2 * This file is part of the GROMACS molecular simulation package.
4 * Copyright (c) 2012,2013,2014, by the GROMACS development team, led by
5 * Mark Abraham, David van der Spoel, Berk Hess, and Erik Lindahl,
6 * and including many others, as listed in the AUTHORS file in the
7 * top-level source directory and at http://www.gromacs.org.
9 * GROMACS is free software; you can redistribute it and/or
10 * modify it under the terms of the GNU Lesser General Public License
11 * as published by the Free Software Foundation; either version 2.1
12 * of the License, or (at your option) any later version.
14 * GROMACS is distributed in the hope that it will be useful,
15 * but WITHOUT ANY WARRANTY; without even the implied warranty of
16 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
17 * Lesser General Public License for more details.
19 * You should have received a copy of the GNU Lesser General Public
20 * License along with GROMACS; if not, see
21 * http://www.gnu.org/licenses, or write to the Free Software Foundation,
22 * Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
24 * If you want to redistribute modifications to GROMACS, please
25 * consider that scientific software is very special. Version
26 * control is crucial - bugs must be traceable. We will be happy to
27 * consider code for inclusion in the official distribution, but
28 * derived work must not be called official GROMACS. Details are found
29 * in the README & COPYING files - if they are missing, get the
30 * official version at http://www.gromacs.org.
32 * To help us fund GROMACS development, we humbly ask that you cite
33 * the research papers on the package. Check out http://www.gromacs.org.
36 * Note: this file was generated by the GROMACS avx_256_single kernel generator.
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_avx_256_single.h"
48 #include "kernelutil_x86_avx_256_single.h"
51 * Gromacs nonbonded kernel: nb_kernel_ElecNone_VdwLJ_GeomP1P1_VF_avx_256_single
52 * Electrostatics interaction: None
53 * VdW interaction: LennardJones
54 * Geometry: Particle-Particle
55 * Calculate force/pot: PotentialAndForce
58 nb_kernel_ElecNone_VdwLJ_GeomP1P1_VF_avx_256_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,E,F,G,H refer to j loop unrolling done with AVX, e.g. for the eight 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 jnrE,jnrF,jnrG,jnrH;
76 int jnrlistA,jnrlistB,jnrlistC,jnrlistD;
77 int jnrlistE,jnrlistF,jnrlistG,jnrlistH;
78 int j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
79 int j_coord_offsetE,j_coord_offsetF,j_coord_offsetG,j_coord_offsetH;
80 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
82 real *shiftvec,*fshift,*x,*f;
83 real *fjptrA,*fjptrB,*fjptrC,*fjptrD,*fjptrE,*fjptrF,*fjptrG,*fjptrH;
85 __m256 tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
86 real * vdwioffsetptr0;
87 __m256 ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
88 int vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D,vdwjidx0E,vdwjidx0F,vdwjidx0G,vdwjidx0H;
89 __m256 jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
90 __m256 dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
92 __m256 rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
95 __m256 one_sixth = _mm256_set1_ps(1.0/6.0);
96 __m256 one_twelfth = _mm256_set1_ps(1.0/12.0);
97 __m256 dummy_mask,cutoff_mask;
98 __m256 signbit = _mm256_castsi256_ps( _mm256_set1_epi32(0x80000000) );
99 __m256 one = _mm256_set1_ps(1.0);
100 __m256 two = _mm256_set1_ps(2.0);
106 jindex = nlist->jindex;
108 shiftidx = nlist->shift;
110 shiftvec = fr->shift_vec[0];
111 fshift = fr->fshift[0];
112 nvdwtype = fr->ntype;
114 vdwtype = mdatoms->typeA;
116 /* Avoid stupid compiler warnings */
117 jnrA = jnrB = jnrC = jnrD = jnrE = jnrF = jnrG = jnrH = 0;
130 for(iidx=0;iidx<4*DIM;iidx++)
135 /* Start outer loop over neighborlists */
136 for(iidx=0; iidx<nri; iidx++)
138 /* Load shift vector for this list */
