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 avx_256_double kernel generator.
42 #include "../nb_kernel.h"
43 #include "types/simple.h"
44 #include "gromacs/math/vec.h"
47 #include "gromacs/simd/math_x86_avx_256_double.h"
48 #include "kernelutil_x86_avx_256_double.h"
51 * Gromacs nonbonded kernel: nb_kernel_ElecRF_VdwLJ_GeomW3W3_VF_avx_256_double
52 * Electrostatics interaction: ReactionField
53 * VdW interaction: LennardJones
54 * Geometry: Water3-Water3
55 * Calculate force/pot: PotentialAndForce
58 nb_kernel_ElecRF_VdwLJ_GeomW3W3_VF_avx_256_double
59 (t_nblist * gmx_restrict nlist,
60 rvec * gmx_restrict xx,
61 rvec * gmx_restrict ff,
62 t_forcerec * gmx_restrict fr,
63 t_mdatoms * gmx_restrict mdatoms,
64 nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
65 t_nrnb * gmx_restrict nrnb)
67 /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
68 * just 0 for non-waters.
69 * Suffixes A,B,C,D refer to j loop unrolling done with AVX, e.g. for the four different
70 * jnr indices corresponding to data put in the four positions in the SIMD register.
72 int i_shift_offset,i_coord_offset,outeriter,inneriter;
73 int j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
74 int jnrA,jnrB,jnrC,jnrD;
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 *iinr,*jindex,*jjnr,*shiftidx,*gid;
80 real *shiftvec,*fshift,*x,*f;
81 real *fjptrA,*fjptrB,*fjptrC,*fjptrD;
83 __m256d tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
84 real * vdwioffsetptr0;
85 __m256d ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
86 real * vdwioffsetptr1;
87 __m256d ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
88 real * vdwioffsetptr2;
89 __m256d ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
90 int vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D;
91 __m256d jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
92 int vdwjidx1A,vdwjidx1B,vdwjidx1C,vdwjidx1D;
93 __m256d jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
94 int vdwjidx2A,vdwjidx2B,vdwjidx2C,vdwjidx2D;
95 __m256d jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
96 __m256d dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
97 __m256d dx01,dy01,dz01,rsq01,rinv01,rinvsq01,r01,qq01,c6_01,c12_01;
98 __m256d dx02,dy02,dz02,rsq02,rinv02,rinvsq02,r02,qq02,c6_02,c12_02;
99 __m256d dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
100 __m256d dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11;
101 __m256d dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12;
102 __m256d dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
103 __m256d dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21;
104 __m256d dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22;
105 __m256d velec,felec,velecsum,facel,crf,krf,krf2;
108 __m256d rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
111 __m256d one_sixth = _mm256_set1_pd(1.0/6.0);
112 __m256d one_twelfth = _mm256_set1_pd(1.0/12.0);
113 __m256d dummy_mask,cutoff_mask;
114 __m128 tmpmask0,tmpmask1;
115 __m256d signbit = _mm256_castsi256_pd( _mm256_set1_epi32(0x80000000) );
116 __m256d one = _mm256_set1_pd(1.0);
117 __m256d two = _mm256_set1_pd(2.0);
123 jindex = nlist->jindex;
125 shiftidx = nlist->shift;
127 shiftvec = fr->shift_vec[0];
128 fshift = fr->fshift[0];
129 facel = _mm256_set1_pd(fr->epsfac);
130 charge = mdatoms->chargeA;
131 krf = _mm256_set1_pd(fr->ic->k_rf);
132 krf2 = _mm256_set1_pd(fr->ic->k_rf*2.0);
133 crf = _mm256_set1_pd(fr->ic->c_rf);
134 nvdwtype = fr->ntype;
136 vdwtype = mdatoms->typeA;
138 /* Setup water-specific parameters */
139 inr = nlist->iinr[0];
140 iq0 = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+0]));
141 iq1 = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+1]));
142 iq2 = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+2]));
143 vdwioffsetptr0 = vdwparam+2*nvdwtype*vdwtype[inr+0];
145 jq0 = _mm256_set1_pd(charge[inr+0]);
146 jq1 = _mm256_set1_pd(charge[inr+1]);
147 jq2 = _mm256_set1_pd(charge[inr+2]);
148 vdwjidx0A = 2*vdwtype[inr+0];
149 qq00 = _mm256_mul_pd(iq0,jq0);
150 c6_00 = _mm256_set1_pd(vdwioffsetptr0[vdwjidx0A]);
151 c12_00 = _mm256_set1_pd(vdwioffsetptr0[vdwjidx0A+1]);
152 qq01 = _mm256_mul_pd(iq0,jq1);
153 qq02 = _mm256_mul_pd(iq0,jq2);
154 qq10 = _mm256_mul_pd(iq1,jq0);
155 qq11 = _mm256_mul_pd(iq1,jq1);
156 qq12 = _mm256_mul_pd(iq1,jq2);
157 qq20 = _mm256_mul_pd(iq2,jq0);
158 qq21 = _mm256_mul_pd(iq2,jq1);
159 qq22 = _mm256_mul_pd(iq2,jq2);
161 /* Avoid stupid compiler warnings */
162 jnrA = jnrB = jnrC = jnrD = 0;
171 for(iidx=0;iidx<4*DIM;iidx++)
176 /* Start outer loop over neighborlists */
177 for(iidx=0; iidx<nri; iidx++)
179 /* Load shift vector for this list */
180 i_shift_offset = DIM*shiftidx[iidx];
182 /* Load limits for loop over neighbors */
183 j_index_start = jindex[iidx];
184 j_index_end = jindex[iidx+1];
186 /* Get outer coordinate index */
188 i_coord_offset = DIM*inr;
190 /* Load i particle coords and add shift vector */
191 gmx_mm256_load_shift_and_3rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset,
192 &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2);
194 fix0 = _mm256_setzero_pd();
195 fiy0 = _mm256_setzero_pd();
196 fiz0 = _mm256_setzero_pd();
197 fix1 = _mm256_setzero_pd();
198 fiy1 = _mm256_setzero_pd();
199 fiz1 = _mm256_setzero_pd();
200 fix2 = _mm256_setzero_pd();
201 fiy2 = _mm256_setzero_pd();
202 fiz2 = _mm256_setzero_pd();
204 /* Reset potential sums */
205 velecsum = _mm256_setzero_pd();
206 vvdwsum = _mm256_setzero_pd();
208 /* Start inner kernel loop */
209 for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4)
212 /* Get j neighbor index, and coordinate index */
217 j_coord_offsetA = DIM*jnrA;
218 j_coord_offsetB = DIM*jnrB;
219 j_coord_offsetC = DIM*jnrC;
220 j_coord_offsetD = DIM*jnrD;
222 /* load j atom coordinates */
223 gmx_mm256_load_3rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
224 x+j_coord_offsetC,x+j_coord_offsetD,
225 &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,&jy2,&jz2);
227 /* Calculate displacement vector */
228 dx00 = _mm256_sub_pd(ix0,jx0);
229 dy00 = _mm256_sub_pd(iy0,jy0);
230 dz00 = _mm256_sub_pd(iz0,jz0);
231 dx01 = _mm256_sub_pd(ix0,jx1);
232 dy01 = _mm256_sub_pd(iy0,jy1);
233 dz01 = _mm256_sub_pd(iz0,jz1);
234 dx02 = _mm256_sub_pd(ix0,jx2);
235 dy02 = _mm256_sub_pd(iy0,jy2);
236 dz02 = _mm256_sub_pd(iz0,jz2);
237 dx10 = _mm256_sub_pd(ix1,jx0);
238 dy10 = _mm256_sub_pd(iy1,jy0);
239 dz10 = _mm256_sub_pd(iz1,jz0);
240 dx11 = _mm256_sub_pd(ix1,jx1);
241 dy11 = _mm256_sub_pd(iy1,jy1);
242 dz11 = _mm256_sub_pd(iz1,jz1);
243 dx12 = _mm256_sub_pd(ix1,jx2);
244 dy12 = _mm256_sub_pd(iy1,jy2);
245 dz12 = _mm256_sub_pd(iz1,jz2);
246 dx20 = _mm256_sub_pd(ix2,jx0);
247 dy20 = _mm256_sub_pd(iy2,jy0);
248 dz20 = _mm256_sub_pd(iz2,jz0);
249 dx21 = _mm256_sub_pd(ix2,jx1);
250 dy21 = _mm256_sub_pd(iy2,jy1);
251 dz21 = _mm256_sub_pd(iz2,jz1);
252 dx22 = _mm256_sub_pd(ix2,jx2);
253 dy22 = _mm256_sub_pd(iy2,jy2);
254 dz22 = _mm256_sub_pd(iz2,jz2);
256 /* Calculate squared distance and things based on it */
257 rsq00 = gmx_mm256_calc_rsq_pd(dx00,dy00,dz00);
258 rsq01 = gmx_mm256_calc_rsq_pd(dx01,dy01,dz01);
