?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
Remove nb-parameters from t_forcerec
[gromacs.git] / src / gromacs / gmxlib / nonbonded / nb_kernel_avx_256_double / nb_kernel_ElecRFCut_VdwLJSh_GeomW4W4_avx_256_double.c
blob58d23d5193606c89e0ca30f8bd863d212ccedd05
1 /*
2 * This file is part of the GROMACS molecular simulation package.
3 *
4 * Copyright (c) 2012,2013,2014,2015,2017, 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.
8 *
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.
13 *
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.
18 *
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.
23 *
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.
31 *
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.
34 */
35 /*
36 * Note: this file was generated by the GROMACS avx_256_double kernel generator.
37 */
38 #include "gmxpre.h"
39
40 #include "config.h"
41
42 #include <math.h>
43
44 #include "../nb_kernel.h"
45 #include "gromacs/gmxlib/nrnb.h"
46
47 #include "kernelutil_x86_avx_256_double.h"
48
49 /*
50 * Gromacs nonbonded kernel: nb_kernel_ElecRFCut_VdwLJSh_GeomW4W4_VF_avx_256_double
51 * Electrostatics interaction: ReactionField
52 * VdW interaction: LennardJones
53 * Geometry: Water4-Water4
54 * Calculate force/pot: PotentialAndForce
55 */
56 void
57 nb_kernel_ElecRFCut_VdwLJSh_GeomW4W4_VF_avx_256_double
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)
65 {
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 refer to j loop unrolling done with AVX, e.g. for the four different
69 * jnr indices corresponding to data put in the four positions in the SIMD register.
70 */
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 jnrlistA,jnrlistB,jnrlistC,jnrlistD;
75 int jnrlistE,jnrlistF,jnrlistG,jnrlistH;
76 int j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
77 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
78 real rcutoff_scalar;
79 real *shiftvec,*fshift,*x,*f;
80 real *fjptrA,*fjptrB,*fjptrC,*fjptrD;
81 real scratch[4*DIM];
82 __m256d tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
83 real * vdwioffsetptr0;
84 __m256d ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
85 real * vdwioffsetptr1;
86 __m256d ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
87 real * vdwioffsetptr2;
88 __m256d ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
89 real * vdwioffsetptr3;
90 __m256d ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
91 int vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D;
92 __m256d jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
93 int vdwjidx1A,vdwjidx1B,vdwjidx1C,vdwjidx1D;
94 __m256d jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
95 int vdwjidx2A,vdwjidx2B,vdwjidx2C,vdwjidx2D;
96 __m256d jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
97 int vdwjidx3A,vdwjidx3B,vdwjidx3C,vdwjidx3D;
98 __m256d jx3,jy3,jz3,fjx3,fjy3,fjz3,jq3,isaj3;
99 __m256d dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
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 dx13,dy13,dz13,rsq13,rinv13,rinvsq13,r13,qq13,c6_13,c12_13;
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 dx23,dy23,dz23,rsq23,rinv23,rinvsq23,r23,qq23,c6_23,c12_23;
106 __m256d dx31,dy31,dz31,rsq31,rinv31,rinvsq31,r31,qq31,c6_31,c12_31;
107 __m256d dx32,dy32,dz32,rsq32,rinv32,rinvsq32,r32,qq32,c6_32,c12_32;
108 __m256d dx33,dy33,dz33,rsq33,rinv33,rinvsq33,r33,qq33,c6_33,c12_33;
109 __m256d velec,felec,velecsum,facel,crf,krf,krf2;
110 real *charge;
111 int nvdwtype;
112 __m256d rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
113 int *vdwtype;
114 real *vdwparam;
115 __m256d one_sixth = _mm256_set1_pd(1.0/6.0);
116 __m256d one_twelfth = _mm256_set1_pd(1.0/12.0);
117 __m256d dummy_mask,cutoff_mask;
118 __m128 tmpmask0,tmpmask1;
119 __m256d signbit = _mm256_castsi256_pd( _mm256_set1_epi32(0x80000000) );
120 __m256d one = _mm256_set1_pd(1.0);
121 __m256d two = _mm256_set1_pd(2.0);
122 x = xx[0];
123 f = ff[0];
124
125 nri = nlist->nri;
126 iinr = nlist->iinr;
127 jindex = nlist->jindex;
128 jjnr = nlist->jjnr;
129 shiftidx = nlist->shift;
130 gid = nlist->gid;
131 shiftvec = fr->shift_vec[0];
132 fshift = fr->fshift[0];
133 facel = _mm256_set1_pd(fr->ic->epsfac);
134 charge = mdatoms->chargeA;
135 krf = _mm256_set1_pd(fr->ic->k_rf);
136 krf2 = _mm256_set1_pd(fr->ic->k_rf*2.0);
137 crf = _mm256_set1_pd(fr->ic->c_rf);
138 nvdwtype = fr->ntype;
139 vdwparam = fr->nbfp;
140 vdwtype = mdatoms->typeA;
141
142 /* Setup water-specific parameters */
143 inr = nlist->iinr[0];
144 iq1 = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+1]));
145 iq2 = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+2]));
146 iq3 = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+3]));
147 vdwioffsetptr0 = vdwparam+2*nvdwtype*vdwtype[inr+0];
148
149 jq1 = _mm256_set1_pd(charge[inr+1]);
150 jq2 = _mm256_set1_pd(charge[inr+2]);
151 jq3 = _mm256_set1_pd(charge[inr+3]);
152 vdwjidx0A = 2*vdwtype[inr+0];
153 c6_00 = _mm256_set1_pd(vdwioffsetptr0[vdwjidx0A]);
154 c12_00 = _mm256_set1_pd(vdwioffsetptr0[vdwjidx0A+1]);
155 qq11 = _mm256_mul_pd(iq1,jq1);
156 qq12 = _mm256_mul_pd(iq1,jq2);
157 qq13 = _mm256_mul_pd(iq1,jq3);
158 qq21 = _mm256_mul_pd(iq2,jq1);
159 qq22 = _mm256_mul_pd(iq2,jq2);
160 qq23 = _mm256_mul_pd(iq2,jq3);
161 qq31 = _mm256_mul_pd(iq3,jq1);
162 qq32 = _mm256_mul_pd(iq3,jq2);
163 qq33 = _mm256_mul_pd(iq3,jq3);
164
165 /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
166 rcutoff_scalar = fr->ic->rcoulomb;
167 rcutoff = _mm256_set1_pd(rcutoff_scalar);
168 rcutoff2 = _mm256_mul_pd(rcutoff,rcutoff);
169
170 sh_vdw_invrcut6 = _mm256_set1_pd(fr->ic->sh_invrc6);
171 rvdw = _mm256_set1_pd(fr->ic->rvdw);
172
173 /* Avoid stupid compiler warnings */
174 jnrA = jnrB = jnrC = jnrD = 0;
175 j_coord_offsetA = 0;
176 j_coord_offsetB = 0;
177 j_coord_offsetC = 0;
178 j_coord_offsetD = 0;
179
180 outeriter = 0;
181 inneriter = 0;
182
183 for(iidx=0;iidx<4*DIM;iidx++)
184 {
185 scratch[iidx] = 0.0;
186 }
187
188 /* Start outer loop over neighborlists */
189 for(iidx=0; iidx<nri; iidx++)
190 {
191 /* Load shift vector for this list */
192 i_shift_offset = DIM*shiftidx[iidx];
193
194 /* Load limits for loop over neighbors */
195 j_index_start = jindex[iidx];
196 j_index_end = jindex[iidx+1];
197
198 /* Get outer coordinate index */
199 inr = iinr[iidx];
200 i_coord_offset = DIM*inr;
201
202 /* Load i particle coords and add shift vector */
203 gmx_mm256_load_shift_and_4rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset,
204 &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
205
206 fix0 = _mm256_setzero_pd();
207 fiy0 = _mm256_setzero_pd();
208 fiz0 = _mm256_setzero_pd();
209 fix1 = _mm256_setzero_pd();
210 fiy1 = _mm256_setzero_pd();
211 fiz1 = _mm256_setzero_pd();
212 fix2 = _mm256_setzero_pd();
213 fiy2 = _mm256_setzero_pd();
214 fiz2 = _mm256_setzero_pd();
215 fix3 = _mm256_setzero_pd();
216 fiy3 = _mm256_setzero_pd();
217 fiz3 = _mm256_setzero_pd();
218
219 /* Reset potential sums */
220 velecsum = _mm256_setzero_pd();
221 vvdwsum = _mm256_setzero_pd();
222
223 /* Start inner kernel loop */
224 for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4)
225 {
226
227 /* Get j neighbor index, and coordinate index */
228 jnrA = jjnr[jidx];
229 jnrB = jjnr[jidx+1];
230 jnrC = jjnr[jidx+2];
231 jnrD = jjnr[jidx+3];
232 j_coord_offsetA = DIM*jnrA;
233 j_coord_offsetB = DIM*jnrB;
234 j_coord_offsetC = DIM*jnrC;
235 j_coord_offsetD = DIM*jnrD;
236
237 /* load j atom coordinates */
238 gmx_mm256_load_4rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
239 x+j_coord_offsetC,x+j_coord_offsetD,
240 &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,
241 &jy2,&jz2,&jx3,&jy3,&jz3);
242
243 /* Calculate displacement vector */
244 dx00 = _mm256_sub_pd(ix0,jx0);
245 dy00 = _mm256_sub_pd(iy0,jy0);
246 dz00 = _mm256_sub_pd(iz0,jz0);
247 dx11 = _mm256_sub_pd(ix1,jx1);
248 dy11 = _mm256_sub_pd(iy1,jy1);
249 dz11 = _mm256_sub_pd(iz1,jz1);
250 dx12 = _mm256_sub_pd(ix1,jx2);
251 dy12 = _mm256_sub_pd(iy1,jy2);
252 dz12 = _mm256_sub_pd(iz1,jz2);
253 dx13 = _mm256_sub_pd(ix1,jx3);
254 dy13 = _mm256_sub_pd(iy1,jy3);
255 dz13 = _mm256_sub_pd(iz1,jz3);
256 dx21 = _mm256_sub_pd(ix2,jx1);
257 dy21 = _mm256_sub_pd(iy2,jy1);
258 dz21 = _mm256_sub_pd(iz2,jz1);
259 dx22 = _mm256_sub_pd(ix2,jx2);
260 dy22 = _mm256_sub_pd(iy2,jy2);
261 dz22 = _mm256_sub_pd(iz2,jz2);
262 dx23 = _mm256_sub_pd(ix2,jx3);
263 dy23 = _mm256_sub_pd(iy2,jy3);
264 dz23 = _mm256_sub_pd(iz2,jz3);
265 dx31 = _mm256_sub_pd(ix3,jx1);
266 dy31 = _mm256_sub_pd(iy3,jy1);
267 dz31 = _mm256_sub_pd(iz3,jz1);
268 dx32 = _mm256_sub_pd(ix3,jx2);
269 dy32 = _mm256_sub_pd(iy3,jy2);
270 dz32 = _mm256_sub_pd(iz3,jz2);
271 dx33 = _mm256_sub_pd(ix3,jx3);
272 dy33 = _mm256_sub_pd(iy3,jy3);
273 dz33 = _mm256_sub_pd(iz3,jz3);
274
275 /* Calculate squared distance and things based on it */
276 rsq00 = gmx_mm256_calc_rsq_pd(dx00,dy00,dz00);
277 rsq11 = gmx_mm256_calc_rsq_pd(dx11,dy11,dz11);
278 rsq12 = gmx_mm256_calc_rsq_pd(dx12,dy12,dz12);
279 rsq13 = gmx_mm256_calc_rsq_pd(dx13,dy13,dz13);
280 rsq21 = gmx_mm256_calc_rsq_pd(dx21,dy21,dz21);
281 rsq22 = gmx_mm256_calc_rsq_pd(dx22,dy22,dz22);
282 rsq23 = gmx_mm256_calc_rsq_pd(dx23,dy23,dz23);
283 rsq31 = gmx_mm256_calc_rsq_pd(dx31,dy31,dz31);
284 rsq32 = gmx_mm256_calc_rsq_pd(dx32,dy32,dz32);