139 i_shift_offset = DIM*shiftidx[iidx];
141 /* Load limits for loop over neighbors */
142 j_index_start = jindex[iidx];
143 j_index_end = jindex[iidx+1];
145 /* Get outer coordinate index */
147 i_coord_offset = DIM*inr;
149 /* Load i particle coords and add shift vector */
150 gmx_mm256_load_shift_and_1rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,&ix0,&iy0,&iz0);
152 fix0 = _mm256_setzero_ps();
153 fiy0 = _mm256_setzero_ps();
154 fiz0 = _mm256_setzero_ps();
156 /* Load parameters for i particles */
157 vdwioffsetptr0 = vdwparam+2*nvdwtype*vdwtype[inr+0];
159 /* Reset potential sums */
160 vvdwsum = _mm256_setzero_ps();
162 /* Start inner kernel loop */
163 for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+7]>=0; jidx+=8)
166 /* Get j neighbor index, and coordinate index */
175 j_coord_offsetA = DIM*jnrA;
176 j_coord_offsetB = DIM*jnrB;
177 j_coord_offsetC = DIM*jnrC;
178 j_coord_offsetD = DIM*jnrD;
179 j_coord_offsetE = DIM*jnrE;
180 j_coord_offsetF = DIM*jnrF;
181 j_coord_offsetG = DIM*jnrG;
182 j_coord_offsetH = DIM*jnrH;
184 /* load j atom coordinates */
185 gmx_mm256_load_1rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
186 x+j_coord_offsetC,x+j_coord_offsetD,
187 x+j_coord_offsetE,x+j_coord_offsetF,
188 x+j_coord_offsetG,x+j_coord_offsetH,
191 /* Calculate displacement vector */
192 dx00 = _mm256_sub_ps(ix0,jx0);
193 dy00 = _mm256_sub_ps(iy0,jy0);
194 dz00 = _mm256_sub_ps(iz0,jz0);
196 /* Calculate squared distance and things based on it */
197 rsq00 = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
199 rinvsq00 = gmx_mm256_inv_ps(rsq00);
201 /* Load parameters for j particles */
202 vdwjidx0A = 2*vdwtype[jnrA+0];
203 vdwjidx0B = 2*vdwtype[jnrB+0];
204 vdwjidx0C = 2*vdwtype[jnrC+0];
205 vdwjidx0D = 2*vdwtype[jnrD+0];
206 vdwjidx0E = 2*vdwtype[jnrE+0];
207 vdwjidx0F = 2*vdwtype[jnrF+0];
208 vdwjidx0G = 2*vdwtype[jnrG+0];
209 vdwjidx0H = 2*vdwtype[jnrH+0];
211 /**************************
212 * CALCULATE INTERACTIONS *
213 **************************/
215 /* Compute parameters for interactions between i and j atoms */
216 gmx_mm256_load_8pair_swizzle_ps(vdwioffsetptr0+vdwjidx0A,
217 vdwioffsetptr0+vdwjidx0B,
218 vdwioffsetptr0+vdwjidx0C,
219 vdwioffsetptr0+vdwjidx0D,
220 vdwioffsetptr0+vdwjidx0E,
221 vdwioffsetptr0+vdwjidx0F,
222 vdwioffsetptr0+vdwjidx0G,
223 vdwioffsetptr0+vdwjidx0H,
226 /* LENNARD-JONES DISPERSION/REPULSION */
228 rinvsix = _mm256_mul_ps(_mm256_mul_ps(rinvsq00,rinvsq00),rinvsq00);
229 vvdw6 = _mm256_mul_ps(c6_00,rinvsix);
230 vvdw12 = _mm256_mul_ps(c12_00,_mm256_mul_ps(rinvsix,rinvsix));
231 vvdw = _mm256_sub_ps( _mm256_mul_ps(vvdw12,one_twelfth) , _mm256_mul_ps(vvdw6,one_sixth) );
232 fvdw = _mm256_mul_ps(_mm256_sub_ps(vvdw12,vvdw6),rinvsq00);
234 /* Update potential sum for this i atom from the interaction with this j atom. */
235 vvdwsum = _mm256_add_ps(vvdwsum,vvdw);
239 /* Calculate temporary vectorial force */
240 tx = _mm256_mul_ps(fscal,dx00);
241 ty = _mm256_mul_ps(fscal,dy00);
242 tz = _mm256_mul_ps(fscal,dz00);
244 /* Update vectorial force */
245 fix0 = _mm256_add_ps(fix0,tx);
246 fiy0 = _mm256_add_ps(fiy0,ty);
247 fiz0 = _mm256_add_ps(fiz0,tz);
249 fjptrA = f+j_coord_offsetA;
250 fjptrB = f+j_coord_offsetB;
251 fjptrC = f+j_coord_offsetC;
252 fjptrD = f+j_coord_offsetD;
253 fjptrE = f+j_coord_offsetE;
254 fjptrF = f+j_coord_offsetF;
255 fjptrG = f+j_coord_offsetG;
256 fjptrH = f+j_coord_offsetH;
257 gmx_mm256_decrement_1rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,tx,ty,tz);
259 /* Inner loop uses 32 flops */
265 /* Get j neighbor index, and coordinate index */
266 jnrlistA = jjnr[jidx];
267 jnrlistB = jjnr[jidx+1];
268 jnrlistC = jjnr[jidx+2];
269 jnrlistD = jjnr[jidx+3];
270 jnrlistE = jjnr[jidx+4];
271 jnrlistF = jjnr[jidx+5];
272 jnrlistG = jjnr[jidx+6];
273 jnrlistH = jjnr[jidx+7];
274 /* Sign of each element will be negative for non-real atoms.
