khash-example.c

#include <stdio.h>
#include <assert.h>
#include <time.h>
#include <stdlib.h>
#include <string.h>

#if 1
#include "khash.h"
// causes output: $ gcc -O3 khash-example.c && time ./a.out 32000000
// causes output: --- generating data... 32000000 * 9 = 288000000 bytes done!
// causes output: [ht_khash_str_put] size: 32000000
// causes output: [ht_timing] 46.030 sec or 695198.8 per second
// causes output: [ht_khash_str_get] size: 32000000
// causes output: [ht_timing] 14.180 sec or 2256699.6 per second
#else
#include "khash.h.old"
// causes output: $ gcc -O3 khash-example.c && time ./a.out 32000000
// causes output: --- generating data... 32000000 * 9 = 288000000 bytes done!
// causes output: [ht_khash_str_put] size: 32000000
// causes output: [ht_timing] 10.710 sec or 2987861.8 per second
// causes output: [ht_khash_str_get] size: 32000000
// causes output: [ht_timing] 13.590 sec or 2354672.6 per second
#endif

KHASH_MAP_INIT_STR(str,int)

struct HEXSTR {
    char hex[8 + 1];
} __attribute__((packed));


static int data_size = 75000000;
static struct HEXSTR * hexstr;
void ht_init_data()
{
    int i;
    char hexkey[] = "FFFFFFFF";
    printf("--- generating data... ");
    hexstr = (struct HEXSTR * )calloc(data_size, sizeof(*hexstr));
    printf("%u * %lu = %lu bytes ", data_size, sizeof(*hexstr), data_size * sizeof(*hexstr));
    for (i = 0; i < data_size; ++i) {
        int p = 8 - 1;
        while (1) { if ('0' == hexkey[p]) { hexkey[p] = 'F'; p --; } else { if ('A' == hexkey[p]) { hexkey[p] = '9'; break; } else { hexkey[p] --; break; } } }
        strcpy(&hexstr[i].hex[0], &hexkey[0]);
    }
    printf("done!\n");
}
khash_t(str) *h;
void ht_khash_str_put()
{
    int i, ret;
    unsigned k;

    for (i = 0; i < data_size; ++i) {
        k = kh_put(str, h, &hexstr[i].hex[0], &ret);
        //if (!ret) kh_del(str, h, k);
        if (1 != ret) {
            printf("ERROR: unexpected non-empty bucket (i=%d, ret=%d, key=%s)\n", i, ret, &hexstr[i].hex[0]);
            exit(1);
        }
        kh_value(h, k) = i;
    }
    printf("[ht_khash_str_put] size: %u\n", kh_size(h));
}
void ht_khash_str_get()
{
    int i;
    unsigned k;

    int j = 0;
    for (i = 0; i < data_size; ++i) {
        j += 65;
        j = j % data_size;
        k = kh_get(str, h, &hexstr[j].hex[0]);
        int is_missing = (k == kh_end(h));
        if (is_missing) {
            printf("ERROR: unexpected missing key (i=%d, key=%s)\n", j, &hexstr[j].hex[0]);
            exit(1);
        }
        else if (j != kh_value(h, k)) {
            printf("ERROR: unexpected value (i=%d, key=%s)\n", j, &hexstr[j].hex[0]);
            exit(1);
        }
    }
    printf("[ht_khash_str_get] size: %u\n", kh_size(h));
}
void ht_timing(void (*f)(void))
{
    clock_t t = clock();
    (*f)();
    printf("[ht_timing] %.3lf sec or %.1f per second\n", (double)(clock() - t) / CLOCKS_PER_SEC, data_size / ((double)(clock() - t) / CLOCKS_PER_SEC));
}
int main(int argc, char *argv[])
{
    if (argc > 1) data_size = atoi(argv[1]);
    ht_init_data();
    h = kh_init(str);
    ht_timing(ht_khash_str_put);
    ht_timing(ht_khash_str_get);
    kh_destroy(str, h);
    return 0;
}