285 rsq33 = gmx_mm256_calc_rsq_pd(dx33,dy33,dz33);
286
287 rinv11 = avx256_invsqrt_d(rsq11);
288 rinv12 = avx256_invsqrt_d(rsq12);
289 rinv13 = avx256_invsqrt_d(rsq13);
290 rinv21 = avx256_invsqrt_d(rsq21);
291 rinv22 = avx256_invsqrt_d(rsq22);
292 rinv23 = avx256_invsqrt_d(rsq23);
293 rinv31 = avx256_invsqrt_d(rsq31);
294 rinv32 = avx256_invsqrt_d(rsq32);
295 rinv33 = avx256_invsqrt_d(rsq33);
296
297 rinvsq00 = avx256_inv_d(rsq00);
298 rinvsq11 = _mm256_mul_pd(rinv11,rinv11);
299 rinvsq12 = _mm256_mul_pd(rinv12,rinv12);
300 rinvsq13 = _mm256_mul_pd(rinv13,rinv13);
301 rinvsq21 = _mm256_mul_pd(rinv21,rinv21);
302 rinvsq22 = _mm256_mul_pd(rinv22,rinv22);
303 rinvsq23 = _mm256_mul_pd(rinv23,rinv23);
304 rinvsq31 = _mm256_mul_pd(rinv31,rinv31);
305 rinvsq32 = _mm256_mul_pd(rinv32,rinv32);
306 rinvsq33 = _mm256_mul_pd(rinv33,rinv33);
307
308 fjx0 = _mm256_setzero_pd();
309 fjy0 = _mm256_setzero_pd();
310 fjz0 = _mm256_setzero_pd();
311 fjx1 = _mm256_setzero_pd();
312 fjy1 = _mm256_setzero_pd();
313 fjz1 = _mm256_setzero_pd();
314 fjx2 = _mm256_setzero_pd();
315 fjy2 = _mm256_setzero_pd();
316 fjz2 = _mm256_setzero_pd();
317 fjx3 = _mm256_setzero_pd();
318 fjy3 = _mm256_setzero_pd();
319 fjz3 = _mm256_setzero_pd();
320
321 /**************************
322 * CALCULATE INTERACTIONS *
323 **************************/
324
325 if (gmx_mm256_any_lt(rsq00,rcutoff2))
326 {
327
328 /* LENNARD-JONES DISPERSION/REPULSION */
329
330 rinvsix = _mm256_mul_pd(_mm256_mul_pd(rinvsq00,rinvsq00),rinvsq00);
331 vvdw6 = _mm256_mul_pd(c6_00,rinvsix);
332 vvdw12 = _mm256_mul_pd(c12_00,_mm256_mul_pd(rinvsix,rinvsix));
333 vvdw = _mm256_sub_pd(_mm256_mul_pd( _mm256_sub_pd(vvdw12 , _mm256_mul_pd(c12_00,_mm256_mul_pd(sh_vdw_invrcut6,sh_vdw_invrcut6))), one_twelfth) ,
334 _mm256_mul_pd( _mm256_sub_pd(vvdw6,_mm256_mul_pd(c6_00,sh_vdw_invrcut6)),one_sixth));
335 fvdw = _mm256_mul_pd(_mm256_sub_pd(vvdw12,vvdw6),rinvsq00);
336
337 cutoff_mask = _mm256_cmp_pd(rsq00,rcutoff2,_CMP_LT_OQ);
338
339 /* Update potential sum for this i atom from the interaction with this j atom. */
340 vvdw = _mm256_and_pd(vvdw,cutoff_mask);
341 vvdwsum = _mm256_add_pd(vvdwsum,vvdw);
342
343 fscal = fvdw;
344
345 fscal = _mm256_and_pd(fscal,cutoff_mask);
346
347 /* Calculate temporary vectorial force */
348 tx = _mm256_mul_pd(fscal,dx00);
349 ty = _mm256_mul_pd(fscal,dy00);
350 tz = _mm256_mul_pd(fscal,dz00);
351
352 /* Update vectorial force */
353 fix0 = _mm256_add_pd(fix0,tx);
354 fiy0 = _mm256_add_pd(fiy0,ty);
355 fiz0 = _mm256_add_pd(fiz0,tz);
356
357 fjx0 = _mm256_add_pd(fjx0,tx);
358 fjy0 = _mm256_add_pd(fjy0,ty);
359 fjz0 = _mm256_add_pd(fjz0,tz);
360
361 }
362
363 /**************************
364 * CALCULATE INTERACTIONS *
365 **************************/
366
367 if (gmx_mm256_any_lt(rsq11,rcutoff2))
368 {
369
370 /* REACTION-FIELD ELECTROSTATICS */
371 velec = _mm256_mul_pd(qq11,_mm256_sub_pd(_mm256_add_pd(rinv11,_mm256_mul_pd(krf,rsq11)),crf));
372 felec = _mm256_mul_pd(qq11,_mm256_sub_pd(_mm256_mul_pd(rinv11,rinvsq11),krf2));
373
374 cutoff_mask = _mm256_cmp_pd(rsq11,rcutoff2,_CMP_LT_OQ);
375
376 /* Update potential sum for this i atom from the interaction with this j atom. */
377 velec = _mm256_and_pd(velec,cutoff_mask);
378 velecsum = _mm256_add_pd(velecsum,velec);
379
380 fscal = felec;
381
382 fscal = _mm256_and_pd(fscal,cutoff_mask);
383
384 /* Calculate temporary vectorial force */
385 tx = _mm256_mul_pd(fscal,dx11);
386 ty = _mm256_mul_pd(fscal,dy11);
387 tz = _mm256_mul_pd(fscal,dz11);
388
389 /* Update vectorial force */
390 fix1 = _mm256_add_pd(fix1,tx);
391 fiy1 = _mm256_add_pd(fiy1,ty);
392 fiz1 = _mm256_add_pd(fiz1,tz);
393
394 fjx1 = _mm256_add_pd(fjx1,tx);
395 fjy1 = _mm256_add_pd(fjy1,ty);
396 fjz1 = _mm256_add_pd(fjz1,tz);
397
398 }
399
400 /**************************
401 * CALCULATE INTERACTIONS *
402 **************************/
403
404 if (gmx_mm256_any_lt(rsq12,rcutoff2))
405 {
406
407 /* REACTION-FIELD ELECTROSTATICS */
408 velec = _mm256_mul_pd(qq12,_mm256_sub_pd(_mm256_add_pd(rinv12,_mm256_mul_pd(krf,rsq12)),crf));
409 felec = _mm256_mul_pd(qq12,_mm256_sub_pd(_mm256_mul_pd(rinv12,rinvsq12),krf2));
410
411 cutoff_mask = _mm256_cmp_pd(rsq12,rcutoff2,_CMP_LT_OQ);
412
413 /* Update potential sum for this i atom from the interaction with this j atom. */
414 velec = _mm256_and_pd(velec,cutoff_mask);
415 velecsum = _mm256_add_pd(velecsum,velec);
416
417 fscal = felec;
418
419 fscal = _mm256_and_pd(fscal,cutoff_mask);
420
421 /* Calculate temporary vectorial force */
422 tx = _mm256_mul_pd(fscal,dx12);
423 ty = _mm256_mul_pd(fscal,dy12);
424 tz = _mm256_mul_pd(fscal,dz12);
425
426 /* Update vectorial force */
427 fix1 = _mm256_add_pd(fix1,tx);
428 fiy1 = _mm256_add_pd(fiy1,ty);
429 fiz1 = _mm256_add_pd(fiz1,tz);
430
431 fjx2 = _mm256_add_pd(fjx2,tx);
432 fjy2 = _mm256_add_pd(fjy2,ty);
433 fjz2 = _mm256_add_pd(fjz2,tz);
434
435 }
436
437 /**************************
438 * CALCULATE INTERACTIONS *
439 **************************/
440
441 if (gmx_mm256_any_lt(rsq13,rcutoff2))
442 {
443
444 /* REACTION-FIELD ELECTROSTATICS */
445 velec = _mm256_mul_pd(qq13,_mm256_sub_pd(_mm256_add_pd(rinv13,_mm256_mul_pd(krf,rsq13)),crf));
446 felec = _mm256_mul_pd(qq13,_mm256_sub_pd(_mm256_mul_pd(rinv13,rinvsq13),krf2));
447
448 cutoff_mask = _mm256_cmp_pd(rsq13,rcutoff2,_CMP_LT_OQ);
449
450 /* Update potential sum for this i atom from the interaction with this j atom. */
451 velec = _mm256_and_pd(velec,cutoff_mask);
452 velecsum = _mm256_add_pd(velecsum,velec);
453
454 fscal = felec;
455
456 fscal = _mm256_and_pd(fscal,cutoff_mask);
457
458 /* Calculate temporary vectorial force */
459 tx = _mm256_mul_pd(fscal,dx13);
460 ty = _mm256_mul_pd(fscal,dy13);
461 tz = _mm256_mul_pd(fscal,dz13);
462
463 /* Update vectorial force */
464 fix1 = _mm256_add_pd(fix1,tx);
465 fiy1 = _mm256_add_pd(fiy1,ty);
466 fiz1 = _mm256_add_pd(fiz1,tz);
467
468 fjx3 = _mm256_add_pd(fjx3,tx);
469 fjy3 = _mm256_add_pd(fjy3,ty);
470 fjz3 = _mm256_add_pd(fjz3,tz);
471
472 }
473
474 /**************************
475 * CALCULATE INTERACTIONS *
476 **************************/
477
478 if (gmx_mm256_any_lt(rsq21,rcutoff2))
479 {
480
481 /* REACTION-FIELD ELECTROSTATICS */
482 velec = _mm256_mul_pd(qq21,_mm256_sub_pd(_mm256_add_pd(rinv21,_mm256_mul_pd(krf,rsq21)),crf));
483 felec = _mm256_mul_pd(qq21,_mm256_sub_pd(_mm256_mul_pd(rinv21,rinvsq21),krf2));
484
485 cutoff_mask = _mm256_cmp_pd(rsq21,rcutoff2,_CMP_LT_OQ);
486
487 /* Update potential sum for this i atom from the interaction with this j atom. */
488 velec = _mm256_and_pd(velec,cutoff_mask);
489 velecsum = _mm256_add_pd(velecsum,velec);
490
491 fscal = felec;
492
493 fscal = _mm256_and_pd(fscal,cutoff_mask);
494
495 /* Calculate temporary vectorial force */
496 tx = _mm256_mul_pd(fscal,dx21);
497 ty = _mm256_mul_pd(fscal,dy21);
498 tz = _mm256_mul_pd(fscal,dz21);
499
500 /* Update vectorial force */
501 fix2 = _mm256_add_pd(fix2,tx);
502 fiy2 = _mm256_add_pd(fiy2,ty);
503 fiz2 = _mm256_add_pd(fiz2,tz);
504
505 fjx1 = _mm256_add_pd(fjx1,tx);
506 fjy1 = _mm256_add_pd(fjy1,ty);
507 fjz1 = _mm256_add_pd(fjz1,tz);
508
509 }
510
511 /**************************
512 * CALCULATE INTERACTIONS *
513 **************************/
514
515 if (gmx_mm256_any_lt(rsq22,rcutoff2))
516 {
517
518 /* REACTION-FIELD ELECTROSTATICS */
519 velec = _mm256_mul_pd(qq22,_mm256_sub_pd(_mm256_add_pd(rinv22,_mm256_mul_pd(krf,rsq22)),crf));
520 felec = _mm256_mul_pd(qq22,_mm256_sub_pd(_mm256_mul_pd(rinv22,rinvsq22),krf2));
521
522 cutoff_mask = _mm256_cmp_pd(rsq22,rcutoff2,_CMP_LT_OQ);
523
524 /* Update potential sum for this i atom from the interaction with this j atom. */
525 velec = _mm256_and_pd(velec,cutoff_mask);
526 velecsum = _mm256_add_pd(velecsum,velec);
527
528 fscal = felec;
529
530 fscal = _mm256_and_pd(fscal,cutoff_mask);
531
532 /* Calculate temporary vectorial force */
533 tx = _mm256_mul_pd(fscal,dx22);
534 ty = _mm256_mul_pd(fscal,dy22);
535 tz = _mm256_mul_pd(fscal,dz22);
536
537 /* Update vectorial force */
538 fix2 = _mm256_add_pd(fix2,tx);
539 fiy2 = _mm256_add_pd(fiy2,ty);
540 fiz2 = _mm256_add_pd(fiz2,tz);
541
542 fjx2 = _mm256_add_pd(fjx2,tx);
543 fjy2 = _mm256_add_pd(fjy2,ty);
544 fjz2 = _mm256_add_pd(fjz2,tz);
545
546 }
547
548 /**************************
549 * CALCULATE INTERACTIONS *
550 **************************/
551
552 if (gmx_mm256_any_lt(rsq23,rcutoff2))
553 {
554
555 /* REACTION-FIELD ELECTROSTATICS */
556 velec = _mm256_mul_pd(qq23,_mm256_sub_pd(_mm256_add_pd(rinv23,_mm256_mul_pd(krf,rsq23)),crf));
557 felec = _mm256_mul_pd(qq23,_mm256_sub_pd(_mm256_mul_pd(rinv23,rinvsq23),krf2));
558
559 cutoff_mask = _mm256_cmp_pd(rsq23,rcutoff2,_CMP_LT_OQ);
560
561 /* Update potential sum for this i atom from the interaction with this j atom. */
562 velec = _mm256_and_pd(velec,cutoff_mask);
563 velecsum = _mm256_add_pd(velecsum,velec);
564
565 fscal = felec;
566
567 fscal = _mm256_and_pd(fscal,cutoff_mask);
568
569 /* Calculate temporary vectorial force */
570 tx = _mm256_mul_pd(fscal,dx23);
571 ty = _mm256_mul_pd(fscal,dy23);
572 tz = _mm256_mul_pd(fscal,dz23);
573
574 /* Update vectorial force */
575 fix2 = _mm256_add_pd(fix2,tx);
576 fiy2 = _mm256_add_pd(fiy2,ty);
577 fiz2 = _mm256_add_pd(fiz2,tz);
578
579 fjx3 = _mm256_add_pd(fjx3,tx);
580 fjy3 = _mm256_add_pd(fjy3,ty);
581 fjz3 = _mm256_add_pd(fjz3,tz);