275 * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
276 * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
278 dummy_mask = gmx_mm256_set_m128(gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx+4)),_mm_setzero_si128())),
279 gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128())));
281 jnrA = (jnrlistA>=0) ? jnrlistA : 0;
282 jnrB = (jnrlistB>=0) ? jnrlistB : 0;
283 jnrC = (jnrlistC>=0) ? jnrlistC : 0;
284 jnrD = (jnrlistD>=0) ? jnrlistD : 0;
285 jnrE = (jnrlistE>=0) ? jnrlistE : 0;
286 jnrF = (jnrlistF>=0) ? jnrlistF : 0;
287 jnrG = (jnrlistG>=0) ? jnrlistG : 0;
288 jnrH = (jnrlistH>=0) ? jnrlistH : 0;
289 j_coord_offsetA = DIM*jnrA;
290 j_coord_offsetB = DIM*jnrB;
291 j_coord_offsetC = DIM*jnrC;
292 j_coord_offsetD = DIM*jnrD;
293 j_coord_offsetE = DIM*jnrE;
294 j_coord_offsetF = DIM*jnrF;
295 j_coord_offsetG = DIM*jnrG;
296 j_coord_offsetH = DIM*jnrH;
298 /* load j atom coordinates */
299 gmx_mm256_load_1rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
300 x+j_coord_offsetC,x+j_coord_offsetD,
301 x+j_coord_offsetE,x+j_coord_offsetF,
302 x+j_coord_offsetG,x+j_coord_offsetH,
305 /* Calculate displacement vector */
306 dx00 = _mm256_sub_ps(ix0,jx0);
307 dy00 = _mm256_sub_ps(iy0,jy0);
308 dz00 = _mm256_sub_ps(iz0,jz0);
310 /* Calculate squared distance and things based on it */
311 rsq00 = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
313 rinvsq00 = gmx_mm256_inv_ps(rsq00);
315 /* Load parameters for j particles */
316 vdwjidx0A = 2*vdwtype[jnrA+0];
317 vdwjidx0B = 2*vdwtype[jnrB+0];
318 vdwjidx0C = 2*vdwtype[jnrC+0];
319 vdwjidx0D = 2*vdwtype[jnrD+0];
320 vdwjidx0E = 2*vdwtype[jnrE+0];
321 vdwjidx0F = 2*vdwtype[jnrF+0];
322 vdwjidx0G = 2*vdwtype[jnrG+0];
323 vdwjidx0H = 2*vdwtype[jnrH+0];
325 /**************************
326 * CALCULATE INTERACTIONS *
327 **************************/
329 /* Compute parameters for interactions between i and j atoms */
330 gmx_mm256_load_8pair_swizzle_ps(vdwioffsetptr0+vdwjidx0A,
331 vdwioffsetptr0+vdwjidx0B,
332 vdwioffsetptr0+vdwjidx0C,
333 vdwioffsetptr0+vdwjidx0D,
334 vdwioffsetptr0+vdwjidx0E,
335 vdwioffsetptr0+vdwjidx0F,
336 vdwioffsetptr0+vdwjidx0G,
337 vdwioffsetptr0+vdwjidx0H,
340 /* LENNARD-JONES DISPERSION/REPULSION */
342 rinvsix = _mm256_mul_ps(_mm256_mul_ps(rinvsq00,rinvsq00),rinvsq00);
343 vvdw6 = _mm256_mul_ps(c6_00,rinvsix);
344 vvdw12 = _mm256_mul_ps(c12_00,_mm256_mul_ps(rinvsix,rinvsix));
345 vvdw = _mm256_sub_ps( _mm256_mul_ps(vvdw12,one_twelfth) , _mm256_mul_ps(vvdw6,one_sixth) );
