我开发了一个MPI程序，它可以循环发送和接收消息。消息传递由进程0启动：它发送到进程1，并从进程p-1接收。其他进程从其左邻居收到并发送给他们的右邻居。请注意，（p-1）-s右邻居为0.
以下是我的代码：
#include <stdio.h>
#include "mpi.h"
#include <stdlib.h>
/* cycle
In this program a message is sent around a circle of processes 0 - (p-1), ]
where p-1 sends to 0.
*/
int main(int argc, char **argv)
{
double startwtime, endwtime;
float elapsed_time, bandwidth;
int my_id, next_id; /* process id-s */
int p; /* number of processes */
char* message; /* storage for the message */
int i, k, max_msgs, msg_size, v;
MPI_Status status; /* return status for receive */
MPI_Init( &argc, &argv );
MPI_Comm_rank( MPI_COMM_WORLD, &my_id );
MPI_Comm_size( MPI_COMM_WORLD, &p );
if (argc < 3)
{
fprintf (stderr, "need msg count and msg size as params\n");
goto EXIT;
}
if ((sscanf (argv[1], "%d", &max_msgs) < 1) ||
(sscanf (argv[2], "%d", &msg_size) < 1))
{
fprintf (stderr, "need msg count and msg size as params\n");
goto EXIT;
}
message = (char*)malloc (msg_size);
char yahoo = 'C';
message =&yahoo;
if (argc > 3) v=1; else v=0; /*are we in verbose mode*/
/* don't start timer until everybody is ok */
MPI_Barrier(MPI_COMM_WORLD);
int t=0;
if( my_id == 0 ) {
startwtime = MPI_Wtime();
// do max_msgs times:
// send message of size msg_size chars to process 1
// receive message of size msg_size chars from process p-1
while(t<max_msgs) {
MPI_Send((char *) message, msg_size, MPI_CHAR, my_id+1, 0, MPI_COMM_WORLD);
printf("Message ,source and destination ids:%c,%d,%d \n ",*message, 0 ,my_id+1);
fflush(stdout);
MPI_Recv((char *) message, msg_size, MPI_CHAR, my_id-1, 0, MPI_COMM_WORLD, &status);
printf("Message,source and destination ids:%c,%d,%d \n",*message, my_id-1,
0);
fflush(stdout);
t++;
}
MPI_Barrier(MPI_COMM_WORLD);
endwtime = MPI_Wtime();
elapsed_time = endwtime-startwtime;
bandwidth = 2.0 * max_msgs * msg_size / (elapsed_time);
printf("Number, size of messages: %3d , %3d \n", max_msgs, msg_size);
fflush(stdout);
printf("Wallclock time = %f seconds\n", elapsed_time );
fflush(stdout);
printf("Bandwidth = %f bytes per second\n", bandwidth);
fflush(stdout);
} else if( my_id == p-1 ) {
// do max_msgs times:
// receive message of size msg_size from process to the left
// send message of size msg_size to process to the right (p-1 sends to 0)
while(t<max_msgs) {
MPI_Send((char *) message, msg_size, MPI_CHAR, 0, 0, MPI_COMM_WORLD);
MPI_Recv((char *) message, msg_size, MPI_CHAR, my_id-1, 0, MPI_COMM_WORLD, &status);
t++;
}
} else {
while(t<max_msgs) {
MPI_Send((char *) message, msg_size, MPI_CHAR, my_id+1, 0, MPI_COMM_WORLD);
MPI_Recv((char *) message, msg_size, MPI_CHAR, my_id-1, 0, MPI_COMM_WORLD, &status);
t++;
}
}
MPI_Barrier(MPI_COMM_WORLD);
EXIT:
MPI_Finalize();
return 0;
}
我无法获得正确的源和目标ID。运行1/2次后的程序进入无限循环。这可能有什么问题？
答案 0 :(得分：3)
您的代码的第一个问题是它有可能与某些MPI实现发生死锁。查看these注意事项了解详细信息。
不考虑死锁问题，此代码存在两个问题。第一个涉及
if( my_id == 0 ) {
条件。
MPI_Recv((char *) message, msg_size, MPI_CHAR, my_id-1, 0, MPI_COMM_WORLD, &status);
设置从PE -1接收消息的源，该消息不起作用。您想收到p-1。
MPI_Barrier来电也是不对的。我必须考虑一下......得到它。您可以在MPI_Barrier分支中调用my_id==0，但在其他分支之外调用PE 0 calls barrier
all conditionals finish
PE 1 to p-1 calls barrier
PE 0 calls barrier again
frozen program
。这将产生以下内容
#include <stdio.h>
#include "mpi.h"
#include <stdlib.h>
/* cycle
In this program a message is sent around a circle of processes 0 - (p-1), ]
where p-1 sends to 0.
