2 * This file is part of the GROMACS molecular simulation package.
4 * Copyright (c) 2012,2013,2014,2015,2017,2018, 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 avx_256_single kernel generator.
44 #include "../nb_kernel.h"
45 #include "gromacs/gmxlib/nrnb.h"
47 #include "kernelutil_x86_avx_256_single.h"
50 * Gromacs nonbonded kernel: nb_kernel_ElecRFCut_VdwNone_GeomW4P1_VF_avx_256_single
51 * Electrostatics interaction: ReactionField
52 * VdW interaction: None
53 * Geometry: Water4-Particle
54 * Calculate force/pot: PotentialAndForce
57 nb_kernel_ElecRFCut_VdwNone_GeomW4P1_VF_avx_256_single
58 (t_nblist * gmx_restrict nlist,
59 rvec * gmx_restrict xx,
60 rvec * gmx_restrict ff,
61 struct 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,C,D,E,F,G,H refer to j loop unrolling done with AVX, e.g. for the eight 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;
73 int jnrA,jnrB,jnrC,jnrD;
74 int jnrE,jnrF,jnrG,jnrH;
75 int jnrlistA,jnrlistB,jnrlistC,jnrlistD;
76 int jnrlistE,jnrlistF,jnrlistG,jnrlistH;
77 int j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
78 int j_coord_offsetE,j_coord_offsetF,j_coord_offsetG,j_coord_offsetH;
79 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
81 real *shiftvec,*fshift,*x,*f;
82 real *fjptrA,*fjptrB,*fjptrC,*fjptrD,*fjptrE,*fjptrF,*fjptrG,*fjptrH;
84 __m256 tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
85 real * vdwioffsetptr1;
86 __m256 ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
87 real * vdwioffsetptr2;
88 __m256 ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
89 real * vdwioffsetptr3;
90 __m256 ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
91 int vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D,vdwjidx0E,vdwjidx0F,vdwjidx0G,vdwjidx0H;
92 __m256 jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
93 __m256 dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
94 __m256 dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
95 __m256 dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30;
96 __m256 velec,felec,velecsum,facel,crf,krf,krf2;
98 __m256 dummy_mask,cutoff_mask;
99 __m256 signbit = _mm256_castsi256_ps( _mm256_set1_epi32(0x80000000) );
100 __m256 one = _mm256_set1_ps(1.0);
101 __m256 two = _mm256_set1_ps(2.0);
107 jindex = nlist->jindex;
109 shiftidx = nlist->shift;
111 shiftvec = fr->shift_vec[0];
112 fshift = fr->fshift[0];
113 facel = _mm256_set1_ps(fr->ic->epsfac);
114 charge = mdatoms->chargeA;
115 krf = _mm256_set1_ps(fr->ic->k_rf);
116 krf2 = _mm256_set1_ps(fr->ic->k_rf*2.0);
117 crf = _mm256_set1_ps(fr->ic->c_rf);
119 /* Setup water-specific parameters */
120 inr = nlist->iinr[0];
121 iq1 = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+1]));
122 iq2 = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+2]));
123 iq3 = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+3]));
125 /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
126 rcutoff_scalar = fr->ic->rcoulomb;
127 rcutoff = _mm256_set1_ps(rcutoff_scalar);
128 rcutoff2 = _mm256_mul_ps(rcutoff,rcutoff);
130 /* Avoid stupid compiler warnings */
131 jnrA = jnrB = jnrC = jnrD = jnrE = jnrF = jnrG = jnrH = 0;
144 for(iidx=0;iidx<4*DIM;iidx++)
149 /* Start outer loop over neighborlists */
150 for(iidx=0; iidx<nri; iidx++)
152 /* Load shift vector for this list */
153 i_shift_offset = DIM*shiftidx[iidx];
155 /* Load limits for loop over neighbors */
156 j_index_start = jindex[iidx];
157 j_index_end = jindex[iidx+1];
159 /* Get outer coordinate index */
161 i_coord_offset = DIM*inr;
163 /* Load i particle coords and add shift vector */
164 gmx_mm256_load_shift_and_3rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset+DIM,
165 &ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
167 fix1 = _mm256_setzero_ps();
168 fiy1 = _mm256_setzero_ps();
169 fiz1 = _mm256_setzero_ps();
170 fix2 = _mm256_setzero_ps();
171 fiy2 = _mm256_setzero_ps();
172 fiz2 = _mm256_setzero_ps();
173 fix3 = _mm256_setzero_ps();
174 fiy3 = _mm256_setzero_ps();
175 fiz3 = _mm256_setzero_ps();
177 /* Reset potential sums */
178 velecsum = _mm256_setzero_ps();