346 fvdw = _mm256_mul_ps(_mm256_sub_ps(vvdw12,vvdw6),rinvsq00);
348 /* Update potential sum for this i atom from the interaction with this j atom. */
349 vvdw = _mm256_andnot_ps(dummy_mask,vvdw);
350 vvdwsum = _mm256_add_ps(vvdwsum,vvdw);
354 fscal = _mm256_andnot_ps(dummy_mask,fscal);
356 /* Calculate temporary vectorial force */
357 tx = _mm256_mul_ps(fscal,dx00);
358 ty = _mm256_mul_ps(fscal,dy00);
359 tz = _mm256_mul_ps(fscal,dz00);
361 /* Update vectorial force */
362 fix0 = _mm256_add_ps(fix0,tx);
363 fiy0 = _mm256_add_ps(fiy0,ty);
364 fiz0 = _mm256_add_ps(fiz0,tz);
366 fjptrA = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
367 fjptrB = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
368 fjptrC = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
369 fjptrD = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
370 fjptrE = (jnrlistE>=0) ? f+j_coord_offsetE : scratch;
371 fjptrF = (jnrlistF>=0) ? f+j_coord_offsetF : scratch;
372 fjptrG = (jnrlistG>=0) ? f+j_coord_offsetG : scratch;
373 fjptrH = (jnrlistH>=0) ? f+j_coord_offsetH : scratch;
374 gmx_mm256_decrement_1rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,tx,ty,tz);
376 /* Inner loop uses 32 flops */
379 /* End of innermost loop */
381 gmx_mm256_update_iforce_1atom_swizzle_ps(fix0,fiy0,fiz0,
382 f+i_coord_offset,fshift+i_shift_offset);
385 /* Update potential energies */
386 gmx_mm256_update_1pot_ps(vvdwsum,kernel_data->energygrp_vdw+ggid);
388 /* Increment number of inner iterations */
389 inneriter += j_index_end - j_index_start;
391 /* Outer loop uses 7 flops */
394 /* Increment number of outer iterations */
397 /* Update outer/inner flops */
399 inc_nrnb(nrnb,eNR_NBKERNEL_VDW_VF,outeriter*7 + inneriter*32);
402 * Gromacs nonbonded kernel: nb_kernel_ElecNone_VdwLJ_GeomP1P1_F_avx_256_single
403 * Electrostatics interaction: None
404 * VdW interaction: LennardJones
405 * Geometry: Particle-Particle
406 * Calculate force/pot: Force
409 nb_kernel_ElecNone_VdwLJ_GeomP1P1_F_avx_256_single
410 (t_nblist * gmx_restrict nlist,
411 rvec * gmx_restrict xx,
412 rvec * gmx_restrict ff,
413 t_forcerec * gmx_restrict fr,
414 t_mdatoms * gmx_restrict mdatoms,
415 nb_kernel_data_t gmx_unused * gmx_restrict kernel_data,
416 t_nrnb * gmx_restrict nrnb)
418 /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or
419 * just 0 for non-waters.
420 * Suffixes A,B,C,D,E,F,G,H refer to j loop unrolling done with AVX, e.g. for the eight different
421 * jnr indices corresponding to data put in the four positions in the SIMD register.