582
583 }
584
585 /**************************
586 * CALCULATE INTERACTIONS *
587 **************************/
588
589 if (gmx_mm256_any_lt(rsq31,rcutoff2))
590 {
591
592 /* REACTION-FIELD ELECTROSTATICS */
593 velec = _mm256_mul_pd(qq31,_mm256_sub_pd(_mm256_add_pd(rinv31,_mm256_mul_pd(krf,rsq31)),crf));
594 felec = _mm256_mul_pd(qq31,_mm256_sub_pd(_mm256_mul_pd(rinv31,rinvsq31),krf2));
595
596 cutoff_mask = _mm256_cmp_pd(rsq31,rcutoff2,_CMP_LT_OQ);
597
598 /* Update potential sum for this i atom from the interaction with this j atom. */
599 velec = _mm256_and_pd(velec,cutoff_mask);
600 velecsum = _mm256_add_pd(velecsum,velec);
601
602 fscal = felec;
603
604 fscal = _mm256_and_pd(fscal,cutoff_mask);
605
606 /* Calculate temporary vectorial force */
607 tx = _mm256_mul_pd(fscal,dx31);
608 ty = _mm256_mul_pd(fscal,dy31);
609 tz = _mm256_mul_pd(fscal,dz31);
610
611 /* Update vectorial force */
612 fix3 = _mm256_add_pd(fix3,tx);
613 fiy3 = _mm256_add_pd(fiy3,ty);
614 fiz3 = _mm256_add_pd(fiz3,tz);
615
616 fjx1 = _mm256_add_pd(fjx1,tx);
617 fjy1 = _mm256_add_pd(fjy1,ty);
618 fjz1 = _mm256_add_pd(fjz1,tz);
619
620 }
621
622 /**************************
623 * CALCULATE INTERACTIONS *
624 **************************/
625
626 if (gmx_mm256_any_lt(rsq32,rcutoff2))
627 {
628
629 /* REACTION-FIELD ELECTROSTATICS */
630 velec = _mm256_mul_pd(qq32,_mm256_sub_pd(_mm256_add_pd(rinv32,_mm256_mul_pd(krf,rsq32)),crf));
631 felec = _mm256_mul_pd(qq32,_mm256_sub_pd(_mm256_mul_pd(rinv32,rinvsq32),krf2));
632
633 cutoff_mask = _mm256_cmp_pd(rsq32,rcutoff2,_CMP_LT_OQ);
634
635 /* Update potential sum for this i atom from the interaction with this j atom. */
636 velec = _mm256_and_pd(velec,cutoff_mask);
637 velecsum = _mm256_add_pd(velecsum,velec);
638
639 fscal = felec;
640
641 fscal = _mm256_and_pd(fscal,cutoff_mask);
642
643 /* Calculate temporary vectorial force */
644 tx = _mm256_mul_pd(fscal,dx32);
645 ty = _mm256_mul_pd(fscal,dy32);
646 tz = _mm256_mul_pd(fscal,dz32);
647
648 /* Update vectorial force */
649 fix3 = _mm256_add_pd(fix3,tx);
650 fiy3 = _mm256_add_pd(fiy3,ty);
651 fiz3 = _mm256_add_pd(fiz3,tz);
652
653 fjx2 = _mm256_add_pd(fjx2,tx);
654 fjy2 = _mm256_add_pd(fjy2,ty);
655 fjz2 = _mm256_add_pd(fjz2,tz);
656
657 }
658
659 /**************************
660 * CALCULATE INTERACTIONS *
661 **************************/
662
663 if (gmx_mm256_any_lt(rsq33,rcutoff2))
664 {
665
666 /* REACTION-FIELD ELECTROSTATICS */
667 velec = _mm256_mul_pd(qq33,_mm256_sub_pd(_mm256_add_pd(rinv33,_mm256_mul_pd(krf,rsq33)),crf));
668 felec = _mm256_mul_pd(qq33,_mm256_sub_pd(_mm256_mul_pd(rinv33,rinvsq33),krf2));
669
670 cutoff_mask = _mm256_cmp_pd(rsq33,rcutoff2,_CMP_LT_OQ);
671
672 /* Update potential sum for this i atom from the interaction with this j atom. */
673 velec = _mm256_and_pd(velec,cutoff_mask);
674 velecsum = _mm256_add_pd(velecsum,velec);
675
676 fscal = felec;
677
678 fscal = _mm256_and_pd(fscal,cutoff_mask);
679
680 /* Calculate temporary vectorial force */
681 tx = _mm256_mul_pd(fscal,dx33);
682 ty = _mm256_mul_pd(fscal,dy33);
683 tz = _mm256_mul_pd(fscal,dz33);
684
685 /* Update vectorial force */
686 fix3 = _mm256_add_pd(fix3,tx);
687 fiy3 = _mm256_add_pd(fiy3,ty);
688 fiz3 = _mm256_add_pd(fiz3,tz);
689
690 fjx3 = _mm256_add_pd(fjx3,tx);
691 fjy3 = _mm256_add_pd(fjy3,ty);
692 fjz3 = _mm256_add_pd(fjz3,tz);
693
694 }
695
696 fjptrA = f+j_coord_offsetA;
697 fjptrB = f+j_coord_offsetB;
698 fjptrC = f+j_coord_offsetC;
699 fjptrD = f+j_coord_offsetD;
700
701 gmx_mm256_decrement_4rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,
702 fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,
703 fjx2,fjy2,fjz2,fjx3,fjy3,fjz3);
704
705 /* Inner loop uses 368 flops */
706 }
707
708 if(jidx<j_index_end)
709 {
710
711 /* Get j neighbor index, and coordinate index */
712 jnrlistA = jjnr[jidx];
713 jnrlistB = jjnr[jidx+1];
714 jnrlistC = jjnr[jidx+2];
715 jnrlistD = jjnr[jidx+3];
716 /* Sign of each element will be negative for non-real atoms.
717 * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
718 * so use it as val = _mm_andnot_pd(mask,val) to clear dummy entries.
719 */
720 tmpmask0 = gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128()));
721
722 tmpmask1 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(3,3,2,2));
723 tmpmask0 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(1,1,0,0));
724 dummy_mask = _mm256_castps_pd(gmx_mm256_set_m128(tmpmask1,tmpmask0));
725
726 jnrA = (jnrlistA>=0) ? jnrlistA : 0;
727 jnrB = (jnrlistB>=0) ? jnrlistB : 0;
728 jnrC = (jnrlistC>=0) ? jnrlistC : 0;
729 jnrD = (jnrlistD>=0) ? jnrlistD : 0;
730 j_coord_offsetA = DIM*jnrA;
731 j_coord_offsetB = DIM*jnrB;
732 j_coord_offsetC = DIM*jnrC;
733 j_coord_offsetD = DIM*jnrD;
734
735 /* load j atom coordinates */
736 gmx_mm256_load_4rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
737 x+j_coord_offsetC,x+j_coord_offsetD,
738 &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,
739 &jy2,&jz2,&jx3,&jy3,&jz3);
740
741 /* Calculate displacement vector */
742 dx00 = _mm256_sub_pd(ix0,jx0);
743 dy00 = _mm256_sub_pd(iy0,jy0);
744 dz00 = _mm256_sub_pd(iz0,jz0);
745 dx11 = _mm256_sub_pd(ix1,jx1);
746 dy11 = _mm256_sub_pd(iy1,jy1);
747 dz11 = _mm256_sub_pd(iz1,jz1);
748 dx12 = _mm256_sub_pd(ix1,jx2);
749 dy12 = _mm256_sub_pd(iy1,jy2);
750 dz12 = _mm256_sub_pd(iz1,jz2);
751 dx13 = _mm256_sub_pd(ix1,jx3);
752 dy13 = _mm256_sub_pd(iy1,jy3);
753 dz13 = _mm256_sub_pd(iz1,jz3);
754 dx21 = _mm256_sub_pd(ix2,jx1);
755 dy21 = _mm256_sub_pd(iy2,jy1);
756 dz21 = _mm256_sub_pd(iz2,jz1);
757 dx22 = _mm256_sub_pd(ix2,jx2);
758 dy22 = _mm256_sub_pd(iy2,jy2);
759 dz22 = _mm256_sub_pd(iz2,jz2);
760 dx23 = _mm256_sub_pd(ix2,jx3);
761 dy23 = _mm256_sub_pd(iy2,jy3);
762 dz23 = _mm256_sub_pd(iz2,jz3);
763 dx31 = _mm256_sub_pd(ix3,jx1);
764 dy31 = _mm256_sub_pd(iy3,jy1);
765 dz31 = _mm256_sub_pd(iz3,jz1);
766 dx32 = _mm256_sub_pd(ix3,jx2);
767 dy32 = _mm256_sub_pd(iy3,jy2);
768 dz32 = _mm256_sub_pd(iz3,jz2);
769 dx33 = _mm256_sub_pd(ix3,jx3);
770 dy33 = _mm256_sub_pd(iy3,jy3);
771 dz33 = _mm256_sub_pd(iz3,jz3);
772
773 /* Calculate squared distance and things based on it */
774 rsq00 = gmx_mm256_calc_rsq_pd(dx00,dy00,dz00);
775 rsq11 = gmx_mm256_calc_rsq_pd(dx11,dy11,dz11);
776 rsq12 = gmx_mm256_calc_rsq_pd(dx12,dy12,dz12);
777 rsq13 = gmx_mm256_calc_rsq_pd(dx13,dy13,dz13);
778 rsq21 = gmx_mm256_calc_rsq_pd(dx21,dy21,dz21);
779 rsq22 = gmx_mm256_calc_rsq_pd(dx22,dy22,dz22);
780 rsq23 = gmx_mm256_calc_rsq_pd(dx23,dy23,dz23);
781 rsq31 = gmx_mm256_calc_rsq_pd(dx31,dy31,dz31);
782 rsq32 = gmx_mm256_calc_rsq_pd(dx32,dy32,dz32);
783 rsq33 = gmx_mm256_calc_rsq_pd(dx33,dy33,dz33);
784
785 rinv11 = avx256_invsqrt_d(rsq11);
786 rinv12 = avx256_invsqrt_d(rsq12);
787 rinv13 = avx256_invsqrt_d(rsq13);
788 rinv21 = avx256_invsqrt_d(rsq21);
789 rinv22 = avx256_invsqrt_d(rsq22);
790 rinv23 = avx256_invsqrt_d(rsq23);
791 rinv31 = avx256_invsqrt_d(rsq31);
792 rinv32 = avx256_invsqrt_d(rsq32);
793 rinv33 = avx256_invsqrt_d(rsq33);
794
795 rinvsq00 = avx256_inv_d(rsq00);
796 rinvsq11 = _mm256_mul_pd(rinv11,rinv11);
797 rinvsq12 = _mm256_mul_pd(rinv12,rinv12);
798 rinvsq13 = _mm256_mul_pd(rinv13,rinv13);
799 rinvsq21 = _mm256_mul_pd(rinv21,rinv21);
800 rinvsq22 = _mm256_mul_pd(rinv22,rinv22);
801 rinvsq23 = _mm256_mul_pd(rinv23,rinv23);
802 rinvsq31 = _mm256_mul_pd(rinv31,rinv31);
803 rinvsq32 = _mm256_mul_pd(rinv32,rinv32);
804 rinvsq33 = _mm256_mul_pd(rinv33,rinv33);
805
806 fjx0 = _mm256_setzero_pd();
807 fjy0 = _mm256_setzero_pd();
808 fjz0 = _mm256_setzero_pd();
809 fjx1 = _mm256_setzero_pd();
810 fjy1 = _mm256_setzero_pd();
811 fjz1 = _mm256_setzero_pd();
812 fjx2 = _mm256_setzero_pd();
813 fjy2 = _mm256_setzero_pd();
814 fjz2 = _mm256_setzero_pd();
815 fjx3 = _mm256_setzero_pd();
816 fjy3 = _mm256_setzero_pd();
817 fjz3 = _mm256_setzero_pd();
818
819 /**************************
820 * CALCULATE INTERACTIONS *
821 **************************/
822
823 if (gmx_mm256_any_lt(rsq00,rcutoff2))
824 {
825
826 /* LENNARD-JONES DISPERSION/REPULSION */
827
828 rinvsix = _mm256_mul_pd(_mm256_mul_pd(rinvsq00,rinvsq00),rinvsq00);
829 vvdw6 = _mm256_mul_pd(c6_00,rinvsix);
830 vvdw12 = _mm256_mul_pd(c12_00,_mm256_mul_pd(rinvsix,rinvsix));
831 vvdw = _mm256_sub_pd(_mm256_mul_pd( _mm256_sub_pd(vvdw12 , _mm256_mul_pd(c12_00,_mm256_mul_pd(sh_vdw_invrcut6,sh_vdw_invrcut6))), one_twelfth) ,
832 _mm256_mul_pd( _mm256_sub_pd(vvdw6,_mm256_mul_pd(c6_00,sh_vdw_invrcut6)),one_sixth));
833 fvdw = _mm256_mul_pd(_mm256_sub_pd(vvdw12,vvdw6),rinvsq00);
834
835 cutoff_mask = _mm256_cmp_pd(rsq00,rcutoff2,_CMP_LT_OQ);
836
837 /* Update potential sum for this i atom from the interaction with this j atom. */
838 vvdw = _mm256_and_pd(vvdw,cutoff_mask);
839 vvdw = _mm256_andnot_pd(dummy_mask,vvdw);
840 vvdwsum = _mm256_add_pd(vvdwsum,vvdw);
841
842 fscal = fvdw;
843
844 fscal = _mm256_and_pd(fscal,cutoff_mask);
845
846 fscal = _mm256_andnot_pd(dummy_mask,fscal);
847
848 /* Calculate temporary vectorial force */
849 tx = _mm256_mul_pd(fscal,dx00);
850 ty = _mm256_mul_pd(fscal,dy00);
851 tz = _mm256_mul_pd(fscal,dz00);
852
853 /* Update vectorial force */
854 fix0 = _mm256_add_pd(fix0,tx);
855 fiy0 = _mm256_add_pd(fiy0,ty);
856 fiz0 = _mm256_add_pd(fiz0,tz);
857
858 fjx0 = _mm256_add_pd(fjx0,tx);
859 fjy0 = _mm256_add_pd(fjy0,ty);
