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_128_fma_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_128_fma_double.h"
48 #include "kernelutil_x86_avx_128_fma_double.h"
51 * Gromacs nonbonded kernel: nb_kernel_ElecCSTab_VdwLJ_GeomW3P1_VF_avx_128_fma_double
52 * Electrostatics interaction: CubicSplineTable
53 * VdW interaction: LennardJones
54 * Geometry: Water3-Particle
55 * Calculate force/pot: PotentialAndForce
58 nb_kernel_ElecCSTab_VdwLJ_GeomW3P1_VF_avx_128_fma_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 refer to j loop unrolling done with SSE double precision, e.g. for the two 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;
75 int j_coord_offsetA,j_coord_offsetB;
76 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
78 real *shiftvec,*fshift,*x,*f;
79 __m128d tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
81 __m128d ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
83 __m128d ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
85 __m128d ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
86 int vdwjidx0A,vdwjidx0B;
87 __m128d jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
88 __m128d dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
89 __m128d dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
90 __m128d dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
91 __m128d velec,felec,velecsum,facel,crf,krf,krf2;
94 __m128d rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
97 __m128d one_sixth = _mm_set1_pd(1.0/6.0);
98 __m128d one_twelfth = _mm_set1_pd(1.0/12.0);
100 __m128i ifour = _mm_set1_epi32(4);
101 __m128d rt,vfeps,vftabscale,Y,F,G,H,Heps,Fp,VV,FF,twovfeps;
103 __m128d dummy_mask,cutoff_mask;
104 __m128d signbit = gmx_mm_castsi128_pd( _mm_set_epi32(0x80000000,0x00000000,0x80000000,0x00000000) );
105 __m128d one = _mm_set1_pd(1.0);
106 __m128d two = _mm_set1_pd(2.0);
112 jindex = nlist->jindex;
114 shiftidx = nlist->shift;
116 shiftvec = fr->shift_vec[0];
117 fshift = fr->fshift[0];
118 facel = _mm_set1_pd(fr->epsfac);
119 charge = mdatoms->chargeA;
120 nvdwtype = fr->ntype;
122 vdwtype = mdatoms->typeA;
124 vftab = kernel_data->table_elec->data;
125 vftabscale = _mm_set1_pd(kernel_data->table_elec->scale);
127 /* Setup water-specific parameters */
128 inr = nlist->iinr[0];
129 iq0 = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+0]));
130 iq1 = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+1]));
131 iq2 = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+2]));
132 vdwioffset0 = 2*nvdwtype*vdwtype[inr+0];
134 /* Avoid stupid compiler warnings */
142 /* Start outer loop over neighborlists */
143 for(iidx=0; iidx<nri; iidx++)
145 /* Load shift vector for this list */
146 i_shift_offset = DIM*shiftidx[iidx];
148 /* Load limits for loop over neighbors */
149 j_index_start = jindex[iidx];
150 j_index_end = jindex[iidx+1];
152 /* Get outer coordinate index */
154 i_coord_offset = DIM*inr;
156 /* Load i particle coords and add shift vector */
157 gmx_mm_load_shift_and_3rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset,
158 &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2);
160 fix0 = _mm_setzero_pd();
161 fiy0 = _mm_setzero_pd();
162 fiz0 = _mm_setzero_pd();
163 fix1 = _mm_setzero_pd();
164 fiy1 = _mm_setzero_pd();
165 fiz1 = _mm_setzero_pd();
166 fix2 = _mm_setzero_pd();
167 fiy2 = _mm_setzero_pd();
168 fiz2 = _mm_setzero_pd();
170 /* Reset potential sums */
171 velecsum = _mm_setzero_pd();
172 vvdwsum = _mm_setzero_pd();
174 /* Start inner kernel loop */
175 for(jidx=j_index_start; jidx<j_index_end-1; jidx+=2)
178 /* Get j neighbor index, and coordinate index */
181 j_coord_offsetA = DIM*jnrA;
182 j_coord_offsetB = DIM*jnrB;
184 /* load j atom coordinates */
185 gmx_mm_load_1rvec_2ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
188 /* Calculate displacement vector */
189 dx00 = _mm_sub_pd(ix0,jx0);
190 dy00 = _mm_sub_pd(iy0,jy0);
191 dz00 = _mm_sub_pd(iz0,jz0);
192 dx10 = _mm_sub_pd(ix1,jx0);
193 dy10 = _mm_sub_pd(iy1,jy0);
194 dz10 = _mm_sub_pd(iz1,jz0);
195 dx20 = _mm_sub_pd(ix2,jx0);
196 dy20 = _mm_sub_pd(iy2,jy0);
197 dz20 = _mm_sub_pd(iz2,jz0);
199 /* Calculate squared distance and things based on it */
