[fio.git] / t / genzipf.c

/*
 * Generate/analyze pareto/zipf distributions to better understand
 * what an access pattern would look like.
 *
 * For instance, the following would generate a zipf distribution
 * with theta 1.2, using 262144 (1 GB / 4096) values and split the reporting into
 * 20 buckets:
 *
 *	./t/fio-genzipf -t zipf -i 1.2 -g 1 -b 4096 -o 20
 *
 * Only the distribution type (zipf or pareto) and spread input need
 * to be given, if not given defaults are used.
 *
 */
#include <stdio.h>
#include <stdlib.h>
#include <fcntl.h>
#include <string.h>
#include <unistd.h>

#include "../lib/zipf.h"
#include "../lib/gauss.h"
#include "../flist.h"
#include "../hash.h"

#define DEF_NR_OUTPUT	20

struct node {
	struct flist_head list;
	unsigned long long val;
	unsigned long hits;
};

static struct flist_head *hash;
static unsigned long hash_bits = 24;
static unsigned long hash_size = 1 << 24;

enum {
	TYPE_NONE = 0,
	TYPE_ZIPF,
	TYPE_PARETO,
	TYPE_NORMAL,
};
static const char *dist_types[] = { "None", "Zipf", "Pareto", "Normal" };

enum {
	OUTPUT_NORMAL,
	OUTPUT_CSV,
};

static int dist_type = TYPE_ZIPF;
static unsigned long gb_size = 500;
static unsigned long block_size = 4096;
static unsigned long output_nranges = DEF_NR_OUTPUT;
static double percentage;
static double dist_val;
static int output_type = OUTPUT_NORMAL;

#define DEF_ZIPF_VAL	1.2
#define DEF_PARETO_VAL	0.3

static struct node *hash_lookup(unsigned long long val)
{
	struct flist_head *l = &hash[hash_long(val, hash_bits)];
	struct flist_head *entry;
	struct node *n;

	flist_for_each(entry, l) {
		n = flist_entry(entry, struct node, list);
		if (n->val == val)
			return n;
	}

	return NULL;
}

static struct node *hash_insert(struct node *n, unsigned long long val)
{
	struct flist_head *l = &hash[hash_long(val, hash_bits)];

	n->val = val;
	n->hits = 1;
	flist_add_tail(&n->list, l);
	return n;
}

static void usage(void)
{
	printf("genzipf: test zipf/pareto values for fio input\n");
	printf("\t-h\tThis help screen\n");
	printf("\t-p\tGenerate size of data set that are hit by this percentage\n");
	printf("\t-t\tDistribution type (zipf, pareto, or normal)\n");
	printf("\t-i\tDistribution algorithm input (zipf theta, pareto power,\n"
		"\t\tor normal %% deviation)\n");
	printf("\t-b\tBlock size of a given range (in bytes)\n");
	printf("\t-g\tSize of data set (in gigabytes)\n");
	printf("\t-o\tNumber of output rows\n");
	printf("\t-c\tOutput ranges in CSV format\n");
}

static int parse_options(int argc, char *argv[])
{
	const char *optstring = "t:g:i:o:b:p:ch";
	int c, dist_val_set = 0;

	while ((c = getopt(argc, argv, optstring)) != -1) {
		switch (c) {
		case 'h':
			usage();
			return 1;
		case 'p':
			percentage = atof(optarg);
			break;
		case 'b':
			block_size = strtoul(optarg, NULL, 10);
			break;
		case 't':
			if (!strncmp(optarg, "zipf", 4))
				dist_type = TYPE_ZIPF;
			else if (!strncmp(optarg, "pareto", 6))
				dist_type = TYPE_PARETO;
			else if (!strncmp(optarg, "normal", 6))
				dist_type = TYPE_NORMAL;
			else {
				printf("wrong dist type: %s\n", optarg);
				return 1;
			}
			break;
		case 'g':
			gb_size = strtoul(optarg, NULL, 10);
			break;
		case 'i':
			dist_val = atof(optarg);
			dist_val_set = 1;
			break;
		case 'o':
			output_nranges = strtoul(optarg, NULL, 10);
			break;
		case 'c':
			output_type = OUTPUT_CSV;
			break;
		default:
			printf("bad option %c\n", c);
			return 1;
		}
	}