259 rsq02 = gmx_mm256_calc_rsq_pd(dx02,dy02,dz02);
260 rsq10 = gmx_mm256_calc_rsq_pd(dx10,dy10,dz10);
261 rsq11 = gmx_mm256_calc_rsq_pd(dx11,dy11,dz11);
262 rsq12 = gmx_mm256_calc_rsq_pd(dx12,dy12,dz12);
263 rsq20 = gmx_mm256_calc_rsq_pd(dx20,dy20,dz20);
264 rsq21 = gmx_mm256_calc_rsq_pd(dx21,dy21,dz21);
265 rsq22 = gmx_mm256_calc_rsq_pd(dx22,dy22,dz22);
267 rinv00 = gmx_mm256_invsqrt_pd(rsq00);
268 rinv01 = gmx_mm256_invsqrt_pd(rsq01);
269 rinv02 = gmx_mm256_invsqrt_pd(rsq02);
270 rinv10 = gmx_mm256_invsqrt_pd(rsq10);
271 rinv11 = gmx_mm256_invsqrt_pd(rsq11);
272 rinv12 = gmx_mm256_invsqrt_pd(rsq12);
273 rinv20 = gmx_mm256_invsqrt_pd(rsq20);
274 rinv21 = gmx_mm256_invsqrt_pd(rsq21);
275 rinv22 = gmx_mm256_invsqrt_pd(rsq22);
277 rinvsq00 = _mm256_mul_pd(rinv00,rinv00);
278 rinvsq01 = _mm256_mul_pd(rinv01,rinv01);
279 rinvsq02 = _mm256_mul_pd(rinv02,rinv02);
280 rinvsq10 = _mm256_mul_pd(rinv10,rinv10);
281 rinvsq11 = _mm256_mul_pd(rinv11,rinv11);
282 rinvsq12 = _mm256_mul_pd(rinv12,rinv12);
283 rinvsq20 = _mm256_mul_pd(rinv20,rinv20);
284 rinvsq21 = _mm256_mul_pd(rinv21,rinv21);
285 rinvsq22 = _mm256_mul_pd(rinv22,rinv22);
287 fjx0 = _mm256_setzero_pd();
288 fjy0 = _mm256_setzero_pd();
289 fjz0 = _mm256_setzero_pd();
290 fjx1 = _mm256_setzero_pd();
291 fjy1 = _mm256_setzero_pd();
292 fjz1 = _mm256_setzero_pd();
293 fjx2 = _mm256_setzero_pd();
294 fjy2 = _mm256_setzero_pd();
295 fjz2 = _mm256_setzero_pd();
297 /**************************
298 * CALCULATE INTERACTIONS *
299 **************************/
301 /* REACTION-FIELD ELECTROSTATICS */
302 velec = _mm256_mul_pd(qq00,_mm256_sub_pd(_mm256_add_pd(rinv00,_mm256_mul_pd(krf,rsq00)),crf));
303 felec = _mm256_mul_pd(qq00,_mm256_sub_pd(_mm256_mul_pd(rinv00,rinvsq00),krf2));
305 /* LENNARD-JONES DISPERSION/REPULSION */
307 rinvsix = _mm256_mul_pd(_mm256_mul_pd(rinvsq00,rinvsq00),rinvsq00);
308 vvdw6 = _mm256_mul_pd(c6_00,rinvsix);
309 vvdw12 = _mm256_mul_pd(c12_00,_mm256_mul_pd(rinvsix,rinvsix));
310 vvdw = _mm256_sub_pd( _mm256_mul_pd(vvdw12,one_twelfth) , _mm256_mul_pd(vvdw6,one_sixth) );
311 fvdw = _mm256_mul_pd(_mm256_sub_pd(vvdw12,vvdw6),rinvsq00);
313 /* Update potential sum for this i atom from the interaction with this j atom. */
314 velecsum = _mm256_add_pd(velecsum,velec);
315 vvdwsum = _mm256_add_pd(vvdwsum,vvdw);
317 fscal = _mm256_add_pd(felec,fvdw);
319 /* Calculate temporary vectorial force */
320 tx = _mm256_mul_pd(fscal,dx00);
321 ty = _mm256_mul_pd(fscal,dy00);
322 tz = _mm256_mul_pd(fscal,dz00);
324 /* Update vectorial force */
325 fix0 = _mm256_add_pd(fix0,tx);
326 fiy0 = _mm256_add_pd(fiy0,ty);
327 fiz0 = _mm256_add_pd(fiz0,tz);
329 fjx0 = _mm256_add_pd(fjx0,tx);
330 fjy0 = _mm256_add_pd(fjy0,ty);
331 fjz0 = _mm256_add_pd(fjz0,tz);
333 /**************************
334 * CALCULATE INTERACTIONS *
335 **************************/
337 /* REACTION-FIELD ELECTROSTATICS */
338 velec = _mm256_mul_pd(qq01,_mm256_sub_pd(_mm256_add_pd(rinv01,_mm256_mul_pd(krf,rsq01)),crf));
339 felec = _mm256_mul_pd(qq01,_mm256_sub_pd(_mm256_mul_pd(rinv01,rinvsq01),krf2));
341 /* Update potential sum for this i atom from the interaction with this j atom. */
342 velecsum = _mm256_add_pd(velecsum,velec);
346 /* Calculate temporary vectorial force */
347 tx = _mm256_mul_pd(fscal,dx01);
348 ty = _mm256_mul_pd(fscal,dy01);
349 tz = _mm256_mul_pd(fscal,dz01);
351 /* Update vectorial force */
352 fix0 = _mm256_add_pd(fix0,tx);
353 fiy0 = _mm256_add_pd(fiy0,ty);
354 fiz0 = _mm256_add_pd(fiz0,tz);
356 fjx1 = _mm256_add_pd(fjx1,tx);
357 fjy1 = _mm256_add_pd(fjy1,ty);
358 fjz1 = _mm256_add_pd(fjz1,tz);
360 /**************************
361 * CALCULATE INTERACTIONS *
362 **************************/
364 /* REACTION-FIELD ELECTROSTATICS */
365 velec = _mm256_mul_pd(qq02,_mm256_sub_pd(_mm256_add_pd(rinv02,_mm256_mul_pd(krf,rsq02)),crf));
366 felec = _mm256_mul_pd(qq02,_mm256_sub_pd(_mm256_mul_pd(rinv02,rinvsq02),krf2));
368 /* Update potential sum for this i atom from the interaction with this j atom. */
369 velecsum = _mm256_add_pd(velecsum,velec);
373 /* Calculate temporary vectorial force */
374 tx = _mm256_mul_pd(fscal,dx02);
375 ty = _mm256_mul_pd(fscal,dy02);
376 tz = _mm256_mul_pd(fscal,dz02);
378 /* Update vectorial force */
379 fix0 = _mm256_add_pd(fix0,tx);
380 fiy0 = _mm256_add_pd(fiy0,ty);
381 fiz0 = _mm256_add_pd(fiz0,tz);
383 fjx2 = _mm256_add_pd(fjx2,tx);
384 fjy2 = _mm256_add_pd(fjy2,ty);
385 fjz2 = _mm256_add_pd(fjz2,tz);
387 /**************************
388 * CALCULATE INTERACTIONS *
389 **************************/
391 /* REACTION-FIELD ELECTROSTATICS */
392 velec = _mm256_mul_pd(qq10,_mm256_sub_pd(_mm256_add_pd(rinv10,_mm256_mul_pd(krf,rsq10)),crf));
393 felec = _mm256_mul_pd(qq10,_mm256_sub_pd(_mm256_mul_pd(rinv10,rinvsq10),krf2));
395 /* Update potential sum for this i atom from the interaction with this j atom. */
396 velecsum = _mm256_add_pd(velecsum,velec);
400 /* Calculate temporary vectorial force */
401 tx = _mm256_mul_pd(fscal,dx10);
402 ty = _mm256_mul_pd(fscal,dy10);
403 tz = _mm256_mul_pd(fscal,dz10);
405 /* Update vectorial force */
406 fix1 = _mm256_add_pd(fix1,tx);
407 fiy1 = _mm256_add_pd(fiy1,ty);
408 fiz1 = _mm256_add_pd(fiz1,tz);
410 fjx0 = _mm256_add_pd(fjx0,tx);
411 fjy0 = _mm256_add_pd(fjy0,ty);
412 fjz0 = _mm256_add_pd(fjz0,tz);
414 /**************************
415 * CALCULATE INTERACTIONS *
416 **************************/
418 /* REACTION-FIELD ELECTROSTATICS */
419 velec = _mm256_mul_pd(qq11,_mm256_sub_pd(_mm256_add_pd(rinv11,_mm256_mul_pd(krf,rsq11)),crf));
420 felec = _mm256_mul_pd(qq11,_mm256_sub_pd(_mm256_mul_pd(rinv11,rinvsq11),krf2));
422 /* Update potential sum for this i atom from the interaction with this j atom. */
423 velecsum = _mm256_add_pd(velecsum,velec);
427 /* Calculate temporary vectorial force */
428 tx = _mm256_mul_pd(fscal,dx11);
429 ty = _mm256_mul_pd(fscal,dy11);
430 tz = _mm256_mul_pd(fscal,dz11);
432 /* Update vectorial force */
433 fix1 = _mm256_add_pd(fix1,tx);
434 fiy1 = _mm256_add_pd(fiy1,ty);
435 fiz1 = _mm256_add_pd(fiz1,tz);
437 fjx1 = _mm256_add_pd(fjx1,tx);
438 fjy1 = _mm256_add_pd(fjy1,ty);
439 fjz1 = _mm256_add_pd(fjz1,tz);
441 /**************************
442 * CALCULATE INTERACTIONS *
443 **************************/
445 /* REACTION-FIELD ELECTROSTATICS */
446 velec = _mm256_mul_pd(qq12,_mm256_sub_pd(_mm256_add_pd(rinv12,_mm256_mul_pd(krf,rsq12)),crf));
447 felec = _mm256_mul_pd(qq12,_mm256_sub_pd(_mm256_mul_pd(rinv12,rinvsq12),krf2));
449 /* Update potential sum for this i atom from the interaction with this j atom. */
450 velecsum = _mm256_add_pd(velecsum,velec);
454 /* Calculate temporary vectorial force */
455 tx = _mm256_mul_pd(fscal,dx12);
456 ty = _mm256_mul_pd(fscal,dy12);
457 tz = _mm256_mul_pd(fscal,dz12);
459 /* Update vectorial force */
460 fix1 = _mm256_add_pd(fix1,tx);
461 fiy1 = _mm256_add_pd(fiy1,ty);
462 fiz1 = _mm256_add_pd(fiz1,tz);
464 fjx2 = _mm256_add_pd(fjx2,tx);
465 fjy2 = _mm256_add_pd(fjy2,ty);
466 fjz2 = _mm256_add_pd(fjz2,tz);
468 /**************************
469 * CALCULATE INTERACTIONS *
470 **************************/
472 /* REACTION-FIELD ELECTROSTATICS */
