1
Fork 0
turbonss/src/jenkins_hash.c

191 lines
6.1 KiB
C
Raw Normal View History

2004-12-23 15:16:30 +02:00
#include "jenkins_hash.h"
#include <stdlib.h>
#ifdef WIN32
#define _USE_MATH_DEFINES //For M_LOG2E
#endif
2004-12-23 15:16:30 +02:00
#include <math.h>
#include <limits.h>
#include <string.h>
//#define DEBUG
#include "debug.h"
2005-01-18 23:06:08 +02:00
#define hashsize(n) ((cmph_uint32)1<<(n))
2004-12-23 15:16:30 +02:00
#define hashmask(n) (hashsize(n)-1)
//#define NM2 /* Define this if you do not want power of 2 table sizes*/
/*
--------------------------------------------------------------------
mix -- mix 3 32-bit values reversibly.
For every delta with one or two bits set, and the deltas of all three
high bits or all three low bits, whether the original value of a,b,c
is almost all zero or is uniformly distributed,
* If mix() is run forward or backward, at least 32 bits in a,b,c
have at least 1/4 probability of changing.
* If mix() is run forward, every bit of c will change between 1/3 and
2/3 of the time. (Well, 22/100 and 78/100 for some 2-bit deltas.)
mix() was built out of 36 single-cycle latency instructions in a
structure that could supported 2x parallelism, like so:
a -= b;
a -= c; x = (c>>13);
b -= c; a ^= x;
b -= a; x = (a<<8);
c -= a; b ^= x;
c -= b; x = (b>>13);
...
Unfortunately, superscalar Pentiums and Sparcs can't take advantage
of that parallelism. They've also turned some of those single-cycle
latency instructions into multi-cycle latency instructions. Still,
this is the fastest good hash I could find. There were about 2^^68
to choose from. I only looked at a billion or so.
--------------------------------------------------------------------
*/
#define mix(a,b,c) \
{ \
a -= b; a -= c; a ^= (c>>13); \
b -= c; b -= a; b ^= (a<<8); \
c -= a; c -= b; c ^= (b>>13); \
a -= b; a -= c; a ^= (c>>12); \
b -= c; b -= a; b ^= (a<<16); \
c -= a; c -= b; c ^= (b>>5); \
a -= b; a -= c; a ^= (c>>3); \
b -= c; b -= a; b ^= (a<<10); \
c -= a; c -= b; c ^= (b>>15); \
}
/*
--------------------------------------------------------------------
hash() -- hash a variable-length key into a 32-bit value
k : the key (the unaligned variable-length array of bytes)
len : the length of the key, counting by bytes
initval : can be any 4-byte value
Returns a 32-bit value. Every bit of the key affects every bit of
the return value. Every 1-bit and 2-bit delta achieves avalanche.
About 6*len+35 instructions.
The best hash table sizes are powers of 2. There is no need to do
mod a prime (mod is sooo slow!). If you need less than 32 bits,
use a bitmask. For example, if you need only 10 bits, do
h = (h & hashmask(10));
In which case, the hash table should have hashsize(10) elements.
2005-01-18 23:06:08 +02:00
If you are hashing n strings (cmph_uint8 **)k, do it like this:
2004-12-23 15:16:30 +02:00
for (i=0, h=0; i<n; ++i) h = hash( k[i], len[i], h);
By Bob Jenkins, 1996. bob_jenkins@burtleburtle.net. You may use this
code any way you wish, private, educational, or commercial. It's free.
See http://burtleburtle.net/bob/hash/evahash.html
Use for hash table lookup, or anything where one collision in 2^^32 is
acceptable. Do NOT use for cryptographic purposes.
