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.
44 #include "../nb_kernel.h"
45 #include "types/simple.h"
46 #include "gromacs/legacyheaders/vec.h"
49 #include "kernelutil_sparc64_hpc_ace_double.h"
52 * Gromacs nonbonded kernel: nb_kernel_ElecRF_VdwNone_GeomP1P1_VF_sparc64_hpc_ace_double
53 * Electrostatics interaction: ReactionField
54 * VdW interaction: None
55 * Geometry: Particle-Particle
56 * Calculate force/pot: PotentialAndForce
59 nb_kernel_ElecRF_VdwNone_GeomP1P1_VF_sparc64_hpc_ace_double
60 (t_nblist * gmx_restrict nlist,
61 rvec * gmx_restrict xx,
62 rvec * gmx_restrict ff,
63 t_forcerec * gmx_restrict fr,
64 t_mdatoms * gmx_restrict mdatoms,
65 nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
66 t_nrnb * gmx_restrict nrnb)
68 /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
69 * just 0 for non-waters.
70 * Suffixes A,B refer to j loop unrolling done with double precision SIMD, e.g. for the two different
71 * jnr indices corresponding to data put in the four positions in the SIMD register.
73 int i_shift_offset,i_coord_offset,outeriter,inneriter;
74 int j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
76 int j_coord_offsetA,j_coord_offsetB;
77 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
79 real *shiftvec,*fshift,*x,*f;
80 _fjsp_v2r8 tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
82 _fjsp_v2r8 ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
83 int vdwjidx0A,vdwjidx0B;
84 _fjsp_v2r8 jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
85 _fjsp_v2r8 dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
86 _fjsp_v2r8 velec,felec,velecsum,facel,crf,krf,krf2;
89 _fjsp_v2r8 dummy_mask,cutoff_mask;
90 _fjsp_v2r8 one = gmx_fjsp_set1_v2r8(1.0);
91 _fjsp_v2r8 two = gmx_fjsp_set1_v2r8(2.0);
92 union { _fjsp_v2r8 simd; long long int i[2]; } vfconv,gbconv,ewconv;
99 jindex = nlist->jindex;
101 shiftidx = nlist->shift;
103 shiftvec = fr->shift_vec[0];
104 fshift = fr->fshift[0];
105 facel = gmx_fjsp_set1_v2r8(fr->epsfac);
106 charge = mdatoms->chargeA;
107 krf = gmx_fjsp_set1_v2r8(fr->ic->k_rf);
108 krf2 = gmx_fjsp_set1_v2r8(fr->ic->k_rf*2.0);
109 crf = gmx_fjsp_set1_v2r8(fr->ic->c_rf);
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 iq0 = _fjsp_mul_v2r8(facel,gmx_fjsp_load1_v2r8(charge+inr+0));
143 /* Reset potential sums */
144 velecsum = _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 rinv00 = gmx_fjsp_invsqrt_v2r8(rsq00);
170 rinvsq00 = _fjsp_mul_v2r8(rinv00,rinv00);
172 /* Load parameters for j particles */
173 jq0 = gmx_fjsp_load_2real_swizzle_v2r8(charge+jnrA+0,charge+jnrB+0);
175 /**************************
176 * CALCULATE INTERACTIONS *
177 **************************/
179 /* Compute parameters for interactions between i and j atoms */
180 qq00 = _fjsp_mul_v2r8(iq0,jq0);
182 /* REACTION-FIELD ELECTROSTATICS */
183 velec = _fjsp_mul_v2r8(qq00,_fjsp_sub_v2r8(_fjsp_madd_v2r8(krf,rsq00,rinv00),crf));
184 felec = _fjsp_mul_v2r8(qq00,_fjsp_msub_v2r8(rinv00,rinvsq00,krf2));
186 /* Update potential sum for this i atom from the interaction with this j atom. */
187 velecsum = _fjsp_add_v2r8(velecsum,velec);
191 /* Update vectorial force */
192 fix0 = _fjsp_madd_v2r8(dx00,fscal,fix0);
193 fiy0 = _fjsp_madd_v2r8(dy00,fscal,fiy0);
194 fiz0 = _fjsp_madd_v2r8(dz00,fscal,fiz0);
196 gmx_fjsp_decrement_fma_1rvec_2ptr_swizzle_v2r8(f+j_coord_offsetA,f+j_coord_offsetB,fscal,dx00,dy00,dz00);
198 /* Inner loop uses 35 flops */
205 j_coord_offsetA = DIM*jnrA;
207 /* load j atom coordinates */
208 gmx_fjsp_load_1rvec_1ptr_swizzle_v2r8(x+j_coord_offsetA,
211 /* Calculate displacement vector */
212 dx00 = _fjsp_sub_v2r8(ix0,jx0);
213 dy00 = _fjsp_sub_v2r8(iy0,jy0);
214 dz00 = _fjsp_sub_v2r8(iz0,jz0);
216 /* Calculate squared distance and things based on it */
