2 * This file is part of the GROMACS molecular simulation package.
4 * Copyright (c) 2012,2013,2014, by the GROMACS development team, led by
5 * Mark Abraham, David van der Spoel, Berk Hess, and Erik Lindahl,
6 * and including many others, as listed in the AUTHORS file in the
7 * top-level source directory and at http://www.gromacs.org.
9 * GROMACS is free software; you can redistribute it and/or
10 * modify it under the terms of the GNU Lesser General Public License
11 * as published by the Free Software Foundation; either version 2.1
12 * of the License, or (at your option) any later version.
14 * GROMACS is distributed in the hope that it will be useful,
15 * but WITHOUT ANY WARRANTY; without even the implied warranty of
16 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
17 * Lesser General Public License for more details.
19 * You should have received a copy of the GNU Lesser General Public
20 * License along with GROMACS; if not, see
21 * http://www.gnu.org/licenses, or write to the Free Software Foundation,
22 * Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
24 * If you want to redistribute modifications to GROMACS, please
25 * consider that scientific software is very special. Version
26 * control is crucial - bugs must be traceable. We will be happy to
27 * consider code for inclusion in the official distribution, but
28 * derived work must not be called official GROMACS. Details are found
29 * in the README & COPYING files - if they are missing, get the
30 * official version at http://www.gromacs.org.
32 * To help us fund GROMACS development, we humbly ask that you cite
33 * the research papers on the package. Check out http://www.gromacs.org.
36 * Note: this file was generated by the GROMACS avx_256_double kernel generator.
42 #include "../nb_kernel.h"
43 #include "types/simple.h"
44 #include "gromacs/math/vec.h"
47 #include "gromacs/simd/math_x86_avx_256_double.h"
48 #include "kernelutil_x86_avx_256_double.h"
51 * Gromacs nonbonded kernel: nb_kernel_ElecEw_VdwNone_GeomW4P1_VF_avx_256_double
52 * Electrostatics interaction: Ewald
53 * VdW interaction: None
54 * Geometry: Water4-Particle
55 * Calculate force/pot: PotentialAndForce
58 nb_kernel_ElecEw_VdwNone_GeomW4P1_VF_avx_256_double
59 (t_nblist * gmx_restrict nlist,
60 rvec * gmx_restrict xx,
61 rvec * gmx_restrict ff,
62 t_forcerec * gmx_restrict fr,
63 t_mdatoms * gmx_restrict mdatoms,
64 nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
65 t_nrnb * gmx_restrict nrnb)
67 /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
68 * just 0 for non-waters.
69 * Suffixes A,B,C,D refer to j loop unrolling done with AVX, e.g. for the four different
70 * jnr indices corresponding to data put in the four positions in the SIMD register.
72 int i_shift_offset,i_coord_offset,outeriter,inneriter;
73 int j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
74 int jnrA,jnrB,jnrC,jnrD;
75 int jnrlistA,jnrlistB,jnrlistC,jnrlistD;
76 int jnrlistE,jnrlistF,jnrlistG,jnrlistH;
77 int j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
78 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
80 real *shiftvec,*fshift,*x,*f;
81 real *fjptrA,*fjptrB,*fjptrC,*fjptrD;
83 __m256d tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
84 real * vdwioffsetptr1;
85 __m256d ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
86 real * vdwioffsetptr2;
87 __m256d ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
88 real * vdwioffsetptr3;
89 __m256d ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
90 int vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D;
91 __m256d jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
92 __m256d dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
93 __m256d dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
94 __m256d dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30;
95 __m256d velec,felec,velecsum,facel,crf,krf,krf2;
98 __m256d ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
99 __m256d beta,beta2,beta3,zeta2,pmecorrF,pmecorrV,rinv3;
101 __m256d dummy_mask,cutoff_mask;
102 __m128 tmpmask0,tmpmask1;
103 __m256d signbit = _mm256_castsi256_pd( _mm256_set1_epi32(0x80000000) );
104 __m256d one = _mm256_set1_pd(1.0);
105 __m256d two = _mm256_set1_pd(2.0);
111 jindex = nlist->jindex;
113 shiftidx = nlist->shift;
115 shiftvec = fr->shift_vec[0];
116 fshift = fr->fshift[0];
117 facel = _mm256_set1_pd(fr->epsfac);
118 charge = mdatoms->chargeA;
120 sh_ewald = _mm256_set1_pd(fr->ic->sh_ewald);
121 beta = _mm256_set1_pd(fr->ic->ewaldcoeff_q);
122 beta2 = _mm256_mul_pd(beta,beta);
123 beta3 = _mm256_mul_pd(beta,beta2);
125 ewtab = fr->ic->tabq_coul_FDV0;
126 ewtabscale = _mm256_set1_pd(fr->ic->tabq_scale);
127 ewtabhalfspace = _mm256_set1_pd(0.5/fr->ic->tabq_scale);
129 /* Setup water-specific parameters */
