File: | gromacs/gmxlib/nonbonded/nb_kernel_sse4_1_single/nb_kernel_ElecEw_VdwLJEw_GeomW4W4_sse4_1_single.c |
Location: | line 194, column 5 |
Description: | Value stored to 'j_coord_offsetC' is never read |
1 | /* |
2 | * This file is part of the GROMACS molecular simulation package. |
3 | * |
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. |
8 | * |
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. |
13 | * |
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. |
18 | * |
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. |
23 | * |
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. |
31 | * |
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. |
34 | */ |
35 | /* |
36 | * Note: this file was generated by the GROMACS sse4_1_single kernel generator. |
37 | */ |
38 | #ifdef HAVE_CONFIG_H1 |
39 | #include <config.h> |
40 | #endif |
41 | |
42 | #include <math.h> |
43 | |
44 | #include "../nb_kernel.h" |
45 | #include "types/simple.h" |
46 | #include "gromacs/math/vec.h" |
47 | #include "nrnb.h" |
48 | |
49 | #include "gromacs/simd/math_x86_sse4_1_single.h" |
50 | #include "kernelutil_x86_sse4_1_single.h" |
51 | |
52 | /* |
53 | * Gromacs nonbonded kernel: nb_kernel_ElecEw_VdwLJEw_GeomW4W4_VF_sse4_1_single |
54 | * Electrostatics interaction: Ewald |
55 | * VdW interaction: LJEwald |
56 | * Geometry: Water4-Water4 |
57 | * Calculate force/pot: PotentialAndForce |
58 | */ |
59 | void |
60 | nb_kernel_ElecEw_VdwLJEw_GeomW4W4_VF_sse4_1_single |
61 | (t_nblist * gmx_restrict nlist, |
62 | rvec * gmx_restrict xx, |
63 | rvec * gmx_restrict ff, |
64 | t_forcerec * gmx_restrict fr, |
65 | t_mdatoms * gmx_restrict mdatoms, |
66 | nb_kernel_data_t gmx_unused__attribute__ ((unused)) * gmx_restrict kernel_data, |
67 | t_nrnb * gmx_restrict nrnb) |
68 | { |
69 | /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or |
70 | * just 0 for non-waters. |
71 | * Suffixes A,B,C,D refer to j loop unrolling done with SSE, e.g. for the four different |
72 | * jnr indices corresponding to data put in the four positions in the SIMD register. |
73 | */ |
74 | int i_shift_offset,i_coord_offset,outeriter,inneriter; |
75 | int j_index_start,j_index_end,jidx,nri,inr,ggid,iidx; |
76 | int jnrA,jnrB,jnrC,jnrD; |
77 | int jnrlistA,jnrlistB,jnrlistC,jnrlistD; |
78 | int j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD; |
79 | int *iinr,*jindex,*jjnr,*shiftidx,*gid; |
80 | real rcutoff_scalar; |
81 | real *shiftvec,*fshift,*x,*f; |
82 | real *fjptrA,*fjptrB,*fjptrC,*fjptrD; |
83 | real scratch[4*DIM3]; |
84 | __m128 tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall; |
85 | int vdwioffset0; |
86 | __m128 ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0; |
87 | int vdwioffset1; |
88 | __m128 ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1; |
89 | int vdwioffset2; |
90 | __m128 ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2; |
91 | int vdwioffset3; |
92 | __m128 ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3; |
93 | int vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D; |
94 | __m128 jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0; |
95 | int vdwjidx1A,vdwjidx1B,vdwjidx1C,vdwjidx1D; |
96 | __m128 jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1; |
97 | int vdwjidx2A,vdwjidx2B,vdwjidx2C,vdwjidx2D; |
98 | __m128 jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2; |
99 | int vdwjidx3A,vdwjidx3B,vdwjidx3C,vdwjidx3D; |
100 | __m128 jx3,jy3,jz3,fjx3,fjy3,fjz3,jq3,isaj3; |
101 | __m128 dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00; |
102 | __m128 dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11; |
103 | __m128 dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12; |
104 | __m128 dx13,dy13,dz13,rsq13,rinv13,rinvsq13,r13,qq13,c6_13,c12_13; |
105 | __m128 dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21; |
106 | __m128 dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22; |
107 | __m128 dx23,dy23,dz23,rsq23,rinv23,rinvsq23,r23,qq23,c6_23,c12_23; |
108 | __m128 dx31,dy31,dz31,rsq31,rinv31,rinvsq31,r31,qq31,c6_31,c12_31; |
109 | __m128 dx32,dy32,dz32,rsq32,rinv32,rinvsq32,r32,qq32,c6_32,c12_32; |
110 | __m128 dx33,dy33,dz33,rsq33,rinv33,rinvsq33,r33,qq33,c6_33,c12_33; |
111 | __m128 velec,felec,velecsum,facel,crf,krf,krf2; |
112 | real *charge; |
113 | int nvdwtype; |
114 | __m128 rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6; |
115 | int *vdwtype; |
116 | real *vdwparam; |
117 | __m128 one_sixth = _mm_set1_ps(1.0/6.0); |
118 | __m128 one_twelfth = _mm_set1_ps(1.0/12.0); |
119 | __m128 c6grid_00; |
120 | __m128 c6grid_11; |
121 | __m128 c6grid_12; |
122 | __m128 c6grid_13; |
123 | __m128 c6grid_21; |
124 | __m128 c6grid_22; |
125 | __m128 c6grid_23; |
126 | __m128 c6grid_31; |
127 | __m128 c6grid_32; |
128 | __m128 c6grid_33; |
129 | __m128 ewclj,ewclj2,ewclj6,ewcljrsq,poly,exponent,f6A,f6B,sh_lj_ewald; |
130 | real *vdwgridparam; |
131 | __m128 one_half = _mm_set1_ps(0.5); |
132 | __m128 minus_one = _mm_set1_ps(-1.0); |
133 | __m128i ewitab; |
134 | __m128 ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV; |
135 | real *ewtab; |
136 | __m128 dummy_mask,cutoff_mask; |
137 | __m128 signbit = _mm_castsi128_ps( _mm_set1_epi32(0x80000000) ); |
138 | __m128 one = _mm_set1_ps(1.0); |
139 | __m128 two = _mm_set1_ps(2.0); |
140 | x = xx[0]; |
141 | f = ff[0]; |
142 | |
143 | nri = nlist->nri; |
144 | iinr = nlist->iinr; |
145 | jindex = nlist->jindex; |
146 | jjnr = nlist->jjnr; |
147 | shiftidx = nlist->shift; |
148 | gid = nlist->gid; |
149 | shiftvec = fr->shift_vec[0]; |
150 | fshift = fr->fshift[0]; |
151 | facel = _mm_set1_ps(fr->epsfac); |
152 | charge = mdatoms->chargeA; |
153 | nvdwtype = fr->ntype; |
154 | vdwparam = fr->nbfp; |
155 | vdwtype = mdatoms->typeA; |
156 | vdwgridparam = fr->ljpme_c6grid; |
157 | sh_lj_ewald = _mm_set1_ps(fr->ic->sh_lj_ewald); |
158 | ewclj = _mm_set1_ps(fr->ewaldcoeff_lj); |
159 | ewclj2 = _mm_mul_ps(minus_one,_mm_mul_ps(ewclj,ewclj)); |
160 | |
161 | sh_ewald = _mm_set1_ps(fr->ic->sh_ewald); |
162 | ewtab = fr->ic->tabq_coul_FDV0; |
163 | ewtabscale = _mm_set1_ps(fr->ic->tabq_scale); |
164 | ewtabhalfspace = _mm_set1_ps(0.5/fr->ic->tabq_scale); |
165 | |
166 | /* Setup water-specific parameters */ |
167 | inr = nlist->iinr[0]; |
168 | iq1 = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+1])); |
169 | iq2 = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+2])); |
170 | iq3 = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+3])); |
171 | vdwioffset0 = 2*nvdwtype*vdwtype[inr+0]; |
172 | |
173 | jq1 = _mm_set1_ps(charge[inr+1]); |
174 | jq2 = _mm_set1_ps(charge[inr+2]); |
175 | jq3 = _mm_set1_ps(charge[inr+3]); |
176 | vdwjidx0A = 2*vdwtype[inr+0]; |
177 | c6_00 = _mm_set1_ps(vdwparam[vdwioffset0+vdwjidx0A]); |
178 | c12_00 = _mm_set1_ps(vdwparam[vdwioffset0+vdwjidx0A+1]); |
179 | c6grid_00 = _mm_set1_ps(vdwgridparam[vdwioffset0+vdwjidx0A]); |
180 | qq11 = _mm_mul_ps(iq1,jq1); |
181 | qq12 = _mm_mul_ps(iq1,jq2); |
182 | qq13 = _mm_mul_ps(iq1,jq3); |
183 | qq21 = _mm_mul_ps(iq2,jq1); |
184 | qq22 = _mm_mul_ps(iq2,jq2); |
185 | qq23 = _mm_mul_ps(iq2,jq3); |
186 | qq31 = _mm_mul_ps(iq3,jq1); |
187 | qq32 = _mm_mul_ps(iq3,jq2); |
188 | qq33 = _mm_mul_ps(iq3,jq3); |
189 | |
190 | /* Avoid stupid compiler warnings */ |
191 | jnrA = jnrB = jnrC = jnrD = 0; |
192 | j_coord_offsetA = 0; |
193 | j_coord_offsetB = 0; |
194 | j_coord_offsetC = 0; |
Value stored to 'j_coord_offsetC' is never read | |
195 | j_coord_offsetD = 0; |
196 | |
197 | outeriter = 0; |
198 | inneriter = 0; |
199 | |
200 | for(iidx=0;iidx<4*DIM3;iidx++) |
201 | { |
202 | scratch[iidx] = 0.0; |
203 | } |
204 | |
205 | /* Start outer loop over neighborlists */ |
206 | for(iidx=0; iidx<nri; iidx++) |
207 | { |
208 | /* Load shift vector for this list */ |
209 | i_shift_offset = DIM3*shiftidx[iidx]; |
210 | |
211 | /* Load limits for loop over neighbors */ |
212 | j_index_start = jindex[iidx]; |
213 | j_index_end = jindex[iidx+1]; |
214 | |
215 | /* Get outer coordinate index */ |
216 | inr = iinr[iidx]; |
217 | i_coord_offset = DIM3*inr; |
218 | |
219 | /* Load i particle coords and add shift vector */ |
220 | gmx_mm_load_shift_and_4rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset, |
221 | &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3); |
222 | |
223 | fix0 = _mm_setzero_ps(); |
224 | fiy0 = _mm_setzero_ps(); |
225 | fiz0 = _mm_setzero_ps(); |
226 | fix1 = _mm_setzero_ps(); |
227 | fiy1 = _mm_setzero_ps(); |
228 | fiz1 = _mm_setzero_ps(); |
229 | fix2 = _mm_setzero_ps(); |
230 | fiy2 = _mm_setzero_ps(); |
231 | fiz2 = _mm_setzero_ps(); |
232 | fix3 = _mm_setzero_ps(); |
233 | fiy3 = _mm_setzero_ps(); |
234 | fiz3 = _mm_setzero_ps(); |
235 | |
236 | /* Reset potential sums */ |
237 | velecsum = _mm_setzero_ps(); |
238 | vvdwsum = _mm_setzero_ps(); |
239 | |
240 | /* Start inner kernel loop */ |
241 | for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4) |
242 | { |
243 | |
244 | /* Get j neighbor index, and coordinate index */ |
245 | jnrA = jjnr[jidx]; |
246 | jnrB = jjnr[jidx+1]; |
247 | jnrC = jjnr[jidx+2]; |
248 | jnrD = jjnr[jidx+3]; |
249 | j_coord_offsetA = DIM3*jnrA; |
250 | j_coord_offsetB = DIM3*jnrB; |
251 | j_coord_offsetC = DIM3*jnrC; |
252 | j_coord_offsetD = DIM3*jnrD; |
253 | |
254 | /* load j atom coordinates */ |
255 | gmx_mm_load_4rvec_4ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB, |
256 | x+j_coord_offsetC,x+j_coord_offsetD, |
257 | &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2, |
258 | &jy2,&jz2,&jx3,&jy3,&jz3); |
259 | |
260 | /* Calculate displacement vector */ |
261 | dx00 = _mm_sub_ps(ix0,jx0); |
262 | dy00 = _mm_sub_ps(iy0,jy0); |
263 | dz00 = _mm_sub_ps(iz0,jz0); |
264 | dx11 = _mm_sub_ps(ix1,jx1); |
265 | dy11 = _mm_sub_ps(iy1,jy1); |
266 | dz11 = _mm_sub_ps(iz1,jz1); |
267 | dx12 = _mm_sub_ps(ix1,jx2); |
268 | dy12 = _mm_sub_ps(iy1,jy2); |
269 | dz12 = _mm_sub_ps(iz1,jz2); |
270 | dx13 = _mm_sub_ps(ix1,jx3); |
271 | dy13 = _mm_sub_ps(iy1,jy3); |
272 | dz13 = _mm_sub_ps(iz1,jz3); |
273 | dx21 = _mm_sub_ps(ix2,jx1); |
274 | dy21 = _mm_sub_ps(iy2,jy1); |
275 | dz21 = _mm_sub_ps(iz2,jz1); |
276 | dx22 = _mm_sub_ps(ix2,jx2); |
277 | dy22 = _mm_sub_ps(iy2,jy2); |
278 | dz22 = _mm_sub_ps(iz2,jz2); |
279 | dx23 = _mm_sub_ps(ix2,jx3); |
280 | dy23 = _mm_sub_ps(iy2,jy3); |
281 | dz23 = _mm_sub_ps(iz2,jz3); |
282 | dx31 = _mm_sub_ps(ix3,jx1); |
283 | dy31 = _mm_sub_ps(iy3,jy1); |
284 | dz31 = _mm_sub_ps(iz3,jz1); |
285 | dx32 = _mm_sub_ps(ix3,jx2); |
286 | dy32 = _mm_sub_ps(iy3,jy2); |
287 | dz32 = _mm_sub_ps(iz3,jz2); |
288 | dx33 = _mm_sub_ps(ix3,jx3); |
289 | dy33 = _mm_sub_ps(iy3,jy3); |
290 | dz33 = _mm_sub_ps(iz3,jz3); |
291 | |
292 | /* Calculate squared distance and things based on it */ |
293 | rsq00 = gmx_mm_calc_rsq_ps(dx00,dy00,dz00); |
294 | rsq11 = gmx_mm_calc_rsq_ps(dx11,dy11,dz11); |
295 | rsq12 = gmx_mm_calc_rsq_ps(dx12,dy12,dz12); |
296 | rsq13 = gmx_mm_calc_rsq_ps(dx13,dy13,dz13); |
297 | rsq21 = gmx_mm_calc_rsq_ps(dx21,dy21,dz21); |
298 | rsq22 = gmx_mm_calc_rsq_ps(dx22,dy22,dz22); |
299 | rsq23 = gmx_mm_calc_rsq_ps(dx23,dy23,dz23); |
300 | rsq31 = gmx_mm_calc_rsq_ps(dx31,dy31,dz31); |
301 | rsq32 = gmx_mm_calc_rsq_ps(dx32,dy32,dz32); |
302 | rsq33 = gmx_mm_calc_rsq_ps(dx33,dy33,dz33); |
303 | |
304 | rinv00 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq00); |
305 | rinv11 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq11); |
306 | rinv12 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq12); |
307 | rinv13 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq13); |
308 | rinv21 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq21); |
309 | rinv22 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq22); |
310 | rinv23 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq23); |
311 | rinv31 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq31); |
312 | rinv32 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq32); |
313 | rinv33 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq33); |
314 | |
315 | rinvsq00 = _mm_mul_ps(rinv00,rinv00); |
316 | rinvsq11 = _mm_mul_ps(rinv11,rinv11); |
317 | rinvsq12 = _mm_mul_ps(rinv12,rinv12); |
318 | rinvsq13 = _mm_mul_ps(rinv13,rinv13); |
319 | rinvsq21 = _mm_mul_ps(rinv21,rinv21); |
320 | rinvsq22 = _mm_mul_ps(rinv22,rinv22); |
321 | rinvsq23 = _mm_mul_ps(rinv23,rinv23); |
322 | rinvsq31 = _mm_mul_ps(rinv31,rinv31); |
323 | rinvsq32 = _mm_mul_ps(rinv32,rinv32); |
324 | rinvsq33 = _mm_mul_ps(rinv33,rinv33); |
325 | |
326 | fjx0 = _mm_setzero_ps(); |
327 | fjy0 = _mm_setzero_ps(); |
328 | fjz0 = _mm_setzero_ps(); |
329 | fjx1 = _mm_setzero_ps(); |
330 | fjy1 = _mm_setzero_ps(); |
331 | fjz1 = _mm_setzero_ps(); |
332 | fjx2 = _mm_setzero_ps(); |
333 | fjy2 = _mm_setzero_ps(); |
334 | fjz2 = _mm_setzero_ps(); |
335 | fjx3 = _mm_setzero_ps(); |
336 | fjy3 = _mm_setzero_ps(); |
337 | fjz3 = _mm_setzero_ps(); |
338 | |
339 | /************************** |
340 | * CALCULATE INTERACTIONS * |
341 | **************************/ |
342 | |
343 | r00 = _mm_mul_ps(rsq00,rinv00); |
344 | |
345 | /* Analytical LJ-PME */ |
346 | rinvsix = _mm_mul_ps(_mm_mul_ps(rinvsq00,rinvsq00),rinvsq00); |
347 | ewcljrsq = _mm_mul_ps(ewclj2,rsq00); |
348 | ewclj6 = _mm_mul_ps(ewclj2,_mm_mul_ps(ewclj2,ewclj2)); |