180 /* Start inner kernel loop */
181 for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+7]>=0; jidx+=8)
184 /* Get j neighbor index, and coordinate index */
193 j_coord_offsetA = DIM*jnrA;
194 j_coord_offsetB = DIM*jnrB;
195 j_coord_offsetC = DIM*jnrC;
196 j_coord_offsetD = DIM*jnrD;
197 j_coord_offsetE = DIM*jnrE;
198 j_coord_offsetF = DIM*jnrF;
199 j_coord_offsetG = DIM*jnrG;
200 j_coord_offsetH = DIM*jnrH;
202 /* load j atom coordinates */
203 gmx_mm256_load_1rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
204 x+j_coord_offsetC,x+j_coord_offsetD,
205 x+j_coord_offsetE,x+j_coord_offsetF,
206 x+j_coord_offsetG,x+j_coord_offsetH,
209 /* Calculate displacement vector */
210 dx10 = _mm256_sub_ps(ix1,jx0);
211 dy10 = _mm256_sub_ps(iy1,jy0);
212 dz10 = _mm256_sub_ps(iz1,jz0);
213 dx20 = _mm256_sub_ps(ix2,jx0);
214 dy20 = _mm256_sub_ps(iy2,jy0);
215 dz20 = _mm256_sub_ps(iz2,jz0);
216 dx30 = _mm256_sub_ps(ix3,jx0);
217 dy30 = _mm256_sub_ps(iy3,jy0);
218 dz30 = _mm256_sub_ps(iz3,jz0);
220 /* Calculate squared distance and things based on it */
221 rsq10 = gmx_mm256_calc_rsq_ps(dx10,dy10,dz10);
222 rsq20 = gmx_mm256_calc_rsq_ps(dx20,dy20,dz20);
223 rsq30 = gmx_mm256_calc_rsq_ps(dx30,dy30,dz30);
225 rinv10 = avx256_invsqrt_f(rsq10);
226 rinv20 = avx256_invsqrt_f(rsq20);
227 rinv30 = avx256_invsqrt_f(rsq30);
229 rinvsq10 = _mm256_mul_ps(rinv10,rinv10);
230 rinvsq20 = _mm256_mul_ps(rinv20,rinv20);
231 rinvsq30 = _mm256_mul_ps(rinv30,rinv30);
233 /* Load parameters for j particles */
234 jq0 = gmx_mm256_load_8real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
235 charge+jnrC+0,charge+jnrD+0,
236 charge+jnrE+0,charge+jnrF+0,
237 charge+jnrG+0,charge+jnrH+0);
239 fjx0 = _mm256_setzero_ps();
240 fjy0 = _mm256_setzero_ps();
241 fjz0 = _mm256_setzero_ps();
243 /**************************
244 * CALCULATE INTERACTIONS *
245 **************************/
247 if (gmx_mm256_any_lt(rsq10,rcutoff2))
250 /* Compute parameters for interactions between i and j atoms */
251 qq10 = _mm256_mul_ps(iq1,jq0);
253 /* REACTION-FIELD ELECTROSTATICS */
254 velec = _mm256_mul_ps(qq10,_mm256_sub_ps(_mm256_add_ps(rinv10,_mm256_mul_ps(krf,rsq10)),crf));
255 felec = _mm256_mul_ps(qq10,_mm256_sub_ps(_mm256_mul_ps(rinv10,rinvsq10),krf2));
257 cutoff_mask = _mm256_cmp_ps(rsq10,rcutoff2,_CMP_LT_OQ);
259 /* Update potential sum for this i atom from the interaction with this j atom. */
260 velec = _mm256_and_ps(velec,cutoff_mask);
261 velecsum = _mm256_add_ps(velecsum,velec);
265 fscal = _mm256_and_ps(fscal,cutoff_mask);
267 /* Calculate temporary vectorial force */
268 tx = _mm256_mul_ps(fscal,dx10);
269 ty = _mm256_mul_ps(fscal,dy10);
270 tz = _mm256_mul_ps(fscal,dz10);
272 /* Update vectorial force */
273 fix1 = _mm256_add_ps(fix1,tx);
274 fiy1 = _mm256_add_ps(fiy1,ty);
275 fiz1 = _mm256_add_ps(fiz1,tz);
277 fjx0 = _mm256_add_ps(fjx0,tx);
278 fjy0 = _mm256_add_ps(fjy0,ty);
279 fjz0 = _mm256_add_ps(fjz0,tz);
283 /**************************
284 * CALCULATE INTERACTIONS *
285 **************************/
287 if (gmx_mm256_any_lt(rsq20,rcutoff2))
290 /* Compute parameters for interactions between i and j atoms */
291 qq20 = _mm256_mul_ps(iq2,jq0);
293 /* REACTION-FIELD ELECTROSTATICS */
294 velec = _mm256_mul_ps(qq20,_mm256_sub_ps(_mm256_add_ps(rinv20,_mm256_mul_ps(krf,rsq20)),crf));
295 felec = _mm256_mul_ps(qq20,_mm256_sub_ps(_mm256_mul_ps(rinv20,rinvsq20),krf2));
297 cutoff_mask = _mm256_cmp_ps(rsq20,rcutoff2,_CMP_LT_OQ);
299 /* Update potential sum for this i atom from the interaction with this j atom. */
300 velec = _mm256_and_ps(velec,cutoff_mask);
301 velecsum = _mm256_add_ps(velecsum,velec);
305 fscal = _mm256_and_ps(fscal,cutoff_mask);
307 /* Calculate temporary vectorial force */
308 tx = _mm256_mul_ps(fscal,dx20);
309 ty = _mm256_mul_ps(fscal,dy20);
310 tz = _mm256_mul_ps(fscal,dz20);
312 /* Update vectorial force */
313 fix2 = _mm256_add_ps(fix2,tx);
314 fiy2 = _mm256_add_ps(fiy2,ty);
315 fiz2 = _mm256_add_ps(fiz2,tz);
317 fjx0 = _mm256_add_ps(fjx0,tx);
318 fjy0 = _mm256_add_ps(fjy0,ty);
319 fjz0 = _mm256_add_ps(fjz0,tz);
