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_single kernel generator.
44 #include "../nb_kernel.h"
45 #include "types/simple.h"
49 #include "gromacs/simd/math_x86_avx_256_single.h"
50 #include "kernelutil_x86_avx_256_single.h"
53 * Gromacs nonbonded kernel: nb_kernel_ElecEw_VdwLJEw_GeomW4P1_VF_avx_256_single
54 * Electrostatics interaction: Ewald
55 * VdW interaction: LJEwald
56 * Geometry: Water4-Particle
57 * Calculate force/pot: PotentialAndForce
60 nb_kernel_ElecEw_VdwLJEw_GeomW4P1_VF_avx_256_single
61 (t_nblist * gmx_restrict nlist,
62 rvec * gmx_restrict xx,
63 rvec * gmx_restrict ff,
64 t_forcerec * gmx_restrict fr,
65 t_mdatoms * gmx_restrict mdatoms,
66 nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
67 t_nrnb * gmx_restrict nrnb)
69 /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
70 * just 0 for non-waters.
71 * Suffixes A,B,C,D,E,F,G,H refer to j loop unrolling done with AVX, e.g. for the eight different
72 * jnr indices corresponding to data put in the four positions in the SIMD register.
74 int i_shift_offset,i_coord_offset,outeriter,inneriter;
75 int j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
76 int jnrA,jnrB,jnrC,jnrD;
77 int jnrE,jnrF,jnrG,jnrH;
78 int jnrlistA,jnrlistB,jnrlistC,jnrlistD;
79 int jnrlistE,jnrlistF,jnrlistG,jnrlistH;
80 int j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
81 int j_coord_offsetE,j_coord_offsetF,j_coord_offsetG,j_coord_offsetH;
82 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
84 real *shiftvec,*fshift,*x,*f;
85 real *fjptrA,*fjptrB,*fjptrC,*fjptrD,*fjptrE,*fjptrF,*fjptrG,*fjptrH;
87 __m256 tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
88 real * vdwioffsetptr0;
89 real * vdwgridioffsetptr0;
90 __m256 ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
91 real * vdwioffsetptr1;
92 real * vdwgridioffsetptr1;
93 __m256 ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
94 real * vdwioffsetptr2;
95 real * vdwgridioffsetptr2;
96 __m256 ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
97 real * vdwioffsetptr3;
98 real * vdwgridioffsetptr3;
99 __m256 ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
100 int vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D,vdwjidx0E,vdwjidx0F,vdwjidx0G,vdwjidx0H;
101 __m256 jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
102 __m256 dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
103 __m256 dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
104 __m256 dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
105 __m256 dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30;
106 __m256 velec,felec,velecsum,facel,crf,krf,krf2;
109 __m256 rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
112 __m256 one_sixth = _mm256_set1_ps(1.0/6.0);
113 __m256 one_twelfth = _mm256_set1_ps(1.0/12.0);
119 __m256 ewclj,ewclj2,ewclj6,ewcljrsq,poly,exponent,f6A,f6B,sh_lj_ewald;
120 __m256 one_half = _mm256_set1_ps(0.5);
121 __m256 minus_one = _mm256_set1_ps(-1.0);
123 __m128i ewitab_lo,ewitab_hi;
124 __m256 ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
125 __m256 beta,beta2,beta3,zeta2,pmecorrF,pmecorrV,rinv3;
127 __m256 dummy_mask,cutoff_mask;
128 __m256 signbit = _mm256_castsi256_ps( _mm256_set1_epi32(0x80000000) );
129 __m256 one = _mm256_set1_ps(1.0);
130 __m256 two = _mm256_set1_ps(2.0);
136 jindex = nlist->jindex;
138 shiftidx = nlist->shift;
140 shiftvec = fr->shift_vec[0];
141 fshift = fr->fshift[0];
142 facel = _mm256_set1_ps(fr->epsfac);
143 charge = mdatoms->chargeA;
144 nvdwtype = fr->ntype;
146 vdwtype = mdatoms->typeA;
147 vdwgridparam = fr->ljpme_c6grid;
148 sh_lj_ewald = _mm256_set1_ps(fr->ic->sh_lj_ewald);
149 ewclj = _mm256_set1_ps(fr->ewaldcoeff_lj);
150 ewclj2 = _mm256_mul_ps(minus_one,_mm256_mul_ps(ewclj,ewclj));
152 sh_ewald = _mm256_set1_ps(fr->ic->sh_ewald);
153 beta = _mm256_set1_ps(fr->ic->ewaldcoeff_q);
154 beta2 = _mm256_mul_ps(beta,beta);
155 beta3 = _mm256_mul_ps(beta,beta2);
157 ewtab = fr->ic->tabq_coul_FDV0;
158 ewtabscale = _mm256_set1_ps(fr->ic->tabq_scale);
159 ewtabhalfspace = _mm256_set1_ps(0.5/fr->ic->tabq_scale);
161 /* Setup water-specific parameters */
162 inr = nlist->iinr[0];
163 iq1 = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+1]));
164 iq2 = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+2]));
165 iq3 = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+3]));
166 vdwioffsetptr0 = vdwparam+2*nvdwtype*vdwtype[inr+0];
167 vdwgridioffsetptr0 = vdwgridparam+2*nvdwtype*vdwtype[inr+0];
169 /* Avoid stupid compiler warnings */
170 jnrA = jnrB = jnrC = jnrD = jnrE = jnrF = jnrG = jnrH = 0;
183 for(iidx=0;iidx<4*DIM;iidx++)
188 /* Start outer loop over neighborlists */
189 for(iidx=0; iidx<nri; iidx++)
191 /* Load shift vector for this list */
192 i_shift_offset = DIM*shiftidx[iidx];
194 /* Load limits for loop over neighbors */
195 j_index_start = jindex[iidx];
196 j_index_end = jindex[iidx+1];
198 /* Get outer coordinate index */
200 i_coord_offset = DIM*inr;
202 /* Load i particle coords and add shift vector */
203 gmx_mm256_load_shift_and_4rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,
204 &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
206 fix0 = _mm256_setzero_ps();
207 fiy0 = _mm256_setzero_ps();
208 fiz0 = _mm256_setzero_ps();
209 fix1 = _mm256_setzero_ps();
210 fiy1 = _mm256_setzero_ps();
211 fiz1 = _mm256_setzero_ps();
212 fix2 = _mm256_setzero_ps();
213 fiy2 = _mm256_setzero_ps();
214 fiz2 = _mm256_setzero_ps();
215 fix3 = _mm256_setzero_ps();
216 fiy3 = _mm256_setzero_ps();
217 fiz3 = _mm256_setzero_ps();
219 /* Reset potential sums */
220 velecsum = _mm256_setzero_ps();
221 vvdwsum = _mm256_setzero_ps();
223 /* Start inner kernel loop */
224 for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+7]>=0; jidx+=8)
227 /* Get j neighbor index, and coordinate index */
236 j_coord_offsetA = DIM*jnrA;
237 j_coord_offsetB = DIM*jnrB;
238 j_coord_offsetC = DIM*jnrC;
239 j_coord_offsetD = DIM*jnrD;
240 j_coord_offsetE = DIM*jnrE;
241 j_coord_offsetF = DIM*jnrF;
242 j_coord_offsetG = DIM*jnrG;
243 j_coord_offsetH = DIM*jnrH;
245 /* load j atom coordinates */
246 gmx_mm256_load_1rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
247 x+j_coord_offsetC,x+j_coord_offsetD,
248 x+j_coord_offsetE,x+j_coord_offsetF,
249 x+j_coord_offsetG,x+j_coord_offsetH,
252 /* Calculate displacement vector */
253 dx00 = _mm256_sub_ps(ix0,jx0);
254 dy00 = _mm256_sub_ps(iy0,jy0);
255 dz00 = _mm256_sub_ps(iz0,jz0);
256 dx10 = _mm256_sub_ps(ix1,jx0);
257 dy10 = _mm256_sub_ps(iy1,jy0);
