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
|
/*
* ADIOS is freely available under the terms of the BSD license described
* in the COPYING file in the top level directory of this source distribution.
*
* Copyright (c) 2008 - 2009. UT-BATTELLE, LLC. All rights reserved.
*/
/* ADIOS C Example: write a 2D table as local array from N processors and
* let ADIOS present it at reading as a joined global 2D array
*
* How to run: mpirun -np <N> joined-array
* Output: joined-array.bp
*
*/
/* This example will write out a 2D variable called 'table'. Processes don't
define the offsets of their pieces in a global 2D array, instead they let
ADIOS merge the tables at the slow dimension to present one global 2D table
at read time.
Running on 4 processes the example output is the following:
$ mpirun -np 4 ./joined-array
$ bpls -la joined-array.bp -d table
double table {18, 6} = 0 / 3.5 / null / null
( 0,0) 0 0.1 0.2 0.3 0.4 0.5
( 1,0) 0 0.1 0.2 0.3 0.4 0.5
( 2,0) 0 0.1 0.2 0.3 0.4 0.5
( 3,0) 1 1.1 1.2 1.3 1.4 1.5
( 4,0) 1 1.1 1.2 1.3 1.4 1.5
( 5,0) 1 1.1 1.2 1.3 1.4 1.5
( 6,0) 1 1.1 1.2 1.3 1.4 1.5
( 7,0) 2 2.1 2.2 2.3 2.4 2.5
( 8,0) 2 2.1 2.2 2.3 2.4 2.5
( 9,0) 2 2.1 2.2 2.3 2.4 2.5
(10,0) 2 2.1 2.2 2.3 2.4 2.5
(11,0) 2 2.1 2.2 2.3 2.4 2.5
(12,0) 3 3.1 3.2 3.3 3.4 3.5
(13,0) 3 3.1 3.2 3.3 3.4 3.5
(14,0) 3 3.1 3.2 3.3 3.4 3.5
(15,0) 3 3.1 3.2 3.3 3.4 3.5
(16,0) 3 3.1 3.2 3.3 3.4 3.5
(17,0) 3 3.1 3.2 3.3 3.4 3.5
The details (per process data) are the following:
$ bpls -la joined-array.bp -dD cols
integer cols scalar = 6
step 0: 4 instances available
6 6 6 6
$ bpls -la joined-array.bp -dD rows
integer rows scalar = 3
step 0: 4 instances available
3 4 5 6
$ bpls -la joined-array.bp -dD table
double table {18, 6} = 0 / 3.5 / null / null
step 0:
block 0: [ 0: 2, 0:5] = 0 / 0.5/ N/A / N/A
(0,0) 0 0.1 0.2 0.3 0.4 0.5
(1,0) 0 0.1 0.2 0.3 0.4 0.5
(2,0) 0 0.1 0.2 0.3 0.4 0.5
block 1: [ 3: 6, 0:5] = 1 / 1.5/ N/A / N/A
(0,0) 1 1.1 1.2 1.3 1.4 1.5
(1,0) 1 1.1 1.2 1.3 1.4 1.5
(2,0) 1 1.1 1.2 1.3 1.4 1.5
(3,0) 1 1.1 1.2 1.3 1.4 1.5
block 2: [ 7:11, 0:5] = 2 / 2.5/ N/A / N/A
(0,0) 2 2.1 2.2 2.3 2.4 2.5
(1,0) 2 2.1 2.2 2.3 2.4 2.5
(2,0) 2 2.1 2.2 2.3 2.4 2.5
(3,0) 2 2.1 2.2 2.3 2.4 2.5
(4,0) 2 2.1 2.2 2.3 2.4 2.5
block 3: [12:17, 0:5] = 3 / 3.5/ N/A / N/A
(0,0) 3 3.1 3.2 3.3 3.4 3.5
(1,0) 3 3.1 3.2 3.3 3.4 3.5
(2,0) 3 3.1 3.2 3.3 3.4 3.5
(3,0) 3 3.1 3.2 3.3 3.4 3.5
(4,0) 3 3.1 3.2 3.3 3.4 3.5
(5,0) 3 3.1 3.2 3.3 3.4 3.5
*/
#include <stdio.h>
#include <stdlib.h>
#include <inttypes.h>
#include <string.h>
#include "mpi.h"
#include "adios.h"
#include "adios_types.h"
int main (int argc, char ** argv)
{
char filename[] = "joined-array.bp";
int rank, size, i,j,n;
int ROWS; // different number of rows per process
int COLS = 6;
double *table;
int64_t varid;
MPI_Comm comm = MPI_COMM_WORLD;
MPI_Init (&argc, &argv);
MPI_Comm_rank (comm, &rank);
MPI_Comm_size (comm, &size);
ROWS = rank + 3;
table = (double *) malloc (ROWS * COLS * sizeof(double));
adios_init_noxml (comm);
int64_t m_adios_group;
int64_t m_adios_file;
adios_declare_group (&m_adios_group, "joined", "", adios_stat_default);
adios_select_method (m_adios_group, "MPI", "verbose=3", "");
adios_define_var (m_adios_group, "rows" ,"", adios_integer ,0, 0, 0);
adios_define_var (m_adios_group, "cols" ,"", adios_integer ,0, 0, 0);
varid = adios_define_var (m_adios_group, "table" ,"", adios_double
,"rows,cols", "JoinedDim,cols", "0,0");
adios_open (&m_adios_file, "joined", filename, "w", comm);
n = 0;
for (i = 0; i < ROWS; i++)
{
for (j = 0; j < COLS; j++)
{
table[n] = (double)rank + (double)j/10.0;
n++;
}
}
adios_write(m_adios_file, "rows", &ROWS);
adios_write(m_adios_file, "cols", &COLS);
adios_write(m_adios_file, "table", table);
adios_close (m_adios_file);
MPI_Barrier (comm);
adios_finalize (rank);
free (table);
MPI_Finalize ();
return 0;
}
|