130 inr = nlist->iinr[0];
131 iq1 = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+1]));
132 iq2 = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+2]));
133 iq3 = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+3]));
135 /* Avoid stupid compiler warnings */
136 jnrA = jnrB = jnrC = jnrD = 0;
145 for(iidx=0;iidx<4*DIM;iidx++)
150 /* Start outer loop over neighborlists */
151 for(iidx=0; iidx<nri; iidx++)
153 /* Load shift vector for this list */
154 i_shift_offset = DIM*shiftidx[iidx];
156 /* Load limits for loop over neighbors */
157 j_index_start = jindex[iidx];
158 j_index_end = jindex[iidx+1];
160 /* Get outer coordinate index */
162 i_coord_offset = DIM*inr;
164 /* Load i particle coords and add shift vector */
165 gmx_mm256_load_shift_and_3rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset+DIM,
166 &ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
168 fix1 = _mm256_setzero_pd();
169 fiy1 = _mm256_setzero_pd();
170 fiz1 = _mm256_setzero_pd();
171 fix2 = _mm256_setzero_pd();
172 fiy2 = _mm256_setzero_pd();
173 fiz2 = _mm256_setzero_pd();
174 fix3 = _mm256_setzero_pd();
175 fiy3 = _mm256_setzero_pd();
176 fiz3 = _mm256_setzero_pd();
178 /* Reset potential sums */
179 velecsum = _mm256_setzero_pd();
181 /* Start inner kernel loop */
182 for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4)
185 /* Get j neighbor index, and coordinate index */
190 j_coord_offsetA = DIM*jnrA;
191 j_coord_offsetB = DIM*jnrB;
192 j_coord_offsetC = DIM*jnrC;
193 j_coord_offsetD = DIM*jnrD;
195 /* load j atom coordinates */
196 gmx_mm256_load_1rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
197 x+j_coord_offsetC,x+j_coord_offsetD,
200 /* Calculate displacement vector */
201 dx10 = _mm256_sub_pd(ix1,jx0);
202 dy10 = _mm256_sub_pd(iy1,jy0);
203 dz10 = _mm256_sub_pd(iz1,jz0);
204 dx20 = _mm256_sub_pd(ix2,jx0);
205 dy20 = _mm256_sub_pd(iy2,jy0);
206 dz20 = _mm256_sub_pd(iz2,jz0);
207 dx30 = _mm256_sub_pd(ix3,jx0);
208 dy30 = _mm256_sub_pd(iy3,jy0);
209 dz30 = _mm256_sub_pd(iz3,jz0);
211 /* Calculate squared distance and things based on it */
212 rsq10 = gmx_mm256_calc_rsq_pd(dx10,dy10,dz10);
213 rsq20 = gmx_mm256_calc_rsq_pd(dx20,dy20,dz20);
214 rsq30 = gmx_mm256_calc_rsq_pd(dx30,dy30,dz30);
216 rinv10 = gmx_mm256_invsqrt_pd(rsq10);
217 rinv20 = gmx_mm256_invsqrt_pd(rsq20);
218 rinv30 = gmx_mm256_invsqrt_pd(rsq30);
220 rinvsq10 = _mm256_mul_pd(rinv10,rinv10);
221 rinvsq20 = _mm256_mul_pd(rinv20,rinv20);
222 rinvsq30 = _mm256_mul_pd(rinv30,rinv30);
224 /* Load parameters for j particles */
225 jq0 = gmx_mm256_load_4real_swizzle_pd(charge+jnrA+0,charge+jnrB+0,
226 charge+jnrC+0,charge+jnrD+0);
228 fjx0 = _mm256_setzero_pd();
229 fjy0 = _mm256_setzero_pd();
230 fjz0 = _mm256_setzero_pd();
232 /**************************
233 * CALCULATE INTERACTIONS *
234 **************************/
236 r10 = _mm256_mul_pd(rsq10,rinv10);
238 /* Compute parameters for interactions between i and j atoms */
239 qq10 = _mm256_mul_pd(iq1,jq0);
241 /* EWALD ELECTROSTATICS */
243 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
244 ewrt = _mm256_mul_pd(r10,ewtabscale);
245 ewitab = _mm256_cvttpd_epi32(ewrt);
246 eweps = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
247 ewitab = _mm_slli_epi32(ewitab,2);
248 ewtabF = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
249 ewtabD = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
250 ewtabV = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,2) );
251 ewtabFn = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,3) );
252 GMX_MM256_FULLTRANSPOSE4_PD(ewtabF,ewtabD,ewtabV,ewtabFn);
253 felec = _mm256_add_pd(ewtabF,_mm256_mul_pd(eweps,ewtabD));
254 velec = _mm256_sub_pd(ewtabV,_mm256_mul_pd(_mm256_mul_pd(ewtabhalfspace,eweps),_mm256_add_pd(ewtabF,felec)));
255 velec = _mm256_mul_pd(qq10,_mm256_sub_pd(rinv10,velec));
256 felec = _mm256_mul_pd(_mm256_mul_pd(qq10,rinv10),_mm256_sub_pd(rinvsq10,felec));
258 /* Update potential sum for this i atom from the interaction with this j atom. */
259 velecsum = _mm256_add_pd(velecsum,velec);
263 /* Calculate temporary vectorial force */
264 tx = _mm256_mul_pd(fscal,dx10);
265 ty = _mm256_mul_pd(fscal,dy10);
266 tz = _mm256_mul_pd(fscal,dz10);
268 /* Update vectorial force */
269 fix1 = _mm256_add_pd(fix1,tx);
270 fiy1 = _mm256_add_pd(fiy1,ty);
271 fiz1 = _mm256_add_pd(fiz1,tz);
273 fjx0 = _mm256_add_pd(fjx0,tx);
274 fjy0 = _mm256_add_pd(fjy0,ty);
275 fjz0 = _mm256_add_pd(fjz0,tz);
277 /**************************
278 * CALCULATE INTERACTIONS *
279 **************************/
281 r20 = _mm256_mul_pd(rsq20,rinv20);
283 /* Compute parameters for interactions between i and j atoms */
284 qq20 = _mm256_mul_pd(iq2,jq0);
286 /* EWALD ELECTROSTATICS */
288 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
