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 sse4_1_single kernel generator.
42 #include "../nb_kernel.h"
43 #include "gromacs/legacyheaders/types/simple.h"
44 #include "gromacs/math/vec.h"
45 #include "gromacs/legacyheaders/nrnb.h"
47 #include "gromacs/simd/math_x86_sse4_1_single.h"
48 #include "kernelutil_x86_sse4_1_single.h"
51 * Gromacs nonbonded kernel: nb_kernel_ElecRF_VdwCSTab_GeomW4P1_VF_sse4_1_single
52 * Electrostatics interaction: ReactionField
53 * VdW interaction: CubicSplineTable
54 * Geometry: Water4-Particle
55 * Calculate force/pot: PotentialAndForce
58 nb_kernel_ElecRF_VdwCSTab_GeomW4P1_VF_sse4_1_single
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 SSE, 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 j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
77 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
79 real *shiftvec,*fshift,*x,*f;
80 real *fjptrA,*fjptrB,*fjptrC,*fjptrD;
82 __m128 tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
84 __m128 ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
86 __m128 ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
88 __m128 ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
90 __m128 ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
91 int vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D;
92 __m128 jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
93 __m128 dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
94 __m128 dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
95 __m128 dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
96 __m128 dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30;
97 __m128 velec,felec,velecsum,facel,crf,krf,krf2;
100 __m128 rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
103 __m128 one_sixth = _mm_set1_ps(1.0/6.0);
104 __m128 one_twelfth = _mm_set1_ps(1.0/12.0);
106 __m128i ifour = _mm_set1_epi32(4);
107 __m128 rt,vfeps,vftabscale,Y,F,G,H,Heps,Fp,VV,FF;
109 __m128 dummy_mask,cutoff_mask;
110 __m128 signbit = _mm_castsi128_ps( _mm_set1_epi32(0x80000000) );
111 __m128 one = _mm_set1_ps(1.0);
112 __m128 two = _mm_set1_ps(2.0);
118 jindex = nlist->jindex;
120 shiftidx = nlist->shift;
122 shiftvec = fr->shift_vec[0];
123 fshift = fr->fshift[0];
124 facel = _mm_set1_ps(fr->epsfac);
125 charge = mdatoms->chargeA;
126 krf = _mm_set1_ps(fr->ic->k_rf);
127 krf2 = _mm_set1_ps(fr->ic->k_rf*2.0);
128 crf = _mm_set1_ps(fr->ic->c_rf);
129 nvdwtype = fr->ntype;
131 vdwtype = mdatoms->typeA;
133 vftab = kernel_data->table_vdw->data;
134 vftabscale = _mm_set1_ps(kernel_data->table_vdw->scale);
136 /* Setup water-specific parameters */
137 inr = nlist->iinr[0];
138 iq1 = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+1]));
139 iq2 = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+2]));
140 iq3 = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+3]));
141 vdwioffset0 = 2*nvdwtype*vdwtype[inr+0];
143 /* Avoid stupid compiler warnings */
144 jnrA = jnrB = jnrC = jnrD = 0;
153 for(iidx=0;iidx<4*DIM;iidx++)
158 /* Start outer loop over neighborlists */
159 for(iidx=0; iidx<nri; iidx++)
161 /* Load shift vector for this list */
162 i_shift_offset = DIM*shiftidx[iidx];
164 /* Load limits for loop over neighbors */
165 j_index_start = jindex[iidx];
166 j_index_end = jindex[iidx+1];
168 /* Get outer coordinate index */
170 i_coord_offset = DIM*inr;
172 /* Load i particle coords and add shift vector */
173 gmx_mm_load_shift_and_4rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,
174 &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
176 fix0 = _mm_setzero_ps();
177 fiy0 = _mm_setzero_ps();
178 fiz0 = _mm_setzero_ps();
179 fix1 = _mm_setzero_ps();
180 fiy1 = _mm_setzero_ps();
181 fiz1 = _mm_setzero_ps();
182 fix2 = _mm_setzero_ps();
183 fiy2 = _mm_setzero_ps();
184 fiz2 = _mm_setzero_ps();
185 fix3 = _mm_setzero_ps();
186 fiy3 = _mm_setzero_ps();
187 fiz3 = _mm_setzero_ps();
189 /* Reset potential sums */
190 velecsum = _mm_setzero_ps();
191 vvdwsum = _mm_setzero_ps();
193 /* Start inner kernel loop */
194 for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4)
197 /* Get j neighbor index, and coordinate index */
202 j_coord_offsetA = DIM*jnrA;
203 j_coord_offsetB = DIM*jnrB;
204 j_coord_offsetC = DIM*jnrC;
205 j_coord_offsetD = DIM*jnrD;
207 /* load j atom coordinates */
208 gmx_mm_load_1rvec_4ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
209 x+j_coord_offsetC,x+j_coord_offsetD,
212 /* Calculate displacement vector */
213 dx00 = _mm_sub_ps(ix0,jx0);
214 dy00 = _mm_sub_ps(iy0,jy0);
215 dz00 = _mm_sub_ps(iz0,jz0);
216 dx10 = _mm_sub_ps(ix1,jx0);
217 dy10 = _mm_sub_ps(iy1,jy0);
218 dz10 = _mm_sub_ps(iz1,jz0);
219 dx20 = _mm_sub_ps(ix2,jx0);
220 dy20 = _mm_sub_ps(iy2,jy0);
221 dz20 = _mm_sub_ps(iz2,jz0);
222 dx30 = _mm_sub_ps(ix3,jx0);
223 dy30 = _mm_sub_ps(iy3,jy0);
224 dz30 = _mm_sub_ps(iz3,jz0);
226 /* Calculate squared distance and things based on it */
227 rsq00 = gmx_mm_calc_rsq_ps(dx00,dy00,dz00);
228 rsq10 = gmx_mm_calc_rsq_ps(dx10,dy10,dz10);
229 rsq20 = gmx_mm_calc_rsq_ps(dx20,dy20,dz20);
230 rsq30 = gmx_mm_calc_rsq_ps(dx30,dy30,dz30);
232 rinv00 = gmx_mm_invsqrt_ps(rsq00);
233 rinv10 = gmx_mm_invsqrt_ps(rsq10);