258 dz10 = _mm256_sub_ps(iz1,jz0);
259 dx20 = _mm256_sub_ps(ix2,jx0);
260 dy20 = _mm256_sub_ps(iy2,jy0);
261 dz20 = _mm256_sub_ps(iz2,jz0);
262 dx30 = _mm256_sub_ps(ix3,jx0);
263 dy30 = _mm256_sub_ps(iy3,jy0);
264 dz30 = _mm256_sub_ps(iz3,jz0);
266 /* Calculate squared distance and things based on it */
267 rsq00 = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
268 rsq10 = gmx_mm256_calc_rsq_ps(dx10,dy10,dz10);
269 rsq20 = gmx_mm256_calc_rsq_ps(dx20,dy20,dz20);
270 rsq30 = gmx_mm256_calc_rsq_ps(dx30,dy30,dz30);
272 rinv00 = gmx_mm256_invsqrt_ps(rsq00);
273 rinv10 = gmx_mm256_invsqrt_ps(rsq10);
274 rinv20 = gmx_mm256_invsqrt_ps(rsq20);
275 rinv30 = gmx_mm256_invsqrt_ps(rsq30);
277 rinvsq00 = _mm256_mul_ps(rinv00,rinv00);
278 rinvsq10 = _mm256_mul_ps(rinv10,rinv10);
279 rinvsq20 = _mm256_mul_ps(rinv20,rinv20);
280 rinvsq30 = _mm256_mul_ps(rinv30,rinv30);
282 /* Load parameters for j particles */
283 jq0 = gmx_mm256_load_8real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
284 charge+jnrC+0,charge+jnrD+0,
285 charge+jnrE+0,charge+jnrF+0,
286 charge+jnrG+0,charge+jnrH+0);
287 vdwjidx0A = 2*vdwtype[jnrA+0];
288 vdwjidx0B = 2*vdwtype[jnrB+0];
289 vdwjidx0C = 2*vdwtype[jnrC+0];
290 vdwjidx0D = 2*vdwtype[jnrD+0];
291 vdwjidx0E = 2*vdwtype[jnrE+0];
292 vdwjidx0F = 2*vdwtype[jnrF+0];
293 vdwjidx0G = 2*vdwtype[jnrG+0];
294 vdwjidx0H = 2*vdwtype[jnrH+0];
296 fjx0 = _mm256_setzero_ps();
297 fjy0 = _mm256_setzero_ps();
298 fjz0 = _mm256_setzero_ps();
300 /**************************
301 * CALCULATE INTERACTIONS *
302 **************************/
304 r00 = _mm256_mul_ps(rsq00,rinv00);
306 /* Compute parameters for interactions between i and j atoms */
307 gmx_mm256_load_8pair_swizzle_ps(vdwioffsetptr0+vdwjidx0A,
308 vdwioffsetptr0+vdwjidx0B,
309 vdwioffsetptr0+vdwjidx0C,
310 vdwioffsetptr0+vdwjidx0D,
311 vdwioffsetptr0+vdwjidx0E,
312 vdwioffsetptr0+vdwjidx0F,
313 vdwioffsetptr0+vdwjidx0G,
314 vdwioffsetptr0+vdwjidx0H,
317 c6grid_00 = gmx_mm256_load_8real_swizzle_ps(vdwgridioffsetptr0+vdwjidx0A,
318 vdwgridioffsetptr0+vdwjidx0B,
319 vdwgridioffsetptr0+vdwjidx0C,
320 vdwgridioffsetptr0+vdwjidx0D,
321 vdwgridioffsetptr0+vdwjidx0E,
322 vdwgridioffsetptr0+vdwjidx0F,
323 vdwgridioffsetptr0+vdwjidx0G,
324 vdwgridioffsetptr0+vdwjidx0H);
326 /* Analytical LJ-PME */
327 rinvsix = _mm256_mul_ps(_mm256_mul_ps(rinvsq00,rinvsq00),rinvsq00);
328 ewcljrsq = _mm256_mul_ps(ewclj2,rsq00);
329 ewclj6 = _mm256_mul_ps(ewclj2,_mm256_mul_ps(ewclj2,ewclj2));
330 exponent = gmx_simd_exp_r(ewcljrsq);
331 /* poly = exp(-(beta*r)^2) * (1 + (beta*r)^2 + (beta*r)^4 /2) */
332 poly = _mm256_mul_ps(exponent,_mm256_add_ps(_mm256_sub_ps(one,ewcljrsq),_mm256_mul_ps(_mm256_mul_ps(ewcljrsq,ewcljrsq),one_half)));
333 /* vvdw6 = [C6 - C6grid * (1-poly)]/r6 */
334 vvdw6 = _mm256_mul_ps(_mm256_sub_ps(c6_00,_mm256_mul_ps(c6grid_00,_mm256_sub_ps(one,poly))),rinvsix);
335 vvdw12 = _mm256_mul_ps(c12_00,_mm256_mul_ps(rinvsix,rinvsix));
336 vvdw = _mm256_sub_ps(_mm256_mul_ps(vvdw12,one_twelfth),_mm256_mul_ps(vvdw6,one_sixth));
337 /* fvdw = vvdw12/r - (vvdw6/r + (C6grid * exponent * beta^6)/r) */
338 fvdw = _mm256_mul_ps(_mm256_sub_ps(vvdw12,_mm256_sub_ps(vvdw6,_mm256_mul_ps(_mm256_mul_ps(c6grid_00,one_sixth),_mm256_mul_ps(exponent,ewclj6)))),rinvsq00);
340 /* Update potential sum for this i atom from the interaction with this j atom. */
341 vvdwsum = _mm256_add_ps(vvdwsum,vvdw);
345 /* Calculate temporary vectorial force */
346 tx = _mm256_mul_ps(fscal,dx00);
347 ty = _mm256_mul_ps(fscal,dy00);
348 tz = _mm256_mul_ps(fscal,dz00);
350 /* Update vectorial force */
351 fix0 = _mm256_add_ps(fix0,tx);
352 fiy0 = _mm256_add_ps(fiy0,ty);
353 fiz0 = _mm256_add_ps(fiz0,tz);
355 fjx0 = _mm256_add_ps(fjx0,tx);
356 fjy0 = _mm256_add_ps(fjy0,ty);
357 fjz0 = _mm256_add_ps(fjz0,tz);
359 /**************************
360 * CALCULATE INTERACTIONS *
361 **************************/
363 r10 = _mm256_mul_ps(rsq10,rinv10);
365 /* Compute parameters for interactions between i and j atoms */
366 qq10 = _mm256_mul_ps(iq1,jq0);
368 /* EWALD ELECTROSTATICS */
370 /* Analytical PME correction */
371 zeta2 = _mm256_mul_ps(beta2,rsq10);
372 rinv3 = _mm256_mul_ps(rinvsq10,rinv10);
373 pmecorrF = gmx_mm256_pmecorrF_ps(zeta2);
374 felec = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
375 felec = _mm256_mul_ps(qq10,felec);
376 pmecorrV = gmx_mm256_pmecorrV_ps(zeta2);
377 pmecorrV = _mm256_mul_ps(pmecorrV,beta);
378 velec = _mm256_sub_ps(rinv10,pmecorrV);
379 velec = _mm256_mul_ps(qq10,velec);
381 /* Update potential sum for this i atom from the interaction with this j atom. */
382 velecsum = _mm256_add_ps(velecsum,velec);
386 /* Calculate temporary vectorial force */
387 tx = _mm256_mul_ps(fscal,dx10);
388 ty = _mm256_mul_ps(fscal,dy10);
389 tz = _mm256_mul_ps(fscal,dz10);
391 /* Update vectorial force */
392 fix1 = _mm256_add_ps(fix1,tx);
393 fiy1 = _mm256_add_ps(fiy1,ty);
394 fiz1 = _mm256_add_ps(fiz1,tz);
396 fjx0 = _mm256_add_ps(fjx0,tx);
397 fjy0 = _mm256_add_ps(fjy0,ty);
398 fjz0 = _mm256_add_ps(fjz0,tz);
400 /**************************
401 * CALCULATE INTERACTIONS *
402 **************************/
404 r20 = _mm256_mul_ps(rsq20,rinv20);
406 /* Compute parameters for interactions between i and j atoms */
407 qq20 = _mm256_mul_ps(iq2,jq0);
409 /* EWALD ELECTROSTATICS */
411 /* Analytical PME correction */
412 zeta2 = _mm256_mul_ps(beta2,rsq20);
413 rinv3 = _mm256_mul_ps(rinvsq20,rinv20);
414 pmecorrF = gmx_mm256_pmecorrF_ps(zeta2);
415 felec = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
416 felec = _mm256_mul_ps(qq20,felec);
417 pmecorrV = gmx_mm256_pmecorrV_ps(zeta2);
418 pmecorrV = _mm256_mul_ps(pmecorrV,beta);
419 velec = _mm256_sub_ps(rinv20,pmecorrV);
420 velec = _mm256_mul_ps(qq20,velec);
422 /* Update potential sum for this i atom from the interaction with this j atom. */
423 velecsum = _mm256_add_ps(velecsum,velec);
427 /* Calculate temporary vectorial force */
428 tx = _mm256_mul_ps(fscal,dx20);
429 ty = _mm256_mul_ps(fscal,dy20);
430 tz = _mm256_mul_ps(fscal,dz20);
432 /* Update vectorial force */
433 fix2 = _mm256_add_ps(fix2,tx);
434 fiy2 = _mm256_add_ps(fiy2,ty);
435 fiz2 = _mm256_add_ps(fiz2,tz);
437 fjx0 = _mm256_add_ps(fjx0,tx);
438 fjy0 = _mm256_add_ps(fjy0,ty);
439 fjz0 = _mm256_add_ps(fjz0,tz);
441 /**************************
442 * CALCULATE INTERACTIONS *
443 **************************/
445 r30 = _mm256_mul_ps(rsq30,rinv30);
447 /* Compute parameters for interactions between i and j atoms */
448 qq30 = _mm256_mul_ps(iq3,jq0);