289 ewrt = _mm256_mul_pd(r20,ewtabscale);
290 ewitab = _mm256_cvttpd_epi32(ewrt);
291 eweps = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
292 ewitab = _mm_slli_epi32(ewitab,2);
293 ewtabF = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
294 ewtabD = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
295 ewtabV = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,2) );
296 ewtabFn = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,3) );
297 GMX_MM256_FULLTRANSPOSE4_PD(ewtabF,ewtabD,ewtabV,ewtabFn);
298 felec = _mm256_add_pd(ewtabF,_mm256_mul_pd(eweps,ewtabD));
299 velec = _mm256_sub_pd(ewtabV,_mm256_mul_pd(_mm256_mul_pd(ewtabhalfspace,eweps),_mm256_add_pd(ewtabF,felec)));
300 velec = _mm256_mul_pd(qq20,_mm256_sub_pd(rinv20,velec));
301 felec = _mm256_mul_pd(_mm256_mul_pd(qq20,rinv20),_mm256_sub_pd(rinvsq20,felec));
303 /* Update potential sum for this i atom from the interaction with this j atom. */
304 velecsum = _mm256_add_pd(velecsum,velec);
308 /* Calculate temporary vectorial force */
309 tx = _mm256_mul_pd(fscal,dx20);
310 ty = _mm256_mul_pd(fscal,dy20);
311 tz = _mm256_mul_pd(fscal,dz20);
313 /* Update vectorial force */
314 fix2 = _mm256_add_pd(fix2,tx);
315 fiy2 = _mm256_add_pd(fiy2,ty);
316 fiz2 = _mm256_add_pd(fiz2,tz);
318 fjx0 = _mm256_add_pd(fjx0,tx);
319 fjy0 = _mm256_add_pd(fjy0,ty);
320 fjz0 = _mm256_add_pd(fjz0,tz);
322 /**************************
323 * CALCULATE INTERACTIONS *
324 **************************/
326 r30 = _mm256_mul_pd(rsq30,rinv30);
328 /* Compute parameters for interactions between i and j atoms */
329 qq30 = _mm256_mul_pd(iq3,jq0);
331 /* EWALD ELECTROSTATICS */
333 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
334 ewrt = _mm256_mul_pd(r30,ewtabscale);
335 ewitab = _mm256_cvttpd_epi32(ewrt);
336 eweps = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
337 ewitab = _mm_slli_epi32(ewitab,2);
338 ewtabF = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
339 ewtabD = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
340 ewtabV = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,2) );
341 ewtabFn = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,3) );
342 GMX_MM256_FULLTRANSPOSE4_PD(ewtabF,ewtabD,ewtabV,ewtabFn);
343 felec = _mm256_add_pd(ewtabF,_mm256_mul_pd(eweps,ewtabD));
344 velec = _mm256_sub_pd(ewtabV,_mm256_mul_pd(_mm256_mul_pd(ewtabhalfspace,eweps),_mm256_add_pd(ewtabF,felec)));
345 velec = _mm256_mul_pd(qq30,_mm256_sub_pd(rinv30,velec));
346 felec = _mm256_mul_pd(_mm256_mul_pd(qq30,rinv30),_mm256_sub_pd(rinvsq30,felec));
348 /* Update potential sum for this i atom from the interaction with this j atom. */
349 velecsum = _mm256_add_pd(velecsum,velec);
353 /* Calculate temporary vectorial force */
354 tx = _mm256_mul_pd(fscal,dx30);
355 ty = _mm256_mul_pd(fscal,dy30);
356 tz = _mm256_mul_pd(fscal,dz30);
358 /* Update vectorial force */
359 fix3 = _mm256_add_pd(fix3,tx);
360 fiy3 = _mm256_add_pd(fiy3,ty);
361 fiz3 = _mm256_add_pd(fiz3,tz);
363 fjx0 = _mm256_add_pd(fjx0,tx);
364 fjy0 = _mm256_add_pd(fjy0,ty);
365 fjz0 = _mm256_add_pd(fjz0,tz);
367 fjptrA = f+j_coord_offsetA;
368 fjptrB = f+j_coord_offsetB;
369 fjptrC = f+j_coord_offsetC;
370 fjptrD = f+j_coord_offsetD;
372 gmx_mm256_decrement_1rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0);
374 /* Inner loop uses 126 flops */
380 /* Get j neighbor index, and coordinate index */
381 jnrlistA = jjnr[jidx];
382 jnrlistB = jjnr[jidx+1];
383 jnrlistC = jjnr[jidx+2];
384 jnrlistD = jjnr[jidx+3];
385 /* Sign of each element will be negative for non-real atoms.
386 * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
387 * so use it as val = _mm_andnot_pd(mask,val) to clear dummy entries.
389 tmpmask0 = gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128()));
391 tmpmask1 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(3,3,2,2));
392 tmpmask0 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(1,1,0,0));
393 dummy_mask = _mm256_castps_pd(gmx_mm256_set_m128(tmpmask1,tmpmask0));
395 jnrA = (jnrlistA>=0) ? jnrlistA : 0;
396 jnrB = (jnrlistB>=0) ? jnrlistB : 0;
397 jnrC = (jnrlistC>=0) ? jnrlistC : 0;
398 jnrD = (jnrlistD>=0) ? jnrlistD : 0;
399 j_coord_offsetA = DIM*jnrA;
400 j_coord_offsetB = DIM*jnrB;
401 j_coord_offsetC = DIM*jnrC;
402 j_coord_offsetD = DIM*jnrD;
404 /* load j atom coordinates */
405 gmx_mm256_load_1rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
406 x+j_coord_offsetC,x+j_coord_offsetD,
409 /* Calculate displacement vector */
410 dx10 = _mm256_sub_pd(ix1,jx0);
411 dy10 = _mm256_sub_pd(iy1,jy0);
412 dz10 = _mm256_sub_pd(iz1,jz0);
413 dx20 = _mm256_sub_pd(ix2,jx0);
414 dy20 = _mm256_sub_pd(iy2,jy0);
415 dz20 = _mm256_sub_pd(iz2,jz0);
416 dx30 = _mm256_sub_pd(ix3,jx0);
417 dy30 = _mm256_sub_pd(iy3,jy0);
418 dz30 = _mm256_sub_pd(iz3,jz0);