200 rsq00 = gmx_mm_calc_rsq_pd(dx00,dy00,dz00);
201 rsq10 = gmx_mm_calc_rsq_pd(dx10,dy10,dz10);
202 rsq20 = gmx_mm_calc_rsq_pd(dx20,dy20,dz20);
204 rinv00 = gmx_mm_invsqrt_pd(rsq00);
205 rinv10 = gmx_mm_invsqrt_pd(rsq10);
206 rinv20 = gmx_mm_invsqrt_pd(rsq20);
208 rinvsq00 = _mm_mul_pd(rinv00,rinv00);
210 /* Load parameters for j particles */
211 jq0 = gmx_mm_load_2real_swizzle_pd(charge+jnrA+0,charge+jnrB+0);
212 vdwjidx0A = 2*vdwtype[jnrA+0];
213 vdwjidx0B = 2*vdwtype[jnrB+0];
215 fjx0 = _mm_setzero_pd();
216 fjy0 = _mm_setzero_pd();
217 fjz0 = _mm_setzero_pd();
219 /**************************
220 * CALCULATE INTERACTIONS *
221 **************************/
223 r00 = _mm_mul_pd(rsq00,rinv00);
225 /* Compute parameters for interactions between i and j atoms */
226 qq00 = _mm_mul_pd(iq0,jq0);
227 gmx_mm_load_2pair_swizzle_pd(vdwparam+vdwioffset0+vdwjidx0A,
228 vdwparam+vdwioffset0+vdwjidx0B,&c6_00,&c12_00);
230 /* Calculate table index by multiplying r with table scale and truncate to integer */
231 rt = _mm_mul_pd(r00,vftabscale);
232 vfitab = _mm_cvttpd_epi32(rt);
234 vfeps = _mm_frcz_pd(rt);
236 vfeps = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
238 twovfeps = _mm_add_pd(vfeps,vfeps);
239 vfitab = _mm_slli_epi32(vfitab,2);
241 /* CUBIC SPLINE TABLE ELECTROSTATICS */
242 Y = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
243 F = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) );
244 GMX_MM_TRANSPOSE2_PD(Y,F);
245 G = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
246 H = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) +2);
247 GMX_MM_TRANSPOSE2_PD(G,H);
248 Fp = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
249 VV = _mm_macc_pd(vfeps,Fp,Y);
250 velec = _mm_mul_pd(qq00,VV);
251 FF = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
252 felec = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq00,FF),_mm_mul_pd(vftabscale,rinv00)));
254 /* LENNARD-JONES DISPERSION/REPULSION */
256 rinvsix = _mm_mul_pd(_mm_mul_pd(rinvsq00,rinvsq00),rinvsq00);
257 vvdw6 = _mm_mul_pd(c6_00,rinvsix);
258 vvdw12 = _mm_mul_pd(c12_00,_mm_mul_pd(rinvsix,rinvsix));
259 vvdw = _mm_msub_pd( vvdw12,one_twelfth, _mm_mul_pd(vvdw6,one_sixth) );
260 fvdw = _mm_mul_pd(_mm_sub_pd(vvdw12,vvdw6),rinvsq00);
262 /* Update potential sum for this i atom from the interaction with this j atom. */
263 velecsum = _mm_add_pd(velecsum,velec);
264 vvdwsum = _mm_add_pd(vvdwsum,vvdw);
266 fscal = _mm_add_pd(felec,fvdw);
268 /* Update vectorial force */
269 fix0 = _mm_macc_pd(dx00,fscal,fix0);
270 fiy0 = _mm_macc_pd(dy00,fscal,fiy0);
271 fiz0 = _mm_macc_pd(dz00,fscal,fiz0);
273 fjx0 = _mm_macc_pd(dx00,fscal,fjx0);
274 fjy0 = _mm_macc_pd(dy00,fscal,fjy0);
275 fjz0 = _mm_macc_pd(dz00,fscal,fjz0);
277 /**************************
278 * CALCULATE INTERACTIONS *
279 **************************/
281 r10 = _mm_mul_pd(rsq10,rinv10);
283 /* Compute parameters for interactions between i and j atoms */
284 qq10 = _mm_mul_pd(iq1,jq0);
286 /* Calculate table index by multiplying r with table scale and truncate to integer */
287 rt = _mm_mul_pd(r10,vftabscale);
288 vfitab = _mm_cvttpd_epi32(rt);
290 vfeps = _mm_frcz_pd(rt);
292 vfeps = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
294 twovfeps = _mm_add_pd(vfeps,vfeps);
295 vfitab = _mm_slli_epi32(vfitab,2);
297 /* CUBIC SPLINE TABLE ELECTROSTATICS */
298 Y = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
299 F = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) );
300 GMX_MM_TRANSPOSE2_PD(Y,F);
301 G = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
302 H = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) +2);
303 GMX_MM_TRANSPOSE2_PD(G,H);
304 Fp = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
305 VV = _mm_macc_pd(vfeps,Fp,Y);
306 velec = _mm_mul_pd(qq10,VV);
307 FF = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
308 felec = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq10,FF),_mm_mul_pd(vftabscale,rinv10)));
310 /* Update potential sum for this i atom from the interaction with this j atom. */
311 velecsum = _mm_add_pd(velecsum,velec);
315 /* Update vectorial force */
316 fix1 = _mm_macc_pd(dx10,fscal,fix1);
317 fiy1 = _mm_macc_pd(dy10,fscal,fiy1);