450 /* EWALD ELECTROSTATICS */
452 /* Analytical PME correction */
453 zeta2 = _mm256_mul_ps(beta2,rsq30);
454 rinv3 = _mm256_mul_ps(rinvsq30,rinv30);
455 pmecorrF = gmx_mm256_pmecorrF_ps(zeta2);
456 felec = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
457 felec = _mm256_mul_ps(qq30,felec);
458 pmecorrV = gmx_mm256_pmecorrV_ps(zeta2);
459 pmecorrV = _mm256_mul_ps(pmecorrV,beta);
460 velec = _mm256_sub_ps(rinv30,pmecorrV);
461 velec = _mm256_mul_ps(qq30,velec);
463 /* Update potential sum for this i atom from the interaction with this j atom. */
464 velecsum = _mm256_add_ps(velecsum,velec);
468 /* Calculate temporary vectorial force */
469 tx = _mm256_mul_ps(fscal,dx30);
470 ty = _mm256_mul_ps(fscal,dy30);
471 tz = _mm256_mul_ps(fscal,dz30);
473 /* Update vectorial force */
474 fix3 = _mm256_add_ps(fix3,tx);
475 fiy3 = _mm256_add_ps(fiy3,ty);
476 fiz3 = _mm256_add_ps(fiz3,tz);
478 fjx0 = _mm256_add_ps(fjx0,tx);
479 fjy0 = _mm256_add_ps(fjy0,ty);
480 fjz0 = _mm256_add_ps(fjz0,tz);
482 fjptrA = f+j_coord_offsetA;
483 fjptrB = f+j_coord_offsetB;
484 fjptrC = f+j_coord_offsetC;
485 fjptrD = f+j_coord_offsetD;
486 fjptrE = f+j_coord_offsetE;
487 fjptrF = f+j_coord_offsetF;
488 fjptrG = f+j_coord_offsetG;
489 fjptrH = f+j_coord_offsetH;
491 gmx_mm256_decrement_1rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,fjx0,fjy0,fjz0);
493 /* Inner loop uses 306 flops */
499 /* Get j neighbor index, and coordinate index */
500 jnrlistA = jjnr[jidx];
501 jnrlistB = jjnr[jidx+1];
502 jnrlistC = jjnr[jidx+2];
503 jnrlistD = jjnr[jidx+3];
504 jnrlistE = jjnr[jidx+4];
505 jnrlistF = jjnr[jidx+5];
506 jnrlistG = jjnr[jidx+6];
507 jnrlistH = jjnr[jidx+7];
508 /* Sign of each element will be negative for non-real atoms.
509 * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
510 * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
512 dummy_mask = gmx_mm256_set_m128(gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx+4)),_mm_setzero_si128())),
513 gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128())));
515 jnrA = (jnrlistA>=0) ? jnrlistA : 0;
516 jnrB = (jnrlistB>=0) ? jnrlistB : 0;
517 jnrC = (jnrlistC>=0) ? jnrlistC : 0;
518 jnrD = (jnrlistD>=0) ? jnrlistD : 0;
519 jnrE = (jnrlistE>=0) ? jnrlistE : 0;
520 jnrF = (jnrlistF>=0) ? jnrlistF : 0;
521 jnrG = (jnrlistG>=0) ? jnrlistG : 0;
522 jnrH = (jnrlistH>=0) ? jnrlistH : 0;
523 j_coord_offsetA = DIM*jnrA;
524 j_coord_offsetB = DIM*jnrB;
525 j_coord_offsetC = DIM*jnrC;
526 j_coord_offsetD = DIM*jnrD;
527 j_coord_offsetE = DIM*jnrE;
528 j_coord_offsetF = DIM*jnrF;
529 j_coord_offsetG = DIM*jnrG;
530 j_coord_offsetH = DIM*jnrH;
532 /* load j atom coordinates */
533 gmx_mm256_load_1rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
534 x+j_coord_offsetC,x+j_coord_offsetD,
535 x+j_coord_offsetE,x+j_coord_offsetF,
536 x+j_coord_offsetG,x+j_coord_offsetH,
539 /* Calculate displacement vector */
540 dx00 = _mm256_sub_ps(ix0,jx0);
541 dy00 = _mm256_sub_ps(iy0,jy0);
542 dz00 = _mm256_sub_ps(iz0,jz0);
543 dx10 = _mm256_sub_ps(ix1,jx0);
544 dy10 = _mm256_sub_ps(iy1,jy0);
545 dz10 = _mm256_sub_ps(iz1,jz0);
546 dx20 = _mm256_sub_ps(ix2,jx0);
547 dy20 = _mm256_sub_ps(iy2,jy0);
548 dz20 = _mm256_sub_ps(iz2,jz0);
549 dx30 = _mm256_sub_ps(ix3,jx0);
550 dy30 = _mm256_sub_ps(iy3,jy0);
551 dz30 = _mm256_sub_ps(iz3,jz0);
553 /* Calculate squared distance and things based on it */
554 rsq00 = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
555 rsq10 = gmx_mm256_calc_rsq_ps(dx10,dy10,dz10);
556 rsq20 = gmx_mm256_calc_rsq_ps(dx20,dy20,dz20);
557 rsq30 = gmx_mm256_calc_rsq_ps(dx30,dy30,dz30);
559 rinv00 = gmx_mm256_invsqrt_ps(rsq00);
560 rinv10 = gmx_mm256_invsqrt_ps(rsq10);
561 rinv20 = gmx_mm256_invsqrt_ps(rsq20);
562 rinv30 = gmx_mm256_invsqrt_ps(rsq30);
564 rinvsq00 = _mm256_mul_ps(rinv00,rinv00);
565 rinvsq10 = _mm256_mul_ps(rinv10,rinv10);
566 rinvsq20 = _mm256_mul_ps(rinv20,rinv20);
567 rinvsq30 = _mm256_mul_ps(rinv30,rinv30);
569 /* Load parameters for j particles */
570 jq0 = gmx_mm256_load_8real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
571 charge+jnrC+0,charge+jnrD+0,
572 charge+jnrE+0,charge+jnrF+0,
573 charge+jnrG+0,charge+jnrH+0);
574 vdwjidx0A = 2*vdwtype[jnrA+0];
575 vdwjidx0B = 2*vdwtype[jnrB+0];
576 vdwjidx0C = 2*vdwtype[jnrC+0];
577 vdwjidx0D = 2*vdwtype[jnrD+0];
578 vdwjidx0E = 2*vdwtype[jnrE+0];
579 vdwjidx0F = 2*vdwtype[jnrF+0];
580 vdwjidx0G = 2*vdwtype[jnrG+0];
581 vdwjidx0H = 2*vdwtype[jnrH+0];
583 fjx0 = _mm256_setzero_ps();
584 fjy0 = _mm256_setzero_ps();
585 fjz0 = _mm256_setzero_ps();
587 /**************************
588 * CALCULATE INTERACTIONS *
589 **************************/
591 r00 = _mm256_mul_ps(rsq00,rinv00);
592 r00 = _mm256_andnot_ps(dummy_mask,r00);
594 /* Compute parameters for interactions between i and j atoms */
595 gmx_mm256_load_8pair_swizzle_ps(vdwioffsetptr0+vdwjidx0A,
596 vdwioffsetptr0+vdwjidx0B,
597 vdwioffsetptr0+vdwjidx0C,
598 vdwioffsetptr0+vdwjidx0D,
599 vdwioffsetptr0+vdwjidx0E,
600 vdwioffsetptr0+vdwjidx0F,
601 vdwioffsetptr0+vdwjidx0G,
602 vdwioffsetptr0+vdwjidx0H,
605 c6grid_00 = gmx_mm256_load_8real_swizzle_ps(vdwgridioffsetptr0+vdwjidx0A,
606 vdwgridioffsetptr0+vdwjidx0B,
607 vdwgridioffsetptr0+vdwjidx0C,
608 vdwgridioffsetptr0+vdwjidx0D,
609 vdwgridioffsetptr0+vdwjidx0E,
610 vdwgridioffsetptr0+vdwjidx0F,
611 vdwgridioffsetptr0+vdwjidx0G,
612 vdwgridioffsetptr0+vdwjidx0H);
614 /* Analytical LJ-PME */
615 rinvsix = _mm256_mul_ps(_mm256_mul_ps(rinvsq00,rinvsq00),rinvsq00);
616 ewcljrsq = _mm256_mul_ps(ewclj2,rsq00);
617 ewclj6 = _mm256_mul_ps(ewclj2,_mm256_mul_ps(ewclj2,ewclj2));
618 exponent = gmx_simd_exp_r(ewcljrsq);
619 /* poly = exp(-(beta*r)^2) * (1 + (beta*r)^2 + (beta*r)^4 /2) */
620 poly = _mm256_mul_ps(exponent,_mm256_add_ps(_mm256_sub_ps(one,ewcljrsq),_mm256_mul_ps(_mm256_mul_ps(ewcljrsq,ewcljrsq),one_half)));
621 /* vvdw6 = [C6 - C6grid * (1-poly)]/r6 */
622 vvdw6 = _mm256_mul_ps(_mm256_sub_ps(c6_00,_mm256_mul_ps(c6grid_00,_mm256_sub_ps(one,poly))),rinvsix);
623 vvdw12 = _mm256_mul_ps(c12_00,_mm256_mul_ps(rinvsix,rinvsix));
624 vvdw = _mm256_sub_ps(_mm256_mul_ps(vvdw12,one_twelfth),_mm256_mul_ps(vvdw6,one_sixth));
625 /* fvdw = vvdw12/r - (vvdw6/r + (C6grid * exponent * beta^6)/r) */
626 fvdw = _mm256_mul_ps(_mm256_sub_ps(vvdw12,_mm256_sub_ps(vvdw6,_mm256_mul_ps(_mm256_mul_ps(c6grid_00,one_sixth),_mm256_mul_ps(exponent,ewclj6)))),rinvsq00);
628 /* Update potential sum for this i atom from the interaction with this j atom. */
629 vvdw = _mm256_andnot_ps(dummy_mask,vvdw);
