-
Notifications
You must be signed in to change notification settings - Fork 5
/
qws.cc
1074 lines (1000 loc) · 37.9 KB
/
qws.cc
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
//****************************************************************************************
//
// Copyright (c) 2015-2020, Yoshifumi Nakamura <[email protected]>
// Copyright (c) 2015-2020, Yuta Mukai <[email protected]>
// Copyright (c) 2018-2020, Ken-Ichi Ishikawa <[email protected]>
// Copyright (c) 2019-2020, Issaku Kanamori <[email protected]>
//
//
// All rights reserved.
//
// Redistribution and use in source and binary forms, with or without
// modification, are permitted provided that the following conditions are
// met:
//
// * Redistributions of source code must retain the above copyright
// notice, this list of conditions and the following disclaimer.
//
// * Redistributions in binary form must reproduce the above copyright
// notice, this list of conditions and the following disclaimer listed
// in this license in the documentation and/or other materials
// provided with the distribution.
//
// * Neither the name of the copyright holders nor the names of its
// contributors may be used to endorse or promote products derived from
// this software without specific prior written permission.
//
// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
// "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
// LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
// A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
// OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
// SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
// LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
// DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
// THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
// (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
// OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
//
//----------------------------------------------------------------------------------------
// ACKNOWLEDGMENT
//
// This software has been developed in a co-design working group for the lattice QCD
// supported by MEXT's programs for the Development and Improvement for the Next
// Generation Ultra High-Speed Computer System, under its Subsidies for Operating the
// Specific Advanced Large Research Facilities, and Priority Issue 9
// (Elucidation of the Fundamental Laws and Evolution of the Universe) to be tackled by
// using the Supercomputer Fugaku.
//
//****************************************************************************************
#include "qws.h"
#define _PFI1 2
#define _PFI2 1
#define _PFI3 1
#define _PFI4 2
#include "wilson_d.h"
#include "wilson_s.h"
#include "clover_d.h"
#include "clover_s.h"
#include "clover_def.h"
#include "prefetch.h"
#ifdef _OPENMP
#include <omp.h>
#endif
#define min(a,b) (a)>(b)?(b):(a)
#include "timing.h"
#include "eml_lib.h"
#ifdef __cplusplus
extern "C"{
#endif
__attribute__((aligned(64))) pglud_t glud;
__attribute__((aligned(64))) pglus_t glus;
__attribute__((aligned(64))) pclvd_t clvd;
__attribute__((aligned(64))) pclvs_t clvs;
//projscd_t *xfd_buff;
//projscd_t *xbd_buff;
projscd1_t *xfd_send, *xfd_recv;
projscd1_t *xbd_send, *xbd_recv;
projscd_t *yfd_send, *yfd_recv;
projscd_t *ybd_send, *ybd_recv;
projscd_t *zfd_send, *zfd_recv;
projscd_t *zbd_send, *zbd_recv;
projscd_t *tfd_send, *tfd_recv;
projscd_t *tbd_send, *tbd_recv;
projscs1_t *xfs_send, *xfs_recv;
projscs1_t *xbs_send, *xbs_recv;
projscs_t *yfs_send, *yfs_recv;
projscs_t *ybs_send, *ybs_recv;
projscs_t *zfs_send, *zfs_recv;
projscs_t *zbs_send, *zbs_recv;
projscs_t *tfs_send, *tfs_recv;
projscs_t *tbs_send, *tbs_recv;
scs_t *qs;
double kappa, kappa2, mkappa;
// nh=nx/2, nd=nx/2/8, nxs=nx/2/16
int nx, ny, nz, nt, nxh, nxd, nxs;
int vold;// = nxd * ny * nz * nt;
int vols;// = nxs * ny * nz * nt;
int npe[4];// number of proccesses
double fbc[4][2];// fermion boundary condiion [mu][0=fwd, 1=bwd]
int rank, size, px, py, pz, pt;// process coordinate
