-
Notifications
You must be signed in to change notification settings - Fork 7
/
partio.c
693 lines (592 loc) · 20.9 KB
/
partio.c
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
#include <stdlib.h>
#include <stdio.h>
#include <math.h>
#include <assert.h>
#include <fftw3.h>
#include <mpi.h>
#include <gsl/gsl_math.h>
#include <gsl/gsl_sort_long.h>
#include <fitsio.h>
#include <unistd.h>
#include "raytrace.h"
#include "read_lensplanes_hdf5.h"
#include "read_lensplanes_pixLC.h"
static int compPartNest(const void *a, const void *b)
{
if(((const Part*)a)->nest > ((const Part*)b)->nest)
return 1;
else if(((const Part*)a)->nest < ((const Part*)b)->nest)
return -1;
else
return 0;
}
struct nctg {
long nest;
long task;
};
static int compNCTGTask(const void *a, const void *b)
{
if(((const struct nctg*)a)->task > ((const struct nctg*)b)->task)
return 1;
else if(((const struct nctg*)a)->task < ((const struct nctg*)b)->task)
return -1;
else
return 0;
}
/* generic io interface */
void readRayTracingPlaneAtPeanoInds(long planeNum, long HEALPixOrder, long *PeanoIndsToRead, long NumPeanoIndsToRead, Part **LCParts, long *NumLCParts)
{
void (*read_lens_plane)(long, long, long *, long, Part **, long *) = NULL;
if(strcmp_caseinsens(rayTraceData.LensPlaneType,"HDF5") == 0)
{
read_lens_plane = &readRayTracingPlaneAtPeanoInds_HDF5;
}
else if(strcmp_caseinsens(rayTraceData.LensPlaneType,"pixLC") == 0)
{
read_lens_plane = &readRayTracingPlaneAtPeanoInds_pixLC;
}
else
{
fprintf(stderr,"%d: readRayTracingPlaneAtPeanoInds - could not find I/O code for lens plane type '%s'!\n",ThisTask,rayTraceData.LensPlaneType);
MPI_Abort(MPI_COMM_WORLD,666);
}
read_lens_plane(planeNum,HEALPixOrder,PeanoIndsToRead,NumPeanoIndsToRead,LCParts,NumLCParts);
}
/* reads light cone particles into bundleCells for the given planeNum */
void read_lcparts_at_planenum(long planeNum)
{
long i,n;
long *PeanoIndsToRead;
long NumPeanoIndsToRead;
long bundleNest;
double vec[3];
long shift;
long NumGroups,myGroup,currGroup,readFromPlane;
double t0;
shift = 2*(HEALPIX_UTILS_MAXORDER-rayTraceData.bundleOrder);
/* set up reading vars
1) get all cells which are either assigned to this task (bit 0 set) or are buffer cells from which we need particles (bit 1 set)
2) find their Peano inds for reading from lens planes
3) make vector which stores how many particles are currently allocated for a given bundle cell - used later for moving parts into bundleCells
*/
NumPeanoIndsToRead = 0;
for(i=0;i<NbundleCells;++i)
if(ISSETBITFLAG(bundleCells[i].active,PRIMARY_BUNDLECELL))
++NumPeanoIndsToRead;
PeanoIndsToRead = (long*)malloc(sizeof(long)*NumPeanoIndsToRead);
assert(PeanoIndsToRead != NULL);
n = 0;
for(i=0;i<NbundleCells;++i)
if(ISSETBITFLAG(bundleCells[i].active,PRIMARY_BUNDLECELL))
{
PeanoIndsToRead[n] = nest2peano(i,rayTraceData.bundleOrder);
++n;
}
assert(n == NumPeanoIndsToRead);
/* init and free old parts if needed */
destroy_parts();
/* read particles from lens plane
1) uses peano inds computed above
2) init bundle cells and numPartsAlloc for allocating particle mem
3) read in groups to limit I/O usage
*/