	if (dist_type == TYPE_PARETO) {
		if ((dist_val >= 1.00 || dist_val < 0.00)) {
			printf("pareto input must be > 0.00 and < 1.00\n");
			return 1;
		}
		if (!dist_val_set)
			dist_val = DEF_PARETO_VAL;
	} else if (dist_type == TYPE_ZIPF) {
		if (dist_val == 1.0) {
			printf("zipf input must be different than 1.0\n");
			return 1;
		}
		if (!dist_val_set)
			dist_val = DEF_ZIPF_VAL;
	}

	return 0;
}

struct output_sum {
	double output;
	unsigned int nranges;
};

static int node_cmp(const void *p1, const void *p2)
{
	const struct node *n1 = p1;
	const struct node *n2 = p2;

	return n2->hits - n1->hits;
}

static void output_csv(struct node *nodes, unsigned long nnodes)
{
	unsigned long i;

	printf("rank, count\n");
	for (i = 0; i < nnodes; i++)
		printf("%lu, %lu\n", i, nodes[i].hits);
}

static void output_normal(struct node *nodes, unsigned long nnodes,
			  unsigned long nranges)
{
	unsigned long i, j, cur_vals, interval_step, next_interval, total_vals;
	unsigned long blocks = percentage * nnodes / 100;
	double hit_percent_sum = 0;
	unsigned long long hit_sum = 0;
	double perc, perc_i;
	struct output_sum *output_sums;

	interval_step = (nnodes - 1) / output_nranges + 1;
	next_interval = interval_step;
	output_sums = malloc(output_nranges * sizeof(struct output_sum));

	for (i = 0; i < output_nranges; i++) {
		output_sums[i].output = 0.0;
		output_sums[i].nranges = 0;
	}

	j = total_vals = cur_vals = 0;

	for (i = 0; i < nnodes; i++) {
		struct output_sum *os = &output_sums[j];
		struct node *node = &nodes[i];
		cur_vals += node->hits;
		total_vals += node->hits;
		os->nranges += node->hits;
		if (i == (next_interval) -1 || i == nnodes - 1) {
			os->output = (double) cur_vals / (double) nranges;
			os->output *= 100.0;
			cur_vals = 0;
			next_interval += interval_step;
			j++;
		}

		if (percentage) {
			if (total_vals >= blocks) {
				double cs = i * block_size / (1024 * 1024);
				char p = 'M';

				if (cs > 1024.0) {
					cs /= 1024.0;
					p = 'G';
				}
				if (cs > 1024.0) {
					cs /= 1024.0;
					p = 'T';
				}

				printf("%.2f%% of hits satisfied in %.3f%cB of cache\n", percentage, cs, p);
				percentage = 0.0;
			}
		}
	}

	perc_i = 100.0 / (double)output_nranges;
	perc = 0.0;

	printf("\n   Rows           Hits %%         Sum %%           # Hits          Size\n");
	printf("-----------------------------------------------------------------------\n");
	for (i = 0; i < output_nranges; i++) {
		struct output_sum *os = &output_sums[i];
		double gb = (double)os->nranges * block_size / 1024.0;
		char p = 'K';

		if (gb > 1024.0) {
			p = 'M';
			gb /= 1024.0;
		}
		if (gb > 1024.0) {
			p = 'G';
			gb /= 1024.0;
		}

		perc += perc_i;
		hit_percent_sum += os->output;
		hit_sum += os->nranges;
		printf("%s %6.2f%%\t%6.2f%%\t\t%6.2f%%\t\t%8u\t%6.2f%c\n",
			i ? "|->" : "Top", perc, os->output, hit_percent_sum,
			os->nranges, gb, p);
	}

	printf("-----------------------------------------------------------------------\n");
	printf("Total\t\t\t\t\t\t%8llu\n", hit_sum);
	free(output_sums);
}