630 vvdwsum = _mm256_add_ps(vvdwsum,vvdw);
634 fscal = _mm256_andnot_ps(dummy_mask,fscal);
636 /* Calculate temporary vectorial force */
637 tx = _mm256_mul_ps(fscal,dx00);
638 ty = _mm256_mul_ps(fscal,dy00);
639 tz = _mm256_mul_ps(fscal,dz00);
641 /* Update vectorial force */
642 fix0 = _mm256_add_ps(fix0,tx);
643 fiy0 = _mm256_add_ps(fiy0,ty);
644 fiz0 = _mm256_add_ps(fiz0,tz);
646 fjx0 = _mm256_add_ps(fjx0,tx);
647 fjy0 = _mm256_add_ps(fjy0,ty);
648 fjz0 = _mm256_add_ps(fjz0,tz);
650 /**************************
651 * CALCULATE INTERACTIONS *
652 **************************/
654 r10 = _mm256_mul_ps(rsq10,rinv10);
655 r10 = _mm256_andnot_ps(dummy_mask,r10);
657 /* Compute parameters for interactions between i and j atoms */
658 qq10 = _mm256_mul_ps(iq1,jq0);
660 /* EWALD ELECTROSTATICS */
662 /* Analytical PME correction */
663 zeta2 = _mm256_mul_ps(beta2,rsq10);
664 rinv3 = _mm256_mul_ps(rinvsq10,rinv10);
665 pmecorrF = gmx_mm256_pmecorrF_ps(zeta2);
666 felec = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
667 felec = _mm256_mul_ps(qq10,felec);
668 pmecorrV = gmx_mm256_pmecorrV_ps(zeta2);
669 pmecorrV = _mm256_mul_ps(pmecorrV,beta);
670 velec = _mm256_sub_ps(rinv10,pmecorrV);
671 velec = _mm256_mul_ps(qq10,velec);
673 /* Update potential sum for this i atom from the interaction with this j atom. */
674 velec = _mm256_andnot_ps(dummy_mask,velec);
675 velecsum = _mm256_add_ps(velecsum,velec);
679 fscal = _mm256_andnot_ps(dummy_mask,fscal);
681 /* Calculate temporary vectorial force */
682 tx = _mm256_mul_ps(fscal,dx10);
683 ty = _mm256_mul_ps(fscal,dy10);
684 tz = _mm256_mul_ps(fscal,dz10);
686 /* Update vectorial force */
687 fix1 = _mm256_add_ps(fix1,tx);
688 fiy1 = _mm256_add_ps(fiy1,ty);
689 fiz1 = _mm256_add_ps(fiz1,tz);
691 fjx0 = _mm256_add_ps(fjx0,tx);
692 fjy0 = _mm256_add_ps(fjy0,ty);
693 fjz0 = _mm256_add_ps(fjz0,tz);
695 /**************************
696 * CALCULATE INTERACTIONS *
697 **************************/
699 r20 = _mm256_mul_ps(rsq20,rinv20);
700 r20 = _mm256_andnot_ps(dummy_mask,r20);
702 /* Compute parameters for interactions between i and j atoms */
703 qq20 = _mm256_mul_ps(iq2,jq0);
705 /* EWALD ELECTROSTATICS */
707 /* Analytical PME correction */
708 zeta2 = _mm256_mul_ps(beta2,rsq20);
709 rinv3 = _mm256_mul_ps(rinvsq20,rinv20);
710 pmecorrF = gmx_mm256_pmecorrF_ps(zeta2);
711 felec = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
712 felec = _mm256_mul_ps(qq20,felec);
713 pmecorrV = gmx_mm256_pmecorrV_ps(zeta2);
714 pmecorrV = _mm256_mul_ps(pmecorrV,beta);
715 velec = _mm256_sub_ps(rinv20,pmecorrV);
716 velec = _mm256_mul_ps(qq20,velec);
718 /* Update potential sum for this i atom from the interaction with this j atom. */
719 velec = _mm256_andnot_ps(dummy_mask,velec);
720 velecsum = _mm256_add_ps(velecsum,velec);
724 fscal = _mm256_andnot_ps(dummy_mask,fscal);
726 /* Calculate temporary vectorial force */
727 tx = _mm256_mul_ps(fscal,dx20);
728 ty = _mm256_mul_ps(fscal,dy20);
729 tz = _mm256_mul_ps(fscal,dz20);
731 /* Update vectorial force */
732 fix2 = _mm256_add_ps(fix2,tx);
733 fiy2 = _mm256_add_ps(fiy2,ty);
734 fiz2 = _mm256_add_ps(fiz2,tz);
736 fjx0 = _mm256_add_ps(fjx0,tx);
737 fjy0 = _mm256_add_ps(fjy0,ty);
738 fjz0 = _mm256_add_ps(fjz0,tz);
740 /**************************
741 * CALCULATE INTERACTIONS *
742 **************************/
744 r30 = _mm256_mul_ps(rsq30,rinv30);
745 r30 = _mm256_andnot_ps(dummy_mask,r30);
747 /* Compute parameters for interactions between i and j atoms */
748 qq30 = _mm256_mul_ps(iq3,jq0);
750 /* EWALD ELECTROSTATICS */
752 /* Analytical PME correction */
753 zeta2 = _mm256_mul_ps(beta2,rsq30);
754 rinv3 = _mm256_mul_ps(rinvsq30,rinv30);
755 pmecorrF = gmx_mm256_pmecorrF_ps(zeta2);
756 felec = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
757 felec = _mm256_mul_ps(qq30,felec);
758 pmecorrV = gmx_mm256_pmecorrV_ps(zeta2);
759 pmecorrV = _mm256_mul_ps(pmecorrV,beta);
760 velec = _mm256_sub_ps(rinv30,pmecorrV);
761 velec = _mm256_mul_ps(qq30,velec);
763 /* Update potential sum for this i atom from the interaction with this j atom. */
764 velec = _mm256_andnot_ps(dummy_mask,velec);
765 velecsum = _mm256_add_ps(velecsum,velec);
769 fscal = _mm256_andnot_ps(dummy_mask,fscal);
771 /* Calculate temporary vectorial force */
772 tx = _mm256_mul_ps(fscal,dx30);
773 ty = _mm256_mul_ps(fscal,dy30);
774 tz = _mm256_mul_ps(fscal,dz30);
776 /* Update vectorial force */
777 fix3 = _mm256_add_ps(fix3,tx);
778 fiy3 = _mm256_add_ps(fiy3,ty);
779 fiz3 = _mm256_add_ps(fiz3,tz);
781 fjx0 = _mm256_add_ps(fjx0,tx);
782 fjy0 = _mm256_add_ps(fjy0,ty);
783 fjz0 = _mm256_add_ps(fjz0,tz);
785 fjptrA = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
786 fjptrB = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
787 fjptrC = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
788 fjptrD = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
789 fjptrE = (jnrlistE>=0) ? f+j_coord_offsetE : scratch;
790 fjptrF = (jnrlistF>=0) ? f+j_coord_offsetF : scratch;
791 fjptrG = (jnrlistG>=0) ? f+j_coord_offsetG : scratch;
792 fjptrH = (jnrlistH>=0) ? f+j_coord_offsetH : scratch;
794 gmx_mm256_decrement_1rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,fjx0,fjy0,fjz0);
796 /* Inner loop uses 310 flops */
799 /* End of innermost loop */
801 gmx_mm256_update_iforce_4atom_swizzle_ps(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
802 f+i_coord_offset,fshift+i_shift_offset);
805 /* Update potential energies */
806 gmx_mm256_update_1pot_ps(velecsum,kernel_data->energygrp_elec+ggid);
807 gmx_mm256_update_1pot_ps(vvdwsum,kernel_data->energygrp_vdw+ggid);
809 /* Increment number of inner iterations */
810 inneriter += j_index_end - j_index_start;
812 /* Outer loop uses 26 flops */
815 /* Increment number of outer iterations */
818 /* Update outer/inner flops */
820 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_VF,outeriter*26 + inneriter*310);
823 * Gromacs nonbonded kernel: nb_kernel_ElecEw_VdwLJEw_GeomW4P1_F_avx_256_single
824 * Electrostatics interaction: Ewald
825 * VdW interaction: LJEwald
826 * Geometry: Water4-Particle
827 * Calculate force/pot: Force
830 nb_kernel_ElecEw_VdwLJEw_GeomW4P1_F_avx_256_single
831 (t_nblist * gmx_restrict nlist,
832 rvec * gmx_restrict xx,
833 rvec * gmx_restrict ff,
834 t_forcerec * gmx_restrict fr,
835 t_mdatoms * gmx_restrict mdatoms,
836 nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
837 t_nrnb * gmx_restrict nrnb)
839 /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
840 * just 0 for non-waters.