349 | exponent = gmx_simd_exp_rgmx_simd_exp_f(ewcljrsq); |
350 | /* poly = exp(-(beta*r)^2) * (1 + (beta*r)^2 + (beta*r)^4 /2) */ |
351 | poly = _mm_mul_ps(exponent,_mm_add_ps(_mm_sub_ps(one,ewcljrsq),_mm_mul_ps(_mm_mul_ps(ewcljrsq,ewcljrsq),one_half))); |
352 | /* vvdw6 = [C6 - C6grid * (1-poly)]/r6 */ |
353 | vvdw6 = _mm_mul_ps(_mm_sub_ps(c6_00,_mm_mul_ps(c6grid_00,_mm_sub_ps(one,poly))),rinvsix); |
354 | vvdw12 = _mm_mul_ps(c12_00,_mm_mul_ps(rinvsix,rinvsix)); |
355 | vvdw = _mm_sub_ps(_mm_mul_ps(vvdw12,one_twelfth),_mm_mul_ps(vvdw6,one_sixth)); |
356 | /* fvdw = vvdw12/r - (vvdw6/r + (C6grid * exponent * beta^6)/r) */ |
357 | fvdw = _mm_mul_ps(_mm_sub_ps(vvdw12,_mm_sub_ps(vvdw6,_mm_mul_ps(_mm_mul_ps(c6grid_00,one_sixth),_mm_mul_ps(exponent,ewclj6)))),rinvsq00); |
358 | |
359 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
360 | vvdwsum = _mm_add_ps(vvdwsum,vvdw); |
361 | |
362 | fscal = fvdw; |
363 | |
364 | /* Calculate temporary vectorial force */ |
365 | tx = _mm_mul_ps(fscal,dx00); |
366 | ty = _mm_mul_ps(fscal,dy00); |
367 | tz = _mm_mul_ps(fscal,dz00); |
368 | |
369 | /* Update vectorial force */ |
370 | fix0 = _mm_add_ps(fix0,tx); |
371 | fiy0 = _mm_add_ps(fiy0,ty); |
372 | fiz0 = _mm_add_ps(fiz0,tz); |
373 | |
374 | fjx0 = _mm_add_ps(fjx0,tx); |
375 | fjy0 = _mm_add_ps(fjy0,ty); |
376 | fjz0 = _mm_add_ps(fjz0,tz); |
377 | |
378 | /************************** |
379 | * CALCULATE INTERACTIONS * |
380 | **************************/ |
381 | |
382 | r11 = _mm_mul_ps(rsq11,rinv11); |
383 | |
384 | /* EWALD ELECTROSTATICS */ |
385 | |
386 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
387 | ewrt = _mm_mul_ps(r11,ewtabscale); |
388 | ewitab = _mm_cvttps_epi32(ewrt); |
389 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
390 | ewitab = _mm_slli_epi32(ewitab,2); |
391 | ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})) ); |
392 | ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})) ); |
393 | ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})) ); |
394 | ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})) ); |
395 | _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn)do { __m128 tmp3, tmp2, tmp1, tmp0; tmp0 = _mm_unpacklo_ps((ewtabF ), (ewtabD)); tmp2 = _mm_unpacklo_ps((ewtabV), (ewtabFn)); tmp1 = _mm_unpackhi_ps((ewtabF), (ewtabD)); tmp3 = _mm_unpackhi_ps ((ewtabV), (ewtabFn)); (ewtabF) = _mm_movelh_ps(tmp0, tmp2); ( ewtabD) = _mm_movehl_ps(tmp2, tmp0); (ewtabV) = _mm_movelh_ps (tmp1, tmp3); (ewtabFn) = _mm_movehl_ps(tmp3, tmp1); } while ( 0); |
396 | felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD)); |
397 | velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec))); |
398 | velec = _mm_mul_ps(qq11,_mm_sub_ps(rinv11,velec)); |
399 | felec = _mm_mul_ps(_mm_mul_ps(qq11,rinv11),_mm_sub_ps(rinvsq11,felec)); |
400 | |
401 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
402 | velecsum = _mm_add_ps(velecsum,velec); |
403 | |
404 | fscal = felec; |
405 | |
406 | /* Calculate temporary vectorial force */ |
407 | tx = _mm_mul_ps(fscal,dx11); |
408 | ty = _mm_mul_ps(fscal,dy11); |
409 | tz = _mm_mul_ps(fscal,dz11); |
410 | |
411 | /* Update vectorial force */ |
412 | fix1 = _mm_add_ps(fix1,tx); |
413 | fiy1 = _mm_add_ps(fiy1,ty); |
414 | fiz1 = _mm_add_ps(fiz1,tz); |
415 | |
416 | fjx1 = _mm_add_ps(fjx1,tx); |
417 | fjy1 = _mm_add_ps(fjy1,ty); |
418 | fjz1 = _mm_add_ps(fjz1,tz); |
419 | |
420 | /************************** |
421 | * CALCULATE INTERACTIONS * |
422 | **************************/ |
423 | |
424 | r12 = _mm_mul_ps(rsq12,rinv12); |
425 | |
426 | /* EWALD ELECTROSTATICS */ |
427 | |
428 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
429 | ewrt = _mm_mul_ps(r12,ewtabscale); |
430 | ewitab = _mm_cvttps_epi32(ewrt); |
431 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
432 | ewitab = _mm_slli_epi32(ewitab,2); |
433 | ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})) ); |
434 | ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})) ); |
435 | ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})) ); |
436 | ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})) ); |
437 | _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn)do { __m128 tmp3, tmp2, tmp1, tmp0; tmp0 = _mm_unpacklo_ps((ewtabF ), (ewtabD)); tmp2 = _mm_unpacklo_ps((ewtabV), (ewtabFn)); tmp1 = _mm_unpackhi_ps((ewtabF), (ewtabD)); tmp3 = _mm_unpackhi_ps ((ewtabV), (ewtabFn)); (ewtabF) = _mm_movelh_ps(tmp0, tmp2); ( ewtabD) = _mm_movehl_ps(tmp2, tmp0); (ewtabV) = _mm_movelh_ps (tmp1, tmp3); (ewtabFn) = _mm_movehl_ps(tmp3, tmp1); } while ( 0); |
438 | felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD)); |
439 | velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec))); |
440 | velec = _mm_mul_ps(qq12,_mm_sub_ps(rinv12,velec)); |
441 | felec = _mm_mul_ps(_mm_mul_ps(qq12,rinv12),_mm_sub_ps(rinvsq12,felec)); |
442 | |
443 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
444 | velecsum = _mm_add_ps(velecsum,velec); |
445 | |
446 | fscal = felec; |
447 | |
448 | /* Calculate temporary vectorial force */ |
449 | tx = _mm_mul_ps(fscal,dx12); |
450 | ty = _mm_mul_ps(fscal,dy12); |
451 | tz = _mm_mul_ps(fscal,dz12); |
452 | |
453 | /* Update vectorial force */ |
454 | fix1 = _mm_add_ps(fix1,tx); |
455 | fiy1 = _mm_add_ps(fiy1,ty); |
456 | fiz1 = _mm_add_ps(fiz1,tz); |
457 | |
458 | fjx2 = _mm_add_ps(fjx2,tx); |
459 | fjy2 = _mm_add_ps(fjy2,ty); |
460 | fjz2 = _mm_add_ps(fjz2,tz); |
461 | |
462 | /************************** |
463 | * CALCULATE INTERACTIONS * |
464 | **************************/ |
465 | |
466 | r13 = _mm_mul_ps(rsq13,rinv13); |
467 | |
468 | /* EWALD ELECTROSTATICS */ |
469 | |
470 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
471 | ewrt = _mm_mul_ps(r13,ewtabscale); |
472 | ewitab = _mm_cvttps_epi32(ewrt); |
473 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
474 | ewitab = _mm_slli_epi32(ewitab,2); |
475 | ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})) ); |
476 | ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})) ); |
477 | ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})) ); |
478 | ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})) ); |
479 | _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn)do { __m128 tmp3, tmp2, tmp1, tmp0; tmp0 = _mm_unpacklo_ps((ewtabF ), (ewtabD)); tmp2 = _mm_unpacklo_ps((ewtabV), (ewtabFn)); tmp1 = _mm_unpackhi_ps((ewtabF), (ewtabD)); tmp3 = _mm_unpackhi_ps ((ewtabV), (ewtabFn)); (ewtabF) = _mm_movelh_ps(tmp0, tmp2); ( ewtabD) = _mm_movehl_ps(tmp2, tmp0); (ewtabV) = _mm_movelh_ps (tmp1, tmp3); (ewtabFn) = _mm_movehl_ps(tmp3, tmp1); } while ( 0); |
480 | felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD)); |
481 | velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec))); |
482 | velec = _mm_mul_ps(qq13,_mm_sub_ps(rinv13,velec)); |
483 | felec = _mm_mul_ps(_mm_mul_ps(qq13,rinv13),_mm_sub_ps(rinvsq13,felec)); |
484 | |
485 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
486 | velecsum = _mm_add_ps(velecsum,velec); |
487 | |
488 | fscal = felec; |
489 | |
490 | /* Calculate temporary vectorial force */ |
491 | tx = _mm_mul_ps(fscal,dx13); |
492 | ty = _mm_mul_ps(fscal,dy13); |
493 | tz = _mm_mul_ps(fscal,dz13); |
494 | |
495 | /* Update vectorial force */ |
496 | fix1 = _mm_add_ps(fix1,tx); |
497 | fiy1 = _mm_add_ps(fiy1,ty); |
498 | fiz1 = _mm_add_ps(fiz1,tz); |
499 | |
500 | fjx3 = _mm_add_ps(fjx3,tx); |
501 | fjy3 = _mm_add_ps(fjy3,ty); |
502 | fjz3 = _mm_add_ps(fjz3,tz); |
503 | |
504 | /************************** |
505 | * CALCULATE INTERACTIONS * |
506 | **************************/ |
507 | |
508 | r21 = _mm_mul_ps(rsq21,rinv21); |
509 | |
510 | /* EWALD ELECTROSTATICS */ |
511 | |
512 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
513 | ewrt = _mm_mul_ps(r21,ewtabscale); |
514 | ewitab = _mm_cvttps_epi32(ewrt); |
515 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
516 | ewitab = _mm_slli_epi32(ewitab,2); |
517 | ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})) ); |
518 | ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})) ); |
519 | ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})) ); |
520 | ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})) ); |
521 | _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn)do { __m128 tmp3, tmp2, tmp1, tmp0; tmp0 = _mm_unpacklo_ps((ewtabF ), (ewtabD)); tmp2 = _mm_unpacklo_ps((ewtabV), (ewtabFn)); tmp1 = _mm_unpackhi_ps((ewtabF), (ewtabD)); tmp3 = _mm_unpackhi_ps ((ewtabV), (ewtabFn)); (ewtabF) = _mm_movelh_ps(tmp0, tmp2); ( ewtabD) = _mm_movehl_ps(tmp2, tmp0); (ewtabV) = _mm_movelh_ps (tmp1, tmp3); (ewtabFn) = _mm_movehl_ps(tmp3, tmp1); } while ( 0); |
522 | felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD)); |
523 | velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec))); |
524 | velec = _mm_mul_ps(qq21,_mm_sub_ps(rinv21,velec)); |
525 | felec = _mm_mul_ps(_mm_mul_ps(qq21,rinv21),_mm_sub_ps(rinvsq21,felec)); |
526 | |
527 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
528 | velecsum = _mm_add_ps(velecsum,velec); |
529 | |
530 | fscal = felec; |
531 | |
532 | /* Calculate temporary vectorial force */ |
533 | tx = _mm_mul_ps(fscal,dx21); |
534 | ty = _mm_mul_ps(fscal,dy21); |
535 | tz = _mm_mul_ps(fscal,dz21); |
536 | |
537 | /* Update vectorial force */ |
538 | fix2 = _mm_add_ps(fix2,tx); |
539 | fiy2 = _mm_add_ps(fiy2,ty); |
540 | fiz2 = _mm_add_ps(fiz2,tz); |
541 | |
542 | fjx1 = _mm_add_ps(fjx1,tx); |
543 | fjy1 = _mm_add_ps(fjy1,ty); |
544 | fjz1 = _mm_add_ps(fjz1,tz); |
545 | |
546 | /************************** |
547 | * CALCULATE INTERACTIONS * |
548 | **************************/ |
549 | |
550 | r22 = _mm_mul_ps(rsq22,rinv22); |
551 | |
552 | /* EWALD ELECTROSTATICS */ |
553 | |
554 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
555 | ewrt = _mm_mul_ps(r22,ewtabscale); |
556 | ewitab = _mm_cvttps_epi32(ewrt); |
557 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
558 | ewitab = _mm_slli_epi32(ewitab,2); |
559 | ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})) ); |
560 | ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})) ); |
561 | ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})) ); |
562 | ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})) ); |
563 | _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn)do { __m128 tmp3, tmp2, tmp1, tmp0; tmp0 = _mm_unpacklo_ps((ewtabF ), (ewtabD)); tmp2 = _mm_unpacklo_ps((ewtabV), (ewtabFn)); tmp1 = _mm_unpackhi_ps((ewtabF), (ewtabD)); tmp3 = _mm_unpackhi_ps ((ewtabV), (ewtabFn)); (ewtabF) = _mm_movelh_ps(tmp0, tmp2); ( ewtabD) = _mm_movehl_ps(tmp2, tmp0); (ewtabV) = _mm_movelh_ps (tmp1, tmp3); (ewtabFn) = _mm_movehl_ps(tmp3, tmp1); } while ( 0); |
564 | felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD)); |
565 | velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec))); |
566 | velec = _mm_mul_ps(qq22,_mm_sub_ps(rinv22,velec)); |
567 | felec = _mm_mul_ps(_mm_mul_ps(qq22,rinv22),_mm_sub_ps(rinvsq22,felec)); |
568 | |
569 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
570 | velecsum = _mm_add_ps(velecsum,velec); |
571 | |
572 | fscal = felec; |
573 | |
574 | /* Calculate temporary vectorial force */ |
575 | tx = _mm_mul_ps(fscal,dx22); |
576 | ty = _mm_mul_ps(fscal,dy22); |
577 | tz = _mm_mul_ps(fscal,dz22); |
578 | |
579 | /* Update vectorial force */ |
580 | fix2 = _mm_add_ps(fix2,tx); |
581 | fiy2 = _mm_add_ps(fiy2,ty); |
582 | fiz2 = _mm_add_ps(fiz2,tz); |
583 | |
584 | fjx2 = _mm_add_ps(fjx2,tx); |
585 | fjy2 = _mm_add_ps(fjy2,ty); |
586 | fjz2 = _mm_add_ps(fjz2,tz); |
587 | |
588 | /************************** |
589 | * CALCULATE INTERACTIONS * |
590 | **************************/ |
591 | |
592 | r23 = _mm_mul_ps(rsq23,rinv23); |
593 | |
594 | /* EWALD ELECTROSTATICS */ |
595 | |
596 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
597 | ewrt = _mm_mul_ps(r23,ewtabscale); |
598 | ewitab = _mm_cvttps_epi32(ewrt); |
599 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
600 | ewitab = _mm_slli_epi32(ewitab,2); |
601 | ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})) ); |
602 | ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})) ); |
603 | ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})) ); |
604 | ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})) ); |
605 | _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn)do { __m128 tmp3, tmp2, tmp1, tmp0; tmp0 = _mm_unpacklo_ps((ewtabF ), (ewtabD)); tmp2 = _mm_unpacklo_ps((ewtabV), (ewtabFn)); tmp1 = _mm_unpackhi_ps((ewtabF), (ewtabD)); tmp3 = _mm_unpackhi_ps ((ewtabV), (ewtabFn)); (ewtabF) = _mm_movelh_ps(tmp0, tmp2); ( ewtabD) = _mm_movehl_ps(tmp2, tmp0); (ewtabV) = _mm_movelh_ps (tmp1, tmp3); (ewtabFn) = _mm_movehl_ps(tmp3, tmp1); } while ( 0); |
606 | felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD)); |
607 | velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec))); |
608 | velec = _mm_mul_ps(qq23,_mm_sub_ps(rinv23,velec)); |
609 | felec = _mm_mul_ps(_mm_mul_ps(qq23,rinv23),_mm_sub_ps(rinvsq23,felec)); |
610 | |
611 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
612 | velecsum = _mm_add_ps(velecsum,velec); |
613 | |
614 | fscal = felec; |
615 | |
616 | /* Calculate temporary vectorial force */ |
617 | tx = _mm_mul_ps(fscal,dx23); |
618 | ty = _mm_mul_ps(fscal,dy23); |