473 velec = _mm256_mul_pd(qq20,_mm256_sub_pd(_mm256_add_pd(rinv20,_mm256_mul_pd(krf,rsq20)),crf));
474 felec = _mm256_mul_pd(qq20,_mm256_sub_pd(_mm256_mul_pd(rinv20,rinvsq20),krf2));
476 /* Update potential sum for this i atom from the interaction with this j atom. */
477 velecsum = _mm256_add_pd(velecsum,velec);
481 /* Calculate temporary vectorial force */
482 tx = _mm256_mul_pd(fscal,dx20);
483 ty = _mm256_mul_pd(fscal,dy20);
484 tz = _mm256_mul_pd(fscal,dz20);
486 /* Update vectorial force */
487 fix2 = _mm256_add_pd(fix2,tx);
488 fiy2 = _mm256_add_pd(fiy2,ty);
489 fiz2 = _mm256_add_pd(fiz2,tz);
491 fjx0 = _mm256_add_pd(fjx0,tx);
492 fjy0 = _mm256_add_pd(fjy0,ty);
493 fjz0 = _mm256_add_pd(fjz0,tz);
495 /**************************
496 * CALCULATE INTERACTIONS *
497 **************************/
499 /* REACTION-FIELD ELECTROSTATICS */
500 velec = _mm256_mul_pd(qq21,_mm256_sub_pd(_mm256_add_pd(rinv21,_mm256_mul_pd(krf,rsq21)),crf));
501 felec = _mm256_mul_pd(qq21,_mm256_sub_pd(_mm256_mul_pd(rinv21,rinvsq21),krf2));
503 /* Update potential sum for this i atom from the interaction with this j atom. */
504 velecsum = _mm256_add_pd(velecsum,velec);
508 /* Calculate temporary vectorial force */
509 tx = _mm256_mul_pd(fscal,dx21);
510 ty = _mm256_mul_pd(fscal,dy21);
511 tz = _mm256_mul_pd(fscal,dz21);
513 /* Update vectorial force */
514 fix2 = _mm256_add_pd(fix2,tx);
515 fiy2 = _mm256_add_pd(fiy2,ty);
516 fiz2 = _mm256_add_pd(fiz2,tz);
518 fjx1 = _mm256_add_pd(fjx1,tx);
519 fjy1 = _mm256_add_pd(fjy1,ty);
520 fjz1 = _mm256_add_pd(fjz1,tz);
522 /**************************
523 * CALCULATE INTERACTIONS *
524 **************************/
526 /* REACTION-FIELD ELECTROSTATICS */
527 velec = _mm256_mul_pd(qq22,_mm256_sub_pd(_mm256_add_pd(rinv22,_mm256_mul_pd(krf,rsq22)),crf));
528 felec = _mm256_mul_pd(qq22,_mm256_sub_pd(_mm256_mul_pd(rinv22,rinvsq22),krf2));
530 /* Update potential sum for this i atom from the interaction with this j atom. */
531 velecsum = _mm256_add_pd(velecsum,velec);
535 /* Calculate temporary vectorial force */
536 tx = _mm256_mul_pd(fscal,dx22);
537 ty = _mm256_mul_pd(fscal,dy22);
538 tz = _mm256_mul_pd(fscal,dz22);
540 /* Update vectorial force */
541 fix2 = _mm256_add_pd(fix2,tx);
542 fiy2 = _mm256_add_pd(fiy2,ty);
543 fiz2 = _mm256_add_pd(fiz2,tz);
545 fjx2 = _mm256_add_pd(fjx2,tx);
546 fjy2 = _mm256_add_pd(fjy2,ty);
547 fjz2 = _mm256_add_pd(fjz2,tz);
549 fjptrA = f+j_coord_offsetA;
550 fjptrB = f+j_coord_offsetB;
551 fjptrC = f+j_coord_offsetC;
552 fjptrD = f+j_coord_offsetD;
554 gmx_mm256_decrement_3rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,
555 fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2);
557 /* Inner loop uses 300 flops */
563 /* Get j neighbor index, and coordinate index */
564 jnrlistA = jjnr[jidx];
565 jnrlistB = jjnr[jidx+1];
566 jnrlistC = jjnr[jidx+2];
567 jnrlistD = jjnr[jidx+3];
568 /* Sign of each element will be negative for non-real atoms.
569 * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
570 * so use it as val = _mm_andnot_pd(mask,val) to clear dummy entries.
572 tmpmask0 = gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128()));
574 tmpmask1 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(3,3,2,2));
575 tmpmask0 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(1,1,0,0));
576 dummy_mask = _mm256_castps_pd(gmx_mm256_set_m128(tmpmask1,tmpmask0));
578 jnrA = (jnrlistA>=0) ? jnrlistA : 0;
579 jnrB = (jnrlistB>=0) ? jnrlistB : 0;
580 jnrC = (jnrlistC>=0) ? jnrlistC : 0;
581 jnrD = (jnrlistD>=0) ? jnrlistD : 0;
582 j_coord_offsetA = DIM*jnrA;
583 j_coord_offsetB = DIM*jnrB;
584 j_coord_offsetC = DIM*jnrC;
585 j_coord_offsetD = DIM*jnrD;
587 /* load j atom coordinates */
588 gmx_mm256_load_3rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
589 x+j_coord_offsetC,x+j_coord_offsetD,
590 &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,&jy2,&jz2);
592 /* Calculate displacement vector */
593 dx00 = _mm256_sub_pd(ix0,jx0);
594 dy00 = _mm256_sub_pd(iy0,jy0);
595 dz00 = _mm256_sub_pd(iz0,jz0);
596 dx01 = _mm256_sub_pd(ix0,jx1);
597 dy01 = _mm256_sub_pd(iy0,jy1);
598 dz01 = _mm256_sub_pd(iz0,jz1);
599 dx02 = _mm256_sub_pd(ix0,jx2);
600 dy02 = _mm256_sub_pd(iy0,jy2);
601 dz02 = _mm256_sub_pd(iz0,jz2);
602 dx10 = _mm256_sub_pd(ix1,jx0);
603 dy10 = _mm256_sub_pd(iy1,jy0);
604 dz10 = _mm256_sub_pd(iz1,jz0);
605 dx11 = _mm256_sub_pd(ix1,jx1);
606 dy11 = _mm256_sub_pd(iy1,jy1);
607 dz11 = _mm256_sub_pd(iz1,jz1);
608 dx12 = _mm256_sub_pd(ix1,jx2);
609 dy12 = _mm256_sub_pd(iy1,jy2);
610 dz12 = _mm256_sub_pd(iz1,jz2);
611 dx20 = _mm256_sub_pd(ix2,jx0);
612 dy20 = _mm256_sub_pd(iy2,jy0);
613 dz20 = _mm256_sub_pd(iz2,jz0);
614 dx21 = _mm256_sub_pd(ix2,jx1);
615 dy21 = _mm256_sub_pd(iy2,jy1);
616 dz21 = _mm256_sub_pd(iz2,jz1);
617 dx22 = _mm256_sub_pd(ix2,jx2);
618 dy22 = _mm256_sub_pd(iy2,jy2);
619 dz22 = _mm256_sub_pd(iz2,jz2);
621 /* Calculate squared distance and things based on it */
622 rsq00 = gmx_mm256_calc_rsq_pd(dx00,dy00,dz00);
623 rsq01 = gmx_mm256_calc_rsq_pd(dx01,dy01,dz01);
624 rsq02 = gmx_mm256_calc_rsq_pd(dx02,dy02,dz02);
625 rsq10 = gmx_mm256_calc_rsq_pd(dx10,dy10,dz10);
626 rsq11 = gmx_mm256_calc_rsq_pd(dx11,dy11,dz11);
627 rsq12 = gmx_mm256_calc_rsq_pd(dx12,dy12,dz12);
628 rsq20 = gmx_mm256_calc_rsq_pd(dx20,dy20,dz20);
629 rsq21 = gmx_mm256_calc_rsq_pd(dx21,dy21,dz21);
630 rsq22 = gmx_mm256_calc_rsq_pd(dx22,dy22,dz22);
632 rinv00 = gmx_mm256_invsqrt_pd(rsq00);
633 rinv01 = gmx_mm256_invsqrt_pd(rsq01);
634 rinv02 = gmx_mm256_invsqrt_pd(rsq02);
635 rinv10 = gmx_mm256_invsqrt_pd(rsq10);
636 rinv11 = gmx_mm256_invsqrt_pd(rsq11);
637 rinv12 = gmx_mm256_invsqrt_pd(rsq12);
638 rinv20 = gmx_mm256_invsqrt_pd(rsq20);
639 rinv21 = gmx_mm256_invsqrt_pd(rsq21);
640 rinv22 = gmx_mm256_invsqrt_pd(rsq22);
642 rinvsq00 = _mm256_mul_pd(rinv00,rinv00);
643 rinvsq01 = _mm256_mul_pd(rinv01,rinv01);
644 rinvsq02 = _mm256_mul_pd(rinv02,rinv02);
645 rinvsq10 = _mm256_mul_pd(rinv10,rinv10);
646 rinvsq11 = _mm256_mul_pd(rinv11,rinv11);
647 rinvsq12 = _mm256_mul_pd(rinv12,rinv12);
648 rinvsq20 = _mm256_mul_pd(rinv20,rinv20);
649 rinvsq21 = _mm256_mul_pd(rinv21,rinv21);
650 rinvsq22 = _mm256_mul_pd(rinv22,rinv22);
652 fjx0 = _mm256_setzero_pd();
653 fjy0 = _mm256_setzero_pd();
654 fjz0 = _mm256_setzero_pd();
655 fjx1 = _mm256_setzero_pd();
656 fjy1 = _mm256_setzero_pd();
657 fjz1 = _mm256_setzero_pd();
658 fjx2 = _mm256_setzero_pd();
659 fjy2 = _mm256_setzero_pd();
660 fjz2 = _mm256_setzero_pd();
662 /**************************
663 * CALCULATE INTERACTIONS *
664 **************************/
666 /* REACTION-FIELD ELECTROSTATICS */
667 velec = _mm256_mul_pd(qq00,_mm256_sub_pd(_mm256_add_pd(rinv00,_mm256_mul_pd(krf,rsq00)),crf));
668 felec = _mm256_mul_pd(qq00,_mm256_sub_pd(_mm256_mul_pd(rinv00,rinvsq00),krf2));
670 /* LENNARD-JONES DISPERSION/REPULSION */
672 rinvsix = _mm256_mul_pd(_mm256_mul_pd(rinvsq00,rinvsq00),rinvsq00);
673 vvdw6 = _mm256_mul_pd(c6_00,rinvsix);
674 vvdw12 = _mm256_mul_pd(c12_00,_mm256_mul_pd(rinvsix,rinvsix));
675 vvdw = _mm256_sub_pd( _mm256_mul_pd(vvdw12,one_twelfth) , _mm256_mul_pd(vvdw6,one_sixth) );
676 fvdw = _mm256_mul_pd(_mm256_sub_pd(vvdw12,vvdw6),rinvsq00);