860 fjz0 = _mm256_add_pd(fjz0,tz);
861
862 }
863
864 /**************************
865 * CALCULATE INTERACTIONS *
866 **************************/
867
868 if (gmx_mm256_any_lt(rsq11,rcutoff2))
869 {
870
871 /* REACTION-FIELD ELECTROSTATICS */
872 velec = _mm256_mul_pd(qq11,_mm256_sub_pd(_mm256_add_pd(rinv11,_mm256_mul_pd(krf,rsq11)),crf));
873 felec = _mm256_mul_pd(qq11,_mm256_sub_pd(_mm256_mul_pd(rinv11,rinvsq11),krf2));
874
875 cutoff_mask = _mm256_cmp_pd(rsq11,rcutoff2,_CMP_LT_OQ);
876
877 /* Update potential sum for this i atom from the interaction with this j atom. */
878 velec = _mm256_and_pd(velec,cutoff_mask);
879 velec = _mm256_andnot_pd(dummy_mask,velec);
880 velecsum = _mm256_add_pd(velecsum,velec);
881
882 fscal = felec;
883
884 fscal = _mm256_and_pd(fscal,cutoff_mask);
885
886 fscal = _mm256_andnot_pd(dummy_mask,fscal);
887
888 /* Calculate temporary vectorial force */
889 tx = _mm256_mul_pd(fscal,dx11);
890 ty = _mm256_mul_pd(fscal,dy11);
891 tz = _mm256_mul_pd(fscal,dz11);
892
893 /* Update vectorial force */
894 fix1 = _mm256_add_pd(fix1,tx);
895 fiy1 = _mm256_add_pd(fiy1,ty);
896 fiz1 = _mm256_add_pd(fiz1,tz);
897
898 fjx1 = _mm256_add_pd(fjx1,tx);
899 fjy1 = _mm256_add_pd(fjy1,ty);
900 fjz1 = _mm256_add_pd(fjz1,tz);
901
902 }
903
904 /**************************
905 * CALCULATE INTERACTIONS *
906 **************************/
907
908 if (gmx_mm256_any_lt(rsq12,rcutoff2))
909 {
910
911 /* REACTION-FIELD ELECTROSTATICS */
912 velec = _mm256_mul_pd(qq12,_mm256_sub_pd(_mm256_add_pd(rinv12,_mm256_mul_pd(krf,rsq12)),crf));
913 felec = _mm256_mul_pd(qq12,_mm256_sub_pd(_mm256_mul_pd(rinv12,rinvsq12),krf2));
914
915 cutoff_mask = _mm256_cmp_pd(rsq12,rcutoff2,_CMP_LT_OQ);
916
917 /* Update potential sum for this i atom from the interaction with this j atom. */
918 velec = _mm256_and_pd(velec,cutoff_mask);
919 velec = _mm256_andnot_pd(dummy_mask,velec);
920 velecsum = _mm256_add_pd(velecsum,velec);
921
922 fscal = felec;
923
924 fscal = _mm256_and_pd(fscal,cutoff_mask);
925
926 fscal = _mm256_andnot_pd(dummy_mask,fscal);
927
928 /* Calculate temporary vectorial force */
929 tx = _mm256_mul_pd(fscal,dx12);
930 ty = _mm256_mul_pd(fscal,dy12);
931 tz = _mm256_mul_pd(fscal,dz12);
932
933 /* Update vectorial force */
934 fix1 = _mm256_add_pd(fix1,tx);
935 fiy1 = _mm256_add_pd(fiy1,ty);
936 fiz1 = _mm256_add_pd(fiz1,tz);
937
938 fjx2 = _mm256_add_pd(fjx2,tx);
939 fjy2 = _mm256_add_pd(fjy2,ty);
940 fjz2 = _mm256_add_pd(fjz2,tz);
941
942 }
943
944 /**************************
945 * CALCULATE INTERACTIONS *
946 **************************/
947
948 if (gmx_mm256_any_lt(rsq13,rcutoff2))
949 {
950
951 /* REACTION-FIELD ELECTROSTATICS */
952 velec = _mm256_mul_pd(qq13,_mm256_sub_pd(_mm256_add_pd(rinv13,_mm256_mul_pd(krf,rsq13)),crf));
953 felec = _mm256_mul_pd(qq13,_mm256_sub_pd(_mm256_mul_pd(rinv13,rinvsq13),krf2));
954
955 cutoff_mask = _mm256_cmp_pd(rsq13,rcutoff2,_CMP_LT_OQ);
956
957 /* Update potential sum for this i atom from the interaction with this j atom. */
958 velec = _mm256_and_pd(velec,cutoff_mask);
959 velec = _mm256_andnot_pd(dummy_mask,velec);
960 velecsum = _mm256_add_pd(velecsum,velec);
961
962 fscal = felec;
963
964 fscal = _mm256_and_pd(fscal,cutoff_mask);
965
966 fscal = _mm256_andnot_pd(dummy_mask,fscal);
967
968 /* Calculate temporary vectorial force */
969 tx = _mm256_mul_pd(fscal,dx13);
970 ty = _mm256_mul_pd(fscal,dy13);
971 tz = _mm256_mul_pd(fscal,dz13);
972
973 /* Update vectorial force */
974 fix1 = _mm256_add_pd(fix1,tx);
975 fiy1 = _mm256_add_pd(fiy1,ty);
976 fiz1 = _mm256_add_pd(fiz1,tz);
977
978 fjx3 = _mm256_add_pd(fjx3,tx);
979 fjy3 = _mm256_add_pd(fjy3,ty);
980 fjz3 = _mm256_add_pd(fjz3,tz);
981
982 }
983
984 /**************************
985 * CALCULATE INTERACTIONS *
986 **************************/
987
988 if (gmx_mm256_any_lt(rsq21,rcutoff2))
989 {
990
991 /* REACTION-FIELD ELECTROSTATICS */
992 velec = _mm256_mul_pd(qq21,_mm256_sub_pd(_mm256_add_pd(rinv21,_mm256_mul_pd(krf,rsq21)),crf));
993 felec = _mm256_mul_pd(qq21,_mm256_sub_pd(_mm256_mul_pd(rinv21,rinvsq21),krf2));
994
995 cutoff_mask = _mm256_cmp_pd(rsq21,rcutoff2,_CMP_LT_OQ);
996
997 /* Update potential sum for this i atom from the interaction with this j atom. */
998 velec = _mm256_and_pd(velec,cutoff_mask);
999 velec = _mm256_andnot_pd(dummy_mask,velec);
1000 velecsum = _mm256_add_pd(velecsum,velec);
1001
1002 fscal = felec;
1003
1004 fscal = _mm256_and_pd(fscal,cutoff_mask);
1005
1006 fscal = _mm256_andnot_pd(dummy_mask,fscal);
1007
1008 /* Calculate temporary vectorial force */
1009 tx = _mm256_mul_pd(fscal,dx21);
1010 ty = _mm256_mul_pd(fscal,dy21);
1011 tz = _mm256_mul_pd(fscal,dz21);
1012
1013 /* Update vectorial force */
1014 fix2 = _mm256_add_pd(fix2,tx);
1015 fiy2 = _mm256_add_pd(fiy2,ty);
1016 fiz2 = _mm256_add_pd(fiz2,tz);
1017
1018 fjx1 = _mm256_add_pd(fjx1,tx);
1019 fjy1 = _mm256_add_pd(fjy1,ty);
1020 fjz1 = _mm256_add_pd(fjz1,tz);
1021
1022 }
1023
1024 /**************************
1025 * CALCULATE INTERACTIONS *
1026 **************************/
1027
1028 if (gmx_mm256_any_lt(rsq22,rcutoff2))
1029 {
1030
1031 /* REACTION-FIELD ELECTROSTATICS */
1032 velec = _mm256_mul_pd(qq22,_mm256_sub_pd(_mm256_add_pd(rinv22,_mm256_mul_pd(krf,rsq22)),crf));
1033 felec = _mm256_mul_pd(qq22,_mm256_sub_pd(_mm256_mul_pd(rinv22,rinvsq22),krf2));
1034
1035 cutoff_mask = _mm256_cmp_pd(rsq22,rcutoff2,_CMP_LT_OQ);
1036
1037 /* Update potential sum for this i atom from the interaction with this j atom. */
1038 velec = _mm256_and_pd(velec,cutoff_mask);
1039 velec = _mm256_andnot_pd(dummy_mask,velec);
1040 velecsum = _mm256_add_pd(velecsum,velec);
1041
1042 fscal = felec;
1043
1044 fscal = _mm256_and_pd(fscal,cutoff_mask);
1045
1046 fscal = _mm256_andnot_pd(dummy_mask,fscal);
1047
1048 /* Calculate temporary vectorial force */
1049 tx = _mm256_mul_pd(fscal,dx22);
1050 ty = _mm256_mul_pd(fscal,dy22);
1051 tz = _mm256_mul_pd(fscal,dz22);
1052
1053 /* Update vectorial force */
1054 fix2 = _mm256_add_pd(fix2,tx);
1055 fiy2 = _mm256_add_pd(fiy2,ty);
1056 fiz2 = _mm256_add_pd(fiz2,tz);
1057
1058 fjx2 = _mm256_add_pd(fjx2,tx);
1059 fjy2 = _mm256_add_pd(fjy2,ty);
1060 fjz2 = _mm256_add_pd(fjz2,tz);
1061
1062 }
1063
1064 /**************************
1065 * CALCULATE INTERACTIONS *
1066 **************************/
1067
1068 if (gmx_mm256_any_lt(rsq23,rcutoff2))
1069 {
1070
1071 /* REACTION-FIELD ELECTROSTATICS */
1072 velec = _mm256_mul_pd(qq23,_mm256_sub_pd(_mm256_add_pd(rinv23,_mm256_mul_pd(krf,rsq23)),crf));
1073 felec = _mm256_mul_pd(qq23,_mm256_sub_pd(_mm256_mul_pd(rinv23,rinvsq23),krf2));
1074
1075 cutoff_mask = _mm256_cmp_pd(rsq23,rcutoff2,_CMP_LT_OQ);
1076
1077 /* Update potential sum for this i atom from the interaction with this j atom. */
1078 velec = _mm256_and_pd(velec,cutoff_mask);
1079 velec = _mm256_andnot_pd(dummy_mask,velec);
1080 velecsum = _mm256_add_pd(velecsum,velec);
1081
1082 fscal = felec;
1083
1084 fscal = _mm256_and_pd(fscal,cutoff_mask);
1085
1086 fscal = _mm256_andnot_pd(dummy_mask,fscal);
1087
1088 /* Calculate temporary vectorial force */
1089 tx = _mm256_mul_pd(fscal,dx23);
1090 ty = _mm256_mul_pd(fscal,dy23);
1091 tz = _mm256_mul_pd(fscal,dz23);
1092
1093 /* Update vectorial force */
1094 fix2 = _mm256_add_pd(fix2,tx);
1095 fiy2 = _mm256_add_pd(fiy2,ty);
1096 fiz2 = _mm256_add_pd(fiz2,tz);
1097
1098 fjx3 = _mm256_add_pd(fjx3,tx);
1099 fjy3 = _mm256_add_pd(fjy3,ty);
1100 fjz3 = _mm256_add_pd(fjz3,tz);
1101
1102 }
1103
1104 /**************************
1105 * CALCULATE INTERACTIONS *
1106 **************************/
1107
1108 if (gmx_mm256_any_lt(rsq31,rcutoff2))
1109 {
1110
1111 /* REACTION-FIELD ELECTROSTATICS */
1112 velec = _mm256_mul_pd(qq31,_mm256_sub_pd(_mm256_add_pd(rinv31,_mm256_mul_pd(krf,rsq31)),crf));
1113 felec = _mm256_mul_pd(qq31,_mm256_sub_pd(_mm256_mul_pd(rinv31,rinvsq31),krf2));
1114
1115 cutoff_mask = _mm256_cmp_pd(rsq31,rcutoff2,_CMP_LT_OQ);
1116
1117 /* Update potential sum for this i atom from the interaction with this j atom. */
1118 velec = _mm256_and_pd(velec,cutoff_mask);
1119 velec = _mm256_andnot_pd(dummy_mask,velec);
1120 velecsum = _mm256_add_pd(velecsum,velec);
1121
1122 fscal = felec;
1123
1124 fscal = _mm256_and_pd(fscal,cutoff_mask);
1125
1126 fscal = _mm256_andnot_pd(dummy_mask,fscal);
1127
1128 /* Calculate temporary vectorial force */
1129 tx = _mm256_mul_pd(fscal,dx31);
1130 ty = _mm256_mul_pd(fscal,dy31);
1131 tz = _mm256_mul_pd(fscal,dz31);
1132
1133 /* Update vectorial force */
1134 fix3 = _mm256_add_pd(fix3,tx);
1135 fiy3 = _mm256_add_pd(fiy3,ty);
1136 fiz3 = _mm256_add_pd(fiz3,tz);
1137
1138 fjx1 = _mm256_add_pd(fjx1,tx);
1139 fjy1 = _mm256_add_pd(fjy1,ty);
1140 fjz1 = _mm256_add_pd(fjz1,tz);
1141
1142 }
1143
1144 /**************************
1145 * CALCULATE INTERACTIONS *
1146 **************************/
1147
1148 if (gmx_mm256_any_lt(rsq32,rcutoff2))
1149 {
1150
1151 /* REACTION-FIELD ELECTROSTATICS */
1152 velec = _mm256_mul_pd(qq32,_mm256_sub_pd(_mm256_add_pd(rinv32,_mm256_mul_pd(krf,rsq32)),crf));
1153 felec = _mm256_mul_pd(qq32,_mm256_sub_pd(_mm256_mul_pd(rinv32,rinvsq32),krf2));
1154
1155 cutoff_mask = _mm256_cmp_pd(rsq32,rcutoff2,_CMP_LT_OQ);
1156
1157 /* Update potential sum for this i atom from the interaction with this j atom. */
1158 velec = _mm256_and_pd(velec,cutoff_mask);
1159 velec = _mm256_andnot_pd(dummy_mask,velec);
