File: put_base.c

package info (click to toggle)
mpich 4.0.2-3
  • links: PTS, VCS
  • area: main
  • in suites: bookworm
  • size: 423,384 kB
  • sloc: ansic: 1,088,434; cpp: 71,364; javascript: 40,763; f90: 22,829; sh: 17,463; perl: 14,773; xml: 14,418; python: 10,265; makefile: 9,246; fortran: 8,008; java: 4,355; asm: 324; ruby: 176; lisp: 19; php: 8; sed: 4
file content (148 lines) | stat: -rw-r--r-- 4,683 bytes parent folder | download | duplicates (4)
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
/*
 * Copyright (C) by Argonne National Laboratory
 *     See COPYRIGHT in top-level directory
 */

/* One-Sided MPI 2-D Strided Put Test
 *
 * Author: James Dinan <dinan@mcs.anl.gov>
 * Date  : March, 2011
 *
 * This code performs N strided put operations into a 2d patch of a shared
 * array.  The array has dimensions [X, Y] and the subarray has dimensions
 * [SUB_X, SUB_Y] and begins at index [0, 0].  The input and output buffers are
 * specified using an MPI datatype.
 *
 * This test generates a datatype that is relative to an arbitrary base address
 * in memory and tests the RMA implementation's ability to perform the correct
 * transfer.
 */

#include <stdio.h>
#include <stdlib.h>
#include <stdint.h>
#include <mpi.h>
#include "mpitest.h"
#include "squelch.h"

#define XDIM 1024
#define YDIM 1024
#define SUB_XDIM 1024
#define SUB_YDIM 1024
#define ITERATIONS 10

int main(int argc, char **argv)
{
    int i, j, rank, nranks, peer, bufsize, errs;
    double *win_buf, *src_buf, *dst_buf;
    MPI_Win buf_win;

    MTest_Init(&argc, &argv);

    MPI_Comm_rank(MPI_COMM_WORLD, &rank);
    MPI_Comm_size(MPI_COMM_WORLD, &nranks);

    bufsize = XDIM * YDIM * sizeof(double);
    MPI_Alloc_mem(bufsize, MPI_INFO_NULL, &win_buf);
    /* Alloc_mem is not required for the origin buffers for RMA operations -
     * just for the Win_create memory */
    MPI_Alloc_mem(bufsize, MPI_INFO_NULL, &src_buf);
    MPI_Alloc_mem(bufsize, MPI_INFO_NULL, &dst_buf);

    for (i = 0; i < XDIM * YDIM; i++) {
        *(win_buf + i) = 1.0 + rank;
        *(src_buf + i) = 1.0 + rank;
    }

    MPI_Win_create(win_buf, bufsize, 1, MPI_INFO_NULL, MPI_COMM_WORLD, &buf_win);

    peer = (rank + 1) % nranks;

    /* Perform ITERATIONS strided put operations */

    for (i = 0; i < ITERATIONS; i++) {
        MPI_Aint idx_loc[SUB_YDIM];
        int idx_rem[SUB_YDIM];
        int blk_len[SUB_YDIM];
        MPI_Datatype src_type, dst_type;

        void *base_ptr = dst_buf;
        MPI_Aint base_int;

        MPI_Get_address(base_ptr, &base_int);

        for (j = 0; j < SUB_YDIM; j++) {
            MPI_Get_address(&src_buf[j * XDIM], &idx_loc[j]);
            idx_loc[j] = idx_loc[j] - base_int;
            idx_rem[j] = j * XDIM * sizeof(double);
            blk_len[j] = SUB_XDIM * sizeof(double);
        }

        MPI_Type_create_hindexed(SUB_YDIM, blk_len, idx_loc, MPI_BYTE, &src_type);
        MPI_Type_create_indexed_block(SUB_YDIM, SUB_XDIM * sizeof(double), idx_rem, MPI_BYTE,
                                      &dst_type);

        MPI_Type_commit(&src_type);
        MPI_Type_commit(&dst_type);

        MPI_Win_lock(MPI_LOCK_EXCLUSIVE, peer, 0, buf_win);
        MPI_Put(base_ptr, 1, src_type, peer, 0, 1, dst_type, buf_win);
        MPI_Win_unlock(peer, buf_win);

        MPI_Type_free(&src_type);
        MPI_Type_free(&dst_type);
    }

    MPI_Barrier(MPI_COMM_WORLD);

    /* Verify that the results are correct */

    MPI_Win_lock(MPI_LOCK_EXCLUSIVE, rank, 0, buf_win);
    errs = 0;
    for (i = 0; i < SUB_XDIM; i++) {
        for (j = 0; j < SUB_YDIM; j++) {
            const double actual = *(win_buf + i + j * XDIM);
            const double expected = (1.0 + ((rank + nranks - 1) % nranks));
            if (actual - expected > 1e-10) {
                SQUELCH(printf("%d: Data validation failed at [%d, %d] expected=%f actual=%f\n",
                               rank, j, i, expected, actual););
                errs++;
                fflush(stdout);
            }
        }
    }
    for (i = SUB_XDIM; i < XDIM; i++) {
        for (j = 0; j < SUB_YDIM; j++) {
            const double actual = *(win_buf + i + j * XDIM);
            const double expected = 1.0 + rank;
            if (actual - expected > 1e-10) {
                SQUELCH(printf("%d: Data validation failed at [%d, %d] expected=%f actual=%f\n",
                               rank, j, i, expected, actual););
                errs++;
                fflush(stdout);
            }
        }
    }
    for (i = 0; i < XDIM; i++) {
        for (j = SUB_YDIM; j < YDIM; j++) {
            const double actual = *(win_buf + i + j * XDIM);
            const double expected = 1.0 + rank;
            if (actual - expected > 1e-10) {
                SQUELCH(printf("%d: Data validation failed at [%d, %d] expected=%f actual=%f\n",
                               rank, j, i, expected, actual););
                errs++;
                fflush(stdout);
            }
        }
    }
    MPI_Win_unlock(rank, buf_win);

    MPI_Win_free(&buf_win);
    MPI_Free_mem(win_buf);
    MPI_Free_mem(src_buf);
    MPI_Free_mem(dst_buf);

    MTest_Finalize(errs);

    return MTestReturnValue(errs);
}