323 /**************************
324 * CALCULATE INTERACTIONS *
325 **************************/
327 if (gmx_mm256_any_lt(rsq30,rcutoff2))
330 /* Compute parameters for interactions between i and j atoms */
331 qq30 = _mm256_mul_ps(iq3,jq0);
333 /* REACTION-FIELD ELECTROSTATICS */
334 velec = _mm256_mul_ps(qq30,_mm256_sub_ps(_mm256_add_ps(rinv30,_mm256_mul_ps(krf,rsq30)),crf));
335 felec = _mm256_mul_ps(qq30,_mm256_sub_ps(_mm256_mul_ps(rinv30,rinvsq30),krf2));
337 cutoff_mask = _mm256_cmp_ps(rsq30,rcutoff2,_CMP_LT_OQ);
339 /* Update potential sum for this i atom from the interaction with this j atom. */
340 velec = _mm256_and_ps(velec,cutoff_mask);
341 velecsum = _mm256_add_ps(velecsum,velec);
345 fscal = _mm256_and_ps(fscal,cutoff_mask);
347 /* Calculate temporary vectorial force */
348 tx = _mm256_mul_ps(fscal,dx30);
349 ty = _mm256_mul_ps(fscal,dy30);
350 tz = _mm256_mul_ps(fscal,dz30);
352 /* Update vectorial force */
353 fix3 = _mm256_add_ps(fix3,tx);
354 fiy3 = _mm256_add_ps(fiy3,ty);
355 fiz3 = _mm256_add_ps(fiz3,tz);
357 fjx0 = _mm256_add_ps(fjx0,tx);
358 fjy0 = _mm256_add_ps(fjy0,ty);
359 fjz0 = _mm256_add_ps(fjz0,tz);
363 fjptrA = f+j_coord_offsetA;
364 fjptrB = f+j_coord_offsetB;
365 fjptrC = f+j_coord_offsetC;
366 fjptrD = f+j_coord_offsetD;
367 fjptrE = f+j_coord_offsetE;
368 fjptrF = f+j_coord_offsetF;
369 fjptrG = f+j_coord_offsetG;
370 fjptrH = f+j_coord_offsetH;
372 gmx_mm256_decrement_1rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,fjx0,fjy0,fjz0);
374 /* Inner loop uses 111 flops */
380 /* Get j neighbor index, and coordinate index */
381 jnrlistA = jjnr[jidx];
382 jnrlistB = jjnr[jidx+1];
383 jnrlistC = jjnr[jidx+2];
384 jnrlistD = jjnr[jidx+3];
385 jnrlistE = jjnr[jidx+4];
386 jnrlistF = jjnr[jidx+5];
387 jnrlistG = jjnr[jidx+6];
388 jnrlistH = jjnr[jidx+7];
389 /* Sign of each element will be negative for non-real atoms.
390 * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
391 * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
393 dummy_mask = gmx_mm256_set_m128(gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx+4)),_mm_setzero_si128())),
394 gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128())));
396 jnrA = (jnrlistA>=0) ? jnrlistA : 0;
397 jnrB = (jnrlistB>=0) ? jnrlistB : 0;
398 jnrC = (jnrlistC>=0) ? jnrlistC : 0;
399 jnrD = (jnrlistD>=0) ? jnrlistD : 0;
400 jnrE = (jnrlistE>=0) ? jnrlistE : 0;
401 jnrF = (jnrlistF>=0) ? jnrlistF : 0;
402 jnrG = (jnrlistG>=0) ? jnrlistG : 0;
403 jnrH = (jnrlistH>=0) ? jnrlistH : 0;
404 j_coord_offsetA = DIM*jnrA;
405 j_coord_offsetB = DIM*jnrB;
406 j_coord_offsetC = DIM*jnrC;
407 j_coord_offsetD = DIM*jnrD;
408 j_coord_offsetE = DIM*jnrE;
409 j_coord_offsetF = DIM*jnrF;
410 j_coord_offsetG = DIM*jnrG;
411 j_coord_offsetH = DIM*jnrH;
413 /* load j atom coordinates */
414 gmx_mm256_load_1rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
415 x+j_coord_offsetC,x+j_coord_offsetD,
416 x+j_coord_offsetE,x+j_coord_offsetF,
417 x+j_coord_offsetG,x+j_coord_offsetH,
420 /* Calculate displacement vector */
421 dx10 = _mm256_sub_ps(ix1,jx0);
422 dy10 = _mm256_sub_ps(iy1,jy0);
423 dz10 = _mm256_sub_ps(iz1,jz0);
424 dx20 = _mm256_sub_ps(ix2,jx0);
425 dy20 = _mm256_sub_ps(iy2,jy0);
426 dz20 = _mm256_sub_ps(iz2,jz0);
427 dx30 = _mm256_sub_ps(ix3,jx0);
428 dy30 = _mm256_sub_ps(iy3,jy0);
429 dz30 = _mm256_sub_ps(iz3,jz0);
431 /* Calculate squared distance and things based on it */
432 rsq10 = gmx_mm256_calc_rsq_ps(dx10,dy10,dz10);
433 rsq20 = gmx_mm256_calc_rsq_ps(dx20,dy20,dz20);
434 rsq30 = gmx_mm256_calc_rsq_ps(dx30,dy30,dz30);
436 rinv10 = avx256_invsqrt_f(rsq10);
437 rinv20 = avx256_invsqrt_f(rsq20);
438 rinv30 = avx256_invsqrt_f(rsq30);
440 rinvsq10 = _mm256_mul_ps(rinv10,rinv10);
441 rinvsq20 = _mm256_mul_ps(rinv20,rinv20);
442 rinvsq30 = _mm256_mul_ps(rinv30,rinv30);
444 /* Load parameters for j particles */
445 jq0 = gmx_mm256_load_8real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
446 charge+jnrC+0,charge+jnrD+0,
447 charge+jnrE+0,charge+jnrF+0,
448 charge+jnrG+0,charge+jnrH+0);
450 fjx0 = _mm256_setzero_ps();