1160 velecsum = _mm256_add_pd(velecsum,velec);
1161
1162 fscal = felec;
1163
1164 fscal = _mm256_and_pd(fscal,cutoff_mask);
1165
1166 fscal = _mm256_andnot_pd(dummy_mask,fscal);
1167
1168 /* Calculate temporary vectorial force */
1169 tx = _mm256_mul_pd(fscal,dx32);
1170 ty = _mm256_mul_pd(fscal,dy32);
1171 tz = _mm256_mul_pd(fscal,dz32);
1172
1173 /* Update vectorial force */
1174 fix3 = _mm256_add_pd(fix3,tx);
1175 fiy3 = _mm256_add_pd(fiy3,ty);
1176 fiz3 = _mm256_add_pd(fiz3,tz);
1177
1178 fjx2 = _mm256_add_pd(fjx2,tx);
1179 fjy2 = _mm256_add_pd(fjy2,ty);
1180 fjz2 = _mm256_add_pd(fjz2,tz);
1181
1182 }
1183
1184 /**************************
1185 * CALCULATE INTERACTIONS *
1186 **************************/
1187
1188 if (gmx_mm256_any_lt(rsq33,rcutoff2))
1189 {
1190
1191 /* REACTION-FIELD ELECTROSTATICS */
1192 velec = _mm256_mul_pd(qq33,_mm256_sub_pd(_mm256_add_pd(rinv33,_mm256_mul_pd(krf,rsq33)),crf));
1193 felec = _mm256_mul_pd(qq33,_mm256_sub_pd(_mm256_mul_pd(rinv33,rinvsq33),krf2));
1194
1195 cutoff_mask = _mm256_cmp_pd(rsq33,rcutoff2,_CMP_LT_OQ);
1196
1197 /* Update potential sum for this i atom from the interaction with this j atom. */
1198 velec = _mm256_and_pd(velec,cutoff_mask);
1199 velec = _mm256_andnot_pd(dummy_mask,velec);
1200 velecsum = _mm256_add_pd(velecsum,velec);
1201
1202 fscal = felec;
1203
1204 fscal = _mm256_and_pd(fscal,cutoff_mask);
1205
1206 fscal = _mm256_andnot_pd(dummy_mask,fscal);
1207
1208 /* Calculate temporary vectorial force */
1209 tx = _mm256_mul_pd(fscal,dx33);
1210 ty = _mm256_mul_pd(fscal,dy33);
1211 tz = _mm256_mul_pd(fscal,dz33);
1212
1213 /* Update vectorial force */
1214 fix3 = _mm256_add_pd(fix3,tx);
1215 fiy3 = _mm256_add_pd(fiy3,ty);
1216 fiz3 = _mm256_add_pd(fiz3,tz);
1217
1218 fjx3 = _mm256_add_pd(fjx3,tx);
1219 fjy3 = _mm256_add_pd(fjy3,ty);
1220 fjz3 = _mm256_add_pd(fjz3,tz);
1221
1222 }
1223
1224 fjptrA = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
1225 fjptrB = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
1226 fjptrC = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
1227 fjptrD = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
1228
1229 gmx_mm256_decrement_4rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,
1230 fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,
1231 fjx2,fjy2,fjz2,fjx3,fjy3,fjz3);
1232
1233 /* Inner loop uses 368 flops */
1234 }
1235
1236 /* End of innermost loop */
1237
1238 gmx_mm256_update_iforce_4atom_swizzle_pd(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
1239 f+i_coord_offset,fshift+i_shift_offset);
1240
1241 ggid = gid[iidx];
1242 /* Update potential energies */
1243 gmx_mm256_update_1pot_pd(velecsum,kernel_data->energygrp_elec+ggid);
1244 gmx_mm256_update_1pot_pd(vvdwsum,kernel_data->energygrp_vdw+ggid);
1245
1246 /* Increment number of inner iterations */
1247 inneriter += j_index_end - j_index_start;
1248
1249 /* Outer loop uses 26 flops */
1250 }
1251
1252 /* Increment number of outer iterations */
1253 outeriter += nri;
1254
1255 /* Update outer/inner flops */
1256
1257 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4W4_VF,outeriter*26 + inneriter*368);
1258 }
1259 /*
1260 * Gromacs nonbonded kernel: nb_kernel_ElecRFCut_VdwLJSh_GeomW4W4_F_avx_256_double
1261 * Electrostatics interaction: ReactionField
1262 * VdW interaction: LennardJones
1263 * Geometry: Water4-Water4
1264 * Calculate force/pot: Force
1265 */
1266 void
1267 nb_kernel_ElecRFCut_VdwLJSh_GeomW4W4_F_avx_256_double
1268 (t_nblist * gmx_restrict nlist,
1269 rvec * gmx_restrict xx,
1270 rvec * gmx_restrict ff,
1271 struct t_forcerec * gmx_restrict fr,
1272 t_mdatoms * gmx_restrict mdatoms,
1273 nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
1274 t_nrnb * gmx_restrict nrnb)
1275 {
1276 /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
1277 * just 0 for non-waters.
1278 * Suffixes A,B,C,D refer to j loop unrolling done with AVX, e.g. for the four different
1279 * jnr indices corresponding to data put in the four positions in the SIMD register.
1280 */
1281 int i_shift_offset,i_coord_offset,outeriter,inneriter;
1282 int j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
1283 int jnrA,jnrB,jnrC,jnrD;
1284 int jnrlistA,jnrlistB,jnrlistC,jnrlistD;
1285 int jnrlistE,jnrlistF,jnrlistG,jnrlistH;
1286 int j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
1287 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
1288 real rcutoff_scalar;
1289 real *shiftvec,*fshift,*x,*f;
1290 real *fjptrA,*fjptrB,*fjptrC,*fjptrD;
1291 real scratch[4*DIM];
1292 __m256d tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
1293 real * vdwioffsetptr0;
1294 __m256d ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
1295 real * vdwioffsetptr1;
1296 __m256d ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
1297 real * vdwioffsetptr2;
1298 __m256d ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
1299 real * vdwioffsetptr3;
1300 __m256d ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
1301 int vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D;
1302 __m256d jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
1303 int vdwjidx1A,vdwjidx1B,vdwjidx1C,vdwjidx1D;
1304 __m256d jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1;
1305 int vdwjidx2A,vdwjidx2B,vdwjidx2C,vdwjidx2D;
1306 __m256d jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2;
1307 int vdwjidx3A,vdwjidx3B,vdwjidx3C,vdwjidx3D;
1308 __m256d jx3,jy3,jz3,fjx3,fjy3,fjz3,jq3,isaj3;
1309 __m256d dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
1310 __m256d dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11;
1311 __m256d dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12;
1312 __m256d dx13,dy13,dz13,rsq13,rinv13,rinvsq13,r13,qq13,c6_13,c12_13;
1313 __m256d dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21;
1314 __m256d dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22;
1315 __m256d dx23,dy23,dz23,rsq23,rinv23,rinvsq23,r23,qq23,c6_23,c12_23;
1316 __m256d dx31,dy31,dz31,rsq31,rinv31,rinvsq31,r31,qq31,c6_31,c12_31;
1317 __m256d dx32,dy32,dz32,rsq32,rinv32,rinvsq32,r32,qq32,c6_32,c12_32;
1318 __m256d dx33,dy33,dz33,rsq33,rinv33,rinvsq33,r33,qq33,c6_33,c12_33;
1319 __m256d velec,felec,velecsum,facel,crf,krf,krf2;
1320 real *charge;
1321 int nvdwtype;
1322 __m256d rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
1323 int *vdwtype;
1324 real *vdwparam;
1325 __m256d one_sixth = _mm256_set1_pd(1.0/6.0);
1326 __m256d one_twelfth = _mm256_set1_pd(1.0/12.0);
1327 __m256d dummy_mask,cutoff_mask;
1328 __m128 tmpmask0,tmpmask1;
1329 __m256d signbit = _mm256_castsi256_pd( _mm256_set1_epi32(0x80000000) );
1330 __m256d one = _mm256_set1_pd(1.0);
1331 __m256d two = _mm256_set1_pd(2.0);
1332 x = xx[0];
1333 f = ff[0];
1334
1335 nri = nlist->nri;
1336 iinr = nlist->iinr;
1337 jindex = nlist->jindex;
1338 jjnr = nlist->jjnr;
1339 shiftidx = nlist->shift;
1340 gid = nlist->gid;
1341 shiftvec = fr->shift_vec[0];
1342 fshift = fr->fshift[0];
1343 facel = _mm256_set1_pd(fr->ic->epsfac);
1344 charge = mdatoms->chargeA;
1345 krf = _mm256_set1_pd(fr->ic->k_rf);
1346 krf2 = _mm256_set1_pd(fr->ic->k_rf*2.0);
1347 crf = _mm256_set1_pd(fr->ic->c_rf);
1348 nvdwtype = fr->ntype;
1349 vdwparam = fr->nbfp;
1350 vdwtype = mdatoms->typeA;
1351
1352 /* Setup water-specific parameters */
1353 inr = nlist->iinr[0];
1354 iq1 = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+1]));
1355 iq2 = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+2]));
1356 iq3 = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+3]));
1357 vdwioffsetptr0 = vdwparam+2*nvdwtype*vdwtype[inr+0];
1358
1359 jq1 = _mm256_set1_pd(charge[inr+1]);
1360 jq2 = _mm256_set1_pd(charge[inr+2]);
1361 jq3 = _mm256_set1_pd(charge[inr+3]);
1362 vdwjidx0A = 2*vdwtype[inr+0];
1363 c6_00 = _mm256_set1_pd(vdwioffsetptr0[vdwjidx0A]);
1364 c12_00 = _mm256_set1_pd(vdwioffsetptr0[vdwjidx0A+1]);
1365 qq11 = _mm256_mul_pd(iq1,jq1);
1366 qq12 = _mm256_mul_pd(iq1,jq2);
1367 qq13 = _mm256_mul_pd(iq1,jq3);
1368 qq21 = _mm256_mul_pd(iq2,jq1);
1369 qq22 = _mm256_mul_pd(iq2,jq2);
1370 qq23 = _mm256_mul_pd(iq2,jq3);
1371 qq31 = _mm256_mul_pd(iq3,jq1);
1372 qq32 = _mm256_mul_pd(iq3,jq2);
1373 qq33 = _mm256_mul_pd(iq3,jq3);
1374
1375 /* When we use explicit cutoffs the value must be identical for elec and VdW, so use elec as an arbitrary choice */
1376 rcutoff_scalar = fr->ic->rcoulomb;
1377 rcutoff = _mm256_set1_pd(rcutoff_scalar);
1378 rcutoff2 = _mm256_mul_pd(rcutoff,rcutoff);
1379
1380 sh_vdw_invrcut6 = _mm256_set1_pd(fr->ic->sh_invrc6);
1381 rvdw = _mm256_set1_pd(fr->ic->rvdw);
1382
1383 /* Avoid stupid compiler warnings */
1384 jnrA = jnrB = jnrC = jnrD = 0;
1385 j_coord_offsetA = 0;
1386 j_coord_offsetB = 0;
1387 j_coord_offsetC = 0;
1388 j_coord_offsetD = 0;
1389
1390 outeriter = 0;
1391 inneriter = 0;
1392
1393 for(iidx=0;iidx<4*DIM;iidx++)
1394 {
1395 scratch[iidx] = 0.0;
1396 }
1397
1398 /* Start outer loop over neighborlists */
1399 for(iidx=0; iidx<nri; iidx++)
1400 {
1401 /* Load shift vector for this list */
1402 i_shift_offset = DIM*shiftidx[iidx];
1403
1404 /* Load limits for loop over neighbors */
1405 j_index_start = jindex[iidx];
1406 j_index_end = jindex[iidx+1];
1407
1408 /* Get outer coordinate index */
1409 inr = iinr[iidx];
1410 i_coord_offset = DIM*inr;
1411