234 rinv20 = gmx_mm_invsqrt_ps(rsq20);
235 rinv30 = gmx_mm_invsqrt_ps(rsq30);
237 rinvsq10 = _mm_mul_ps(rinv10,rinv10);
238 rinvsq20 = _mm_mul_ps(rinv20,rinv20);
239 rinvsq30 = _mm_mul_ps(rinv30,rinv30);
241 /* Load parameters for j particles */
242 jq0 = gmx_mm_load_4real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
243 charge+jnrC+0,charge+jnrD+0);
244 vdwjidx0A = 2*vdwtype[jnrA+0];
245 vdwjidx0B = 2*vdwtype[jnrB+0];
246 vdwjidx0C = 2*vdwtype[jnrC+0];
247 vdwjidx0D = 2*vdwtype[jnrD+0];
249 fjx0 = _mm_setzero_ps();
250 fjy0 = _mm_setzero_ps();
251 fjz0 = _mm_setzero_ps();
253 /**************************
254 * CALCULATE INTERACTIONS *
255 **************************/
257 r00 = _mm_mul_ps(rsq00,rinv00);
259 /* Compute parameters for interactions between i and j atoms */
260 gmx_mm_load_4pair_swizzle_ps(vdwparam+vdwioffset0+vdwjidx0A,
261 vdwparam+vdwioffset0+vdwjidx0B,
262 vdwparam+vdwioffset0+vdwjidx0C,
263 vdwparam+vdwioffset0+vdwjidx0D,
266 /* Calculate table index by multiplying r with table scale and truncate to integer */
267 rt = _mm_mul_ps(r00,vftabscale);
268 vfitab = _mm_cvttps_epi32(rt);
269 vfeps = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
270 vfitab = _mm_slli_epi32(vfitab,3);
272 /* CUBIC SPLINE TABLE DISPERSION */
273 Y = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,0) );
274 F = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,1) );
275 G = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,2) );
276 H = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,3) );
277 _MM_TRANSPOSE4_PS(Y,F,G,H);
278 Heps = _mm_mul_ps(vfeps,H);
279 Fp = _mm_add_ps(F,_mm_mul_ps(vfeps,_mm_add_ps(G,Heps)));
280 VV = _mm_add_ps(Y,_mm_mul_ps(vfeps,Fp));
281 vvdw6 = _mm_mul_ps(c6_00,VV);
282 FF = _mm_add_ps(Fp,_mm_mul_ps(vfeps,_mm_add_ps(G,_mm_add_ps(Heps,Heps))));
283 fvdw6 = _mm_mul_ps(c6_00,FF);
285 /* CUBIC SPLINE TABLE REPULSION */
286 vfitab = _mm_add_epi32(vfitab,ifour);
287 Y = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,0) );
288 F = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,1) );
289 G = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,2) );
290 H = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,3) );
291 _MM_TRANSPOSE4_PS(Y,F,G,H);
292 Heps = _mm_mul_ps(vfeps,H);
293 Fp = _mm_add_ps(F,_mm_mul_ps(vfeps,_mm_add_ps(G,Heps)));
294 VV = _mm_add_ps(Y,_mm_mul_ps(vfeps,Fp));
295 vvdw12 = _mm_mul_ps(c12_00,VV);
296 FF = _mm_add_ps(Fp,_mm_mul_ps(vfeps,_mm_add_ps(G,_mm_add_ps(Heps,Heps))));
297 fvdw12 = _mm_mul_ps(c12_00,FF);
298 vvdw = _mm_add_ps(vvdw12,vvdw6);
299 fvdw = _mm_xor_ps(signbit,_mm_mul_ps(_mm_add_ps(fvdw6,fvdw12),_mm_mul_ps(vftabscale,rinv00)));
301 /* Update potential sum for this i atom from the interaction with this j atom. */
302 vvdwsum = _mm_add_ps(vvdwsum,vvdw);
306 /* Calculate temporary vectorial force */
307 tx = _mm_mul_ps(fscal,dx00);
308 ty = _mm_mul_ps(fscal,dy00);
309 tz = _mm_mul_ps(fscal,dz00);
311 /* Update vectorial force */
312 fix0 = _mm_add_ps(fix0,tx);
313 fiy0 = _mm_add_ps(fiy0,ty);
314 fiz0 = _mm_add_ps(fiz0,tz);
316 fjx0 = _mm_add_ps(fjx0,tx);
317 fjy0 = _mm_add_ps(fjy0,ty);
318 fjz0 = _mm_add_ps(fjz0,tz);
320 /**************************
321 * CALCULATE INTERACTIONS *
322 **************************/
324 /* Compute parameters for interactions between i and j atoms */
325 qq10 = _mm_mul_ps(iq1,jq0);
327 /* REACTION-FIELD ELECTROSTATICS */
328 velec = _mm_mul_ps(qq10,_mm_sub_ps(_mm_add_ps(rinv10,_mm_mul_ps(krf,rsq10)),crf));
329 felec = _mm_mul_ps(qq10,_mm_sub_ps(_mm_mul_ps(rinv10,rinvsq10),krf2));
331 /* Update potential sum for this i atom from the interaction with this j atom. */
332 velecsum = _mm_add_ps(velecsum,velec);
336 /* Calculate temporary vectorial force */
337 tx = _mm_mul_ps(fscal,dx10);
338 ty = _mm_mul_ps(fscal,dy10);
339 tz = _mm_mul_ps(fscal,dz10);
341 /* Update vectorial force */
342 fix1 = _mm_add_ps(fix1,tx);
343 fiy1 = _mm_add_ps(fiy1,ty);
344 fiz1 = _mm_add_ps(fiz1,tz);
346 fjx0 = _mm_add_ps(fjx0,tx);
347 fjy0 = _mm_add_ps(fjy0,ty);
348 fjz0 = _mm_add_ps(fjz0,tz);
350 /**************************
351 * CALCULATE INTERACTIONS *
352 **************************/
354 /* Compute parameters for interactions between i and j atoms */
355 qq20 = _mm_mul_ps(iq2,jq0);
357 /* REACTION-FIELD ELECTROSTATICS */
358 velec = _mm_mul_ps(qq20,_mm_sub_ps(_mm_add_ps(rinv20,_mm_mul_ps(krf,rsq20)),crf));
359 felec = _mm_mul_ps(qq20,_mm_sub_ps(_mm_mul_ps(rinv20,rinvsq20),krf2));
361 /* Update potential sum for this i atom from the interaction with this j atom. */
362 velecsum = _mm_add_ps(velecsum,velec);
366 /* Calculate temporary vectorial force */
367 tx = _mm_mul_ps(fscal,dx20);
368 ty = _mm_mul_ps(fscal,dy20);
369 tz = _mm_mul_ps(fscal,dz20);
371 /* Update vectorial force */
372 fix2 = _mm_add_ps(fix2,tx);
373 fiy2 = _mm_add_ps(fiy2,ty);
374 fiz2 = _mm_add_ps(fiz2,tz);
376 fjx0 = _mm_add_ps(fjx0,tx);
377 fjy0 = _mm_add_ps(fjy0,ty);
378 fjz0 = _mm_add_ps(fjz0,tz);
380 /**************************
381 * CALCULATE INTERACTIONS *
382 **************************/
384 /* Compute parameters for interactions between i and j atoms */
385 qq30 = _mm_mul_ps(iq3,jq0);
387 /* REACTION-FIELD ELECTROSTATICS */