NumGroups = NTasks/rayTraceData.NumFilesIOInParallel;
if(NTasks - NumGroups*rayTraceData.NumFilesIOInParallel > 0)
++NumGroups;
myGroup = ThisTask/rayTraceData.NumFilesIOInParallel;
readFromPlane = 0;
t0 = -MPI_Wtime();
for(currGroup=0;currGroup<NumGroups;++currGroup)
{
if(currGroup == myGroup)
{
readFromPlane = 1;
readRayTracingPlaneAtPeanoInds(planeNum,rayTraceData.bundleOrder,PeanoIndsToRead,NumPeanoIndsToRead,&lensPlaneParts,&NlensPlaneParts);
free(PeanoIndsToRead);
if(NlensPlaneParts > 0)
{
/* reorder parts and build indexing for bundleCells */
for(i=0;i<NlensPlaneParts;++i)
{
vec[0] = (double) (lensPlaneParts[i].pos[0]);
vec[1] = (double) (lensPlaneParts[i].pos[1]);
vec[2] = (double) (lensPlaneParts[i].pos[2]);
lensPlaneParts[i].nest = vec2nest(vec,HEALPIX_UTILS_MAXORDER);
}
qsort(lensPlaneParts,(size_t) NlensPlaneParts,sizeof(Part),compPartNest);
/* now fill in index vals in bundleCells */
for(i=0;i<NlensPlaneParts;++i)
{
bundleNest = lensPlaneParts[i].nest >> shift;
if(bundleCells[bundleNest].Nparts == 0)
bundleCells[bundleNest].firstPart = i;
bundleCells[bundleNest].Nparts += 1;
}
}
}
//////////////////////////////
MPI_Barrier(MPI_COMM_WORLD);
//////////////////////////////
}
assert(readFromPlane == 1);
t0 += MPI_Wtime();
if(ThisTask == 0)
{
fprintf(stderr,"read parts in %f seconds\n",t0);
fflush(stderr);
}
//now do an exchange to get buffer parts
long j,log2NTasks;
long level,sendTask,recvTask;
long Nsend,Nrecv;
long NumBufferCells = 0,NumPartsAlloc,NumBufferParts;
Part *tmpPart;
long rpInd;
struct nctg *nestCellsToGet, *nestCellsToSend=NULL, *tmpNCTG;
long NnestCellsToSend = 0;
long firstNestCellForRecvTask,NnestCellsForRecvTask;
MPI_Status Stat;
int count,NnestCellsToSendAlloc=0;
MPI_Request requestSend,requestRecv;
int didSend,didRecv;
t0 = -MPI_Wtime();
log2NTasks = 0;
while(NTasks > (1 << log2NTasks))
++log2NTasks;
//get the bundle cells for which parts are needed and which task they are on
for(i=0;i<NbundleCells;++i)
if(ISSETBITFLAG(bundleCells[i].active,PARTBUFF_BUNDLECELL))
++NumBufferCells;
nestCellsToGet = (struct nctg*)malloc(sizeof(struct nctg)*NumBufferCells);
assert(nestCellsToGet != NULL);
NumBufferCells = 0;
for(i=0;i<NbundleCells;++i)
if(ISSETBITFLAG(bundleCells[i].active,PARTBUFF_BUNDLECELL))
{
nestCellsToGet[NumBufferCells].nest = i;
rpInd = bundleCellsNest2RestrictedPeanoInd[i];
nestCellsToGet[NumBufferCells].task = -1;
if(rpInd != -1)
{
for(j=0;j<NTasks;++j)
{
if(firstRestrictedPeanoIndTasks[j] <= rpInd && rpInd <= lastRestrictedPeanoIndTasks[j])
{
nestCellsToGet[NumBufferCells].task = j;
break;
}
}
if(nestCellsToGet[NumBufferCells].task == -1)
{
fprintf(stderr,"%d: could not find task for rpInd = %ld\n",ThisTask,rpInd);
MPI_Abort(MPI_COMM_WORLD,123);
}
}
++NumBufferCells;
}
//sort by task
qsort(nestCellsToGet,(size_t) NumBufferCells,sizeof(struct nctg),compNCTGTask);
//make extra room for incoming parts
NumPartsAlloc = NlensPlaneParts;
NumBufferParts = NlensPlaneParts*((double) NumBufferCells)/((double) NumPeanoIndsToRead)*1.1;
if(NumBufferParts*sizeof(Part)/1024.0/1024.0 > 100.0)
NumBufferParts = (100.0*1024.0*1024.0)/sizeof(Part);