1412 /* Load i particle coords and add shift vector */
1413 gmx_mm256_load_shift_and_4rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset,
1414 &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
1415
1416 fix0 = _mm256_setzero_pd();
1417 fiy0 = _mm256_setzero_pd();
1418 fiz0 = _mm256_setzero_pd();
1419 fix1 = _mm256_setzero_pd();
1420 fiy1 = _mm256_setzero_pd();
1421 fiz1 = _mm256_setzero_pd();
1422 fix2 = _mm256_setzero_pd();
1423 fiy2 = _mm256_setzero_pd();
1424 fiz2 = _mm256_setzero_pd();
1425 fix3 = _mm256_setzero_pd();
1426 fiy3 = _mm256_setzero_pd();
1427 fiz3 = _mm256_setzero_pd();
1428
1429 /* Start inner kernel loop */
1430 for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4)
1431 {
1432
1433 /* Get j neighbor index, and coordinate index */
1434 jnrA = jjnr[jidx];
1435 jnrB = jjnr[jidx+1];
1436 jnrC = jjnr[jidx+2];
1437 jnrD = jjnr[jidx+3];
1438 j_coord_offsetA = DIM*jnrA;
1439 j_coord_offsetB = DIM*jnrB;
1440 j_coord_offsetC = DIM*jnrC;
1441 j_coord_offsetD = DIM*jnrD;
1442
1443 /* load j atom coordinates */
1444 gmx_mm256_load_4rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
1445 x+j_coord_offsetC,x+j_coord_offsetD,
1446 &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,
1447 &jy2,&jz2,&jx3,&jy3,&jz3);
1448
1449 /* Calculate displacement vector */
1450 dx00 = _mm256_sub_pd(ix0,jx0);
1451 dy00 = _mm256_sub_pd(iy0,jy0);
1452 dz00 = _mm256_sub_pd(iz0,jz0);
1453 dx11 = _mm256_sub_pd(ix1,jx1);
1454 dy11 = _mm256_sub_pd(iy1,jy1);
1455 dz11 = _mm256_sub_pd(iz1,jz1);
1456 dx12 = _mm256_sub_pd(ix1,jx2);
1457 dy12 = _mm256_sub_pd(iy1,jy2);
1458 dz12 = _mm256_sub_pd(iz1,jz2);
1459 dx13 = _mm256_sub_pd(ix1,jx3);
1460 dy13 = _mm256_sub_pd(iy1,jy3);
1461 dz13 = _mm256_sub_pd(iz1,jz3);
1462 dx21 = _mm256_sub_pd(ix2,jx1);
1463 dy21 = _mm256_sub_pd(iy2,jy1);
1464 dz21 = _mm256_sub_pd(iz2,jz1);
1465 dx22 = _mm256_sub_pd(ix2,jx2);
1466 dy22 = _mm256_sub_pd(iy2,jy2);
1467 dz22 = _mm256_sub_pd(iz2,jz2);
1468 dx23 = _mm256_sub_pd(ix2,jx3);
1469 dy23 = _mm256_sub_pd(iy2,jy3);
1470 dz23 = _mm256_sub_pd(iz2,jz3);
1471 dx31 = _mm256_sub_pd(ix3,jx1);
1472 dy31 = _mm256_sub_pd(iy3,jy1);
1473 dz31 = _mm256_sub_pd(iz3,jz1);
1474 dx32 = _mm256_sub_pd(ix3,jx2);
1475 dy32 = _mm256_sub_pd(iy3,jy2);
1476 dz32 = _mm256_sub_pd(iz3,jz2);
1477 dx33 = _mm256_sub_pd(ix3,jx3);
1478 dy33 = _mm256_sub_pd(iy3,jy3);
1479 dz33 = _mm256_sub_pd(iz3,jz3);
1480
1481 /* Calculate squared distance and things based on it */
1482 rsq00 = gmx_mm256_calc_rsq_pd(dx00,dy00,dz00);
1483 rsq11 = gmx_mm256_calc_rsq_pd(dx11,dy11,dz11);
1484 rsq12 = gmx_mm256_calc_rsq_pd(dx12,dy12,dz12);
1485 rsq13 = gmx_mm256_calc_rsq_pd(dx13,dy13,dz13);
1486 rsq21 = gmx_mm256_calc_rsq_pd(dx21,dy21,dz21);
1487 rsq22 = gmx_mm256_calc_rsq_pd(dx22,dy22,dz22);
1488 rsq23 = gmx_mm256_calc_rsq_pd(dx23,dy23,dz23);
1489 rsq31 = gmx_mm256_calc_rsq_pd(dx31,dy31,dz31);
1490 rsq32 = gmx_mm256_calc_rsq_pd(dx32,dy32,dz32);
1491 rsq33 = gmx_mm256_calc_rsq_pd(dx33,dy33,dz33);
1492
1493 rinv11 = avx256_invsqrt_d(rsq11);
1494 rinv12 = avx256_invsqrt_d(rsq12);
1495 rinv13 = avx256_invsqrt_d(rsq13);
1496 rinv21 = avx256_invsqrt_d(rsq21);
1497 rinv22 = avx256_invsqrt_d(rsq22);
1498 rinv23 = avx256_invsqrt_d(rsq23);
1499 rinv31 = avx256_invsqrt_d(rsq31);
1500 rinv32 = avx256_invsqrt_d(rsq32);
1501 rinv33 = avx256_invsqrt_d(rsq33);
1502
1503 rinvsq00 = avx256_inv_d(rsq00);
1504 rinvsq11 = _mm256_mul_pd(rinv11,rinv11);
1505 rinvsq12 = _mm256_mul_pd(rinv12,rinv12);
1506 rinvsq13 = _mm256_mul_pd(rinv13,rinv13);
1507 rinvsq21 = _mm256_mul_pd(rinv21,rinv21);
1508 rinvsq22 = _mm256_mul_pd(rinv22,rinv22);
1509 rinvsq23 = _mm256_mul_pd(rinv23,rinv23);
1510 rinvsq31 = _mm256_mul_pd(rinv31,rinv31);
1511 rinvsq32 = _mm256_mul_pd(rinv32,rinv32);
1512 rinvsq33 = _mm256_mul_pd(rinv33,rinv33);
1513
1514 fjx0 = _mm256_setzero_pd();
1515 fjy0 = _mm256_setzero_pd();
1516 fjz0 = _mm256_setzero_pd();
1517 fjx1 = _mm256_setzero_pd();
1518 fjy1 = _mm256_setzero_pd();
1519 fjz1 = _mm256_setzero_pd();
1520 fjx2 = _mm256_setzero_pd();
1521 fjy2 = _mm256_setzero_pd();
1522 fjz2 = _mm256_setzero_pd();
1523 fjx3 = _mm256_setzero_pd();
1524 fjy3 = _mm256_setzero_pd();
1525 fjz3 = _mm256_setzero_pd();
1526
1527 /**************************
1528 * CALCULATE INTERACTIONS *
1529 **************************/
1530
1531 if (gmx_mm256_any_lt(rsq00,rcutoff2))
1532 {
1533
1534 /* LENNARD-JONES DISPERSION/REPULSION */
1535
1536 rinvsix = _mm256_mul_pd(_mm256_mul_pd(rinvsq00,rinvsq00),rinvsq00);
1537 fvdw = _mm256_mul_pd(_mm256_sub_pd(_mm256_mul_pd(c12_00,rinvsix),c6_00),_mm256_mul_pd(rinvsix,rinvsq00));
1538
1539 cutoff_mask = _mm256_cmp_pd(rsq00,rcutoff2,_CMP_LT_OQ);
1540
1541 fscal = fvdw;
1542
1543 fscal = _mm256_and_pd(fscal,cutoff_mask);
1544
1545 /* Calculate temporary vectorial force */
1546 tx = _mm256_mul_pd(fscal,dx00);
1547 ty = _mm256_mul_pd(fscal,dy00);
1548 tz = _mm256_mul_pd(fscal,dz00);
1549
1550 /* Update vectorial force */
1551 fix0 = _mm256_add_pd(fix0,tx);
1552 fiy0 = _mm256_add_pd(fiy0,ty);
1553 fiz0 = _mm256_add_pd(fiz0,tz);
1554
1555 fjx0 = _mm256_add_pd(fjx0,tx);
1556 fjy0 = _mm256_add_pd(fjy0,ty);
1557 fjz0 = _mm256_add_pd(fjz0,tz);
1558
1559 }
1560
1561 /**************************
1562 * CALCULATE INTERACTIONS *
1563 **************************/
1564
1565 if (gmx_mm256_any_lt(rsq11,rcutoff2))
1566 {
1567
1568 /* REACTION-FIELD ELECTROSTATICS */
1569 felec = _mm256_mul_pd(qq11,_mm256_sub_pd(_mm256_mul_pd(rinv11,rinvsq11),krf2));
1570
1571 cutoff_mask = _mm256_cmp_pd(rsq11,rcutoff2,_CMP_LT_OQ);
1572
1573 fscal = felec;
1574
1575 fscal = _mm256_and_pd(fscal,cutoff_mask);
1576
1577 /* Calculate temporary vectorial force */
1578 tx = _mm256_mul_pd(fscal,dx11);
1579 ty = _mm256_mul_pd(fscal,dy11);
1580 tz = _mm256_mul_pd(fscal,dz11);
1581
1582 /* Update vectorial force */
1583 fix1 = _mm256_add_pd(fix1,tx);
1584 fiy1 = _mm256_add_pd(fiy1,ty);
1585 fiz1 = _mm256_add_pd(fiz1,tz);
1586
1587 fjx1 = _mm256_add_pd(fjx1,tx);
1588 fjy1 = _mm256_add_pd(fjy1,ty);
1589 fjz1 = _mm256_add_pd(fjz1,tz);
1590
1591 }
1592
1593 /**************************
1594 * CALCULATE INTERACTIONS *
1595 **************************/
1596
1597 if (gmx_mm256_any_lt(rsq12,rcutoff2))
1598 {
1599
1600 /* REACTION-FIELD ELECTROSTATICS */
1601 felec = _mm256_mul_pd(qq12,_mm256_sub_pd(_mm256_mul_pd(rinv12,rinvsq12),krf2));
1602
1603 cutoff_mask = _mm256_cmp_pd(rsq12,rcutoff2,_CMP_LT_OQ);
1604
1605 fscal = felec;
1606
1607 fscal = _mm256_and_pd(fscal,cutoff_mask);
1608
1609 /* Calculate temporary vectorial force */
1610 tx = _mm256_mul_pd(fscal,dx12);
1611 ty = _mm256_mul_pd(fscal,dy12);
1612 tz = _mm256_mul_pd(fscal,dz12);
1613
1614 /* Update vectorial force */
1615 fix1 = _mm256_add_pd(fix1,tx);
1616 fiy1 = _mm256_add_pd(fiy1,ty);
1617 fiz1 = _mm256_add_pd(fiz1,tz);
1618
1619 fjx2 = _mm256_add_pd(fjx2,tx);
1620 fjy2 = _mm256_add_pd(fjy2,ty);
1621 fjz2 = _mm256_add_pd(fjz2,tz);
1622
1623 }
1624
1625 /**************************
1626 * CALCULATE INTERACTIONS *
1627 **************************/
1628
1629 if (gmx_mm256_any_lt(rsq13,rcutoff2))
1630 {
1631
1632 /* REACTION-FIELD ELECTROSTATICS */
1633 felec = _mm256_mul_pd(qq13,_mm256_sub_pd(_mm256_mul_pd(rinv13,rinvsq13),krf2));
1634
1635 cutoff_mask = _mm256_cmp_pd(rsq13,rcutoff2,_CMP_LT_OQ);
1636
1637 fscal = felec;
1638
1639 fscal = _mm256_and_pd(fscal,cutoff_mask);
1640
1641 /* Calculate temporary vectorial force */
1642 tx = _mm256_mul_pd(fscal,dx13);
1643 ty = _mm256_mul_pd(fscal,dy13);
1644 tz = _mm256_mul_pd(fscal,dz13);
1645
1646 /* Update vectorial force */
1647 fix1 = _mm256_add_pd(fix1,tx);
1648 fiy1 = _mm256_add_pd(fiy1,ty);
1649 fiz1 = _mm256_add_pd(fiz1,tz);
1650
1651 fjx3 = _mm256_add_pd(fjx3,tx);
1652 fjy3 = _mm256_add_pd(fjy3,ty);
1653 fjz3 = _mm256_add_pd(fjz3,tz);
1654
1655 }
1656
1657 /**************************
1658 * CALCULATE INTERACTIONS *
1659 **************************/
1660
1661 if (gmx_mm256_any_lt(rsq21,rcutoff2))
1662 {
1663
1664 /* REACTION-FIELD ELECTROSTATICS */
1665 felec = _mm256_mul_pd(qq21,_mm256_sub_pd(_mm256_mul_pd(rinv21,rinvsq21),krf2));
1666
1667 cutoff_mask = _mm256_cmp_pd(rsq21,rcutoff2,_CMP_LT_OQ);
1668
1669 fscal = felec;
1670
1671 fscal = _mm256_and_pd(fscal,cutoff_mask);
1672
1673 /* Calculate temporary vectorial force */
1674 tx = _mm256_mul_pd(fscal,dx21);
1675 ty = _mm256_mul_pd(fscal,dy21);
1676 tz = _mm256_mul_pd(fscal,dz21);
1677
1678 /* Update vectorial force */
1679 fix2 = _mm256_add_pd(fix2,tx);
1680 fiy2 = _mm256_add_pd(fiy2,ty);
1681 fiz2 = _mm256_add_pd(fiz2,tz);
1682
1683 fjx1 = _mm256_add_pd(fjx1,tx);
1684 fjy1 = _mm256_add_pd(fjy1,ty);
1685 fjz1 = _mm256_add_pd(fjz1,tz);
1686
1687 }
1688
1689 /**************************
1690 * CALCULATE INTERACTIONS *
1691 **************************/
1692
1693 if (gmx_mm256_any_lt(rsq22,rcutoff2))
1694 {
1695
1696 /* REACTION-FIELD ELECTROSTATICS */
1697 felec = _mm256_mul_pd(qq22,_mm256_sub_pd(_mm256_mul_pd(rinv22,rinvsq22),krf2));
1698
1699 cutoff_mask = _mm256_cmp_pd(rsq22,rcutoff2,_CMP_LT_OQ);
1700
1701 fscal = felec;
1702
1703 fscal = _mm256_and_pd(fscal,cutoff_mask);
1704
1705 /* Calculate temporary vectorial force */
1706 tx = _mm256_mul_pd(fscal,dx22);
1707 ty = _mm256_mul_pd(fscal,dy22);
1708 tz = _mm256_mul_pd(fscal,dz22);
1709
1710 /* Update vectorial force */
1711 fix2 = _mm256_add_pd(fix2,tx);
1712 fiy2 = _mm256_add_pd(fiy2,ty);