int pxf, pyf, pzf, ptf;
int pxb, pyb, pzb, ptb;
int domain_e, domain_o;
int thmax,volsize;
int iam;
#pragma omp threadprivate(iam)
int volse;
#pragma omp threadprivate(volse)
int *pce, *pco;//even-odd precondition notation
block_map_t* block_map;
int num_blocks;
FILE *para_outputfile=0;
#ifdef _POWER_API_
extern void power_api_init();
extern void power_api_finalize();
#endif
//---------------------------------------------------------------------------------------- qws init
void qws_init_(int* lx, int* ly, int* lz, int* lt,
int* npe_f, int* fbc_f, int* pce_f, int* pco_f, int* block_size){
int i;
pce = pce_f;
pco = pco_f;
for (i=0;i<4;i++){
npe[i]=npe_f[i];
}
nx=*lx; ny=*ly; nz=*lz; nt=*lt;
nxh=nx/2;
nxd=nx/2/VLEND;
nxs=nx/2/VLENS;
if (nxd == 0) {printf("ERROR at %s : nx/2/VLEND must be >=1\n", __func__); exit(1);}
if (nxs == 0) {printf("ERROR at %s : nx/2/VLENS must be >=1\n", __func__); exit(1);}
vold = nxd * ny * nz * nt;
vols = nxs * ny * nz * nt;
#ifdef _OPENMP
#pragma omp parallel
{
iam = omp_get_thread_num();
#pragma omp master
{
thmax = omp_get_num_threads();
volsize = vols / thmax;
}
#pragma omp barrier
volse = volsize * (iam+1);
}
#else
iam=0;
thmax=1;
volsize=vols;
volse=vols;
#endif
glud = (pglud_t)malloc( sizeof(glud_t) * vold*NDIM*NEO); /////////////////
clvd = (pclvd_t)malloc( sizeof(clvd_t) * vold*NEO);
void *tmp;
posix_memalign(&tmp, CLS, sizeof(glus_t) * vols*NDIM*NEO);
glus = (pglus_t)tmp;
posix_memalign(&tmp, CLS, sizeof(clvs_t) * vols*NEO);
clvs = (pclvs_t)tmp;
posix_memalign(&tmp, CLS, sizeof(scs_t)*vols);
qs = (scs_t*)tmp;
for(i=0; i<vols; i++){
for(int j=0; j<24; j++){
for(int v=0; v<VLENS; v++){
qs[i].ccs[j].v[v] = 0.0f;
}
}
}
#ifdef _MPI_
MPI_Comm_rank(MPI_COMM_WORLD, &rank);
MPI_Comm_size(MPI_COMM_WORLD, &size);
#else
rank = 0;
size = 1;
#endif
if(rank==0)printf("QWS git commit hash : %s\n",VERSION);
if (size != npe[0]*npe[1]*npe[2]*npe[3]) {
printf("MPI size /= npe[0]*npe[1]*npe[2]*npe[3]\n size=%d, npe[0]*npe[1]*npe[2]*npe[3]=%d\nStop.\n",size,npe[0]*npe[1]*npe[2]*npe[3]);
#ifdef _MPI_
MPI_Finalize();
#endif
}
#ifdef _CHECK_TIMING
#ifdef _TIMING_EACH_RANK
char filename[21];
sprintf(filename, "output_rank%06d.txt", rank);
para_outputfile = fopen(filename, "w");
#endif
#endif
int map_id=-1;
int rank_coords[4];
int rank_size[4];
int rankf[4];
int rankb[4];
map_id=xbound_get_rankmap(rank, rank_coords, rank_size, rankf, rankb);
if (size == 1){
printf("single %d\n", rank);
fflush(stdout);
px=0; py=0; pz=0; pt=0;
domain_e = ( py + pz + pt)%2;
domain_o = 1 - domain_e;
} else {
if(map_id>=0){ //
if(rank==0){
printf("rankmap is found: map_id=%d\n", map_id);
fflush(stdout);
}
if(npe[0] != rank_size[0] ||npe[1] != rank_size[1] ||
npe[2] != rank_size[2] ||npe[3] != rank_size[3] ){
if(rank==0){
printf(" WARNING!!! new rank size: %d %d %d %d\n", rank_size[0], rank_size[1], rank_size[2], rank_size[3]);
fflush(stdout);
}
}
for(int i=0; i<4; i++){
npe[i]=rank_size[i];
}
px=rank_coords[0];
py=rank_coords[1];
pz=rank_coords[2];
pt=rank_coords[3];
pxf=rankf[0];
pyf=rankf[1];
pzf=rankf[2];
ptf=rankf[3];
pxb=rankb[0];
pyb=rankb[1];
pzb=rankb[2];
ptb=rankb[3];
} else { // no rankmpa is found
px = rank;
pt = px / (npe[0]*npe[1]*npe[2]);
px-= npe[0]*npe[1]*npe[2]*pt;
pz = px / (npe[0]*npe[1]);
px-= npe[0]*npe[1]*pz;
py = px / npe[0];
px-= npe[0]*py;
pxf = (( px+1)%npe[0]) + npe[0]*py + npe[0]*npe[1]*pz + npe[0]*npe[1]*npe[2]*pt;
pxb = ((npe[0]+px-1)%npe[0]) + npe[0]*py + npe[0]*npe[1]*pz + npe[0]*npe[1]*npe[2]*pt;
pyf = px + (( py+1)%npe[1])*npe[0] + npe[0]*npe[1]*pz + npe[0]*npe[1]*npe[2]*pt;
pyb = px + ((npe[1]+py-1)%npe[1])*npe[0] + npe[0]*npe[1]*pz + npe[0]*npe[1]*npe[2]*pt;
pzf = px + npe[0]*py + (( pz+1)%npe[2])*npe[0]*npe[1] + npe[0]*npe[1]*npe[2]*pt;
pzb = px + npe[0]*py + ((npe[2]+pz-1)%npe[2])*npe[0]*npe[1] + npe[0]*npe[1]*npe[2]*pt;
ptf = px + npe[0]*py + npe[0]*npe[1]*pz + (( pt+1)%npe[3])*npe[0]*npe[1]*npe[2];
ptb = px + npe[0]*py + npe[0]*npe[1]*pz + ((npe[3]+pt-1)%npe[3])*npe[0]*npe[1]*npe[2];