388 velec = _mm_mul_ps(qq30,_mm_sub_ps(_mm_add_ps(rinv30,_mm_mul_ps(krf,rsq30)),crf));
389 felec = _mm_mul_ps(qq30,_mm_sub_ps(_mm_mul_ps(rinv30,rinvsq30),krf2));
391 /* Update potential sum for this i atom from the interaction with this j atom. */
392 velecsum = _mm_add_ps(velecsum,velec);
396 /* Calculate temporary vectorial force */
397 tx = _mm_mul_ps(fscal,dx30);
398 ty = _mm_mul_ps(fscal,dy30);
399 tz = _mm_mul_ps(fscal,dz30);
401 /* Update vectorial force */
402 fix3 = _mm_add_ps(fix3,tx);
403 fiy3 = _mm_add_ps(fiy3,ty);
404 fiz3 = _mm_add_ps(fiz3,tz);
406 fjx0 = _mm_add_ps(fjx0,tx);
407 fjy0 = _mm_add_ps(fjy0,ty);
408 fjz0 = _mm_add_ps(fjz0,tz);
410 fjptrA = f+j_coord_offsetA;
411 fjptrB = f+j_coord_offsetB;
412 fjptrC = f+j_coord_offsetC;
413 fjptrD = f+j_coord_offsetD;
415 gmx_mm_decrement_1rvec_4ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0);
417 /* Inner loop uses 152 flops */
423 /* Get j neighbor index, and coordinate index */
424 jnrlistA = jjnr[jidx];
425 jnrlistB = jjnr[jidx+1];
426 jnrlistC = jjnr[jidx+2];
427 jnrlistD = jjnr[jidx+3];
428 /* Sign of each element will be negative for non-real atoms.
429 * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
430 * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
432 dummy_mask = gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128()));
433 jnrA = (jnrlistA>=0) ? jnrlistA : 0;
434 jnrB = (jnrlistB>=0) ? jnrlistB : 0;
435 jnrC = (jnrlistC>=0) ? jnrlistC : 0;
436 jnrD = (jnrlistD>=0) ? jnrlistD : 0;
437 j_coord_offsetA = DIM*jnrA;
438 j_coord_offsetB = DIM*jnrB;
439 j_coord_offsetC = DIM*jnrC;
440 j_coord_offsetD = DIM*jnrD;
442 /* load j atom coordinates */
443 gmx_mm_load_1rvec_4ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
444 x+j_coord_offsetC,x+j_coord_offsetD,
447 /* Calculate displacement vector */
448 dx00 = _mm_sub_ps(ix0,jx0);
449 dy00 = _mm_sub_ps(iy0,jy0);
450 dz00 = _mm_sub_ps(iz0,jz0);
451 dx10 = _mm_sub_ps(ix1,jx0);
452 dy10 = _mm_sub_ps(iy1,jy0);
453 dz10 = _mm_sub_ps(iz1,jz0);
454 dx20 = _mm_sub_ps(ix2,jx0);
455 dy20 = _mm_sub_ps(iy2,jy0);
456 dz20 = _mm_sub_ps(iz2,jz0);
457 dx30 = _mm_sub_ps(ix3,jx0);
458 dy30 = _mm_sub_ps(iy3,jy0);
459 dz30 = _mm_sub_ps(iz3,jz0);
461 /* Calculate squared distance and things based on it */
462 rsq00 = gmx_mm_calc_rsq_ps(dx00,dy00,dz00);
463 rsq10 = gmx_mm_calc_rsq_ps(dx10,dy10,dz10);
464 rsq20 = gmx_mm_calc_rsq_ps(dx20,dy20,dz20);
465 rsq30 = gmx_mm_calc_rsq_ps(dx30,dy30,dz30);
467 rinv00 = gmx_mm_invsqrt_ps(rsq00);
468 rinv10 = gmx_mm_invsqrt_ps(rsq10);
469 rinv20 = gmx_mm_invsqrt_ps(rsq20);
470 rinv30 = gmx_mm_invsqrt_ps(rsq30);
472 rinvsq10 = _mm_mul_ps(rinv10,rinv10);
473 rinvsq20 = _mm_mul_ps(rinv20,rinv20);
474 rinvsq30 = _mm_mul_ps(rinv30,rinv30);
476 /* Load parameters for j particles */
477 jq0 = gmx_mm_load_4real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
478 charge+jnrC+0,charge+jnrD+0);
479 vdwjidx0A = 2*vdwtype[jnrA+0];
480 vdwjidx0B = 2*vdwtype[jnrB+0];
481 vdwjidx0C = 2*vdwtype[jnrC+0];
482 vdwjidx0D = 2*vdwtype[jnrD+0];
484 fjx0 = _mm_setzero_ps();
485 fjy0 = _mm_setzero_ps();
486 fjz0 = _mm_setzero_ps();
488 /**************************
489 * CALCULATE INTERACTIONS *
490 **************************/
492 r00 = _mm_mul_ps(rsq00,rinv00);
493 r00 = _mm_andnot_ps(dummy_mask,r00);
495 /* Compute parameters for interactions between i and j atoms */
496 gmx_mm_load_4pair_swizzle_ps(vdwparam+vdwioffset0+vdwjidx0A,
497 vdwparam+vdwioffset0+vdwjidx0B,
498 vdwparam+vdwioffset0+vdwjidx0C,
499 vdwparam+vdwioffset0+vdwjidx0D,
502 /* Calculate table index by multiplying r with table scale and truncate to integer */
503 rt = _mm_mul_ps(r00,vftabscale);
504 vfitab = _mm_cvttps_epi32(rt);
505 vfeps = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
506 vfitab = _mm_slli_epi32(vfitab,3);
508 /* CUBIC SPLINE TABLE DISPERSION */
509 Y = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,0) );
510 F = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,1) );
511 G = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,2) );
512 H = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,3) );
513 _MM_TRANSPOSE4_PS(Y,F,G,H);
514 Heps = _mm_mul_ps(vfeps,H);
515 Fp = _mm_add_ps(F,_mm_mul_ps(vfeps,_mm_add_ps(G,Heps)));
516 VV = _mm_add_ps(Y,_mm_mul_ps(vfeps,Fp));
517 vvdw6 = _mm_mul_ps(c6_00,VV);
518 FF = _mm_add_ps(Fp,_mm_mul_ps(vfeps,_mm_add_ps(G,_mm_add_ps(Heps,Heps))));
519 fvdw6 = _mm_mul_ps(c6_00,FF);
521 /* CUBIC SPLINE TABLE REPULSION */
522 vfitab = _mm_add_epi32(vfitab,ifour);
523 Y = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,0) );
524 F = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,1) );
525 G = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,2) );
526 H = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,3) );
527 _MM_TRANSPOSE4_PS(Y,F,G,H);
528 Heps = _mm_mul_ps(vfeps,H);
529 Fp = _mm_add_ps(F,_mm_mul_ps(vfeps,_mm_add_ps(G,Heps)));
530 VV = _mm_add_ps(Y,_mm_mul_ps(vfeps,Fp));
531 vvdw12 = _mm_mul_ps(c12_00,VV);
532 FF = _mm_add_ps(Fp,_mm_mul_ps(vfeps,_mm_add_ps(G,_mm_add_ps(Heps,Heps))));
533 fvdw12 = _mm_mul_ps(c12_00,FF);
534 vvdw = _mm_add_ps(vvdw12,vvdw6);
535 fvdw = _mm_xor_ps(signbit,_mm_mul_ps(_mm_add_ps(fvdw6,fvdw12),_mm_mul_ps(vftabscale,rinv00)));