1713 fiz2 = _mm256_add_pd(fiz2,tz);
1714
1715 fjx2 = _mm256_add_pd(fjx2,tx);
1716 fjy2 = _mm256_add_pd(fjy2,ty);
1717 fjz2 = _mm256_add_pd(fjz2,tz);
1718
1719 }
1720
1721 /**************************
1722 * CALCULATE INTERACTIONS *
1723 **************************/
1724
1725 if (gmx_mm256_any_lt(rsq23,rcutoff2))
1726 {
1727
1728 /* REACTION-FIELD ELECTROSTATICS */
1729 felec = _mm256_mul_pd(qq23,_mm256_sub_pd(_mm256_mul_pd(rinv23,rinvsq23),krf2));
1730
1731 cutoff_mask = _mm256_cmp_pd(rsq23,rcutoff2,_CMP_LT_OQ);
1732
1733 fscal = felec;
1734
1735 fscal = _mm256_and_pd(fscal,cutoff_mask);
1736
1737 /* Calculate temporary vectorial force */
1738 tx = _mm256_mul_pd(fscal,dx23);
1739 ty = _mm256_mul_pd(fscal,dy23);
1740 tz = _mm256_mul_pd(fscal,dz23);
1741
1742 /* Update vectorial force */
1743 fix2 = _mm256_add_pd(fix2,tx);
1744 fiy2 = _mm256_add_pd(fiy2,ty);
1745 fiz2 = _mm256_add_pd(fiz2,tz);
1746
1747 fjx3 = _mm256_add_pd(fjx3,tx);
1748 fjy3 = _mm256_add_pd(fjy3,ty);
1749 fjz3 = _mm256_add_pd(fjz3,tz);
1750
1751 }
1752
1753 /**************************
1754 * CALCULATE INTERACTIONS *
1755 **************************/
1756
1757 if (gmx_mm256_any_lt(rsq31,rcutoff2))
1758 {
1759
1760 /* REACTION-FIELD ELECTROSTATICS */
1761 felec = _mm256_mul_pd(qq31,_mm256_sub_pd(_mm256_mul_pd(rinv31,rinvsq31),krf2));
1762
1763 cutoff_mask = _mm256_cmp_pd(rsq31,rcutoff2,_CMP_LT_OQ);
1764
1765 fscal = felec;
1766
1767 fscal = _mm256_and_pd(fscal,cutoff_mask);
1768
1769 /* Calculate temporary vectorial force */
1770 tx = _mm256_mul_pd(fscal,dx31);
1771 ty = _mm256_mul_pd(fscal,dy31);
1772 tz = _mm256_mul_pd(fscal,dz31);
1773
1774 /* Update vectorial force */
1775 fix3 = _mm256_add_pd(fix3,tx);
1776 fiy3 = _mm256_add_pd(fiy3,ty);
1777 fiz3 = _mm256_add_pd(fiz3,tz);
1778
1779 fjx1 = _mm256_add_pd(fjx1,tx);
1780 fjy1 = _mm256_add_pd(fjy1,ty);
1781 fjz1 = _mm256_add_pd(fjz1,tz);
1782
1783 }
1784
1785 /**************************
1786 * CALCULATE INTERACTIONS *
1787 **************************/
1788
1789 if (gmx_mm256_any_lt(rsq32,rcutoff2))
1790 {
1791
1792 /* REACTION-FIELD ELECTROSTATICS */
1793 felec = _mm256_mul_pd(qq32,_mm256_sub_pd(_mm256_mul_pd(rinv32,rinvsq32),krf2));
1794
1795 cutoff_mask = _mm256_cmp_pd(rsq32,rcutoff2,_CMP_LT_OQ);
1796
1797 fscal = felec;
1798
1799 fscal = _mm256_and_pd(fscal,cutoff_mask);
1800
1801 /* Calculate temporary vectorial force */
1802 tx = _mm256_mul_pd(fscal,dx32);
1803 ty = _mm256_mul_pd(fscal,dy32);
1804 tz = _mm256_mul_pd(fscal,dz32);
1805
1806 /* Update vectorial force */
1807 fix3 = _mm256_add_pd(fix3,tx);
1808 fiy3 = _mm256_add_pd(fiy3,ty);
1809 fiz3 = _mm256_add_pd(fiz3,tz);
1810
1811 fjx2 = _mm256_add_pd(fjx2,tx);
1812 fjy2 = _mm256_add_pd(fjy2,ty);
1813 fjz2 = _mm256_add_pd(fjz2,tz);
1814
1815 }
1816
1817 /**************************
1818 * CALCULATE INTERACTIONS *
1819 **************************/
1820
1821 if (gmx_mm256_any_lt(rsq33,rcutoff2))
1822 {
1823
1824 /* REACTION-FIELD ELECTROSTATICS */
1825 felec = _mm256_mul_pd(qq33,_mm256_sub_pd(_mm256_mul_pd(rinv33,rinvsq33),krf2));
1826
1827 cutoff_mask = _mm256_cmp_pd(rsq33,rcutoff2,_CMP_LT_OQ);
1828
1829 fscal = felec;
1830
1831 fscal = _mm256_and_pd(fscal,cutoff_mask);
1832
1833 /* Calculate temporary vectorial force */
1834 tx = _mm256_mul_pd(fscal,dx33);
1835 ty = _mm256_mul_pd(fscal,dy33);
1836 tz = _mm256_mul_pd(fscal,dz33);
1837
1838 /* Update vectorial force */
1839 fix3 = _mm256_add_pd(fix3,tx);
1840 fiy3 = _mm256_add_pd(fiy3,ty);
1841 fiz3 = _mm256_add_pd(fiz3,tz);
1842
1843 fjx3 = _mm256_add_pd(fjx3,tx);
1844 fjy3 = _mm256_add_pd(fjy3,ty);
1845 fjz3 = _mm256_add_pd(fjz3,tz);
1846
1847 }
1848
1849 fjptrA = f+j_coord_offsetA;
1850 fjptrB = f+j_coord_offsetB;
1851 fjptrC = f+j_coord_offsetC;
1852 fjptrD = f+j_coord_offsetD;
1853
1854 gmx_mm256_decrement_4rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,
1855 fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,
1856 fjx2,fjy2,fjz2,fjx3,fjy3,fjz3);
1857
1858 /* Inner loop uses 303 flops */
1859 }
1860
1861 if(jidx<j_index_end)
1862 {
1863
1864 /* Get j neighbor index, and coordinate index */
1865 jnrlistA = jjnr[jidx];
1866 jnrlistB = jjnr[jidx+1];
1867 jnrlistC = jjnr[jidx+2];
1868 jnrlistD = jjnr[jidx+3];
1869 /* Sign of each element will be negative for non-real atoms.
1870 * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
1871 * so use it as val = _mm_andnot_pd(mask,val) to clear dummy entries.
1872 */
1873 tmpmask0 = gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128()));
1874
1875 tmpmask1 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(3,3,2,2));
1876 tmpmask0 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(1,1,0,0));
1877 dummy_mask = _mm256_castps_pd(gmx_mm256_set_m128(tmpmask1,tmpmask0));
1878
1879 jnrA = (jnrlistA>=0) ? jnrlistA : 0;
1880 jnrB = (jnrlistB>=0) ? jnrlistB : 0;
1881 jnrC = (jnrlistC>=0) ? jnrlistC : 0;
1882 jnrD = (jnrlistD>=0) ? jnrlistD : 0;
1883 j_coord_offsetA = DIM*jnrA;
1884 j_coord_offsetB = DIM*jnrB;
1885 j_coord_offsetC = DIM*jnrC;
1886 j_coord_offsetD = DIM*jnrD;
1887
1888 /* load j atom coordinates */
1889 gmx_mm256_load_4rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
1890 x+j_coord_offsetC,x+j_coord_offsetD,
1891 &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2,
1892 &jy2,&jz2,&jx3,&jy3,&jz3);
1893
1894 /* Calculate displacement vector */
1895 dx00 = _mm256_sub_pd(ix0,jx0);
1896 dy00 = _mm256_sub_pd(iy0,jy0);
1897 dz00 = _mm256_sub_pd(iz0,jz0);
1898 dx11 = _mm256_sub_pd(ix1,jx1);
1899 dy11 = _mm256_sub_pd(iy1,jy1);
1900 dz11 = _mm256_sub_pd(iz1,jz1);
1901 dx12 = _mm256_sub_pd(ix1,jx2);
1902 dy12 = _mm256_sub_pd(iy1,jy2);
1903 dz12 = _mm256_sub_pd(iz1,jz2);
1904 dx13 = _mm256_sub_pd(ix1,jx3);
1905 dy13 = _mm256_sub_pd(iy1,jy3);
1906 dz13 = _mm256_sub_pd(iz1,jz3);
1907 dx21 = _mm256_sub_pd(ix2,jx1);
1908 dy21 = _mm256_sub_pd(iy2,jy1);
1909 dz21 = _mm256_sub_pd(iz2,jz1);
1910 dx22 = _mm256_sub_pd(ix2,jx2);
1911 dy22 = _mm256_sub_pd(iy2,jy2);
1912 dz22 = _mm256_sub_pd(iz2,jz2);
1913 dx23 = _mm256_sub_pd(ix2,jx3);
1914 dy23 = _mm256_sub_pd(iy2,jy3);
1915 dz23 = _mm256_sub_pd(iz2,jz3);
1916 dx31 = _mm256_sub_pd(ix3,jx1);
1917 dy31 = _mm256_sub_pd(iy3,jy1);
1918 dz31 = _mm256_sub_pd(iz3,jz1);
1919 dx32 = _mm256_sub_pd(ix3,jx2);
1920 dy32 = _mm256_sub_pd(iy3,jy2);
1921 dz32 = _mm256_sub_pd(iz3,jz2);
1922 dx33 = _mm256_sub_pd(ix3,jx3);
1923 dy33 = _mm256_sub_pd(iy3,jy3);
1924 dz33 = _mm256_sub_pd(iz3,jz3);
1925
1926 /* Calculate squared distance and things based on it */
1927 rsq00 = gmx_mm256_calc_rsq_pd(dx00,dy00,dz00);
1928 rsq11 = gmx_mm256_calc_rsq_pd(dx11,dy11,dz11);
1929 rsq12 = gmx_mm256_calc_rsq_pd(dx12,dy12,dz12);
1930 rsq13 = gmx_mm256_calc_rsq_pd(dx13,dy13,dz13);
1931 rsq21 = gmx_mm256_calc_rsq_pd(dx21,dy21,dz21);
1932 rsq22 = gmx_mm256_calc_rsq_pd(dx22,dy22,dz22);
1933 rsq23 = gmx_mm256_calc_rsq_pd(dx23,dy23,dz23);
1934 rsq31 = gmx_mm256_calc_rsq_pd(dx31,dy31,dz31);
1935 rsq32 = gmx_mm256_calc_rsq_pd(dx32,dy32,dz32);
1936 rsq33 = gmx_mm256_calc_rsq_pd(dx33,dy33,dz33);
1937
1938 rinv11 = avx256_invsqrt_d(rsq11);
1939 rinv12 = avx256_invsqrt_d(rsq12);
1940 rinv13 = avx256_invsqrt_d(rsq13);
1941 rinv21 = avx256_invsqrt_d(rsq21);
1942 rinv22 = avx256_invsqrt_d(rsq22);
1943 rinv23 = avx256_invsqrt_d(rsq23);
1944 rinv31 = avx256_invsqrt_d(rsq31);
1945 rinv32 = avx256_invsqrt_d(rsq32);
1946 rinv33 = avx256_invsqrt_d(rsq33);
1947
1948 rinvsq00 = avx256_inv_d(rsq00);
1949 rinvsq11 = _mm256_mul_pd(rinv11,rinv11);
1950 rinvsq12 = _mm256_mul_pd(rinv12,rinv12);
1951 rinvsq13 = _mm256_mul_pd(rinv13,rinv13);
1952 rinvsq21 = _mm256_mul_pd(rinv21,rinv21);
1953 rinvsq22 = _mm256_mul_pd(rinv22,rinv22);
1954 rinvsq23 = _mm256_mul_pd(rinv23,rinv23);
1955 rinvsq31 = _mm256_mul_pd(rinv31,rinv31);
1956 rinvsq32 = _mm256_mul_pd(rinv32,rinv32);
1957 rinvsq33 = _mm256_mul_pd(rinv33,rinv33);
1958
1959 fjx0 = _mm256_setzero_pd();
1960 fjy0 = _mm256_setzero_pd();
1961 fjz0 = _mm256_setzero_pd();
1962 fjx1 = _mm256_setzero_pd();
1963 fjy1 = _mm256_setzero_pd();
1964 fjz1 = _mm256_setzero_pd();
1965 fjx2 = _mm256_setzero_pd();
1966 fjy2 = _mm256_setzero_pd();
1967 fjz2 = _mm256_setzero_pd();
1968 fjx3 = _mm256_setzero_pd();
1969 fjy3 = _mm256_setzero_pd();
1970 fjz3 = _mm256_setzero_pd();
1971
1972 /**************************
1973 * CALCULATE INTERACTIONS *
1974 **************************/
1975
1976 if (gmx_mm256_any_lt(rsq00,rcutoff2))
1977 {
1978
1979 /* LENNARD-JONES DISPERSION/REPULSION */
1980
1981 rinvsix = _mm256_mul_pd(_mm256_mul_pd(rinvsq00,rinvsq00),rinvsq00);
1982 fvdw = _mm256_mul_pd(_mm256_sub_pd(_mm256_mul_pd(c12_00,rinvsix),c6_00),_mm256_mul_pd(rinvsix,rinvsq00));
1983
1984 cutoff_mask = _mm256_cmp_pd(rsq00,rcutoff2,_CMP_LT_OQ);
1985
1986 fscal = fvdw;
1987
1988 fscal = _mm256_and_pd(fscal,cutoff_mask);
1989
1990 fscal = _mm256_andnot_pd(dummy_mask,fscal);
1991
1992 /* Calculate temporary vectorial force */
1993 tx = _mm256_mul_pd(fscal,dx00);
1994 ty = _mm256_mul_pd(fscal,dy00);
1995 tz = _mm256_mul_pd(fscal,dz00);
1996
1997 /* Update vectorial force */
1998 fix0 = _mm256_add_pd(fix0,tx);
1999 fiy0 = _mm256_add_pd(fiy0,ty);
2000 fiz0 = _mm256_add_pd(fiz0,tz);
2001
2002 fjx0 = _mm256_add_pd(fjx0,tx);
2003 fjy0 = _mm256_add_pd(fjy0,ty);
2004 fjz0 = _mm256_add_pd(fjz0,tz);
2005
2006 }
2007
2008 /**************************
2009 * CALCULATE INTERACTIONS *