420 /* Calculate squared distance and things based on it */
421 rsq10 = gmx_mm256_calc_rsq_pd(dx10,dy10,dz10);
422 rsq20 = gmx_mm256_calc_rsq_pd(dx20,dy20,dz20);
423 rsq30 = gmx_mm256_calc_rsq_pd(dx30,dy30,dz30);
425 rinv10 = gmx_mm256_invsqrt_pd(rsq10);
426 rinv20 = gmx_mm256_invsqrt_pd(rsq20);
427 rinv30 = gmx_mm256_invsqrt_pd(rsq30);
429 rinvsq10 = _mm256_mul_pd(rinv10,rinv10);
430 rinvsq20 = _mm256_mul_pd(rinv20,rinv20);
431 rinvsq30 = _mm256_mul_pd(rinv30,rinv30);
433 /* Load parameters for j particles */
434 jq0 = gmx_mm256_load_4real_swizzle_pd(charge+jnrA+0,charge+jnrB+0,
435 charge+jnrC+0,charge+jnrD+0);
437 fjx0 = _mm256_setzero_pd();
438 fjy0 = _mm256_setzero_pd();
439 fjz0 = _mm256_setzero_pd();
441 /**************************
442 * CALCULATE INTERACTIONS *
443 **************************/
445 r10 = _mm256_mul_pd(rsq10,rinv10);
446 r10 = _mm256_andnot_pd(dummy_mask,r10);
448 /* Compute parameters for interactions between i and j atoms */
449 qq10 = _mm256_mul_pd(iq1,jq0);
451 /* EWALD ELECTROSTATICS */
453 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
454 ewrt = _mm256_mul_pd(r10,ewtabscale);
455 ewitab = _mm256_cvttpd_epi32(ewrt);
456 eweps = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
457 ewitab = _mm_slli_epi32(ewitab,2);
458 ewtabF = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
459 ewtabD = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
460 ewtabV = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,2) );
461 ewtabFn = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,3) );
462 GMX_MM256_FULLTRANSPOSE4_PD(ewtabF,ewtabD,ewtabV,ewtabFn);
463 felec = _mm256_add_pd(ewtabF,_mm256_mul_pd(eweps,ewtabD));
464 velec = _mm256_sub_pd(ewtabV,_mm256_mul_pd(_mm256_mul_pd(ewtabhalfspace,eweps),_mm256_add_pd(ewtabF,felec)));
465 velec = _mm256_mul_pd(qq10,_mm256_sub_pd(rinv10,velec));
466 felec = _mm256_mul_pd(_mm256_mul_pd(qq10,rinv10),_mm256_sub_pd(rinvsq10,felec));
468 /* Update potential sum for this i atom from the interaction with this j atom. */
469 velec = _mm256_andnot_pd(dummy_mask,velec);
470 velecsum = _mm256_add_pd(velecsum,velec);
474 fscal = _mm256_andnot_pd(dummy_mask,fscal);
476 /* Calculate temporary vectorial force */
477 tx = _mm256_mul_pd(fscal,dx10);
478 ty = _mm256_mul_pd(fscal,dy10);
479 tz = _mm256_mul_pd(fscal,dz10);
481 /* Update vectorial force */
482 fix1 = _mm256_add_pd(fix1,tx);
483 fiy1 = _mm256_add_pd(fiy1,ty);
484 fiz1 = _mm256_add_pd(fiz1,tz);
486 fjx0 = _mm256_add_pd(fjx0,tx);
487 fjy0 = _mm256_add_pd(fjy0,ty);
488 fjz0 = _mm256_add_pd(fjz0,tz);
490 /**************************
491 * CALCULATE INTERACTIONS *
492 **************************/
494 r20 = _mm256_mul_pd(rsq20,rinv20);
495 r20 = _mm256_andnot_pd(dummy_mask,r20);
497 /* Compute parameters for interactions between i and j atoms */
498 qq20 = _mm256_mul_pd(iq2,jq0);
500 /* EWALD ELECTROSTATICS */
502 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
503 ewrt = _mm256_mul_pd(r20,ewtabscale);
504 ewitab = _mm256_cvttpd_epi32(ewrt);
505 eweps = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
506 ewitab = _mm_slli_epi32(ewitab,2);
507 ewtabF = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
508 ewtabD = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
509 ewtabV = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,2) );
510 ewtabFn = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,3) );
511 GMX_MM256_FULLTRANSPOSE4_PD(ewtabF,ewtabD,ewtabV,ewtabFn);
512 felec = _mm256_add_pd(ewtabF,_mm256_mul_pd(eweps,ewtabD));
513 velec = _mm256_sub_pd(ewtabV,_mm256_mul_pd(_mm256_mul_pd(ewtabhalfspace,eweps),_mm256_add_pd(ewtabF,felec)));
514 velec = _mm256_mul_pd(qq20,_mm256_sub_pd(rinv20,velec));
515 felec = _mm256_mul_pd(_mm256_mul_pd(qq20,rinv20),_mm256_sub_pd(rinvsq20,felec));
517 /* Update potential sum for this i atom from the interaction with this j atom. */
518 velec = _mm256_andnot_pd(dummy_mask,velec);
519 velecsum = _mm256_add_pd(velecsum,velec);
523 fscal = _mm256_andnot_pd(dummy_mask,fscal);
525 /* Calculate temporary vectorial force */
526 tx = _mm256_mul_pd(fscal,dx20);
527 ty = _mm256_mul_pd(fscal,dy20);
528 tz = _mm256_mul_pd(fscal,dz20);
530 /* Update vectorial force */
531 fix2 = _mm256_add_pd(fix2,tx);
532 fiy2 = _mm256_add_pd(fiy2,ty);
533 fiz2 = _mm256_add_pd(fiz2,tz);
535 fjx0 = _mm256_add_pd(fjx0,tx);
536 fjy0 = _mm256_add_pd(fjy0,ty);
537 fjz0 = _mm256_add_pd(fjz0,tz);
539 /**************************
540 * CALCULATE INTERACTIONS *
541 **************************/
543 r30 = _mm256_mul_pd(rsq30,rinv30);
544 r30 = _mm256_andnot_pd(dummy_mask,r30);
546 /* Compute parameters for interactions between i and j atoms */
547 qq30 = _mm256_mul_pd(iq3,jq0);
549 /* EWALD ELECTROSTATICS */
551 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