}
domain_e = ( py + pz + pt)%2;
domain_o = 1 - domain_e;
#ifdef _DEBUG_
printf("MPI %d\n", px);
printf("MPI rank=%d px=%d py=%d pz=%d pt=%d\n", rank, px, py, pz, pt);
printf("MPI rank=%d pxf=%d pxb=%d (forward and backward ranks in X-dir)\n", rank, pxf, pxb);
printf("MPI rank=%d pyf=%d pyb=%d (forward and backward ranks in Y-dir)\n", rank, pyf, pyb);
printf("MPI rank=%d pzf=%d pzb=%d (forward and backward ranks in Z-dir)\n", rank, pzf, pzb);
printf("MPI rank=%d ptf=%d ptb=%d (forward and backward ranks in T-dir)\n", rank, ptf, ptb);
printf("MPI rank=%d domain_e=%d\n", rank, domain_e);
printf("MPI rank=%d domain_o=%d\n", rank, domain_o);
fflush(stdout);
#endif
}
// initialize communicaiton
xbound_init(rank,
pxf,pyf,pzf,ptf,pxb,pyb,pzb,ptb);
#ifdef _DEBUG_
printf("qws_init end: MPI rank=%d px=%d py=%d pz=%d pt=%d\n", rank, px, py, pz, pt);
#endif
if (pt == npe[3]-1 && pt != 0 && fbc_f[3] != 1){
fbc[3][0]=(double)(fbc_f[3]*2);
fbc[3][1]=(double)2;
} else if (pt != npe[3]-1 && pt == 0 && fbc_f[3] != 1){
fbc[3][0]=(double)2;
fbc[3][1]=(double)(fbc_f[3]*2);
} else if (pt == npe[3]-1 && pt == 0 && fbc_f[3] != 1){
fbc[3][0]=(double)(fbc_f[3]*2);
fbc[3][1]=(double)(fbc_f[3]*2);
} else {
fbc[3][0]=(double)2;
fbc[3][1]=(double)2;
}
#ifdef _POWER_API_
power_api_init();
#endif
}
//---------------------------------------------------------------------------------------- qws init end
void qws_finalize_() {
xbound_finalize();
if(glud) free(glud);
if(glus) free(glus);
if(clvd) free(clvd);
if(clvs) free(clvs);
#ifdef _POWER_API_
power_api_finalize();
#endif
#ifdef _CHECK_TIMING
#ifdef _TIMING_EACH_RANK
if(!para_outputfile)fclose(para_outputfile);
#endif
#endif
}
//---------------------------------------------------------------------------------------- for testing, not using
void deo_test_(int* pe, int* po, scd_t* out, scd_t* in){
__attribute__((aligned(64))) scd_t tmp;
int x, y, z, t;
pglud_t gxf = &glud[vold*0 + NDIM*vold*(*pe)];
pglud_t gxb = &glud[vold*0 + NDIM*vold*(*po)];
pglud_t gyf = &glud[vold*1 + NDIM*vold*(*pe)];
pglud_t gyb = &glud[vold*1 + NDIM*vold*(*po)];
pglud_t gzf = &glud[vold*2 + NDIM*vold*(*pe)];
pglud_t gzb = &glud[vold*2 + NDIM*vold*(*po)];
pglud_t gtf = &glud[vold*3 + NDIM*vold*(*pe)];
pglud_t gtb = &glud[vold*3 + NDIM*vold*(*po)];
for(t=0; t<nt; t++){
for(z=0; z<nz; z++){
for(y=0; y<ny; y++){
int ick= ( (*pe) + y + z + t)%2;
//printf("ick e y z t %d %d %d %d %d\n",ick,e,y,z,t);
for(x=0; x<nxd; x++){
//printf("iiiii x=%d y=%d z=%d t=%d iyf=%d\n",x, y, z, t, iyf);
__zero_sc(tmp.c);
int i0 = x + nxd*y + nxd*ny*z + nxd*ny*nz*t;
int ixf = ( x+1)%nxd + nxd*y + nxd*ny*z + nxd*ny*nz*t;
int ixb = (nxd+x-1)%nxd + nxd*y + nxd*ny*z + nxd*ny*nz*t;
int iyf = x + (( y+1)%ny)*nxd + nxd*ny*z + nxd*ny*nz*t;
int iyb = x + ((ny+y-1)%ny)*nxd + nxd*ny*z + nxd*ny*nz*t;
int izf = x + nxd*y + (( z+1)%nz)*nxd*ny + nxd*ny*nz*t;
int izb = x + nxd*y + ((nz+z-1)%nz)*nxd*ny + nxd*ny*nz*t;
int itf = x + nxd*y + nxd*ny*z + (( t+1)%nt)*nxd*ny*nz;
int itb = x + nxd*y + nxd*ny*z + ((nt+t-1)%nt)*nxd*ny*nz;
if (ick == 0) {
__mult_xfwd0(tmp.c, gxf, in, i0);
__mult_xbwd0(tmp.c, gxb, in, i0, ixb);
} else {
__mult_xfwd1(tmp.c, gxf, in, i0, ixf);
__mult_xbwd1(tmp.c, gxb, in, i0);
}
__mult_yfwd(tmp.c, gyf, in, i0, iyf);
__mult_ybwd(tmp.c, gyb, in, iyb);
__mult_zfwd(tmp.c, gzf, in, i0, izf);
__mult_zbwd(tmp.c, gzb, in, izb);
__mult_tfwd(tmp.c, gtf, in, i0, itf);
__mult_tbwd(tmp.c, gtb, in, itb);
__store_sc(out[i0].c, tmp.c);
}
}
}
}
}
//----------------------------------------------------------------------------------------
void clvd_vm_(int* pe, scd_t* inout){
__attribute__((aligned(64))) scd_t tmp;
int x, y, z, t;
#pragma omp for private(x,y,z,t,tmp) collapse(3) schedule(static)
for(t=0; t<nt; t++){
for(z=0; z<nz; z++){
for(y=0; y<ny; y++){
for(x=0; x<nxd; x++){
int i0 = x + nxd*y + nxd*ny*z + nxd*ny*nz*t;
__load_sc(tmp.c, inout[i0].c);
//__mult_clvd( tmp.cv, clv[i0 + vold*(*pe)].cv);
__mult_clvd_def( tmp.cv, clvd[i0 + vold*(*pe)].cv);
__store_sc(inout[i0].c, tmp.c);
}