318 fiz1 = _mm_macc_pd(dz10,fscal,fiz1);
320 fjx0 = _mm_macc_pd(dx10,fscal,fjx0);
321 fjy0 = _mm_macc_pd(dy10,fscal,fjy0);
322 fjz0 = _mm_macc_pd(dz10,fscal,fjz0);
324 /**************************
325 * CALCULATE INTERACTIONS *
326 **************************/
328 r20 = _mm_mul_pd(rsq20,rinv20);
330 /* Compute parameters for interactions between i and j atoms */
331 qq20 = _mm_mul_pd(iq2,jq0);
333 /* Calculate table index by multiplying r with table scale and truncate to integer */
334 rt = _mm_mul_pd(r20,vftabscale);
335 vfitab = _mm_cvttpd_epi32(rt);
337 vfeps = _mm_frcz_pd(rt);
339 vfeps = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
341 twovfeps = _mm_add_pd(vfeps,vfeps);
342 vfitab = _mm_slli_epi32(vfitab,2);
344 /* CUBIC SPLINE TABLE ELECTROSTATICS */
345 Y = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
346 F = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) );
347 GMX_MM_TRANSPOSE2_PD(Y,F);
348 G = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
349 H = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) +2);
350 GMX_MM_TRANSPOSE2_PD(G,H);
351 Fp = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
352 VV = _mm_macc_pd(vfeps,Fp,Y);
353 velec = _mm_mul_pd(qq20,VV);
354 FF = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
355 felec = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq20,FF),_mm_mul_pd(vftabscale,rinv20)));
357 /* Update potential sum for this i atom from the interaction with this j atom. */
358 velecsum = _mm_add_pd(velecsum,velec);
362 /* Update vectorial force */
363 fix2 = _mm_macc_pd(dx20,fscal,fix2);
364 fiy2 = _mm_macc_pd(dy20,fscal,fiy2);
365 fiz2 = _mm_macc_pd(dz20,fscal,fiz2);
367 fjx0 = _mm_macc_pd(dx20,fscal,fjx0);
368 fjy0 = _mm_macc_pd(dy20,fscal,fjy0);
369 fjz0 = _mm_macc_pd(dz20,fscal,fjz0);
371 gmx_mm_decrement_1rvec_2ptr_swizzle_pd(f+j_coord_offsetA,f+j_coord_offsetB,fjx0,fjy0,fjz0);
373 /* Inner loop uses 154 flops */
380 j_coord_offsetA = DIM*jnrA;
382 /* load j atom coordinates */
383 gmx_mm_load_1rvec_1ptr_swizzle_pd(x+j_coord_offsetA,
386 /* Calculate displacement vector */
387 dx00 = _mm_sub_pd(ix0,jx0);
388 dy00 = _mm_sub_pd(iy0,jy0);
389 dz00 = _mm_sub_pd(iz0,jz0);
390 dx10 = _mm_sub_pd(ix1,jx0);
391 dy10 = _mm_sub_pd(iy1,jy0);
392 dz10 = _mm_sub_pd(iz1,jz0);
393 dx20 = _mm_sub_pd(ix2,jx0);
394 dy20 = _mm_sub_pd(iy2,jy0);
395 dz20 = _mm_sub_pd(iz2,jz0);
397 /* Calculate squared distance and things based on it */
398 rsq00 = gmx_mm_calc_rsq_pd(dx00,dy00,dz00);
399 rsq10 = gmx_mm_calc_rsq_pd(dx10,dy10,dz10);
400 rsq20 = gmx_mm_calc_rsq_pd(dx20,dy20,dz20);
402 rinv00 = gmx_mm_invsqrt_pd(rsq00);
403 rinv10 = gmx_mm_invsqrt_pd(rsq10);
404 rinv20 = gmx_mm_invsqrt_pd(rsq20);
406 rinvsq00 = _mm_mul_pd(rinv00,rinv00);
408 /* Load parameters for j particles */
409 jq0 = _mm_load_sd(charge+jnrA+0);
410 vdwjidx0A = 2*vdwtype[jnrA+0];
412 fjx0 = _mm_setzero_pd();
413 fjy0 = _mm_setzero_pd();
414 fjz0 = _mm_setzero_pd();
416 /**************************
417 * CALCULATE INTERACTIONS *
418 **************************/
420 r00 = _mm_mul_pd(rsq00,rinv00);
422 /* Compute parameters for interactions between i and j atoms */
423 qq00 = _mm_mul_pd(iq0,jq0);
424 gmx_mm_load_1pair_swizzle_pd(vdwparam+vdwioffset0+vdwjidx0A,&c6_00,&c12_00);
426 /* Calculate table index by multiplying r with table scale and truncate to integer */
427 rt = _mm_mul_pd(r00,vftabscale);
428 vfitab = _mm_cvttpd_epi32(rt);
430 vfeps = _mm_frcz_pd(rt);
432 vfeps = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
434 twovfeps = _mm_add_pd(vfeps,vfeps);
435 vfitab = _mm_slli_epi32(vfitab,2);
437 /* CUBIC SPLINE TABLE ELECTROSTATICS */
438 Y = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
439 F = _mm_setzero_pd();
440 GMX_MM_TRANSPOSE2_PD(Y,F);
441 G = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
442 H = _mm_setzero_pd();
443 GMX_MM_TRANSPOSE2_PD(G,H);
444 Fp = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
445 VV = _mm_macc_pd(vfeps,Fp,Y);
446 velec = _mm_mul_pd(qq00,VV);
447 FF = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
448 felec = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq00,FF),_mm_mul_pd(vftabscale,rinv00)));
450 /* LENNARD-JONES DISPERSION/REPULSION */