552 ewrt = _mm256_mul_pd(r30,ewtabscale);
553 ewitab = _mm256_cvttpd_epi32(ewrt);
554 eweps = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
555 ewitab = _mm_slli_epi32(ewitab,2);
556 ewtabF = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,0) );
557 ewtabD = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,1) );
558 ewtabV = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,2) );
559 ewtabFn = _mm256_load_pd( ewtab + _mm_extract_epi32(ewitab,3) );
560 GMX_MM256_FULLTRANSPOSE4_PD(ewtabF,ewtabD,ewtabV,ewtabFn);
561 felec = _mm256_add_pd(ewtabF,_mm256_mul_pd(eweps,ewtabD));
562 velec = _mm256_sub_pd(ewtabV,_mm256_mul_pd(_mm256_mul_pd(ewtabhalfspace,eweps),_mm256_add_pd(ewtabF,felec)));
563 velec = _mm256_mul_pd(qq30,_mm256_sub_pd(rinv30,velec));
564 felec = _mm256_mul_pd(_mm256_mul_pd(qq30,rinv30),_mm256_sub_pd(rinvsq30,felec));
566 /* Update potential sum for this i atom from the interaction with this j atom. */
567 velec = _mm256_andnot_pd(dummy_mask,velec);
568 velecsum = _mm256_add_pd(velecsum,velec);
572 fscal = _mm256_andnot_pd(dummy_mask,fscal);
574 /* Calculate temporary vectorial force */
575 tx = _mm256_mul_pd(fscal,dx30);
576 ty = _mm256_mul_pd(fscal,dy30);
577 tz = _mm256_mul_pd(fscal,dz30);
579 /* Update vectorial force */
580 fix3 = _mm256_add_pd(fix3,tx);
581 fiy3 = _mm256_add_pd(fiy3,ty);
582 fiz3 = _mm256_add_pd(fiz3,tz);
584 fjx0 = _mm256_add_pd(fjx0,tx);
585 fjy0 = _mm256_add_pd(fjy0,ty);
586 fjz0 = _mm256_add_pd(fjz0,tz);
588 fjptrA = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
589 fjptrB = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
590 fjptrC = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
591 fjptrD = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
593 gmx_mm256_decrement_1rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0);
595 /* Inner loop uses 129 flops */
598 /* End of innermost loop */
600 gmx_mm256_update_iforce_3atom_swizzle_pd(fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
601 f+i_coord_offset+DIM,fshift+i_shift_offset);
604 /* Update potential energies */
605 gmx_mm256_update_1pot_pd(velecsum,kernel_data->energygrp_elec+ggid);
607 /* Increment number of inner iterations */
608 inneriter += j_index_end - j_index_start;
610 /* Outer loop uses 19 flops */
613 /* Increment number of outer iterations */
616 /* Update outer/inner flops */
618 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_W4_VF,outeriter*19 + inneriter*129);
621 * Gromacs nonbonded kernel: nb_kernel_ElecEw_VdwNone_GeomW4P1_F_avx_256_double
622 * Electrostatics interaction: Ewald
623 * VdW interaction: None
624 * Geometry: Water4-Particle
625 * Calculate force/pot: Force
628 nb_kernel_ElecEw_VdwNone_GeomW4P1_F_avx_256_double
629 (t_nblist * gmx_restrict nlist,
630 rvec * gmx_restrict xx,
631 rvec * gmx_restrict ff,
632 t_forcerec * gmx_restrict fr,
633 t_mdatoms * gmx_restrict mdatoms,
634 nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
635 t_nrnb * gmx_restrict nrnb)
637 /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
638 * just 0 for non-waters.
639 * Suffixes A,B,C,D refer to j loop unrolling done with AVX, e.g. for the four different
640 * jnr indices corresponding to data put in the four positions in the SIMD register.
642 int i_shift_offset,i_coord_offset,outeriter,inneriter;
643 int j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
644 int jnrA,jnrB,jnrC,jnrD;
645 int jnrlistA,jnrlistB,jnrlistC,jnrlistD;
646 int jnrlistE,jnrlistF,jnrlistG,jnrlistH;
647 int j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
648 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
650 real *shiftvec,*fshift,*x,*f;
651 real *fjptrA,*fjptrB,*fjptrC,*fjptrD;
653 __m256d tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
654 real * vdwioffsetptr1;
655 __m256d ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
656 real * vdwioffsetptr2;
657 __m256d ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
658 real * vdwioffsetptr3;
659 __m256d ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
660 int vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D;
661 __m256d jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
662 __m256d dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
663 __m256d dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
664 __m256d dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30;
665 __m256d velec,felec,velecsum,facel,crf,krf,krf2;
668 __m256d ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV;
669 __m256d beta,beta2,beta3,zeta2,pmecorrF,pmecorrV,rinv3;
671 __m256d dummy_mask,cutoff_mask;
672 __m128 tmpmask0,tmpmask1;
673 __m256d signbit = _mm256_castsi256_pd( _mm256_set1_epi32(0x80000000) );
674 __m256d one = _mm256_set1_pd(1.0);
675 __m256d two = _mm256_set1_pd(2.0);
681 jindex = nlist->jindex;