678 /* Update potential sum for this i atom from the interaction with this j atom. */
679 velec = _mm256_andnot_pd(dummy_mask,velec);
680 velecsum = _mm256_add_pd(velecsum,velec);
681 vvdw = _mm256_andnot_pd(dummy_mask,vvdw);
682 vvdwsum = _mm256_add_pd(vvdwsum,vvdw);
684 fscal = _mm256_add_pd(felec,fvdw);
686 fscal = _mm256_andnot_pd(dummy_mask,fscal);
688 /* Calculate temporary vectorial force */
689 tx = _mm256_mul_pd(fscal,dx00);
690 ty = _mm256_mul_pd(fscal,dy00);
691 tz = _mm256_mul_pd(fscal,dz00);
693 /* Update vectorial force */
694 fix0 = _mm256_add_pd(fix0,tx);
695 fiy0 = _mm256_add_pd(fiy0,ty);
696 fiz0 = _mm256_add_pd(fiz0,tz);
698 fjx0 = _mm256_add_pd(fjx0,tx);
699 fjy0 = _mm256_add_pd(fjy0,ty);
700 fjz0 = _mm256_add_pd(fjz0,tz);
702 /**************************
703 * CALCULATE INTERACTIONS *
704 **************************/
706 /* REACTION-FIELD ELECTROSTATICS */
707 velec = _mm256_mul_pd(qq01,_mm256_sub_pd(_mm256_add_pd(rinv01,_mm256_mul_pd(krf,rsq01)),crf));
708 felec = _mm256_mul_pd(qq01,_mm256_sub_pd(_mm256_mul_pd(rinv01,rinvsq01),krf2));
710 /* Update potential sum for this i atom from the interaction with this j atom. */
711 velec = _mm256_andnot_pd(dummy_mask,velec);
712 velecsum = _mm256_add_pd(velecsum,velec);
716 fscal = _mm256_andnot_pd(dummy_mask,fscal);
718 /* Calculate temporary vectorial force */
719 tx = _mm256_mul_pd(fscal,dx01);
720 ty = _mm256_mul_pd(fscal,dy01);
721 tz = _mm256_mul_pd(fscal,dz01);
723 /* Update vectorial force */
724 fix0 = _mm256_add_pd(fix0,tx);
725 fiy0 = _mm256_add_pd(fiy0,ty);
726 fiz0 = _mm256_add_pd(fiz0,tz);
728 fjx1 = _mm256_add_pd(fjx1,tx);
729 fjy1 = _mm256_add_pd(fjy1,ty);
730 fjz1 = _mm256_add_pd(fjz1,tz);
732 /**************************
733 * CALCULATE INTERACTIONS *
734 **************************/
736 /* REACTION-FIELD ELECTROSTATICS */
737 velec = _mm256_mul_pd(qq02,_mm256_sub_pd(_mm256_add_pd(rinv02,_mm256_mul_pd(krf,rsq02)),crf));
738 felec = _mm256_mul_pd(qq02,_mm256_sub_pd(_mm256_mul_pd(rinv02,rinvsq02),krf2));
740 /* Update potential sum for this i atom from the interaction with this j atom. */
741 velec = _mm256_andnot_pd(dummy_mask,velec);
742 velecsum = _mm256_add_pd(velecsum,velec);
746 fscal = _mm256_andnot_pd(dummy_mask,fscal);
748 /* Calculate temporary vectorial force */
749 tx = _mm256_mul_pd(fscal,dx02);
750 ty = _mm256_mul_pd(fscal,dy02);
751 tz = _mm256_mul_pd(fscal,dz02);
753 /* Update vectorial force */
754 fix0 = _mm256_add_pd(fix0,tx);
755 fiy0 = _mm256_add_pd(fiy0,ty);
756 fiz0 = _mm256_add_pd(fiz0,tz);
758 fjx2 = _mm256_add_pd(fjx2,tx);
759 fjy2 = _mm256_add_pd(fjy2,ty);
760 fjz2 = _mm256_add_pd(fjz2,tz);
762 /**************************
763 * CALCULATE INTERACTIONS *
764 **************************/
766 /* REACTION-FIELD ELECTROSTATICS */
767 velec = _mm256_mul_pd(qq10,_mm256_sub_pd(_mm256_add_pd(rinv10,_mm256_mul_pd(krf,rsq10)),crf));
768 felec = _mm256_mul_pd(qq10,_mm256_sub_pd(_mm256_mul_pd(rinv10,rinvsq10),krf2));
770 /* Update potential sum for this i atom from the interaction with this j atom. */
771 velec = _mm256_andnot_pd(dummy_mask,velec);
772 velecsum = _mm256_add_pd(velecsum,velec);
776 fscal = _mm256_andnot_pd(dummy_mask,fscal);
778 /* Calculate temporary vectorial force */
779 tx = _mm256_mul_pd(fscal,dx10);
780 ty = _mm256_mul_pd(fscal,dy10);
781 tz = _mm256_mul_pd(fscal,dz10);
783 /* Update vectorial force */
784 fix1 = _mm256_add_pd(fix1,tx);
785 fiy1 = _mm256_add_pd(fiy1,ty);
786 fiz1 = _mm256_add_pd(fiz1,tz);
788 fjx0 = _mm256_add_pd(fjx0,tx);
789 fjy0 = _mm256_add_pd(fjy0,ty);
790 fjz0 = _mm256_add_pd(fjz0,tz);
792 /**************************
793 * CALCULATE INTERACTIONS *
794 **************************/
796 /* REACTION-FIELD ELECTROSTATICS */
797 velec = _mm256_mul_pd(qq11,_mm256_sub_pd(_mm256_add_pd(rinv11,_mm256_mul_pd(krf,rsq11)),crf));
798 felec = _mm256_mul_pd(qq11,_mm256_sub_pd(_mm256_mul_pd(rinv11,rinvsq11),krf2));
800 /* Update potential sum for this i atom from the interaction with this j atom. */
801 velec = _mm256_andnot_pd(dummy_mask,velec);
802 velecsum = _mm256_add_pd(velecsum,velec);
806 fscal = _mm256_andnot_pd(dummy_mask,fscal);
808 /* Calculate temporary vectorial force */
809 tx = _mm256_mul_pd(fscal,dx11);
810 ty = _mm256_mul_pd(fscal,dy11);
811 tz = _mm256_mul_pd(fscal,dz11);
813 /* Update vectorial force */
814 fix1 = _mm256_add_pd(fix1,tx);
815 fiy1 = _mm256_add_pd(fiy1,ty);
816 fiz1 = _mm256_add_pd(fiz1,tz);
818 fjx1 = _mm256_add_pd(fjx1,tx);
819 fjy1 = _mm256_add_pd(fjy1,ty);
820 fjz1 = _mm256_add_pd(fjz1,tz);
822 /**************************
823 * CALCULATE INTERACTIONS *
824 **************************/
826 /* REACTION-FIELD ELECTROSTATICS */
827 velec = _mm256_mul_pd(qq12,_mm256_sub_pd(_mm256_add_pd(rinv12,_mm256_mul_pd(krf,rsq12)),crf));
828 felec = _mm256_mul_pd(qq12,_mm256_sub_pd(_mm256_mul_pd(rinv12,rinvsq12),krf2));
830 /* Update potential sum for this i atom from the interaction with this j atom. */
831 velec = _mm256_andnot_pd(dummy_mask,velec);
832 velecsum = _mm256_add_pd(velecsum,velec);
836 fscal = _mm256_andnot_pd(dummy_mask,fscal);
838 /* Calculate temporary vectorial force */
839 tx = _mm256_mul_pd(fscal,dx12);
840 ty = _mm256_mul_pd(fscal,dy12);
841 tz = _mm256_mul_pd(fscal,dz12);
843 /* Update vectorial force */
844 fix1 = _mm256_add_pd(fix1,tx);
845 fiy1 = _mm256_add_pd(fiy1,ty);
846 fiz1 = _mm256_add_pd(fiz1,tz);
848 fjx2 = _mm256_add_pd(fjx2,tx);
849 fjy2 = _mm256_add_pd(fjy2,ty);
850 fjz2 = _mm256_add_pd(fjz2,tz);
852 /**************************
853 * CALCULATE INTERACTIONS *
854 **************************/
856 /* REACTION-FIELD ELECTROSTATICS */
857 velec = _mm256_mul_pd(qq20,_mm256_sub_pd(_mm256_add_pd(rinv20,_mm256_mul_pd(krf,rsq20)),crf));
858 felec = _mm256_mul_pd(qq20,_mm256_sub_pd(_mm256_mul_pd(rinv20,rinvsq20),krf2));
860 /* Update potential sum for this i atom from the interaction with this j atom. */
861 velec = _mm256_andnot_pd(dummy_mask,velec);
862 velecsum = _mm256_add_pd(velecsum,velec);
866 fscal = _mm256_andnot_pd(dummy_mask,fscal);
868 /* Calculate temporary vectorial force */
869 tx = _mm256_mul_pd(fscal,dx20);
870 ty = _mm256_mul_pd(fscal,dy20);
871 tz = _mm256_mul_pd(fscal,dz20);
873 /* Update vectorial force */
874 fix2 = _mm256_add_pd(fix2,tx);
875 fiy2 = _mm256_add_pd(fiy2,ty);
876 fiz2 = _mm256_add_pd(fiz2,tz);
878 fjx0 = _mm256_add_pd(fjx0,tx);
879 fjy0 = _mm256_add_pd(fjy0,ty);
880 fjz0 = _mm256_add_pd(fjz0,tz);
882 /**************************
883 * CALCULATE INTERACTIONS *
884 **************************/
886 /* REACTION-FIELD ELECTROSTATICS */
887 velec = _mm256_mul_pd(qq21,_mm256_sub_pd(_mm256_add_pd(rinv21,_mm256_mul_pd(krf,rsq21)),crf));
888 felec = _mm256_mul_pd(qq21,_mm256_sub_pd(_mm256_mul_pd(rinv21,rinvsq21),krf2));
890 /* Update potential sum for this i atom from the interaction with this j atom. */
891 velec = _mm256_andnot_pd(dummy_mask,velec);
892 velecsum = _mm256_add_pd(velecsum,velec);
896 fscal = _mm256_andnot_pd(dummy_mask,fscal);
898 /* Calculate temporary vectorial force */
899 tx = _mm256_mul_pd(fscal,dx21);
900 ty = _mm256_mul_pd(fscal,dy21);
901 tz = _mm256_mul_pd(fscal,dz21);
903 /* Update vectorial force */
