial force */
252 fix0 = _fjsp_madd_v2r8(dx00,fscal,fix0);
253 fiy0 = _fjsp_madd_v2r8(dy00,fscal,fiy0);
254 fiz0 = _fjsp_madd_v2r8(dz00,fscal,fiz0);
256 gmx_fjsp_decrement_fma_1rvec_1ptr_swizzle_v2r8(f+j_coord_offsetA,fscal,dx00,dy00,dz00);
258 /* Inner loop uses 35 flops */
261 /* End of innermost loop */
263 gmx_fjsp_update_iforce_1atom_swizzle_v2r8(fix0,fiy0,fiz0,
264 f+i_coord_offset,fshift+i_shift_offset);
267 /* Update potential energies */
268 gmx_fjsp_update_1pot_v2r8(vvdwsum,kernel_data->energygrp_vdw+ggid);
270 /* Increment number of inner iterations */
271 inneriter += j_index_end - j_index_start;
273 /* Outer loop uses 7 flops */
276 /* Increment number of outer iterations */
279 /* Update outer/inner flops */
281 inc_nrnb(nrnb,eNR_NBKERNEL_VDW_VF,outeriter*7 + inneriter*35);
284 * Gromacs nonbonded kernel: nb_kernel_ElecNone_VdwLJ_GeomP1P1_F_sparc64_hpc_ace_double
285 * Electrostatics interaction: None
286 * VdW interaction: LennardJones
287 * Geometry: Particle-Particle
288 * Calculate force/pot: Force
291 nb_kernel_ElecNone_VdwLJ_GeomP1P1_F_sparc64_hpc_ace_double
292 (t_nblist * gmx_restrict nlist,
293 rvec * gmx_restrict xx,
294 rvec * gmx_restrict ff,
295 t_forcerec * gmx_restrict fr,
296 t_mdatoms * gmx_restrict mdatoms,
297 nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
298 t_nrnb * gmx_restrict nrnb)
300 /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
301 * just 0 for non-waters.
302 * Suffixes A,B refer to j loop unrolling done with double precision SIMD, e.g. for the two different
303 * jnr indices corresponding to data put in the four positions in the SIMD register.
305 int i_shift_offset,i_coord_offset,outeriter,inneriter;
306 int j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
308 int j_coord_offsetA,j_coord_offsetB;
309 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
311 real *shiftvec,*fshift,*x,*f;
312 _fjsp_v2r8 tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
314 _fjsp_v2r8 ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
315 int vdwjidx0A,vdwjidx0B;
316 _fjsp_v2r8 jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
317 _fjsp_v2r8 dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
319 _fjsp_v2r8 rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
322 _fjsp_v2r8 one_sixth = gmx_fjsp_set1_v2r8(1.0/6.0);
323 _fjsp_v2r8 one_twelfth = gmx_fjsp_set1_v2r8(1.0/12.0);
325 _fjsp_v2r8 dummy_mask,cutoff_mask;
326 _fjsp_v2r8 one = gmx_fjsp_set1_v2r8(1.0);
327 _fjsp_v2r8 two = gmx_fjsp_set1_v2r8(2.0);
328 union { _fjsp_v2r8 simd; long long int i[2]; } vfconv,gbconv,ewconv;
335 jindex = nlist->jindex;
337 shiftidx = nlist->shift;
339 shiftvec = fr->shift_vec[0];
340 fshift = fr->fshift[0];
341 nvdwtype = fr->ntype;
343 vdwtype = mdatoms->typeA;
345 /* Avoid stupid compiler warnings */
353 /* Start outer loop over neighborlists */
354 for(iidx=0; iidx<nri; iidx++)
356 /* Load shift vector for this list */
357 i_shift_offset = DIM*shiftidx[iidx];
359 /* Load limits for loop over neighbors */
360 j_index_start = jindex[iidx];
361 j_index_end = jindex[iidx+1];
363 /* Get outer coordinate index */
365 i_coord_offset = DIM*inr;
367 /* Load i particle coords and add shift vector */
368 gmx_fjsp_load_shift_and_1rvec_broadcast_v2r8(shiftvec+i_shift_offset,x+i_coord_offset,&ix0,&iy0,&iz0);
370 fix0 = _fjsp_setzero_v2r8();
371 fiy0 = _fjsp_setzero_v2r8();
372 fiz0 = _fjsp_setzero_v2r8();