217 rsq00 = gmx_fjsp_calc_rsq_v2r8(dx00,dy00,dz00);
219 rinv00 = gmx_fjsp_invsqrt_v2r8(rsq00);
221 rinvsq00 = _fjsp_mul_v2r8(rinv00,rinv00);
223 /* Load parameters for j particles */
224 jq0 = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(),charge+jnrA+0);
226 /**************************
227 * CALCULATE INTERACTIONS *
228 **************************/
230 /* Compute parameters for interactions between i and j atoms */
231 qq00 = _fjsp_mul_v2r8(iq0,jq0);
233 /* REACTION-FIELD ELECTROSTATICS */
234 velec = _fjsp_mul_v2r8(qq00,_fjsp_sub_v2r8(_fjsp_madd_v2r8(krf,rsq00,rinv00),crf));
235 felec = _fjsp_mul_v2r8(qq00,_fjsp_msub_v2r8(rinv00,rinvsq00,krf2));
237 /* Update potential sum for this i atom from the interaction with this j atom. */
238 velec = _fjsp_unpacklo_v2r8(velec,_fjsp_setzero_v2r8());
239 velecsum = _fjsp_add_v2r8(velecsum,velec);
243 fscal = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
245 /* Update vectorial force */
246 fix0 = _fjsp_madd_v2r8(dx00,fscal,fix0);
247 fiy0 = _fjsp_madd_v2r8(dy00,fscal,fiy0);
248 fiz0 = _fjsp_madd_v2r8(dz00,fscal,fiz0);
250 gmx_fjsp_decrement_fma_1rvec_1ptr_swizzle_v2r8(f+j_coord_offsetA,fscal,dx00,dy00,dz00);
252 /* Inner loop uses 35 flops */
255 /* End of innermost loop */
257 gmx_fjsp_update_iforce_1atom_swizzle_v2r8(fix0,fiy0,fiz0,
258 f+i_coord_offset,fshift+i_shift_offset);
261 /* Update potential energies */
262 gmx_fjsp_update_1pot_v2r8(velecsum,kernel_data->energygrp_elec+ggid);
264 /* Increment number of inner iterations */
265 inneriter += j_index_end - j_index_start;
267 /* Outer loop uses 8 flops */
270 /* Increment number of outer iterations */
273 /* Update outer/inner flops */
275 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VF,outeriter*8 + inneriter*35);
278 * Gromacs nonbonded kernel: nb_kernel_ElecRF_VdwNone_GeomP1P1_F_sparc64_hpc_ace_double
279 * Electrostatics interaction: ReactionField
280 * VdW interaction: None
281 * Geometry: Particle-Particle
282 * Calculate force/pot: Force
285 nb_kernel_ElecRF_VdwNone_GeomP1P1_F_sparc64_hpc_ace_double
286 (t_nblist * gmx_restrict nlist,
287 rvec * gmx_restrict xx,
288 rvec * gmx_restrict ff,
289 t_forcerec * gmx_restrict fr,
290 t_mdatoms * gmx_restrict mdatoms,
291 nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
292 t_nrnb * gmx_restrict nrnb)
294 /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
295 * just 0 for non-waters.
296 * Suffixes A,B refer to j loop unrolling done with double precision SIMD, e.g. for the two different
297 * jnr indices corresponding to data put in the four positions in the SIMD register.
299 int i_shift_offset,i_coord_offset,outeriter,inneriter;
300 int j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
302 int j_coord_offsetA,j_coord_offsetB;
303 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
305 real *shiftvec,*fshift,*x,*f;
306 _fjsp_v2r8 tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
308 _fjsp_v2r8 ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
309 int vdwjidx0A,vdwjidx0B;
310 _fjsp_v2r8 jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
311 _fjsp_v2r8 dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
312 _fjsp_v2r8 velec,felec,velecsum,facel,crf,krf,krf2;
315 _fjsp_v2r8 dummy_mask,cutoff_mask;
316 _fjsp_v2r8 one = gmx_fjsp_set1_v2r8(1.0);
317 _fjsp_v2r8 two = gmx_fjsp_set1_v2r8(2.0);
318 union { _fjsp_v2r8 simd; long long int i[2]; } vfconv,gbconv,ewconv;
325 jindex = nlist->jindex;
327 shiftidx = nlist->shift;
329 shiftvec = fr->shift_vec[0];
330 fshift = fr->fshift[0];
331 facel = gmx_fjsp_set1_v2r8(fr->epsfac);
332 charge = mdatoms->chargeA;
333 krf = gmx_fjsp_set1_v2r8(fr->ic->k_rf);
334 krf2 = gmx_fjsp_set1_v2r8(fr->ic->k_rf*2.0);
335 crf = gmx_fjsp_set1_v2r8(fr->ic->c_rf);
337 /* Avoid stupid compiler warnings */
