Commit | Line | Data |
---|---|---|
e25839d4 JA |
1 | #include <math.h> |
2 | #include <string.h> | |
3 | #include <inttypes.h> | |
4 | #include <stdio.h> | |
5 | #include <unistd.h> | |
6 | #include <sys/types.h> | |
7 | #include <fcntl.h> | |
8 | #include "ieee754.h" | |
e25839d4 JA |
9 | #include "zipf.h" |
10 | #include "../minmax.h" | |
ed1860cd | 11 | #include "../hash.h" |
e25839d4 | 12 | |
e6f735f0 | 13 | #define ZIPF_MAX_GEN 10000000UL |
e25839d4 JA |
14 | |
15 | static void zipf_update(struct zipf_state *zs) | |
16 | { | |
4c9060e3 | 17 | unsigned long to_gen; |
e25839d4 JA |
18 | unsigned int i; |
19 | ||
4c9060e3 JA |
20 | /* |
21 | * It can become very costly to generate long sequences. Just cap it at | |
e4839591 JA |
22 | * 10M max, that should be doable in 1-2s on even slow machines. |
23 | * Precision will take a slight hit, but nothing major. | |
4c9060e3 | 24 | */ |
761c2729 | 25 | to_gen = min(zs->nranges, (uint64_t) ZIPF_MAX_GEN); |
e25839d4 | 26 | |
4c9060e3 JA |
27 | for (i = 0; i < to_gen; i++) |
28 | zs->zetan += pow(1.0 / (double) (i + 1), zs->theta); | |
e25839d4 JA |
29 | } |
30 | ||
2316296a JA |
31 | static void shared_rand_init(struct zipf_state *zs, unsigned long nranges, |
32 | unsigned int seed) | |
b2b0b753 JA |
33 | { |
34 | memset(zs, 0, sizeof(*zs)); | |
35 | zs->nranges = nranges; | |
36 | ||
c3546b53 | 37 | init_rand_seed(&zs->rand, seed, 0); |
b2b0b753 JA |
38 | zs->rand_off = __rand(&zs->rand); |
39 | } | |
40 | ||
2316296a JA |
41 | void zipf_init(struct zipf_state *zs, unsigned long nranges, double theta, |
42 | unsigned int seed) | |
e25839d4 | 43 | { |
2316296a | 44 | shared_rand_init(zs, nranges, seed); |
e25839d4 | 45 | |
e25839d4 | 46 | zs->theta = theta; |
1442ba18 | 47 | zs->zeta2 = pow(1.0, zs->theta) + pow(0.5, zs->theta); |
e25839d4 | 48 | |
4c9060e3 | 49 | zipf_update(zs); |
e25839d4 JA |
50 | } |
51 | ||
52 | unsigned long long zipf_next(struct zipf_state *zs) | |
53 | { | |
e25839d4 JA |
54 | double alpha, eta, rand_uni, rand_z; |
55 | unsigned long long n = zs->nranges; | |
56 | unsigned long long val; | |
57 | ||
58 | alpha = 1.0 / (1.0 - zs->theta); | |
59 | eta = (1.0 - pow(2.0 / n, 1.0 - zs->theta)) / (1.0 - zs->zeta2 / zs->zetan); | |
60 | ||
c3546b53 | 61 | rand_uni = (double) __rand(&zs->rand) / (double) FRAND32_MAX; |
e25839d4 JA |
62 | rand_z = rand_uni * zs->zetan; |
63 | ||
64 | if (rand_z < 1.0) | |
65 | val = 1; | |
66 | else if (rand_z < (1.0 + pow(0.5, zs->theta))) | |
67 | val = 2; | |
68 | else | |
69 | val = 1 + (unsigned long long)(n * pow(eta*rand_uni - eta + 1.0, alpha)); | |
70 | ||
8c5e96a4 JA |
71 | val--; |
72 | ||
8348daf9 | 73 | if (!zs->disable_hash) |
8c5e96a4 | 74 | val = __hash_u64(val); |
8348daf9 | 75 | |
8c5e96a4 | 76 | return (val + zs->rand_off) % zs->nranges; |
e25839d4 | 77 | } |
925fee33 | 78 | |
2316296a JA |
79 | void pareto_init(struct zipf_state *zs, unsigned long nranges, double h, |
80 | unsigned int seed) | |
925fee33 | 81 | { |
2316296a | 82 | shared_rand_init(zs, nranges, seed); |
925fee33 | 83 | zs->pareto_pow = log(h) / log(1.0 - h); |
925fee33 JA |
84 | } |
85 | ||
86 | unsigned long long pareto_next(struct zipf_state *zs) | |
87 | { | |
c3546b53 | 88 | double rand = (double) __rand(&zs->rand) / (double) FRAND32_MAX; |
8c5e96a4 JA |
89 | unsigned long long n; |
90 | ||
91 | n = (zs->nranges - 1) * pow(rand, zs->pareto_pow); | |
925fee33 | 92 | |
8348daf9 | 93 | if (!zs->disable_hash) |
8c5e96a4 | 94 | n = __hash_u64(n); |
8348daf9 | 95 | |
8c5e96a4 | 96 | return (n + zs->rand_off) % zs->nranges; |
8348daf9 JA |
97 | } |
98 | ||
99 | void zipf_disable_hash(struct zipf_state *zs) | |
100 | { | |
101 | zs->disable_hash = true; | |
925fee33 | 102 | } |