452 rinvsix = _mm_mul_pd(_mm_mul_pd(rinvsq00,rinvsq00),rinvsq00);
453 vvdw6 = _mm_mul_pd(c6_00,rinvsix);
454 vvdw12 = _mm_mul_pd(c12_00,_mm_mul_pd(rinvsix,rinvsix));
455 vvdw = _mm_msub_pd( vvdw12,one_twelfth, _mm_mul_pd(vvdw6,one_sixth) );
456 fvdw = _mm_mul_pd(_mm_sub_pd(vvdw12,vvdw6),rinvsq00);
458 /* Update potential sum for this i atom from the interaction with this j atom. */
459 velec = _mm_unpacklo_pd(velec,_mm_setzero_pd());
460 velecsum = _mm_add_pd(velecsum,velec);
461 vvdw = _mm_unpacklo_pd(vvdw,_mm_setzero_pd());
462 vvdwsum = _mm_add_pd(vvdwsum,vvdw);
464 fscal = _mm_add_pd(felec,fvdw);
466 fscal = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
468 /* Update vectorial force */
469 fix0 = _mm_macc_pd(dx00,fscal,fix0);
470 fiy0 = _mm_macc_pd(dy00,fscal,fiy0);
471 fiz0 = _mm_macc_pd(dz00,fscal,fiz0);
473 fjx0 = _mm_macc_pd(dx00,fscal,fjx0);
474 fjy0 = _mm_macc_pd(dy00,fscal,fjy0);
475 fjz0 = _mm_macc_pd(dz00,fscal,fjz0);
477 /**************************
478 * CALCULATE INTERACTIONS *
479 **************************/
481 r10 = _mm_mul_pd(rsq10,rinv10);
483 /* Compute parameters for interactions between i and j atoms */
484 qq10 = _mm_mul_pd(iq1,jq0);
486 /* Calculate table index by multiplying r with table scale and truncate to integer */
487 rt = _mm_mul_pd(r10,vftabscale);
488 vfitab = _mm_cvttpd_epi32(rt);
490 vfeps = _mm_frcz_pd(rt);
492 vfeps = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
494 twovfeps = _mm_add_pd(vfeps,vfeps);
495 vfitab = _mm_slli_epi32(vfitab,2);
497 /* CUBIC SPLINE TABLE ELECTROSTATICS */
498 Y = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
499 F = _mm_setzero_pd();
500 GMX_MM_TRANSPOSE2_PD(Y,F);
501 G = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
502 H = _mm_setzero_pd();
503 GMX_MM_TRANSPOSE2_PD(G,H);
504 Fp = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
505 VV = _mm_macc_pd(vfeps,Fp,Y);
506 velec = _mm_mul_pd(qq10,VV);
507 FF = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
508 felec = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq10,FF),_mm_mul_pd(vftabscale,rinv10)));
510 /* Update potential sum for this i atom from the interaction with this j atom. */
511 velec = _mm_unpacklo_pd(velec,_mm_setzero_pd());
512 velecsum = _mm_add_pd(velecsum,velec);
516 fscal = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
518 /* Update vectorial force */
519 fix1 = _mm_macc_pd(dx10,fscal,fix1);
520 fiy1 = _mm_macc_pd(dy10,fscal,fiy1);
521 fiz1 = _mm_macc_pd(dz10,fscal,fiz1);
523 fjx0 = _mm_macc_pd(dx10,fscal,fjx0);
524 fjy0 = _mm_macc_pd(dy10,fscal,fjy0);
525 fjz0 = _mm_macc_pd(dz10,fscal,fjz0);
527 /**************************
528 * CALCULATE INTERACTIONS *
529 **************************/
531 r20 = _mm_mul_pd(rsq20,rinv20);
533 /* Compute parameters for interactions between i and j atoms */
534 qq20 = _mm_mul_pd(iq2,jq0);
536 /* Calculate table index by multiplying r with table scale and truncate to integer */
537 rt = _mm_mul_pd(r20,vftabscale);
538 vfitab = _mm_cvttpd_epi32(rt);
540 vfeps = _mm_frcz_pd(rt);
542 vfeps = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
544 twovfeps = _mm_add_pd(vfeps,vfeps);
545 vfitab = _mm_slli_epi32(vfitab,2);
547 /* CUBIC SPLINE TABLE ELECTROSTATICS */
548 Y = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
549 F = _mm_setzero_pd();
550 GMX_MM_TRANSPOSE2_PD(Y,F);
551 G = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
552 H = _mm_setzero_pd();
553 GMX_MM_TRANSPOSE2_PD(G,H);
554 Fp = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
555 VV = _mm_macc_pd(vfeps,Fp,Y);
556 velec = _mm_mul_pd(qq20,VV);
557 FF = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
558 felec = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq20,FF),_mm_mul_pd(vftabscale,rinv20)));
560 /* Update potential sum for this i atom from the interaction with this j atom. */
561 velec = _mm_unpacklo_pd(velec,_mm_setzero_pd());
562 velecsum = _mm_add_pd(velecsum,velec);
566 fscal = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
568 /* Update vectorial force */
569 fix2 = _mm_macc_pd(dx20,fscal,fix2);
570 fiy2 = _mm_macc_pd(dy20,fscal,fiy2);
571 fiz2 = _mm_macc_pd(dz20,fscal,fiz2);
573 fjx0 = _mm_macc_pd(dx20,fscal,fjx0);