904 fix2 = _mm256_add_pd(fix2,tx);
905 fiy2 = _mm256_add_pd(fiy2,ty);
906 fiz2 = _mm256_add_pd(fiz2,tz);
908 fjx1 = _mm256_add_pd(fjx1,tx);
909 fjy1 = _mm256_add_pd(fjy1,ty);
910 fjz1 = _mm256_add_pd(fjz1,tz);
912 /**************************
913 * CALCULATE INTERACTIONS *
914 **************************/
916 /* REACTION-FIELD ELECTROSTATICS */
917 velec = _mm256_mul_pd(qq22,_mm256_sub_pd(_mm256_add_pd(rinv22,_mm256_mul_pd(krf,rsq22)),crf));
918 felec = _mm256_mul_pd(qq22,_mm256_sub_pd(_mm256_mul_pd(rinv22,rinvsq22),krf2));
920 /* Update potential sum for this i atom from the interaction with this j atom. */
921 velec = _mm256_andnot_pd(dummy_mask,velec);
922 velecsum = _mm256_add_pd(velecsum,velec);
926 fscal = _mm256_andnot_pd(dummy_mask,fscal);
928 /* Calculate temporary vectorial force */
929 tx = _mm256_mul_pd(fscal,dx22);
930 ty = _mm256_mul_pd(fscal,dy22);
931 tz = _mm256_mul_pd(fscal,dz22);
933 /* Update vectorial force */
934 fix2 = _mm256_add_pd(fix2,tx);
935 fiy2 = _mm256_add_pd(fiy2,ty);
936 fiz2 = _mm256_add_pd(fiz2,tz);
938 fjx2 = _mm256_add_pd(fjx2,tx);
939 fjy2 = _mm256_add_pd(fjy2,ty);
940 fjz2 = _mm256_add_pd(fjz2,tz);
942 fjptrA = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
943 fjptrB = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
944 fjptrC = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
945 fjptrD = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
947 gmx_mm256_decrement_3rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,
948 fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2);
950 /* Inner loop uses 300 flops */
953 /* End of innermost loop */
955 gmx_mm256_update_iforce_3atom_swizzle_pd(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,
956 f+i_coord_offset,fshift+i_shift_offset);
959 /* Update potential energies */
960 gmx_mm256_update_1pot_pd(velecsum,kernel_data->energygrp_elec+ggid);
961 gmx_mm256_update_1pot_pd(vvdwsum,kernel_data->energygrp_vdw+ggid);
963 /* Increment number of inner iterations */
964 inneriter += j_index_end - j_index_start;
966 /* Outer loop uses 20 flops */
969 /* Increment number of outer iterations */
972 /* Update outer/inner flops */
974 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3W3_VF,outeriter*20 + inneriter*300);
977 * Gromacs nonbonded kernel: nb_kernel_ElecRF_VdwLJ_GeomW3W3_F_avx_256_double
978 * Electrostatics interaction: ReactionField
979 * VdW interaction: LennardJones
980 * Geometry: Water3-Water3
981 * Calculate force/pot: Force
984 nb_kernel_ElecRF_VdwLJ_GeomW3W3_F_avx_256_double
985 (t_nblist * gmx_restrict nlist,
986 rvec * gmx_restrict xx,
987 rvec * gmx_restrict ff,
988 t_forcerec * gmx_restrict fr,
989 t_mdatoms * gmx_restrict mdatoms,
990 nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
991 t_nrnb * gmx_restrict nrnb)
993 /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
994 * just 0 for non-waters.
995 * Suffixes A,B,C,D refer to j loop unrolling done with AVX, e.g. for the four different
996 * jnr indices corresponding to data put in the four positions in the SIMD register.
998 int i_shift_offset,i_coord_offset,outeriter,inneriter;
999 int j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
1000 int jnrA,jnrB,jnrC,jnrD;
1001 int jnrlistA,jnrlistB,jnrlistC,jnrlistD;
1002 int jnrlistE,jnrlistF,jnrlistG,jnrlistH;
1003 int j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
1004 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
1005 real rcutoff_scalar;
1006 real *shiftvec,*fshift,*x,*f;
1007 real *fjptrA,*fjptrB,*fjptrC,*fjptrD;
1008 real scratch[4*DIM];
1009 __m256d tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
1010 real * vdwioffsetptr0;
1011 __m256d ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
1012 real * vdwioffsetptr1;
1013 __m256d ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
1014 real * vdwioffsetptr2;
1015 __m256d ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
1016 int vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D;
1017 __m256d jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
1018 int vdwjidx1A,vdwjidx1B,vdwjidx1C,vdwjidx1D;
1019 __m256d jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
1020 int vdwjidx2A,vdwjidx2B,vdwjidx2C,vdwjidx2D;
1021 __m256d jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
1022 __m256d dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
1023 __m256d dx01,dy01,dz01,rsq01,rinv01,rinvsq01,r01,qq01,c6_01,c12_01;
1024 __m256d dx02,dy02,dz02,rsq02,rinv02,rinvsq02,r02,qq02,c6_02,c12_02;
1025 __m256d dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
1026 __m256d dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11;
1027 __m256d dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12;
1028 __m256d dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
1029 __m256d dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21;
1030 __m256d dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22;
1031 __m256d velec,felec,velecsum,facel,crf,krf,krf2;
1034 __m256d rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
1037 __m256d one_sixth = _mm256_set1_pd(1.0/6.0);
1038 __m256d one_twelfth = _mm256_set1_pd(1.0/12.0);
1039 __m256d dummy_mask,cutoff_mask;
1040 __m128 tmpmask0,tmpmask1;
1041 __m256d signbit = _mm256_castsi256_pd( _mm256_set1_epi32(0x80000000) );
1042 __m256d one = _mm256_set1_pd(1.0);
1043 __m256d two = _mm256_set1_pd(2.0);
1049 jindex = nlist->jindex;
1051 shiftidx = nlist->shift;
1053 shiftvec = fr->shift_vec[0];
1054 fshift = fr->fshift[0];
1055 facel = _mm256_set1_pd(fr->epsfac);
1056 charge = mdatoms->chargeA;
1057 krf = _mm256_set1_pd(fr->ic->k_rf);
1058 krf2 = _mm256_set1_pd(fr->ic->k_rf*2.0);
1059 crf = _mm256_set1_pd(fr->ic->c_rf);
1060 nvdwtype = fr->ntype;
1061 vdwparam = fr->nbfp;
1062 vdwtype = mdatoms->typeA;
1064 /* Setup water-specific parameters */
1065 inr = nlist->iinr[0];
1066 iq0 = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+0]));
1067 iq1 = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+1]));
1068 iq2 = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+2]));
1069 vdwioffsetptr0 = vdwparam+2*nvdwtype*vdwtype[inr+0];
1071 jq0 = _mm256_set1_pd(charge[inr+0]);
1072 jq1 = _mm256_set1_pd(charge[inr+1]);
1073 jq2 = _mm256_set1_pd(charge[inr+2]);
1074 vdwjidx0A = 2*vdwtype[inr+0];
1075 qq00 = _mm256_mul_pd(iq0,jq0);
1076 c6_00 = _mm256_set1_pd(vdwioffsetptr0[vdwjidx0A]);
1077 c12_00 = _mm256_set1_pd(vdwioffsetptr0[vdwjidx0A+1]);
1078 qq01 = _mm256_mul_pd(iq0,jq1);
1079 qq02 = _mm256_mul_pd(iq0,jq2);
1080 qq10 = _mm256_mul_pd(iq1,jq0);
1081 qq11 = _mm256_mul_pd(iq1,jq1);
1082 qq12 = _mm256_mul_pd(iq1,jq2);
1083 qq20 = _mm256_mul_pd(iq2,jq0);
1084 qq21 = _mm256_mul_pd(iq2,jq1);
1085 qq22 = _mm256_mul_pd(iq2,jq2);
1087 /* Avoid stupid compiler warnings */
1088 jnrA = jnrB = jnrC = jnrD = 0;
1089 j_coord_offsetA = 0;
1090 j_coord_offsetB = 0;
1091 j_coord_offsetC = 0;
1092 j_coord_offsetD = 0;
1097 for(iidx=0;iidx<4*DIM;iidx++)
1099 scratch[iidx] = 0.0;
1102 /* Start outer loop over neighborlists */
1103 for(iidx=0; iidx<nri; iidx++)
1105 /* Load shift vector for this list */
1106 i_shift_offset = DIM*shiftidx[iidx];