tmpPart = (Part*)realloc(lensPlaneParts,sizeof(Part)*(NlensPlaneParts + NumBufferParts));
if(tmpPart != NULL)
{
lensPlaneParts = tmpPart;
NumPartsAlloc = NlensPlaneParts + NumBufferParts;
}
else
{
fprintf(stderr,"%d: could not realloc lensPlaneParts for getting buffer regions! - wanted %lg MB extra for %lf MB total.\n",
ThisTask,NumBufferParts*sizeof(Part)/1024.0/1024.0,(NlensPlaneParts + NumBufferParts)*sizeof(Part)/1024.0/1024.0);
MPI_Abort(MPI_COMM_WORLD,123);
}
NumBufferParts = 0;
/*algorithm to loop through pairs of tasks linearly
-lifted from Gadget-2 under GPL (http://www.gnu.org/copyleft/gpl.html)
-see pm_periodic.c from Gadget-2 at http://www.mpa-garching.mpg.de/gadget/
*/
for(level = 0; level < (1 << log2NTasks); level++) /* note: for level=0, target is the same task */
{
sendTask = ThisTask;
recvTask = ThisTask ^ level;
if(recvTask < NTasks && sendTask != recvTask) //never need to give parts to ourselves since part buffer cells are not primary cells on same task
{
//first see if we need to get any bundle cells parts from recvTask
firstNestCellForRecvTask = -1;
NnestCellsForRecvTask = 0;
for(j=0;j<NumBufferCells;++j)
{
if(nestCellsToGet[j].task == recvTask)
{
if(firstNestCellForRecvTask == -1)
firstNestCellForRecvTask = j;
++NnestCellsForRecvTask;
}
}
if(NnestCellsForRecvTask == 0)
firstNestCellForRecvTask = 0;
Nsend = NnestCellsForRecvTask;
MPI_Sendrecv(&Nsend,1,MPI_LONG,(int) recvTask,TAG_NUMBUFF_PIO,
&Nrecv,1,MPI_LONG,(int) recvTask,TAG_NUMBUFF_PIO,
MPI_COMM_WORLD,&Stat);
NnestCellsToSend = Nrecv;
if(Nsend > 0 || Nrecv > 0) //there is overlap between tasks so they may need to send stuff
{
//get bundle cells for which parts should be sent back
if(NnestCellsToSendAlloc < NnestCellsToSend)
{
tmpNCTG = (struct nctg*)realloc(nestCellsToSend,sizeof(struct nctg)*NnestCellsToSend);
if(tmpNCTG != NULL)
{
nestCellsToSend = tmpNCTG;
NnestCellsToSendAlloc = NnestCellsToSend;
}
else
{
fprintf(stderr,"%d: could not realloc nestCellsToSend!\n",ThisTask);
MPI_Abort(MPI_COMM_WORLD,123);
}
}
MPI_Sendrecv(nestCellsToGet+firstNestCellForRecvTask,(int) (Nsend*sizeof(struct nctg)),MPI_BYTE,(int) recvTask,TAG_BUFF_PIO,
nestCellsToSend,(int) (NnestCellsToSend*sizeof(struct nctg)),MPI_BYTE,(int) recvTask,TAG_BUFF_PIO,
MPI_COMM_WORLD,&Stat);
//get # of parts to send back
Nsend = 0;
for(i=0;i<NnestCellsToSend;++i)
if(ISSETBITFLAG(bundleCells[nestCellsToSend[i].nest].active,PRIMARY_BUNDLECELL))
Nsend += bundleCells[nestCellsToSend[i].nest].Nparts;
//get # of parts to recv and make sure have room
MPI_Sendrecv(&Nsend,1,MPI_LONG,(int) recvTask,TAG_NUMPBUFF_PIO,
&Nrecv,1,MPI_LONG,(int) recvTask,TAG_NUMPBUFF_PIO,
MPI_COMM_WORLD,&Stat);
if(NlensPlaneParts+NumBufferParts+Nrecv > NumPartsAlloc)
{
tmpPart = (Part*)realloc(lensPlaneParts,sizeof(Part)*(NumPartsAlloc+Nrecv*4));
if(tmpPart != NULL)
{
lensPlaneParts = tmpPart;
NumPartsAlloc += Nrecv*4;
}
else
{
fprintf(stderr,"%d: could not realloc lensPlaneParts for getting buffer regions in loop! - wanted %lg MB extra for %lf MB total.\n",
ThisTask,Nrecv*4*sizeof(Part)/1024.0/1024.0,(NumPartsAlloc+Nrecv*4)*sizeof(Part)/1024.0/1024.0);