}
}
}
}
//----------------------------------------------------------------------------------------
void clv_s_(int* pe, scs_t* inout){
__attribute__((aligned(64))) scs_t tmp;
int x, y, z, t;
#pragma omp for private(x,y,z,t,tmp) collapse(3) schedule(static)
for(t=0; t<nt; t++){
for(z=0; z<nz; z++){
for(y=0; y<ny; y++){
for(x=0; x<nxs; x++){
int i0 = x + nxs*y + nxs*ny*z + nxs*ny*nz*t;
__load_sc_s(tmp.c, inout[i0].c);
__mult_clvs( tmp.cv, clvs[i0 + vols*(*pe)].cv);
__store_sc_s(inout[i0].c, tmp.c);
}
}
}
}
}
//----------------------------------------------------------------------------------------
void clvd_vm2_(int* pe, scd_t* out, scd_t* in){
__attribute__((aligned(64))) scd_t tmp;
int x, y, z, t;
#pragma omp for private(x,y,z,t,tmp) collapse(3) schedule(static)
for(t=0; t<nt; t++){
for(z=0; z<nz; z++){
for(y=0; y<ny; y++){
for(x=0; x<nxd; x++){
int i0 = x + nxd*y + nxd*ny*z + nxd*ny*nz*t;
__load_sc(tmp.c, in[i0].c);
__mult_clvd( tmp.cv, clvd[i0 + vold*(*pe)].cv);
__store_sc(out[i0].c, tmp.c);
}
}
}
}
}
//----------------------------------------------------------------------------
void zero_scd_field(scd_t* inout, int size){
#pragma omp parallel for
for (int i=0;i<size;i++) {
for (int j=0;j<24;j++) {
for (int iv=0;iv<VLEND;iv++) {
inout[i].ccs[j].v[iv]=0;
}
}
}
}
//----------------------------------------------------------------------------
void zero_scs_field(scs_t* inout, int size){
#pragma omp parallel for
for (int i=0;i<size;i++) {
for (int j=0;j<24;j++) {
for (int iv=0;iv<VLENS;iv++) {
inout[i].ccs[j].v[iv]=0;
}
}
}
}
//----------------------------------------------------------------------------
void print_scdnorm2(const char* a, scd_t* in, int size){
double rtmp0 = 0;
#pragma omp parallel for reduction(+:rtmp0)
for(int i=0; i<size; i++){
for(int j=0; j<24; j++){
rvecd_t rvd0 = fmul_d(in[i].ccs[j],in[i].ccs[j]);
rtmp0 += fsum_d(rvd0);
}
}
#ifdef _MPI_
MPI_Allreduce(MPI_IN_PLACE, &rtmp0,1,MPI_DOUBLE,MPI_SUM,MPI_COMM_WORLD);
#endif
if(rank==0){
printf("%30s : %24.14e\n", a, rtmp0);
fflush(stdout);
}
}
//----------------------------------------------------------------------------
void print_scsnorm2(char* a, scs_t* in, int size){
float rtmp0 = 0;
#pragma omp parallel for reduction(+:rtmp0)
for(int i=0; i<size; i++){
for(int j=0; j<24; j++){
vecs_t v = fload1_s(pred_true_all(), &(in[i].ccs[j]), dims_1d, 0);
vecs_t rvd0 = fmul_s(pred_true_all(), v, v);
rtmp0 += fsum_s(pred_true_all(), rvd0);
}
}
#ifdef _MPI_
MPI_Allreduce(MPI_IN_PLACE, &rtmp0,1,MPI_REAL,MPI_SUM,MPI_COMM_WORLD);
#endif
if(rank==0){
printf("%30s : %24.14e\n", a, rtmp0);
fflush(stdout);
}
}
//----------------------------------------------------------------------------------------
//----------------------------------------------------------------------------------------
void deo_in_vm_(int* pe, int* po, scd_t* out, scd_t* in);
void dee_deo_in_vm_(int* pe, int* po, scd_t* out, scd_t* in);
void deo_dag_in_vm_(int* pe, int* po, scd_t* out, scd_t* in);
void dee_deo_dag_in_vm_(int* pe, int* po, scd_t* out, scd_t* in);
void one_minus_dee_deo_in_vm_(int* pe, int* po, scd_t* out, scd_t* in, scd_t* in0, double factor);
void one_minus_deo_dag_in_vm_(int* pe, int* po, scd_t* out, scd_t* in, scd_t* in0, double factor);
void deo_out_pre_vm_(int* pe, int* po, scd_t* out, scd_t* in);
void deo_dag_out_pre_vm_(int* pe, int* po, scd_t* out, scd_t* in);
void deo_out_pos_vm_(int* pe, int* po, scd_t* out, scd_t* in, double factor);
void deo_dag_out_pos_vm_(int* pe, int* po, scd_t* out, scd_t* in, double factor);
void dee_deo_out_pos_vm_(int* pe, int* po, scd_t* out, scd_t* in, double factor);
void dee_deo_dag_out_pos_vm_(int* pe, int* po, scd_t* out, scd_t* in, double factor);
//---------------------------------------------------------------------------------------- double precision
void deo_vm_(int* pe, int* po, scd_t* out, scd_t* in){
deo_out_pre_vm_(pe, po, out, in);
deo_in_vm_(pe, po, out, in);
deo_out_pos_vm_(pe, po, out, in, 1);
}
void dee_deo_vm_(int* pe, int* po, scd_t* out, scd_t* in){