451 fjy0 = _mm256_setzero_ps();
452 fjz0 = _mm256_setzero_ps();
454 /**************************
455 * CALCULATE INTERACTIONS *
456 **************************/
458 if (gmx_mm256_any_lt(rsq10,rcutoff2))
461 /* Compute parameters for interactions between i and j atoms */
462 qq10 = _mm256_mul_ps(iq1,jq0);
464 /* REACTION-FIELD ELECTROSTATICS */
465 velec = _mm256_mul_ps(qq10,_mm256_sub_ps(_mm256_add_ps(rinv10,_mm256_mul_ps(krf,rsq10)),crf));
466 felec = _mm256_mul_ps(qq10,_mm256_sub_ps(_mm256_mul_ps(rinv10,rinvsq10),krf2));
468 cutoff_mask = _mm256_cmp_ps(rsq10,rcutoff2,_CMP_LT_OQ);
470 /* Update potential sum for this i atom from the interaction with this j atom. */
471 velec = _mm256_and_ps(velec,cutoff_mask);
472 velec = _mm256_andnot_ps(dummy_mask,velec);
473 velecsum = _mm256_add_ps(velecsum,velec);
477 fscal = _mm256_and_ps(fscal,cutoff_mask);
479 fscal = _mm256_andnot_ps(dummy_mask,fscal);
481 /* Calculate temporary vectorial force */
482 tx = _mm256_mul_ps(fscal,dx10);
483 ty = _mm256_mul_ps(fscal,dy10);
484 tz = _mm256_mul_ps(fscal,dz10);
486 /* Update vectorial force */
487 fix1 = _mm256_add_ps(fix1,tx);
488 fiy1 = _mm256_add_ps(fiy1,ty);
489 fiz1 = _mm256_add_ps(fiz1,tz);
491 fjx0 = _mm256_add_ps(fjx0,tx);
492 fjy0 = _mm256_add_ps(fjy0,ty);
493 fjz0 = _mm256_add_ps(fjz0,tz);
497 /**************************
498 * CALCULATE INTERACTIONS *
499 **************************/
501 if (gmx_mm256_any_lt(rsq20,rcutoff2))
504 /* Compute parameters for interactions between i and j atoms */
505 qq20 = _mm256_mul_ps(iq2,jq0);
507 /* REACTION-FIELD ELECTROSTATICS */
508 velec = _mm256_mul_ps(qq20,_mm256_sub_ps(_mm256_add_ps(rinv20,_mm256_mul_ps(krf,rsq20)),crf));
509 felec = _mm256_mul_ps(qq20,_mm256_sub_ps(_mm256_mul_ps(rinv20,rinvsq20),krf2));
511 cutoff_mask = _mm256_cmp_ps(rsq20,rcutoff2,_CMP_LT_OQ);
513 /* Update potential sum for this i atom from the interaction with this j atom. */
514 velec = _mm256_and_ps(velec,cutoff_mask);
515 velec = _mm256_andnot_ps(dummy_mask,velec);
516 velecsum = _mm256_add_ps(velecsum,velec);
520 fscal = _mm256_and_ps(fscal,cutoff_mask);
522 fscal = _mm256_andnot_ps(dummy_mask,fscal);
524 /* Calculate temporary vectorial force */
525 tx = _mm256_mul_ps(fscal,dx20);
526 ty = _mm256_mul_ps(fscal,dy20);
527 tz = _mm256_mul_ps(fscal,dz20);
529 /* Update vectorial force */
530 fix2 = _mm256_add_ps(fix2,tx);
531 fiy2 = _mm256_add_ps(fiy2,ty);
532 fiz2 = _mm256_add_ps(fiz2,tz);
534 fjx0 = _mm256_add_ps(fjx0,tx);
535 fjy0 = _mm256_add_ps(fjy0,ty);
536 fjz0 = _mm256_add_ps(fjz0,tz);
540 /**************************
541 * CALCULATE INTERACTIONS *
542 **************************/
544 if (gmx_mm256_any_lt(rsq30,rcutoff2))
547 /* Compute parameters for interactions between i and j atoms */
548 qq30 = _mm256_mul_ps(iq3,jq0);
550 /* REACTION-FIELD ELECTROSTATICS */
551 velec = _mm256_mul_ps(qq30,_mm256_sub_ps(_mm256_add_ps(rinv30,_mm256_mul_ps(krf,rsq30)),crf));
552 felec = _mm256_mul_ps(qq30,_mm256_sub_ps(_mm256_mul_ps(rinv30,rinvsq30),krf2));
554 cutoff_mask = _mm256_cmp_ps(rsq30,rcutoff2,_CMP_LT_OQ);
556 /* Update potential sum for this i atom from the interaction with this j atom. */
557 velec = _mm256_and_ps(velec,cutoff_mask);
558 velec = _mm256_andnot_ps(dummy_mask,velec);
559 velecsum = _mm256_add_ps(velecsum,velec);
563 fscal = _mm256_and_ps(fscal,cutoff_mask);
565 fscal = _mm256_andnot_ps(dummy_mask,fscal);
567 /* Calculate temporary vectorial force */
568 tx = _mm256_mul_ps(fscal,dx30);
569 ty = _mm256_mul_ps(fscal,dy30);
570 tz = _mm256_mul_ps(fscal,dz30);
572 /* Update vectorial force */
573 fix3 = _mm256_add_ps(fix3,tx);
574 fiy3 = _mm256_add_ps(fiy3,ty);
575 fiz3 = _mm256_add_ps(fiz3,tz);
577 fjx0 = _mm256_add_ps(fjx0,tx);
578 fjy0 = _mm256_add_ps(fjy0,ty);
579 fjz0 = _mm256_add_ps(fjz0,tz);
583 fjptrA = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
584 fjptrB = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
585 fjptrC = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
586 fjptrD = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
587 fjptrE = (jnrlistE>=0) ? f+j_coord_offsetE : scratch;
