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
|
#if HAVE_CONFIG_H
# include "config.fh"
#endif
c $Id: test.F,v 1.64.2.11 2007-04-06 22:37:35 d3g293 Exp $
c vector boxes lack arithmetic precision
# define THRESH 1d-13
# define THRESHF 2e-5
#define MISMATCH(x,y) abs(x-y)/max(1d0,abs(x)).gt.THRESH
#define MISMATCHF(x,y) abs(x-y)/max(1.0,abs(x)).gt.THRESHF
c#define NEW_API
c#define MIRROR
#define GA3
#define NGA_GATSCAT
c#define BLOCK_CYCLIC
c#define USE_SCALAPACK_DISTR
c#define USE_RESTRICTED
#ifdef USE_RESTRICTED
# define NEW_API
#endif
#define MEM_INC 1000
#ifdef BLOCK_CYCLIC
# define NEW_API
# undef MIRROR
#else
# undef USE_SCALAPAC_DISTR
#endif
program main
implicit none
#include "mafdecls.fh"
#include "global.fh"
#include "testutil.fh"
integer heap, stack, fudge, ma_heap, me, nproc, map(4096), block
integer g_s, ndim, dim1, i
logical status
parameter (heap=200*200*4, fudge=100, stack=200*200)
c
c*** Intitialize a message passing library
c
#include "mp3.fh"
c
c*** Initialize GA
c
c There are 2 choices: ga_initialize or ga_initialize_ltd.
c In the first case, there is no explicit limit on memory usage.
c In the second, user can set limit (per processor) in bytes.
c
call ga_initialize()
nproc = ga_nnodes()
me = ga_nodeid()
c we can also use GA_set_memory_limit BEFORE first ga_create call
c
ma_heap = heap/nproc + fudge
ma_heap = 2*ma_heap
#ifdef USE_RESTRICTED
ma_heap = 2*ma_heap
#endif
call GA_set_memory_limit(util_mdtob(ma_heap))
c
if(ga_nodeid().eq.0)then
#ifdef MIRROR
print *,' Performing tests on Mirrored Arrays'
#endif
print *,' GA initialized '
call ffflush(6)
endif
c
c*** Initialize the MA package
c MA must be initialized before any global array is allocated
c
status = ma_init(MT_DCPL, stack, ma_heap)
if (.not. status) call ga_error('ma_init failed',-1)
c
c Uncomment the below line to register external memory allocator
c for dynamic arrays inside GA routines.
c call register_ext_memory()
c
if(me.eq.(nproc-1))then
print *, 'using ', nproc,' process(es) ', ga_cluster_nnodes(),
$ ' cluster nodes'
print *,'process ', me, ' is on node ',ga_cluster_nodeid(),
$ ' with ', ga_cluster_nprocs(-1), ' processes'
call ffflush(6)
endif
c
c create array to force staggering of memory and uneven distribution
c of pointers
c
dim1 = MEM_INC
map(1) = 1
do i = 2, nproc
map(i) = MEM_INC*(i-1)+1
dim1 = dim1 + MEM_INC*i
end do
g_s = ga_create_handle()
ndim = 1
call ga_set_data(g_s,ndim,dim1,MT_INT)
call ga_set_array_name(g_s,'s')
call ga_set_irreg_distr(g_s,map,nproc)
c
c*** Check support for single precision complex arrays
c
if (me.eq.0) then
write(6,*)
write(6,*) ' CHECKING SINGLE COMPLEX '
write(6,*)
call ffflush(6)
endif
call check_complex_float()
c