683 shiftidx = nlist->shift;
685 shiftvec = fr->shift_vec[0];
686 fshift = fr->fshift[0];
687 facel = _mm256_set1_pd(fr->epsfac);
688 charge = mdatoms->chargeA;
690 sh_ewald = _mm256_set1_pd(fr->ic->sh_ewald);
691 beta = _mm256_set1_pd(fr->ic->ewaldcoeff_q);
692 beta2 = _mm256_mul_pd(beta,beta);
693 beta3 = _mm256_mul_pd(beta,beta2);
695 ewtab = fr->ic->tabq_coul_F;
696 ewtabscale = _mm256_set1_pd(fr->ic->tabq_scale);
697 ewtabhalfspace = _mm256_set1_pd(0.5/fr->ic->tabq_scale);
699 /* Setup water-specific parameters */
700 inr = nlist->iinr[0];
701 iq1 = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+1]));
702 iq2 = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+2]));
703 iq3 = _mm256_mul_pd(facel,_mm256_set1_pd(charge[inr+3]));
705 /* Avoid stupid compiler warnings */
706 jnrA = jnrB = jnrC = jnrD = 0;
715 for(iidx=0;iidx<4*DIM;iidx++)
720 /* Start outer loop over neighborlists */
721 for(iidx=0; iidx<nri; iidx++)
723 /* Load shift vector for this list */
724 i_shift_offset = DIM*shiftidx[iidx];
726 /* Load limits for loop over neighbors */
727 j_index_start = jindex[iidx];
728 j_index_end = jindex[iidx+1];
730 /* Get outer coordinate index */
732 i_coord_offset = DIM*inr;
734 /* Load i particle coords and add shift vector */
735 gmx_mm256_load_shift_and_3rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset+DIM,
736 &ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
738 fix1 = _mm256_setzero_pd();
739 fiy1 = _mm256_setzero_pd();
740 fiz1 = _mm256_setzero_pd();
741 fix2 = _mm256_setzero_pd();
742 fiy2 = _mm256_setzero_pd();
743 fiz2 = _mm256_setzero_pd();
744 fix3 = _mm256_setzero_pd();
745 fiy3 = _mm256_setzero_pd();
746 fiz3 = _mm256_setzero_pd();
748 /* Start inner kernel loop */
749 for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4)
752 /* Get j neighbor index, and coordinate index */
757 j_coord_offsetA = DIM*jnrA;
758 j_coord_offsetB = DIM*jnrB;
759 j_coord_offsetC = DIM*jnrC;
760 j_coord_offsetD = DIM*jnrD;
762 /* load j atom coordinates */
763 gmx_mm256_load_1rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
764 x+j_coord_offsetC,x+j_coord_offsetD,
767 /* Calculate displacement vector */
768 dx10 = _mm256_sub_pd(ix1,jx0);
769 dy10 = _mm256_sub_pd(iy1,jy0);
770 dz10 = _mm256_sub_pd(iz1,jz0);
771 dx20 = _mm256_sub_pd(ix2,jx0);
772 dy20 = _mm256_sub_pd(iy2,jy0);
773 dz20 = _mm256_sub_pd(iz2,jz0);
774 dx30 = _mm256_sub_pd(ix3,jx0);
775 dy30 = _mm256_sub_pd(iy3,jy0);
776 dz30 = _mm256_sub_pd(iz3,jz0);
778 /* Calculate squared distance and things based on it */
779 rsq10 = gmx_mm256_calc_rsq_pd(dx10,dy10,dz10);
780 rsq20 = gmx_mm256_calc_rsq_pd(dx20,dy20,dz20);
781 rsq30 = gmx_mm256_calc_rsq_pd(dx30,dy30,dz30);
783 rinv10 = gmx_mm256_invsqrt_pd(rsq10);
784 rinv20 = gmx_mm256_invsqrt_pd(rsq20);
785 rinv30 = gmx_mm256_invsqrt_pd(rsq30);
787 rinvsq10 = _mm256_mul_pd(rinv10,rinv10);
788 rinvsq20 = _mm256_mul_pd(rinv20,rinv20);
789 rinvsq30 = _mm256_mul_pd(rinv30,rinv30);
791 /* Load parameters for j particles */
792 jq0 = gmx_mm256_load_4real_swizzle_pd(charge+jnrA+0,charge+jnrB+0,
793 charge+jnrC+0,charge+jnrD+0);
795 fjx0 = _mm256_setzero_pd();
796 fjy0 = _mm256_setzero_pd();
797 fjz0 = _mm256_setzero_pd();
799 /**************************
800 * CALCULATE INTERACTIONS *
801 **************************/
803 r10 = _mm256_mul_pd(rsq10,rinv10);
805 /* Compute parameters for interactions between i and j atoms */
806 qq10 = _mm256_mul_pd(iq1,jq0);
808 /* EWALD ELECTROSTATICS */
810 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
811 ewrt = _mm256_mul_pd(r10,ewtabscale);
812 ewitab = _mm256_cvttpd_epi32(ewrt);
813 eweps = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
814 gmx_mm256_load_4pair_swizzle_pd(ewtab + _mm_extract_epi32(ewitab,0),ewtab + _mm_extract_epi32(ewitab,1),
815 ewtab + _mm_extract_epi32(ewitab,2),ewtab + _mm_extract_epi32(ewitab,3),
817 felec = _mm256_add_pd(_mm256_mul_pd( _mm256_sub_pd(one,eweps),ewtabF),_mm256_mul_pd(eweps,ewtabFn));
818 felec = _mm256_mul_pd(_mm256_mul_pd(qq10,rinv10),_mm256_sub_pd(rinvsq10,felec));
822 /* Calculate temporary vectorial force */
823 tx = _mm256_mul_pd(fscal,dx10);
824 ty = _mm256_mul_pd(fscal,dy10);
825 tz = _mm256_mul_pd(fscal,dz10);
827 /* Update vectorial force */
828 fix1 = _mm256_add_pd(fix1,tx);
829 fiy1 = _mm256_add_pd(fiy1,ty);
830 fiz1 = _mm256_add_pd(fiz1,tz);
832 fjx0 = _mm256_add_pd(fjx0,tx);
833 fjy0 = _mm256_add_pd(fjy0,ty);
834 fjz0 = _mm256_add_pd(fjz0,tz);
836 /**************************
837 * CALCULATE INTERACTIONS *
838 **************************/
840 r20 = _mm256_mul_pd(rsq20,rinv20);
842 /* Compute parameters for interactions between i and j atoms */
843 qq20 = _mm256_mul_pd(iq2,jq0);
845 /* EWALD ELECTROSTATICS */
847 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
848 ewrt = _mm256_mul_pd(r20,ewtabscale);
