paralleling a function call using openmp

Question

I'm trying to design a small program in C. I'm using openMP to parallelise parts of the code, Which reads through a file line by line, stores a number of line reads in an array buffer and calls on a function which for now 'waits' for 10 microseconds. Here's my code. I basically want to execute that particular function in parallel. i.e. the buffer's data is dispatched onto a single thread as soon as it fills up, and then second buffer fills up onto a second thread and so on... . This might not be synchronized, and doesn't need to be at this stage.

int count = 0;             //global
char *array_buffer[8];     //global

char process(int lent)
{
    int row,col;
    #pragma omp single
        for(row = 0; row<count; row++)
    {
        usleep(10); 
                //printf("%s", array_buffer[row]);

                //may be executing another function call from here .... ? 

                free(array_buffer[row]);
    }
    return 1;
}


void line(char *line_char)
{

    int lent = strlen(line_char);

    array_buffer[count] = malloc((lent + 1)*sizeof(char));
    strcpy(array_buffer[count], line_char); 

    #pragma omp parallel
    if (count == 8)
    {
    int returning, returned_all = 0;
    returning = process(lent); //function call, waits for 10 microseconds and returns 1 as o/p

    #pragma omp single
    count = 0;

    #pragma omp atomic
    returned_all = returned_all + returning;
    }

    count++;
}



int main(int argc,char **argv)
{

    FILE *fp = fopen(argv[1], "r");
    if(fp == NULL )
    {
        printf("Couldn't open file %s",argv[1]);
    }
    char buff[512];

    while (fgets(buff, 512, fp) != NULL )
    {
        line(buff);             /*sending out an array having one line*/
    }

    return 0;
}

It's obviously not working. I know I have messed up for omp single, but omp atomic seems right. Any help, suggestions or corrections to get this working ?

Sergey L. · Accepted Answer · 2013-07-08 12:28:44Z

I don't even know where to start.

There is just so much wrong with your code.

First: do you know that your parallel code will only run on every 8th row?

if (count == 8)

You allocate returned_all inside the parallel block, meaning each thread will have it's own private copy of returned_all.

Your use of #pragma omp single is completely out of place. Use of this outside of a #omp parallel block has no effect if it even compiles...

Calling free(array_buffer[row]) inside a parallel function will get you into lot's of trouble with double free's and such.

If you want to process a file line by line in parallel I suggest you make use of the default locking in the stdio library and do something like

#pragma omp parallel
{
    // each thread get's it's own private buffer
    char buff[512];
    while (fgets(buff, 512, fp)) {
        // process the line
    }
}

Collectives™ on Stack Overflow

paralleling a function call using openmp

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related