deo_out_pre_vm_(pe, po, out, in);
dee_deo_in_vm_(pe, po, out, in);
dee_deo_out_pos_vm_(pe, po, out, in, 1);
}
void one_minus_dee_deo_vm_(int* pe, int* po, scd_t* out, scd_t* in, scd_t* in0){
double fac = -kappa2;
deo_out_pre_vm_(pe, po, out, in);
one_minus_dee_deo_in_vm_(pe, po, out, in, in0, fac);
dee_deo_out_pos_vm_(pe, po, out, in, fac );
}
void deo_dag_vm_(int* pe, int* po, scd_t* out, scd_t* in){
deo_dag_out_pre_vm_(pe, po, out, in);
deo_dag_in_vm_(pe, po, out, in);
deo_dag_out_pos_vm_(pe, po, out, in, 1);
}
void dee_deo_dag_vm_(int* pe, int* po, scd_t* out, scd_t* in){
deo_dag_out_pre_vm_(pe, po, out, in);
dee_deo_dag_in_vm_(pe, po, out, in);
dee_deo_dag_out_pos_vm_(pe, po, out, in, 1);
}
void one_minus_deo_dag_vm_(int* pe, int* po, scd_t* out, scd_t* in, scd_t* in0){
double fac = -kappa2;
deo_dag_out_pre_vm_(pe, po, out, in);
one_minus_deo_dag_in_vm_(pe, po, out, in, in0, fac);
deo_dag_out_pos_vm_(pe, po, out, in, fac );
}
//----------------------------------------------------------------------------------------
void mtilde_vm_(scd_t* out, scd_t* in){
scd_t* tmp = (scd_t*)malloc( sizeof(scd_t) * nxd*ny*nz*nt);
dee_deo_vm_(pco, pce, tmp, in);
one_minus_dee_deo_vm_(pce, pco, out, tmp, in);
free(tmp);
}
void mtilde_dag_vm_(scd_t* out, scd_t* in){
scd_t* tmp0 = (scd_t*)malloc( sizeof(scd_t) * nxd*ny*nz*nt);
scd_t* tmp1 = (scd_t*)malloc( sizeof(scd_t) * nxd*ny*nz*nt);
clvd_vm2_(pce, tmp0, in);
dee_deo_dag_vm_(pco, pce, tmp1, tmp0);
one_minus_deo_dag_vm_(pce, pco, out, tmp1, in);
free(tmp0);
free(tmp1);
}
void mtdagmt_vm_(scd_t* out, scd_t* in){
scd_t* tmp = (scd_t*)malloc( sizeof(scd_t) * nxd*ny*nz*nt);
mtilde_vm_(tmp, in);
mtilde_dag_vm_(out, tmp);
free(tmp);
}
//---------------------------------------------------------------------------------------- single precision
void deo_in_s_(int* pe, int* po, scs_t* out, scs_t* in);
void dee_deo_in_s_(int* pe, int* po, scs_t* out, scs_t* in);
void one_minus_dee_deo_in_s_(int* pe, int* po, scs_t* out, scs_t* in, scs_t* in0, float factor);
void deo_out_pre_s_(const int* pe, const int* po, const scs_t* out, const scs_t* in);
void deo_out_pos_s_(const int* pe, const int* po, const scs_t* out, const scs_t* in, float factor);
void dee_deo_out_pos_s_(int* pe, int* po, scs_t* out, scs_t* in, float factor);
//----------------------------------------------------------------------------------------
void deo_s_(int* pe, int* po, scs_t* out, scs_t* in){
float fac = 1;
deo_out_pre_s_(pe, po, out, in);
deo_in_s_(pe, po, out, in);
deo_out_pos_s_(pe, po, out, in, fac);
}
void dee_deo_s_(int* pe, int* po, scs_t* out, scs_t* in){
deo_out_pre_s_(pe, po, out, in);
dee_deo_in_s_(pe, po, out, in);
dee_deo_out_pos_s_(pe, po, out, in, 1);
}
void one_minus_dee_deo_s_(int* pe, int* po, scs_t* out, scs_t* in, scs_t* in0){
float fac = -(float)kappa2;
deo_out_pre_s_(pe, po, out, in);
one_minus_dee_deo_in_s_(pe, po, out, in, in0, fac);
dee_deo_out_pos_s_(pe, po, out, in, fac );
}
//----------------------------------------------------------------------------------------
void mtilde_s_(scs_t* out, scs_t* in){
scs_t* tmp = (scs_t*)malloc( sizeof(scs_t) * nxs*ny*nz*nt);
dee_deo_s_(pco, pce, tmp, in);
one_minus_dee_deo_s_(pce, pco, out, tmp, in);
free(tmp);
}
//---------------------------------------------------------------------------------------- for DD solver
void ddd_out_pre_d_(scd_t* in, int* domain);
void ddd_out_pos_d_(scd_t* out, scd_t* in, int* domain);
void ddd_in_d_(scd_t* out, scd_t* in, int* domain);
void ddd_out_pre_s_(scs_t* in, int* domain);
void ddd_out_pre_s_noprl_(scs_t* in, int* domain);
void ddd_out_pre_s_no_timer_(scs_t* in, int* domain);
void ddd_out_pre_s_noprl_no_timer_(scs_t* in, int* domain);
void ddd_out_pos_s_(scs_t* out, scs_t* in, int* domain, float factor);
void ddd_out_pos_s_no_timer_(scs_t* out, scs_t* in, int* domain, float factor);
void ddd_in_s_(scs_t* out, scs_t* in, int* domain);
void ddd_out_pos_s_noprl_(scs_t* out, scs_t* in, int* domain, float factor);