537 /* Update potential sum for this i atom from the interaction with this j atom. */
538 vvdw = _mm_andnot_ps(dummy_mask,vvdw);
539 vvdwsum = _mm_add_ps(vvdwsum,vvdw);
543 fscal = _mm_andnot_ps(dummy_mask,fscal);
545 /* Calculate temporary vectorial force */
546 tx = _mm_mul_ps(fscal,dx00);
547 ty = _mm_mul_ps(fscal,dy00);
548 tz = _mm_mul_ps(fscal,dz00);
550 /* Update vectorial force */
551 fix0 = _mm_add_ps(fix0,tx);
552 fiy0 = _mm_add_ps(fiy0,ty);
553 fiz0 = _mm_add_ps(fiz0,tz);
555 fjx0 = _mm_add_ps(fjx0,tx);
556 fjy0 = _mm_add_ps(fjy0,ty);
557 fjz0 = _mm_add_ps(fjz0,tz);
559 /**************************
560 * CALCULATE INTERACTIONS *
561 **************************/
563 /* Compute parameters for interactions between i and j atoms */
564 qq10 = _mm_mul_ps(iq1,jq0);
566 /* REACTION-FIELD ELECTROSTATICS */
567 velec = _mm_mul_ps(qq10,_mm_sub_ps(_mm_add_ps(rinv10,_mm_mul_ps(krf,rsq10)),crf));
568 felec = _mm_mul_ps(qq10,_mm_sub_ps(_mm_mul_ps(rinv10,rinvsq10),krf2));
570 /* Update potential sum for this i atom from the interaction with this j atom. */
571 velec = _mm_andnot_ps(dummy_mask,velec);
572 velecsum = _mm_add_ps(velecsum,velec);
576 fscal = _mm_andnot_ps(dummy_mask,fscal);
578 /* Calculate temporary vectorial force */
579 tx = _mm_mul_ps(fscal,dx10);
580 ty = _mm_mul_ps(fscal,dy10);
581 tz = _mm_mul_ps(fscal,dz10);
583 /* Update vectorial force */
584 fix1 = _mm_add_ps(fix1,tx);
585 fiy1 = _mm_add_ps(fiy1,ty);
586 fiz1 = _mm_add_ps(fiz1,tz);
588 fjx0 = _mm_add_ps(fjx0,tx);
589 fjy0 = _mm_add_ps(fjy0,ty);
590 fjz0 = _mm_add_ps(fjz0,tz);
592 /**************************
593 * CALCULATE INTERACTIONS *
594 **************************/
596 /* Compute parameters for interactions between i and j atoms */
597 qq20 = _mm_mul_ps(iq2,jq0);
599 /* REACTION-FIELD ELECTROSTATICS */
600 velec = _mm_mul_ps(qq20,_mm_sub_ps(_mm_add_ps(rinv20,_mm_mul_ps(krf,rsq20)),crf));
601 felec = _mm_mul_ps(qq20,_mm_sub_ps(_mm_mul_ps(rinv20,rinvsq20),krf2));
603 /* Update potential sum for this i atom from the interaction with this j atom. */
604 velec = _mm_andnot_ps(dummy_mask,velec);
605 velecsum = _mm_add_ps(velecsum,velec);
609 fscal = _mm_andnot_ps(dummy_mask,fscal);
611 /* Calculate temporary vectorial force */
612 tx = _mm_mul_ps(fscal,dx20);
613 ty = _mm_mul_ps(fscal,dy20);
614 tz = _mm_mul_ps(fscal,dz20);
616 /* Update vectorial force */
617 fix2 = _mm_add_ps(fix2,tx);
618 fiy2 = _mm_add_ps(fiy2,ty);
619 fiz2 = _mm_add_ps(fiz2,tz);
621 fjx0 = _mm_add_ps(fjx0,tx);
622 fjy0 = _mm_add_ps(fjy0,ty);
623 fjz0 = _mm_add_ps(fjz0,tz);
625 /**************************
626 * CALCULATE INTERACTIONS *
627 **************************/
629 /* Compute parameters for interactions between i and j atoms */
630 qq30 = _mm_mul_ps(iq3,jq0);
632 /* REACTION-FIELD ELECTROSTATICS */
633 velec = _mm_mul_ps(qq30,_mm_sub_ps(_mm_add_ps(rinv30,_mm_mul_ps(krf,rsq30)),crf));
634 felec = _mm_mul_ps(qq30,_mm_sub_ps(_mm_mul_ps(rinv30,rinvsq30),krf2));
636 /* Update potential sum for this i atom from the interaction with this j atom. */
637 velec = _mm_andnot_ps(dummy_mask,velec);
638 velecsum = _mm_add_ps(velecsum,velec);
642 fscal = _mm_andnot_ps(dummy_mask,fscal);
644 /* Calculate temporary vectorial force */
645 tx = _mm_mul_ps(fscal,dx30);
646 ty = _mm_mul_ps(fscal,dy30);
647 tz = _mm_mul_ps(fscal,dz30);
649 /* Update vectorial force */
650 fix3 = _mm_add_ps(fix3,tx);
651 fiy3 = _mm_add_ps(fiy3,ty);
652 fiz3 = _mm_add_ps(fiz3,tz);
654 fjx0 = _mm_add_ps(fjx0,tx);
655 fjy0 = _mm_add_ps(fjy0,ty);
656 fjz0 = _mm_add_ps(fjz0,tz);
658 fjptrA = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
659 fjptrB = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
660 fjptrC = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
661 fjptrD = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
663 gmx_mm_decrement_1rvec_4ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0);
665 /* Inner loop uses 153 flops */
668 /* End of innermost loop */
670 gmx_mm_update_iforce_4atom_swizzle_ps(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
671 f+i_coord_offset,fshift+i_shift_offset);
674 /* Update potential energies */
675 gmx_mm_update_1pot_ps(velecsum,kernel_data->energygrp_elec+ggid);
676 gmx_mm_update_1pot_ps(vvdwsum,kernel_data->energygrp_vdw+ggid);
678 /* Increment number of inner iterations */
679 inneriter += j_index_end - j_index_start;
681 /* Outer loop uses 26 flops */
684 /* Increment number of outer iterations */
687 /* Update outer/inner flops */
689 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_VF,outeriter*26 + inneriter*153);
692 * Gromacs nonbonded kernel: nb_kernel_ElecRF_VdwCSTab_GeomW4P1_F_sse4_1_single
693 * Electrostatics interaction: ReactionField
694 * VdW interaction: CubicSplineTable
695 * Geometry: Water4-Particle
696 * Calculate force/pot: Force
699 nb_kernel_ElecRF_VdwCSTab_GeomW4P1_F_sse4_1_single
700 (t_nblist * gmx_restrict nlist,
701 rvec * gmx_restrict xx,
702 rvec * gmx_restrict ff,
703 t_forcerec * gmx_restrict fr,
704 t_mdatoms * gmx_restrict mdatoms,
705 nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
706 t_nrnb * gmx_restrict nrnb)
708 /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
709 * just 0 for non-waters.