374 /* Load parameters for i particles */
375 vdwioffset0 = 2*nvdwtype*vdwtype[inr+0];
377 /* Start inner kernel loop */
378 for(jidx=j_index_start; jidx<j_index_end-1; jidx+=2)
381 /* Get j neighbor index, and coordinate index */
384 j_coord_offsetA = DIM*jnrA;
385 j_coord_offsetB = DIM*jnrB;
387 /* load j atom coordinates */
388 gmx_fjsp_load_1rvec_2ptr_swizzle_v2r8(x+j_coord_offsetA,x+j_coord_offsetB,
391 /* Calculate displacement vector */
392 dx00 = _fjsp_sub_v2r8(ix0,jx0);
393 dy00 = _fjsp_sub_v2r8(iy0,jy0);
394 dz00 = _fjsp_sub_v2r8(iz0,jz0);
396 /* Calculate squared distance and things based on it */
397 rsq00 = gmx_fjsp_calc_rsq_v2r8(dx00,dy00,dz00);
399 rinvsq00 = gmx_fjsp_inv_v2r8(rsq00);
401 /* Load parameters for j particles */
402 vdwjidx0A = 2*vdwtype[jnrA+0];
403 vdwjidx0B = 2*vdwtype[jnrB+0];
405 /**************************
406 * CALCULATE INTERACTIONS *
407 **************************/
409 /* Compute parameters for interactions between i and j atoms */
410 gmx_fjsp_load_2pair_swizzle_v2r8(vdwparam+vdwioffset0+vdwjidx0A,
411 vdwparam+vdwioffset0+vdwjidx0B,&c6_00,&c12_00);
413 /* LENNARD-JONES DISPERSION/REPULSION */
415 rinvsix = _fjsp_mul_v2r8(_fjsp_mul_v2r8(rinvsq00,rinvsq00),rinvsq00);
416 fvdw = _fjsp_mul_v2r8(_fjsp_msub_v2r8(c12_00,rinvsix,c6_00),_fjsp_mul_v2r8(rinvsix,rinvsq00));
420 /* Update vectorial force */
421 fix0 = _fjsp_madd_v2r8(dx00,fscal,fix0);
422 fiy0 = _fjsp_madd_v2r8(dy00,fscal,fiy0);
423 fiz0 = _fjsp_madd_v2r8(dz00,fscal,fiz0);
425 gmx_fjsp_decrement_fma_1rvec_2ptr_swizzle_v2r8(f+j_coord_offsetA,f+j_coord_offsetB,fscal,dx00,dy00,dz00);
427 /* Inner loop uses 30 flops */
434 j_coord_offsetA = DIM*jnrA;
436 /* load j atom coordinates */
437 gmx_fjsp_load_1rvec_1ptr_swizzle_v2r8(x+j_coord_offsetA,
440 /* Calculate displacement vector */
441 dx00 = _fjsp_sub_v2r8(ix0,jx0);
442 dy00 = _fjsp_sub_v2r8(iy0,jy0);
443 dz00 = _fjsp_sub_v2r8(iz0,jz0);
445 /* Calculate squared distance and things based on it */
446 rsq00 = gmx_fjsp_calc_rsq_v2r8(dx00,dy00,dz00);
448 rinvsq00 = gmx_fjsp_inv_v2r8(rsq00);
450 /* Load parameters for j particles */
451 vdwjidx0A = 2*vdwtype[jnrA+0];
453 /**************************
454 * CALCULATE INTERACTIONS *
455 **************************/
457 /* Compute parameters for interactions between i and j atoms */
458 gmx_fjsp_load_1pair_swizzle_v2r8(vdwparam+vdwioffset0+vdwjidx0A,&c6_00,&c12_00);
460 /* LENNARD-JONES DISPERSION/REPULSION */
462 rinvsix = _fjsp_mul_v2r8(_fjsp_mul_v2r8(rinvsq00,rinvsq00),rinvsq00);
463 fvdw = _fjsp_mul_v2r8(_fjsp_msub_v2r8(c12_00,rinvsix,c6_00),_fjsp_mul_v2r8(rinvsix,rinvsq00));
467 fscal = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
469 /* Update vectorial force */
470 fix0 = _fjsp_madd_v2r8(dx00,fscal,fix0);
471 fiy0 = _fjsp_madd_v2r8(dy00,fscal,fiy0);
472 fiz0 = _fjsp_madd_v2r8(dz00,fscal,fiz0);
474 gmx_fjsp_decrement_fma_1rvec_1ptr_swizzle_v2r8(f+j_coord_offsetA,fscal,dx00,dy00,dz00);
476 /* Inner loop uses 30 flops */
479 /* End of innermost loop */
481 gmx_fjsp_update_iforce_1atom_swizzle_v2r8(fix0,fiy0,fiz0,
482 f+i_coord_offset,fshift+i_shift_offset);
484 /* Increment number of inner iterations */
485 inneriter += j_index_end - j_index_start;
487 /* Outer loop uses 6 flops */
490 /* Increment number of outer iterations */
493 /* Update outer/inner flops */
495 inc_nrnb(nrnb,eNR_NBKERNEL_VDW_F,outeriter*6 + inneriter*30);