int main(int argc, char *argv[])
{
	unsigned long offset;
	unsigned long long nranges;
	unsigned long nnodes;
	struct node *nodes;
	struct zipf_state zs;
	struct gauss_state gs;
	int i, j;

	if (parse_options(argc, argv))
		return 1;

	if (output_type != OUTPUT_CSV)
		printf("Generating %s distribution with %f input and %lu GB size and %lu block_size.\n", dist_types[dist_type], dist_val, gb_size, block_size);

	nranges = gb_size * 1024 * 1024 * 1024ULL;
	nranges /= block_size;

	if (dist_type == TYPE_ZIPF)
		zipf_init(&zs, nranges, dist_val, 1);
	else if (dist_type == TYPE_PARETO)
		pareto_init(&zs, nranges, dist_val, 1);
	else
		gauss_init(&gs, nranges, dist_val, 1);

	hash_bits = 0;
	hash_size = nranges;
	while ((hash_size >>= 1) != 0)
		hash_bits++;

	hash_size = 1 << hash_bits;

	hash = malloc(hash_size * sizeof(struct flist_head));
	for (i = 0; i < hash_size; i++)
		INIT_FLIST_HEAD(&hash[i]);

	nodes = malloc(nranges * sizeof(struct node));

	for (i = j = 0; i < nranges; i++) {
		struct node *n;

		if (dist_type == TYPE_ZIPF)
			offset = zipf_next(&zs);
		else if (dist_type == TYPE_PARETO)
			offset = pareto_next(&zs);
		else
			offset = gauss_next(&gs);

		n = hash_lookup(offset);
		if (n)
			n->hits++;
		else {
			hash_insert(&nodes[j], offset);
			j++;
		}

	}

	qsort(nodes, j, sizeof(struct node), node_cmp);
	nnodes = j;

	if (output_type == OUTPUT_CSV)
		output_csv(nodes, nnodes);
	else
		output_normal(nodes, nnodes, nranges);