619 | tz = _mm_mul_ps(fscal,dz23); |
620 | |
621 | /* Update vectorial force */ |
622 | fix2 = _mm_add_ps(fix2,tx); |
623 | fiy2 = _mm_add_ps(fiy2,ty); |
624 | fiz2 = _mm_add_ps(fiz2,tz); |
625 | |
626 | fjx3 = _mm_add_ps(fjx3,tx); |
627 | fjy3 = _mm_add_ps(fjy3,ty); |
628 | fjz3 = _mm_add_ps(fjz3,tz); |
629 | |
630 | /************************** |
631 | * CALCULATE INTERACTIONS * |
632 | **************************/ |
633 | |
634 | r31 = _mm_mul_ps(rsq31,rinv31); |
635 | |
636 | /* EWALD ELECTROSTATICS */ |
637 | |
638 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
639 | ewrt = _mm_mul_ps(r31,ewtabscale); |
640 | ewitab = _mm_cvttps_epi32(ewrt); |
641 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
642 | ewitab = _mm_slli_epi32(ewitab,2); |
643 | ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})) ); |
644 | ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})) ); |
645 | ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})) ); |
646 | ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})) ); |
647 | _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn)do { __m128 tmp3, tmp2, tmp1, tmp0; tmp0 = _mm_unpacklo_ps((ewtabF ), (ewtabD)); tmp2 = _mm_unpacklo_ps((ewtabV), (ewtabFn)); tmp1 = _mm_unpackhi_ps((ewtabF), (ewtabD)); tmp3 = _mm_unpackhi_ps ((ewtabV), (ewtabFn)); (ewtabF) = _mm_movelh_ps(tmp0, tmp2); ( ewtabD) = _mm_movehl_ps(tmp2, tmp0); (ewtabV) = _mm_movelh_ps (tmp1, tmp3); (ewtabFn) = _mm_movehl_ps(tmp3, tmp1); } while ( 0); |
648 | felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD)); |
649 | velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec))); |
650 | velec = _mm_mul_ps(qq31,_mm_sub_ps(rinv31,velec)); |
651 | felec = _mm_mul_ps(_mm_mul_ps(qq31,rinv31),_mm_sub_ps(rinvsq31,felec)); |
652 | |
653 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
654 | velecsum = _mm_add_ps(velecsum,velec); |
655 | |
656 | fscal = felec; |
657 | |
658 | /* Calculate temporary vectorial force */ |
659 | tx = _mm_mul_ps(fscal,dx31); |
660 | ty = _mm_mul_ps(fscal,dy31); |
661 | tz = _mm_mul_ps(fscal,dz31); |
662 | |
663 | /* Update vectorial force */ |
664 | fix3 = _mm_add_ps(fix3,tx); |
665 | fiy3 = _mm_add_ps(fiy3,ty); |
666 | fiz3 = _mm_add_ps(fiz3,tz); |
667 | |
668 | fjx1 = _mm_add_ps(fjx1,tx); |
669 | fjy1 = _mm_add_ps(fjy1,ty); |
670 | fjz1 = _mm_add_ps(fjz1,tz); |
671 | |
672 | /************************** |
673 | * CALCULATE INTERACTIONS * |
674 | **************************/ |
675 | |
676 | r32 = _mm_mul_ps(rsq32,rinv32); |
677 | |
678 | /* EWALD ELECTROSTATICS */ |
679 | |
680 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
681 | ewrt = _mm_mul_ps(r32,ewtabscale); |
682 | ewitab = _mm_cvttps_epi32(ewrt); |
683 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
684 | ewitab = _mm_slli_epi32(ewitab,2); |
685 | ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})) ); |
686 | ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})) ); |
687 | ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})) ); |
688 | ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})) ); |
689 | _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn)do { __m128 tmp3, tmp2, tmp1, tmp0; tmp0 = _mm_unpacklo_ps((ewtabF ), (ewtabD)); tmp2 = _mm_unpacklo_ps((ewtabV), (ewtabFn)); tmp1 = _mm_unpackhi_ps((ewtabF), (ewtabD)); tmp3 = _mm_unpackhi_ps ((ewtabV), (ewtabFn)); (ewtabF) = _mm_movelh_ps(tmp0, tmp2); ( ewtabD) = _mm_movehl_ps(tmp2, tmp0); (ewtabV) = _mm_movelh_ps (tmp1, tmp3); (ewtabFn) = _mm_movehl_ps(tmp3, tmp1); } while ( 0); |
690 | felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD)); |
691 | velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec))); |
692 | velec = _mm_mul_ps(qq32,_mm_sub_ps(rinv32,velec)); |
693 | felec = _mm_mul_ps(_mm_mul_ps(qq32,rinv32),_mm_sub_ps(rinvsq32,felec)); |
694 | |
695 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
696 | velecsum = _mm_add_ps(velecsum,velec); |
697 | |
698 | fscal = felec; |
699 | |
700 | /* Calculate temporary vectorial force */ |
701 | tx = _mm_mul_ps(fscal,dx32); |
702 | ty = _mm_mul_ps(fscal,dy32); |
703 | tz = _mm_mul_ps(fscal,dz32); |
704 | |
705 | /* Update vectorial force */ |
706 | fix3 = _mm_add_ps(fix3,tx); |
707 | fiy3 = _mm_add_ps(fiy3,ty); |
708 | fiz3 = _mm_add_ps(fiz3,tz); |
709 | |
710 | fjx2 = _mm_add_ps(fjx2,tx); |
711 | fjy2 = _mm_add_ps(fjy2,ty); |
712 | fjz2 = _mm_add_ps(fjz2,tz); |
713 | |
714 | /************************** |
715 | * CALCULATE INTERACTIONS * |
716 | **************************/ |
717 | |
718 | r33 = _mm_mul_ps(rsq33,rinv33); |
719 | |
720 | /* EWALD ELECTROSTATICS */ |
721 | |
722 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
723 | ewrt = _mm_mul_ps(r33,ewtabscale); |
724 | ewitab = _mm_cvttps_epi32(ewrt); |
725 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
726 | ewitab = _mm_slli_epi32(ewitab,2); |
727 | ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})) ); |
728 | ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})) ); |
729 | ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})) ); |
730 | ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})) ); |
731 | _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn)do { __m128 tmp3, tmp2, tmp1, tmp0; tmp0 = _mm_unpacklo_ps((ewtabF ), (ewtabD)); tmp2 = _mm_unpacklo_ps((ewtabV), (ewtabFn)); tmp1 = _mm_unpackhi_ps((ewtabF), (ewtabD)); tmp3 = _mm_unpackhi_ps ((ewtabV), (ewtabFn)); (ewtabF) = _mm_movelh_ps(tmp0, tmp2); ( ewtabD) = _mm_movehl_ps(tmp2, tmp0); (ewtabV) = _mm_movelh_ps (tmp1, tmp3); (ewtabFn) = _mm_movehl_ps(tmp3, tmp1); } while ( 0); |
732 | felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD)); |
733 | velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec))); |
734 | velec = _mm_mul_ps(qq33,_mm_sub_ps(rinv33,velec)); |
735 | felec = _mm_mul_ps(_mm_mul_ps(qq33,rinv33),_mm_sub_ps(rinvsq33,felec)); |
736 | |
737 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
738 | velecsum = _mm_add_ps(velecsum,velec); |
739 | |
740 | fscal = felec; |
741 | |
742 | /* Calculate temporary vectorial force */ |
743 | tx = _mm_mul_ps(fscal,dx33); |
744 | ty = _mm_mul_ps(fscal,dy33); |
745 | tz = _mm_mul_ps(fscal,dz33); |
746 | |
747 | /* Update vectorial force */ |
748 | fix3 = _mm_add_ps(fix3,tx); |
749 | fiy3 = _mm_add_ps(fiy3,ty); |
750 | fiz3 = _mm_add_ps(fiz3,tz); |
751 | |
752 | fjx3 = _mm_add_ps(fjx3,tx); |
753 | fjy3 = _mm_add_ps(fjy3,ty); |
754 | fjz3 = _mm_add_ps(fjz3,tz); |
755 | |
756 | fjptrA = f+j_coord_offsetA; |
757 | fjptrB = f+j_coord_offsetB; |
758 | fjptrC = f+j_coord_offsetC; |
759 | fjptrD = f+j_coord_offsetD; |
760 | |
761 | gmx_mm_decrement_4rvec_4ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD, |
762 | fjx0,fjy0,fjz0,fjx1,fjy1,fjz1, |
763 | fjx2,fjy2,fjz2,fjx3,fjy3,fjz3); |
764 | |
765 | /* Inner loop uses 423 flops */ |
766 | } |
767 | |
768 | if(jidx<j_index_end) |
769 | { |
770 | |
771 | /* Get j neighbor index, and coordinate index */ |
772 | jnrlistA = jjnr[jidx]; |
773 | jnrlistB = jjnr[jidx+1]; |
774 | jnrlistC = jjnr[jidx+2]; |
775 | jnrlistD = jjnr[jidx+3]; |
776 | /* Sign of each element will be negative for non-real atoms. |
777 | * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones, |
778 | * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries. |
779 | */ |
780 | dummy_mask = gmx_mm_castsi128_ps_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128())); |
781 | jnrA = (jnrlistA>=0) ? jnrlistA : 0; |
782 | jnrB = (jnrlistB>=0) ? jnrlistB : 0; |
783 | jnrC = (jnrlistC>=0) ? jnrlistC : 0; |
784 | jnrD = (jnrlistD>=0) ? jnrlistD : 0; |
785 | j_coord_offsetA = DIM3*jnrA; |
786 | j_coord_offsetB = DIM3*jnrB; |
787 | j_coord_offsetC = DIM3*jnrC; |
788 | j_coord_offsetD = DIM3*jnrD; |
789 | |
790 | /* load j atom coordinates */ |
791 | gmx_mm_load_4rvec_4ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB, |
792 | x+j_coord_offsetC,x+j_coord_offsetD, |
793 | &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2, |
794 | &jy2,&jz2,&jx3,&jy3,&jz3); |
795 | |
796 | /* Calculate displacement vector */ |
797 | dx00 = _mm_sub_ps(ix0,jx0); |
798 | dy00 = _mm_sub_ps(iy0,jy0); |
799 | dz00 = _mm_sub_ps(iz0,jz0); |
800 | dx11 = _mm_sub_ps(ix1,jx1); |
801 | dy11 = _mm_sub_ps(iy1,jy1); |
802 | dz11 = _mm_sub_ps(iz1,jz1); |
803 | dx12 = _mm_sub_ps(ix1,jx2); |
804 | dy12 = _mm_sub_ps(iy1,jy2); |
805 | dz12 = _mm_sub_ps(iz1,jz2); |
806 | dx13 = _mm_sub_ps(ix1,jx3); |
807 | dy13 = _mm_sub_ps(iy1,jy3); |
808 | dz13 = _mm_sub_ps(iz1,jz3); |
809 | dx21 = _mm_sub_ps(ix2,jx1); |
810 | dy21 = _mm_sub_ps(iy2,jy1); |
811 | dz21 = _mm_sub_ps(iz2,jz1); |
812 | dx22 = _mm_sub_ps(ix2,jx2); |
813 | dy22 = _mm_sub_ps(iy2,jy2); |
814 | dz22 = _mm_sub_ps(iz2,jz2); |
815 | dx23 = _mm_sub_ps(ix2,jx3); |
816 | dy23 = _mm_sub_ps(iy2,jy3); |
817 | dz23 = _mm_sub_ps(iz2,jz3); |
818 | dx31 = _mm_sub_ps(ix3,jx1); |
819 | dy31 = _mm_sub_ps(iy3,jy1); |
820 | dz31 = _mm_sub_ps(iz3,jz1); |
821 | dx32 = _mm_sub_ps(ix3,jx2); |
822 | dy32 = _mm_sub_ps(iy3,jy2); |
823 | dz32 = _mm_sub_ps(iz3,jz2); |
824 | dx33 = _mm_sub_ps(ix3,jx3); |
825 | dy33 = _mm_sub_ps(iy3,jy3); |
826 | dz33 = _mm_sub_ps(iz3,jz3); |
827 | |
828 | /* Calculate squared distance and things based on it */ |
829 | rsq00 = gmx_mm_calc_rsq_ps(dx00,dy00,dz00); |
830 | rsq11 = gmx_mm_calc_rsq_ps(dx11,dy11,dz11); |
831 | rsq12 = gmx_mm_calc_rsq_ps(dx12,dy12,dz12); |
832 | rsq13 = gmx_mm_calc_rsq_ps(dx13,dy13,dz13); |
833 | rsq21 = gmx_mm_calc_rsq_ps(dx21,dy21,dz21); |
834 | rsq22 = gmx_mm_calc_rsq_ps(dx22,dy22,dz22); |
835 | rsq23 = gmx_mm_calc_rsq_ps(dx23,dy23,dz23); |
836 | rsq31 = gmx_mm_calc_rsq_ps(dx31,dy31,dz31); |
837 | rsq32 = gmx_mm_calc_rsq_ps(dx32,dy32,dz32); |
838 | rsq33 = gmx_mm_calc_rsq_ps(dx33,dy33,dz33); |
839 | |
840 | rinv00 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq00); |
841 | rinv11 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq11); |
842 | rinv12 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq12); |
843 | rinv13 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq13); |
844 | rinv21 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq21); |
845 | rinv22 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq22); |
846 | rinv23 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq23); |
847 | rinv31 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq31); |
848 | rinv32 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq32); |
849 | rinv33 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq33); |
850 | |
851 | rinvsq00 = _mm_mul_ps(rinv00,rinv00); |
852 | rinvsq11 = _mm_mul_ps(rinv11,rinv11); |
853 | rinvsq12 = _mm_mul_ps(rinv12,rinv12); |
854 | rinvsq13 = _mm_mul_ps(rinv13,rinv13); |
855 | rinvsq21 = _mm_mul_ps(rinv21,rinv21); |
856 | rinvsq22 = _mm_mul_ps(rinv22,rinv22); |
857 | rinvsq23 = _mm_mul_ps(rinv23,rinv23); |
858 | rinvsq31 = _mm_mul_ps(rinv31,rinv31); |
859 | rinvsq32 = _mm_mul_ps(rinv32,rinv32); |
860 | rinvsq33 = _mm_mul_ps(rinv33,rinv33); |
861 | |
862 | fjx0 = _mm_setzero_ps(); |
863 | fjy0 = _mm_setzero_ps(); |
864 | fjz0 = _mm_setzero_ps(); |
865 | fjx1 = _mm_setzero_ps(); |
866 | fjy1 = _mm_setzero_ps(); |
867 | fjz1 = _mm_setzero_ps(); |
868 | fjx2 = _mm_setzero_ps(); |
869 | fjy2 = _mm_setzero_ps(); |
870 | fjz2 = _mm_setzero_ps(); |
871 | fjx3 = _mm_setzero_ps(); |
872 | fjy3 = _mm_setzero_ps(); |
873 | fjz3 = _mm_setzero_ps(); |
874 | |
875 | /************************** |
876 | * CALCULATE INTERACTIONS * |
877 | **************************/ |
878 | |
879 | r00 = _mm_mul_ps(rsq00,rinv00); |
880 | r00 = _mm_andnot_ps(dummy_mask,r00); |
881 | |
882 | /* Analytical LJ-PME */ |
883 | rinvsix = _mm_mul_ps(_mm_mul_ps(rinvsq00,rinvsq00),rinvsq00); |
884 | ewcljrsq = _mm_mul_ps(ewclj2,rsq00); |
885 | ewclj6 = _mm_mul_ps(ewclj2,_mm_mul_ps(ewclj2,ewclj2)); |
886 | exponent = gmx_simd_exp_rgmx_simd_exp_f(ewcljrsq); |
887 | /* poly = exp(-(beta*r)^2) * (1 + (beta*r)^2 + (beta*r)^4 /2) */ |
888 | poly = _mm_mul_ps(exponent,_mm_add_ps(_mm_sub_ps(one,ewcljrsq),_mm_mul_ps(_mm_mul_ps(ewcljrsq,ewcljrsq),one_half))); |
889 | /* vvdw6 = [C6 - C6grid * (1-poly)]/r6 */ |
890 | vvdw6 = _mm_mul_ps(_mm_sub_ps(c6_00,_mm_mul_ps(c6grid_00,_mm_sub_ps(one,poly))),rinvsix); |
891 | vvdw12 = _mm_mul_ps(c12_00,_mm_mul_ps(rinvsix,rinvsix)); |
892 | vvdw = _mm_sub_ps(_mm_mul_ps(vvdw12,one_twelfth),_mm_mul_ps(vvdw6,one_sixth)); |
893 | /* fvdw = vvdw12/r - (vvdw6/r + (C6grid * exponent * beta^6)/r) */ |
894 | fvdw = _mm_mul_ps(_mm_sub_ps(vvdw12,_mm_sub_ps(vvdw6,_mm_mul_ps(_mm_mul_ps(c6grid_00,one_sixth),_mm_mul_ps(exponent,ewclj6)))),rinvsq00); |
895 | |
896 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
897 | vvdw = _mm_andnot_ps(dummy_mask,vvdw); |
898 | vvdwsum = _mm_add_ps(vvdwsum,vvdw); |
899 | |
900 | fscal = fvdw; |
901 | |
902 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
903 | |
904 | /* Calculate temporary vectorial force */ |
905 | tx = _mm_mul_ps(fscal,dx00); |
906 | ty = _mm_mul_ps(fscal,dy00); |
907 | tz = _mm_mul_ps(fscal,dz00); |
908 | |
909 | /* Update vectorial force */ |
910 | fix0 = _mm_add_ps(fix0,tx); |
911 | fiy0 = _mm_add_ps(fiy0,ty); |
912 | fiz0 = _mm_add_ps(fiz0,tz); |
913 | |
914 | fjx0 = _mm_add_ps(fjx0,tx); |
915 | fjy0 = _mm_add_ps(fjy0,ty); |
916 | fjz0 = _mm_add_ps(fjz0,tz); |
917 | |
918 | /************************** |
919 | * CALCULATE INTERACTIONS * |
