stdio (c_strings.c)

User time: 109.45000 (sec) system time: 1.05000 (sec)
Average duration: 11.05000 (sec)

iostream (strings.cxx)

User time: 69.85(sec) system time: 0.70(sec)
Average duration: 7.055(sec)

Input.txt was generated with create_input.cxx

/* cc -O5 -Wall -o c_strings c_strings.c */

#include <stdio.h>
#include <string.h>

#include <time.h>
#include <sys/times.h>

#define TOTAL_LOOPS 10

int main(int argc, char *argv[])
{
    char* f_name="input.txt";
    char* search_word="hello";
    FILE* words;
    int i;
    struct tms start;
    struct tms finish;

    /* syntax: c_strings <file_name> <search_word> */
    if(argc > 1) {
        f_name=argv[1];
        if(argc > 2) {
            search_word=argv[2];
        }
    }
    
    words=fopen(f_name, "r");
    
    if(!words) {
        fprintf(stderr, "Unable to open file: %s\n", f_name);
        return -1;
    }

    times(&start);
    for(i=0; i < TOTAL_LOOPS; i++) {
        long start = clock();
        int i=0;
        do {
            char test_word[1024];

            fscanf(words, "%s", test_word); ++i;

            if(!strcmp(test_word, search_word)) {
                printf("Found(%d): %s\n", i, search_word);
            }
        } while(!feof(words));

        printf("Duration: %2.5f\n", (double)(clock()-start)/CLOCKS_PER_SEC);

        fseek(words, 0L, SEEK_SET);
    }

    times(&finish);
    printf("*****\n"
           "User time: %2.5f (sec) system time: %2.5f (sec)\n"
           "Average duration: %2.5f (sec)\n"
           "*****\n",
           (double)(finish.tms_utime-start.tms_utime)/100,
           (double)(finish.tms_stime-start.tms_stime)/100,
           (double)((finish.tms_utime-start.tms_utime)+
                    (finish.tms_stime-start.tms_stime))/100/TOTAL_LOOPS);
    fclose(words);

    return 0;
}

// c++ -O5 -Wall -o strings strings.cxx

#include <string>

#include <iostream.h>
#include <fstream.h>
#include <iomanip.h>
#include <sys/times.h> 
#include <time.h>

using namespace std ;

int main(int argc, char *argv[])
{
    const char* f_name="input.txt";
    string search_word="hello";

    // calling syntax: ./strings <file_name> <search_word>
    if(argc > 1) {
        f_name=argv[1];
        if(argc > 2) {
            search_word=argv[2];
        }
    }
    ifstream words(f_name, ios::in|ios::nocreate);
    
    if(!words.good()) {
        cerr << "Unable to open file: " << f_name << endl;
        return -1;
    }

    struct tms start;
    times(&start); 

    const int total_loops=10;

    cout << setprecision(5);
    for(int i = 0; i < total_loops; i++) {
        long start=clock();
        int i=0;

        do {
            char test_word[1024];

            words >> test_word;
            ++i;

            if(search_word == test_word) {
                cout << "Found(" << i << "): " << test_word << endl;
            }
        } while(!words.eof());

        cout << "Duration: "<<(double)(clock()-start)/CLOCKS_PER_SEC<<endl;

        words.seekg(0, ios::beg);
        words.clear(); // we just clear eof flag.
    }

    struct tms finish;
    times(&finish); 

    cout << "*****\n"
            "User time: " <<
                (double)(finish.tms_utime-start.tms_utime)/100
            << "(sec) system time: " <<
                (double)(finish.tms_stime-start.tms_stime)/100
            << "(sec)\n"
            "Average duration: " << 
                (double)((finish.tms_utime-start.tms_utime)+
                    (finish.tms_stime-start.tms_stime))/100/total_loops <<
            "(sec)\n"
            "*****\n";

    return 0;
}

// c++ -O5 -Wall -o create_input create_input.cxx

#include <iostream.h>
#include <fstream.h>

#include <stdlib.h>

const int output_size=5*1024*1024; // 5 MB
const int block_size=80; // a line

const int common_word_length=5;
#define HELLO_MESSAGES  1

const char f_name[]="input.txt";

int main(int, char*[])
{
    ofstream out_f(f_name);
    if(!out_f) {
        cerr << "Unable to write file: " << f_name << endl;
        return -1;
    }

#if HELLO_MESSAGES==1
    bool already=false;
#endif  
    for(int n=output_size/block_size; n--;) {
        char block_buffer[block_size];
        
        for(unsigned i=0; i < sizeof(block_buffer); i++) {
            if(!(rand()%common_word_length)) {
                block_buffer[i]=' ';
            } else {
                block_buffer[i]=(char)(rand()%('z'-'a'))+'a';
            }
        }
        
        out_f.write(block_buffer, sizeof(block_buffer));
        out_f.put('\n'); // out_f<<endl; causes flush, we do not need it.
        
#if HELLO_MESSAGES != 1
        if(!( rand() % (output_size/block_size/HELLO_MESSAGES) )) {
            out_f << "hello\n"; // We are going to find this word.
        }
#else
        if(!already && !( rand() % ((output_size/block_size)>>1) )) {
            out_f << "hello\n";
            already=true;
        }
#endif
    }

    return 0;
}