通过多个线程使用ifstream读取同一文件（pthread）

Question

I have a function that is executed simultaneously by multiple threads created in my program and which recursively creates more threads to execute the same function again. I have to process a large file in this function. As there would be multiple threads processing the same file, I guess I'll have to seek to the start of the file for each thread. Will this move the other file streams also to the start of the file? Will there be any issues?

void *myFunc(){
    string lin;
    ifstream ifs ("input.txt");
    if(ifs){
        ifs.seekg(0,ifs.beg);
        while(getline(ifs,lin)){
            ...
            do something
            ...
        }
        ifs.close();
    }
    pthread_t ptds[100];
    int cc = 0;
    if(some condition based on the above code){
        for(int i=0;i<100;i++){
            int rc = pthread_create(&ptds[cc++], NULL, myFunc, NULL);
            if (rc){
                cout << "Error:unable to create thread," << rc << endl;
                exit(-1);
            }
            else{
                cout << "Thread created" << endl;
            }
        }
        void* status;
        int rc;
        for(int i=0; i < 100; i++ ){
            rc = pthread_join(ptds[i], &status);
            if (rc){
                cout << "Error:unable to join," << rc << endl;
                exit(-1);
            }
        }
    }
}

I am getting some ambiguous results. The results change every time I run the code. I'm assuming this has something to do with file I/O synchronization. How to resolve this?

Answer 1

If you try to read the same file object in multiple threads, they'll clobber each other's state. I suggest reading the file into memory once (or memory-mapping it) and having each thread read the data from memory, which they can do wait-free.