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
|
/*
* Copyright (C) by Argonne National Laboratory
* See COPYRIGHT in top-level directory
*/
#include <mpi.h>
#include <stdio.h>
#include <assert.h>
#include "mpitest.h"
#define NSTEPS 100
#define N 1000
#define M 10
/* This is example 11.21 from the MPI 3.0 spec:
*
* The following example shows how request-based operations can be used to
* overlap communication with computation. Each process fetches, processes,
* and writes the result for NSTEPS chunks of data. Instead of a single
* buffer, M local buffers are used to allow up to M communication operations
* to overlap with computation.
*/
/* Use a global variable to inhibit compiler optimizations in the compute
* function. */
double junk = 0.0;
static void compute(int step, double *data)
{
int i;
for (i = 0; i < N; i++)
junk += data[i] * (double) step;
}
int main(int argc, char *argv[])
{
int i, rank, nproc;
int errors = 0, all_errors = 0;
MPI_Win win;
MPI_Request put_req[M] = { MPI_REQUEST_NULL };
MPI_Request get_req;
double *baseptr;
double data[M][N]; /* M buffers of length N */
MPI_Info win_info;
MTest_Init(&argc, &argv);
MPI_Comm_rank(MPI_COMM_WORLD, &rank);
MPI_Comm_size(MPI_COMM_WORLD, &nproc);
assert(M < NSTEPS);
MPI_Info_create(&win_info);
#ifdef USE_WIN_ALLOC_SHM
MPI_Info_set(win_info, "alloc_shm", "true");
#else
MPI_Info_set(win_info, "alloc_shm", "false");
#endif
MPI_Win_allocate(NSTEPS * N * sizeof(double), sizeof(double), win_info,
MPI_COMM_WORLD, &baseptr, &win);
MPI_Win_lock_all(0, win);
for (i = 0; i < NSTEPS; i++) {
int target = (rank + 1) % nproc;
int j;
/* Find a free put request */
if (i < M) {
j = i;
} else {
MPI_Waitany(M, put_req, &j, MPI_STATUS_IGNORE);
}
MPI_Rget(data[j], N, MPI_DOUBLE, target, i * N, N, MPI_DOUBLE, win, &get_req);
MPI_Wait(&get_req, MPI_STATUS_IGNORE);
compute(i, data[j]);
MPI_Rput(data[j], N, MPI_DOUBLE, target, i * N, N, MPI_DOUBLE, win, &put_req[j]);
}
MPI_Waitall(M, put_req, MPI_STATUSES_IGNORE);
MPI_Win_unlock_all(win);
MPI_Win_free(&win);
MPI_Info_free(&win_info);
MTest_Finalize(errors);
return MTestReturnValue(all_errors);
}
|