c*** Check support for double precision complex arrays
c
if (me.eq.0) then
write(6,*)
write(6,*) ' CHECKING DOUBLE COMPLEX '
write(6,*)
call ffflush(6)
endif
call check_complex()
if(me.eq.0) call ga_print_stats()
if(me.eq.0) print *,' '
if(me.eq.0) print *,'All tests successful '
status = ga_destroy(g_s)
c
c*** Tidy up the GA package
c
call ga_terminate()
c
c*** Tidy up after message-passing library
c
call MP_FINALIZE()
c
end
subroutine check_complex_float()
implicit none
#include "mafdecls.fh"
#include "global.fh"
#include "testutil.fh"
c
integer n,m
parameter (n = 60)
parameter (m = 2*n)
complex a(n,n), b(n,n), v(m),w(m)
#ifdef MIRROR
integer ndim, dims(2), chunk(2), p_mirror
#else
# ifdef NEW_API
integer ndim, dims(2), chunk(2), p_mirror
# endif
#endif
integer iv(m), jv(m)
logical status
integer g_a, g_b
integer iran, i,j, loop,nloop,maxloop, ilo, ihi, jlo, jhi, itmp
integer nproc, me, int, ij, inc, ii, jj, nnodes
parameter (maxloop = 100)
integer maxproc
parameter (maxproc = 4096)
double precision crap, real
double precision nwords
complex x, sum1, sum2, factor
integer lprocs, inode, iproc, lproc
integer scpl_type, istatus
#ifdef USE_RESTRICTED
integer num_rstrctd
integer rstrctd_list(maxproc/2)
#endif
intrinsic int
iran(i) = int(drand(0)*real(i)) + 1
c
nproc = ga_nnodes()
me = ga_nodeid()
inode = ga_cluster_nodeid()
lprocs = ga_cluster_nprocs(inode)
nnodes = ga_cluster_nnodes()
iproc = mod(me,lprocs)
nloop = Min(maxloop,n)
#ifdef USE_RESTRICTED
num_rstrctd = nproc/2
if (num_rstrctd.eq.0) num_rstrctd = 1
do i = 1, num_rstrctd
rstrctd_list(i) = (num_rstrctd/2) + i-1
end do
#endif
c
c a() is a local copy of what the global array should start as
c
do j = 1, n
do i = 1, n
#ifndef MIRROR
a(i,j) = cmplx(real(i-1), real((j-1)*n))
#else
a(i,j) = cmplx(real(inode),0.0d00)
+ + cmplx(real(i-1), real((j-1)*n))
#endif
b(i,j) = cmplx(-1d0,1d0)
enddo
enddo
c
c Create type
c
scpl_type = nga_register_type(8)
c
c Create a global array
c
c print *,ga_nodeid(), ' creating array'
call ffflush(6)
c call setdbg(1)
#ifdef NEW_API
ndim = 2
dims(1) = n
dims(2) = n
g_a = ga_create_handle()
call ga_set_data(g_a,ndim,dims,scpl_type)
call ga_set_array_name(g_a,'a')
#ifdef USE_RESTRICTED
call ga_set_restricted(g_a, rstrctd_list, num_rstrctd)
#endif
# ifdef MIRROR
p_mirror = ga_pgroup_get_mirror()
call ga_set_pgroup(g_a,p_mirror)
# endif
status = ga_allocate(g_a)
#else
# ifndef MIRROR
status = ga_create(scpl_type, n, n, 'a', 0, 0, g_a)
# else
ndim = 2
dims(1) = n
dims(2) = n
chunk(1) = 0
chunk(2) = 0
p_mirror = ga_pgroup_get_mirror()
status = nga_create_config(scpl_type, ndim, dims, 'a', chunk,
+ p_mirror, g_a)
# endif
#endif
if (.not. status) then
write(6,*) ' ga_create failed'