920 | **************************/ |
921 | |
922 | r11 = _mm_mul_ps(rsq11,rinv11); |
923 | r11 = _mm_andnot_ps(dummy_mask,r11); |
924 | |
925 | /* EWALD ELECTROSTATICS */ |
926 | |
927 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
928 | ewrt = _mm_mul_ps(r11,ewtabscale); |
929 | ewitab = _mm_cvttps_epi32(ewrt); |
930 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
931 | ewitab = _mm_slli_epi32(ewitab,2); |
932 | ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})) ); |
933 | ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})) ); |
934 | ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})) ); |
935 | ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})) ); |
936 | _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn)do { __m128 tmp3, tmp2, tmp1, tmp0; tmp0 = _mm_unpacklo_ps((ewtabF ), (ewtabD)); tmp2 = _mm_unpacklo_ps((ewtabV), (ewtabFn)); tmp1 = _mm_unpackhi_ps((ewtabF), (ewtabD)); tmp3 = _mm_unpackhi_ps ((ewtabV), (ewtabFn)); (ewtabF) = _mm_movelh_ps(tmp0, tmp2); ( ewtabD) = _mm_movehl_ps(tmp2, tmp0); (ewtabV) = _mm_movelh_ps (tmp1, tmp3); (ewtabFn) = _mm_movehl_ps(tmp3, tmp1); } while ( 0); |
937 | felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD)); |
938 | velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec))); |
939 | velec = _mm_mul_ps(qq11,_mm_sub_ps(rinv11,velec)); |
940 | felec = _mm_mul_ps(_mm_mul_ps(qq11,rinv11),_mm_sub_ps(rinvsq11,felec)); |
941 | |
942 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
943 | velec = _mm_andnot_ps(dummy_mask,velec); |
944 | velecsum = _mm_add_ps(velecsum,velec); |
945 | |
946 | fscal = felec; |
947 | |
948 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
949 | |
950 | /* Calculate temporary vectorial force */ |
951 | tx = _mm_mul_ps(fscal,dx11); |
952 | ty = _mm_mul_ps(fscal,dy11); |
953 | tz = _mm_mul_ps(fscal,dz11); |
954 | |
955 | /* Update vectorial force */ |
956 | fix1 = _mm_add_ps(fix1,tx); |
957 | fiy1 = _mm_add_ps(fiy1,ty); |
958 | fiz1 = _mm_add_ps(fiz1,tz); |
959 | |
960 | fjx1 = _mm_add_ps(fjx1,tx); |
961 | fjy1 = _mm_add_ps(fjy1,ty); |
962 | fjz1 = _mm_add_ps(fjz1,tz); |
963 | |
964 | /************************** |
965 | * CALCULATE INTERACTIONS * |
966 | **************************/ |
967 | |
968 | r12 = _mm_mul_ps(rsq12,rinv12); |
969 | r12 = _mm_andnot_ps(dummy_mask,r12); |
970 | |
971 | /* EWALD ELECTROSTATICS */ |
972 | |
973 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
974 | ewrt = _mm_mul_ps(r12,ewtabscale); |
975 | ewitab = _mm_cvttps_epi32(ewrt); |
976 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
977 | ewitab = _mm_slli_epi32(ewitab,2); |
978 | ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})) ); |
979 | ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})) ); |
980 | ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})) ); |
981 | ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})) ); |
982 | _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn)do { __m128 tmp3, tmp2, tmp1, tmp0; tmp0 = _mm_unpacklo_ps((ewtabF ), (ewtabD)); tmp2 = _mm_unpacklo_ps((ewtabV), (ewtabFn)); tmp1 = _mm_unpackhi_ps((ewtabF), (ewtabD)); tmp3 = _mm_unpackhi_ps ((ewtabV), (ewtabFn)); (ewtabF) = _mm_movelh_ps(tmp0, tmp2); ( ewtabD) = _mm_movehl_ps(tmp2, tmp0); (ewtabV) = _mm_movelh_ps (tmp1, tmp3); (ewtabFn) = _mm_movehl_ps(tmp3, tmp1); } while ( 0); |
983 | felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD)); |
984 | velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec))); |
985 | velec = _mm_mul_ps(qq12,_mm_sub_ps(rinv12,velec)); |
986 | felec = _mm_mul_ps(_mm_mul_ps(qq12,rinv12),_mm_sub_ps(rinvsq12,felec)); |
987 | |
988 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
989 | velec = _mm_andnot_ps(dummy_mask,velec); |
990 | velecsum = _mm_add_ps(velecsum,velec); |
991 | |
992 | fscal = felec; |
993 | |
994 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
995 | |
996 | /* Calculate temporary vectorial force */ |
997 | tx = _mm_mul_ps(fscal,dx12); |
998 | ty = _mm_mul_ps(fscal,dy12); |
999 | tz = _mm_mul_ps(fscal,dz12); |
1000 | |
1001 | /* Update vectorial force */ |
1002 | fix1 = _mm_add_ps(fix1,tx); |
1003 | fiy1 = _mm_add_ps(fiy1,ty); |
1004 | fiz1 = _mm_add_ps(fiz1,tz); |
1005 | |
1006 | fjx2 = _mm_add_ps(fjx2,tx); |
1007 | fjy2 = _mm_add_ps(fjy2,ty); |
1008 | fjz2 = _mm_add_ps(fjz2,tz); |
1009 | |
1010 | /************************** |
1011 | * CALCULATE INTERACTIONS * |
1012 | **************************/ |
1013 | |
1014 | r13 = _mm_mul_ps(rsq13,rinv13); |
1015 | r13 = _mm_andnot_ps(dummy_mask,r13); |
1016 | |
1017 | /* EWALD ELECTROSTATICS */ |
1018 | |
1019 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1020 | ewrt = _mm_mul_ps(r13,ewtabscale); |
1021 | ewitab = _mm_cvttps_epi32(ewrt); |
1022 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
1023 | ewitab = _mm_slli_epi32(ewitab,2); |
1024 | ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})) ); |
1025 | ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})) ); |
1026 | ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})) ); |
1027 | ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})) ); |
1028 | _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn)do { __m128 tmp3, tmp2, tmp1, tmp0; tmp0 = _mm_unpacklo_ps((ewtabF ), (ewtabD)); tmp2 = _mm_unpacklo_ps((ewtabV), (ewtabFn)); tmp1 = _mm_unpackhi_ps((ewtabF), (ewtabD)); tmp3 = _mm_unpackhi_ps ((ewtabV), (ewtabFn)); (ewtabF) = _mm_movelh_ps(tmp0, tmp2); ( ewtabD) = _mm_movehl_ps(tmp2, tmp0); (ewtabV) = _mm_movelh_ps (tmp1, tmp3); (ewtabFn) = _mm_movehl_ps(tmp3, tmp1); } while ( 0); |
1029 | felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD)); |
1030 | velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec))); |
1031 | velec = _mm_mul_ps(qq13,_mm_sub_ps(rinv13,velec)); |
1032 | felec = _mm_mul_ps(_mm_mul_ps(qq13,rinv13),_mm_sub_ps(rinvsq13,felec)); |
1033 | |
1034 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
1035 | velec = _mm_andnot_ps(dummy_mask,velec); |
1036 | velecsum = _mm_add_ps(velecsum,velec); |
1037 | |
1038 | fscal = felec; |
1039 | |
1040 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
1041 | |
1042 | /* Calculate temporary vectorial force */ |
1043 | tx = _mm_mul_ps(fscal,dx13); |
1044 | ty = _mm_mul_ps(fscal,dy13); |
1045 | tz = _mm_mul_ps(fscal,dz13); |
1046 | |
1047 | /* Update vectorial force */ |
1048 | fix1 = _mm_add_ps(fix1,tx); |
1049 | fiy1 = _mm_add_ps(fiy1,ty); |
1050 | fiz1 = _mm_add_ps(fiz1,tz); |
1051 | |
1052 | fjx3 = _mm_add_ps(fjx3,tx); |
1053 | fjy3 = _mm_add_ps(fjy3,ty); |
1054 | fjz3 = _mm_add_ps(fjz3,tz); |
1055 | |
1056 | /************************** |
1057 | * CALCULATE INTERACTIONS * |
1058 | **************************/ |
1059 | |
1060 | r21 = _mm_mul_ps(rsq21,rinv21); |
1061 | r21 = _mm_andnot_ps(dummy_mask,r21); |
1062 | |
1063 | /* EWALD ELECTROSTATICS */ |
1064 | |
1065 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1066 | ewrt = _mm_mul_ps(r21,ewtabscale); |
1067 | ewitab = _mm_cvttps_epi32(ewrt); |
1068 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
1069 | ewitab = _mm_slli_epi32(ewitab,2); |
1070 | ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})) ); |
1071 | ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})) ); |
1072 | ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})) ); |
1073 | ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})) ); |
1074 | _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn)do { __m128 tmp3, tmp2, tmp1, tmp0; tmp0 = _mm_unpacklo_ps((ewtabF ), (ewtabD)); tmp2 = _mm_unpacklo_ps((ewtabV), (ewtabFn)); tmp1 = _mm_unpackhi_ps((ewtabF), (ewtabD)); tmp3 = _mm_unpackhi_ps ((ewtabV), (ewtabFn)); (ewtabF) = _mm_movelh_ps(tmp0, tmp2); ( ewtabD) = _mm_movehl_ps(tmp2, tmp0); (ewtabV) = _mm_movelh_ps (tmp1, tmp3); (ewtabFn) = _mm_movehl_ps(tmp3, tmp1); } while ( 0); |
1075 | felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD)); |
1076 | velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec))); |
1077 | velec = _mm_mul_ps(qq21,_mm_sub_ps(rinv21,velec)); |
1078 | felec = _mm_mul_ps(_mm_mul_ps(qq21,rinv21),_mm_sub_ps(rinvsq21,felec)); |
1079 | |
1080 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
1081 | velec = _mm_andnot_ps(dummy_mask,velec); |
1082 | velecsum = _mm_add_ps(velecsum,velec); |
1083 | |
1084 | fscal = felec; |
1085 | |
1086 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
1087 | |
1088 | /* Calculate temporary vectorial force */ |
1089 | tx = _mm_mul_ps(fscal,dx21); |
1090 | ty = _mm_mul_ps(fscal,dy21); |
1091 | tz = _mm_mul_ps(fscal,dz21); |
1092 | |
1093 | /* Update vectorial force */ |
1094 | fix2 = _mm_add_ps(fix2,tx); |
1095 | fiy2 = _mm_add_ps(fiy2,ty); |
1096 | fiz2 = _mm_add_ps(fiz2,tz); |
1097 | |
1098 | fjx1 = _mm_add_ps(fjx1,tx); |
1099 | fjy1 = _mm_add_ps(fjy1,ty); |
1100 | fjz1 = _mm_add_ps(fjz1,tz); |
1101 | |
1102 | /************************** |
1103 | * CALCULATE INTERACTIONS * |
1104 | **************************/ |
1105 | |
1106 | r22 = _mm_mul_ps(rsq22,rinv22); |
1107 | r22 = _mm_andnot_ps(dummy_mask,r22); |
1108 | |
1109 | /* EWALD ELECTROSTATICS */ |
1110 | |
1111 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1112 | ewrt = _mm_mul_ps(r22,ewtabscale); |
1113 | ewitab = _mm_cvttps_epi32(ewrt); |
1114 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
1115 | ewitab = _mm_slli_epi32(ewitab,2); |
1116 | ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})) ); |
1117 | ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})) ); |
1118 | ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})) ); |
1119 | ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})) ); |
1120 | _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn)do { __m128 tmp3, tmp2, tmp1, tmp0; tmp0 = _mm_unpacklo_ps((ewtabF ), (ewtabD)); tmp2 = _mm_unpacklo_ps((ewtabV), (ewtabFn)); tmp1 = _mm_unpackhi_ps((ewtabF), (ewtabD)); tmp3 = _mm_unpackhi_ps ((ewtabV), (ewtabFn)); (ewtabF) = _mm_movelh_ps(tmp0, tmp2); ( ewtabD) = _mm_movehl_ps(tmp2, tmp0); (ewtabV) = _mm_movelh_ps (tmp1, tmp3); (ewtabFn) = _mm_movehl_ps(tmp3, tmp1); } while ( 0); |
1121 | felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD)); |
1122 | velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec))); |
1123 | velec = _mm_mul_ps(qq22,_mm_sub_ps(rinv22,velec)); |
1124 | felec = _mm_mul_ps(_mm_mul_ps(qq22,rinv22),_mm_sub_ps(rinvsq22,felec)); |
1125 | |
1126 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
1127 | velec = _mm_andnot_ps(dummy_mask,velec); |
1128 | velecsum = _mm_add_ps(velecsum,velec); |
1129 | |
1130 | fscal = felec; |
1131 | |
1132 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
1133 | |
1134 | /* Calculate temporary vectorial force */ |
1135 | tx = _mm_mul_ps(fscal,dx22); |
1136 | ty = _mm_mul_ps(fscal,dy22); |
1137 | tz = _mm_mul_ps(fscal,dz22); |
1138 | |
1139 | /* Update vectorial force */ |
1140 | fix2 = _mm_add_ps(fix2,tx); |
1141 | fiy2 = _mm_add_ps(fiy2,ty); |
1142 | fiz2 = _mm_add_ps(fiz2,tz); |
1143 | |
1144 | fjx2 = _mm_add_ps(fjx2,tx); |
1145 | fjy2 = _mm_add_ps(fjy2,ty); |
1146 | fjz2 = _mm_add_ps(fjz2,tz); |
1147 | |
1148 | /************************** |
1149 | * CALCULATE INTERACTIONS * |
1150 | **************************/ |
1151 | |
1152 | r23 = _mm_mul_ps(rsq23,rinv23); |
1153 | r23 = _mm_andnot_ps(dummy_mask,r23); |
1154 | |
1155 | /* EWALD ELECTROSTATICS */ |
1156 | |
1157 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1158 | ewrt = _mm_mul_ps(r23,ewtabscale); |
1159 | ewitab = _mm_cvttps_epi32(ewrt); |
1160 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
1161 | ewitab = _mm_slli_epi32(ewitab,2); |
1162 | ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})) ); |
1163 | ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})) ); |
1164 | ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})) ); |
1165 | ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})) ); |
1166 | _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn)do { __m128 tmp3, tmp2, tmp1, tmp0; tmp0 = _mm_unpacklo_ps((ewtabF ), (ewtabD)); tmp2 = _mm_unpacklo_ps((ewtabV), (ewtabFn)); tmp1 = _mm_unpackhi_ps((ewtabF), (ewtabD)); tmp3 = _mm_unpackhi_ps ((ewtabV), (ewtabFn)); (ewtabF) = _mm_movelh_ps(tmp0, tmp2); ( ewtabD) = _mm_movehl_ps(tmp2, tmp0); (ewtabV) = _mm_movelh_ps (tmp1, tmp3); (ewtabFn) = _mm_movehl_ps(tmp3, tmp1); } while ( 0); |
1167 | felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD)); |
1168 | velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec))); |
1169 | velec = _mm_mul_ps(qq23,_mm_sub_ps(rinv23,velec)); |
1170 | felec = _mm_mul_ps(_mm_mul_ps(qq23,rinv23),_mm_sub_ps(rinvsq23,felec)); |
1171 | |
1172 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
1173 | velec = _mm_andnot_ps(dummy_mask,velec); |
1174 | velecsum = _mm_add_ps(velecsum,velec); |
1175 | |
1176 | fscal = felec; |
1177 | |
1178 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
1179 | |
1180 | /* Calculate temporary vectorial force */ |
1181 | tx = _mm_mul_ps(fscal,dx23); |
1182 | ty = _mm_mul_ps(fscal,dy23); |
1183 | tz = _mm_mul_ps(fscal,dz23); |
1184 | |
1185 | /* Update vectorial force */ |
1186 | fix2 = _mm_add_ps(fix2,tx); |
1187 | fiy2 = _mm_add_ps(fiy2,ty); |
1188 | fiz2 = _mm_add_ps(fiz2,tz); |
1189 | |
1190 | fjx3 = _mm_add_ps(fjx3,tx); |
1191 | fjy3 = _mm_add_ps(fjy3,ty); |
1192 | fjz3 = _mm_add_ps(fjz3,tz); |
