Question

The macro is like this:

#define MAX 10000000
#define CNT 1000000
#define TIMER_INIT             \
    clock_t starttime, endtime;    \
#define TIMER(txt, process)                     \
    starttime = clock();                        \
    process;                            \
    endtime = clock();                      \
    printf("%-20s %20ld\n",                     \
           txt, (endtime - starttime))

The main codes is like this:

char vector2[CNT/8 + 1];

TIMER(  "bitsort",
        for(i = 0; i < CNT; ++i)
            set1(rand() % MAX, vector2));

And the set1 function is like this:

void set1(int pos, char* vector)        /* set the position to 1 */
{
    vector[pos/8] |= (0x1 << pos%8);
}

When I execute the program, the output looks like this..

bitsort                -36035411302143896

Even if I cast the difference to unsigned int like this, it still doesn’t work:

#define TIMER(txt, process)                     \
    starttime = clock();                        \
    process;                            \
    endtime = clock();                      \
    printf("%-20s %20d\n",                      \
           txt, (unsigned int)(endtime - starttime))

Output is bitsort -16766097

I tried to modify %ld to %d, but it doesn’t work.

But if I don’t use for in the parameter, like this:

TIMER("bitsort", set1(rand() % MAX,vector2));

The TIMER works perfectly and gives me a reasonable result..

Does anyone have idea about what’s wrong with my code? Is it caused by macro or by clock() function? Thanks!

Answer 1

I have found your issue:

Look at this documentation http://www.cplusplus.com/reference/clibrary/cstdio/printf/

Notice something about %d and %i? Signed decimal integer

You must use %u or %lu because clock_t is unsigned, and an unsigned int is also obviously unsigned.

Edit: This was only half of the problem, the other half was that he was destroying his stack(MAX is larger than CNT) in the bitsort function, thanks to @nneonneo in the chat for finding that.