call ga_error('... exiting ',0)
endif
#ifdef NEW_API
g_b = ga_create_handle()
call ga_set_data(g_b,ndim,dims,scpl_type)
call ga_set_array_name(g_b,'b')
# ifdef MIRROR
call ga_set_pgroup(g_b,p_mirror)
# endif
if (.not.ga_allocate(g_b)) then
#else
# ifndef MIRROR
if (.not. ga_create(scpl_type, n, n, 'b', 0, 0, g_b)) then
# else
if (.not. nga_create_config(scpl_type, ndim, dims, 'b', chunk,
_ p_mirror, g_b)) then
# endif
#endif
call ga_error('ga_create failed for second array ',0)
endif
#ifndef MIRROR
call ga_distribution(g_a,me,ilo, ihi, jlo, jhi)
#else
lproc = me - ga_cluster_procid(inode,0)
call ga_distribution(g_a, lproc, ilo, ihi, jlo, jhi)
#endif
call ga_sync()
c
c Zero the array
c
if (me .eq. 0) then
write(6,21)
21 format(/'> Checking zero ... ')
call ffflush(6)
endif
call ga_zero(g_a)
c
c Check that it is indeed zero
c
call ga_get(g_a, 1, n, 1, n, b, n)
call ga_sync()
do i = 1, n
do j = 1, n
if(b(i,j).ne.(0d0,0d0)) then
write(6,*) me,' zero ', i, j, b(i,j)
call ffflush(6)
c call ga_error('... exiting ',0)
endif
enddo
enddo
if (me.eq.0) then
write(6,*)
write(6,*) ' ga_zero is OK'
write(6,*)
endif
call ga_sync()
c
c Each node fills in disjoint sections of the array
c
if (me .eq. 0) then
write(6,2)
2 format(/'> Checking disjoint put ... ')
call ffflush(6)
endif
call ga_sync()
c
inc = (n-1)/4 + 1
ij = 0
do j = 1, n, inc
do i = 1, n, inc
#ifndef MIRROR
if (mod(ij,nproc) .eq. me) then
#else
if (mod(ij,lprocs) .eq. iproc) then
#endif
ilo = i
ihi = min(i+inc, n)
jlo = j
jhi = min(j+inc, n)
call ga_put(g_a, ilo, ihi, jlo, jhi, a(ilo, jlo), n)
endif
ij = ij + 1
enddo
enddo
call ga_sync()
c
c All nodes check all of a
c
call ga_get(g_a, 1, n, 1, n, b, n)
c
do i = 1, n
do j = 1, n
if (b(i,j) .ne. a(i,j)) then
write(6,*) ' put ', me, i, j, a(i,j),b(i,j)
call ffflush(6)
call ga_error('... exiting ',0)
endif
enddo
enddo
if (me.eq.0) then
write(6,*)
write(6,*) ' ga_put is OK'
write(6,*)
endif
call ga_sync()
c
c Now check nloop random gets from each node
c
if (me .eq. 0) then
write(6,5) nloop
5 format(/'> Checking random get (',i5,' calls)...')
call ffflush(6)
endif
call ga_sync()
c
nwords = 0
c
crap = drand(ga_nodeid()*51 +1) !Different seed for each proc
do loop = 1, nloop
ilo = iran(loop)
ihi = iran(loop)
if (ihi.lt. ilo) then
itmp = ihi
ihi = ilo
ilo = itmp
endif
jlo = iran(loop)
jhi = iran(loop)
if (jhi.lt. jlo) then
itmp = jhi
jhi = jlo
jlo = itmp
endif
c
nwords = nwords + (ihi-ilo+1)*(jhi-jlo+1)
c
call ga_get(g_a, ilo, ihi, jlo, jhi, b(ilo, jlo), n)
if (me .eq. 0 .and. mod(loop-1, max(1,nloop/20)).eq.0) then
write(6,1) loop, me, ilo, ihi, jlo, jhi, nwords
1 format(' call ',i5, ' node ',i2,' checking get ',4i4,
$ ' total ',d9.2)
call ffflush(6)
endif
do j = jlo, jhi
do i = ilo, ihi