423 int i_shift_offset,i_coord_offset,outeriter,inneriter;
424 int j_index_start,j_index_end,jidx,nri,inr,ggid,iidx;
425 int jnrA,jnrB,jnrC,jnrD;
426 int jnrE,jnrF,jnrG,jnrH;
427 int jnrlistA,jnrlistB,jnrlistC,jnrlistD;
428 int jnrlistE,jnrlistF,jnrlistG,jnrlistH;
429 int j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD;
430 int j_coord_offsetE,j_coord_offsetF,j_coord_offsetG,j_coord_offsetH;
431 int *iinr,*jindex,*jjnr,*shiftidx,*gid;
433 real *shiftvec,*fshift,*x,*f;
434 real *fjptrA,*fjptrB,*fjptrC,*fjptrD,*fjptrE,*fjptrF,*fjptrG,*fjptrH;
436 __m256 tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall;
437 real * vdwioffsetptr0;
438 __m256 ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0;
439 int vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D,vdwjidx0E,vdwjidx0F,vdwjidx0G,vdwjidx0H;
440 __m256 jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0;
441 __m256 dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00;
443 __m256 rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6;
446 __m256 one_sixth = _mm256_set1_ps(1.0/6.0);
447 __m256 one_twelfth = _mm256_set1_ps(1.0/12.0);
448 __m256 dummy_mask,cutoff_mask;
449 __m256 signbit = _mm256_castsi256_ps( _mm256_set1_epi32(0x80000000) );
450 __m256 one = _mm256_set1_ps(1.0);
451 __m256 two = _mm256_set1_ps(2.0);
457 jindex = nlist->jindex;
459 shiftidx = nlist->shift;
461 shiftvec = fr->shift_vec[0];
462 fshift = fr->fshift[0];
463 nvdwtype = fr->ntype;
465 vdwtype = mdatoms->typeA;
467 /* Avoid stupid compiler warnings */
468 jnrA = jnrB = jnrC = jnrD = jnrE = jnrF = jnrG = jnrH = 0;
481 for(iidx=0;iidx<4*DIM;iidx++)
486 /* Start outer loop over neighborlists */
487 for(iidx=0; iidx<nri; iidx++)
489 /* Load shift vector for this list */
490 i_shift_offset = DIM*shiftidx[iidx];
492 /* Load limits for loop over neighbors */
493 j_index_start = jindex[iidx];
494 j_index_end = jindex[iidx+1];
496 /* Get outer coordinate index */
498 i_coord_offset = DIM*inr;
500 /* Load i particle coords and add shift vector */
501 gmx_mm256_load_shift_and_1rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset,&ix0,&iy0,&iz0);
503 fix0 = _mm256_setzero_ps();
504 fiy0 = _mm256_setzero_ps();
505 fiz0 = _mm256_setzero_ps();
507 /* Load parameters for i particles */
508 vdwioffsetptr0 = vdwparam+2*nvdwtype*vdwtype[inr+0];
510 /* Start inner kernel loop */
511 for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+7]>=0; jidx+=8)
514 /* Get j neighbor index, and coordinate index */
523 j_coord_offsetA = DIM*jnrA;
524 j_coord_offsetB = DIM*jnrB;
525 j_coord_offsetC = DIM*jnrC;
526 j_coord_offsetD = DIM*jnrD;
527 j_coord_offsetE = DIM*jnrE;
528 j_coord_offsetF = DIM*jnrF;
529 j_coord_offsetG = DIM*jnrG;
530 j_coord_offsetH = DIM*jnrH;
532 /* load j atom coordinates */
533 gmx_mm256_load_1rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
534 x+j_coord_offsetC,x+j_coord_offsetD,
535 x+j_coord_offsetE,x+j_coord_offsetF,
536 x+j_coord_offsetG,x+j_coord_offsetH,
539 /* Calculate displacement vector */