2010 **************************/
2011
2012 if (gmx_mm256_any_lt(rsq11,rcutoff2))
2013 {
2014
2015 /* REACTION-FIELD ELECTROSTATICS */
2016 felec = _mm256_mul_pd(qq11,_mm256_sub_pd(_mm256_mul_pd(rinv11,rinvsq11),krf2));
2017
2018 cutoff_mask = _mm256_cmp_pd(rsq11,rcutoff2,_CMP_LT_OQ);
2019
2020 fscal = felec;
2021
2022 fscal = _mm256_and_pd(fscal,cutoff_mask);
2023
2024 fscal = _mm256_andnot_pd(dummy_mask,fscal);
2025
2026 /* Calculate temporary vectorial force */
2027 tx = _mm256_mul_pd(fscal,dx11);
2028 ty = _mm256_mul_pd(fscal,dy11);
2029 tz = _mm256_mul_pd(fscal,dz11);
2030
2031 /* Update vectorial force */
2032 fix1 = _mm256_add_pd(fix1,tx);
2033 fiy1 = _mm256_add_pd(fiy1,ty);
2034 fiz1 = _mm256_add_pd(fiz1,tz);
2035
2036 fjx1 = _mm256_add_pd(fjx1,tx);
2037 fjy1 = _mm256_add_pd(fjy1,ty);
2038 fjz1 = _mm256_add_pd(fjz1,tz);
2039
2040 }
2041
2042 /**************************
2043 * CALCULATE INTERACTIONS *
2044 **************************/
2045
2046 if (gmx_mm256_any_lt(rsq12,rcutoff2))
2047 {
2048
2049 /* REACTION-FIELD ELECTROSTATICS */
2050 felec = _mm256_mul_pd(qq12,_mm256_sub_pd(_mm256_mul_pd(rinv12,rinvsq12),krf2));
2051
2052 cutoff_mask = _mm256_cmp_pd(rsq12,rcutoff2,_CMP_LT_OQ);
2053
2054 fscal = felec;
2055
2056 fscal = _mm256_and_pd(fscal,cutoff_mask);
2057
2058 fscal = _mm256_andnot_pd(dummy_mask,fscal);
2059
2060 /* Calculate temporary vectorial force */
2061 tx = _mm256_mul_pd(fscal,dx12);
2062 ty = _mm256_mul_pd(fscal,dy12);
2063 tz = _mm256_mul_pd(fscal,dz12);
2064
2065 /* Update vectorial force */
2066 fix1 = _mm256_add_pd(fix1,tx);
2067 fiy1 = _mm256_add_pd(fiy1,ty);
2068 fiz1 = _mm256_add_pd(fiz1,tz);
2069
2070 fjx2 = _mm256_add_pd(fjx2,tx);
2071 fjy2 = _mm256_add_pd(fjy2,ty);
2072 fjz2 = _mm256_add_pd(fjz2,tz);
2073
2074 }
2075
2076 /**************************
2077 * CALCULATE INTERACTIONS *
2078 **************************/
2079
2080 if (gmx_mm256_any_lt(rsq13,rcutoff2))
2081 {
2082
2083 /* REACTION-FIELD ELECTROSTATICS */
2084 felec = _mm256_mul_pd(qq13,_mm256_sub_pd(_mm256_mul_pd(rinv13,rinvsq13),krf2));
2085
2086 cutoff_mask = _mm256_cmp_pd(rsq13,rcutoff2,_CMP_LT_OQ);
2087
2088 fscal = felec;
2089
2090 fscal = _mm256_and_pd(fscal,cutoff_mask);
2091
2092 fscal = _mm256_andnot_pd(dummy_mask,fscal);
2093
2094 /* Calculate temporary vectorial force */
2095 tx = _mm256_mul_pd(fscal,dx13);
2096 ty = _mm256_mul_pd(fscal,dy13);
2097 tz = _mm256_mul_pd(fscal,dz13);
2098
2099 /* Update vectorial force */
2100 fix1 = _mm256_add_pd(fix1,tx);
2101 fiy1 = _mm256_add_pd(fiy1,ty);
2102 fiz1 = _mm256_add_pd(fiz1,tz);
2103
2104 fjx3 = _mm256_add_pd(fjx3,tx);
2105 fjy3 = _mm256_add_pd(fjy3,ty);
2106 fjz3 = _mm256_add_pd(fjz3,tz);
2107
2108 }
2109
2110 /**************************
2111 * CALCULATE INTERACTIONS *
2112 **************************/
2113
2114 if (gmx_mm256_any_lt(rsq21,rcutoff2))
2115 {
2116
2117 /* REACTION-FIELD ELECTROSTATICS */
2118 felec = _mm256_mul_pd(qq21,_mm256_sub_pd(_mm256_mul_pd(rinv21,rinvsq21),krf2));
2119
2120 cutoff_mask = _mm256_cmp_pd(rsq21,rcutoff2,_CMP_LT_OQ);
2121
2122 fscal = felec;
2123
2124 fscal = _mm256_and_pd(fscal,cutoff_mask);
2125
2126 fscal = _mm256_andnot_pd(dummy_mask,fscal);
2127
2128 /* Calculate temporary vectorial force */
2129 tx = _mm256_mul_pd(fscal,dx21);
2130 ty = _mm256_mul_pd(fscal,dy21);
2131 tz = _mm256_mul_pd(fscal,dz21);
2132
2133 /* Update vectorial force */
2134 fix2 = _mm256_add_pd(fix2,tx);
2135 fiy2 = _mm256_add_pd(fiy2,ty);
2136 fiz2 = _mm256_add_pd(fiz2,tz);
2137
2138 fjx1 = _mm256_add_pd(fjx1,tx);
2139 fjy1 = _mm256_add_pd(fjy1,ty);
2140 fjz1 = _mm256_add_pd(fjz1,tz);
2141
2142 }
2143
2144 /**************************
2145 * CALCULATE INTERACTIONS *
2146 **************************/
2147
2148 if (gmx_mm256_any_lt(rsq22,rcutoff2))
2149 {
2150
2151 /* REACTION-FIELD ELECTROSTATICS */
2152 felec = _mm256_mul_pd(qq22,_mm256_sub_pd(_mm256_mul_pd(rinv22,rinvsq22),krf2));
2153
2154 cutoff_mask = _mm256_cmp_pd(rsq22,rcutoff2,_CMP_LT_OQ);
2155
2156 fscal = felec;
2157
2158 fscal = _mm256_and_pd(fscal,cutoff_mask);
2159
2160 fscal = _mm256_andnot_pd(dummy_mask,fscal);
2161
2162 /* Calculate temporary vectorial force */
2163 tx = _mm256_mul_pd(fscal,dx22);
2164 ty = _mm256_mul_pd(fscal,dy22);
2165 tz = _mm256_mul_pd(fscal,dz22);
2166
2167 /* Update vectorial force */
2168 fix2 = _mm256_add_pd(fix2,tx);
2169 fiy2 = _mm256_add_pd(fiy2,ty);
2170 fiz2 = _mm256_add_pd(fiz2,tz);
2171
2172 fjx2 = _mm256_add_pd(fjx2,tx);
2173 fjy2 = _mm256_add_pd(fjy2,ty);
2174 fjz2 = _mm256_add_pd(fjz2,tz);
2175
2176 }
2177
2178 /**************************
2179 * CALCULATE INTERACTIONS *
2180 **************************/
2181
2182 if (gmx_mm256_any_lt(rsq23,rcutoff2))
2183 {
2184
2185 /* REACTION-FIELD ELECTROSTATICS */
2186 felec = _mm256_mul_pd(qq23,_mm256_sub_pd(_mm256_mul_pd(rinv23,rinvsq23),krf2));
2187
2188 cutoff_mask = _mm256_cmp_pd(rsq23,rcutoff2,_CMP_LT_OQ);
2189
2190 fscal = felec;
2191
2192 fscal = _mm256_and_pd(fscal,cutoff_mask);
2193
2194 fscal = _mm256_andnot_pd(dummy_mask,fscal);
2195
2196 /* Calculate temporary vectorial force */
2197 tx = _mm256_mul_pd(fscal,dx23);
2198 ty = _mm256_mul_pd(fscal,dy23);
2199 tz = _mm256_mul_pd(fscal,dz23);
2200
2201 /* Update vectorial force */
2202 fix2 = _mm256_add_pd(fix2,tx);
2203 fiy2 = _mm256_add_pd(fiy2,ty);
2204 fiz2 = _mm256_add_pd(fiz2,tz);
2205
2206 fjx3 = _mm256_add_pd(fjx3,tx);
2207 fjy3 = _mm256_add_pd(fjy3,ty);
2208 fjz3 = _mm256_add_pd(fjz3,tz);
2209
2210 }
2211
2212 /**************************
2213 * CALCULATE INTERACTIONS *
2214 **************************/
2215
2216 if (gmx_mm256_any_lt(rsq31,rcutoff2))
2217 {
2218
2219 /* REACTION-FIELD ELECTROSTATICS */
2220 felec = _mm256_mul_pd(qq31,_mm256_sub_pd(_mm256_mul_pd(rinv31,rinvsq31),krf2));
2221
2222 cutoff_mask = _mm256_cmp_pd(rsq31,rcutoff2,_CMP_LT_OQ);
2223
2224 fscal = felec;
2225
2226 fscal = _mm256_and_pd(fscal,cutoff_mask);
2227
2228 fscal = _mm256_andnot_pd(dummy_mask,fscal);
2229
2230 /* Calculate temporary vectorial force */
2231 tx = _mm256_mul_pd(fscal,dx31);
2232 ty = _mm256_mul_pd(fscal,dy31);
2233 tz = _mm256_mul_pd(fscal,dz31);
2234
2235 /* Update vectorial force */
2236 fix3 = _mm256_add_pd(fix3,tx);
2237 fiy3 = _mm256_add_pd(fiy3,ty);
2238 fiz3 = _mm256_add_pd(fiz3,tz);
2239
2240 fjx1 = _mm256_add_pd(fjx1,tx);
2241 fjy1 = _mm256_add_pd(fjy1,ty);
2242 fjz1 = _mm256_add_pd(fjz1,tz);
2243
2244 }
2245
2246 /**************************
2247 * CALCULATE INTERACTIONS *
2248 **************************/
2249
2250 if (gmx_mm256_any_lt(rsq32,rcutoff2))
2251 {
2252
2253 /* REACTION-FIELD ELECTROSTATICS */
2254 felec = _mm256_mul_pd(qq32,_mm256_sub_pd(_mm256_mul_pd(rinv32,rinvsq32),krf2));
2255
2256 cutoff_mask = _mm256_cmp_pd(rsq32,rcutoff2,_CMP_LT_OQ);
2257
2258 fscal = felec;
2259
2260 fscal = _mm256_and_pd(fscal,cutoff_mask);
2261
2262 fscal = _mm256_andnot_pd(dummy_mask,fscal);
2263
2264 /* Calculate temporary vectorial force */
2265 tx = _mm256_mul_pd(fscal,dx32);
2266 ty = _mm256_mul_pd(fscal,dy32);
2267 tz = _mm256_mul_pd(fscal,dz32);
2268
2269 /* Update vectorial force */
2270 fix3 = _mm256_add_pd(fix3,tx);
2271 fiy3 = _mm256_add_pd(fiy3,ty);
2272 fiz3 = _mm256_add_pd(fiz3,tz);
2273
2274 fjx2 = _mm256_add_pd(fjx2,tx);
2275 fjy2 = _mm256_add_pd(fjy2,ty);
2276 fjz2 = _mm256_add_pd(fjz2,tz);
2277
2278 }
2279
2280 /**************************
2281 * CALCULATE INTERACTIONS *
2282 **************************/
2283
2284 if (gmx_mm256_any_lt(rsq33,rcutoff2))
2285 {
2286
2287 /* REACTION-FIELD ELECTROSTATICS */
2288 felec = _mm256_mul_pd(qq33,_mm256_sub_pd(_mm256_mul_pd(rinv33,rinvsq33),krf2));
2289
2290 cutoff_mask = _mm256_cmp_pd(rsq33,rcutoff2,_CMP_LT_OQ);
2291
2292 fscal = felec;
2293
2294 fscal = _mm256_and_pd(fscal,cutoff_mask);
2295
2296 fscal = _mm256_andnot_pd(dummy_mask,fscal);
2297
2298 /* Calculate temporary vectorial force */
2299 tx = _mm256_mul_pd(fscal,dx33);
2300 ty = _mm256_mul_pd(fscal,dy33);
2301 tz = _mm256_mul_pd(fscal,dz33);
2302
2303 /* Update vectorial force */
2304 fix3 = _mm256_add_pd(fix3,tx);
2305 fiy3 = _mm256_add_pd(fiy3,ty);
2306 fiz3 = _mm256_add_pd(fiz3,tz);
2307
2308 fjx3 = _mm256_add_pd(fjx3,tx);
2309 fjy3 = _mm256_add_pd(fjy3,ty);
2310 fjz3 = _mm256_add_pd(fjz3,tz);
2311
2312 }
2313
2314 fjptrA = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
2315 fjptrB = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
2316 fjptrC = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
2317 fjptrD = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
2318
2319 gmx_mm256_decrement_4rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,
2320 fjx0,fjy0,fjz0,fjx1,fjy1,fjz1,
2321 fjx2,fjy2,fjz2,fjx3,fjy3,fjz3);
2322
2323 /* Inner loop uses 303 flops */
2324 }
2325
2326 /* End of innermost loop */
2327
2328 gmx_mm256_update_iforce_4atom_swizzle_pd(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
2329 f+i_coord_offset,fshift+i_shift_offset);
2330
2331 /* Increment number of inner iterations */
2332 inneriter += j_index_end - j_index_start;
2333
2334 /* Outer loop uses 24 flops */
2335 }
2336
2337 /* Increment number of outer iterations */
2338 outeriter += nri;
2339
2340 /* Update outer/inner flops */
2341
2342 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4W4_F,outeriter*24 + inneriter*303);
2343 }
The ultimate molecular dynamics simulation package