if (b(i,j) .ne. a(i,j)) then
write(6,*)'error:', i, j, b(i,j), a(i,j)
call ga_error('... exiting ',0)
endif
enddo
enddo
c
enddo
if (me.eq.0) then
write(6,*)
write(6,*) ' ga_get is OK'
write(6,*)
call ffflush(6)
endif
call ga_sync()
c
c Check the ga_copy function
c
if (me .eq. 0) then
write(6,*)
write(6,*)'> Checking copy'
write(6,*)
call ffflush(6)
endif
call ga_sync()
#ifndef MIRROR
if(me.eq.0) call ga_put(g_a, 1, n, 1, n, a, n)
#else
if(iproc.eq.0) call ga_put(g_a, 1, n, 1, n, a, n)
#endif
call ga_copy(g_a, g_b)
call ga_get(g_b, 1, n, 1, n, b, n)
do j = 1, n
do i = 1, n
if (b(i,j) .ne. a(i,j)) then
write(6,*) ' copy ', me, i, j, a(i,j), b(i,j)
call ga_error('... exiting ',0)
endif
enddo
enddo
if (me.eq.0) then
write(6,*)
write(6,*) ' copy is OK '
write(6,*)
endif
c
c Check scatter&gather
c
call ga_sync()
if (me .eq. 0) then
write(6,*) '> Checking scatter/gather (might be slow)... '
call ffflush(6)
if(me.eq.0) call ga_put(g_a, 1, n, 1, n, a, n)
endif
call ga_sync()
c
crap = drand(ga_nodeid()*51 +1) !Different seed for each proc
do j = 1, 10
call ga_sync()
#ifndef MIRROR
itmp = iran(nproc)-1
if(me.eq.itmp) then
#else
itmp = iran(lprocs)-1
if(iproc.eq.itmp) then
#endif
do loop = 1,m
ilo = iran(n)
jlo = iran(n)
iv(loop) = ilo
jv(loop) = jlo
enddo
call ga_gather(g_a, v, iv, jv, m)
do loop = 1,m
ilo= iv(loop)
jlo= jv(loop)
call ga_get(g_a,ilo,ilo,jlo,jlo,v(loop),1)
if(v(loop) .ne. a(ilo,jlo))then
write(6,*)me,' gather ', ilo,',',jlo,',', a(ilo,jlo)
& ,' ',v(loop)
call ffflush(6)
call ga_error('... exiting ',0)
endif
enddo
endif
enddo
c
if (me.eq.0) then
write(6,*)
write(6,*) ' gather is OK'
write(6,*)
call ffflush(6)
endif
c
do j = 1,10
call ga_sync()
#ifndef MIRROR
if(me.eq.iran(ga_nnodes())-1) then
#else
if(me.eq.iran(lprocs)-1) then
#endif
do loop = 1,m
ilo = iran(n)
jlo = iran(n)
iv(loop) = ilo
jv(loop) = jlo
v(loop) = (1d0,-1d0) *(ilo+jlo)
enddo
call ga_scatter(g_a, v, iv, jv, m)
do loop = 1,m
ilo= iv(loop)
jlo= jv(loop)
call ga_get(g_a,ilo,ilo,jlo,jlo,w(loop),1)
if(w(loop) .ne. (1d0,-1d0) *(ilo+jlo) )then
write(6,*)me,' scatter ', ilo,',',jlo,',',w(loop)
& ,' ', (1d0,-1d0) *(ilo+jlo)
call ffflush(6)
endif
enddo
endif
call ga_sync()
enddo
c
if (me.eq.0) then
write(6,*)
write(6,*) ' scatter is OK'
write(6,*)
endif
c
c Delete the global arrays
c
status = ga_destroy(g_b)
status = ga_destroy(g_a)
c
istatus = nga_deregister_type(scpl_type)
c
end
c-----------------------------------------------------------------
subroutine check_complex()
implicit none
#include "mafdecls.fh"
#include "global.fh"
#include "testutil.fh"
c
integer n,m
parameter (n = 60)
parameter (m = 2*n)
double complex a(n,n), b(n,n), v(m),w(m)
#ifdef MIRROR
integer ndim, dims(2), chunk(2), p_mirror