841 * Suffixes A,B,C,D,E,F,G,H refer to j loop unrolling done with AVX, e.g. for the eight different
842 * jnr indices corresponding to data put in the four positions in the SIMD register.
844 int i_shift_offset,i_coord_offset,outeriter,inneriter;
845 int j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
846 int jnrA,jnrB,jnrC,jnrD;
847 int jnrE,jnrF,jnrG,jnrH;
848 int jnrlistA,jnrlistB,jnrlistC,jnrlistD;
849 int jnrlistE,jnrlistF,jnrlistG,jnrlistH;
850 int j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
851 int j_coord_offsetE,j_coord_offsetF,j_coord_offsetG,j_coord_offsetH;
852 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
854 real *shiftvec,*fshift,*x,*f;
855 real *fjptrA,*fjptrB,*fjptrC,*fjptrD,*fjptrE,*fjptrF,*fjptrG,*fjptrH;
857 __m256 tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
858 real * vdwioffsetptr0;
859 real * vdwgridioffsetptr0;
860 __m256 ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
861 real * vdwioffsetptr1;
862 real * vdwgridioffsetptr1;
863 __m256 ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
864 real * vdwioffsetptr2;
865 real * vdwgridioffsetptr2;
866 __m256 ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
867 real * vdwioffsetptr3;
868 real * vdwgridioffsetptr3;
869 __m256 ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
870 int vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D,vdwjidx0E,vdwjidx0F,vdwjidx0G,vdwjidx0H;
871 __m256 jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
872 __m256 dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
873 __m256 dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
874 __m256 dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
875 __m256 dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30;
876 __m256 velec,felec,velecsum,facel,crf,krf,krf2;
879 __m256 rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
882 __m256 one_sixth = _mm256_set1_ps(1.0/6.0);
883 __m256 one_twelfth = _mm256_set1_ps(1.0/12.0);
889 __m256 ewclj,ewclj2,ewclj6,ewcljrsq,poly,exponent,f6A,f6B,sh_lj_ewald;
890 __m256 one_half = _mm256_set1_ps(0.5);
891 __m256 minus_one = _mm256_set1_ps(-1.0);
893 __m128i ewitab_lo,ewitab_hi;
894 __m256 ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
895 __m256 beta,beta2,beta3,zeta2,pmecorrF,pmecorrV,rinv3;
897 __m256 dummy_mask,cutoff_mask;
898 __m256 signbit = _mm256_castsi256_ps( _mm256_set1_epi32(0x80000000) );
899 __m256 one = _mm256_set1_ps(1.0);
900 __m256 two = _mm256_set1_ps(2.0);
906 jindex = nlist->jindex;
908 shiftidx = nlist->shift;
910 shiftvec = fr->shift_vec[0];
911 fshift = fr->fshift[0];
912 facel = _mm256_set1_ps(fr->epsfac);
913 charge = mdatoms->chargeA;
914 nvdwtype = fr->ntype;
916 vdwtype = mdatoms->typeA;
917 vdwgridparam = fr->ljpme_c6grid;
918 sh_lj_ewald = _mm256_set1_ps(fr->ic->sh_lj_ewald);
919 ewclj = _mm256_set1_ps(fr->ewaldcoeff_lj);
920 ewclj2 = _mm256_mul_ps(minus_one,_mm256_mul_ps(ewclj,ewclj));
922 sh_ewald = _mm256_set1_ps(fr->ic->sh_ewald);
923 beta = _mm256_set1_ps(fr->ic->ewaldcoeff_q);
924 beta2 = _mm256_mul_ps(beta,beta);
925 beta3 = _mm256_mul_ps(beta,beta2);
927 ewtab = fr->ic->tabq_coul_F;
928 ewtabscale = _mm256_set1_ps(fr->ic->tabq_scale);
929 ewtabhalfspace = _mm256_set1_ps(0.5/fr->ic->tabq_scale);
931 /* Setup water-specific parameters */
932 inr = nlist->iinr[0];
933 iq1 = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+1]));
934 iq2 = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+2]));
935 iq3 = _mm256_mul_ps(facel,_mm256_set1_ps(charge[inr+3]));
936 vdwioffsetptr0 = vdwparam+2*nvdwtype*vdwtype[inr+0];
937 vdwgridioffsetptr0 = vdwgridparam+2*nvdwtype*vdwtype[inr+0];
939 /* Avoid stupid compiler warnings */
940 jnrA = jnrB = jnrC = jnrD = jnrE = jnrF = jnrG = jnrH = 0;
953 for(iidx=0;iidx<4*DIM;iidx++)
958 /* Start outer loop over neighborlists */
959 for(iidx=0; iidx<nri; iidx++)
961 /* Load shift vector for this list */
962 i_shift_offset = DIM*shiftidx[iidx];
964 /* Load limits for loop over neighbors */
965 j_index_start = jindex[iidx];
966 j_index_end = jindex[iidx+1];
968 /* Get outer coordinate index */
970 i_coord_offset = DIM*inr;
972 /* Load i particle coords and add shift vector */
973 gmx_mm256_load_shift_and_4rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,
974 &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
976 fix0 = _mm256_setzero_ps();
977 fiy0 = _mm256_setzero_ps();
978 fiz0 = _mm256_setzero_ps();
979 fix1 = _mm256_setzero_ps();
980 fiy1 = _mm256_setzero_ps();
981 fiz1 = _mm256_setzero_ps();
982 fix2 = _mm256_setzero_ps();
983 fiy2 = _mm256_setzero_ps();
984 fiz2 = _mm256_setzero_ps();
985 fix3 = _mm256_setzero_ps();
986 fiy3 = _mm256_setzero_ps();
987 fiz3 = _mm256_setzero_ps();
989 /* Start inner kernel loop */
990 for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+7]>=0; jidx+=8)
993 /* Get j neighbor index, and coordinate index */
1000 jnrG = jjnr[jidx+6];
1001 jnrH = jjnr[jidx+7];
1002 j_coord_offsetA = DIM*jnrA;
1003 j_coord_offsetB = DIM*jnrB;
1004 j_coord_offsetC = DIM*jnrC;
1005 j_coord_offsetD = DIM*jnrD;
1006 j_coord_offsetE = DIM*jnrE;
1007 j_coord_offsetF = DIM*jnrF;
1008 j_coord_offsetG = DIM*jnrG;
1009 j_coord_offsetH = DIM*jnrH;
1011 /* load j atom coordinates */
1012 gmx_mm256_load_1rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
1013 x+j_coord_offsetC,x+j_coord_offsetD,
1014 x+j_coord_offsetE,x+j_coord_offsetF,
1015 x+j_coord_offsetG,x+j_coord_offsetH,
1018 /* Calculate displacement vector */
1019 dx00 = _mm256_sub_ps(ix0,jx0);
1020 dy00 = _mm256_sub_ps(iy0,jy0);
1021 dz00 = _mm256_sub_ps(iz0,jz0);