574 fjy0 = _mm_macc_pd(dy20,fscal,fjy0);
575 fjz0 = _mm_macc_pd(dz20,fscal,fjz0);
577 gmx_mm_decrement_1rvec_1ptr_swizzle_pd(f+j_coord_offsetA,fjx0,fjy0,fjz0);
579 /* Inner loop uses 154 flops */
582 /* End of innermost loop */
584 gmx_mm_update_iforce_3atom_swizzle_pd(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,
585 f+i_coord_offset,fshift+i_shift_offset);
588 /* Update potential energies */
589 gmx_mm_update_1pot_pd(velecsum,kernel_data->energygrp_elec+ggid);
590 gmx_mm_update_1pot_pd(vvdwsum,kernel_data->energygrp_vdw+ggid);
592 /* Increment number of inner iterations */
593 inneriter += j_index_end - j_index_start;
595 /* Outer loop uses 20 flops */
598 /* Increment number of outer iterations */
601 /* Update outer/inner flops */
603 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3_VF,outeriter*20 + inneriter*154);
606 * Gromacs nonbonded kernel: nb_kernel_ElecCSTab_VdwLJ_GeomW3P1_F_avx_128_fma_double
607 * Electrostatics interaction: CubicSplineTable
608 * VdW interaction: LennardJones
609 * Geometry: Water3-Particle
610 * Calculate force/pot: Force
613 nb_kernel_ElecCSTab_VdwLJ_GeomW3P1_F_avx_128_fma_double
614 (t_nblist * gmx_restrict nlist,
615 rvec * gmx_restrict xx,
616 rvec * gmx_restrict ff,
617 t_forcerec * gmx_restrict fr,
618 t_mdatoms * gmx_restrict mdatoms,
619 nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
620 t_nrnb * gmx_restrict nrnb)
622 /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
623 * just 0 for non-waters.
624 * Suffixes A,B refer to j loop unrolling done with SSE double precision, e.g. for the two different
625 * jnr indices corresponding to data put in the four positions in the SIMD register.
627 int i_shift_offset,i_coord_offset,outeriter,inneriter;
628 int j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
630 int j_coord_offsetA,j_coord_offsetB;
631 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
633 real *shiftvec,*fshift,*x,*f;
634 __m128d tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
636 __m128d ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
638 __m128d ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
640 __m128d ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
641 int vdwjidx0A,vdwjidx0B;
642 __m128d jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
643 __m128d dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
644 __m128d dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
645 __m128d dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
646 __m128d velec,felec,velecsum,facel,crf,krf,krf2;
649 __m128d rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
652 __m128d one_sixth = _mm_set1_pd(1.0/6.0);
653 __m128d one_twelfth = _mm_set1_pd(1.0/12.0);
655 __m128i ifour = _mm_set1_epi32(4);
656 __m128d rt,vfeps,vftabscale,Y,F,G,H,Heps,Fp,VV,FF,twovfeps;
658 __m128d dummy_mask,cutoff_mask;
659 __m128d signbit = gmx_mm_castsi128_pd( _mm_set_epi32(0x80000000,0x00000000,0x80000000,0x00000000) );
660 __m128d one = _mm_set1_pd(1.0);
661 __m128d two = _mm_set1_pd(2.0);
667 jindex = nlist->jindex;
669 shiftidx = nlist->shift;
671 shiftvec = fr->shift_vec[0];
672 fshift = fr->fshift[0];
673 facel = _mm_set1_pd(fr->epsfac);
674 charge = mdatoms->chargeA;
675 nvdwtype = fr->ntype;
677 vdwtype = mdatoms->typeA;
679 vftab = kernel_data->table_elec->data;
680 vftabscale = _mm_set1_pd(kernel_data->table_elec->scale);
682 /* Setup water-specific parameters */
683 inr = nlist->iinr[0];
684 iq0 = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+0]));
685 iq1 = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+1]));
686 iq2 = _mm_mul_pd(facel,_mm_set1_pd(charge[inr+2]));
687 vdwioffset0 = 2*nvdwtype*vdwtype[inr+0];
689 /* Avoid stupid compiler warnings */
697 /* Start outer loop over neighborlists */
698 for(iidx=0; iidx<nri; iidx++)
700 /* Load shift vector for this list */
701 i_shift_offset = DIM*shiftidx[iidx];
703 /* Load limits for loop over neighbors */
704 j_index_start = jindex[iidx];
705 j_index_end = jindex[iidx+1];
707 /* Get outer coordinate index */
709 i_coord_offset = DIM*inr;
711 /* Load i particle coords and add shift vector */
712 gmx_mm_load_shift_and_3rvec_broadcast_pd(shiftvec+i_shift_offset,x+i_coord_offset,