#else
# ifdef NEW_API
integer ndim, dims(2), chunk(2), p_mirror
# endif
#endif
integer iv(m), jv(m)
logical status
integer g_a, g_b
integer iran, i,j, loop,nloop,maxloop, ilo, ihi, jlo, jhi, itmp
integer nproc, me, int, ij, inc, ii, jj, nnodes
parameter (maxloop = 100)
integer maxproc
parameter (maxproc = 4096)
double precision crap, real
double precision nwords
double complex x, sum1, sum2, factor
integer lprocs, inode, iproc, lproc
integer dcpl_type, istatus
#ifdef USE_RESTRICTED
integer num_rstrctd
integer rstrctd_list(maxproc/2)
#endif
#ifdef BLOCK_CYCLIC
integer block_size(2), proc_grid(2)
#endif
intrinsic int
iran(i) = int(drand(0)*real(i)) + 1
c
nproc = ga_nnodes()
me = ga_nodeid()
inode = ga_cluster_nodeid()
lprocs = ga_cluster_nprocs(inode)
nnodes = ga_cluster_nnodes()
iproc = mod(me,lprocs)
nloop = Min(maxloop,n)
#ifdef USE_RESTRICTED
num_rstrctd = nproc/2
if (num_rstrctd.eq.0) num_rstrctd = 1
do i = 1, num_rstrctd
rstrctd_list(i) = (num_rstrctd/2) + i-1
end do
#endif
#ifdef BLOCK_CYCLIC
block_size(1) = 32
block_size(2) = 32
#ifdef USE_SCALAPACK_DISTR
if (mod(nproc,2).ne.0)
+ call ga_error("Available procs must be divisible by 2",0)
proc_grid(1) = 2
proc_grid(2) = nproc/2
#endif
#endif
c
c a() is a local copy of what the global array should start as
c
do j = 1, n
do i = 1, n
#ifndef MIRROR
a(i,j) = cmplx(dble(i-1), dble((j-1)*n))
#else
a(i,j) = cmplx(dble(inode),0.0d00)
+ + cmplx(dble(i-1), dble((j-1)*n))
#endif
b(i,j) = cmplx(-1d0,1d0)
enddo
enddo
c
c Create type
c
dcpl_type = nga_register_type(16)
c
c Create a global array
c
c print *,ga_nodeid(), ' creating array'
call ffflush(6)
c call setdbg(1)
#ifdef NEW_API
ndim = 2
dims(1) = n
dims(2) = n
g_a = ga_create_handle()
call ga_set_data(g_a,ndim,dims,dcpl_type)
call ga_set_array_name(g_a,'a')
#ifdef USE_RESTRICTED
call ga_set_restricted(g_a, rstrctd_list, num_rstrctd)
#endif
#ifdef BLOCK_CYCLIC
#ifdef USE_SCALAPACK_DISTR
call ga_set_block_cyclic_proc_grid(g_a,block_size,proc_grid)
#else
call ga_set_block_cyclic(g_a,block_size)
#endif
#endif
# ifdef MIRROR
p_mirror = ga_pgroup_get_mirror()
call ga_set_pgroup(g_a,p_mirror)
# endif
status = ga_allocate(g_a)
#else
# ifndef MIRROR
status = ga_create(dcpl_type, n, n, 'a', 0, 0, g_a)
# else
ndim = 2
dims(1) = n
dims(2) = n
chunk(1) = 0
chunk(2) = 0
p_mirror = ga_pgroup_get_mirror()
status = nga_create_config(dcpl_type, ndim, dims, 'a', chunk,
+ p_mirror, g_a)
# endif
#endif
if (.not. status) then
write(6,*) ' ga_create failed'
call ga_error('... exiting ',0)
endif
#ifdef NEW_API
g_b = ga_create_handle()
call ga_set_data(g_b,ndim,dims,dcpl_type)
call ga_set_array_name(g_b,'b')
#ifdef BLOCK_CYCLIC
#ifdef USE_SCALAPACK_DISTR