1108 /* Load limits for loop over neighbors */
1109 j_index_start = jindex[iidx];
1110 j_index_end = jindex[iidx+1];
1112 /* Get outer coordinate index */
1114 i_coord_offset = DIM*inr;
1116 /* Load i particle coords and add shift vector */
1117 gmx_mm256_load_shift_and_3rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset,
1118 &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2);
1120 fix0 = _mm256_setzero_pd();
1121 fiy0 = _mm256_setzero_pd();
1122 fiz0 = _mm256_setzero_pd();
1123 fix1 = _mm256_setzero_pd();
1124 fiy1 = _mm256_setzero_pd();
1125 fiz1 = _mm256_setzero_pd();
1126 fix2 = _mm256_setzero_pd();
1127 fiy2 = _mm256_setzero_pd();
1128 fiz2 = _mm256_setzero_pd();
1130 /* Start inner kernel loop */
1131 for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4)
1134 /* Get j neighbor index, and coordinate index */
1136 jnrB = jjnr[jidx+1];
1137 jnrC = jjnr[jidx+2];
1138 jnrD = jjnr[jidx+3];
1139 j_coord_offsetA = DIM*jnrA;
1140 j_coord_offsetB = DIM*jnrB;
1141 j_coord_offsetC = DIM*jnrC;
1142 j_coord_offsetD = DIM*jnrD;
1144 /* load j atom coordinates */
1145 gmx_mm256_load_3rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
1146 x+j_coord_offsetC,x+j_coord_offsetD,
1147 &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,&jy2,&jz2);
1149 /* Calculate displacement vector */
1150 dx00 = _mm256_sub_pd(ix0,jx0);
1151 dy00 = _mm256_sub_pd(iy0,jy0);
1152 dz00 = _mm256_sub_pd(iz0,jz0);
1153 dx01 = _mm256_sub_pd(ix0,jx1);
1154 dy01 = _mm256_sub_pd(iy0,jy1);
1155 dz01 = _mm256_sub_pd(iz0,jz1);
1156 dx02 = _mm256_sub_pd(ix0,jx2);
1157 dy02 = _mm256_sub_pd(iy0,jy2);
1158 dz02 = _mm256_sub_pd(iz0,jz2);
1159 dx10 = _mm256_sub_pd(ix1,jx0);
1160 dy10 = _mm256_sub_pd(iy1,jy0);
1161 dz10 = _mm256_sub_pd(iz1,jz0);
1162 dx11 = _mm256_sub_pd(ix1,jx1);
1163 dy11 = _mm256_sub_pd(iy1,jy1);
1164 dz11 = _mm256_sub_pd(iz1,jz1);
1165 dx12 = _mm256_sub_pd(ix1,jx2);
1166 dy12 = _mm256_sub_pd(iy1,jy2);
1167 dz12 = _mm256_sub_pd(iz1,jz2);
1168 dx20 = _mm256_sub_pd(ix2,jx0);
1169 dy20 = _mm256_sub_pd(iy2,jy0);
1170 dz20 = _mm256_sub_pd(iz2,jz0);
1171 dx21 = _mm256_sub_pd(ix2,jx1);
1172 dy21 = _mm256_sub_pd(iy2,jy1);
1173 dz21 = _mm256_sub_pd(iz2,jz1);
1174 dx22 = _mm256_sub_pd(ix2,jx2);
1175 dy22 = _mm256_sub_pd(iy2,jy2);
1176 dz22 = _mm256_sub_pd(iz2,jz2);
1178 /* Calculate squared distance and things based on it */
1179 rsq00 = gmx_mm256_calc_rsq_pd(dx00,dy00,dz00);
1180 rsq01 = gmx_mm256_calc_rsq_pd(dx01,dy01,dz01);
1181 rsq02 = gmx_mm256_calc_rsq_pd(dx02,dy02,dz02);
1182 rsq10 = gmx_mm256_calc_rsq_pd(dx10,dy10,dz10);
1183 rsq11 = gmx_mm256_calc_rsq_pd(dx11,dy11,dz11);
1184 rsq12 = gmx_mm256_calc_rsq_pd(dx12,dy12,dz12);
1185 rsq20 = gmx_mm256_calc_rsq_pd(dx20,dy20,dz20);
1186 rsq21 = gmx_mm256_calc_rsq_pd(dx21,dy21,dz21);
1187 rsq22 = gmx_mm256_calc_rsq_pd(dx22,dy22,dz22);
1189 rinv00 = gmx_mm256_invsqrt_pd(rsq00);
1190 rinv01 = gmx_mm256_invsqrt_pd(rsq01);
1191 rinv02 = gmx_mm256_invsqrt_pd(rsq02);
1192 rinv10 = gmx_mm256_invsqrt_pd(rsq10);
1193 rinv11 = gmx_mm256_invsqrt_pd(rsq11);
1194 rinv12 = gmx_mm256_invsqrt_pd(rsq12);
1195 rinv20 = gmx_mm256_invsqrt_pd(rsq20);
1196 rinv21 = gmx_mm256_invsqrt_pd(rsq21);
1197 rinv22 = gmx_mm256_invsqrt_pd(rsq22);
1199 rinvsq00 = _mm256_mul_pd(rinv00,rinv00);
1200 rinvsq01 = _mm256_mul_pd(rinv01,rinv01);
1201 rinvsq02 = _mm256_mul_pd(rinv02,rinv02);
1202 rinvsq10 = _mm256_mul_pd(rinv10,rinv10);
1203 rinvsq11 = _mm256_mul_pd(rinv11,rinv11);
1204 rinvsq12 = _mm256_mul_pd(rinv12,rinv12);
1205 rinvsq20 = _mm256_mul_pd(rinv20,rinv20);
1206 rinvsq21 = _mm256_mul_pd(rinv21,rinv21);
1207 rinvsq22 = _mm256_mul_pd(rinv22,rinv22);
1209 fjx0 = _mm256_setzero_pd();
1210 fjy0 = _mm256_setzero_pd();
1211 fjz0 = _mm256_setzero_pd();
1212 fjx1 = _mm256_setzero_pd();
1213 fjy1 = _mm256_setzero_pd();
1214 fjz1 = _mm256_setzero_pd();
1215 fjx2 = _mm256_setzero_pd();
1216 fjy2 = _mm256_setzero_pd();
1217 fjz2 = _mm256_setzero_pd();
1219 /**************************
1220 * CALCULATE INTERACTIONS *
1221 **************************/
1223 /* REACTION-FIELD ELECTROSTATICS */
1224 felec = _mm256_mul_pd(qq00,_mm256_sub_pd(_mm256_mul_pd(rinv00,rinvsq00),krf2));
1226 /* LENNARD-JONES DISPERSION/REPULSION */
1228 rinvsix = _mm256_mul_pd(_mm256_mul_pd(rinvsq00,rinvsq00),rinvsq00);
1229 fvdw = _mm256_mul_pd(_mm256_sub_pd(_mm256_mul_pd(c12_00,rinvsix),c6_00),_mm256_mul_pd(rinvsix,rinvsq00));
1231 fscal = _mm256_add_pd(felec,fvdw);
1233 /* Calculate temporary vectorial force */
1234 tx = _mm256_mul_pd(fscal,dx00);
1235 ty = _mm256_mul_pd(fscal,dy00);
1236 tz = _mm256_mul_pd(fscal,dz00);
1238 /* Update vectorial force */
1239 fix0 = _mm256_add_pd(fix0,tx);
1240 fiy0 = _mm256_add_pd(fiy0,ty);
1241 fiz0 = _mm256_add_pd(fiz0,tz);
1243 fjx0 = _mm256_add_pd(fjx0,tx);
1244 fjy0 = _mm256_add_pd(fjy0,ty);
1245 fjz0 = _mm256_add_pd(fjz0,tz);
1247 /**************************
1248 * CALCULATE INTERACTIONS *
1249 **************************/
1251 /* REACTION-FIELD ELECTROSTATICS */
1252 felec = _mm256_mul_pd(qq01,_mm256_sub_pd(_mm256_mul_pd(rinv01,rinvsq01),krf2));
1256 /* Calculate temporary vectorial force */
1257 tx = _mm256_mul_pd(fscal,dx01);
1258 ty = _mm256_mul_pd(fscal,dy01);
1259 tz = _mm256_mul_pd(fscal,dz01);
1261 /* Update vectorial force */
1262 fix0 = _mm256_add_pd(fix0,tx);
1263 fiy0 = _mm256_add_pd(fiy0,ty);
1264 fiz0 = _mm256_add_pd(fiz0,tz);
1266 fjx1 = _mm256_add_pd(fjx1,tx);
1267 fjy1 = _mm256_add_pd(fjy1,ty);
1268 fjz1 = _mm256_add_pd(fjz1,tz);
1270 /**************************
1271 * CALCULATE INTERACTIONS *
1272 **************************/
1274 /* REACTION-FIELD ELECTROSTATICS */
1275 felec = _mm256_mul_pd(qq02,_mm256_sub_pd(_mm256_mul_pd(rinv02,rinvsq02),krf2));
1279 /* Calculate temporary vectorial force */
1280 tx = _mm256_mul_pd(fscal,dx02);
1281 ty = _mm256_mul_pd(fscal,dy02);
1282 tz = _mm256_mul_pd(fscal,dz02);
1284 /* Update vectorial force */
1285 fix0 = _mm256_add_pd(fix0,tx);
1286 fiy0 = _mm256_add_pd(fiy0,ty);
1287 fiz0 = _mm256_add_pd(fiz0,tz);
1289 fjx2 = _mm256_add_pd(fjx2,tx);
1290 fjy2 = _mm256_add_pd(fjy2,ty);
1291 fjz2 = _mm256_add_pd(fjz2,tz);
1293 /**************************
1294 * CALCULATE INTERACTIONS *
1295 **************************/
1297 /* REACTION-FIELD ELECTROSTATICS */
1298 felec = _mm256_mul_pd(qq10,_mm256_sub_pd(_mm256_mul_pd(rinv10,rinvsq10),krf2));
1302 /* Calculate temporary vectorial force */
1303 tx = _mm256_mul_pd(fscal,dx10);
1304 ty = _mm256_mul_pd(fscal,dy10);
1305 tz = _mm256_mul_pd(fscal,dz10);
1307 /* Update vectorial force */
1308 fix1 = _mm256_add_pd(fix1,tx);
1309 fiy1 = _mm256_add_pd(fiy1,ty);
1310 fiz1 = _mm256_add_pd(fiz1,tz);
1312 fjx0 = _mm256_add_pd(fjx0,tx);
1313 fjy0 = _mm256_add_pd(fjy0,ty);
1314 fjz0 = _mm256_add_pd(fjz0,tz);
1316 /**************************
1317 * CALCULATE INTERACTIONS *
1318 **************************/
1320 /* REACTION-FIELD ELECTROSTATICS */
1321 felec = _mm256_mul_pd(qq11,_mm256_sub_pd(_mm256_mul_pd(rinv11,rinvsq11),krf2));
1325 /* Calculate temporary vectorial force */
1326 tx = _mm256_mul_pd(fscal,dx11);
1327 ty = _mm256_mul_pd(fscal,dy11);
1328 tz = _mm256_mul_pd(fscal,dz11);
1330 /* Update vectorial force */