--------------------------------------------------------------------
*/
2005-01-18 23:06:08 +02:00
cmph_jenkins_state_t *cmph_jenkins_state_new(cmph_uint32 size) //size of hash table
2004-12-23 15:16:30 +02:00
{
2005-01-18 23:06:08 +02:00
cmph_jenkins_state_t *state = (cmph_jenkins_state_t *)malloc(sizeof(cmph_jenkins_state_t));
2004-12-23 15:16:30 +02:00
DEBUGP("Initializing jenkins hash\n");
state->seed = rand() % size;
2005-01-18 23:06:08 +02:00
state->nbits = (cmph_uint32)ceil(log(size)/M_LOG2E);
2004-12-23 15:16:30 +02:00
state->size = size;
DEBUGP("Initialized jenkins with size %u, nbits %u and seed %u\n", size, state->nbits, state->seed);
return state;
}
2005-01-18 23:06:08 +02:00
void cmph_jenkins_state_destroy(cmph_jenkins_state_t *state)
2004-12-23 15:16:30 +02:00
{
free(state);
}
2005-01-18 23:06:08 +02:00
cmph_uint32 cmph_jenkins_hash(cmph_jenkins_state_t *state, const char *k, cmph_uint32 keylen)
2004-12-23 15:16:30 +02:00
{
2005-01-18 23:06:08 +02:00
cmph_uint32 a, b, c;
cmph_uint32 len, length;
2004-12-23 15:16:30 +02:00
/* Set up the internal state */
length = keylen;
len = length;
a = b = 0x9e3779b9; /* the golden ratio; an arbitrary value */
c = state->seed; /* the previous hash value - seed in our case */
/*---------------------------------------- handle most of the key */
while (len >= 12)
{
2005-01-18 23:06:08 +02:00
a += (k[0] +((cmph_uint32)k[1]<<8) +((cmph_uint32)k[2]<<16) +((cmph_uint32)k[3]<<24));
b += (k[4] +((cmph_uint32)k[5]<<8) +((cmph_uint32)k[6]<<16) +((cmph_uint32)k[7]<<24));
c += (k[8] +((cmph_uint32)k[9]<<8) +((cmph_uint32)k[10]<<16)+((cmph_uint32)k[11]<<24));
2004-12-23 15:16:30 +02:00
mix(a,b,c);
k += 12; len -= 12;
}
/*------------------------------------- handle the last 11 bytes */
c += length;
switch(len) /* all the case statements fall through */
{
case 11:
2005-01-18 23:06:08 +02:00
c +=((cmph_uint32)k[10]<<24);
2004-12-23 15:16:30 +02:00
case 10:
2005-01-18 23:06:08 +02:00
c +=((cmph_uint32)k[9]<<16);
2004-12-23 15:16:30 +02:00
case 9 :
2005-01-18 23:06:08 +02:00
c +=((cmph_uint32)k[8]<<8);
2004-12-23 15:16:30 +02:00
/* the first byte of c is reserved for the length */
case 8 :
2005-01-18 23:06:08 +02:00
b +=((cmph_uint32)k[7]<<24);
2004-12-23 15:16:30 +02:00
case 7 :
2005-01-18 23:06:08 +02:00
b +=((cmph_uint32)k[6]<<16);
2004-12-23 15:16:30 +02:00
case 6 :
2005-01-18 23:06:08 +02:00
b +=((cmph_uint32)k[5]<<8);
2004-12-23 15:16:30 +02:00
case 5 :
b +=k[4];
case 4 :
2005-01-18 23:06:08 +02:00
a +=((cmph_uint32)k[3]<<24);
2004-12-23 15:16:30 +02:00
case 3 :
2005-01-18 23:06:08 +02:00
a +=((cmph_uint32)k[2]<<16);
2004-12-23 15:16:30 +02:00
case 2 :
2005-01-18 23:06:08 +02:00
a +=((cmph_uint32)k[1]<<8);
2004-12-23 15:16:30 +02:00
case 1 :
a +=k[0];
/* case 0: nothing left to add */
}
mix(a,b,c);
/*-------------------------------------------- report the result */
//c = (c & hashmask(state->size));
//c = (c >= state->size) ? c ^ state->size: c;
//state->last_hash = c; Do not update last_hash because we use a fixed
//seed
return c;
}
2005-01-18 23:06:08 +02:00
void cmph_jenkins_state_dump(cmph_jenkins_state_t *state, char **buf, cmph_uint32 *buflen)
2004-12-23 15:16:30 +02:00
{
2005-01-18 23:06:08 +02:00
*buflen = sizeof(cmph_uint32)*3;
2004-12-23 15:16:30 +02:00
*buf = malloc(*buflen);
if (!*buf)
{
*buflen = UINT_MAX;
return;
}
2005-01-18 23:06:08 +02:00
memcpy(*buf, &(state->seed), sizeof(cmph_uint32));
memcpy(*buf + sizeof(cmph_uint32), &(state->nbits), sizeof(cmph_uint32));
memcpy(*buf + sizeof(cmph_uint32)*2, &(state->size), sizeof(cmph_uint32));
2004-12-23 15:16:30 +02:00
DEBUGP("Dumped jenkins state with seed %u\n", state->seed);
return;
}
2005-01-18 23:06:08 +02:00
cmph_jenkins_state_t *cmph_jenkins_state_load(const char *buf, cmph_uint32 buflen)
2004-12-23 15:16:30 +02:00
{
2005-01-18 23:06:08 +02:00
cmph_jenkins_state_t *state = (cmph_jenkins_state_t *)malloc(sizeof(cmph_jenkins_state_t));
state->seed = *(cmph_uint32 *)buf;
state->nbits = *(((cmph_uint32 *)buf) + 1);
state->size = *(((cmph_uint32 *)buf) + 2);
state->hashfunc = CMPH_HASH_JENKINS;
2004-12-23 15:16:30 +02:00
DEBUGP("Loaded jenkins state with seed %u\n", state->seed);
return state;
}