void ddd_out_pos_s_noprl_no_timer_(scs_t* out, scs_t* in, int* domain, float factor);
void ddd_in_s_noprl(scs_t* out, scs_t* in, int* domain);
int domain0 = 0;
int domain1 = 1;
//----------------------------------------------------------------------------------------
void ddd_d_(scd_t* out, scd_t* in){
ddd_out_pre_d_(in, &domain0);
ddd_in_d_( &out[vold*0], &in[vold*0], &domain0);
ddd_out_pos_d_(&out[vold*0], &in[vold*0], &domain0);
ddd_out_pre_d_(in, &domain1);
ddd_in_d_( &out[vold*1], &in[vold*1], &domain1);
ddd_out_pos_d_(&out[vold*1], &in[vold*0], &domain1);
}
//----------------------------------------------------------------------------------------
void ddd_eo_s_(scs_t* out, scs_t* in, int* domain){
ddd_out_pre_s_no_timer_(in, domain );
ddd_in_s_( &out[vols*(*domain)], &in[vols*(*domain)], domain);
ddd_out_pos_s_no_timer_(&out[vols*(*domain)], in, domain, (float)mkappa);
}
//----------------------------------------------------------------------------------------
void ddd_s_(scs_t* out, scs_t* in){
ddd_eo_s_(out, in, &domain_e);
ddd_eo_s_(out, in, &domain_o);
}
void jinv_ddd_in_s_(scs_t* x, scs_t* b, int *domain, int* maxiter);
void jinv_ddd_in_s_noprl_(scs_t* x, scs_t* b, int *domain, int* maxiter);
//----------------------------------------------------------------------------------------
//----------------------------------------------------------------------------------------
//----------------------------------------------------------------------------------------
inline void copy_vols2_s_noprl_(scs_t* out, scs_t* in){
#ifdef COMPILE_TIME_DIM_SIZE
const int vols = VOLS;
#else
const int vols = ::vols;
#endif
int i, j;
#pragma omp for private(i,j)
for(i=0; i<vols*2; i++){
if (i+_PFI1 < volse*2) {
__builtin_prefetch(&(((in+i+_PFI1 )->c_prefetch)[0]),0,3);
__builtin_prefetch(&(((in+i+_PFI1 )->c_prefetch)[64]),0,3);
__builtin_prefetch(&(((in+i+_PFI1 )->c_prefetch)[128]),0,3);
__builtin_prefetch(&(((out+i+_PFI1 )->c_prefetch)[0]),1,3);
__builtin_prefetch(&(((out+i+_PFI1 )->c_prefetch)[64]),1,3);
__builtin_prefetch(&(((out+i+_PFI1 )->c_prefetch)[128]),1,3);
}
for(j=0; j<24; j++){
for(int v=0; v<VLENS; v++){
out[i].ccs[j].v[v] = in[i].ccs[j].v[v];
}
}
}
}
//----------------------------------------------------------------------------------------
inline void zero_vols_s_noprl_(scs_t* out){
#ifdef COMPILE_TIME_DIM_SIZE
const int vols = VOLS;
#else
const int vols = ::vols;
#endif
int i, j;
#pragma omp for private(i,j)
for(i=0; i<vols; i++){
if (i+_PFI4 < volse) {
__builtin_prefetch(&((( out+i+_PFI4 )->c_prefetch)[0]),1,3);
__builtin_prefetch(&((( out+i+_PFI4 )->c_prefetch)[64]),1,3);
__builtin_prefetch(&((( out+i+_PFI4 )->c_prefetch)[128]),1,3);
}
for(j=0; j<24; j++){
for(int v=0; v<VLENS; v++){
out[i].ccs[j].v[v] = 0;
}
}
}
}
//----------------------------------------------------------------------------------------
inline void accum_add_vols_s_noprl_(scs_t* out, scs_t* in){
#ifdef COMPILE_TIME_DIM_SIZE
const int vols = VOLS;
#else
const int vols = ::vols;
#endif
int i, j;
#pragma omp for private(i,j)
for(i=0; i<vols; i++){
if (i+_PFI3 < volse) {
__builtin_prefetch(&((( out+i+_PFI3 )->c_prefetch)[0]),1,3);
__builtin_prefetch(&((( out+i+_PFI3 )->c_prefetch)[64]),1,3);
__builtin_prefetch(&((( out+i+_PFI3 )->c_prefetch)[128]),1,3);
__builtin_prefetch(&((( in+i+_PFI3 )->c_prefetch)[0]),0,3);
__builtin_prefetch(&((( in+i+_PFI3 )->c_prefetch)[64]),0,3);
__builtin_prefetch(&((( in+i+_PFI3 )->c_prefetch)[128]),0,3);
}
for(j=0; j<24; j++){
for(int v=0; v<VLENS; v++){
out[i].ccs[j].v[v] += in[i].ccs[j].v[v];
}
}
}
}
//----------------------------------------------------------------------------------------
inline void accum_sub_vols_s_noprl_(scs_t* out, scs_t* in){
#ifdef COMPILE_TIME_DIM_SIZE
const int vols = VOLS;
#else
const int vols = ::vols;
#endif
int i, j;
#pragma omp for private(i,j)
for(i=0; i<vols; i++){
if (i+_PFI3 < volse) {