588 fjptrF = (jnrlistF>=0) ? f+j_coord_offsetF : scratch;
589 fjptrG = (jnrlistG>=0) ? f+j_coord_offsetG : scratch;
590 fjptrH = (jnrlistH>=0) ? f+j_coord_offsetH : scratch;
592 gmx_mm256_decrement_1rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,fjx0,fjy0,fjz0);
594 /* Inner loop uses 111 flops */
597 /* End of innermost loop */
599 gmx_mm256_update_iforce_3atom_swizzle_ps(fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
600 f+i_coord_offset+DIM,fshift+i_shift_offset);
603 /* Update potential energies */
604 gmx_mm256_update_1pot_ps(velecsum,kernel_data->energygrp_elec+ggid);
606 /* Increment number of inner iterations */
607 inneriter += j_index_end - j_index_start;
609 /* Outer loop uses 19 flops */
612 /* Increment number of outer iterations */
615 /* Update outer/inner flops */
617 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_W4_VF,outeriter*19 + inneriter*111);
620 * Gromacs nonbonded kernel: nb_kernel_ElecRFCut_VdwNone_GeomW4P1_F_avx_256_single
621 * Electrostatics interaction: ReactionField
622 * VdW interaction: None
623 * Geometry: Water4-Particle
624 * Calculate force/pot: Force
627 nb_kernel_ElecRFCut_VdwNone_GeomW4P1_F_avx_256_single
628 (t_nblist * gmx_restrict nlist,
629 rvec * gmx_restrict xx,
630 rvec * gmx_restrict ff,
631 struct t_forcerec * gmx_restrict fr,
632 t_mdatoms * gmx_restrict mdatoms,
633 nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
634 t_nrnb * gmx_restrict nrnb)
636 /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
637 * just 0 for non-waters.
638 * Suffixes A,B,C,D,E,F,G,H refer to j loop unrolling done with AVX, e.g. for the eight different
639 * jnr indices corresponding to data put in the four positions in the SIMD register.
641 int i_shift_offset,i_coord_offset,outeriter,inneriter;
642 int j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
643 int jnrA,jnrB,jnrC,jnrD;
644 int jnrE,jnrF,jnrG,jnrH;
645 int jnrlistA,jnrlistB,jnrlistC,jnrlistD;
646 int jnrlistE,jnrlistF,jnrlistG,jnrlistH;
647 int j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
648 int j_coord_offsetE,j_coord_offsetF,j_coord_offsetG,j_coord_offsetH;
649 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
651 real *shiftvec,*fshift,*x,*f;
652 real *fjptrA,*fjptrB,*fjptrC,*fjptrD,*fjptrE,*fjptrF,*fjptrG,*fjptrH;
654 __m256 tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
655 real * vdwioffsetptr1;
656 __m256 ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
657 real * vdwioffsetptr2;
658 __m256 ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
659 real * vdwioffsetptr3;
660 __m256 ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
661 int vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D,vdwjidx0E,vdwjidx0F,vdwjidx0G,vdwjidx0H;
662 __m256 jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
663 __m256 dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
664 __m256 dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
665 __m256 dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30;
666 __m256 velec,felec,velecsum,facel,crf,krf,krf2;
668 __m256 dummy_mask,cutoff_mask;
669 __m256 signbit = _mm256_castsi256_ps( _mm256_set1_epi32(0x80000000) );
670 __m256 one = _mm256_set1_ps(1.0);
671 __m256 two = _mm256_set1_ps(2.0);
677 jindex = nlist->jindex;
679 shiftidx = nlist->shift;
681 shiftvec = fr->shift_vec[0];
682 fshift = fr->fshift[0];
683 facel = _mm256_set1_ps(fr->ic->epsfac);
684 charge = mdatoms->chargeA;
685 krf = _mm256_set1_ps(fr->ic->k_rf);
686 krf2 = _mm256_set1_ps(fr->ic->k_rf*2.0);
687 crf = _mm256_set1_ps(fr->ic->c_rf);
689 /* Setup water-specific parameters */
690 inr = nlist->iinr[0];
691 iq1 = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+1]));
692 iq2 = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+2]));
693 iq3 = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+3]));
695 /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
696 rcutoff_scalar = fr->ic->rcoulomb;
697 rcutoff = _mm256_set1_ps(rcutoff_scalar);
698 rcutoff2 = _mm256_mul_ps(rcutoff,rcutoff);
700 /* Avoid stupid compiler warnings */