call ga_set_block_cyclic_proc_grid(g_b,block_size,proc_grid)
#else
call ga_set_block_cyclic(g_b,block_size)
#endif
#endif
# ifdef MIRROR
call ga_set_pgroup(g_b,p_mirror)
# endif
if (.not.ga_allocate(g_b)) then
#else
# ifndef MIRROR
if (.not. ga_create(dcpl_type, n, n, 'b', 0, 0, g_b)) then
# else
if (.not. nga_create_config(dcpl_type, ndim, dims, 'b', chunk,
_ p_mirror, g_b)) then
# endif
#endif
call ga_error('ga_create failed for second array ',0)
endif
#ifndef MIRROR
call ga_distribution(g_a,me,ilo, ihi, jlo, jhi)
#else
lproc = me - ga_cluster_procid(inode,0)
call ga_distribution(g_a, lproc, ilo, ihi, jlo, jhi)
#endif
call ga_sync()
c
c Zero the array
c
if (me .eq. 0) then
write(6,21)
21 format('> Checking zero ... ')
call ffflush(6)
endif
call ga_zero(g_a)
c
c Check that it is indeed zero
c
call ga_get(g_a, 1, n, 1, n, b, n)
call ga_sync()
do i = 1, n
do j = 1, n
if(b(i,j).ne.(0d0,0d0)) then
write(6,*) me,' zero ', i, j, b(i,j)
call ffflush(6)
c call ga_error('... exiting ',0)
endif
enddo
enddo
if (me.eq.0) then
write(6,*)
write(6,*) ' ga_zero is OK'
write(6,*)
endif
call ga_sync()
c
c Each node fills in disjoint sections of the array
c
if (me .eq. 0) then
write(6,2)
2 format('> Checking disjoint put ... ')
call ffflush(6)
endif
call ga_sync()
c
inc = (n-1)/20 + 1
ij = 0
do j = 1, n, inc
do i = 1, n, inc
#ifndef MIRROR
if (mod(ij,nproc) .eq. me) then
#else
if (mod(ij,lprocs) .eq. iproc) then
#endif
ilo = i
ihi = min(i+inc, n)
jlo = j
jhi = min(j+inc, n)
call ga_put(g_a, ilo, ihi, jlo, jhi, a(ilo, jlo), n)
endif
ij = ij + 1
enddo
enddo
call ga_sync()
c
c All nodes check all of a
c
call util_qfill(n*n, (0d0,0d0), b, 1)
call ga_get(g_a, 1, n, 1, n, b, n)
c
do i = 1, n
do j = 1, n
if (b(i,j) .ne. a(i,j)) then
write(6,*) ' put ', me, i, j, a(i,j),b(i,j)
call ffflush(6)
call ga_error('... exiting ',0)
endif
enddo
enddo
if (me.eq.0) then
write(6,*)
write(6,*) ' ga_put is OK'
write(6,*)
endif
call ga_sync()
c
c Now check nloop random gets from each node
c
if (me .eq. 0) then
write(6,5) nloop
5 format('> Checking random get (',i5,' calls)...')
call ffflush(6)
endif
call ga_sync()
c
nwords = 0
c
crap = drand(ga_nodeid()*51 +1) !Different seed for each proc
do loop = 1, nloop
ilo = iran(loop)
ihi = iran(loop)
if (ihi.lt. ilo) then
itmp = ihi
ihi = ilo
ilo = itmp
endif
jlo = iran(loop)
jhi = iran(loop)
if (jhi.lt. jlo) then
itmp = jhi
jhi = jlo
jlo = itmp
endif
c
nwords = nwords + (ihi-ilo+1)*(jhi-jlo+1)
c
call util_qfill(n*n, (0.0d0,0d0), b, 1)
call ga_get(g_a, ilo, ihi, jlo, jhi, b(ilo, jlo), n)
if (me .eq. 0 .and. mod(loop-1, max(1,nloop/20)).eq.0) then
write(6,1) loop, me, ilo, ihi, jlo, jhi, nwords
1 format(' call ',i5, ' node ',i2,' checking get ',4i4,
$ ' total ',d9.2)