1193 | |
1194 | /************************** |
1195 | * CALCULATE INTERACTIONS * |
1196 | **************************/ |
1197 | |
1198 | r31 = _mm_mul_ps(rsq31,rinv31); |
1199 | r31 = _mm_andnot_ps(dummy_mask,r31); |
1200 | |
1201 | /* EWALD ELECTROSTATICS */ |
1202 | |
1203 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1204 | ewrt = _mm_mul_ps(r31,ewtabscale); |
1205 | ewitab = _mm_cvttps_epi32(ewrt); |
1206 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
1207 | ewitab = _mm_slli_epi32(ewitab,2); |
1208 | ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})) ); |
1209 | ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})) ); |
1210 | ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})) ); |
1211 | ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})) ); |
1212 | _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn)do { __m128 tmp3, tmp2, tmp1, tmp0; tmp0 = _mm_unpacklo_ps((ewtabF ), (ewtabD)); tmp2 = _mm_unpacklo_ps((ewtabV), (ewtabFn)); tmp1 = _mm_unpackhi_ps((ewtabF), (ewtabD)); tmp3 = _mm_unpackhi_ps ((ewtabV), (ewtabFn)); (ewtabF) = _mm_movelh_ps(tmp0, tmp2); ( ewtabD) = _mm_movehl_ps(tmp2, tmp0); (ewtabV) = _mm_movelh_ps (tmp1, tmp3); (ewtabFn) = _mm_movehl_ps(tmp3, tmp1); } while ( 0); |
1213 | felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD)); |
1214 | velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec))); |
1215 | velec = _mm_mul_ps(qq31,_mm_sub_ps(rinv31,velec)); |
1216 | felec = _mm_mul_ps(_mm_mul_ps(qq31,rinv31),_mm_sub_ps(rinvsq31,felec)); |
1217 | |
1218 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
1219 | velec = _mm_andnot_ps(dummy_mask,velec); |
1220 | velecsum = _mm_add_ps(velecsum,velec); |
1221 | |
1222 | fscal = felec; |
1223 | |
1224 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
1225 | |
1226 | /* Calculate temporary vectorial force */ |
1227 | tx = _mm_mul_ps(fscal,dx31); |
1228 | ty = _mm_mul_ps(fscal,dy31); |
1229 | tz = _mm_mul_ps(fscal,dz31); |
1230 | |
1231 | /* Update vectorial force */ |
1232 | fix3 = _mm_add_ps(fix3,tx); |
1233 | fiy3 = _mm_add_ps(fiy3,ty); |
1234 | fiz3 = _mm_add_ps(fiz3,tz); |
1235 | |
1236 | fjx1 = _mm_add_ps(fjx1,tx); |
1237 | fjy1 = _mm_add_ps(fjy1,ty); |
1238 | fjz1 = _mm_add_ps(fjz1,tz); |
1239 | |
1240 | /************************** |
1241 | * CALCULATE INTERACTIONS * |
1242 | **************************/ |
1243 | |
1244 | r32 = _mm_mul_ps(rsq32,rinv32); |
1245 | r32 = _mm_andnot_ps(dummy_mask,r32); |
1246 | |
1247 | /* EWALD ELECTROSTATICS */ |
1248 | |
1249 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1250 | ewrt = _mm_mul_ps(r32,ewtabscale); |
1251 | ewitab = _mm_cvttps_epi32(ewrt); |
1252 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
1253 | ewitab = _mm_slli_epi32(ewitab,2); |
1254 | ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})) ); |
1255 | ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})) ); |
1256 | ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})) ); |
1257 | ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})) ); |
1258 | _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn)do { __m128 tmp3, tmp2, tmp1, tmp0; tmp0 = _mm_unpacklo_ps((ewtabF ), (ewtabD)); tmp2 = _mm_unpacklo_ps((ewtabV), (ewtabFn)); tmp1 = _mm_unpackhi_ps((ewtabF), (ewtabD)); tmp3 = _mm_unpackhi_ps ((ewtabV), (ewtabFn)); (ewtabF) = _mm_movelh_ps(tmp0, tmp2); ( ewtabD) = _mm_movehl_ps(tmp2, tmp0); (ewtabV) = _mm_movelh_ps (tmp1, tmp3); (ewtabFn) = _mm_movehl_ps(tmp3, tmp1); } while ( 0); |
1259 | felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD)); |
1260 | velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec))); |
1261 | velec = _mm_mul_ps(qq32,_mm_sub_ps(rinv32,velec)); |
1262 | felec = _mm_mul_ps(_mm_mul_ps(qq32,rinv32),_mm_sub_ps(rinvsq32,felec)); |
1263 | |
1264 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
1265 | velec = _mm_andnot_ps(dummy_mask,velec); |
1266 | velecsum = _mm_add_ps(velecsum,velec); |
1267 | |
1268 | fscal = felec; |
1269 | |
1270 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
1271 | |
1272 | /* Calculate temporary vectorial force */ |
1273 | tx = _mm_mul_ps(fscal,dx32); |
1274 | ty = _mm_mul_ps(fscal,dy32); |
1275 | tz = _mm_mul_ps(fscal,dz32); |
1276 | |
1277 | /* Update vectorial force */ |
1278 | fix3 = _mm_add_ps(fix3,tx); |
1279 | fiy3 = _mm_add_ps(fiy3,ty); |
1280 | fiz3 = _mm_add_ps(fiz3,tz); |
1281 | |
1282 | fjx2 = _mm_add_ps(fjx2,tx); |
1283 | fjy2 = _mm_add_ps(fjy2,ty); |
1284 | fjz2 = _mm_add_ps(fjz2,tz); |
1285 | |
1286 | /************************** |
1287 | * CALCULATE INTERACTIONS * |
1288 | **************************/ |
1289 | |
1290 | r33 = _mm_mul_ps(rsq33,rinv33); |
1291 | r33 = _mm_andnot_ps(dummy_mask,r33); |
1292 | |
1293 | /* EWALD ELECTROSTATICS */ |
1294 | |
1295 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1296 | ewrt = _mm_mul_ps(r33,ewtabscale); |
1297 | ewitab = _mm_cvttps_epi32(ewrt); |
1298 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
1299 | ewitab = _mm_slli_epi32(ewitab,2); |
1300 | ewtabF = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})) ); |
1301 | ewtabD = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})) ); |
1302 | ewtabV = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})) ); |
1303 | ewtabFn = _mm_load_ps( ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})) ); |
1304 | _MM_TRANSPOSE4_PS(ewtabF,ewtabD,ewtabV,ewtabFn)do { __m128 tmp3, tmp2, tmp1, tmp0; tmp0 = _mm_unpacklo_ps((ewtabF ), (ewtabD)); tmp2 = _mm_unpacklo_ps((ewtabV), (ewtabFn)); tmp1 = _mm_unpackhi_ps((ewtabF), (ewtabD)); tmp3 = _mm_unpackhi_ps ((ewtabV), (ewtabFn)); (ewtabF) = _mm_movelh_ps(tmp0, tmp2); ( ewtabD) = _mm_movehl_ps(tmp2, tmp0); (ewtabV) = _mm_movelh_ps (tmp1, tmp3); (ewtabFn) = _mm_movehl_ps(tmp3, tmp1); } while ( 0); |
1305 | felec = _mm_add_ps(ewtabF,_mm_mul_ps(eweps,ewtabD)); |
1306 | velec = _mm_sub_ps(ewtabV,_mm_mul_ps(_mm_mul_ps(ewtabhalfspace,eweps),_mm_add_ps(ewtabF,felec))); |
1307 | velec = _mm_mul_ps(qq33,_mm_sub_ps(rinv33,velec)); |
1308 | felec = _mm_mul_ps(_mm_mul_ps(qq33,rinv33),_mm_sub_ps(rinvsq33,felec)); |
1309 | |
1310 | /* Update potential sum for this i atom from the interaction with this j atom. */ |
1311 | velec = _mm_andnot_ps(dummy_mask,velec); |
1312 | velecsum = _mm_add_ps(velecsum,velec); |
1313 | |
1314 | fscal = felec; |
1315 | |
1316 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
1317 | |
1318 | /* Calculate temporary vectorial force */ |
1319 | tx = _mm_mul_ps(fscal,dx33); |
1320 | ty = _mm_mul_ps(fscal,dy33); |
1321 | tz = _mm_mul_ps(fscal,dz33); |
1322 | |
1323 | /* Update vectorial force */ |
1324 | fix3 = _mm_add_ps(fix3,tx); |
1325 | fiy3 = _mm_add_ps(fiy3,ty); |
1326 | fiz3 = _mm_add_ps(fiz3,tz); |
1327 | |
1328 | fjx3 = _mm_add_ps(fjx3,tx); |
1329 | fjy3 = _mm_add_ps(fjy3,ty); |
1330 | fjz3 = _mm_add_ps(fjz3,tz); |
1331 | |
1332 | fjptrA = (jnrlistA>=0) ? f+j_coord_offsetA : scratch; |
1333 | fjptrB = (jnrlistB>=0) ? f+j_coord_offsetB : scratch; |
1334 | fjptrC = (jnrlistC>=0) ? f+j_coord_offsetC : scratch; |
1335 | fjptrD = (jnrlistD>=0) ? f+j_coord_offsetD : scratch; |
1336 | |
1337 | gmx_mm_decrement_4rvec_4ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD, |
1338 | fjx0,fjy0,fjz0,fjx1,fjy1,fjz1, |
1339 | fjx2,fjy2,fjz2,fjx3,fjy3,fjz3); |
1340 | |
1341 | /* Inner loop uses 433 flops */ |
1342 | } |
1343 | |
1344 | /* End of innermost loop */ |
1345 | |
1346 | gmx_mm_update_iforce_4atom_swizzle_ps(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3, |
1347 | f+i_coord_offset,fshift+i_shift_offset); |
1348 | |
1349 | ggid = gid[iidx]; |
1350 | /* Update potential energies */ |
1351 | gmx_mm_update_1pot_ps(velecsum,kernel_data->energygrp_elec+ggid); |
1352 | gmx_mm_update_1pot_ps(vvdwsum,kernel_data->energygrp_vdw+ggid); |
1353 | |
1354 | /* Increment number of inner iterations */ |
1355 | inneriter += j_index_end - j_index_start; |
1356 | |
1357 | /* Outer loop uses 26 flops */ |
1358 | } |
1359 | |
1360 | /* Increment number of outer iterations */ |
1361 | outeriter += nri; |
1362 | |
1363 | /* Update outer/inner flops */ |
1364 | |
1365 | inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4W4_VF,outeriter*26 + inneriter*433)(nrnb)->n[eNR_NBKERNEL_ELEC_VDW_W4W4_VF] += outeriter*26 + inneriter*433; |
1366 | } |
1367 | /* |
1368 | * Gromacs nonbonded kernel: nb_kernel_ElecEw_VdwLJEw_GeomW4W4_F_sse4_1_single |
1369 | * Electrostatics interaction: Ewald |
1370 | * VdW interaction: LJEwald |
1371 | * Geometry: Water4-Water4 |
1372 | * Calculate force/pot: Force |
1373 | */ |
1374 | void |
1375 | nb_kernel_ElecEw_VdwLJEw_GeomW4W4_F_sse4_1_single |
1376 | (t_nblist * gmx_restrict nlist, |
1377 | rvec * gmx_restrict xx, |
1378 | rvec * gmx_restrict ff, |
1379 | t_forcerec * gmx_restrict fr, |
1380 | t_mdatoms * gmx_restrict mdatoms, |
1381 | nb_kernel_data_t gmx_unused__attribute__ ((unused)) * gmx_restrict kernel_data, |
1382 | t_nrnb * gmx_restrict nrnb) |
1383 | { |
1384 | /* Suffixes 0,1,2,3 refer to particle indices for waters in the inner or outer loop, or |
1385 | * just 0 for non-waters. |
1386 | * Suffixes A,B,C,D refer to j loop unrolling done with SSE, e.g. for the four different |
1387 | * jnr indices corresponding to data put in the four positions in the SIMD register. |
1388 | */ |
1389 | int i_shift_offset,i_coord_offset,outeriter,inneriter; |
1390 | int j_index_start,j_index_end,jidx,nri,inr,ggid,iidx; |
1391 | int jnrA,jnrB,jnrC,jnrD; |
1392 | int jnrlistA,jnrlistB,jnrlistC,jnrlistD; |
1393 | int j_coord_offsetA,j_coord_offsetB,j_coord_offsetC,j_coord_offsetD; |
1394 | int *iinr,*jindex,*jjnr,*shiftidx,*gid; |
1395 | real rcutoff_scalar; |
1396 | real *shiftvec,*fshift,*x,*f; |
1397 | real *fjptrA,*fjptrB,*fjptrC,*fjptrD; |
1398 | real scratch[4*DIM3]; |
1399 | __m128 tx,ty,tz,fscal,rcutoff,rcutoff2,jidxall; |
1400 | int vdwioffset0; |
1401 | __m128 ix0,iy0,iz0,fix0,fiy0,fiz0,iq0,isai0; |
1402 | int vdwioffset1; |
1403 | __m128 ix1,iy1,iz1,fix1,fiy1,fiz1,iq1,isai1; |
1404 | int vdwioffset2; |
1405 | __m128 ix2,iy2,iz2,fix2,fiy2,fiz2,iq2,isai2; |
1406 | int vdwioffset3; |
1407 | __m128 ix3,iy3,iz3,fix3,fiy3,fiz3,iq3,isai3; |
1408 | int vdwjidx0A,vdwjidx0B,vdwjidx0C,vdwjidx0D; |
1409 | __m128 jx0,jy0,jz0,fjx0,fjy0,fjz0,jq0,isaj0; |
1410 | int vdwjidx1A,vdwjidx1B,vdwjidx1C,vdwjidx1D; |
1411 | __m128 jx1,jy1,jz1,fjx1,fjy1,fjz1,jq1,isaj1; |
1412 | int vdwjidx2A,vdwjidx2B,vdwjidx2C,vdwjidx2D; |
1413 | __m128 jx2,jy2,jz2,fjx2,fjy2,fjz2,jq2,isaj2; |
1414 | int vdwjidx3A,vdwjidx3B,vdwjidx3C,vdwjidx3D; |
1415 | __m128 jx3,jy3,jz3,fjx3,fjy3,fjz3,jq3,isaj3; |
1416 | __m128 dx00,dy00,dz00,rsq00,rinv00,rinvsq00,r00,qq00,c6_00,c12_00; |
1417 | __m128 dx11,dy11,dz11,rsq11,rinv11,rinvsq11,r11,qq11,c6_11,c12_11; |
1418 | __m128 dx12,dy12,dz12,rsq12,rinv12,rinvsq12,r12,qq12,c6_12,c12_12; |
1419 | __m128 dx13,dy13,dz13,rsq13,rinv13,rinvsq13,r13,qq13,c6_13,c12_13; |
1420 | __m128 dx21,dy21,dz21,rsq21,rinv21,rinvsq21,r21,qq21,c6_21,c12_21; |
1421 | __m128 dx22,dy22,dz22,rsq22,rinv22,rinvsq22,r22,qq22,c6_22,c12_22; |
1422 | __m128 dx23,dy23,dz23,rsq23,rinv23,rinvsq23,r23,qq23,c6_23,c12_23; |
1423 | __m128 dx31,dy31,dz31,rsq31,rinv31,rinvsq31,r31,qq31,c6_31,c12_31; |
1424 | __m128 dx32,dy32,dz32,rsq32,rinv32,rinvsq32,r32,qq32,c6_32,c12_32; |
1425 | __m128 dx33,dy33,dz33,rsq33,rinv33,rinvsq33,r33,qq33,c6_33,c12_33; |
1426 | __m128 velec,felec,velecsum,facel,crf,krf,krf2; |
1427 | real *charge; |
1428 | int nvdwtype; |
1429 | __m128 rinvsix,rvdw,vvdw,vvdw6,vvdw12,fvdw,fvdw6,fvdw12,vvdwsum,sh_vdw_invrcut6; |
1430 | int *vdwtype; |
1431 | real *vdwparam; |
1432 | __m128 one_sixth = _mm_set1_ps(1.0/6.0); |
1433 | __m128 one_twelfth = _mm_set1_ps(1.0/12.0); |
1434 | __m128 c6grid_00; |
1435 | __m128 c6grid_11; |
1436 | __m128 c6grid_12; |
1437 | __m128 c6grid_13; |
1438 | __m128 c6grid_21; |
1439 | __m128 c6grid_22; |
1440 | __m128 c6grid_23; |
1441 | __m128 c6grid_31; |
1442 | __m128 c6grid_32; |
1443 | __m128 c6grid_33; |
1444 | __m128 ewclj,ewclj2,ewclj6,ewcljrsq,poly,exponent,f6A,f6B,sh_lj_ewald; |
1445 | real *vdwgridparam; |
1446 | __m128 one_half = _mm_set1_ps(0.5); |
1447 | __m128 minus_one = _mm_set1_ps(-1.0); |
1448 | __m128i ewitab; |
1449 | __m128 ewtabscale,eweps,sh_ewald,ewrt,ewtabhalfspace,ewtabF,ewtabFn,ewtabD,ewtabV; |
1450 | real *ewtab; |
1451 | __m128 dummy_mask,cutoff_mask; |
1452 | __m128 signbit = _mm_castsi128_ps( _mm_set1_epi32(0x80000000) ); |
1453 | __m128 one = _mm_set1_ps(1.0); |
1454 | __m128 two = _mm_set1_ps(2.0); |
1455 | x = xx[0]; |
1456 | f = ff[0]; |
1457 | |
1458 | nri = nlist->nri; |
1459 | iinr = nlist->iinr; |
1460 | jindex = nlist->jindex; |
1461 | jjnr = nlist->jjnr; |
1462 | shiftidx = nlist->shift; |
1463 | gid = nlist->gid; |
1464 | shiftvec = fr->shift_vec[0]; |
1465 | fshift = fr->fshift[0]; |
1466 | facel = _mm_set1_ps(fr->epsfac); |
1467 | charge = mdatoms->chargeA; |
1468 | nvdwtype = fr->ntype; |
1469 | vdwparam = fr->nbfp; |
1470 | vdwtype = mdatoms->typeA; |
1471 | vdwgridparam = fr->ljpme_c6grid; |
1472 | sh_lj_ewald = _mm_set1_ps(fr->ic->sh_lj_ewald); |
1473 | ewclj = _mm_set1_ps(fr->ewaldcoeff_lj); |
1474 | ewclj2 = _mm_mul_ps(minus_one,_mm_mul_ps(ewclj,ewclj)); |
1475 | |
1476 | sh_ewald = _mm_set1_ps(fr->ic->sh_ewald); |
1477 | ewtab = fr->ic->tabq_coul_F; |
1478 | ewtabscale = _mm_set1_ps(fr->ic->tabq_scale); |