713 &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2);
715 fix0 = _mm_setzero_pd();
716 fiy0 = _mm_setzero_pd();
717 fiz0 = _mm_setzero_pd();
718 fix1 = _mm_setzero_pd();
719 fiy1 = _mm_setzero_pd();
720 fiz1 = _mm_setzero_pd();
721 fix2 = _mm_setzero_pd();
722 fiy2 = _mm_setzero_pd();
723 fiz2 = _mm_setzero_pd();
725 /* Start inner kernel loop */
726 for(jidx=j_index_start; jidx<j_index_end-1; jidx+=2)
729 /* Get j neighbor index, and coordinate index */
732 j_coord_offsetA = DIM*jnrA;
733 j_coord_offsetB = DIM*jnrB;
735 /* load j atom coordinates */
736 gmx_mm_load_1rvec_2ptr_swizzle_pd(x+j_coord_offsetA,x+j_coord_offsetB,
739 /* Calculate displacement vector */
740 dx00 = _mm_sub_pd(ix0,jx0);
741 dy00 = _mm_sub_pd(iy0,jy0);
742 dz00 = _mm_sub_pd(iz0,jz0);
743 dx10 = _mm_sub_pd(ix1,jx0);
744 dy10 = _mm_sub_pd(iy1,jy0);
745 dz10 = _mm_sub_pd(iz1,jz0);
746 dx20 = _mm_sub_pd(ix2,jx0);
747 dy20 = _mm_sub_pd(iy2,jy0);
748 dz20 = _mm_sub_pd(iz2,jz0);
750 /* Calculate squared distance and things based on it */
751 rsq00 = gmx_mm_calc_rsq_pd(dx00,dy00,dz00);
752 rsq10 = gmx_mm_calc_rsq_pd(dx10,dy10,dz10);
753 rsq20 = gmx_mm_calc_rsq_pd(dx20,dy20,dz20);
755 rinv00 = gmx_mm_invsqrt_pd(rsq00);
756 rinv10 = gmx_mm_invsqrt_pd(rsq10);
757 rinv20 = gmx_mm_invsqrt_pd(rsq20);
759 rinvsq00 = _mm_mul_pd(rinv00,rinv00);
761 /* Load parameters for j particles */
762 jq0 = gmx_mm_load_2real_swizzle_pd(charge+jnrA+0,charge+jnrB+0);
763 vdwjidx0A = 2*vdwtype[jnrA+0];
764 vdwjidx0B = 2*vdwtype[jnrB+0];
766 fjx0 = _mm_setzero_pd();
767 fjy0 = _mm_setzero_pd();
768 fjz0 = _mm_setzero_pd();
770 /**************************
771 * CALCULATE INTERACTIONS *
772 **************************/
774 r00 = _mm_mul_pd(rsq00,rinv00);
776 /* Compute parameters for interactions between i and j atoms */
777 qq00 = _mm_mul_pd(iq0,jq0);
778 gmx_mm_load_2pair_swizzle_pd(vdwparam+vdwioffset0+vdwjidx0A,
779 vdwparam+vdwioffset0+vdwjidx0B,&c6_00,&c12_00);
781 /* Calculate table index by multiplying r with table scale and truncate to integer */
782 rt = _mm_mul_pd(r00,vftabscale);
783 vfitab = _mm_cvttpd_epi32(rt);
785 vfeps = _mm_frcz_pd(rt);
787 vfeps = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
789 twovfeps = _mm_add_pd(vfeps,vfeps);
790 vfitab = _mm_slli_epi32(vfitab,2);
792 /* CUBIC SPLINE TABLE ELECTROSTATICS */
793 Y = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
794 F = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) );
795 GMX_MM_TRANSPOSE2_PD(Y,F);
796 G = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
797 H = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) +2);
798 GMX_MM_TRANSPOSE2_PD(G,H);
799 Fp = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
800 FF = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
801 felec = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq00,FF),_mm_mul_pd(vftabscale,rinv00)));
803 /* LENNARD-JONES DISPERSION/REPULSION */
805 rinvsix = _mm_mul_pd(_mm_mul_pd(rinvsq00,rinvsq00),rinvsq00);
806 fvdw = _mm_mul_pd(_mm_msub_pd(c12_00,rinvsix,c6_00),_mm_mul_pd(rinvsix,rinvsq00));
808 fscal = _mm_add_pd(felec,fvdw);
810 /* Update vectorial force */
811 fix0 = _mm_macc_pd(dx00,fscal,fix0);
812 fiy0 = _mm_macc_pd(dy00,fscal,fiy0);
813 fiz0 = _mm_macc_pd(dz00,fscal,fiz0);
815 fjx0 = _mm_macc_pd(dx00,fscal,fjx0);
816 fjy0 = _mm_macc_pd(dy00,fscal,fjy0);
817 fjz0 = _mm_macc_pd(dz00,fscal,fjz0);
819 /**************************
820 * CALCULATE INTERACTIONS *
821 **************************/
823 r10 = _mm_mul_pd(rsq10,rinv10);
825 /* Compute parameters for interactions between i and j atoms */
826 qq10 = _mm_mul_pd(iq1,jq0);
828 /* Calculate table index by multiplying r with table scale and truncate to integer */
829 rt = _mm_mul_pd(r10,vftabscale);
830 vfitab = _mm_cvttpd_epi32(rt);
832 vfeps = _mm_frcz_pd(rt);
834 vfeps = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
836 twovfeps = _mm_add_pd(vfeps,vfeps);
837 vfitab = _mm_slli_epi32(vfitab,2);
839 /* CUBIC SPLINE TABLE ELECTROSTATICS */
840 Y = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
841 F = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) );
842 GMX_MM_TRANSPOSE2_PD(Y,F);
843 G = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