1331 fix1 = _mm256_add_pd(fix1,tx);
1332 fiy1 = _mm256_add_pd(fiy1,ty);
1333 fiz1 = _mm256_add_pd(fiz1,tz);
1335 fjx1 = _mm256_add_pd(fjx1,tx);
1336 fjy1 = _mm256_add_pd(fjy1,ty);
1337 fjz1 = _mm256_add_pd(fjz1,tz);
1339 /**************************
1340 * CALCULATE INTERACTIONS *
1341 **************************/
1343 /* REACTION-FIELD ELECTROSTATICS */
1344 felec = _mm256_mul_pd(qq12,_mm256_sub_pd(_mm256_mul_pd(rinv12,rinvsq12),krf2));
1348 /* Calculate temporary vectorial force */
1349 tx = _mm256_mul_pd(fscal,dx12);
1350 ty = _mm256_mul_pd(fscal,dy12);
1351 tz = _mm256_mul_pd(fscal,dz12);
1353 /* Update vectorial force */
1354 fix1 = _mm256_add_pd(fix1,tx);
1355 fiy1 = _mm256_add_pd(fiy1,ty);
1356 fiz1 = _mm256_add_pd(fiz1,tz);
1358 fjx2 = _mm256_add_pd(fjx2,tx);
1359 fjy2 = _mm256_add_pd(fjy2,ty);
1360 fjz2 = _mm256_add_pd(fjz2,tz);
1362 /**************************
1363 * CALCULATE INTERACTIONS *
1364 **************************/
1366 /* REACTION-FIELD ELECTROSTATICS */
1367 felec = _mm256_mul_pd(qq20,_mm256_sub_pd(_mm256_mul_pd(rinv20,rinvsq20),krf2));
1371 /* Calculate temporary vectorial force */
1372 tx = _mm256_mul_pd(fscal,dx20);
1373 ty = _mm256_mul_pd(fscal,dy20);
1374 tz = _mm256_mul_pd(fscal,dz20);
1376 /* Update vectorial force */
1377 fix2 = _mm256_add_pd(fix2,tx);
1378 fiy2 = _mm256_add_pd(fiy2,ty);
1379 fiz2 = _mm256_add_pd(fiz2,tz);
1381 fjx0 = _mm256_add_pd(fjx0,tx);
1382 fjy0 = _mm256_add_pd(fjy0,ty);
1383 fjz0 = _mm256_add_pd(fjz0,tz);
1385 /**************************
1386 * CALCULATE INTERACTIONS *
1387 **************************/
1389 /* REACTION-FIELD ELECTROSTATICS */
1390 felec = _mm256_mul_pd(qq21,_mm256_sub_pd(_mm256_mul_pd(rinv21,rinvsq21),krf2));
1394 /* Calculate temporary vectorial force */
1395 tx = _mm256_mul_pd(fscal,dx21);
1396 ty = _mm256_mul_pd(fscal,dy21);
1397 tz = _mm256_mul_pd(fscal,dz21);
1399 /* Update vectorial force */
1400 fix2 = _mm256_add_pd(fix2,tx);
1401 fiy2 = _mm256_add_pd(fiy2,ty);
1402 fiz2 = _mm256_add_pd(fiz2,tz);
1404 fjx1 = _mm256_add_pd(fjx1,tx);
1405 fjy1 = _mm256_add_pd(fjy1,ty);
1406 fjz1 = _mm256_add_pd(fjz1,tz);
1408 /**************************
1409 * CALCULATE INTERACTIONS *
1410 **************************/
1412 /* REACTION-FIELD ELECTROSTATICS */
1413 felec = _mm256_mul_pd(qq22,_mm256_sub_pd(_mm256_mul_pd(rinv22,rinvsq22),krf2));
1417 /* Calculate temporary vectorial force */
1418 tx = _mm256_mul_pd(fscal,dx22);
1419 ty = _mm256_mul_pd(fscal,dy22);
1420 tz = _mm256_mul_pd(fscal,dz22);
1422 /* Update vectorial force */
1423 fix2 = _mm256_add_pd(fix2,tx);
1424 fiy2 = _mm256_add_pd(fiy2,ty);
1425 fiz2 = _mm256_add_pd(fiz2,tz);
1427 fjx2 = _mm256_add_pd(fjx2,tx);
1428 fjy2 = _mm256_add_pd(fjy2,ty);
1429 fjz2 = _mm256_add_pd(fjz2,tz);
1431 fjptrA = f+j_coord_offsetA;
1432 fjptrB = f+j_coord_offsetB;
1433 fjptrC = f+j_coord_offsetC;
1434 fjptrD = f+j_coord_offsetD;
1436 gmx_mm256_decrement_3rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,
1437 fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2);
1439 /* Inner loop uses 250 flops */
1442 if(jidx<j_index_end)
1445 /* Get j neighbor index, and coordinate index */
1446 jnrlistA = jjnr[jidx];
1447 jnrlistB = jjnr[jidx+1];
1448 jnrlistC = jjnr[jidx+2];
1449 jnrlistD = jjnr[jidx+3];
1450 /* Sign of each element will be negative for non-real atoms.
1451 * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
1452 * so use it as val = _mm_andnot_pd(mask,val) to clear dummy entries.
1454 tmpmask0 = gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128()));
1456 tmpmask1 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(3,3,2,2));
1457 tmpmask0 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(1,1,0,0));
1458 dummy_mask = _mm256_castps_pd(gmx_mm256_set_m128(tmpmask1,tmpmask0));
1460 jnrA = (jnrlistA>=0) ? jnrlistA : 0;
1461 jnrB = (jnrlistB>=0) ? jnrlistB : 0;
1462 jnrC = (jnrlistC>=0) ? jnrlistC : 0;
1463 jnrD = (jnrlistD>=0) ? jnrlistD : 0;
1464 j_coord_offsetA = DIM*jnrA;
1465 j_coord_offsetB = DIM*jnrB;
1466 j_coord_offsetC = DIM*jnrC;
1467 j_coord_offsetD = DIM*jnrD;
1469 /* load j atom coordinates */
1470 gmx_mm256_load_3rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
1471 x+j_coord_offsetC,x+j_coord_offsetD,
1472 &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,&jy2,&jz2);
1474 /* Calculate displacement vector */
1475 dx00 = _mm256_sub_pd(ix0,jx0);
1476 dy00 = _mm256_sub_pd(iy0,jy0);
1477 dz00 = _mm256_sub_pd(iz0,jz0);
1478 dx01 = _mm256_sub_pd(ix0,jx1);
1479 dy01 = _mm256_sub_pd(iy0,jy1);
1480 dz01 = _mm256_sub_pd(iz0,jz1);
1481 dx02 = _mm256_sub_pd(ix0,jx2);
1482 dy02 = _mm256_sub_pd(iy0,jy2);
1483 dz02 = _mm256_sub_pd(iz0,jz2);
1484 dx10 = _mm256_sub_pd(ix1,jx0);
1485 dy10 = _mm256_sub_pd(iy1,jy0);
1486 dz10 = _mm256_sub_pd(iz1,jz0);
1487 dx11 = _mm256_sub_pd(ix1,jx1);
1488 dy11 = _mm256_sub_pd(iy1,jy1);
1489 dz11 = _mm256_sub_pd(iz1,jz1);
1490 dx12 = _mm256_sub_pd(ix1,jx2);
1491 dy12 = _mm256_sub_pd(iy1,jy2);
1492 dz12 = _mm256_sub_pd(iz1,jz2);
1493 dx20 = _mm256_sub_pd(ix2,jx0);
1494 dy20 = _mm256_sub_pd(iy2,jy0);
1495 dz20 = _mm256_sub_pd(iz2,jz0);
1496 dx21 = _mm256_sub_pd(ix2,jx1);
1497 dy21 = _mm256_sub_pd(iy2,jy1);
1498 dz21 = _mm256_sub_pd(iz2,jz1);
1499 dx22 = _mm256_sub_pd(ix2,jx2);
1500 dy22 = _mm256_sub_pd(iy2,jy2);
1501 dz22 = _mm256_sub_pd(iz2,jz2);
1503 /* Calculate squared distance and things based on it */
1504 rsq00 = gmx_mm256_calc_rsq_pd(dx00,dy00,dz00);
1505 rsq01 = gmx_mm256_calc_rsq_pd(dx01,dy01,dz01);
1506 rsq02 = gmx_mm256_calc_rsq_pd(dx02,dy02,dz02);
1507 rsq10 = gmx_mm256_calc_rsq_pd(dx10,dy10,dz10);
1508 rsq11 = gmx_mm256_calc_rsq_pd(dx11,dy11,dz11);
1509 rsq12 = gmx_mm256_calc_rsq_pd(dx12,dy12,dz12);
1510 rsq20 = gmx_mm256_calc_rsq_pd(dx20,dy20,dz20);
1511 rsq21 = gmx_mm256_calc_rsq_pd(dx21,dy21,dz21);
1512 rsq22 = gmx_mm256_calc_rsq_pd(dx22,dy22,dz22);
1514 rinv00 = gmx_mm256_invsqrt_pd(rsq00);
1515 rinv01 = gmx_mm256_invsqrt_pd(rsq01);
1516 rinv02 = gmx_mm256_invsqrt_pd(rsq02);
1517 rinv10 = gmx_mm256_invsqrt_pd(rsq10);
1518 rinv11 = gmx_mm256_invsqrt_pd(rsq11);
1519 rinv12 = gmx_mm256_invsqrt_pd(rsq12);
1520 rinv20 = gmx_mm256_invsqrt_pd(rsq20);
1521 rinv21 = gmx_mm256_invsqrt_pd(rsq21);
1522 rinv22 = gmx_mm256_invsqrt_pd(rsq22);
1524 rinvsq00 = _mm256_mul_pd(rinv00,rinv00);
1525 rinvsq01 = _mm256_mul_pd(rinv01,rinv01);
1526 rinvsq02 = _mm256_mul_pd(rinv02,rinv02);
1527 rinvsq10 = _mm256_mul_pd(rinv10,rinv10);
1528 rinvsq11 = _mm256_mul_pd(rinv11,rinv11);
1529 rinvsq12 = _mm256_mul_pd(rinv12,rinv12);
1530 rinvsq20 = _mm256_mul_pd(rinv20,rinv20);
1531 rinvsq21 = _mm256_mul_pd(rinv21,rinv21);
1532 rinvsq22 = _mm256_mul_pd(rinv22,rinv22);
1534 fjx0 = _mm256_setzero_pd();
1535 fjy0 = _mm256_setzero_pd();
1536 fjz0 = _mm256_setzero_pd();
1537 fjx1 = _mm256_setzero_pd();
1538 fjy1 = _mm256_setzero_pd();
1539 fjz1 = _mm256_setzero_pd();
1540 fjx2 = _mm256_setzero_pd();
1541 fjy2 = _mm256_setzero_pd();
1542 fjz2 = _mm256_setzero_pd();
1544 /**************************
1545 * CALCULATE INTERACTIONS *
1546 **************************/
1548 /* REACTION-FIELD ELECTROSTATICS */
1549 felec = _mm256_mul_pd(qq00,_mm256_sub_pd(_mm256_mul_pd(rinv00,rinvsq00),krf2));