1479 | ewtabhalfspace = _mm_set1_ps(0.5/fr->ic->tabq_scale); |
1480 | |
1481 | /* Setup water-specific parameters */ |
1482 | inr = nlist->iinr[0]; |
1483 | iq1 = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+1])); |
1484 | iq2 = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+2])); |
1485 | iq3 = _mm_mul_ps(facel,_mm_set1_ps(charge[inr+3])); |
1486 | vdwioffset0 = 2*nvdwtype*vdwtype[inr+0]; |
1487 | |
1488 | jq1 = _mm_set1_ps(charge[inr+1]); |
1489 | jq2 = _mm_set1_ps(charge[inr+2]); |
1490 | jq3 = _mm_set1_ps(charge[inr+3]); |
1491 | vdwjidx0A = 2*vdwtype[inr+0]; |
1492 | c6_00 = _mm_set1_ps(vdwparam[vdwioffset0+vdwjidx0A]); |
1493 | c12_00 = _mm_set1_ps(vdwparam[vdwioffset0+vdwjidx0A+1]); |
1494 | c6grid_00 = _mm_set1_ps(vdwgridparam[vdwioffset0+vdwjidx0A]); |
1495 | qq11 = _mm_mul_ps(iq1,jq1); |
1496 | qq12 = _mm_mul_ps(iq1,jq2); |
1497 | qq13 = _mm_mul_ps(iq1,jq3); |
1498 | qq21 = _mm_mul_ps(iq2,jq1); |
1499 | qq22 = _mm_mul_ps(iq2,jq2); |
1500 | qq23 = _mm_mul_ps(iq2,jq3); |
1501 | qq31 = _mm_mul_ps(iq3,jq1); |
1502 | qq32 = _mm_mul_ps(iq3,jq2); |
1503 | qq33 = _mm_mul_ps(iq3,jq3); |
1504 | |
1505 | /* Avoid stupid compiler warnings */ |
1506 | jnrA = jnrB = jnrC = jnrD = 0; |
1507 | j_coord_offsetA = 0; |
1508 | j_coord_offsetB = 0; |
1509 | j_coord_offsetC = 0; |
1510 | j_coord_offsetD = 0; |
1511 | |
1512 | outeriter = 0; |
1513 | inneriter = 0; |
1514 | |
1515 | for(iidx=0;iidx<4*DIM3;iidx++) |
1516 | { |
1517 | scratch[iidx] = 0.0; |
1518 | } |
1519 | |
1520 | /* Start outer loop over neighborlists */ |
1521 | for(iidx=0; iidx<nri; iidx++) |
1522 | { |
1523 | /* Load shift vector for this list */ |
1524 | i_shift_offset = DIM3*shiftidx[iidx]; |
1525 | |
1526 | /* Load limits for loop over neighbors */ |
1527 | j_index_start = jindex[iidx]; |
1528 | j_index_end = jindex[iidx+1]; |
1529 | |
1530 | /* Get outer coordinate index */ |
1531 | inr = iinr[iidx]; |
1532 | i_coord_offset = DIM3*inr; |
1533 | |
1534 | /* Load i particle coords and add shift vector */ |
1535 | gmx_mm_load_shift_and_4rvec_broadcast_ps(shiftvec+i_shift_offset,x+i_coord_offset, |
1536 | &ix0,&iy0,&iz0,&ix1,&iy1,&iz1,&ix2,&iy2,&iz2,&ix3,&iy3,&iz3); |
1537 | |
1538 | fix0 = _mm_setzero_ps(); |
1539 | fiy0 = _mm_setzero_ps(); |
1540 | fiz0 = _mm_setzero_ps(); |
1541 | fix1 = _mm_setzero_ps(); |
1542 | fiy1 = _mm_setzero_ps(); |
1543 | fiz1 = _mm_setzero_ps(); |
1544 | fix2 = _mm_setzero_ps(); |
1545 | fiy2 = _mm_setzero_ps(); |
1546 | fiz2 = _mm_setzero_ps(); |
1547 | fix3 = _mm_setzero_ps(); |
1548 | fiy3 = _mm_setzero_ps(); |
1549 | fiz3 = _mm_setzero_ps(); |
1550 | |
1551 | /* Start inner kernel loop */ |
1552 | for(jidx=j_index_start; jidx<j_index_end && jjnr[jidx+3]>=0; jidx+=4) |
1553 | { |
1554 | |
1555 | /* Get j neighbor index, and coordinate index */ |
1556 | jnrA = jjnr[jidx]; |
1557 | jnrB = jjnr[jidx+1]; |
1558 | jnrC = jjnr[jidx+2]; |
1559 | jnrD = jjnr[jidx+3]; |
1560 | j_coord_offsetA = DIM3*jnrA; |
1561 | j_coord_offsetB = DIM3*jnrB; |
1562 | j_coord_offsetC = DIM3*jnrC; |
1563 | j_coord_offsetD = DIM3*jnrD; |
1564 | |
1565 | /* load j atom coordinates */ |
1566 | gmx_mm_load_4rvec_4ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB, |
1567 | x+j_coord_offsetC,x+j_coord_offsetD, |
1568 | &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2, |
1569 | &jy2,&jz2,&jx3,&jy3,&jz3); |
1570 | |
1571 | /* Calculate displacement vector */ |
1572 | dx00 = _mm_sub_ps(ix0,jx0); |
1573 | dy00 = _mm_sub_ps(iy0,jy0); |
1574 | dz00 = _mm_sub_ps(iz0,jz0); |
1575 | dx11 = _mm_sub_ps(ix1,jx1); |
1576 | dy11 = _mm_sub_ps(iy1,jy1); |
1577 | dz11 = _mm_sub_ps(iz1,jz1); |
1578 | dx12 = _mm_sub_ps(ix1,jx2); |
1579 | dy12 = _mm_sub_ps(iy1,jy2); |
1580 | dz12 = _mm_sub_ps(iz1,jz2); |
1581 | dx13 = _mm_sub_ps(ix1,jx3); |
1582 | dy13 = _mm_sub_ps(iy1,jy3); |
1583 | dz13 = _mm_sub_ps(iz1,jz3); |
1584 | dx21 = _mm_sub_ps(ix2,jx1); |
1585 | dy21 = _mm_sub_ps(iy2,jy1); |
1586 | dz21 = _mm_sub_ps(iz2,jz1); |
1587 | dx22 = _mm_sub_ps(ix2,jx2); |
1588 | dy22 = _mm_sub_ps(iy2,jy2); |
1589 | dz22 = _mm_sub_ps(iz2,jz2); |
1590 | dx23 = _mm_sub_ps(ix2,jx3); |
1591 | dy23 = _mm_sub_ps(iy2,jy3); |
1592 | dz23 = _mm_sub_ps(iz2,jz3); |
1593 | dx31 = _mm_sub_ps(ix3,jx1); |
1594 | dy31 = _mm_sub_ps(iy3,jy1); |
1595 | dz31 = _mm_sub_ps(iz3,jz1); |
1596 | dx32 = _mm_sub_ps(ix3,jx2); |
1597 | dy32 = _mm_sub_ps(iy3,jy2); |
1598 | dz32 = _mm_sub_ps(iz3,jz2); |
1599 | dx33 = _mm_sub_ps(ix3,jx3); |
1600 | dy33 = _mm_sub_ps(iy3,jy3); |
1601 | dz33 = _mm_sub_ps(iz3,jz3); |
1602 | |
1603 | /* Calculate squared distance and things based on it */ |
1604 | rsq00 = gmx_mm_calc_rsq_ps(dx00,dy00,dz00); |
1605 | rsq11 = gmx_mm_calc_rsq_ps(dx11,dy11,dz11); |
1606 | rsq12 = gmx_mm_calc_rsq_ps(dx12,dy12,dz12); |
1607 | rsq13 = gmx_mm_calc_rsq_ps(dx13,dy13,dz13); |
1608 | rsq21 = gmx_mm_calc_rsq_ps(dx21,dy21,dz21); |
1609 | rsq22 = gmx_mm_calc_rsq_ps(dx22,dy22,dz22); |
1610 | rsq23 = gmx_mm_calc_rsq_ps(dx23,dy23,dz23); |
1611 | rsq31 = gmx_mm_calc_rsq_ps(dx31,dy31,dz31); |
1612 | rsq32 = gmx_mm_calc_rsq_ps(dx32,dy32,dz32); |
1613 | rsq33 = gmx_mm_calc_rsq_ps(dx33,dy33,dz33); |
1614 | |
1615 | rinv00 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq00); |
1616 | rinv11 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq11); |
1617 | rinv12 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq12); |
1618 | rinv13 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq13); |
1619 | rinv21 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq21); |
1620 | rinv22 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq22); |
1621 | rinv23 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq23); |
1622 | rinv31 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq31); |
1623 | rinv32 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq32); |
1624 | rinv33 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq33); |
1625 | |
1626 | rinvsq00 = _mm_mul_ps(rinv00,rinv00); |
1627 | rinvsq11 = _mm_mul_ps(rinv11,rinv11); |
1628 | rinvsq12 = _mm_mul_ps(rinv12,rinv12); |
1629 | rinvsq13 = _mm_mul_ps(rinv13,rinv13); |
1630 | rinvsq21 = _mm_mul_ps(rinv21,rinv21); |
1631 | rinvsq22 = _mm_mul_ps(rinv22,rinv22); |
1632 | rinvsq23 = _mm_mul_ps(rinv23,rinv23); |
1633 | rinvsq31 = _mm_mul_ps(rinv31,rinv31); |
1634 | rinvsq32 = _mm_mul_ps(rinv32,rinv32); |
1635 | rinvsq33 = _mm_mul_ps(rinv33,rinv33); |
1636 | |
1637 | fjx0 = _mm_setzero_ps(); |
1638 | fjy0 = _mm_setzero_ps(); |
1639 | fjz0 = _mm_setzero_ps(); |
1640 | fjx1 = _mm_setzero_ps(); |
1641 | fjy1 = _mm_setzero_ps(); |
1642 | fjz1 = _mm_setzero_ps(); |
1643 | fjx2 = _mm_setzero_ps(); |
1644 | fjy2 = _mm_setzero_ps(); |
1645 | fjz2 = _mm_setzero_ps(); |
1646 | fjx3 = _mm_setzero_ps(); |
1647 | fjy3 = _mm_setzero_ps(); |
1648 | fjz3 = _mm_setzero_ps(); |
1649 | |
1650 | /************************** |
1651 | * CALCULATE INTERACTIONS * |
1652 | **************************/ |
1653 | |
1654 | r00 = _mm_mul_ps(rsq00,rinv00); |
1655 | |
1656 | /* Analytical LJ-PME */ |
1657 | rinvsix = _mm_mul_ps(_mm_mul_ps(rinvsq00,rinvsq00),rinvsq00); |
1658 | ewcljrsq = _mm_mul_ps(ewclj2,rsq00); |
1659 | ewclj6 = _mm_mul_ps(ewclj2,_mm_mul_ps(ewclj2,ewclj2)); |
1660 | exponent = gmx_simd_exp_rgmx_simd_exp_f(ewcljrsq); |
1661 | /* poly = exp(-(beta*r)^2) * (1 + (beta*r)^2 + (beta*r)^4 /2) */ |
1662 | poly = _mm_mul_ps(exponent,_mm_add_ps(_mm_sub_ps(one,ewcljrsq),_mm_mul_ps(_mm_mul_ps(ewcljrsq,ewcljrsq),one_half))); |
1663 | /* f6A = 6 * C6grid * (1 - poly) */ |
1664 | f6A = _mm_mul_ps(c6grid_00,_mm_sub_ps(one,poly)); |
1665 | /* f6B = C6grid * exponent * beta^6 */ |
1666 | f6B = _mm_mul_ps(_mm_mul_ps(c6grid_00,one_sixth),_mm_mul_ps(exponent,ewclj6)); |
1667 | /* fvdw = 12*C12/r13 - ((6*C6 - f6A)/r6 + f6B)/r */ |
1668 | fvdw = _mm_mul_ps(_mm_add_ps(_mm_mul_ps(_mm_sub_ps(_mm_mul_ps(c12_00,rinvsix),_mm_sub_ps(c6_00,f6A)),rinvsix),f6B),rinvsq00); |
1669 | |
1670 | fscal = fvdw; |
1671 | |
1672 | /* Calculate temporary vectorial force */ |
1673 | tx = _mm_mul_ps(fscal,dx00); |
1674 | ty = _mm_mul_ps(fscal,dy00); |
1675 | tz = _mm_mul_ps(fscal,dz00); |
1676 | |
1677 | /* Update vectorial force */ |
1678 | fix0 = _mm_add_ps(fix0,tx); |
1679 | fiy0 = _mm_add_ps(fiy0,ty); |
1680 | fiz0 = _mm_add_ps(fiz0,tz); |
1681 | |
1682 | fjx0 = _mm_add_ps(fjx0,tx); |
1683 | fjy0 = _mm_add_ps(fjy0,ty); |
1684 | fjz0 = _mm_add_ps(fjz0,tz); |
1685 | |
1686 | /************************** |
1687 | * CALCULATE INTERACTIONS * |
1688 | **************************/ |
1689 | |
1690 | r11 = _mm_mul_ps(rsq11,rinv11); |
1691 | |
1692 | /* EWALD ELECTROSTATICS */ |
1693 | |
1694 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1695 | ewrt = _mm_mul_ps(r11,ewtabscale); |
1696 | ewitab = _mm_cvttps_epi32(ewrt); |
1697 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
1698 | gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})), |
1699 | ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})), |
1700 | &ewtabF,&ewtabFn); |
1701 | felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn)); |
1702 | felec = _mm_mul_ps(_mm_mul_ps(qq11,rinv11),_mm_sub_ps(rinvsq11,felec)); |
1703 | |
1704 | fscal = felec; |
1705 | |
1706 | /* Calculate temporary vectorial force */ |
1707 | tx = _mm_mul_ps(fscal,dx11); |
1708 | ty = _mm_mul_ps(fscal,dy11); |
1709 | tz = _mm_mul_ps(fscal,dz11); |
1710 | |
1711 | /* Update vectorial force */ |
1712 | fix1 = _mm_add_ps(fix1,tx); |
1713 | fiy1 = _mm_add_ps(fiy1,ty); |
1714 | fiz1 = _mm_add_ps(fiz1,tz); |
1715 | |
1716 | fjx1 = _mm_add_ps(fjx1,tx); |
1717 | fjy1 = _mm_add_ps(fjy1,ty); |
1718 | fjz1 = _mm_add_ps(fjz1,tz); |
1719 | |
1720 | /************************** |
1721 | * CALCULATE INTERACTIONS * |
1722 | **************************/ |
1723 | |
1724 | r12 = _mm_mul_ps(rsq12,rinv12); |
1725 | |
1726 | /* EWALD ELECTROSTATICS */ |
1727 | |
1728 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1729 | ewrt = _mm_mul_ps(r12,ewtabscale); |
1730 | ewitab = _mm_cvttps_epi32(ewrt); |
1731 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
1732 | gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})), |
1733 | ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})), |
1734 | &ewtabF,&ewtabFn); |
1735 | felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn)); |
1736 | felec = _mm_mul_ps(_mm_mul_ps(qq12,rinv12),_mm_sub_ps(rinvsq12,felec)); |
1737 | |
1738 | fscal = felec; |
1739 | |
1740 | /* Calculate temporary vectorial force */ |
1741 | tx = _mm_mul_ps(fscal,dx12); |
1742 | ty = _mm_mul_ps(fscal,dy12); |
1743 | tz = _mm_mul_ps(fscal,dz12); |
1744 | |
1745 | /* Update vectorial force */ |
1746 | fix1 = _mm_add_ps(fix1,tx); |
1747 | fiy1 = _mm_add_ps(fiy1,ty); |
1748 | fiz1 = _mm_add_ps(fiz1,tz); |
1749 | |
1750 | fjx2 = _mm_add_ps(fjx2,tx); |
1751 | fjy2 = _mm_add_ps(fjy2,ty); |
1752 | fjz2 = _mm_add_ps(fjz2,tz); |
1753 | |
1754 | /************************** |
1755 | * CALCULATE INTERACTIONS * |
1756 | **************************/ |
1757 | |
1758 | r13 = _mm_mul_ps(rsq13,rinv13); |
1759 | |
1760 | /* EWALD ELECTROSTATICS */ |
1761 | |
1762 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1763 | ewrt = _mm_mul_ps(r13,ewtabscale); |
1764 | ewitab = _mm_cvttps_epi32(ewrt); |
1765 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
1766 | gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})), |
1767 | ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})), |
1768 | &ewtabF,&ewtabFn); |
1769 | felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn)); |
1770 | felec = _mm_mul_ps(_mm_mul_ps(qq13,rinv13),_mm_sub_ps(rinvsq13,felec)); |
1771 | |
1772 | fscal = felec; |
1773 | |
1774 | /* Calculate temporary vectorial force */ |
1775 | tx = _mm_mul_ps(fscal,dx13); |
1776 | ty = _mm_mul_ps(fscal,dy13); |
1777 | tz = _mm_mul_ps(fscal,dz13); |
1778 | |
1779 | /* Update vectorial force */ |
1780 | fix1 = _mm_add_ps(fix1,tx); |
1781 | fiy1 = _mm_add_ps(fiy1,ty); |
1782 | fiz1 = _mm_add_ps(fiz1,tz); |
1783 | |
1784 | fjx3 = _mm_add_ps(fjx3,tx); |
1785 | fjy3 = _mm_add_ps(fjy3,ty); |
1786 | fjz3 = _mm_add_ps(fjz3,tz); |
1787 | |
1788 | /************************** |
1789 | * CALCULATE INTERACTIONS * |
1790 | **************************/ |
1791 | |
1792 | r21 = _mm_mul_ps(rsq21,rinv21); |
1793 | |
1794 | /* EWALD ELECTROSTATICS */ |
1795 | |
1796 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1797 | ewrt = _mm_mul_ps(r21,ewtabscale); |
1798 | ewitab = _mm_cvttps_epi32(ewrt); |
1799 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
1800 | gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})), |
1801 | ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})), |
1802 | &ewtabF,&ewtabFn); |
1803 | felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn)); |
1804 | felec = _mm_mul_ps(_mm_mul_ps(qq21,rinv21),_mm_sub_ps(rinvsq21,felec)); |
1805 | |
1806 | fscal = felec; |
1807 | |
1808 | /* Calculate temporary vectorial force */ |
1809 | tx = _mm_mul_ps(fscal,dx21); |
1810 | ty = _mm_mul_ps(fscal,dy21); |
1811 | tz = _mm_mul_ps(fscal,dz21); |
1812 | |
1813 | /* Update vectorial force */ |
1814 | fix2 = _mm_add_ps(fix2,tx); |
1815 | fiy2 = _mm_add_ps(fiy2,ty); |
1816 | fiz2 = _mm_add_ps(fiz2,tz); |
1817 | |
1818 | fjx1 = _mm_add_ps(fjx1,tx); |
1819 | fjy1 = _mm_add_ps(fjy1,ty); |
1820 | fjz1 = _mm_add_ps(fjz1,tz); |
1821 | |