701 jnrA = jnrB = jnrC = jnrD = jnrE = jnrF = jnrG = jnrH = 0;
714 for(iidx=0;iidx<4*DIM;iidx++)
719 /* Start outer loop over neighborlists */
720 for(iidx=0; iidx<nri; iidx++)
722 /* Load shift vector for this list */
723 i_shift_offset = DIM*shiftidx[iidx];
725 /* Load limits for loop over neighbors */
726 j_index_start = jindex[iidx];
727 j_index_end = jindex[iidx+1];
729 /* Get outer coordinate index */
731 i_coord_offset = DIM*inr;
733 /* Load i particle coords and add shift vector */
734 gmx_mm256_load_shift_and_3rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset+DIM,
735 &ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
737 fix1 = _mm256_setzero_ps();
738 fiy1 = _mm256_setzero_ps();
739 fiz1 = _mm256_setzero_ps();
740 fix2 = _mm256_setzero_ps();
741 fiy2 = _mm256_setzero_ps();
742 fiz2 = _mm256_setzero_ps();
743 fix3 = _mm256_setzero_ps();
744 fiy3 = _mm256_setzero_ps();
745 fiz3 = _mm256_setzero_ps();
747 /* Start inner kernel loop */
748 for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+7]>=0; jidx+=8)
751 /* Get j neighbor index, and coordinate index */
760 j_coord_offsetA = DIM*jnrA;
761 j_coord_offsetB = DIM*jnrB;
762 j_coord_offsetC = DIM*jnrC;
763 j_coord_offsetD = DIM*jnrD;
764 j_coord_offsetE = DIM*jnrE;
765 j_coord_offsetF = DIM*jnrF;
766 j_coord_offsetG = DIM*jnrG;
767 j_coord_offsetH = DIM*jnrH;
769 /* load j atom coordinates */
770 gmx_mm256_load_1rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
771 x+j_coord_offsetC,x+j_coord_offsetD,
772 x+j_coord_offsetE,x+j_coord_offsetF,
773 x+j_coord_offsetG,x+j_coord_offsetH,
776 /* Calculate displacement vector */
777 dx10 = _mm256_sub_ps(ix1,jx0);
778 dy10 = _mm256_sub_ps(iy1,jy0);
779 dz10 = _mm256_sub_ps(iz1,jz0);
780 dx20 = _mm256_sub_ps(ix2,jx0);
781 dy20 = _mm256_sub_ps(iy2,jy0);
782 dz20 = _mm256_sub_ps(iz2,jz0);
783 dx30 = _mm256_sub_ps(ix3,jx0);
784 dy30 = _mm256_sub_ps(iy3,jy0);
785 dz30 = _mm256_sub_ps(iz3,jz0);
787 /* Calculate squared distance and things based on it */
788 rsq10 = gmx_mm256_calc_rsq_ps(dx10,dy10,dz10);
789 rsq20 = gmx_mm256_calc_rsq_ps(dx20,dy20,dz20);
790 rsq30 = gmx_mm256_calc_rsq_ps(dx30,dy30,dz30);
792 rinv10 = avx256_invsqrt_f(rsq10);
793 rinv20 = avx256_invsqrt_f(rsq20);
794 rinv30 = avx256_invsqrt_f(rsq30);
796 rinvsq10 = _mm256_mul_ps(rinv10,rinv10);
797 rinvsq20 = _mm256_mul_ps(rinv20,rinv20);
798 rinvsq30 = _mm256_mul_ps(rinv30,rinv30);
800 /* Load parameters for j particles */
801 jq0 = gmx_mm256_load_8real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
802 charge+jnrC+0,charge+jnrD+0,
803 charge+jnrE+0,charge+jnrF+0,
804 charge+jnrG+0,charge+jnrH+0);
806 fjx0 = _mm256_setzero_ps();
807 fjy0 = _mm256_setzero_ps();
808 fjz0 = _mm256_setzero_ps();
810 /**************************
811 * CALCULATE INTERACTIONS *
812 **************************/
814 if (gmx_mm256_any_lt(rsq10,rcutoff2))
817 /* Compute parameters for interactions between i and j atoms */
818 qq10 = _mm256_mul_ps(iq1,jq0);
820 /* REACTION-FIELD ELECTROSTATICS */
821 felec = _mm256_mul_ps(qq10,_mm256_sub_ps(_mm256_mul_ps(rinv10,rinvsq10),krf2));
823 cutoff_mask = _mm256_cmp_ps(rsq10,rcutoff2,_CMP_LT_OQ);
827 fscal = _mm256_and_ps(fscal,cutoff_mask);
829 /* Calculate temporary vectorial force */
830 tx = _mm256_mul_ps(fscal,dx10);
831 ty = _mm256_mul_ps(fscal,dy10);
832 tz = _mm256_mul_ps(fscal,dz10);
834 /* Update vectorial force */
835 fix1 = _mm256_add_ps(fix1,tx);
836 fiy1 = _mm256_add_ps(fiy1,ty);
837 fiz1 = _mm256_add_ps(fiz1,tz);
839 fjx0 = _mm256_add_ps(fjx0,tx);
840 fjy0 = _mm256_add_ps(fjy0,ty);
841 fjz0 = _mm256_add_ps(fjz0,tz);
845 /**************************
846 * CALCULATE INTERACTIONS *
847 **************************/
849 if (gmx_mm256_any_lt(rsq20,rcutoff2))
852 /* Compute parameters for interactions between i and j atoms */
853 qq20 = _mm256_mul_ps(iq2,jq0);
855 /* REACTION-FIELD ELECTROSTATICS */
856 felec = _mm256_mul_ps(qq20,_mm256_sub_ps(_mm256_mul_ps(rinv20,rinvsq20),krf2));
858 cutoff_mask = _mm256_cmp_ps(rsq20,rcutoff2,_CMP_LT_OQ);
862 fscal = _mm256_and_ps(fscal,cutoff_mask);
864 /* Calculate temporary vectorial force */