1022 dx10 = _mm256_sub_ps(ix1,jx0);
1023 dy10 = _mm256_sub_ps(iy1,jy0);
1024 dz10 = _mm256_sub_ps(iz1,jz0);
1025 dx20 = _mm256_sub_ps(ix2,jx0);
1026 dy20 = _mm256_sub_ps(iy2,jy0);
1027 dz20 = _mm256_sub_ps(iz2,jz0);
1028 dx30 = _mm256_sub_ps(ix3,jx0);
1029 dy30 = _mm256_sub_ps(iy3,jy0);
1030 dz30 = _mm256_sub_ps(iz3,jz0);
1032 /* Calculate squared distance and things based on it */
1033 rsq00 = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
1034 rsq10 = gmx_mm256_calc_rsq_ps(dx10,dy10,dz10);
1035 rsq20 = gmx_mm256_calc_rsq_ps(dx20,dy20,dz20);
1036 rsq30 = gmx_mm256_calc_rsq_ps(dx30,dy30,dz30);
1038 rinv00 = gmx_mm256_invsqrt_ps(rsq00);
1039 rinv10 = gmx_mm256_invsqrt_ps(rsq10);
1040 rinv20 = gmx_mm256_invsqrt_ps(rsq20);
1041 rinv30 = gmx_mm256_invsqrt_ps(rsq30);
1043 rinvsq00 = _mm256_mul_ps(rinv00,rinv00);
1044 rinvsq10 = _mm256_mul_ps(rinv10,rinv10);
1045 rinvsq20 = _mm256_mul_ps(rinv20,rinv20);
1046 rinvsq30 = _mm256_mul_ps(rinv30,rinv30);
1048 /* Load parameters for j particles */
1049 jq0 = gmx_mm256_load_8real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
1050 charge+jnrC+0,charge+jnrD+0,
1051 charge+jnrE+0,charge+jnrF+0,
1052 charge+jnrG+0,charge+jnrH+0);
1053 vdwjidx0A = 2*vdwtype[jnrA+0];
1054 vdwjidx0B = 2*vdwtype[jnrB+0];
1055 vdwjidx0C = 2*vdwtype[jnrC+0];
1056 vdwjidx0D = 2*vdwtype[jnrD+0];
1057 vdwjidx0E = 2*vdwtype[jnrE+0];
1058 vdwjidx0F = 2*vdwtype[jnrF+0];
1059 vdwjidx0G = 2*vdwtype[jnrG+0];
1060 vdwjidx0H = 2*vdwtype[jnrH+0];
1062 fjx0 = _mm256_setzero_ps();
1063 fjy0 = _mm256_setzero_ps();
1064 fjz0 = _mm256_setzero_ps();
1066 /**************************
1067 * CALCULATE INTERACTIONS *
1068 **************************/
1070 r00 = _mm256_mul_ps(rsq00,rinv00);
1072 /* Compute parameters for interactions between i and j atoms */
1073 gmx_mm256_load_8pair_swizzle_ps(vdwioffsetptr0+vdwjidx0A,
1074 vdwioffsetptr0+vdwjidx0B,
1075 vdwioffsetptr0+vdwjidx0C,
1076 vdwioffsetptr0+vdwjidx0D,
1077 vdwioffsetptr0+vdwjidx0E,
1078 vdwioffsetptr0+vdwjidx0F,
1079 vdwioffsetptr0+vdwjidx0G,
1080 vdwioffsetptr0+vdwjidx0H,
1083 c6grid_00 = gmx_mm256_load_8real_swizzle_ps(vdwgridioffsetptr0+vdwjidx0A,
1084 vdwgridioffsetptr0+vdwjidx0B,
1085 vdwgridioffsetptr0+vdwjidx0C,
1086 vdwgridioffsetptr0+vdwjidx0D,
1087 vdwgridioffsetptr0+vdwjidx0E,
1088 vdwgridioffsetptr0+vdwjidx0F,
1089 vdwgridioffsetptr0+vdwjidx0G,
1090 vdwgridioffsetptr0+vdwjidx0H);
1092 /* Analytical LJ-PME */
1093 rinvsix = _mm256_mul_ps(_mm256_mul_ps(rinvsq00,rinvsq00),rinvsq00);
1094 ewcljrsq = _mm256_mul_ps(ewclj2,rsq00);
1095 ewclj6 = _mm256_mul_ps(ewclj2,_mm256_mul_ps(ewclj2,ewclj2));
1096 exponent = gmx_simd_exp_r(ewcljrsq);
1097 /* poly = exp(-(beta*r)^2) * (1 + (beta*r)^2 + (beta*r)^4 /2) */
1098 poly = _mm256_mul_ps(exponent,_mm256_add_ps(_mm256_sub_ps(one,ewcljrsq),_mm256_mul_ps(_mm256_mul_ps(ewcljrsq,ewcljrsq),one_half)));
1099 /* f6A = 6 * C6grid * (1 - poly) */
1100 f6A = _mm256_mul_ps(c6grid_00,_mm256_sub_ps(one,poly));
1101 /* f6B = C6grid * exponent * beta^6 */
1102 f6B = _mm256_mul_ps(_mm256_mul_ps(c6grid_00,one_sixth),_mm256_mul_ps(exponent,ewclj6));
1103 /* fvdw = 12*C12/r13 - ((6*C6 - f6A)/r6 + f6B)/r */
1104 fvdw = _mm256_mul_ps(_mm256_add_ps(_mm256_mul_ps(_mm256_sub_ps(_mm256_mul_ps(c12_00,rinvsix),_mm256_sub_ps(c6_00,f6A)),rinvsix),f6B),rinvsq00);
1108 /* Calculate temporary vectorial force */
1109 tx = _mm256_mul_ps(fscal,dx00);
1110 ty = _mm256_mul_ps(fscal,dy00);
1111 tz = _mm256_mul_ps(fscal,dz00);
1113 /* Update vectorial force */
1114 fix0 = _mm256_add_ps(fix0,tx);
1115 fiy0 = _mm256_add_ps(fiy0,ty);
1116 fiz0 = _mm256_add_ps(fiz0,tz);
1118 fjx0 = _mm256_add_ps(fjx0,tx);
1119 fjy0 = _mm256_add_ps(fjy0,ty);
1120 fjz0 = _mm256_add_ps(fjz0,tz);
1122 /**************************
1123 * CALCULATE INTERACTIONS *
1124 **************************/
1126 r10 = _mm256_mul_ps(rsq10,rinv10);
1128 /* Compute parameters for interactions between i and j atoms */
1129 qq10 = _mm256_mul_ps(iq1,jq0);
1131 /* EWALD ELECTROSTATICS */
1133 /* Analytical PME correction */
1134 zeta2 = _mm256_mul_ps(beta2,rsq10);
1135 rinv3 = _mm256_mul_ps(rinvsq10,rinv10);
1136 pmecorrF = gmx_mm256_pmecorrF_ps(zeta2);
1137 felec = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1138 felec = _mm256_mul_ps(qq10,felec);
1142 /* Calculate temporary vectorial force */
1143 tx = _mm256_mul_ps(fscal,dx10);
1144 ty = _mm256_mul_ps(fscal,dy10);
1145 tz = _mm256_mul_ps(fscal,dz10);
1147 /* Update vectorial force */
1148 fix1 = _mm256_add_ps(fix1,tx);
1149 fiy1 = _mm256_add_ps(fiy1,ty);
1150 fiz1 = _mm256_add_ps(fiz1,tz);
1152 fjx0 = _mm256_add_ps(fjx0,tx);
1153 fjy0 = _mm256_add_ps(fjy0,ty);
1154 fjz0 = _mm256_add_ps(fjz0,tz);
1156 /**************************
1157 * CALCULATE INTERACTIONS *
1158 **************************/
1160 r20 = _mm256_mul_ps(rsq20,rinv20);
1162 /* Compute parameters for interactions between i and j atoms */
1163 qq20 = _mm256_mul_ps(iq2,jq0);
1165 /* EWALD ELECTROSTATICS */
1167 /* Analytical PME correction */
1168 zeta2 = _mm256_mul_ps(beta2,rsq20);
1169 rinv3 = _mm256_mul_ps(rinvsq20,rinv20);
1170 pmecorrF = gmx_mm256_pmecorrF_ps(zeta2);
1171 felec = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1172 felec = _mm256_mul_ps(qq20,felec);
1176 /* Calculate temporary vectorial force */
1177 tx = _mm256_mul_ps(fscal,dx20);
1178 ty = _mm256_mul_ps(fscal,dy20);
1179 tz = _mm256_mul_ps(fscal,dz20);
1181 /* Update vectorial force */
1182 fix2 = _mm256_add_ps(fix2,tx);
1183 fiy2 = _mm256_add_ps(fiy2,ty);
1184 fiz2 = _mm256_add_ps(fiz2,tz);
1186 fjx0 = _mm256_add_ps(fjx0,tx);
1187 fjy0 = _mm256_add_ps(fjy0,ty);