__builtin_prefetch(&((( out+i+_PFI3 )->c_prefetch)[0]),1,3);
__builtin_prefetch(&((( out+i+_PFI3 )->c_prefetch)[64]),1,3);
__builtin_prefetch(&((( out+i+_PFI3 )->c_prefetch)[128]),1,3);
__builtin_prefetch(&((( in+i+_PFI3 )->c_prefetch)[0]),0,3);
__builtin_prefetch(&((( in+i+_PFI3 )->c_prefetch)[64]),0,3);
__builtin_prefetch(&((( in+i+_PFI3 )->c_prefetch)[128]),0,3);
}
for(j=0; j<24; j++){
for(int v=0; v<VLENS; v++){
out[i].ccs[j].v[v] -= in[i].ccs[j].v[v];
}
}
}
}
//----------------------------------------------------------------------------------------
inline void accum_addsub_vols_s_noprl_(scs_t* out, scs_t* in1, scs_t* in2){
#ifdef COMPILE_TIME_DIM_SIZE
const int vols = VOLS;
#else
const int vols = ::vols;
#endif
int i, j;
#pragma omp for private(i,j)
for(i=0; i<vols; i++){
if (i+_PFI2 < volse) {
__builtin_prefetch(&(((in1+i+_PFI2 )->c_prefetch)[0]),0,3);
__builtin_prefetch(&(((in1+i+_PFI2 )->c_prefetch)[64]),0,3);
__builtin_prefetch(&(((in1+i+_PFI2 )->c_prefetch)[128]),0,3);
__builtin_prefetch(&(((in2+i+_PFI2 )->c_prefetch)[0]),0,3);
__builtin_prefetch(&(((in2+i+_PFI2 )->c_prefetch)[64]),0,3);
__builtin_prefetch(&(((in2+i+_PFI2 )->c_prefetch)[128]),0,3);
__builtin_prefetch(&(((out+i+_PFI2 )->c_prefetch)[0]),1,3);
__builtin_prefetch(&(((out+i+_PFI2 )->c_prefetch)[64]),1,3);
__builtin_prefetch(&(((out+i+_PFI2 )->c_prefetch)[128]),1,3);
}
for(j=0; j<24; j++){
for(int v=0; v<VLENS; v++){
out[i].ccs[j].v[v] += in1[i].ccs[j].v[v] - in2[i].ccs[j].v[v];
}
}
}
}
//----------------------------------------------------------------------------------------
//----------------------------------------------------------------------------------------
void prec_ddd_s_(scs_t* out, scs_t* in, int* nsap, int* nm){
static scs_t *s, *q;
int ret;
ret=0;
if(s==0) {
void *p;
ret = posix_memalign(&p, CLS, sizeof(scs_t)*vols*2);
s = (scs_t*)p;
}
if(q==0) {
void *p;
ret = posix_memalign(&p, CLS, sizeof(scs_t)*vols*2);
q = (scs_t*)p;
}
#ifdef COMPILE_TIME_DIM_SIZE
const int vols = VOLS;
#else
const int vols = ::vols;
#endif
float kappa = ::kappa, mkappa = ::mkappa;
if (*nsap > 10) { printf("prec_ddd_s_ nsap > 10\n"); exit(1); }
if (*nm > 10) { printf("prec_ddd_s_ nn > 10\n"); exit(1); }
_PREC_DDD_S_TIC_;
#pragma omp parallel
{
copy_vols2_s_noprl_(s, in);
if (npe[1]==1 || npe[2]==1 || npe[3]==1){
zero_vols_s_noprl_(&q[vols*domain_o]);
}
_OTHER_CALC_TOC_;
for (int isap=0; isap < *nsap; isap++) {
_JINV_DDD_IN_S_TIC_;
jinv_ddd_in_s_noprl_(&q[vols*domain_e], &s[vols*domain_e], &domain_e, nm);
_JINV_DDD_IN_S_TOC_;
_DDD_OUT_PRE_S_TIC_;
ddd_out_pre_s_noprl_( q, &domain_o);
_DDD_OUT_PRE_S_TOC_;
_DDD_IN_S_TIC_;
ddd_in_s_noprl(&q[vols*domain_o], &q[vols*domain_e], &domain_e);
_DDD_IN_S_TOC_;
_OTHER_CALC_TIC_;
accum_addsub_vols_s_noprl_(&s[vols*domain_e], &in[vols*domain_e], &q[vols*domain_o]);
_OTHER_CALC_TOC_;
_DDD_OUT_POS_S_TIC_;
ddd_out_pos_s_noprl_(&s[vols*domain_o], q, &domain_o, (float)kappa);
_DDD_OUT_POS_S_TOC_;
_JINV_DDD_IN_S_TIC_;
jinv_ddd_in_s_noprl_(&q[vols*domain_o], &s[vols*domain_o], &domain_o, nm);
_JINV_DDD_IN_S_TOC_;
_DDD_OUT_PRE_S_TIC_;
ddd_out_pre_s_noprl_( q, &domain_e);
_DDD_OUT_PRE_S_TOC_;
_DDD_IN_S_TIC_;
ddd_in_s_noprl(&q[vols*domain_e], &q[vols*domain_o], &domain_o);
_DDD_IN_S_TOC_;
_OTHER_CALC_TIC_;
accum_addsub_vols_s_noprl_(&s[vols*domain_o], &in[vols*domain_o], &q[vols*domain_e]);
_OTHER_CALC_TOC_;
_DDD_OUT_POS_S_TIC_;
ddd_out_pos_s_noprl_(&s[vols*domain_e], q, &domain_e, (float)kappa);
_DDD_OUT_POS_S_TOC_;
}
if (npe[1]==1 || npe[2]==1 || npe[3]==1){
_OTHER_CALC_TIC_;
zero_vols_s_noprl_(&q[vols*domain_o]);
_OTHER_CALC_TOC_;
}
_JINV_DDD_IN_S_TIC_;
jinv_ddd_in_s_noprl_(&q[vols*domain_e], &s[vols*domain_e], &domain_e, nm);
_JINV_DDD_IN_S_TOC_;
_DDD_OUT_PRE_S_TIC_;
ddd_out_pre_s_noprl_( q, &domain_o);
_DDD_OUT_PRE_S_TOC_;
_DDD_IN_S_TIC_;
ddd_in_s_noprl( &out[vols*domain_e], &q[vols*domain_e], &domain_e);
_DDD_IN_S_TOC_;
if (npe[1]==1 || npe[2]==1 || npe[3]==1){
_DDD_OUT_POS_S_TIC_;
ddd_out_pos_s_noprl_( &s[vols*domain_o], q, &domain_o, (float)kappa);
_DDD_OUT_POS_S_TOC_;
_JINV_DDD_IN_S_TIC_;
jinv_ddd_in_s_noprl_(&q[vols*domain_o], &s[vols*domain_o], &domain_o, nm);