540 dx00 = _mm256_sub_ps(ix0,jx0);
541 dy00 = _mm256_sub_ps(iy0,jy0);
542 dz00 = _mm256_sub_ps(iz0,jz0);
544 /* Calculate squared distance and things based on it */
545 rsq00 = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
547 rinvsq00 = gmx_mm256_inv_ps(rsq00);
549 /* Load parameters for j particles */
550 vdwjidx0A = 2*vdwtype[jnrA+0];
551 vdwjidx0B = 2*vdwtype[jnrB+0];
552 vdwjidx0C = 2*vdwtype[jnrC+0];
553 vdwjidx0D = 2*vdwtype[jnrD+0];
554 vdwjidx0E = 2*vdwtype[jnrE+0];
555 vdwjidx0F = 2*vdwtype[jnrF+0];
556 vdwjidx0G = 2*vdwtype[jnrG+0];
557 vdwjidx0H = 2*vdwtype[jnrH+0];
559 /**************************
560 * CALCULATE INTERACTIONS *
561 **************************/
563 /* Compute parameters for interactions between i and j atoms */
564 gmx_mm256_load_8pair_swizzle_ps(vdwioffsetptr0+vdwjidx0A,
565 vdwioffsetptr0+vdwjidx0B,
566 vdwioffsetptr0+vdwjidx0C,
567 vdwioffsetptr0+vdwjidx0D,
568 vdwioffsetptr0+vdwjidx0E,
569 vdwioffsetptr0+vdwjidx0F,
570 vdwioffsetptr0+vdwjidx0G,
571 vdwioffsetptr0+vdwjidx0H,
574 /* LENNARD-JONES DISPERSION/REPULSION */
576 rinvsix = _mm256_mul_ps(_mm256_mul_ps(rinvsq00,rinvsq00),rinvsq00);
577 fvdw = _mm256_mul_ps(_mm256_sub_ps(_mm256_mul_ps(c12_00,rinvsix),c6_00),_mm256_mul_ps(rinvsix,rinvsq00));
581 /* Calculate temporary vectorial force */
582 tx = _mm256_mul_ps(fscal,dx00);
583 ty = _mm256_mul_ps(fscal,dy00);
584 tz = _mm256_mul_ps(fscal,dz00);
586 /* Update vectorial force */
587 fix0 = _mm256_add_ps(fix0,tx);
588 fiy0 = _mm256_add_ps(fiy0,ty);
589 fiz0 = _mm256_add_ps(fiz0,tz);
591 fjptrA = f+j_coord_offsetA;
592 fjptrB = f+j_coord_offsetB;
593 fjptrC = f+j_coord_offsetC;
594 fjptrD = f+j_coord_offsetD;
595 fjptrE = f+j_coord_offsetE;
596 fjptrF = f+j_coord_offsetF;
597 fjptrG = f+j_coord_offsetG;
598 fjptrH = f+j_coord_offsetH;
599 gmx_mm256_decrement_1rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,tx,ty,tz);
601 /* Inner loop uses 27 flops */
607 /* Get j neighbor index, and coordinate index */
608 jnrlistA = jjnr[jidx];
609 jnrlistB = jjnr[jidx+1];
610 jnrlistC = jjnr[jidx+2];
611 jnrlistD = jjnr[jidx+3];
612 jnrlistE = jjnr[jidx+4];
613 jnrlistF = jjnr[jidx+5];
614 jnrlistG = jjnr[jidx+6];
615 jnrlistH = jjnr[jidx+7];
616 /* Sign of each element will be negative for non-real atoms.
617 * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones,
618 * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries.
620 dummy_mask = gmx_mm256_set_m128(gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx+4)),_mm_setzero_si128())),
621 gmx_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128())));
623 jnrA = (jnrlistA>=0) ? jnrlistA : 0;
624 jnrB = (jnrlistB>=0) ? jnrlistB : 0;
625 jnrC = (jnrlistC>=0) ? jnrlistC : 0;
626 jnrD = (jnrlistD>=0) ? jnrlistD : 0;
627 jnrE = (jnrlistE>=0) ? jnrlistE : 0;
628 jnrF = (jnrlistF>=0) ? jnrlistF : 0;
629 jnrG = (jnrlistG>=0) ? jnrlistG : 0;
630 jnrH = (jnrlistH>=0) ? jnrlistH : 0;
631 j_coord_offsetA = DIM*jnrA;
632 j_coord_offsetB = DIM*jnrB;
633 j_coord_offsetC = DIM*jnrC;
634 j_coord_offsetD = DIM*jnrD;