1188 fjz0 = _mm256_add_ps(fjz0,tz);
1190 /**************************
1191 * CALCULATE INTERACTIONS *
1192 **************************/
1194 r30 = _mm256_mul_ps(rsq30,rinv30);
1196 /* Compute parameters for interactions between i and j atoms */
1197 qq30 = _mm256_mul_ps(iq3,jq0);
1199 /* EWALD ELECTROSTATICS */
1201 /* Analytical PME correction */
1202 zeta2 = _mm256_mul_ps(beta2,rsq30);
1203 rinv3 = _mm256_mul_ps(rinvsq30,rinv30);
1204 pmecorrF = gmx_mm256_pmecorrF_ps(zeta2);
1205 felec = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1206 felec = _mm256_mul_ps(qq30,felec);
1210 /* Calculate temporary vectorial force */
1211 tx = _mm256_mul_ps(fscal,dx30);
1212 ty = _mm256_mul_ps(fscal,dy30);
1213 tz = _mm256_mul_ps(fscal,dz30);
1215 /* Update vectorial force */
1216 fix3 = _mm256_add_ps(fix3,tx);
1217 fiy3 = _mm256_add_ps(fiy3,ty);
1218 fiz3 = _mm256_add_ps(fiz3,tz);
1220 fjx0 = _mm256_add_ps(fjx0,tx);
1221 fjy0 = _mm256_add_ps(fjy0,ty);
1222 fjz0 = _mm256_add_ps(fjz0,tz);
1224 fjptrA = f+j_coord_offsetA;
1225 fjptrB = f+j_coord_offsetB;
1226 fjptrC = f+j_coord_offsetC;
1227 fjptrD = f+j_coord_offsetD;
1228 fjptrE = f+j_coord_offsetE;
1229 fjptrF = f+j_coord_offsetF;
1230 fjptrG = f+j_coord_offsetG;
1231 fjptrH = f+j_coord_offsetH;
1233 gmx_mm256_decrement_1rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,fjx0,fjy0,fjz0);
1235 /* Inner loop uses 217 flops */
1238 if(jidx<j_index_end)
1241 /* Get j neighbor index, and coordinate index */
1242 jnrlistA = jjnr[jidx];
1243 jnrlistB = jjnr[jidx+1];
1244 jnrlistC = jjnr[jidx+2];
1245 jnrlistD = jjnr[jidx+3];
1246 jnrlistE = jjnr[jidx+4];
1247 jnrlistF = jjnr[jidx+5];
1248 jnrlistG = jjnr[jidx+6];
1249 jnrlistH = jjnr[jidx+7];
1250 /* Sign of each element will be negative for non-real atoms.
1251 * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
1252 * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
1254 dummy_mask = gmx_mm256_set_m128(gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx+4)),_mm_setzero_si128())),
1255 gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128())));
1257 jnrA = (jnrlistA>=0) ? jnrlistA : 0;
1258 jnrB = (jnrlistB>=0) ? jnrlistB : 0;
1259 jnrC = (jnrlistC>=0) ? jnrlistC : 0;
1260 jnrD = (jnrlistD>=0) ? jnrlistD : 0;
1261 jnrE = (jnrlistE>=0) ? jnrlistE : 0;
1262 jnrF = (jnrlistF>=0) ? jnrlistF : 0;
1263 jnrG = (jnrlistG>=0) ? jnrlistG : 0;
1264 jnrH = (jnrlistH>=0) ? jnrlistH : 0;
1265 j_coord_offsetA = DIM*jnrA;
1266 j_coord_offsetB = DIM*jnrB;
1267 j_coord_offsetC = DIM*jnrC;
1268 j_coord_offsetD = DIM*jnrD;
1269 j_coord_offsetE = DIM*jnrE;
1270 j_coord_offsetF = DIM*jnrF;
1271 j_coord_offsetG = DIM*jnrG;
1272 j_coord_offsetH = DIM*jnrH;
1274 /* load j atom coordinates */
1275 gmx_mm256_load_1rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
1276 x+j_coord_offsetC,x+j_coord_offsetD,
1277 x+j_coord_offsetE,x+j_coord_offsetF,
1278 x+j_coord_offsetG,x+j_coord_offsetH,
1281 /* Calculate displacement vector */
1282 dx00 = _mm256_sub_ps(ix0,jx0);
1283 dy00 = _mm256_sub_ps(iy0,jy0);
1284 dz00 = _mm256_sub_ps(iz0,jz0);
1285 dx10 = _mm256_sub_ps(ix1,jx0);
1286 dy10 = _mm256_sub_ps(iy1,jy0);
1287 dz10 = _mm256_sub_ps(iz1,jz0);
1288 dx20 = _mm256_sub_ps(ix2,jx0);
1289 dy20 = _mm256_sub_ps(iy2,jy0);
1290 dz20 = _mm256_sub_ps(iz2,jz0);
1291 dx30 = _mm256_sub_ps(ix3,jx0);
1292 dy30 = _mm256_sub_ps(iy3,jy0);
1293 dz30 = _mm256_sub_ps(iz3,jz0);
1295 /* Calculate squared distance and things based on it */
1296 rsq00 = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
1297 rsq10 = gmx_mm256_calc_rsq_ps(dx10,dy10,dz10);
1298 rsq20 = gmx_mm256_calc_rsq_ps(dx20,dy20,dz20);
1299 rsq30 = gmx_mm256_calc_rsq_ps(dx30,dy30,dz30);
1301 rinv00 = gmx_mm256_invsqrt_ps(rsq00);
1302 rinv10 = gmx_mm256_invsqrt_ps(rsq10);
1303 rinv20 = gmx_mm256_invsqrt_ps(rsq20);
1304 rinv30 = gmx_mm256_invsqrt_ps(rsq30);
1306 rinvsq00 = _mm256_mul_ps(rinv00,rinv00);
1307 rinvsq10 = _mm256_mul_ps(rinv10,rinv10);
1308 rinvsq20 = _mm256_mul_ps(rinv20,rinv20);
1309 rinvsq30 = _mm256_mul_ps(rinv30,rinv30);
1311 /* Load parameters for j particles */
1312 jq0 = gmx_mm256_load_8real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
1313 charge+jnrC+0,charge+jnrD+0,
1314 charge+jnrE+0,charge+jnrF+0,
1315 charge+jnrG+0,charge+jnrH+0);
1316 vdwjidx0A = 2*vdwtype[jnrA+0];
1317 vdwjidx0B = 2*vdwtype[jnrB+0];
1318 vdwjidx0C = 2*vdwtype[jnrC+0];
1319 vdwjidx0D = 2*vdwtype[jnrD+0];
1320 vdwjidx0E = 2*vdwtype[jnrE+0];
1321 vdwjidx0F = 2*vdwtype[jnrF+0];
1322 vdwjidx0G = 2*vdwtype[jnrG+0];
1323 vdwjidx0H = 2*vdwtype[jnrH+0];
1325 fjx0 = _mm256_setzero_ps();
1326 fjy0 = _mm256_setzero_ps();
1327 fjz0 = _mm256_setzero_ps();
1329 /**************************
1330 * CALCULATE INTERACTIONS *
1331 **************************/
1333 r00 = _mm256_mul_ps(rsq00,rinv00);
1334 r00 = _mm256_andnot_ps(dummy_mask,r00);
1336 /* Compute parameters for interactions between i and j atoms */
1337 gmx_mm256_load_8pair_swizzle_ps(vdwioffsetptr0+vdwjidx0A,
1338 vdwioffsetptr0+vdwjidx0B,
1339 vdwioffsetptr0+vdwjidx0C,
1340 vdwioffsetptr0+vdwjidx0D,
1341 vdwioffsetptr0+vdwjidx0E,
1342 vdwioffsetptr0+vdwjidx0F,
1343 vdwioffsetptr0+vdwjidx0G,
1344 vdwioffsetptr0+vdwjidx0H,
1347 c6grid_00 = gmx_mm256_load_8real_swizzle_ps(vdwgridioffsetptr0+vdwjidx0A,
1348 vdwgridioffsetptr0+vdwjidx0B,
1349 vdwgridioffsetptr0+vdwjidx0C,
1350 vdwgridioffsetptr0+vdwjidx0D,
1351 vdwgridioffsetptr0+vdwjidx0E,
1352 vdwgridioffsetptr0+vdwjidx0F,
1353 vdwgridioffsetptr0+vdwjidx0G,
1354 vdwgridioffsetptr0+vdwjidx0H);
1356 /* Analytical LJ-PME */