865 tx = _mm256_mul_ps(fscal,dx20);
866 ty = _mm256_mul_ps(fscal,dy20);
867 tz = _mm256_mul_ps(fscal,dz20);
869 /* Update vectorial force */
870 fix2 = _mm256_add_ps(fix2,tx);
871 fiy2 = _mm256_add_ps(fiy2,ty);
872 fiz2 = _mm256_add_ps(fiz2,tz);
874 fjx0 = _mm256_add_ps(fjx0,tx);
875 fjy0 = _mm256_add_ps(fjy0,ty);
876 fjz0 = _mm256_add_ps(fjz0,tz);
880 /**************************
881 * CALCULATE INTERACTIONS *
882 **************************/
884 if (gmx_mm256_any_lt(rsq30,rcutoff2))
887 /* Compute parameters for interactions between i and j atoms */
888 qq30 = _mm256_mul_ps(iq3,jq0);
890 /* REACTION-FIELD ELECTROSTATICS */
891 felec = _mm256_mul_ps(qq30,_mm256_sub_ps(_mm256_mul_ps(rinv30,rinvsq30),krf2));
893 cutoff_mask = _mm256_cmp_ps(rsq30,rcutoff2,_CMP_LT_OQ);
897 fscal = _mm256_and_ps(fscal,cutoff_mask);
899 /* Calculate temporary vectorial force */
900 tx = _mm256_mul_ps(fscal,dx30);
901 ty = _mm256_mul_ps(fscal,dy30);
902 tz = _mm256_mul_ps(fscal,dz30);
904 /* Update vectorial force */
905 fix3 = _mm256_add_ps(fix3,tx);
906 fiy3 = _mm256_add_ps(fiy3,ty);
907 fiz3 = _mm256_add_ps(fiz3,tz);
909 fjx0 = _mm256_add_ps(fjx0,tx);
910 fjy0 = _mm256_add_ps(fjy0,ty);
911 fjz0 = _mm256_add_ps(fjz0,tz);
915 fjptrA = f+j_coord_offsetA;
916 fjptrB = f+j_coord_offsetB;
917 fjptrC = f+j_coord_offsetC;
918 fjptrD = f+j_coord_offsetD;
919 fjptrE = f+j_coord_offsetE;
920 fjptrF = f+j_coord_offsetF;
921 fjptrG = f+j_coord_offsetG;
922 fjptrH = f+j_coord_offsetH;
924 gmx_mm256_decrement_1rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,fjx0,fjy0,fjz0);
926 /* Inner loop uses 93 flops */
932 /* Get j neighbor index, and coordinate index */
933 jnrlistA = jjnr[jidx];
934 jnrlistB = jjnr[jidx+1];
935 jnrlistC = jjnr[jidx+2];
936 jnrlistD = jjnr[jidx+3];
937 jnrlistE = jjnr[jidx+4];
938 jnrlistF = jjnr[jidx+5];
939 jnrlistG = jjnr[jidx+6];
940 jnrlistH = jjnr[jidx+7];
941 /* Sign of each element will be negative for non-real atoms.
942 * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
943 * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
945 dummy_mask = gmx_mm256_set_m128(gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx+4)),_mm_setzero_si128())),
946 gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128())));
948 jnrA = (jnrlistA>=0) ? jnrlistA : 0;
949 jnrB = (jnrlistB>=0) ? jnrlistB : 0;
950 jnrC = (jnrlistC>=0) ? jnrlistC : 0;
951 jnrD = (jnrlistD>=0) ? jnrlistD : 0;
952 jnrE = (jnrlistE>=0) ? jnrlistE : 0;
953 jnrF = (jnrlistF>=0) ? jnrlistF : 0;
954 jnrG = (jnrlistG>=0) ? jnrlistG : 0;
955 jnrH = (jnrlistH>=0) ? jnrlistH : 0;
956 j_coord_offsetA = DIM*jnrA;
957 j_coord_offsetB = DIM*jnrB;
958 j_coord_offsetC = DIM*jnrC;
959 j_coord_offsetD = DIM*jnrD;
960 j_coord_offsetE = DIM*jnrE;
961 j_coord_offsetF = DIM*jnrF;
962 j_coord_offsetG = DIM*jnrG;
963 j_coord_offsetH = DIM*jnrH;
965 /* load j atom coordinates */
966 gmx_mm256_load_1rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
967 x+j_coord_offsetC,x+j_coord_offsetD,
968 x+j_coord_offsetE,x+j_coord_offsetF,
969 x+j_coord_offsetG,x+j_coord_offsetH,
972 /* Calculate displacement vector */
973 dx10 = _mm256_sub_ps(ix1,jx0);
974 dy10 = _mm256_sub_ps(iy1,jy0);
975 dz10 = _mm256_sub_ps(iz1,jz0);
976 dx20 = _mm256_sub_ps(ix2,jx0);
977 dy20 = _mm256_sub_ps(iy2,jy0);
978 dz20 = _mm256_sub_ps(iz2,jz0);
979 dx30 = _mm256_sub_ps(ix3,jx0);
980 dy30 = _mm256_sub_ps(iy3,jy0);
981 dz30 = _mm256_sub_ps(iz3,jz0);
983 /* Calculate squared distance and things based on it */
984 rsq10 = gmx_mm256_calc_rsq_ps(dx10,dy10,dz10);
985 rsq20 = gmx_mm256_calc_rsq_ps(dx20,dy20,dz20);
986 rsq30 = gmx_mm256_calc_rsq_ps(dx30,dy30,dz30);
988 rinv10 = avx256_invsqrt_f(rsq10);
989 rinv20 = avx256_invsqrt_f(rsq20);
990 rinv30 = avx256_invsqrt_f(rsq30);
992 rinvsq10 = _mm256_mul_ps(rinv10,rinv10);
993 rinvsq20 = _mm256_mul_ps(rinv20,rinv20);
994 rinvsq30 = _mm256_mul_ps(rinv30,rinv30);
996 /* Load parameters for j particles */
997 jq0 = gmx_mm256_load_8real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
998 charge+jnrC+0,charge+jnrD+0,
999 charge+jnrE+0,charge+jnrF+0,
