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
|
/*
* This file is a part of TiledArray.
* Copyright (C) 2018 Virginia Tech
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation, either version 3 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program. If not, see <http://www.gnu.org/licenses/>.
*
* Created by Chong Peng on 7/19/18.
*
*/
#include <cuda_runtime.h>
#include <madness/world/safempi.h>
#include <assert.h>
#include <iostream>
#include <stdexcept>
/**
* Test CUDA-aware MPI
*/
const std::size_t N = 100;
int main(int argc, char* argv[]) {
MPI_Init(&argc, &argv);
int mpi_size = -1;
MPI_Comm_size(MPI_COMM_WORLD, &mpi_size);
if (mpi_size != 2) {
throw std::runtime_error("Must run this program with 2 mpi processes");
}
int mpi_rank = -1;
MPI_Comm_rank(MPI_COMM_WORLD, &mpi_rank);
/*
* Test MPI Send & Recv with GPU memory
*/
{
// allocate host vector
int* vector_host = (int*)std::malloc(sizeof(int) * N);
cudaError_t cuda_error;
int* vector_device;
cuda_error = cudaMalloc(&vector_device, sizeof(int) * N);
assert(cuda_error == cudaSuccess);
// initialize data on node 0
if (mpi_rank == 0) {
for (std::size_t i = 0; i < N; i++) {
vector_host[i] = i;
}
cudaMemcpy(vector_device, vector_host, sizeof(int) * N,
cudaMemcpyHostToDevice);
// MPI SEND to NODE 1
MPI_Send(vector_device, N, MPI_INT, 1, 100, MPI_COMM_WORLD);
}
// receive data on node 1
else {
MPI_Status status;
MPI_Recv(vector_device, N, MPI_INT, 0, 100, MPI_COMM_WORLD, &status);
cudaMemcpy(vector_host, vector_device, sizeof(int) * N,
cudaMemcpyDeviceToHost);
// verify the data
for (std::size_t i = 0; i < N; i++) {
assert(vector_host[i] == i);
}
}
free(vector_host);
cudaFree(vector_device);
if (mpi_rank == 0) {
std::cout << "MPI Send & Recv SUCCESS on GPU memory.\n";
}
}
/**
* Test MPI
*/
{
// allocate host vector
int* vector_um;
cudaError_t cuda_error;
cuda_error = cudaMallocManaged(&vector_um, sizeof(int) * N);
assert(cuda_error == cudaSuccess);
// initialize data on node 0
if (mpi_rank == 0) {
for (std::size_t i = 0; i < N; i++) {
vector_um[i] = i;
}
// MPI SEND to NODE 1
MPI_Send(vector_um, N, MPI_INT, 1, 100, MPI_COMM_WORLD);
}
// receive data on node 1
else {
MPI_Status status;
MPI_Recv(vector_um, N, MPI_INT, 0, 100, MPI_COMM_WORLD, &status);
// verify the data
for (std::size_t i = 0; i < N; i++) {
assert(vector_um[i] == i);
}
}
cudaFree(vector_um);
if (mpi_rank == 0) {
std::cout << "MPI Send & Recv SUCCESS on CUDA Unified memory.\n";
}
}
MPI_Finalize();
return 0;
}
|