849 ewitab = _mm256_cvttpd_epi32(ewrt);
850 eweps = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
851 gmx_mm256_load_4pair_swizzle_pd(ewtab + _mm_extract_epi32(ewitab,0),ewtab + _mm_extract_epi32(ewitab,1),
852 ewtab + _mm_extract_epi32(ewitab,2),ewtab + _mm_extract_epi32(ewitab,3),
854 felec = _mm256_add_pd(_mm256_mul_pd( _mm256_sub_pd(one,eweps),ewtabF),_mm256_mul_pd(eweps,ewtabFn));
855 felec = _mm256_mul_pd(_mm256_mul_pd(qq20,rinv20),_mm256_sub_pd(rinvsq20,felec));
859 /* Calculate temporary vectorial force */
860 tx = _mm256_mul_pd(fscal,dx20);
861 ty = _mm256_mul_pd(fscal,dy20);
862 tz = _mm256_mul_pd(fscal,dz20);
864 /* Update vectorial force */
865 fix2 = _mm256_add_pd(fix2,tx);
866 fiy2 = _mm256_add_pd(fiy2,ty);
867 fiz2 = _mm256_add_pd(fiz2,tz);
869 fjx0 = _mm256_add_pd(fjx0,tx);
870 fjy0 = _mm256_add_pd(fjy0,ty);
871 fjz0 = _mm256_add_pd(fjz0,tz);
873 /**************************
874 * CALCULATE INTERACTIONS *
875 **************************/
877 r30 = _mm256_mul_pd(rsq30,rinv30);
879 /* Compute parameters for interactions between i and j atoms */
880 qq30 = _mm256_mul_pd(iq3,jq0);
882 /* EWALD ELECTROSTATICS */
884 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
885 ewrt = _mm256_mul_pd(r30,ewtabscale);
886 ewitab = _mm256_cvttpd_epi32(ewrt);
887 eweps = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
888 gmx_mm256_load_4pair_swizzle_pd(ewtab + _mm_extract_epi32(ewitab,0),ewtab + _mm_extract_epi32(ewitab,1),
889 ewtab + _mm_extract_epi32(ewitab,2),ewtab + _mm_extract_epi32(ewitab,3),
891 felec = _mm256_add_pd(_mm256_mul_pd( _mm256_sub_pd(one,eweps),ewtabF),_mm256_mul_pd(eweps,ewtabFn));
892 felec = _mm256_mul_pd(_mm256_mul_pd(qq30,rinv30),_mm256_sub_pd(rinvsq30,felec));
896 /* Calculate temporary vectorial force */
897 tx = _mm256_mul_pd(fscal,dx30);
898 ty = _mm256_mul_pd(fscal,dy30);
899 tz = _mm256_mul_pd(fscal,dz30);
901 /* Update vectorial force */
902 fix3 = _mm256_add_pd(fix3,tx);
903 fiy3 = _mm256_add_pd(fiy3,ty);
904 fiz3 = _mm256_add_pd(fiz3,tz);
906 fjx0 = _mm256_add_pd(fjx0,tx);
907 fjy0 = _mm256_add_pd(fjy0,ty);
908 fjz0 = _mm256_add_pd(fjz0,tz);
910 fjptrA = f+j_coord_offsetA;
911 fjptrB = f+j_coord_offsetB;
912 fjptrC = f+j_coord_offsetC;
913 fjptrD = f+j_coord_offsetD;
915 gmx_mm256_decrement_1rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0);
917 /* Inner loop uses 111 flops */
923 /* Get j neighbor index, and coordinate index */
924 jnrlistA = jjnr[jidx];
925 jnrlistB = jjnr[jidx+1];
926 jnrlistC = jjnr[jidx+2];
927 jnrlistD = jjnr[jidx+3];
928 /* Sign of each element will be negative for non-real atoms.
929 * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
930 * so use it as val = _mm_andnot_pd(mask,val) to clear dummy entries.
932 tmpmask0 = gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128()));
934 tmpmask1 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(3,3,2,2));
935 tmpmask0 = _mm_permute_ps(tmpmask0,_GMX_MM_PERMUTE(1,1,0,0));
936 dummy_mask = _mm256_castps_pd(gmx_mm256_set_m128(tmpmask1,tmpmask0));
938 jnrA = (jnrlistA>=0) ? jnrlistA : 0;
939 jnrB = (jnrlistB>=0) ? jnrlistB : 0;
940 jnrC = (jnrlistC>=0) ? jnrlistC : 0;
941 jnrD = (jnrlistD>=0) ? jnrlistD : 0;
942 j_coord_offsetA = DIM*jnrA;
943 j_coord_offsetB = DIM*jnrB;
944 j_coord_offsetC = DIM*jnrC;
945 j_coord_offsetD = DIM*jnrD;
947 /* load j atom coordinates */
948 gmx_mm256_load_1rvec_4ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
949 x+j_coord_offsetC,x+j_coord_offsetD,
952 /* Calculate displacement vector */
953 dx10 = _mm256_sub_pd(ix1,jx0);
954 dy10 = _mm256_sub_pd(iy1,jy0);
955 dz10 = _mm256_sub_pd(iz1,jz0);
956 dx20 = _mm256_sub_pd(ix2,jx0);
957 dy20 = _mm256_sub_pd(iy2,jy0);
958 dz20 = _mm256_sub_pd(iz2,jz0);
959 dx30 = _mm256_sub_pd(ix3,jx0);
960 dy30 = _mm256_sub_pd(iy3,jy0);
961 dz30 = _mm256_sub_pd(iz3,jz0);
963 /* Calculate squared distance and things based on it */
964 rsq10 = gmx_mm256_calc_rsq_pd(dx10,dy10,dz10);
965 rsq20 = gmx_mm256_calc_rsq_pd(dx20,dy20,dz20);
966 rsq30 = gmx_mm256_calc_rsq_pd(dx30,dy30,dz30);
968 rinv10 = gmx_mm256_invsqrt_pd(rsq10);
969 rinv20 = gmx_mm256_invsqrt_pd(rsq20);
970 rinv30 = gmx_mm256_invsqrt_pd(rsq30);
972 rinvsq10 = _mm256_mul_pd(rinv10,rinv10);
973 rinvsq20 = _mm256_mul_pd(rinv20,rinv20);
974 rinvsq30 = _mm256_mul_pd(rinv30,rinv30);
976 /* Load parameters for j particles */
977 jq0 = gmx_mm256_load_4real_swizzle_pd(charge+jnrA+0,charge+jnrB+0,
978 charge+jnrC+0,charge+jnrD+0);
980 fjx0 = _mm256_setzero_pd();
981 fjy0 = _mm256_setzero_pd();
982 fjz0 = _mm256_setzero_pd();
984 /**************************
985 * CALCULATE INTERACTIONS *
986 **************************/
988 r10 = _mm256_mul_pd(rsq10,rinv10);
989 r10 = _mm256_andnot_pd(dummy_mask,r10);
991 /* Compute parameters for interactions between i and j atoms */