	free(hash);
	free(nodes);
	return 0;
}
Commit	Line	Data
6ff38856 JA	1	/*
	2	* Generate/analyze pareto/zipf distributions to better understand
	3	* what an access pattern would look like.
	4	*
	5	* For instance, the following would generate a zipf distribution
2c9b3ccd	6	* with theta 1.2, using 262144 (1 GB / 4096) values and split the reporting into
6ff38856 JA	7	* 20 buckets:
6ff38856 JA	8	*
2c9b3ccd	9	* ./t/fio-genzipf -t zipf -i 1.2 -g 1 -b 4096 -o 20
6ff38856	10	*
f880b1f6 JA	11	* Only the distribution type (zipf or pareto) and spread input need
	12	* to be given, if not given defaults are used.
	13	*
6ff38856 JA	14	*/
	15	#include <stdio.h>
	16	#include <stdlib.h>
	17	#include <fcntl.h>
	18	#include <string.h>
921d17ba	19	#include <unistd.h>
6ff38856 JA	20
6ff38856 JA	21	#include "../lib/zipf.h"
c4ba4d4f	22	#include "../lib/gauss.h"
fd6f237d JA	23	#include "../flist.h"
fd6f237d JA	24	#include "../hash.h"
6ff38856	25
2c9b3ccd	26	#define DEF_NR_OUTPUT 20
f880b1f6	27
fd6f237d JA	28	struct node {
fd6f237d JA	29	struct flist_head list;
fd6f237d JA	30	unsigned long long val;
	31	unsigned long hits;
	32	};
	33
	34	static struct flist_head *hash;
	35	static unsigned long hash_bits = 24;
	36	static unsigned long hash_size = 1 << 24;
fd6f237d	37
921d17ba JA	38	enum {
	39	TYPE_NONE = 0,
	40	TYPE_ZIPF,
	41	TYPE_PARETO,
c4ba4d4f	42	TYPE_NORMAL,
921d17ba	43	};
c4ba4d4f	44	static const char *dist_types[] = { "None", "Zipf", "Pareto", "Normal" };
921d17ba	45
85629b7f JA	46	enum {
	47	OUTPUT_NORMAL,
	48	OUTPUT_CSV,
	49	};
	50
921d17ba JA	51	static int dist_type = TYPE_ZIPF;
921d17ba JA	52	static unsigned long gb_size = 500;
444256ea	53	static unsigned long block_size = 4096;
921d17ba	54	static unsigned long output_nranges = DEF_NR_OUTPUT;
444256ea	55	static double percentage;
921d17ba	56	static double dist_val;
85629b7f	57	static int output_type = OUTPUT_NORMAL;
921d17ba JA	58
	59	#define DEF_ZIPF_VAL 1.2
	60	#define DEF_PARETO_VAL 0.3
	61
fd6f237d JA	62	static struct node *hash_lookup(unsigned long long val)
	63	{
	64	struct flist_head *l = &hash[hash_long(val, hash_bits)];
	65	struct flist_head *entry;
	66	struct node *n;
	67
	68	flist_for_each(entry, l) {
	69	n = flist_entry(entry, struct node, list);
	70	if (n->val == val)
	71	return n;
	72	}
	73
	74	return NULL;
	75	}
	76
24baa4c7	77	static struct node hash_insert(struct node n, unsigned long long val)
fd6f237d JA	78	{
fd6f237d JA	79	struct flist_head *l = &hash[hash_long(val, hash_bits)];
fd6f237d JA	80
	81	n->val = val;
	82	n->hits = 1;
	83	flist_add_tail(&n->list, l);
24baa4c7	84	return n;
6ff38856 JA	85	}
6ff38856 JA	86
4e98a450 JA	87	static void usage(void)
	88	{
	89	printf("genzipf: test zipf/pareto values for fio input\n");
	90	printf("\t-h\tThis help screen\n");
	91	printf("\t-p\tGenerate size of data set that are hit by this percentage\n");
c4ba4d4f JA	92	printf("\t-t\tDistribution type (zipf, pareto, or normal)\n");
	93	printf("\t-i\tDistribution algorithm input (zipf theta, pareto power,\n"
	94	"\t\tor normal %% deviation)\n");
4e98a450 JA	95	printf("\t-b\tBlock size of a given range (in bytes)\n");
4e98a450 JA	96	printf("\t-g\tSize of data set (in gigabytes)\n");
2c9b3ccd	97	printf("\t-o\tNumber of output rows\n");
4e98a450 JA	98	printf("\t-c\tOutput ranges in CSV format\n");
	99	}
	100
921d17ba JA	101	static int parse_options(int argc, char *argv[])
921d17ba JA	102	{