1822 | /************************** |
1823 | * CALCULATE INTERACTIONS * |
1824 | **************************/ |
1825 | |
1826 | r22 = _mm_mul_ps(rsq22,rinv22); |
1827 | |
1828 | /* EWALD ELECTROSTATICS */ |
1829 | |
1830 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1831 | ewrt = _mm_mul_ps(r22,ewtabscale); |
1832 | ewitab = _mm_cvttps_epi32(ewrt); |
1833 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
1834 | gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})), |
1835 | ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})), |
1836 | &ewtabF,&ewtabFn); |
1837 | felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn)); |
1838 | felec = _mm_mul_ps(_mm_mul_ps(qq22,rinv22),_mm_sub_ps(rinvsq22,felec)); |
1839 | |
1840 | fscal = felec; |
1841 | |
1842 | /* Calculate temporary vectorial force */ |
1843 | tx = _mm_mul_ps(fscal,dx22); |
1844 | ty = _mm_mul_ps(fscal,dy22); |
1845 | tz = _mm_mul_ps(fscal,dz22); |
1846 | |
1847 | /* Update vectorial force */ |
1848 | fix2 = _mm_add_ps(fix2,tx); |
1849 | fiy2 = _mm_add_ps(fiy2,ty); |
1850 | fiz2 = _mm_add_ps(fiz2,tz); |
1851 | |
1852 | fjx2 = _mm_add_ps(fjx2,tx); |
1853 | fjy2 = _mm_add_ps(fjy2,ty); |
1854 | fjz2 = _mm_add_ps(fjz2,tz); |
1855 | |
1856 | /************************** |
1857 | * CALCULATE INTERACTIONS * |
1858 | **************************/ |
1859 | |
1860 | r23 = _mm_mul_ps(rsq23,rinv23); |
1861 | |
1862 | /* EWALD ELECTROSTATICS */ |
1863 | |
1864 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1865 | ewrt = _mm_mul_ps(r23,ewtabscale); |
1866 | ewitab = _mm_cvttps_epi32(ewrt); |
1867 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
1868 | gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})), |
1869 | ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})), |
1870 | &ewtabF,&ewtabFn); |
1871 | felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn)); |
1872 | felec = _mm_mul_ps(_mm_mul_ps(qq23,rinv23),_mm_sub_ps(rinvsq23,felec)); |
1873 | |
1874 | fscal = felec; |
1875 | |
1876 | /* Calculate temporary vectorial force */ |
1877 | tx = _mm_mul_ps(fscal,dx23); |
1878 | ty = _mm_mul_ps(fscal,dy23); |
1879 | tz = _mm_mul_ps(fscal,dz23); |
1880 | |
1881 | /* Update vectorial force */ |
1882 | fix2 = _mm_add_ps(fix2,tx); |
1883 | fiy2 = _mm_add_ps(fiy2,ty); |
1884 | fiz2 = _mm_add_ps(fiz2,tz); |
1885 | |
1886 | fjx3 = _mm_add_ps(fjx3,tx); |
1887 | fjy3 = _mm_add_ps(fjy3,ty); |
1888 | fjz3 = _mm_add_ps(fjz3,tz); |
1889 | |
1890 | /************************** |
1891 | * CALCULATE INTERACTIONS * |
1892 | **************************/ |
1893 | |
1894 | r31 = _mm_mul_ps(rsq31,rinv31); |
1895 | |
1896 | /* EWALD ELECTROSTATICS */ |
1897 | |
1898 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1899 | ewrt = _mm_mul_ps(r31,ewtabscale); |
1900 | ewitab = _mm_cvttps_epi32(ewrt); |
1901 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
1902 | gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})), |
1903 | ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})), |
1904 | &ewtabF,&ewtabFn); |
1905 | felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn)); |
1906 | felec = _mm_mul_ps(_mm_mul_ps(qq31,rinv31),_mm_sub_ps(rinvsq31,felec)); |
1907 | |
1908 | fscal = felec; |
1909 | |
1910 | /* Calculate temporary vectorial force */ |
1911 | tx = _mm_mul_ps(fscal,dx31); |
1912 | ty = _mm_mul_ps(fscal,dy31); |
1913 | tz = _mm_mul_ps(fscal,dz31); |
1914 | |
1915 | /* Update vectorial force */ |
1916 | fix3 = _mm_add_ps(fix3,tx); |
1917 | fiy3 = _mm_add_ps(fiy3,ty); |
1918 | fiz3 = _mm_add_ps(fiz3,tz); |
1919 | |
1920 | fjx1 = _mm_add_ps(fjx1,tx); |
1921 | fjy1 = _mm_add_ps(fjy1,ty); |
1922 | fjz1 = _mm_add_ps(fjz1,tz); |
1923 | |
1924 | /************************** |
1925 | * CALCULATE INTERACTIONS * |
1926 | **************************/ |
1927 | |
1928 | r32 = _mm_mul_ps(rsq32,rinv32); |
1929 | |
1930 | /* EWALD ELECTROSTATICS */ |
1931 | |
1932 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1933 | ewrt = _mm_mul_ps(r32,ewtabscale); |
1934 | ewitab = _mm_cvttps_epi32(ewrt); |
1935 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
1936 | gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})), |
1937 | ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})), |
1938 | &ewtabF,&ewtabFn); |
1939 | felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn)); |
1940 | felec = _mm_mul_ps(_mm_mul_ps(qq32,rinv32),_mm_sub_ps(rinvsq32,felec)); |
1941 | |
1942 | fscal = felec; |
1943 | |
1944 | /* Calculate temporary vectorial force */ |
1945 | tx = _mm_mul_ps(fscal,dx32); |
1946 | ty = _mm_mul_ps(fscal,dy32); |
1947 | tz = _mm_mul_ps(fscal,dz32); |
1948 | |
1949 | /* Update vectorial force */ |
1950 | fix3 = _mm_add_ps(fix3,tx); |
1951 | fiy3 = _mm_add_ps(fiy3,ty); |
1952 | fiz3 = _mm_add_ps(fiz3,tz); |
1953 | |
1954 | fjx2 = _mm_add_ps(fjx2,tx); |
1955 | fjy2 = _mm_add_ps(fjy2,ty); |
1956 | fjz2 = _mm_add_ps(fjz2,tz); |
1957 | |
1958 | /************************** |
1959 | * CALCULATE INTERACTIONS * |
1960 | **************************/ |
1961 | |
1962 | r33 = _mm_mul_ps(rsq33,rinv33); |
1963 | |
1964 | /* EWALD ELECTROSTATICS */ |
1965 | |
1966 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
1967 | ewrt = _mm_mul_ps(r33,ewtabscale); |
1968 | ewitab = _mm_cvttps_epi32(ewrt); |
1969 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
1970 | gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})), |
1971 | ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})), |
1972 | &ewtabF,&ewtabFn); |
1973 | felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn)); |
1974 | felec = _mm_mul_ps(_mm_mul_ps(qq33,rinv33),_mm_sub_ps(rinvsq33,felec)); |
1975 | |
1976 | fscal = felec; |
1977 | |
1978 | /* Calculate temporary vectorial force */ |
1979 | tx = _mm_mul_ps(fscal,dx33); |
1980 | ty = _mm_mul_ps(fscal,dy33); |
1981 | tz = _mm_mul_ps(fscal,dz33); |
1982 | |
1983 | /* Update vectorial force */ |
1984 | fix3 = _mm_add_ps(fix3,tx); |
1985 | fiy3 = _mm_add_ps(fiy3,ty); |
1986 | fiz3 = _mm_add_ps(fiz3,tz); |
1987 | |
1988 | fjx3 = _mm_add_ps(fjx3,tx); |
1989 | fjy3 = _mm_add_ps(fjy3,ty); |
1990 | fjz3 = _mm_add_ps(fjz3,tz); |
1991 | |
1992 | fjptrA = f+j_coord_offsetA; |
1993 | fjptrB = f+j_coord_offsetB; |
1994 | fjptrC = f+j_coord_offsetC; |
1995 | fjptrD = f+j_coord_offsetD; |
1996 | |
1997 | gmx_mm_decrement_4rvec_4ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD, |
1998 | fjx0,fjy0,fjz0,fjx1,fjy1,fjz1, |
1999 | fjx2,fjy2,fjz2,fjx3,fjy3,fjz3); |
2000 | |
2001 | /* Inner loop uses 373 flops */ |
2002 | } |
2003 | |
2004 | if(jidx<j_index_end) |
2005 | { |
2006 | |
2007 | /* Get j neighbor index, and coordinate index */ |
2008 | jnrlistA = jjnr[jidx]; |
2009 | jnrlistB = jjnr[jidx+1]; |
2010 | jnrlistC = jjnr[jidx+2]; |
2011 | jnrlistD = jjnr[jidx+3]; |
2012 | /* Sign of each element will be negative for non-real atoms. |
2013 | * This mask will be 0xFFFFFFFF for dummy entries and 0x0 for real ones, |
2014 | * so use it as val = _mm_andnot_ps(mask,val) to clear dummy entries. |
2015 | */ |
2016 | dummy_mask = gmx_mm_castsi128_ps_mm_castsi128_ps(_mm_cmplt_epi32(_mm_loadu_si128((const __m128i *)(jjnr+jidx)),_mm_setzero_si128())); |
2017 | jnrA = (jnrlistA>=0) ? jnrlistA : 0; |
2018 | jnrB = (jnrlistB>=0) ? jnrlistB : 0; |
2019 | jnrC = (jnrlistC>=0) ? jnrlistC : 0; |
2020 | jnrD = (jnrlistD>=0) ? jnrlistD : 0; |
2021 | j_coord_offsetA = DIM3*jnrA; |
2022 | j_coord_offsetB = DIM3*jnrB; |
2023 | j_coord_offsetC = DIM3*jnrC; |
2024 | j_coord_offsetD = DIM3*jnrD; |
2025 | |
2026 | /* load j atom coordinates */ |
2027 | gmx_mm_load_4rvec_4ptr_swizzle_ps(x+j_coord_offsetA,x+j_coord_offsetB, |
2028 | x+j_coord_offsetC,x+j_coord_offsetD, |
2029 | &jx0,&jy0,&jz0,&jx1,&jy1,&jz1,&jx2, |
2030 | &jy2,&jz2,&jx3,&jy3,&jz3); |
2031 | |
2032 | /* Calculate displacement vector */ |
2033 | dx00 = _mm_sub_ps(ix0,jx0); |
2034 | dy00 = _mm_sub_ps(iy0,jy0); |
2035 | dz00 = _mm_sub_ps(iz0,jz0); |
2036 | dx11 = _mm_sub_ps(ix1,jx1); |
2037 | dy11 = _mm_sub_ps(iy1,jy1); |
2038 | dz11 = _mm_sub_ps(iz1,jz1); |
2039 | dx12 = _mm_sub_ps(ix1,jx2); |
2040 | dy12 = _mm_sub_ps(iy1,jy2); |
2041 | dz12 = _mm_sub_ps(iz1,jz2); |
2042 | dx13 = _mm_sub_ps(ix1,jx3); |
2043 | dy13 = _mm_sub_ps(iy1,jy3); |
2044 | dz13 = _mm_sub_ps(iz1,jz3); |
2045 | dx21 = _mm_sub_ps(ix2,jx1); |
2046 | dy21 = _mm_sub_ps(iy2,jy1); |
2047 | dz21 = _mm_sub_ps(iz2,jz1); |
2048 | dx22 = _mm_sub_ps(ix2,jx2); |
2049 | dy22 = _mm_sub_ps(iy2,jy2); |
2050 | dz22 = _mm_sub_ps(iz2,jz2); |
2051 | dx23 = _mm_sub_ps(ix2,jx3); |
2052 | dy23 = _mm_sub_ps(iy2,jy3); |
2053 | dz23 = _mm_sub_ps(iz2,jz3); |
2054 | dx31 = _mm_sub_ps(ix3,jx1); |
2055 | dy31 = _mm_sub_ps(iy3,jy1); |
2056 | dz31 = _mm_sub_ps(iz3,jz1); |
2057 | dx32 = _mm_sub_ps(ix3,jx2); |
2058 | dy32 = _mm_sub_ps(iy3,jy2); |
2059 | dz32 = _mm_sub_ps(iz3,jz2); |
2060 | dx33 = _mm_sub_ps(ix3,jx3); |
2061 | dy33 = _mm_sub_ps(iy3,jy3); |
2062 | dz33 = _mm_sub_ps(iz3,jz3); |
2063 | |
2064 | /* Calculate squared distance and things based on it */ |
2065 | rsq00 = gmx_mm_calc_rsq_ps(dx00,dy00,dz00); |
2066 | rsq11 = gmx_mm_calc_rsq_ps(dx11,dy11,dz11); |
2067 | rsq12 = gmx_mm_calc_rsq_ps(dx12,dy12,dz12); |
2068 | rsq13 = gmx_mm_calc_rsq_ps(dx13,dy13,dz13); |
2069 | rsq21 = gmx_mm_calc_rsq_ps(dx21,dy21,dz21); |
2070 | rsq22 = gmx_mm_calc_rsq_ps(dx22,dy22,dz22); |
2071 | rsq23 = gmx_mm_calc_rsq_ps(dx23,dy23,dz23); |
2072 | rsq31 = gmx_mm_calc_rsq_ps(dx31,dy31,dz31); |
2073 | rsq32 = gmx_mm_calc_rsq_ps(dx32,dy32,dz32); |
2074 | rsq33 = gmx_mm_calc_rsq_ps(dx33,dy33,dz33); |
2075 | |
2076 | rinv00 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq00); |
2077 | rinv11 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq11); |
2078 | rinv12 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq12); |
2079 | rinv13 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq13); |
2080 | rinv21 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq21); |
2081 | rinv22 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq22); |
2082 | rinv23 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq23); |
2083 | rinv31 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq31); |
2084 | rinv32 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq32); |
2085 | rinv33 = gmx_mm_invsqrt_psgmx_simd_invsqrt_f(rsq33); |
2086 | |
2087 | rinvsq00 = _mm_mul_ps(rinv00,rinv00); |
2088 | rinvsq11 = _mm_mul_ps(rinv11,rinv11); |
2089 | rinvsq12 = _mm_mul_ps(rinv12,rinv12); |
2090 | rinvsq13 = _mm_mul_ps(rinv13,rinv13); |
2091 | rinvsq21 = _mm_mul_ps(rinv21,rinv21); |
2092 | rinvsq22 = _mm_mul_ps(rinv22,rinv22); |
2093 | rinvsq23 = _mm_mul_ps(rinv23,rinv23); |
2094 | rinvsq31 = _mm_mul_ps(rinv31,rinv31); |
2095 | rinvsq32 = _mm_mul_ps(rinv32,rinv32); |
2096 | rinvsq33 = _mm_mul_ps(rinv33,rinv33); |
2097 | |
2098 | fjx0 = _mm_setzero_ps(); |
2099 | fjy0 = _mm_setzero_ps(); |
2100 | fjz0 = _mm_setzero_ps(); |
2101 | fjx1 = _mm_setzero_ps(); |
2102 | fjy1 = _mm_setzero_ps(); |
2103 | fjz1 = _mm_setzero_ps(); |
2104 | fjx2 = _mm_setzero_ps(); |
2105 | fjy2 = _mm_setzero_ps(); |
2106 | fjz2 = _mm_setzero_ps(); |
2107 | fjx3 = _mm_setzero_ps(); |
2108 | fjy3 = _mm_setzero_ps(); |
2109 | fjz3 = _mm_setzero_ps(); |
2110 | |
2111 | /************************** |
2112 | * CALCULATE INTERACTIONS * |
2113 | **************************/ |
2114 | |
2115 | r00 = _mm_mul_ps(rsq00,rinv00); |
2116 | r00 = _mm_andnot_ps(dummy_mask,r00); |
2117 | |
2118 | /* Analytical LJ-PME */ |
2119 | rinvsix = _mm_mul_ps(_mm_mul_ps(rinvsq00,rinvsq00),rinvsq00); |
2120 | ewcljrsq = _mm_mul_ps(ewclj2,rsq00); |
2121 | ewclj6 = _mm_mul_ps(ewclj2,_mm_mul_ps(ewclj2,ewclj2)); |
2122 | exponent = gmx_simd_exp_rgmx_simd_exp_f(ewcljrsq); |
2123 | /* poly = exp(-(beta*r)^2) * (1 + (beta*r)^2 + (beta*r)^4 /2) */ |
2124 | poly = _mm_mul_ps(exponent,_mm_add_ps(_mm_sub_ps(one,ewcljrsq),_mm_mul_ps(_mm_mul_ps(ewcljrsq,ewcljrsq),one_half))); |
2125 | /* f6A = 6 * C6grid * (1 - poly) */ |
2126 | f6A = _mm_mul_ps(c6grid_00,_mm_sub_ps(one,poly)); |
2127 | /* f6B = C6grid * exponent * beta^6 */ |
2128 | f6B = _mm_mul_ps(_mm_mul_ps(c6grid_00,one_sixth),_mm_mul_ps(exponent,ewclj6)); |
2129 | /* fvdw = 12*C12/r13 - ((6*C6 - f6A)/r6 + f6B)/r */ |
2130 | fvdw = _mm_mul_ps(_mm_add_ps(_mm_mul_ps(_mm_sub_ps(_mm_mul_ps(c12_00,rinvsix),_mm_sub_ps(c6_00,f6A)),rinvsix),f6B),rinvsq00); |
2131 | |
2132 | fscal = fvdw; |
2133 | |
2134 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
2135 | |
2136 | /* Calculate temporary vectorial force */ |
2137 | tx = _mm_mul_ps(fscal,dx00); |
2138 | ty = _mm_mul_ps(fscal,dy00); |
2139 | tz = _mm_mul_ps(fscal,dz00); |
2140 | |
2141 | /* Update vectorial force */ |
2142 | fix0 = _mm_add_ps(fix0,tx); |
2143 | fiy0 = _mm_add_ps(fiy0,ty); |
2144 | fiz0 = _mm_add_ps(fiz0,tz); |
2145 | |
2146 | fjx0 = _mm_add_ps(fjx0,tx); |
2147 | fjy0 = _mm_add_ps(fjy0,ty); |
2148 | fjz0 = _mm_add_ps(fjz0,tz); |
2149 | |
2150 | /************************** |
2151 | * CALCULATE INTERACTIONS * |
2152 | **************************/ |
2153 | |
2154 | r11 = _mm_mul_ps(rsq11,rinv11); |
2155 | r11 = _mm_andnot_ps(dummy_mask,r11); |
2156 | |
2157 | /* EWALD ELECTROSTATICS */ |
2158 | |
2159 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
2160 | ewrt = _mm_mul_ps(r11,ewtabscale); |
2161 | ewitab = _mm_cvttps_epi32(ewrt); |
2162 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