992 qq10 = _mm256_mul_pd(iq1,jq0);
994 /* EWALD ELECTROSTATICS */
996 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
997 ewrt = _mm256_mul_pd(r10,ewtabscale);
998 ewitab = _mm256_cvttpd_epi32(ewrt);
999 eweps = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
1000 gmx_mm256_load_4pair_swizzle_pd(ewtab + _mm_extract_epi32(ewitab,0),ewtab + _mm_extract_epi32(ewitab,1),
1001 ewtab + _mm_extract_epi32(ewitab,2),ewtab + _mm_extract_epi32(ewitab,3),
1003 felec = _mm256_add_pd(_mm256_mul_pd( _mm256_sub_pd(one,eweps),ewtabF),_mm256_mul_pd(eweps,ewtabFn));
1004 felec = _mm256_mul_pd(_mm256_mul_pd(qq10,rinv10),_mm256_sub_pd(rinvsq10,felec));
1008 fscal = _mm256_andnot_pd(dummy_mask,fscal);
1010 /* Calculate temporary vectorial force */
1011 tx = _mm256_mul_pd(fscal,dx10);
1012 ty = _mm256_mul_pd(fscal,dy10);
1013 tz = _mm256_mul_pd(fscal,dz10);
1015 /* Update vectorial force */
1016 fix1 = _mm256_add_pd(fix1,tx);
1017 fiy1 = _mm256_add_pd(fiy1,ty);
1018 fiz1 = _mm256_add_pd(fiz1,tz);
1020 fjx0 = _mm256_add_pd(fjx0,tx);
1021 fjy0 = _mm256_add_pd(fjy0,ty);
1022 fjz0 = _mm256_add_pd(fjz0,tz);
1024 /**************************
1025 * CALCULATE INTERACTIONS *
1026 **************************/
1028 r20 = _mm256_mul_pd(rsq20,rinv20);
1029 r20 = _mm256_andnot_pd(dummy_mask,r20);
1031 /* Compute parameters for interactions between i and j atoms */
1032 qq20 = _mm256_mul_pd(iq2,jq0);
1034 /* EWALD ELECTROSTATICS */
1036 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1037 ewrt = _mm256_mul_pd(r20,ewtabscale);
1038 ewitab = _mm256_cvttpd_epi32(ewrt);
1039 eweps = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
1040 gmx_mm256_load_4pair_swizzle_pd(ewtab + _mm_extract_epi32(ewitab,0),ewtab + _mm_extract_epi32(ewitab,1),
1041 ewtab + _mm_extract_epi32(ewitab,2),ewtab + _mm_extract_epi32(ewitab,3),
1043 felec = _mm256_add_pd(_mm256_mul_pd( _mm256_sub_pd(one,eweps),ewtabF),_mm256_mul_pd(eweps,ewtabFn));
1044 felec = _mm256_mul_pd(_mm256_mul_pd(qq20,rinv20),_mm256_sub_pd(rinvsq20,felec));
1048 fscal = _mm256_andnot_pd(dummy_mask,fscal);
1050 /* Calculate temporary vectorial force */
1051 tx = _mm256_mul_pd(fscal,dx20);
1052 ty = _mm256_mul_pd(fscal,dy20);
1053 tz = _mm256_mul_pd(fscal,dz20);
1055 /* Update vectorial force */
1056 fix2 = _mm256_add_pd(fix2,tx);
1057 fiy2 = _mm256_add_pd(fiy2,ty);
1058 fiz2 = _mm256_add_pd(fiz2,tz);
1060 fjx0 = _mm256_add_pd(fjx0,tx);
1061 fjy0 = _mm256_add_pd(fjy0,ty);
1062 fjz0 = _mm256_add_pd(fjz0,tz);
1064 /**************************
1065 * CALCULATE INTERACTIONS *
1066 **************************/
1068 r30 = _mm256_mul_pd(rsq30,rinv30);
1069 r30 = _mm256_andnot_pd(dummy_mask,r30);
1071 /* Compute parameters for interactions between i and j atoms */
1072 qq30 = _mm256_mul_pd(iq3,jq0);
1074 /* EWALD ELECTROSTATICS */
1076 /* Calculate Ewald table index by multiplying r with scale and truncate to integer */
1077 ewrt = _mm256_mul_pd(r30,ewtabscale);
1078 ewitab = _mm256_cvttpd_epi32(ewrt);
1079 eweps = _mm256_sub_pd(ewrt,_mm256_round_pd(ewrt, _MM_FROUND_FLOOR));
1080 gmx_mm256_load_4pair_swizzle_pd(ewtab + _mm_extract_epi32(ewitab,0),ewtab + _mm_extract_epi32(ewitab,1),
1081 ewtab + _mm_extract_epi32(ewitab,2),ewtab + _mm_extract_epi32(ewitab,3),
1083 felec = _mm256_add_pd(_mm256_mul_pd( _mm256_sub_pd(one,eweps),ewtabF),_mm256_mul_pd(eweps,ewtabFn));
1084 felec = _mm256_mul_pd(_mm256_mul_pd(qq30,rinv30),_mm256_sub_pd(rinvsq30,felec));
1088 fscal = _mm256_andnot_pd(dummy_mask,fscal);
1090 /* Calculate temporary vectorial force */
1091 tx = _mm256_mul_pd(fscal,dx30);
1092 ty = _mm256_mul_pd(fscal,dy30);
1093 tz = _mm256_mul_pd(fscal,dz30);
1095 /* Update vectorial force */
1096 fix3 = _mm256_add_pd(fix3,tx);
1097 fiy3 = _mm256_add_pd(fiy3,ty);
1098 fiz3 = _mm256_add_pd(fiz3,tz);
1100 fjx0 = _mm256_add_pd(fjx0,tx);
1101 fjy0 = _mm256_add_pd(fjy0,ty);
1102 fjz0 = _mm256_add_pd(fjz0,tz);
1104 fjptrA = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
1105 fjptrB = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
1106 fjptrC = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
1107 fjptrD = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
1109 gmx_mm256_decrement_1rvec_4ptr_swizzle_pd(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0);
1111 /* Inner loop uses 114 flops */
1114 /* End of innermost loop */
1116 gmx_mm256_update_iforce_3atom_swizzle_pd(fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
1117 f+i_coord_offset+DIM,fshift+i_shift_offset);
1119 /* Increment number of inner iterations */
1120 inneriter += j_index_end - j_index_start;
1122 /* Outer loop uses 18 flops */
1125 /* Increment number of outer iterations */
1128 /* Update outer/inner flops */
1130 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_W4_F,outeriter*18 + inneriter*114);