MPI_Abort(MPI_COMM_WORLD,123);
}
}
//if have parts to send or recv, do it
if(Nsend > 0 || Nrecv > 0)
{
i = 0;
while(Nsend > 0 || Nrecv > 0)
{
if(Nrecv > 0) //recv parts
{
MPI_Irecv(lensPlaneParts+NlensPlaneParts+NumBufferParts,
(int) (sizeof(Part)*(NumPartsAlloc-(NlensPlaneParts+NumBufferParts))),MPI_BYTE,
(int) recvTask,TAG_PBUFF_PIO,MPI_COMM_WORLD,&requestRecv);
didRecv = 1;
}
else
didRecv = 0;
if(Nsend > 0)
{
while(i < NnestCellsToSend && !(ISSETBITFLAG(bundleCells[nestCellsToSend[i].nest].active,PRIMARY_BUNDLECELL) && bundleCells[nestCellsToSend[i].nest].Nparts > 0))
++i;
if(i >= NnestCellsToSend)
{
fprintf(stderr,"%d: out of nest cells in while Nsend, Nrecv loop\n",ThisTask);
MPI_Abort(MPI_COMM_WORLD,123);
}
MPI_Issend(lensPlaneParts+bundleCells[nestCellsToSend[i].nest].firstPart,
(int) (sizeof(Part)*bundleCells[nestCellsToSend[i].nest].Nparts),MPI_BYTE,
(int) recvTask,TAG_PBUFF_PIO,MPI_COMM_WORLD,&requestSend);
Nsend -= bundleCells[nestCellsToSend[i].nest].Nparts;
didSend = 1;
++i;
}
else
didSend = 0;
if(didRecv)
{
MPI_Wait(&requestRecv,&Stat);
MPI_Get_count(&Stat,MPI_BYTE,&count);
count /= sizeof(Part);
NumBufferParts += count;
Nrecv -= count;
}
if(didSend)
MPI_Wait(&requestSend,&Stat);
}
}
}
}
}
//update total number of parts
NlensPlaneParts += NumBufferParts;
//clean up
if(NnestCellsToSendAlloc > 0)
free(nestCellsToSend);
free(nestCellsToGet);
t0 += MPI_Wtime();
if(ThisTask == 0)
{
fprintf(stderr,"got buffer parts in %f seconds\n",t0);
fflush(stderr);
}
//if using the full light cone part distribution,
//need to read in buffer parts which are not primary cells in the domain
#ifdef USE_FULLSKY_PARTDIST
Part *buffParts = NULL;
long NumBuffParts = 0;
long peanoInd;
t0 = -MPI_Wtime();
NumPeanoIndsToRead = 1;
for(i=0;i<NbundleCells;++i)
{
if(ISSETBITFLAG(bundleCells[i].active,PARTBUFF_BUNDLECELL) && bundleCellsNest2RestrictedPeanoInd[i] < 0)
{
//read parts from file
peanoInd = nest2peano(i,rayTraceData.bundleOrder);
readRayTracingPlaneAtPeanoInds(planeNum,rayTraceData.bundleOrder,&peanoInd,NumPeanoIndsToRead,&buffParts,&NumBuffParts);
//now add to current parts vector if needed
if(NumBuffParts > 0)
{
//make sure have mem
if(NlensPlaneParts+NumBuffParts > NumPartsAlloc)
{
tmpPart = (Part*)realloc(lensPlaneParts,sizeof(Part)*(NumPartsAlloc+NumBuffParts*4));
if(tmpPart != NULL)
{
lensPlaneParts = tmpPart;
NumPartsAlloc += NumBuffParts*4;
}
else
{
fprintf(stderr,"%d: could not realloc lensPlaneParts for getting USE_FULLSKY_PARTDIST buffer regions in loop! - wanted %lg MB extra for %lf MB total.\n",
ThisTask,NumBuffParts*4*sizeof(Part)/1024.0/1024.0,(NumPartsAlloc+NumBuffParts*4)*sizeof(Part)/1024.0/1024.0);
MPI_Abort(MPI_COMM_WORLD,123);
}
}
//add to code
for(j=0;j<NumBuffParts;++j)
{
vec[0] = (double) (buffParts[j].pos[0]);
vec[1] = (double) (buffParts[j].pos[1]);
vec[2] = (double) (buffParts[j].pos[2]);
buffParts[j].nest = vec2nest(vec,HEALPIX_UTILS_MAXORDER);
lensPlaneParts[NlensPlaneParts+j] = buffParts[j];
}
NlensPlaneParts += NumBuffParts;
//clean up
free(buffParts);
buffParts = NULL;
NumBuffParts = 0;
}
}
}
t0 += MPI_Wtime();