844 H = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) +2);
845 GMX_MM_TRANSPOSE2_PD(G,H);
846 Fp = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
847 FF = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
848 felec = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq10,FF),_mm_mul_pd(vftabscale,rinv10)));
852 /* Update vectorial force */
853 fix1 = _mm_macc_pd(dx10,fscal,fix1);
854 fiy1 = _mm_macc_pd(dy10,fscal,fiy1);
855 fiz1 = _mm_macc_pd(dz10,fscal,fiz1);
857 fjx0 = _mm_macc_pd(dx10,fscal,fjx0);
858 fjy0 = _mm_macc_pd(dy10,fscal,fjy0);
859 fjz0 = _mm_macc_pd(dz10,fscal,fjz0);
861 /**************************
862 * CALCULATE INTERACTIONS *
863 **************************/
865 r20 = _mm_mul_pd(rsq20,rinv20);
867 /* Compute parameters for interactions between i and j atoms */
868 qq20 = _mm_mul_pd(iq2,jq0);
870 /* Calculate table index by multiplying r with table scale and truncate to integer */
871 rt = _mm_mul_pd(r20,vftabscale);
872 vfitab = _mm_cvttpd_epi32(rt);
874 vfeps = _mm_frcz_pd(rt);
876 vfeps = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
878 twovfeps = _mm_add_pd(vfeps,vfeps);
879 vfitab = _mm_slli_epi32(vfitab,2);
881 /* CUBIC SPLINE TABLE ELECTROSTATICS */
882 Y = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
883 F = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) );
884 GMX_MM_TRANSPOSE2_PD(Y,F);
885 G = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
886 H = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,1) +2);
887 GMX_MM_TRANSPOSE2_PD(G,H);
888 Fp = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
889 FF = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
890 felec = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq20,FF),_mm_mul_pd(vftabscale,rinv20)));
894 /* Update vectorial force */
895 fix2 = _mm_macc_pd(dx20,fscal,fix2);
896 fiy2 = _mm_macc_pd(dy20,fscal,fiy2);
897 fiz2 = _mm_macc_pd(dz20,fscal,fiz2);
899 fjx0 = _mm_macc_pd(dx20,fscal,fjx0);
900 fjy0 = _mm_macc_pd(dy20,fscal,fjy0);
901 fjz0 = _mm_macc_pd(dz20,fscal,fjz0);
903 gmx_mm_decrement_1rvec_2ptr_swizzle_pd(f+j_coord_offsetA,f+j_coord_offsetB,fjx0,fjy0,fjz0);
905 /* Inner loop uses 137 flops */
912 j_coord_offsetA = DIM*jnrA;
914 /* load j atom coordinates */
915 gmx_mm_load_1rvec_1ptr_swizzle_pd(x+j_coord_offsetA,
918 /* Calculate displacement vector */
919 dx00 = _mm_sub_pd(ix0,jx0);
920 dy00 = _mm_sub_pd(iy0,jy0);
921 dz00 = _mm_sub_pd(iz0,jz0);
922 dx10 = _mm_sub_pd(ix1,jx0);
923 dy10 = _mm_sub_pd(iy1,jy0);
924 dz10 = _mm_sub_pd(iz1,jz0);
925 dx20 = _mm_sub_pd(ix2,jx0);
926 dy20 = _mm_sub_pd(iy2,jy0);
927 dz20 = _mm_sub_pd(iz2,jz0);
929 /* Calculate squared distance and things based on it */
930 rsq00 = gmx_mm_calc_rsq_pd(dx00,dy00,dz00);
931 rsq10 = gmx_mm_calc_rsq_pd(dx10,dy10,dz10);
932 rsq20 = gmx_mm_calc_rsq_pd(dx20,dy20,dz20);
934 rinv00 = gmx_mm_invsqrt_pd(rsq00);
935 rinv10 = gmx_mm_invsqrt_pd(rsq10);
936 rinv20 = gmx_mm_invsqrt_pd(rsq20);
938 rinvsq00 = _mm_mul_pd(rinv00,rinv00);
940 /* Load parameters for j particles */
941 jq0 = _mm_load_sd(charge+jnrA+0);
942 vdwjidx0A = 2*vdwtype[jnrA+0];
944 fjx0 = _mm_setzero_pd();
945 fjy0 = _mm_setzero_pd();
946 fjz0 = _mm_setzero_pd();
948 /**************************
949 * CALCULATE INTERACTIONS *
950 **************************/
952 r00 = _mm_mul_pd(rsq00,rinv00);
954 /* Compute parameters for interactions between i and j atoms */
955 qq00 = _mm_mul_pd(iq0,jq0);
956 gmx_mm_load_1pair_swizzle_pd(vdwparam+vdwioffset0+vdwjidx0A,&c6_00,&c12_00);
958 /* Calculate table index by multiplying r with table scale and truncate to integer */
959 rt = _mm_mul_pd(r00,vftabscale);
960 vfitab = _mm_cvttpd_epi32(rt);
962 vfeps = _mm_frcz_pd(rt);
964 vfeps = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
966 twovfeps = _mm_add_pd(vfeps,vfeps);
967 vfitab = _mm_slli_epi32(vfitab,2);
969 /* CUBIC SPLINE TABLE ELECTROSTATICS */
970 Y = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
971 F = _mm_setzero_pd();
972 GMX_MM_TRANSPOSE2_PD(Y,F);
973 G = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
974 H = _mm_setzero_pd();
975 GMX_MM_TRANSPOSE2_PD(G,H);
976 Fp = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