call ffflush(6)
endif
do j = jlo, jhi
do i = ilo, ihi
if (b(i,j) .ne. a(i,j)) then
write(6,*)'error:', i, j, b(i,j), a(i,j)
call ga_error('... exiting ',0)
endif
enddo
enddo
c
enddo
if (me.eq.0) then
write(6,*)
write(6,*) ' ga_get is OK'
write(6,*)
call ffflush(6)
endif
call ga_sync()
c
c Check the ga_copy function
c
if (me .eq. 0) then
write(6,*)
write(6,*)'> Checking copy'
write(6,*)
call ffflush(6)
endif
call ga_sync()
#ifndef MIRROR
if(me.eq.0) call ga_put(g_a, 1, n, 1, n, a, n)
#else
if(iproc.eq.0) call ga_put(g_a, 1, n, 1, n, a, n)
#endif
call ga_copy(g_a, g_b)
call ga_get(g_b, 1, n, 1, n, b, n)
do j = 1, n
do i = 1, n
if (b(i,j) .ne. a(i,j)) then
write(6,*) ' copy ', me, i, j, a(i,j), b(i,j)
call ga_error('... exiting ',0)
endif
enddo
enddo
if (me.eq.0) then
write(6,*)
write(6,*) ' copy is OK '
write(6,*)
endif
c
c Check scatter&gather
c
call ga_sync()
if (me .eq. 0) then
write(6,*) '> Checking scatter/gather (might be slow)... '
call ffflush(6)
if(me.eq.0) call ga_put(g_a, 1, n, 1, n, a, n)
endif
call ga_sync()
c
crap = drand(ga_nodeid()*51 +1) !Different seed for each proc
do j = 1, 10
call ga_sync()
#ifndef MIRROR
itmp = iran(nproc)-1
if(me.eq.itmp) then
#else
itmp = iran(lprocs)-1
if(iproc.eq.itmp) then
#endif
do loop = 1,m
ilo = iran(n)
jlo = iran(n)
iv(loop) = ilo
jv(loop) = jlo
enddo
call ga_gather(g_a, v, iv, jv, m)
do loop = 1,m
ilo= iv(loop)
jlo= jv(loop)
call ga_get(g_a,ilo,ilo,jlo,jlo,v(loop),1)
if(v(loop) .ne. a(ilo,jlo))then
write(6,*)me,' gather ', ilo,',',jlo,',', a(ilo,jlo)
& ,' ',v(loop)
call ffflush(6)
call ga_error('... exiting ',0)
endif
enddo
endif
enddo
c
if (me.eq.0) then
write(6,*)
write(6,*) ' gather is OK'
write(6,*)
call ffflush(6)
endif
c
do j = 1,10
call ga_sync()
#ifndef MIRROR
if(me.eq.iran(ga_nnodes())-1) then
#else
if(me.eq.iran(lprocs)-1) then
#endif
do loop = 1,m
ilo = iran(n)
jlo = iran(n)
iv(loop) = ilo
jv(loop) = jlo
v(loop) = (1d0,-1d0) *(ilo+jlo)
enddo
call ga_scatter(g_a, v, iv, jv, m)
do loop = 1,m
ilo= iv(loop)
jlo= jv(loop)
call ga_get(g_a,ilo,ilo,jlo,jlo,w(loop),1)
if(w(loop) .ne. (1d0,-1d0) *(ilo+jlo) )then
write(6,*)me,' scatter ', ilo,',',jlo,',',w(loop)
& ,' ', (1d0,-1d0) *(ilo+jlo)
call ffflush(6)
endif
enddo
endif
call ga_sync()
enddo
c
if (me.eq.0) then
write(6,*)
write(6,*) ' scatter is OK'
write(6,*)
endif
c
c Delete the global arrays
c
status = ga_destroy(g_b)
status = ga_destroy(g_a)
c
istatus = nga_deregister_type(dcpl_type)
c
end
subroutine util_qfill(n,val,a,ia)
implicit none
double complex a(*), val
integer n, ia, i
c
c initialise double complex array to scalar value
c
if (ia.eq.1) then
do 10 i = 1, n
a(i) = val
10 continue
else
do 20 i = 1,(n-1)*ia+1,ia
a(i) = val
20 continue
endif
c
end
|