if(ThisTask == 0)
{
fprintf(stderr,"got fullsky buffer parts in %f seconds\n",t0);
fflush(stderr);
}
#endif
//free extra mem
if(NlensPlaneParts < NumPartsAlloc)
{
NumPartsAlloc = NlensPlaneParts;
tmpPart = (Part*)realloc(lensPlaneParts,sizeof(Part)*NlensPlaneParts);
if(tmpPart != NULL)
{
lensPlaneParts = tmpPart;
}
else
{
fprintf(stderr,"%d: could not do final realloc lensPlaneParts for getting buffer regions!\n",ThisTask);
MPI_Abort(MPI_COMM_WORLD,123);
}
}
//redo index vals in bundleCells
for(i=0;i<NbundleCells;++i)
{
bundleCells[i].Nparts = 0;
bundleCells[i].firstPart = -1;
}
qsort(lensPlaneParts,(size_t) NlensPlaneParts,sizeof(Part),compPartNest);
shift = 2*(HEALPIX_UTILS_MAXORDER-rayTraceData.bundleOrder);
for(i=0;i<NlensPlaneParts;++i)
{
bundleNest = lensPlaneParts[i].nest >> shift;
if(bundleCells[bundleNest].Nparts == 0)
bundleCells[bundleNest].firstPart = i;
bundleCells[bundleNest].Nparts += 1;
}
}
/* reads light cone particles into bundleCells for the given planeNum */
void read_lcparts_at_planenum_fullsky_partdist(long planeNum)
{
long i,n;
long *PeanoIndsToRead;
long NumPeanoIndsToRead;
long bundleNest;
double vec[3];
long shift;
long NumGroups,myGroup,currGroup,readFromPlane;
double t0;
shift = 2*(HEALPIX_UTILS_MAXORDER - rayTraceData.bundleOrder);
/* set up reading vars
1) get all cells which are either assigned to this task (bit 0 set) or are buffer cells from which we need particles (bit 1 set)
2) find their Peano inds for reading from lens planes
3) make vector which stores how many particles are currently allocated for a given bundle cell - used later for moving parts into bundleCells
*/
NumPeanoIndsToRead = 0;
for(i=0;i<NbundleCells;++i)
if(ISSETBITFLAG(bundleCells[i].active,FULLSKY_PARTDIST_PRIMARY_BUNDLECELL))
++NumPeanoIndsToRead;
PeanoIndsToRead = (long*)malloc(sizeof(long)*NumPeanoIndsToRead);
assert(PeanoIndsToRead != NULL);
n = 0;
for(i=0;i<NbundleCells;++i)
if(ISSETBITFLAG(bundleCells[i].active,FULLSKY_PARTDIST_PRIMARY_BUNDLECELL))
{
PeanoIndsToRead[n] = nest2peano(i,rayTraceData.bundleOrder);
++n;
}
assert(n == NumPeanoIndsToRead);
/* init and free old parts if needed */
destroy_parts();
/* read particles from lens plane
1) uses peano inds computed above
2) init bundle cells and numPartsAlloc for allocating particle mem
3) read in groups to limit I/O usage
*/
NumGroups = NTasks/rayTraceData.NumFilesIOInParallel;
if(NTasks - NumGroups*rayTraceData.NumFilesIOInParallel > 0)
++NumGroups;
myGroup = ThisTask/rayTraceData.NumFilesIOInParallel;
readFromPlane = 0;
t0 = -MPI_Wtime();
for(currGroup=0;currGroup<NumGroups;++currGroup)
{
if(currGroup == myGroup)
{
readFromPlane = 1;
readRayTracingPlaneAtPeanoInds(planeNum,rayTraceData.bundleOrder,PeanoIndsToRead,NumPeanoIndsToRead,&lensPlaneParts,&NlensPlaneParts);
free(PeanoIndsToRead);
if(NlensPlaneParts > 0)
{
/* reorder parts and build indexing for bundleCells */
for(i=0;i<NlensPlaneParts;++i)
{
vec[0] = (double) (lensPlaneParts[i].pos[0]);
vec[1] = (double) (lensPlaneParts[i].pos[1]);
vec[2] = (double) (lensPlaneParts[i].pos[2]);
lensPlaneParts[i].nest = vec2nest(vec,HEALPIX_UTILS_MAXORDER);