635 j_coord_offsetE = DIM*jnrE;
636 j_coord_offsetF = DIM*jnrF;
637 j_coord_offsetG = DIM*jnrG;
638 j_coord_offsetH = DIM*jnrH;
640 /* load j atom coordinates */
641 gmx_mm256_load_1rvec_8ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB,
642 x+j_coord_offsetC,x+j_coord_offsetD,
643 x+j_coord_offsetE,x+j_coord_offsetF,
644 x+j_coord_offsetG,x+j_coord_offsetH,
647 /* Calculate displacement vector */
648 dx00 = _mm256_sub_ps(ix0,jx0);
649 dy00 = _mm256_sub_ps(iy0,jy0);
650 dz00 = _mm256_sub_ps(iz0,jz0);
652 /* Calculate squared distance and things based on it */
653 rsq00 = gmx_mm256_calc_rsq_ps(dx00,dy00,dz00);
655 rinvsq00 = gmx_mm256_inv_ps(rsq00);
657 /* Load parameters for j particles */
658 vdwjidx0A = 2*vdwtype[jnrA+0];
659 vdwjidx0B = 2*vdwtype[jnrB+0];
660 vdwjidx0C = 2*vdwtype[jnrC+0];
661 vdwjidx0D = 2*vdwtype[jnrD+0];
662 vdwjidx0E = 2*vdwtype[jnrE+0];
663 vdwjidx0F = 2*vdwtype[jnrF+0];
664 vdwjidx0G = 2*vdwtype[jnrG+0];
665 vdwjidx0H = 2*vdwtype[jnrH+0];
667 /**************************
668 * CALCULATE INTERACTIONS *
669 **************************/
671 /* Compute parameters for interactions between i and j atoms */
672 gmx_mm256_load_8pair_swizzle_ps(vdwioffsetptr0+vdwjidx0A,
673 vdwioffsetptr0+vdwjidx0B,
674 vdwioffsetptr0+vdwjidx0C,
675 vdwioffsetptr0+vdwjidx0D,
676 vdwioffsetptr0+vdwjidx0E,
677 vdwioffsetptr0+vdwjidx0F,
678 vdwioffsetptr0+vdwjidx0G,
679 vdwioffsetptr0+vdwjidx0H,
682 /* LENNARD-JONES DISPERSION/REPULSION */
684 rinvsix = _mm256_mul_ps(_mm256_mul_ps(rinvsq00,rinvsq00),rinvsq00);
685 fvdw = _mm256_mul_ps(_mm256_sub_ps(_mm256_mul_ps(c12_00,rinvsix),c6_00),_mm256_mul_ps(rinvsix,rinvsq00));
689 fscal = _mm256_andnot_ps(dummy_mask,fscal);
691 /* Calculate temporary vectorial force */
692 tx = _mm256_mul_ps(fscal,dx00);
693 ty = _mm256_mul_ps(fscal,dy00);
694 tz = _mm256_mul_ps(fscal,dz00);
696 /* Update vectorial force */
697 fix0 = _mm256_add_ps(fix0,tx);
698 fiy0 = _mm256_add_ps(fiy0,ty);
699 fiz0 = _mm256_add_ps(fiz0,tz);
701 fjptrA = (jnrlistA>=0) ? f+j_coord_offsetA : scratch;
702 fjptrB = (jnrlistB>=0) ? f+j_coord_offsetB : scratch;
703 fjptrC = (jnrlistC>=0) ? f+j_coord_offsetC : scratch;
704 fjptrD = (jnrlistD>=0) ? f+j_coord_offsetD : scratch;
705 fjptrE = (jnrlistE>=0) ? f+j_coord_offsetE : scratch;
706 fjptrF = (jnrlistF>=0) ? f+j_coord_offsetF : scratch;
707 fjptrG = (jnrlistG>=0) ? f+j_coord_offsetG : scratch;
708 fjptrH = (jnrlistH>=0) ? f+j_coord_offsetH : scratch;
709 gmx_mm256_decrement_1rvec_8ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD,fjptrE,fjptrF,fjptrG,fjptrH,tx,ty,tz);
711 /* Inner loop uses 27 flops */
714 /* End of innermost loop */
716 gmx_mm256_update_iforce_1atom_swizzle_ps(fix0,fiy0,fiz0,
717 f+i_coord_offset,fshift+i_shift_offset);
719 /* Increment number of inner iterations */
720 inneriter += j_index_end - j_index_start;
722 /* Outer loop uses 6 flops */
725 /* Increment number of outer iterations */
728 /* Update outer/inner flops */
730 inc_nrnb(nrnb,eNR_NBKERNEL_VDW_F,outeriter*6 + inneriter*27);