1357 rinvsix = _mm256_mul_ps(_mm256_mul_ps(rinvsq00,rinvsq00),rinvsq00);
1358 ewcljrsq = _mm256_mul_ps(ewclj2,rsq00);
1359 ewclj6 = _mm256_mul_ps(ewclj2,_mm256_mul_ps(ewclj2,ewclj2));
1360 exponent = gmx_simd_exp_r(ewcljrsq);
1361 /* poly = exp(-(beta*r)^2) * (1 + (beta*r)^2 + (beta*r)^4 /2) */
1362 poly = _mm256_mul_ps(exponent,_mm256_add_ps(_mm256_sub_ps(one,ewcljrsq),_mm256_mul_ps(_mm256_mul_ps(ewcljrsq,ewcljrsq),one_half)));
1363 /* f6A = 6 * C6grid * (1 - poly) */
1364 f6A = _mm256_mul_ps(c6grid_00,_mm256_sub_ps(one,poly));
1365 /* f6B = C6grid * exponent * beta^6 */
1366 f6B = _mm256_mul_ps(_mm256_mul_ps(c6grid_00,one_sixth),_mm256_mul_ps(exponent,ewclj6));
1367 /* fvdw = 12*C12/r13 - ((6*C6 - f6A)/r6 + f6B)/r */
1368 fvdw = _mm256_mul_ps(_mm256_add_ps(_mm256_mul_ps(_mm256_sub_ps(_mm256_mul_ps(c12_00,rinvsix),_mm256_sub_ps(c6_00,f6A)),rinvsix),f6B),rinvsq00);
1372 fscal = _mm256_andnot_ps(dummy_mask,fscal);
1374 /* Calculate temporary vectorial force */
1375 tx = _mm256_mul_ps(fscal,dx00);
1376 ty = _mm256_mul_ps(fscal,dy00);
1377 tz = _mm256_mul_ps(fscal,dz00);
1379 /* Update vectorial force */
1380 fix0 = _mm256_add_ps(fix0,tx);
1381 fiy0 = _mm256_add_ps(fiy0,ty);
1382 fiz0 = _mm256_add_ps(fiz0,tz);
1384 fjx0 = _mm256_add_ps(fjx0,tx);
1385 fjy0 = _mm256_add_ps(fjy0,ty);
1386 fjz0 = _mm256_add_ps(fjz0,tz);
1388 /**************************
1389 * CALCULATE INTERACTIONS *
1390 **************************/
1392 r10 = _mm256_mul_ps(rsq10,rinv10);
1393 r10 = _mm256_andnot_ps(dummy_mask,r10);
1395 /* Compute parameters for interactions between i and j atoms */
1396 qq10 = _mm256_mul_ps(iq1,jq0);
1398 /* EWALD ELECTROSTATICS */
1400 /* Analytical PME correction */
1401 zeta2 = _mm256_mul_ps(beta2,rsq10);
1402 rinv3 = _mm256_mul_ps(rinvsq10,rinv10);
1403 pmecorrF = gmx_mm256_pmecorrF_ps(zeta2);
1404 felec = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1405 felec = _mm256_mul_ps(qq10,felec);
1409 fscal = _mm256_andnot_ps(dummy_mask,fscal);
1411 /* Calculate temporary vectorial force */
1412 tx = _mm256_mul_ps(fscal,dx10);
1413 ty = _mm256_mul_ps(fscal,dy10);
1414 tz = _mm256_mul_ps(fscal,dz10);
1416 /* Update vectorial force */
1417 fix1 = _mm256_add_ps(fix1,tx);
1418 fiy1 = _mm256_add_ps(fiy1,ty);
1419 fiz1 = _mm256_add_ps(fiz1,tz);
1421 fjx0 = _mm256_add_ps(fjx0,tx);
1422 fjy0 = _mm256_add_ps(fjy0,ty);
1423 fjz0 = _mm256_add_ps(fjz0,tz);
1425 /**************************
1426 * CALCULATE INTERACTIONS *
1427 **************************/
1429 r20 = _mm256_mul_ps(rsq20,rinv20);
1430 r20 = _mm256_andnot_ps(dummy_mask,r20);
1432 /* Compute parameters for interactions between i and j atoms */
1433 qq20 = _mm256_mul_ps(iq2,jq0);
1435 /* EWALD ELECTROSTATICS */
1437 /* Analytical PME correction */
1438 zeta2 = _mm256_mul_ps(beta2,rsq20);
1439 rinv3 = _mm256_mul_ps(rinvsq20,rinv20);
1440 pmecorrF = gmx_mm256_pmecorrF_ps(zeta2);
1441 felec = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1442 felec = _mm256_mul_ps(qq20,felec);
1446 fscal = _mm256_andnot_ps(dummy_mask,fscal);
1448 /* Calculate temporary vectorial force */
1449 tx = _mm256_mul_ps(fscal,dx20);
1450 ty = _mm256_mul_ps(fscal,dy20);
1451 tz = _mm256_mul_ps(fscal,dz20);
1453 /* Update vectorial force */
1454 fix2 = _mm256_add_ps(fix2,tx);
1455 fiy2 = _mm256_add_ps(fiy2,ty);
1456 fiz2 = _mm256_add_ps(fiz2,tz);
1458 fjx0 = _mm256_add_ps(fjx0,tx);
1459 fjy0 = _mm256_add_ps(fjy0,ty);
1460 fjz0 = _mm256_add_ps(fjz0,tz);
1462 /**************************
1463 * CALCULATE INTERACTIONS *
1464 **************************/
1466 r30 = _mm256_mul_ps(rsq30,rinv30);
1467 r30 = _mm256_andnot_ps(dummy_mask,r30);
1469 /* Compute parameters for interactions between i and j atoms */
1470 qq30 = _mm256_mul_ps(iq3,jq0);
1472 /* EWALD ELECTROSTATICS */
1474 /* Analytical PME correction */
1475 zeta2 = _mm256_mul_ps(beta2,rsq30);
1476 rinv3 = _mm256_mul_ps(rinvsq30,rinv30);
1477 pmecorrF = gmx_mm256_pmecorrF_ps(zeta2);
1478 felec = _mm256_add_ps( _mm256_mul_ps(pmecorrF,beta3), rinv3);
1479 felec = _mm256_mul_ps(qq30,felec);
1483 fscal = _mm256_andnot_ps(dummy_mask,fscal);
1485 /* Calculate temporary vectorial force */
1486 tx = _mm256_mul_ps(fscal,dx30);
1487 ty = _mm256_mul_ps(fscal,dy30);
1488 tz = _mm256_mul_ps(fscal,dz30);
1490 /* Update vectorial force */
1491 fix3 = _mm256_add_ps(fix3,tx);
1492 fiy3 = _mm256_add_ps(fiy3,ty);
1493 fiz3 = _mm256_add_ps(fiz3,tz);
1495 fjx0 = _mm256_add_ps(fjx0,tx);
1496 fjy0 = _mm256_add_ps(fjy0,ty);
1497 fjz0 = _mm256_add_ps(fjz0,tz);
1499 fjptrA = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
1500 fjptrB = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
1501 fjptrC = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
1502 fjptrD = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
1503 fjptrE = (jnrlistE>=0) ? f+j_coord_offsetE : scratch;
1504 fjptrF = (jnrlistF>=0) ? f+j_coord_offsetF : scratch;
1505 fjptrG = (jnrlistG>=0) ? f+j_coord_offsetG : scratch;
1506 fjptrH = (jnrlistH>=0) ? f+j_coord_offsetH : scratch;
1508 gmx_mm256_decrement_1rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,fjx0,fjy0,fjz0);
1510 /* Inner loop uses 221 flops */
1513 /* End of innermost loop */
1515 gmx_mm256_update_iforce_4atom_swizzle_ps(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
1516 f+i_coord_offset,fshift+i_shift_offset);
1518 /* Increment number of inner iterations */
1519 inneriter += j_index_end - j_index_start;
1521 /* Outer loop uses 24 flops */
1524 /* Increment number of outer iterations */
1527 /* Update outer/inner flops */
1529 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_F,outeriter*24 + inneriter*221);