2163 | gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})), |
2164 | ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})), |
2165 | &ewtabF,&ewtabFn); |
2166 | felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn)); |
2167 | felec = _mm_mul_ps(_mm_mul_ps(qq11,rinv11),_mm_sub_ps(rinvsq11,felec)); |
2168 | |
2169 | fscal = felec; |
2170 | |
2171 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
2172 | |
2173 | /* Calculate temporary vectorial force */ |
2174 | tx = _mm_mul_ps(fscal,dx11); |
2175 | ty = _mm_mul_ps(fscal,dy11); |
2176 | tz = _mm_mul_ps(fscal,dz11); |
2177 | |
2178 | /* Update vectorial force */ |
2179 | fix1 = _mm_add_ps(fix1,tx); |
2180 | fiy1 = _mm_add_ps(fiy1,ty); |
2181 | fiz1 = _mm_add_ps(fiz1,tz); |
2182 | |
2183 | fjx1 = _mm_add_ps(fjx1,tx); |
2184 | fjy1 = _mm_add_ps(fjy1,ty); |
2185 | fjz1 = _mm_add_ps(fjz1,tz); |
2186 | |
2187 | /************************** |
2188 | * CALCULATE INTERACTIONS * |
2189 | **************************/ |
2190 | |
2191 | r12 = _mm_mul_ps(rsq12,rinv12); |
2192 | r12 = _mm_andnot_ps(dummy_mask,r12); |
2193 | |
2194 | /* EWALD ELECTROSTATICS */ |
2195 | |
2196 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
2197 | ewrt = _mm_mul_ps(r12,ewtabscale); |
2198 | ewitab = _mm_cvttps_epi32(ewrt); |
2199 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
2200 | gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})), |
2201 | ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})), |
2202 | &ewtabF,&ewtabFn); |
2203 | felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn)); |
2204 | felec = _mm_mul_ps(_mm_mul_ps(qq12,rinv12),_mm_sub_ps(rinvsq12,felec)); |
2205 | |
2206 | fscal = felec; |
2207 | |
2208 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
2209 | |
2210 | /* Calculate temporary vectorial force */ |
2211 | tx = _mm_mul_ps(fscal,dx12); |
2212 | ty = _mm_mul_ps(fscal,dy12); |
2213 | tz = _mm_mul_ps(fscal,dz12); |
2214 | |
2215 | /* Update vectorial force */ |
2216 | fix1 = _mm_add_ps(fix1,tx); |
2217 | fiy1 = _mm_add_ps(fiy1,ty); |
2218 | fiz1 = _mm_add_ps(fiz1,tz); |
2219 | |
2220 | fjx2 = _mm_add_ps(fjx2,tx); |
2221 | fjy2 = _mm_add_ps(fjy2,ty); |
2222 | fjz2 = _mm_add_ps(fjz2,tz); |
2223 | |
2224 | /************************** |
2225 | * CALCULATE INTERACTIONS * |
2226 | **************************/ |
2227 | |
2228 | r13 = _mm_mul_ps(rsq13,rinv13); |
2229 | r13 = _mm_andnot_ps(dummy_mask,r13); |
2230 | |
2231 | /* EWALD ELECTROSTATICS */ |
2232 | |
2233 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
2234 | ewrt = _mm_mul_ps(r13,ewtabscale); |
2235 | ewitab = _mm_cvttps_epi32(ewrt); |
2236 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
2237 | gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})), |
2238 | ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})), |
2239 | &ewtabF,&ewtabFn); |
2240 | felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn)); |
2241 | felec = _mm_mul_ps(_mm_mul_ps(qq13,rinv13),_mm_sub_ps(rinvsq13,felec)); |
2242 | |
2243 | fscal = felec; |
2244 | |
2245 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
2246 | |
2247 | /* Calculate temporary vectorial force */ |
2248 | tx = _mm_mul_ps(fscal,dx13); |
2249 | ty = _mm_mul_ps(fscal,dy13); |
2250 | tz = _mm_mul_ps(fscal,dz13); |
2251 | |
2252 | /* Update vectorial force */ |
2253 | fix1 = _mm_add_ps(fix1,tx); |
2254 | fiy1 = _mm_add_ps(fiy1,ty); |
2255 | fiz1 = _mm_add_ps(fiz1,tz); |
2256 | |
2257 | fjx3 = _mm_add_ps(fjx3,tx); |
2258 | fjy3 = _mm_add_ps(fjy3,ty); |
2259 | fjz3 = _mm_add_ps(fjz3,tz); |
2260 | |
2261 | /************************** |
2262 | * CALCULATE INTERACTIONS * |
2263 | **************************/ |
2264 | |
2265 | r21 = _mm_mul_ps(rsq21,rinv21); |
2266 | r21 = _mm_andnot_ps(dummy_mask,r21); |
2267 | |
2268 | /* EWALD ELECTROSTATICS */ |
2269 | |
2270 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
2271 | ewrt = _mm_mul_ps(r21,ewtabscale); |
2272 | ewitab = _mm_cvttps_epi32(ewrt); |
2273 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
2274 | gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})), |
2275 | ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})), |
2276 | &ewtabF,&ewtabFn); |
2277 | felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn)); |
2278 | felec = _mm_mul_ps(_mm_mul_ps(qq21,rinv21),_mm_sub_ps(rinvsq21,felec)); |
2279 | |
2280 | fscal = felec; |
2281 | |
2282 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
2283 | |
2284 | /* Calculate temporary vectorial force */ |
2285 | tx = _mm_mul_ps(fscal,dx21); |
2286 | ty = _mm_mul_ps(fscal,dy21); |
2287 | tz = _mm_mul_ps(fscal,dz21); |
2288 | |
2289 | /* Update vectorial force */ |
2290 | fix2 = _mm_add_ps(fix2,tx); |
2291 | fiy2 = _mm_add_ps(fiy2,ty); |
2292 | fiz2 = _mm_add_ps(fiz2,tz); |
2293 | |
2294 | fjx1 = _mm_add_ps(fjx1,tx); |
2295 | fjy1 = _mm_add_ps(fjy1,ty); |
2296 | fjz1 = _mm_add_ps(fjz1,tz); |
2297 | |
2298 | /************************** |
2299 | * CALCULATE INTERACTIONS * |
2300 | **************************/ |
2301 | |
2302 | r22 = _mm_mul_ps(rsq22,rinv22); |
2303 | r22 = _mm_andnot_ps(dummy_mask,r22); |
2304 | |
2305 | /* EWALD ELECTROSTATICS */ |
2306 | |
2307 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
2308 | ewrt = _mm_mul_ps(r22,ewtabscale); |
2309 | ewitab = _mm_cvttps_epi32(ewrt); |
2310 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
2311 | gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})), |
2312 | ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})), |
2313 | &ewtabF,&ewtabFn); |
2314 | felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn)); |
2315 | felec = _mm_mul_ps(_mm_mul_ps(qq22,rinv22),_mm_sub_ps(rinvsq22,felec)); |
2316 | |
2317 | fscal = felec; |
2318 | |
2319 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
2320 | |
2321 | /* Calculate temporary vectorial force */ |
2322 | tx = _mm_mul_ps(fscal,dx22); |
2323 | ty = _mm_mul_ps(fscal,dy22); |
2324 | tz = _mm_mul_ps(fscal,dz22); |
2325 | |
2326 | /* Update vectorial force */ |
2327 | fix2 = _mm_add_ps(fix2,tx); |
2328 | fiy2 = _mm_add_ps(fiy2,ty); |
2329 | fiz2 = _mm_add_ps(fiz2,tz); |
2330 | |
2331 | fjx2 = _mm_add_ps(fjx2,tx); |
2332 | fjy2 = _mm_add_ps(fjy2,ty); |
2333 | fjz2 = _mm_add_ps(fjz2,tz); |
2334 | |
2335 | /************************** |
2336 | * CALCULATE INTERACTIONS * |
2337 | **************************/ |
2338 | |
2339 | r23 = _mm_mul_ps(rsq23,rinv23); |
2340 | r23 = _mm_andnot_ps(dummy_mask,r23); |
2341 | |
2342 | /* EWALD ELECTROSTATICS */ |
2343 | |
2344 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
2345 | ewrt = _mm_mul_ps(r23,ewtabscale); |
2346 | ewitab = _mm_cvttps_epi32(ewrt); |
2347 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
2348 | gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})), |
2349 | ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})), |
2350 | &ewtabF,&ewtabFn); |
2351 | felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn)); |
2352 | felec = _mm_mul_ps(_mm_mul_ps(qq23,rinv23),_mm_sub_ps(rinvsq23,felec)); |
2353 | |
2354 | fscal = felec; |
2355 | |
2356 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
2357 | |
2358 | /* Calculate temporary vectorial force */ |
2359 | tx = _mm_mul_ps(fscal,dx23); |
2360 | ty = _mm_mul_ps(fscal,dy23); |
2361 | tz = _mm_mul_ps(fscal,dz23); |
2362 | |
2363 | /* Update vectorial force */ |
2364 | fix2 = _mm_add_ps(fix2,tx); |
2365 | fiy2 = _mm_add_ps(fiy2,ty); |
2366 | fiz2 = _mm_add_ps(fiz2,tz); |
2367 | |
2368 | fjx3 = _mm_add_ps(fjx3,tx); |
2369 | fjy3 = _mm_add_ps(fjy3,ty); |
2370 | fjz3 = _mm_add_ps(fjz3,tz); |
2371 | |
2372 | /************************** |
2373 | * CALCULATE INTERACTIONS * |
2374 | **************************/ |
2375 | |
2376 | r31 = _mm_mul_ps(rsq31,rinv31); |
2377 | r31 = _mm_andnot_ps(dummy_mask,r31); |
2378 | |
2379 | /* EWALD ELECTROSTATICS */ |
2380 | |
2381 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
2382 | ewrt = _mm_mul_ps(r31,ewtabscale); |
2383 | ewitab = _mm_cvttps_epi32(ewrt); |
2384 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
2385 | gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})), |
2386 | ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})), |
2387 | &ewtabF,&ewtabFn); |
2388 | felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn)); |
2389 | felec = _mm_mul_ps(_mm_mul_ps(qq31,rinv31),_mm_sub_ps(rinvsq31,felec)); |
2390 | |
2391 | fscal = felec; |
2392 | |
2393 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
2394 | |
2395 | /* Calculate temporary vectorial force */ |
2396 | tx = _mm_mul_ps(fscal,dx31); |
2397 | ty = _mm_mul_ps(fscal,dy31); |
2398 | tz = _mm_mul_ps(fscal,dz31); |
2399 | |
2400 | /* Update vectorial force */ |
2401 | fix3 = _mm_add_ps(fix3,tx); |
2402 | fiy3 = _mm_add_ps(fiy3,ty); |
2403 | fiz3 = _mm_add_ps(fiz3,tz); |
2404 | |
2405 | fjx1 = _mm_add_ps(fjx1,tx); |
2406 | fjy1 = _mm_add_ps(fjy1,ty); |
2407 | fjz1 = _mm_add_ps(fjz1,tz); |
2408 | |
2409 | /************************** |
2410 | * CALCULATE INTERACTIONS * |
2411 | **************************/ |
2412 | |
2413 | r32 = _mm_mul_ps(rsq32,rinv32); |
2414 | r32 = _mm_andnot_ps(dummy_mask,r32); |
2415 | |
2416 | /* EWALD ELECTROSTATICS */ |
2417 | |
2418 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
2419 | ewrt = _mm_mul_ps(r32,ewtabscale); |
2420 | ewitab = _mm_cvttps_epi32(ewrt); |
2421 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
2422 | gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})), |
2423 | ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})), |
2424 | &ewtabF,&ewtabFn); |
2425 | felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn)); |
2426 | felec = _mm_mul_ps(_mm_mul_ps(qq32,rinv32),_mm_sub_ps(rinvsq32,felec)); |
2427 | |
2428 | fscal = felec; |
2429 | |
2430 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
2431 | |
2432 | /* Calculate temporary vectorial force */ |
2433 | tx = _mm_mul_ps(fscal,dx32); |
2434 | ty = _mm_mul_ps(fscal,dy32); |
2435 | tz = _mm_mul_ps(fscal,dz32); |
2436 | |
2437 | /* Update vectorial force */ |
2438 | fix3 = _mm_add_ps(fix3,tx); |
2439 | fiy3 = _mm_add_ps(fiy3,ty); |
2440 | fiz3 = _mm_add_ps(fiz3,tz); |
2441 | |
2442 | fjx2 = _mm_add_ps(fjx2,tx); |
2443 | fjy2 = _mm_add_ps(fjy2,ty); |
2444 | fjz2 = _mm_add_ps(fjz2,tz); |
2445 | |
2446 | /************************** |
2447 | * CALCULATE INTERACTIONS * |
2448 | **************************/ |
2449 | |
2450 | r33 = _mm_mul_ps(rsq33,rinv33); |
2451 | r33 = _mm_andnot_ps(dummy_mask,r33); |
2452 | |
2453 | /* EWALD ELECTROSTATICS */ |
2454 | |
2455 | /* Calculate Ewald table index by multiplying r with scale and truncate to integer */ |
2456 | ewrt = _mm_mul_ps(r33,ewtabscale); |
2457 | ewitab = _mm_cvttps_epi32(ewrt); |
2458 | eweps = _mm_sub_ps(ewrt,_mm_round_ps(ewrt, _MM_FROUND_FLOOR)__extension__ ({ __m128 __X = (ewrt); (__m128) __builtin_ia32_roundps ((__v4sf)__X, ((0x00 | 0x01))); })); |
2459 | gmx_mm_load_4pair_swizzle_ps(ewtab + gmx_mm_extract_epi32(ewitab,0)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(0) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,1)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(1) & 3];})), |
2460 | ewtab + gmx_mm_extract_epi32(ewitab,2)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(2) & 3];})),ewtab + gmx_mm_extract_epi32(ewitab,3)(__extension__ ({ __v4si __a = (__v4si)(ewitab); __a[(3) & 3];})), |
2461 | &ewtabF,&ewtabFn); |
2462 | felec = _mm_add_ps(_mm_mul_ps( _mm_sub_ps(one,eweps),ewtabF),_mm_mul_ps(eweps,ewtabFn)); |
2463 | felec = _mm_mul_ps(_mm_mul_ps(qq33,rinv33),_mm_sub_ps(rinvsq33,felec)); |
2464 | |
2465 | fscal = felec; |
2466 | |
2467 | fscal = _mm_andnot_ps(dummy_mask,fscal); |
2468 | |
2469 | /* Calculate temporary vectorial force */ |
2470 | tx = _mm_mul_ps(fscal,dx33); |
2471 | ty = _mm_mul_ps(fscal,dy33); |
2472 | tz = _mm_mul_ps(fscal,dz33); |
2473 | |
2474 | /* Update vectorial force */ |
2475 | fix3 = _mm_add_ps(fix3,tx); |
2476 | fiy3 = _mm_add_ps(fiy3,ty); |
2477 | fiz3 = _mm_add_ps(fiz3,tz); |
2478 | |
2479 | fjx3 = _mm_add_ps(fjx3,tx); |
2480 | fjy3 = _mm_add_ps(fjy3,ty); |
2481 | fjz3 = _mm_add_ps(fjz3,tz); |
2482 | |
2483 | fjptrA = (jnrlistA>=0) ? f+j_coord_offsetA : scratch; |
2484 | fjptrB = (jnrlistB>=0) ? f+j_coord_offsetB : scratch; |
2485 | fjptrC = (jnrlistC>=0) ? f+j_coord_offsetC : scratch; |
2486 | fjptrD = (jnrlistD>=0) ? f+j_coord_offsetD : scratch; |
2487 | |
2488 | gmx_mm_decrement_4rvec_4ptr_swizzle_ps(fjptrA,fjptrB,fjptrC,fjptrD, |
2489 | fjx0,fjy0,fjz0,fjx1,fjy1,fjz1, |
2490 | fjx2,fjy2,fjz2,fjx3,fjy3,fjz3); |
2491 | |
2492 | /* Inner loop uses 383 flops */ |
2493 | } |
2494 | |
2495 | /* End of innermost loop */ |
2496 | |
2497 | gmx_mm_update_iforce_4atom_swizzle_ps(fix0,fiy0,fiz0,fix1,fiy1,fiz1,fix2,fiy2,fiz2,fix3,fiy3,fiz3, |
2498 | f+i_coord_offset,fshift+i_shift_offset); |
2499 | |
2500 | /* Increment number of inner iterations */ |
2501 | inneriter += j_index_end - j_index_start; |
2502 | |
2503 | /* Outer loop uses 24 flops */ |
2504 | } |
2505 | |
2506 | /* Increment number of outer iterations */ |
2507 | outeriter += nri; |
2508 | |
2509 | /* Update outer/inner flops */ |
2510 | |
2511 | inc_nrnb(nrnb,eNR_NBKERNEL_ELEC_VDW_W4W4_F,outeriter*24 + inneriter*383)(nrnb)->n[eNR_NBKERNEL_ELEC_VDW_W4W4_F] += outeriter*24 + inneriter *383; |
2512 | } |