345 /* Start outer loop over neighborlists */
346 for(iidx=0; iidx<nri; iidx++)
348 /* Load shift vector for this list */
349 i_shift_offset = DIM*shiftidx[iidx];
351 /* Load limits for loop over neighbors */
352 j_index_start = jindex[iidx];
353 j_index_end = jindex[iidx+1];
355 /* Get outer coordinate index */
357 i_coord_offset = DIM*inr;
359 /* Load i particle coords and add shift vector */
360 gmx_fjsp_load_shift_and_1rvec_broadcast_v2r8(shiftvec+i_shift_offset,x+i_coord_offset,&ix0,&iy0,&iz0);
362 fix0 = _fjsp_setzero_v2r8();
363 fiy0 = _fjsp_setzero_v2r8();
364 fiz0 = _fjsp_setzero_v2r8();
366 /* Load parameters for i particles */
367 iq0 = _fjsp_mul_v2r8(facel,gmx_fjsp_load1_v2r8(charge+inr+0));
369 /* Start inner kernel loop */
370 for(jidx=j_index_start; jidx<j_index_end-1; jidx+=2)
373 /* Get j neighbor index, and coordinate index */
376 j_coord_offsetA = DIM*jnrA;
377 j_coord_offsetB = DIM*jnrB;
379 /* load j atom coordinates */
380 gmx_fjsp_load_1rvec_2ptr_swizzle_v2r8(x+j_coord_offsetA,x+j_coord_offsetB,
383 /* Calculate displacement vector */
384 dx00 = _fjsp_sub_v2r8(ix0,jx0);
385 dy00 = _fjsp_sub_v2r8(iy0,jy0);
386 dz00 = _fjsp_sub_v2r8(iz0,jz0);
388 /* Calculate squared distance and things based on it */
389 rsq00 = gmx_fjsp_calc_rsq_v2r8(dx00,dy00,dz00);
391 rinv00 = gmx_fjsp_invsqrt_v2r8(rsq00);
393 rinvsq00 = _fjsp_mul_v2r8(rinv00,rinv00);
395 /* Load parameters for j particles */
396 jq0 = gmx_fjsp_load_2real_swizzle_v2r8(charge+jnrA+0,charge+jnrB+0);
398 /**************************
399 * CALCULATE INTERACTIONS *
400 **************************/
402 /* Compute parameters for interactions between i and j atoms */
403 qq00 = _fjsp_mul_v2r8(iq0,jq0);
405 /* REACTION-FIELD ELECTROSTATICS */
406 felec = _fjsp_mul_v2r8(qq00,_fjsp_msub_v2r8(rinv00,rinvsq00,krf2));
410 /* Update vectorial force */
411 fix0 = _fjsp_madd_v2r8(dx00,fscal,fix0);
412 fiy0 = _fjsp_madd_v2r8(dy00,fscal,fiy0);
413 fiz0 = _fjsp_madd_v2r8(dz00,fscal,fiz0);
415 gmx_fjsp_decrement_fma_1rvec_2ptr_swizzle_v2r8(f+j_coord_offsetA,f+j_coord_offsetB,fscal,dx00,dy00,dz00);
417 /* Inner loop uses 30 flops */
424 j_coord_offsetA = DIM*jnrA;
426 /* load j atom coordinates */
427 gmx_fjsp_load_1rvec_1ptr_swizzle_v2r8(x+j_coord_offsetA,
430 /* Calculate displacement vector */
431 dx00 = _fjsp_sub_v2r8(ix0,jx0);
432 dy00 = _fjsp_sub_v2r8(iy0,jy0);
433 dz00 = _fjsp_sub_v2r8(iz0,jz0);
435 /* Calculate squared distance and things based on it */
436 rsq00 = gmx_fjsp_calc_rsq_v2r8(dx00,dy00,dz00);
438 rinv00 = gmx_fjsp_invsqrt_v2r8(rsq00);
440 rinvsq00 = _fjsp_mul_v2r8(rinv00,rinv00);
442 /* Load parameters for j particles */
443 jq0 = _fjsp_loadl_v2r8(_fjsp_setzero_v2r8(),charge+jnrA+0);
445 /**************************
446 * CALCULATE INTERACTIONS *
447 **************************/
449 /* Compute parameters for interactions between i and j atoms */
450 qq00 = _fjsp_mul_v2r8(iq0,jq0);
452 /* REACTION-FIELD ELECTROSTATICS */
453 felec = _fjsp_mul_v2r8(qq00,_fjsp_msub_v2r8(rinv00,rinvsq00,krf2));
457 fscal = _fjsp_unpacklo_v2r8(fscal,_fjsp_setzero_v2r8());
459 /* Update vectorial force */
460 fix0 = _fjsp_madd_v2r8(dx00,fscal,fix0);
461 fiy0 = _fjsp_madd_v2r8(dy00,fscal,fiy0);
462 fiz0 = _fjsp_madd_v2r8(dz00,fscal,fiz0);
464 gmx_fjsp_decrement_fma_1rvec_1ptr_swizzle_v2r8(f+j_coord_offsetA,fscal,dx00,dy00,dz00);
466 /* Inner loop uses 30 flops */
469 /* End of innermost loop */
471 gmx_fjsp_update_iforce_1atom_swizzle_v2r8(fix0,fiy0,fiz0,
472 f+i_coord_offset,fshift+i_shift_offset);
474 /* Increment number of inner iterations */
475 inneriter += j_index_end - j_index_start;
477 /* Outer loop uses 7 flops */
480 /* Increment number of outer iterations */
483 /* Update outer/inner flops */
485 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_F,outeriter*7 + inneriter*30);