1551 /* LENNARD-JONES DISPERSION/REPULSION */
1553 rinvsix = _mm256_mul_pd(_mm256_mul_pd(rinvsq00,rinvsq00),rinvsq00);
1554 fvdw = _mm256_mul_pd(_mm256_sub_pd(_mm256_mul_pd(c12_00,rinvsix),c6_00),_mm256_mul_pd(rinvsix,rinvsq00));
1556 fscal = _mm256_add_pd(felec,fvdw);
1558 fscal = _mm256_andnot_pd(dummy_mask,fscal);
1560 /* Calculate temporary vectorial force */
1561 tx = _mm256_mul_pd(fscal,dx00);
1562 ty = _mm256_mul_pd(fscal,dy00);
1563 tz = _mm256_mul_pd(fscal,dz00);
1565 /* Update vectorial force */
1566 fix0 = _mm256_add_pd(fix0,tx);
1567 fiy0 = _mm256_add_pd(fiy0,ty);
1568 fiz0 = _mm256_add_pd(fiz0,tz);
1570 fjx0 = _mm256_add_pd(fjx0,tx);
1571 fjy0 = _mm256_add_pd(fjy0,ty);
1572 fjz0 = _mm256_add_pd(fjz0,tz);
1574 /**************************
1575 * CALCULATE INTERACTIONS *
1576 **************************/
1578 /* REACTION-FIELD ELECTROSTATICS */
1579 felec = _mm256_mul_pd(qq01,_mm256_sub_pd(_mm256_mul_pd(rinv01,rinvsq01),krf2));
1583 fscal = _mm256_andnot_pd(dummy_mask,fscal);
1585 /* Calculate temporary vectorial force */
1586 tx = _mm256_mul_pd(fscal,dx01);
1587 ty = _mm256_mul_pd(fscal,dy01);
1588 tz = _mm256_mul_pd(fscal,dz01);
1590 /* Update vectorial force */
1591 fix0 = _mm256_add_pd(fix0,tx);
1592 fiy0 = _mm256_add_pd(fiy0,ty);
1593 fiz0 = _mm256_add_pd(fiz0,tz);
1595 fjx1 = _mm256_add_pd(fjx1,tx);
1596 fjy1 = _mm256_add_pd(fjy1,ty);
1597 fjz1 = _mm256_add_pd(fjz1,tz);
1599 /**************************
1600 * CALCULATE INTERACTIONS *
1601 **************************/
1603 /* REACTION-FIELD ELECTROSTATICS */
1604 felec = _mm256_mul_pd(qq02,_mm256_sub_pd(_mm256_mul_pd(rinv02,rinvsq02),krf2));
1608 fscal = _mm256_andnot_pd(dummy_mask,fscal);
1610 /* Calculate temporary vectorial force */
1611 tx = _mm256_mul_pd(fscal,dx02);
1612 ty = _mm256_mul_pd(fscal,dy02);
1613 tz = _mm256_mul_pd(fscal,dz02);
1615 /* Update vectorial force */
1616 fix0 = _mm256_add_pd(fix0,tx);
1617 fiy0 = _mm256_add_pd(fiy0,ty);
1618 fiz0 = _mm256_add_pd(fiz0,tz);
1620 fjx2 = _mm256_add_pd(fjx2,tx);
1621 fjy2 = _mm256_add_pd(fjy2,ty);
1622 fjz2 = _mm256_add_pd(fjz2,tz);
1624 /**************************
1625 * CALCULATE INTERACTIONS *
1626 **************************/
1628 /* REACTION-FIELD ELECTROSTATICS */
1629 felec = _mm256_mul_pd(qq10,_mm256_sub_pd(_mm256_mul_pd(rinv10,rinvsq10),krf2));
1633 fscal = _mm256_andnot_pd(dummy_mask,fscal);
1635 /* Calculate temporary vectorial force */
1636 tx = _mm256_mul_pd(fscal,dx10);
1637 ty = _mm256_mul_pd(fscal,dy10);
1638 tz = _mm256_mul_pd(fscal,dz10);
1640 /* Update vectorial force */
1641 fix1 = _mm256_add_pd(fix1,tx);
1642 fiy1 = _mm256_add_pd(fiy1,ty);
1643 fiz1 = _mm256_add_pd(fiz1,tz);
1645 fjx0 = _mm256_add_pd(fjx0,tx);
1646 fjy0 = _mm256_add_pd(fjy0,ty);
1647 fjz0 = _mm256_add_pd(fjz0,tz);
1649 /**************************
1650 * CALCULATE INTERACTIONS *
1651 **************************/
1653 /* REACTION-FIELD ELECTROSTATICS */
1654 felec = _mm256_mul_pd(qq11,_mm256_sub_pd(_mm256_mul_pd(rinv11,rinvsq11),krf2));
1658 fscal = _mm256_andnot_pd(dummy_mask,fscal);
1660 /* Calculate temporary vectorial force */
1661 tx = _mm256_mul_pd(fscal,dx11);
1662 ty = _mm256_mul_pd(fscal,dy11);
1663 tz = _mm256_mul_pd(fscal,dz11);
1665 /* Update vectorial force */
1666 fix1 = _mm256_add_pd(fix1,tx);
1667 fiy1 = _mm256_add_pd(fiy1,ty);
1668 fiz1 = _mm256_add_pd(fiz1,tz);
1670 fjx1 = _mm256_add_pd(fjx1,tx);
1671 fjy1 = _mm256_add_pd(fjy1,ty);
1672 fjz1 = _mm256_add_pd(fjz1,tz);
1674 /**************************
1675 * CALCULATE INTERACTIONS *
1676 **************************/
1678 /* REACTION-FIELD ELECTROSTATICS */
1679 felec = _mm256_mul_pd(qq12,_mm256_sub_pd(_mm256_mul_pd(rinv12,rinvsq12),krf2));
1683 fscal = _mm256_andnot_pd(dummy_mask,fscal);
1685 /* Calculate temporary vectorial force */
1686 tx = _mm256_mul_pd(fscal,dx12);
1687 ty = _mm256_mul_pd(fscal,dy12);
1688 tz = _mm256_mul_pd(fscal,dz12);
1690 /* Update vectorial force */
1691 fix1 = _mm256_add_pd(fix1,tx);
1692 fiy1 = _mm256_add_pd(fiy1,ty);
1693 fiz1 = _mm256_add_pd(fiz1,tz);
1695 fjx2 = _mm256_add_pd(fjx2,tx);
1696 fjy2 = _mm256_add_pd(fjy2,ty);
1697 fjz2 = _mm256_add_pd(fjz2,tz);
1699 /**************************
1700 * CALCULATE INTERACTIONS *
1701 **************************/
1703 /* REACTION-FIELD ELECTROSTATICS */
1704 felec = _mm256_mul_pd(qq20,_mm256_sub_pd(_mm256_mul_pd(rinv20,rinvsq20),krf2));
1708 fscal = _mm256_andnot_pd(dummy_mask,fscal);
1710 /* Calculate temporary vectorial force */
1711 tx = _mm256_mul_pd(fscal,dx20);
1712 ty = _mm256_mul_pd(fscal,dy20);
1713 tz = _mm256_mul_pd(fscal,dz20);
1715 /* Update vectorial force */
1716 fix2 = _mm256_add_pd(fix2,tx);
1717 fiy2 = _mm256_add_pd(fiy2,ty);
1718 fiz2 = _mm256_add_pd(fiz2,tz);
1720 fjx0 = _mm256_add_pd(fjx0,tx);
1721 fjy0 = _mm256_add_pd(fjy0,ty);
1722 fjz0 = _mm256_add_pd(fjz0,tz);
1724 /**************************
1725 * CALCULATE INTERACTIONS *
1726 **************************/
1728 /* REACTION-FIELD ELECTROSTATICS */
1729 felec = _mm256_mul_pd(qq21,_mm256_sub_pd(_mm256_mul_pd(rinv21,rinvsq21),krf2));
1733 fscal = _mm256_andnot_pd(dummy_mask,fscal);
1735 /* Calculate temporary vectorial force */
1736 tx = _mm256_mul_pd(fscal,dx21);
1737 ty = _mm256_mul_pd(fscal,dy21);
1738 tz = _mm256_mul_pd(fscal,dz21);
1740 /* Update vectorial force */
1741 fix2 = _mm256_add_pd(fix2,tx);
1742 fiy2 = _mm256_add_pd(fiy2,ty);
1743 fiz2 = _mm256_add_pd(fiz2,tz);
1745 fjx1 = _mm256_add_pd(fjx1,tx);
1746 fjy1 = _mm256_add_pd(fjy1,ty);
1747 fjz1 = _mm256_add_pd(fjz1,tz);
1749 /**************************
1750 * CALCULATE INTERACTIONS *
1751 **************************/
1753 /* REACTION-FIELD ELECTROSTATICS */
1754 felec = _mm256_mul_pd(qq22,_mm256_sub_pd(_mm256_mul_pd(rinv22,rinvsq22),krf2));
1758 fscal = _mm256_andnot_pd(dummy_mask,fscal);
1760 /* Calculate temporary vectorial force */
1761 tx = _mm256_mul_pd(fscal,dx22);
1762 ty = _mm256_mul_pd(fscal,dy22);
1763 tz = _mm256_mul_pd(fscal,dz22);
1765 /* Update vectorial force */
1766 fix2 = _mm256_add_pd(fix2,tx);
1767 fiy2 = _mm256_add_pd(fiy2,ty);
1768 fiz2 = _mm256_add_pd(fiz2,tz);
1770 fjx2 = _mm256_add_pd(fjx2,tx);
1771 fjy2 = _mm256_add_pd(fjy2,ty);
1772 fjz2 = _mm256_add_pd(fjz2,tz);
1774 fjptrA = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
1775 fjptrB = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
1776 fjptrC = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
1777 fjptrD = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
1779 gmx_mm256_decrement_3rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,
1780 fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,fjx2,fjy2,fjz2);
1782 /* Inner loop uses 250 flops */
1785 /* End of innermost loop */
1787 gmx_mm256_update_iforce_3atom_swizzle_pd(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,
1788 f+i_coord_offset,fshift+i_shift_offset);
1790 /* Increment number of inner iterations */
1791 inneriter += j_index_end - j_index_start;
1793 /* Outer loop uses 18 flops */
1796 /* Increment number of outer iterations */
1799 /* Update outer/inner flops */
1801 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3W3_F,outeriter*18 + inneriter*250);