1000 charge+jnrG+0,charge+jnrH+0);
1002 fjx0 = _mm256_setzero_ps();
1003 fjy0 = _mm256_setzero_ps();
1004 fjz0 = _mm256_setzero_ps();
1006 /**************************
1007 * CALCULATE INTERACTIONS *
1008 **************************/
1010 if (gmx_mm256_any_lt(rsq10,rcutoff2))
1013 /* Compute parameters for interactions between i and j atoms */
1014 qq10 = _mm256_mul_ps(iq1,jq0);
1016 /* REACTION-FIELD ELECTROSTATICS */
1017 felec = _mm256_mul_ps(qq10,_mm256_sub_ps(_mm256_mul_ps(rinv10,rinvsq10),krf2));
1019 cutoff_mask = _mm256_cmp_ps(rsq10,rcutoff2,_CMP_LT_OQ);
1023 fscal = _mm256_and_ps(fscal,cutoff_mask);
1025 fscal = _mm256_andnot_ps(dummy_mask,fscal);
1027 /* Calculate temporary vectorial force */
1028 tx = _mm256_mul_ps(fscal,dx10);
1029 ty = _mm256_mul_ps(fscal,dy10);
1030 tz = _mm256_mul_ps(fscal,dz10);
1032 /* Update vectorial force */
1033 fix1 = _mm256_add_ps(fix1,tx);
1034 fiy1 = _mm256_add_ps(fiy1,ty);
1035 fiz1 = _mm256_add_ps(fiz1,tz);
1037 fjx0 = _mm256_add_ps(fjx0,tx);
1038 fjy0 = _mm256_add_ps(fjy0,ty);
1039 fjz0 = _mm256_add_ps(fjz0,tz);
1043 /**************************
1044 * CALCULATE INTERACTIONS *
1045 **************************/
1047 if (gmx_mm256_any_lt(rsq20,rcutoff2))
1050 /* Compute parameters for interactions between i and j atoms */
1051 qq20 = _mm256_mul_ps(iq2,jq0);
1053 /* REACTION-FIELD ELECTROSTATICS */
1054 felec = _mm256_mul_ps(qq20,_mm256_sub_ps(_mm256_mul_ps(rinv20,rinvsq20),krf2));
1056 cutoff_mask = _mm256_cmp_ps(rsq20,rcutoff2,_CMP_LT_OQ);
1060 fscal = _mm256_and_ps(fscal,cutoff_mask);
1062 fscal = _mm256_andnot_ps(dummy_mask,fscal);
1064 /* Calculate temporary vectorial force */
1065 tx = _mm256_mul_ps(fscal,dx20);
1066 ty = _mm256_mul_ps(fscal,dy20);
1067 tz = _mm256_mul_ps(fscal,dz20);
1069 /* Update vectorial force */
1070 fix2 = _mm256_add_ps(fix2,tx);
1071 fiy2 = _mm256_add_ps(fiy2,ty);
1072 fiz2 = _mm256_add_ps(fiz2,tz);
1074 fjx0 = _mm256_add_ps(fjx0,tx);
1075 fjy0 = _mm256_add_ps(fjy0,ty);
1076 fjz0 = _mm256_add_ps(fjz0,tz);
1080 /**************************
1081 * CALCULATE INTERACTIONS *
1082 **************************/
1084 if (gmx_mm256_any_lt(rsq30,rcutoff2))
1087 /* Compute parameters for interactions between i and j atoms */
1088 qq30 = _mm256_mul_ps(iq3,jq0);
1090 /* REACTION-FIELD ELECTROSTATICS */
1091 felec = _mm256_mul_ps(qq30,_mm256_sub_ps(_mm256_mul_ps(rinv30,rinvsq30),krf2));
1093 cutoff_mask = _mm256_cmp_ps(rsq30,rcutoff2,_CMP_LT_OQ);
1097 fscal = _mm256_and_ps(fscal,cutoff_mask);
1099 fscal = _mm256_andnot_ps(dummy_mask,fscal);
1101 /* Calculate temporary vectorial force */
1102 tx = _mm256_mul_ps(fscal,dx30);
1103 ty = _mm256_mul_ps(fscal,dy30);
1104 tz = _mm256_mul_ps(fscal,dz30);
1106 /* Update vectorial force */
1107 fix3 = _mm256_add_ps(fix3,tx);
1108 fiy3 = _mm256_add_ps(fiy3,ty);
1109 fiz3 = _mm256_add_ps(fiz3,tz);
1111 fjx0 = _mm256_add_ps(fjx0,tx);
1112 fjy0 = _mm256_add_ps(fjy0,ty);
1113 fjz0 = _mm256_add_ps(fjz0,tz);
1117 fjptrA = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
1118 fjptrB = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
1119 fjptrC = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
1120 fjptrD = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
1121 fjptrE = (jnrlistE>=0) ? f+j_coord_offsetE : scratch;
1122 fjptrF = (jnrlistF>=0) ? f+j_coord_offsetF : scratch;
1123 fjptrG = (jnrlistG>=0) ? f+j_coord_offsetG : scratch;
1124 fjptrH = (jnrlistH>=0) ? f+j_coord_offsetH : scratch;
1126 gmx_mm256_decrement_1rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,fjx0,fjy0,fjz0);
1128 /* Inner loop uses 93 flops */
1131 /* End of innermost loop */
1133 gmx_mm256_update_iforce_3atom_swizzle_ps(fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
1134 f+i_coord_offset+DIM,fshift+i_shift_offset);
1136 /* Increment number of inner iterations */
1137 inneriter += j_index_end - j_index_start;
1139 /* Outer loop uses 18 flops */
1142 /* Increment number of outer iterations */
1145 /* Update outer/inner flops */
1147 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_W4_F,outeriter*18 + inneriter*93);