710 * Suffixes A,B,C,D refer to j loop unrolling done with SSE, e.g. for the four different
711 * jnr indices corresponding to data put in the four positions in the SIMD register.
713 int i_shift_offset,i_coord_offset,outeriter,inneriter;
714 int j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
715 int jnrA,jnrB,jnrC,jnrD;
716 int jnrlistA,jnrlistB,jnrlistC,jnrlistD;
717 int j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
718 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
720 real *shiftvec,*fshift,*x,*f;
721 real *fjptrA,*fjptrB,*fjptrC,*fjptrD;
723 __m128 tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
725 __m128 ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
727 __m128 ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1;
729 __m128 ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2;
731 __m128 ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3;
732 int vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D;
733 __m128 jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
734 __m128 dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
735 __m128 dx10,dy10,dz10,rsq10,rinv10,rinvsq10,r10,qq10,c6_10,c12_10;
736 __m128 dx20,dy20,dz20,rsq20,rinv20,rinvsq20,r20,qq20,c6_20,c12_20;
737 __m128 dx30,dy30,dz30,rsq30,rinv30,rinvsq30,r30,qq30,c6_30,c12_30;
738 __m128 velec,felec,velecsum,facel,crf,krf,krf2;
741 __m128 rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
744 __m128 one_sixth = _mm_set1_ps(1.0/6.0);
745 __m128 one_twelfth = _mm_set1_ps(1.0/12.0);
747 __m128i ifour = _mm_set1_epi32(4);
748 __m128 rt,vfeps,vftabscale,Y,F,G,H,Heps,Fp,VV,FF;
750 __m128 dummy_mask,cutoff_mask;
751 __m128 signbit = _mm_castsi128_ps( _mm_set1_epi32(0x80000000) );
752 __m128 one = _mm_set1_ps(1.0);
753 __m128 two = _mm_set1_ps(2.0);
759 jindex = nlist->jindex;
761 shiftidx = nlist->shift;
763 shiftvec = fr->shift_vec[0];
764 fshift = fr->fshift[0];
765 facel = _mm_set1_ps(fr->epsfac);
766 charge = mdatoms->chargeA;
767 krf = _mm_set1_ps(fr->ic->k_rf);
768 krf2 = _mm_set1_ps(fr->ic->k_rf*2.0);
769 crf = _mm_set1_ps(fr->ic->c_rf);
770 nvdwtype = fr->ntype;
772 vdwtype = mdatoms->typeA;
774 vftab = kernel_data->table_vdw->data;
775 vftabscale = _mm_set1_ps(kernel_data->table_vdw->scale);
777 /* Setup water-specific parameters */
778 inr = nlist->iinr[0];
779 iq1 = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+1]));
780 iq2 = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+2]));
781 iq3 = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+3]));
782 vdwioffset0 = 2*nvdwtype*vdwtype[inr+0];
784 /* Avoid stupid compiler warnings */
785 jnrA = jnrB = jnrC = jnrD = 0;
794 for(iidx=0;iidx<4*DIM;iidx++)
799 /* Start outer loop over neighborlists */
800 for(iidx=0; iidx<nri; iidx++)
802 /* Load shift vector for this list */
803 i_shift_offset = DIM*shiftidx[iidx];
805 /* Load limits for loop over neighbors */
806 j_index_start = jindex[iidx];
807 j_index_end = jindex[iidx+1];
809 /* Get outer coordinate index */
811 i_coord_offset = DIM*inr;
813 /* Load i particle coords and add shift vector */
814 gmx_mm_load_shift_and_4rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,
815 &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3);
817 fix0 = _mm_setzero_ps();
818 fiy0 = _mm_setzero_ps();
819 fiz0 = _mm_setzero_ps();
820 fix1 = _mm_setzero_ps();
821 fiy1 = _mm_setzero_ps();
822 fiz1 = _mm_setzero_ps();
823 fix2 = _mm_setzero_ps();
824 fiy2 = _mm_setzero_ps();
825 fiz2 = _mm_setzero_ps();
826 fix3 = _mm_setzero_ps();
827 fiy3 = _mm_setzero_ps();
828 fiz3 = _mm_setzero_ps();
830 /* Start inner kernel loop */
831 for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4)
834 /* Get j neighbor index, and coordinate index */
839 j_coord_offsetA = DIM*jnrA;
840 j_coord_offsetB = DIM*jnrB;
841 j_coord_offsetC = DIM*jnrC;
842 j_coord_offsetD = DIM*jnrD;
844 /* load j atom coordinates */
845 gmx_mm_load_1rvec_4ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
846 x+j_coord_offsetC,x+j_coord_offsetD,
849 /* Calculate displacement vector */
850 dx00 = _mm_sub_ps(ix0,jx0);
851 dy00 = _mm_sub_ps(iy0,jy0);
852 dz00 = _mm_sub_ps(iz0,jz0);
853 dx10 = _mm_sub_ps(ix1,jx0);
854 dy10 = _mm_sub_ps(iy1,jy0);
855 dz10 = _mm_sub_ps(iz1,jz0);
856 dx20 = _mm_sub_ps(ix2,jx0);
857 dy20 = _mm_sub_ps(iy2,jy0);
858 dz20 = _mm_sub_ps(iz2,jz0);
859 dx30 = _mm_sub_ps(ix3,jx0);
860 dy30 = _mm_sub_ps(iy3,jy0);
861 dz30 = _mm_sub_ps(iz3,jz0);