_JINV_DDD_IN_S_TOC_;
_DDD_OUT_PRE_S_TIC_;
ddd_out_pre_s_noprl_( q , &domain_e);
_DDD_OUT_PRE_S_TOC_;
_DDD_IN_S_TIC_;
ddd_in_s_noprl( &out[vols*domain_o], &q[vols*domain_o], &domain_o);
_DDD_IN_S_TOC_;
_DDD_OUT_POS_S_TIC_;
ddd_out_pos_s_noprl_(&out[vols*domain_e], q, &domain_e, (float)mkappa);
_DDD_OUT_POS_S_TOC_;
_DDD_OUT_PRE_S_TIC_;
ddd_out_pre_s_noprl_( q , &domain_o);
_DDD_OUT_PRE_S_TOC_;
_DDD_OUT_POS_S_TIC_;
ddd_out_pos_s_noprl_(&out[vols*domain_o], q, &domain_o, (float)mkappa);
_DDD_OUT_POS_S_TOC_;
}else{
_OTHER_CALC_TIC_;
zero_vols_s_noprl_(&out[vols*domain_o]);
_OTHER_CALC_TOC_;
_DDD_OUT_POS_S_TIC_;
ddd_out_pos_s_noprl_( &out[vols*domain_o], q, &domain_o, (float)mkappa);
_DDD_OUT_POS_S_TOC_;
_OTHER_CALC_TIC_;
accum_sub_vols_s_noprl_(&s[vols*domain_o], &out[vols*domain_o]);
_OTHER_CALC_TOC_;
_JINV_DDD_IN_S_TIC_;
jinv_ddd_in_s_noprl_(&q[vols*domain_o], &s[vols*domain_o], &domain_o, nm);
_JINV_DDD_IN_S_TOC_;
_DDD_OUT_PRE_S_TIC_;
ddd_out_pre_s_noprl_( q , &domain_e);
_DDD_OUT_PRE_S_TOC_;
_DDD_IN_S_TIC_;
ddd_in_s_noprl( &s[vols*domain_o], &q[vols*domain_o], &domain_o);
_DDD_IN_S_TOC_;
_OTHER_CALC_TIC_;
accum_add_vols_s_noprl_(&out[vols*domain_o], &s[vols*domain_o]);
_OTHER_CALC_TOC_;
_DDD_OUT_POS_S_TIC_;
ddd_out_pos_s_noprl_(&out[vols*domain_e], q, &domain_e, (float)mkappa);
_DDD_OUT_POS_S_TOC_;
}
_OTHER_CALC_TIC_;
} //end parallel
_PREC_DDD_S_TOC_;
}
//----------------------------------------------------------------------------------------
void prec_s_(scs_t* out, scs_t* in, int* nsap, int* nm){
static scs_t *s, *q;
int ret;
ret=0;
if(s==0) {
void *p;
ret = posix_memalign(&p, CLS, sizeof(scs_t)*vols*2);
s = (scs_t*)p;
}
if(q==0) {
void *p;
ret = posix_memalign(&p, CLS, sizeof(scs_t)*vols*2);
q = (scs_t*)p;
}
#ifdef COMPILE_TIME_DIM_SIZE
const int vols = VOLS;
#else
const int vols = ::vols;
#endif
float kappa = ::kappa;
_PREC_S_TIC_;
#pragma omp parallel
{
copy_vols2_s_noprl_(s, in);
if (npe[1]==1 || npe[2]==1 || npe[3]==1)zero_vols_s_noprl_(&q[vols*domain_o]);
for (int isap=0; isap < *nsap; isap++) {
jinv_ddd_in_s_noprl_(&q[vols*domain_e], &s[vols*domain_e], &domain_e, nm);
ddd_out_pre_s_noprl_no_timer_( q, &domain_o);
ddd_in_s_noprl(&q[vols*domain_o], &q[vols*domain_e], &domain_e);
accum_addsub_vols_s_noprl_(&s[vols*domain_e], &in[vols*domain_e], &q[vols*domain_o]);
ddd_out_pos_s_noprl_no_timer_(&s[vols*domain_o], q, &domain_o, (float)kappa);
jinv_ddd_in_s_noprl_(&q[vols*domain_o], &s[vols*domain_o], &domain_o, nm);
ddd_out_pre_s_noprl_no_timer_( q, &domain_e);
ddd_in_s_noprl(&q[vols*domain_e], &q[vols*domain_o], &domain_o);
accum_addsub_vols_s_noprl_(&s[vols*domain_o], &in[vols*domain_o], &q[vols*domain_e]);
ddd_out_pos_s_noprl_no_timer_(&s[vols*domain_e], q, &domain_e, (float)kappa);
}
if (npe[1]==1 || npe[2]==1 || npe[3]==1)zero_vols_s_noprl_(&out[vols*domain_o]);
jinv_ddd_in_s_noprl_(&out[vols*domain_e], &s[vols*domain_e], &domain_e, nm);
ddd_out_pre_s_noprl_no_timer_(out, &domain_o);
ddd_out_pos_s_noprl_no_timer_(&s[vols*domain_o], out, &domain_o, (float)kappa);
jinv_ddd_in_s_noprl_(&out[vols*domain_o], &s[vols*domain_o], &domain_o, nm);
} //end parallel
_PREC_S_TOC_;
}
// //----------------------------------------------------------------------------------------
// void copy_s(float* out, float* in, int n){
// int i;
//#pragma omp parallel for private(i)
// for (i=0; i<n; i++){
// out[i] = in[i];
// }
// }
// //----------------------------------------------------------------------------------------
// void zero_s(float* out, int n){
// int i;
//#pragma omp parallel for private(i)
// for (i=0; i<n; i++){
// out[i] = 0;
// }
// }
// //----------------------------------------------------------------------------------------
// void accum_addsub_s(float* out, float* in1, float* in2, int n){
// int i;
//#pragma omp parallel for private(i)
// for (i=0; i<n; i++){
// out[i] = in1[i] - in2[i] + out[i];
// }
// }
// //----------------------------------------------------------------------------------------