*/
int main(int argc, char **argv)
{
double startwtime, endwtime;
float elapsed_time, bandwidth;
int my_id, next_id; /* process id-s */
int p; /* number of processes */
char* message; /* storage for the message */
int i, k, max_msgs, msg_size, v;
MPI_Status status; /* return status for receive */
MPI_Init( &argc, &argv );
MPI_Comm_rank( MPI_COMM_WORLD, &my_id );
MPI_Comm_size( MPI_COMM_WORLD, &p );
if (argc < 3)
{
fprintf (stderr, "need msg count and msg size as params\n");
goto EXIT;
}
if ((sscanf (argv[1], "%d", &max_msgs) < 1) ||
(sscanf (argv[2], "%d", &msg_size) < 1))
{
fprintf (stderr, "need msg count and msg size as params\n");
goto EXIT;
}
fprintf(stdout, "my_id: %d, max_msgs: %d\n", my_id, max_msgs);
message = (char*)malloc (msg_size);
char yahoo = 'C';
message =&yahoo;
if (argc > 3) v=1; else v=0; /*are we in verbose mode*/
/* don't start timer until everybody is ok */
MPI_Barrier(MPI_COMM_WORLD);
fprintf(stdout, "my_id: %d, barrier 1\n", my_id);
int t=0;
if( my_id == 0 ) {
startwtime = MPI_Wtime();
// do max_msgs times:
// send message of size msg_size chars to process 1
// receive message of size msg_size chars from process p-1
while(t<max_msgs) {
printf("PE %d about to send\n ", my_id);
fflush(stdout);
MPI_Send((char *) message, msg_size, MPI_CHAR, my_id+1, 0, MPI_COMM_WORLD);
printf("PE %d completed send\n ", my_id);
fflush(stdout);
printf("PE %d about to recv\n ", my_id);
fflush(stdout);
MPI_Recv((char *) message, msg_size, MPI_CHAR, p-1, 0, MPI_COMM_WORLD, &status);
printf("PE %d completed recv\n ", my_id);
fflush(stdout);
t++;
}
MPI_Barrier(MPI_COMM_WORLD);
endwtime = MPI_Wtime();
elapsed_time = endwtime-startwtime;
bandwidth = 2.0 * max_msgs * msg_size / (elapsed_time);
printf("Number, size of messages: %3d , %3d \n", max_msgs, msg_size);
fflush(stdout);
printf("Wallclock time = %f seconds\n", elapsed_time );
fflush(stdout);
printf("Bandwidth = %f bytes per second\n", bandwidth);
fflush(stdout);
} else if( my_id == p-1 ) {
// do max_msgs times:
// receive message of size msg_size from process to the left
// send message of size msg_size to process to the right (p-1 sends to 0)
while(t<max_msgs) {
printf("PE %d about to send\n ", my_id);
fflush(stdout);
MPI_Send((char *) message, msg_size, MPI_CHAR, 0, 0, MPI_COMM_WORLD);
printf("PE %d completed send\n ", my_id);
fflush(stdout);
printf("PE %d about to recv\n ", my_id);
fflush(stdout);
MPI_Recv((char *) message, msg_size, MPI_CHAR, my_id-1, 0, MPI_COMM_WORLD, &status);
printf("PE %d completed recv\n ", my_id);
fflush(stdout);
t++;
}
MPI_Barrier(MPI_COMM_WORLD);
} else {
while(t<max_msgs) {
printf("*PE %d about to send\n ", my_id);
fflush(stdout);
MPI_Send((char *) message, msg_size, MPI_CHAR, my_id+1, 0, MPI_COMM_WORLD);
printf("*PE %d completed send\n ", my_id);
fflush(stdout);
printf("*PE %d about to recv\n ", my_id);
fflush(stdout);
MPI_Recv((char *) message, msg_size, MPI_CHAR, my_id-1, 0, MPI_COMM_WORLD, &status);
printf("*PE %d completed recv\n ", my_id);
fflush(stdout);
t++;
}
MPI_Barrier(MPI_COMM_WORLD);
}
EXIT:
MPI_Finalize();
return 0;
}
在所有条件中放置一个障碍，或者没有障碍。
> mpirun -n 3 ./a.out 1 1
my_id: 0, max_msgs: 1
my_id: 1, max_msgs: 1
my_id: 2, max_msgs: 1
my_id: 0, barrier 1
PE 0 about to send
PE 0 completed send
my_id: 1, barrier 1
*PE 1 about to send
my_id: 2, barrier 1
PE 2 about to send
PE 2 completed send
PE 2 about to recv
PE 0 about to recv
PE 0 completed recv
*PE 1 completed send
*PE 1 about to recv
*PE 1 completed recv
PE 2 completed recv
Number, size of messages: 1 , 1
Wallclock time = 0.000050 seconds
Bandwidth = 40136.878906 bytes per second
{{1}}