863 /* Calculate squared distance and things based on it */
864 rsq00 = gmx_mm_calc_rsq_ps(dx00,dy00,dz00);
865 rsq10 = gmx_mm_calc_rsq_ps(dx10,dy10,dz10);
866 rsq20 = gmx_mm_calc_rsq_ps(dx20,dy20,dz20);
867 rsq30 = gmx_mm_calc_rsq_ps(dx30,dy30,dz30);
869 rinv00 = gmx_mm_invsqrt_ps(rsq00);
870 rinv10 = gmx_mm_invsqrt_ps(rsq10);
871 rinv20 = gmx_mm_invsqrt_ps(rsq20);
872 rinv30 = gmx_mm_invsqrt_ps(rsq30);
874 rinvsq10 = _mm_mul_ps(rinv10,rinv10);
875 rinvsq20 = _mm_mul_ps(rinv20,rinv20);
876 rinvsq30 = _mm_mul_ps(rinv30,rinv30);
878 /* Load parameters for j particles */
879 jq0 = gmx_mm_load_4real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
880 charge+jnrC+0,charge+jnrD+0);
881 vdwjidx0A = 2*vdwtype[jnrA+0];
882 vdwjidx0B = 2*vdwtype[jnrB+0];
883 vdwjidx0C = 2*vdwtype[jnrC+0];
884 vdwjidx0D = 2*vdwtype[jnrD+0];
886 fjx0 = _mm_setzero_ps();
887 fjy0 = _mm_setzero_ps();
888 fjz0 = _mm_setzero_ps();
890 /**************************
891 * CALCULATE INTERACTIONS *
892 **************************/
894 r00 = _mm_mul_ps(rsq00,rinv00);
896 /* Compute parameters for interactions between i and j atoms */
897 gmx_mm_load_4pair_swizzle_ps(vdwparam+vdwioffset0+vdwjidx0A,
898 vdwparam+vdwioffset0+vdwjidx0B,
899 vdwparam+vdwioffset0+vdwjidx0C,
900 vdwparam+vdwioffset0+vdwjidx0D,
903 /* Calculate table index by multiplying r with table scale and truncate to integer */
904 rt = _mm_mul_ps(r00,vftabscale);
905 vfitab = _mm_cvttps_epi32(rt);
906 vfeps = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
907 vfitab = _mm_slli_epi32(vfitab,3);
909 /* CUBIC SPLINE TABLE DISPERSION */
910 Y = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,0) );
911 F = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,1) );
912 G = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,2) );
913 H = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,3) );
914 _MM_TRANSPOSE4_PS(Y,F,G,H);
915 Heps = _mm_mul_ps(vfeps,H);
916 Fp = _mm_add_ps(F,_mm_mul_ps(vfeps,_mm_add_ps(G,Heps)));
917 FF = _mm_add_ps(Fp,_mm_mul_ps(vfeps,_mm_add_ps(G,_mm_add_ps(Heps,Heps))));
918 fvdw6 = _mm_mul_ps(c6_00,FF);
920 /* CUBIC SPLINE TABLE REPULSION */
921 vfitab = _mm_add_epi32(vfitab,ifour);
922 Y = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,0) );
923 F = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,1) );
924 G = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,2) );
925 H = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,3) );
926 _MM_TRANSPOSE4_PS(Y,F,G,H);
927 Heps = _mm_mul_ps(vfeps,H);
928 Fp = _mm_add_ps(F,_mm_mul_ps(vfeps,_mm_add_ps(G,Heps)));
929 FF = _mm_add_ps(Fp,_mm_mul_ps(vfeps,_mm_add_ps(G,_mm_add_ps(Heps,Heps))));
930 fvdw12 = _mm_mul_ps(c12_00,FF);
931 fvdw = _mm_xor_ps(signbit,_mm_mul_ps(_mm_add_ps(fvdw6,fvdw12),_mm_mul_ps(vftabscale,rinv00)));
935 /* Calculate temporary vectorial force */
936 tx = _mm_mul_ps(fscal,dx00);
937 ty = _mm_mul_ps(fscal,dy00);
938 tz = _mm_mul_ps(fscal,dz00);
940 /* Update vectorial force */
941 fix0 = _mm_add_ps(fix0,tx);
942 fiy0 = _mm_add_ps(fiy0,ty);
943 fiz0 = _mm_add_ps(fiz0,tz);
945 fjx0 = _mm_add_ps(fjx0,tx);
946 fjy0 = _mm_add_ps(fjy0,ty);
947 fjz0 = _mm_add_ps(fjz0,tz);
949 /**************************
950 * CALCULATE INTERACTIONS *
951 **************************/
953 /* Compute parameters for interactions between i and j atoms */
954 qq10 = _mm_mul_ps(iq1,jq0);
956 /* REACTION-FIELD ELECTROSTATICS */
957 felec = _mm_mul_ps(qq10,_mm_sub_ps(_mm_mul_ps(rinv10,rinvsq10),krf2));
961 /* Calculate temporary vectorial force */
962 tx = _mm_mul_ps(fscal,dx10);
963 ty = _mm_mul_ps(fscal,dy10);
964 tz = _mm_mul_ps(fscal,dz10);
966 /* Update vectorial force */
967 fix1 = _mm_add_ps(fix1,tx);
968 fiy1 = _mm_add_ps(fiy1,ty);
969 fiz1 = _mm_add_ps(fiz1,tz);
971 fjx0 = _mm_add_ps(fjx0,tx);
972 fjy0 = _mm_add_ps(fjy0,ty);
973 fjz0 = _mm_add_ps(fjz0,tz);
975 /**************************
976 * CALCULATE INTERACTIONS *
977 **************************/
979 /* Compute parameters for interactions between i and j atoms */
980 qq20 = _mm_mul_ps(iq2,jq0);
982 /* REACTION-FIELD ELECTROSTATICS */
983 felec = _mm_mul_ps(qq20,_mm_sub_ps(_mm_mul_ps(rinv20,rinvsq20),krf2));
987 /* Calculate temporary vectorial force */
988 tx = _mm_mul_ps(fscal,dx20);
989 ty = _mm_mul_ps(fscal,dy20);
990 tz = _mm_mul_ps(fscal,dz20);
992 /* Update vectorial force */
993 fix2 = _mm_add_ps(fix2,tx);
994 fiy2 = _mm_add_ps(fiy2,ty);
995 fiz2 = _mm_add_ps(fiz2,tz);