4e98a450	103	const char *optstring = "t:g:i:o:b:p:ch";
921d17ba JA	104	int c, dist_val_set = 0;
	105
	106	while ((c = getopt(argc, argv, optstring)) != -1) {
	107	switch (c) {
4e98a450 JA	108	case 'h':
	109	usage();
	110	return 1;
444256ea JA	111	case 'p':
	112	percentage = atof(optarg);
	113	break;
	114	case 'b':
	115	block_size = strtoul(optarg, NULL, 10);
	116	break;
921d17ba JA	117	case 't':
	118	if (!strncmp(optarg, "zipf", 4))
	119	dist_type = TYPE_ZIPF;
	120	else if (!strncmp(optarg, "pareto", 6))
	121	dist_type = TYPE_PARETO;
c4ba4d4f JA	122	else if (!strncmp(optarg, "normal", 6))
c4ba4d4f JA	123	dist_type = TYPE_NORMAL;
921d17ba JA	124	else {
	125	printf("wrong dist type: %s\n", optarg);
	126	return 1;
	127	}
	128	break;
	129	case 'g':
	130	gb_size = strtoul(optarg, NULL, 10);
	131	break;
	132	case 'i':
	133	dist_val = atof(optarg);
	134	dist_val_set = 1;
	135	break;
921d17ba JA	136	case 'o':
	137	output_nranges = strtoul(optarg, NULL, 10);
	138	break;
355934b7	139	case 'c':
85629b7f	140	output_type = OUTPUT_CSV;
355934b7	141	break;
921d17ba JA	142	default:
	143	printf("bad option %c\n", c);
	144	return 1;
	145	}
	146	}
	147
	148	if (dist_type == TYPE_PARETO) {
	149	if ((dist_val >= 1.00 \|\| dist_val < 0.00)) {
	150	printf("pareto input must be > 0.00 and < 1.00\n");
	151	return 1;
	152	}
	153	if (!dist_val_set)
	154	dist_val = DEF_PARETO_VAL;
	155	} else if (dist_type == TYPE_ZIPF) {
	156	if (dist_val == 1.0) {
	157	printf("zipf input must be different than 1.0\n");
	158	return 1;
	159	}
	160	if (!dist_val_set)
	161	dist_val = DEF_ZIPF_VAL;
	162	}
	163
	164	return 0;
	165	}
	166
444256ea JA	167	struct output_sum {
	168	double output;
	169	unsigned int nranges;
	170	};
	171
24baa4c7 JA	172	static int node_cmp(const void p1, const void p2)
	173	{
	174	const struct node *n1 = p1;
	175	const struct node *n2 = p2;
	176
	177	return n2->hits - n1->hits;
	178	}
	179
85629b7f JA	180	static void output_csv(struct node *nodes, unsigned long nnodes)
	181	{
	182	unsigned long i;
	183
	184	printf("rank, count\n");
	185	for (i = 0; i < nnodes; i++)
	186	printf("%lu, %lu\n", i, nodes[i].hits);
	187	}
	188
	189	static void output_normal(struct node *nodes, unsigned long nnodes,
	190	unsigned long nranges)
	191	{
	192	unsigned long i, j, cur_vals, interval_step, next_interval, total_vals;
	193	unsigned long blocks = percentage * nnodes / 100;
	194	double hit_percent_sum = 0;
	195	unsigned long long hit_sum = 0;
	196	double perc, perc_i;
	197	struct output_sum *output_sums;
	198
	199	interval_step = (nnodes - 1) / output_nranges + 1;
	200	next_interval = interval_step;
	201	output_sums = malloc(output_nranges * sizeof(struct output_sum));
	202
	203	for (i = 0; i < output_nranges; i++) {
	204	output_sums[i].output = 0.0;
	205	output_sums[i].nranges = 0;
	206	}
	207
	208	j = total_vals = cur_vals = 0;
	209
	210	for (i = 0; i < nnodes; i++) {
	211	struct output_sum *os = &output_sums[j];
	212	struct node *node = &nodes[i];
	213	cur_vals += node->hits;
	214	total_vals += node->hits;
	215	os->nranges += node->hits;
	216	if (i == (next_interval) -1 \|\| i == nnodes - 1) {
	217	os->output = (double) cur_vals / (double) nranges;
	218	os->output *= 100.0;
	219	cur_vals = 0;
	220	next_interval += interval_step;
	221	j++;
	222	}
	223
	224	if (percentage) {
	225	if (total_vals >= blocks) {
	226	double cs = i * block_size / (1024 * 1024);
	227	char p = 'M';
	228