977 FF = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
978 felec = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq00,FF),_mm_mul_pd(vftabscale,rinv00)));
980 /* LENNARD-JONES DISPERSION/REPULSION */
982 rinvsix = _mm_mul_pd(_mm_mul_pd(rinvsq00,rinvsq00),rinvsq00);
983 fvdw = _mm_mul_pd(_mm_msub_pd(c12_00,rinvsix,c6_00),_mm_mul_pd(rinvsix,rinvsq00));
985 fscal = _mm_add_pd(felec,fvdw);
987 fscal = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
989 /* Update vectorial force */
990 fix0 = _mm_macc_pd(dx00,fscal,fix0);
991 fiy0 = _mm_macc_pd(dy00,fscal,fiy0);
992 fiz0 = _mm_macc_pd(dz00,fscal,fiz0);
994 fjx0 = _mm_macc_pd(dx00,fscal,fjx0);
995 fjy0 = _mm_macc_pd(dy00,fscal,fjy0);
996 fjz0 = _mm_macc_pd(dz00,fscal,fjz0);
998 /**************************
999 * CALCULATE INTERACTIONS *
1000 **************************/
1002 r10 = _mm_mul_pd(rsq10,rinv10);
1004 /* Compute parameters for interactions between i and j atoms */
1005 qq10 = _mm_mul_pd(iq1,jq0);
1007 /* Calculate table index by multiplying r with table scale and truncate to integer */
1008 rt = _mm_mul_pd(r10,vftabscale);
1009 vfitab = _mm_cvttpd_epi32(rt);
1011 vfeps = _mm_frcz_pd(rt);
1013 vfeps = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
1015 twovfeps = _mm_add_pd(vfeps,vfeps);
1016 vfitab = _mm_slli_epi32(vfitab,2);
1018 /* CUBIC SPLINE TABLE ELECTROSTATICS */
1019 Y = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
1020 F = _mm_setzero_pd();
1021 GMX_MM_TRANSPOSE2_PD(Y,F);
1022 G = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
1023 H = _mm_setzero_pd();
1024 GMX_MM_TRANSPOSE2_PD(G,H);
1025 Fp = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
1026 FF = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
1027 felec = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq10,FF),_mm_mul_pd(vftabscale,rinv10)));
1031 fscal = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
1033 /* Update vectorial force */
1034 fix1 = _mm_macc_pd(dx10,fscal,fix1);
1035 fiy1 = _mm_macc_pd(dy10,fscal,fiy1);
1036 fiz1 = _mm_macc_pd(dz10,fscal,fiz1);
1038 fjx0 = _mm_macc_pd(dx10,fscal,fjx0);
1039 fjy0 = _mm_macc_pd(dy10,fscal,fjy0);
1040 fjz0 = _mm_macc_pd(dz10,fscal,fjz0);
1042 /**************************
1043 * CALCULATE INTERACTIONS *
1044 **************************/
1046 r20 = _mm_mul_pd(rsq20,rinv20);
1048 /* Compute parameters for interactions between i and j atoms */
1049 qq20 = _mm_mul_pd(iq2,jq0);
1051 /* Calculate table index by multiplying r with table scale and truncate to integer */
1052 rt = _mm_mul_pd(r20,vftabscale);
1053 vfitab = _mm_cvttpd_epi32(rt);
1055 vfeps = _mm_frcz_pd(rt);
1057 vfeps = _mm_sub_pd(rt,_mm_round_pd(rt, _MM_FROUND_FLOOR));
1059 twovfeps = _mm_add_pd(vfeps,vfeps);
1060 vfitab = _mm_slli_epi32(vfitab,2);
1062 /* CUBIC SPLINE TABLE ELECTROSTATICS */
1063 Y = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) );
1064 F = _mm_setzero_pd();
1065 GMX_MM_TRANSPOSE2_PD(Y,F);
1066 G = _mm_load_pd( vftab + _mm_extract_epi32(vfitab,0) +2);
1067 H = _mm_setzero_pd();
1068 GMX_MM_TRANSPOSE2_PD(G,H);
1069 Fp = _mm_macc_pd(vfeps,_mm_macc_pd(vfeps,H,G),F);
1070 FF = _mm_macc_pd(_mm_macc_pd(twovfeps,H,G),vfeps,Fp);
1071 felec = _mm_xor_pd(signbit,_mm_mul_pd(_mm_mul_pd(qq20,FF),_mm_mul_pd(vftabscale,rinv20)));
1075 fscal = _mm_unpacklo_pd(fscal,_mm_setzero_pd());
1077 /* Update vectorial force */
1078 fix2 = _mm_macc_pd(dx20,fscal,fix2);
1079 fiy2 = _mm_macc_pd(dy20,fscal,fiy2);
1080 fiz2 = _mm_macc_pd(dz20,fscal,fiz2);
1082 fjx0 = _mm_macc_pd(dx20,fscal,fjx0);
1083 fjy0 = _mm_macc_pd(dy20,fscal,fjy0);
1084 fjz0 = _mm_macc_pd(dz20,fscal,fjz0);
1086 gmx_mm_decrement_1rvec_1ptr_swizzle_pd(f+j_coord_offsetA,fjx0,fjy0,fjz0);
1088 /* Inner loop uses 137 flops */
1091 /* End of innermost loop */
1093 gmx_mm_update_iforce_3atom_swizzle_pd(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,
1094 f+i_coord_offset,fshift+i_shift_offset);
1096 /* Increment number of inner iterations */
1097 inneriter += j_index_end - j_index_start;
1099 /* Outer loop uses 18 flops */
1102 /* Increment number of outer iterations */
1105 /* Update outer/inner flops */
1107 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W3_F,outeriter*18 + inneriter*137);