997 fjx0 = _mm_add_ps(fjx0,tx);
998 fjy0 = _mm_add_ps(fjy0,ty);
999 fjz0 = _mm_add_ps(fjz0,tz);
1001 /**************************
1002 * CALCULATE INTERACTIONS *
1003 **************************/
1005 /* Compute parameters for interactions between i and j atoms */
1006 qq30 = _mm_mul_ps(iq3,jq0);
1008 /* REACTION-FIELD ELECTROSTATICS */
1009 felec = _mm_mul_ps(qq30,_mm_sub_ps(_mm_mul_ps(rinv30,rinvsq30),krf2));
1013 /* Calculate temporary vectorial force */
1014 tx = _mm_mul_ps(fscal,dx30);
1015 ty = _mm_mul_ps(fscal,dy30);
1016 tz = _mm_mul_ps(fscal,dz30);
1018 /* Update vectorial force */
1019 fix3 = _mm_add_ps(fix3,tx);
1020 fiy3 = _mm_add_ps(fiy3,ty);
1021 fiz3 = _mm_add_ps(fiz3,tz);
1023 fjx0 = _mm_add_ps(fjx0,tx);
1024 fjy0 = _mm_add_ps(fjy0,ty);
1025 fjz0 = _mm_add_ps(fjz0,tz);
1027 fjptrA = f+j_coord_offsetA;
1028 fjptrB = f+j_coord_offsetB;
1029 fjptrC = f+j_coord_offsetC;
1030 fjptrD = f+j_coord_offsetD;
1032 gmx_mm_decrement_1rvec_4ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0);
1034 /* Inner loop uses 129 flops */
1037 if(jidx<j_index_end)
1040 /* Get j neighbor index, and coordinate index */
1041 jnrlistA = jjnr[jidx];
1042 jnrlistB = jjnr[jidx+1];
1043 jnrlistC = jjnr[jidx+2];
1044 jnrlistD = jjnr[jidx+3];
1045 /* Sign of each element will be negative for non-real atoms.
1046 * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
1047 * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
1049 dummy_mask = gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128()));
1050 jnrA = (jnrlistA>=0) ? jnrlistA : 0;
1051 jnrB = (jnrlistB>=0) ? jnrlistB : 0;
1052 jnrC = (jnrlistC>=0) ? jnrlistC : 0;
1053 jnrD = (jnrlistD>=0) ? jnrlistD : 0;
1054 j_coord_offsetA = DIM*jnrA;
1055 j_coord_offsetB = DIM*jnrB;
1056 j_coord_offsetC = DIM*jnrC;
1057 j_coord_offsetD = DIM*jnrD;
1059 /* load j atom coordinates */
1060 gmx_mm_load_1rvec_4ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
1061 x+j_coord_offsetC,x+j_coord_offsetD,
1064 /* Calculate displacement vector */
1065 dx00 = _mm_sub_ps(ix0,jx0);
1066 dy00 = _mm_sub_ps(iy0,jy0);
1067 dz00 = _mm_sub_ps(iz0,jz0);
1068 dx10 = _mm_sub_ps(ix1,jx0);
1069 dy10 = _mm_sub_ps(iy1,jy0);
1070 dz10 = _mm_sub_ps(iz1,jz0);
1071 dx20 = _mm_sub_ps(ix2,jx0);
1072 dy20 = _mm_sub_ps(iy2,jy0);
1073 dz20 = _mm_sub_ps(iz2,jz0);
1074 dx30 = _mm_sub_ps(ix3,jx0);
1075 dy30 = _mm_sub_ps(iy3,jy0);
1076 dz30 = _mm_sub_ps(iz3,jz0);
1078 /* Calculate squared distance and things based on it */
1079 rsq00 = gmx_mm_calc_rsq_ps(dx00,dy00,dz00);
1080 rsq10 = gmx_mm_calc_rsq_ps(dx10,dy10,dz10);
1081 rsq20 = gmx_mm_calc_rsq_ps(dx20,dy20,dz20);
1082 rsq30 = gmx_mm_calc_rsq_ps(dx30,dy30,dz30);
1084 rinv00 = gmx_mm_invsqrt_ps(rsq00);
1085 rinv10 = gmx_mm_invsqrt_ps(rsq10);
1086 rinv20 = gmx_mm_invsqrt_ps(rsq20);
1087 rinv30 = gmx_mm_invsqrt_ps(rsq30);
1089 rinvsq10 = _mm_mul_ps(rinv10,rinv10);
1090 rinvsq20 = _mm_mul_ps(rinv20,rinv20);
1091 rinvsq30 = _mm_mul_ps(rinv30,rinv30);
1093 /* Load parameters for j particles */
1094 jq0 = gmx_mm_load_4real_swizzle_ps(charge+jnrA+0,charge+jnrB+0,
1095 charge+jnrC+0,charge+jnrD+0);
1096 vdwjidx0A = 2*vdwtype[jnrA+0];
1097 vdwjidx0B = 2*vdwtype[jnrB+0];
1098 vdwjidx0C = 2*vdwtype[jnrC+0];
1099 vdwjidx0D = 2*vdwtype[jnrD+0];
1101 fjx0 = _mm_setzero_ps();
1102 fjy0 = _mm_setzero_ps();
1103 fjz0 = _mm_setzero_ps();
1105 /**************************
1106 * CALCULATE INTERACTIONS *
1107 **************************/
1109 r00 = _mm_mul_ps(rsq00,rinv00);
1110 r00 = _mm_andnot_ps(dummy_mask,r00);
1112 /* Compute parameters for interactions between i and j atoms */
1113 gmx_mm_load_4pair_swizzle_ps(vdwparam+vdwioffset0+vdwjidx0A,
1114 vdwparam+vdwioffset0+vdwjidx0B,
1115 vdwparam+vdwioffset0+vdwjidx0C,
1116 vdwparam+vdwioffset0+vdwjidx0D,
1119 /* Calculate table index by multiplying r with table scale and truncate to integer */
1120 rt = _mm_mul_ps(r00,vftabscale);
1121 vfitab = _mm_cvttps_epi32(rt);
1122 vfeps = _mm_sub_ps(rt,_mm_round_ps(rt, _MM_FROUND_FLOOR));
1123 vfitab = _mm_slli_epi32(vfitab,3);
1125 /* CUBIC SPLINE TABLE DISPERSION */
1126 Y = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,0) );
1127 F = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,1) );
1128 G = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,2) );
1129 H = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,3) );
1130 _MM_TRANSPOSE4_PS(Y,F,G,H);
1131 Heps = _mm_mul_ps(vfeps,H);