}
qsort(lensPlaneParts,(size_t) NlensPlaneParts,sizeof(Part),compPartNest);
/* now fill in index vals in bundleCells */
for(i=0;i<NlensPlaneParts;++i)
{
bundleNest = lensPlaneParts[i].nest >> shift;
if(bundleCells[bundleNest].Nparts == 0)
bundleCells[bundleNest].firstPart = i;
bundleCells[bundleNest].Nparts += 1;
}
}
}
//////////////////////////////
MPI_Barrier(MPI_COMM_WORLD);
//////////////////////////////
}
assert(readFromPlane == 1);
t0 += MPI_Wtime();
if(ThisTask == 0)
{
fprintf(stderr,"read parts in %f seconds\n",t0);
fflush(stderr);
}
}
/* reads light cone particles into bundleCells for the given planeNum */
void read_lcparts_at_planenum_all(long planeNum)
{
long i,n;
long *PeanoIndsToRead;
long NumPeanoIndsToRead;
long bundleNest;
double vec[3];
long shift;
long NumGroups,myGroup,currGroup,readFromPlane;
double t0;
shift = 2*(HEALPIX_UTILS_MAXORDER - rayTraceData.bundleOrder);
/* set up reading vars
1) get all cells which are either assigned to this task (bit 0 set) or are buffer cells from which we need particles (bit 1 set)
2) find their Peano inds for reading from lens planes
3) make vector which stores how many particles are currently allocated for a given bundle cell - used later for moving parts into bundleCells
*/
NumPeanoIndsToRead = NbundleCells;
PeanoIndsToRead = (long*)malloc(sizeof(long)*NumPeanoIndsToRead);
assert(PeanoIndsToRead != NULL);
n = 0;
for(i=0;i<NbundleCells;++i)
{
PeanoIndsToRead[n] = nest2peano(i,rayTraceData.bundleOrder);
++n;
}
assert(n == NumPeanoIndsToRead);
/* init and free old parts if needed */
destroy_parts();
/* read particles from lens plane
1) uses peano inds computed above
2) init bundle cells and numPartsAlloc for allocating particle mem
3) read in groups to limit I/O usage
*/
NumGroups = NTasks/rayTraceData.NumFilesIOInParallel;
if(NTasks - NumGroups*rayTraceData.NumFilesIOInParallel > 0)
++NumGroups;
myGroup = ThisTask/rayTraceData.NumFilesIOInParallel;
readFromPlane = 0;
t0 = -MPI_Wtime();
for(currGroup=0;currGroup<NumGroups;++currGroup)
{
if(currGroup == myGroup)
{
readFromPlane = 1;
readRayTracingPlaneAtPeanoInds(planeNum,rayTraceData.bundleOrder,PeanoIndsToRead,NumPeanoIndsToRead,&lensPlaneParts,&NlensPlaneParts);
free(PeanoIndsToRead);
if(NlensPlaneParts > 0)
{
/* reorder parts and build indexing for bundleCells */
for(i=0;i<NlensPlaneParts;++i)
{
vec[0] = (double) (lensPlaneParts[i].pos[0]);
vec[1] = (double) (lensPlaneParts[i].pos[1]);
vec[2] = (double) (lensPlaneParts[i].pos[2]);
lensPlaneParts[i].nest = vec2nest(vec,HEALPIX_UTILS_MAXORDER);
}
qsort(lensPlaneParts,(size_t) NlensPlaneParts,sizeof(Part),compPartNest);
/* now fill in index vals in bundleCells */
for(i=0;i<NlensPlaneParts;++i)
{
bundleNest = lensPlaneParts[i].nest >> shift;
if(bundleCells[bundleNest].Nparts == 0)
bundleCells[bundleNest].firstPart = i;
bundleCells[bundleNest].Nparts += 1;
}
}
}
//////////////////////////////
MPI_Barrier(MPI_COMM_WORLD);
//////////////////////////////
}
assert(readFromPlane == 1);
t0 += MPI_Wtime();
if(ThisTask == 0)
{
fprintf(stderr,"read parts in %f seconds\n",t0);
fflush(stderr);
}
}