	229	if (cs > 1024.0) {
	230	cs /= 1024.0;
	231	p = 'G';
	232	}
	233	if (cs > 1024.0) {
	234	cs /= 1024.0;
	235	p = 'T';
	236	}
	237
	238	printf("%.2f%% of hits satisfied in %.3f%cB of cache\n", percentage, cs, p);
	239	percentage = 0.0;
	240	}
	241	}
	242	}
	243
244	perc_i = 100.0 / (double)output_nranges;
245	perc = 0.0;
246
247	printf("\n Rows Hits %% Sum %% # Hits Size\n");
248	printf("-----------------------------------------------------------------------\n");
249	for (i = 0; i < output_nranges; i++) {
250	struct output_sum *os = &output_sums[i];
251	double gb = (double)os->nranges * block_size / 1024.0;
252	char p = 'K';
253
254	if (gb > 1024.0) {
255	p = 'M';
256	gb /= 1024.0;
257	}
258	if (gb > 1024.0) {
259	p = 'G';
260	gb /= 1024.0;
261	}
262
263	perc += perc_i;
264	hit_percent_sum += os->output;
265	hit_sum += os->nranges;
266	printf("%s %6.2f%%\t%6.2f%%\t\t%6.2f%%\t\t%8u\t%6.2f%c\n",
267	i ? "\|->" : "Top", perc, os->output, hit_percent_sum,
268	os->nranges, gb, p);
269	}
270
271	printf("-----------------------------------------------------------------------\n");
272	printf("Total\t\t\t\t\t\t%8llu\n", hit_sum);
273	free(output_sums);
274	}
275
6ff38856 JA	276	int main(int argc, char *argv[])
6ff38856 JA	277	{
fd6f237d	278	unsigned long offset;
444256ea	279	unsigned long long nranges;
85629b7f	280	unsigned long nnodes;
24baa4c7	281	struct node *nodes;
6ff38856	282	struct zipf_state zs;
c4ba4d4f	283	struct gauss_state gs;
85629b7f	284	int i, j;
6ff38856	285
921d17ba	286	if (parse_options(argc, argv))
6ff38856	287	return 1;
6ff38856	288
85629b7f	289	if (output_type != OUTPUT_CSV)
355934b7	290	printf("Generating %s distribution with %f input and %lu GB size and %lu block_size.\n", dist_types[dist_type], dist_val, gb_size, block_size);
444256ea JA	291
	292	nranges = gb_size * 1024 * 1024 * 1024ULL;
	293	nranges /= block_size;
f880b1f6	294
921d17ba JA	295	if (dist_type == TYPE_ZIPF)
921d17ba JA	296	zipf_init(&zs, nranges, dist_val, 1);
c4ba4d4f	297	else if (dist_type == TYPE_PARETO)
921d17ba	298	pareto_init(&zs, nranges, dist_val, 1);
c4ba4d4f JA	299	else
c4ba4d4f JA	300	gauss_init(&gs, nranges, dist_val, 1);
6ff38856	301
c71224e7 JA	302	hash_bits = 0;
	303	hash_size = nranges;
	304	while ((hash_size >>= 1) != 0)
	305	hash_bits++;
	306
	307	hash_size = 1 << hash_bits;
	308
fd6f237d JA	309	hash = malloc(hash_size * sizeof(struct flist_head));
	310	for (i = 0; i < hash_size; i++)
	311	INIT_FLIST_HEAD(&hash[i]);
	312
24baa4c7 JA	313	nodes = malloc(nranges * sizeof(struct node));
24baa4c7 JA	314
2c9b3ccd	315	for (i = j = 0; i < nranges; i++) {
fd6f237d	316	struct node *n;
6ff38856	317
921d17ba	318	if (dist_type == TYPE_ZIPF)
fd6f237d	319	offset = zipf_next(&zs);
c4ba4d4f	320	else if (dist_type == TYPE_PARETO)
fd6f237d	321	offset = pareto_next(&zs);
c4ba4d4f JA	322	else
c4ba4d4f JA	323	offset = gauss_next(&gs);
fd6f237d JA	324
	325	n = hash_lookup(offset);
	326	if (n)
	327	n->hits++;
24baa4c7 JA	328	else {
	329	hash_insert(&nodes[j], offset);
	330	j++;
	331	}
6ff38856	332
6ff38856 JA	333	}
6ff38856 JA	334
24baa4c7 JA	335	qsort(nodes, j, sizeof(struct node), node_cmp);
24baa4c7 JA	336	nnodes = j;
921d17ba	337
85629b7f JA	338	if (output_type == OUTPUT_CSV)
	339	output_csv(nodes, nnodes);
	340	else
	341	output_normal(nodes, nnodes, nranges);
6ff38856	342
fd6f237d	343	free(hash);
355934b7	344	free(nodes);
6ff38856 JA	345	return 0;
6ff38856 JA	346	}