1132 Fp = _mm_add_ps(F,_mm_mul_ps(vfeps,_mm_add_ps(G,Heps)));
1133 FF = _mm_add_ps(Fp,_mm_mul_ps(vfeps,_mm_add_ps(G,_mm_add_ps(Heps,Heps))));
1134 fvdw6 = _mm_mul_ps(c6_00,FF);
1136 /* CUBIC SPLINE TABLE REPULSION */
1137 vfitab = _mm_add_epi32(vfitab,ifour);
1138 Y = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,0) );
1139 F = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,1) );
1140 G = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,2) );
1141 H = _mm_load_ps( vftab + gmx_mm_extract_epi32(vfitab,3) );
1142 _MM_TRANSPOSE4_PS(Y,F,G,H);
1143 Heps = _mm_mul_ps(vfeps,H);
1144 Fp = _mm_add_ps(F,_mm_mul_ps(vfeps,_mm_add_ps(G,Heps)));
1145 FF = _mm_add_ps(Fp,_mm_mul_ps(vfeps,_mm_add_ps(G,_mm_add_ps(Heps,Heps))));
1146 fvdw12 = _mm_mul_ps(c12_00,FF);
1147 fvdw = _mm_xor_ps(signbit,_mm_mul_ps(_mm_add_ps(fvdw6,fvdw12),_mm_mul_ps(vftabscale,rinv00)));
1151 fscal = _mm_andnot_ps(dummy_mask,fscal);
1153 /* Calculate temporary vectorial force */
1154 tx = _mm_mul_ps(fscal,dx00);
1155 ty = _mm_mul_ps(fscal,dy00);
1156 tz = _mm_mul_ps(fscal,dz00);
1158 /* Update vectorial force */
1159 fix0 = _mm_add_ps(fix0,tx);
1160 fiy0 = _mm_add_ps(fiy0,ty);
1161 fiz0 = _mm_add_ps(fiz0,tz);
1163 fjx0 = _mm_add_ps(fjx0,tx);
1164 fjy0 = _mm_add_ps(fjy0,ty);
1165 fjz0 = _mm_add_ps(fjz0,tz);
1167 /**************************
1168 * CALCULATE INTERACTIONS *
1169 **************************/
1171 /* Compute parameters for interactions between i and j atoms */
1172 qq10 = _mm_mul_ps(iq1,jq0);
1174 /* REACTION-FIELD ELECTROSTATICS */
1175 felec = _mm_mul_ps(qq10,_mm_sub_ps(_mm_mul_ps(rinv10,rinvsq10),krf2));
1179 fscal = _mm_andnot_ps(dummy_mask,fscal);
1181 /* Calculate temporary vectorial force */
1182 tx = _mm_mul_ps(fscal,dx10);
1183 ty = _mm_mul_ps(fscal,dy10);
1184 tz = _mm_mul_ps(fscal,dz10);
1186 /* Update vectorial force */
1187 fix1 = _mm_add_ps(fix1,tx);
1188 fiy1 = _mm_add_ps(fiy1,ty);
1189 fiz1 = _mm_add_ps(fiz1,tz);
1191 fjx0 = _mm_add_ps(fjx0,tx);
1192 fjy0 = _mm_add_ps(fjy0,ty);
1193 fjz0 = _mm_add_ps(fjz0,tz);
1195 /**************************
1196 * CALCULATE INTERACTIONS *
1197 **************************/
1199 /* Compute parameters for interactions between i and j atoms */
1200 qq20 = _mm_mul_ps(iq2,jq0);
1202 /* REACTION-FIELD ELECTROSTATICS */
1203 felec = _mm_mul_ps(qq20,_mm_sub_ps(_mm_mul_ps(rinv20,rinvsq20),krf2));
1207 fscal = _mm_andnot_ps(dummy_mask,fscal);
1209 /* Calculate temporary vectorial force */
1210 tx = _mm_mul_ps(fscal,dx20);
1211 ty = _mm_mul_ps(fscal,dy20);
1212 tz = _mm_mul_ps(fscal,dz20);
1214 /* Update vectorial force */
1215 fix2 = _mm_add_ps(fix2,tx);
1216 fiy2 = _mm_add_ps(fiy2,ty);
1217 fiz2 = _mm_add_ps(fiz2,tz);
1219 fjx0 = _mm_add_ps(fjx0,tx);
1220 fjy0 = _mm_add_ps(fjy0,ty);
1221 fjz0 = _mm_add_ps(fjz0,tz);
1223 /**************************
1224 * CALCULATE INTERACTIONS *
1225 **************************/
1227 /* Compute parameters for interactions between i and j atoms */
1228 qq30 = _mm_mul_ps(iq3,jq0);
1230 /* REACTION-FIELD ELECTROSTATICS */
1231 felec = _mm_mul_ps(qq30,_mm_sub_ps(_mm_mul_ps(rinv30,rinvsq30),krf2));
1235 fscal = _mm_andnot_ps(dummy_mask,fscal);
1237 /* Calculate temporary vectorial force */
1238 tx = _mm_mul_ps(fscal,dx30);
1239 ty = _mm_mul_ps(fscal,dy30);
1240 tz = _mm_mul_ps(fscal,dz30);
1242 /* Update vectorial force */
1243 fix3 = _mm_add_ps(fix3,tx);
1244 fiy3 = _mm_add_ps(fiy3,ty);
1245 fiz3 = _mm_add_ps(fiz3,tz);
1247 fjx0 = _mm_add_ps(fjx0,tx);
1248 fjy0 = _mm_add_ps(fjy0,ty);
1249 fjz0 = _mm_add_ps(fjz0,tz);
1251 fjptrA = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
1252 fjptrB = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
1253 fjptrC = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
1254 fjptrD = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
1256 gmx_mm_decrement_1rvec_4ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjx0,fjy0,fjz0);
1258 /* Inner loop uses 130 flops */
1261 /* End of innermost loop */
1263 gmx_mm_update_iforce_4atom_swizzle_ps(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3,
1264 f+i_coord_offset,fshift+i_shift_offset);
1266 /* Increment number of inner iterations */
1267 inneriter += j_index_end - j_index_start;
1269 /* Outer loop uses 24 flops */
1272 /* Increment number of outer iterations */
1275 /* Update outer/inner flops */
1277 inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4_F,outeriter*24 + inneriter*130);