From 7ecf94d4b339bc2197f449f0f1538f2915313087 Mon Sep 17 00:00:00 2001 From: pixel Date: Mon, 28 Jan 2008 12:49:06 +0000 Subject: Trying with a hashtable instead of a dumb structure. --- lookupa.c | 239 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 239 insertions(+) create mode 100644 lookupa.c (limited to 'lookupa.c') diff --git a/lookupa.c b/lookupa.c new file mode 100644 index 0000000..731163b --- /dev/null +++ b/lookupa.c @@ -0,0 +1,239 @@ +/* +-------------------------------------------------------------------- +lookupa.c, by Bob Jenkins, December 1996. Same as lookup2.c +Use this code however you wish. Public Domain. No warranty. +Source is http://burtleburtle.net/bob/c/lookupa.c +-------------------------------------------------------------------- +*/ +#include "lookupa.h" + +/* +-------------------------------------------------------------------- +mix -- mix 3 32-bit values reversibly. +For every delta with one or two bit 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); \ +} + +/* +-------------------------------------------------------------------- +lookup() -- 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 + level : 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 6len+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. + +If you are hashing n strings (ub1 **)k, do it like this: + for (i=0, h=0; i= 12) + { + a += (k[0] +((uint32_t)k[1]<<8) +((uint32_t)k[2]<<16) +((uint32_t)k[3]<<24)); + b += (k[4] +((uint32_t)k[5]<<8) +((uint32_t)k[6]<<16) +((uint32_t)k[7]<<24)); + c += (k[8] +((uint32_t)k[9]<<8) +((uint32_t)k[10]<<16)+((uint32_t)k[11]<<24)); + 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: c+=((uint32_t)k[10]<<24); + case 10: c+=((uint32_t)k[9]<<16); + case 9 : c+=((uint32_t)k[8]<<8); + /* the first byte of c is reserved for the length */ + case 8 : b+=((uint32_t)k[7]<<24); + case 7 : b+=((uint32_t)k[6]<<16); + case 6 : b+=((uint32_t)k[5]<<8); + case 5 : b+=k[4]; + case 4 : a+=((uint32_t)k[3]<<24); + case 3 : a+=((uint32_t)k[2]<<16); + case 2 : a+=((uint32_t)k[1]<<8); + case 1 : a+=k[0]; + /* case 0: nothing left to add */ + } + mix(a,b,c); + /*-------------------------------------------- report the result */ + return c; +} + + +/* +-------------------------------------------------------------------- +mixc -- mixc 8 4-bit values as quickly and thoroughly as possible. +Repeating mix() three times achieves avalanche. +Repeating mix() four times eliminates all funnels and all + characteristics stronger than 2^{-11}. +-------------------------------------------------------------------- +*/ +#define mixc(a,b,c,d,e,f,g,h) \ +{ \ + a^=b<<11; d+=a; b+=c; \ + b^=c>>2; e+=b; c+=d; \ + c^=d<<8; f+=c; d+=e; \ + d^=e>>16; g+=d; e+=f; \ + e^=f<<10; h+=e; f+=g; \ + f^=g>>4; a+=f; g+=h; \ + g^=h<<8; b+=g; h+=a; \ + h^=a>>9; c+=h; a+=b; \ +} + +/* +-------------------------------------------------------------------- +checksum() -- hash a variable-length key into a 256-bit value + k : the key (the unaligned variable-length array of bytes) + len : the length of the key, counting by bytes + state : an array of CHECKSTATE 4-byte values (256 bits) +The state is the checksum. Every bit of the key affects every bit of +the state. There are no funnels. About 112+6.875len instructions. + +If you are hashing n strings (ub1 **)k, do it like this: + for (i=0; i<8; ++i) state[i] = 0x9e3779b9; + for (i=0, h=0; i= 32) + { + a += (k[0] +(k[1]<<8) +(k[2]<<16) +(k[3]<<24)); + b += (k[4] +(k[5]<<8) +(k[6]<<16) +(k[7]<<24)); + c += (k[8] +(k[9]<<8) +(k[10]<<16)+(k[11]<<24)); + d += (k[12]+(k[13]<<8)+(k[14]<<16)+(k[15]<<24)); + e += (k[16]+(k[17]<<8)+(k[18]<<16)+(k[19]<<24)); + f += (k[20]+(k[21]<<8)+(k[22]<<16)+(k[23]<<24)); + g += (k[24]+(k[25]<<8)+(k[26]<<16)+(k[27]<<24)); + h += (k[28]+(k[29]<<8)+(k[30]<<16)+(k[31]<<24)); + mixc(a,b,c,d,e,f,g,h); + mixc(a,b,c,d,e,f,g,h); + mixc(a,b,c,d,e,f,g,h); + mixc(a,b,c,d,e,f,g,h); + k += 32; len -= 32; + } + + /*------------------------------------- handle the last 31 bytes */ + h += length; + switch(len) + { + case 31: h+=(k[30]<<24); + case 30: h+=(k[29]<<16); + case 29: h+=(k[28]<<8); + case 28: g+=(k[27]<<24); + case 27: g+=(k[26]<<16); + case 26: g+=(k[25]<<8); + case 25: g+=k[24]; + case 24: f+=(k[23]<<24); + case 23: f+=(k[22]<<16); + case 22: f+=(k[21]<<8); + case 21: f+=k[20]; + case 20: e+=(k[19]<<24); + case 19: e+=(k[18]<<16); + case 18: e+=(k[17]<<8); + case 17: e+=k[16]; + case 16: d+=(k[15]<<24); + case 15: d+=(k[14]<<16); + case 14: d+=(k[13]<<8); + case 13: d+=k[12]; + case 12: c+=(k[11]<<24); + case 11: c+=(k[10]<<16); + case 10: c+=(k[9]<<8); + case 9 : c+=k[8]; + case 8 : b+=(k[7]<<24); + case 7 : b+=(k[6]<<16); + case 6 : b+=(k[5]<<8); + case 5 : b+=k[4]; + case 4 : a+=(k[3]<<24); + case 3 : a+=(k[2]<<16); + case 2 : a+=(k[1]<<8); + case 1 : a+=k[0]; + } + mixc(a,b,c,d,e,f,g,h); + mixc(a,b,c,d,e,f,g,h); + mixc(a,b,c,d,e,f,g,h); + mixc(a,b,c,d,e,f,g,h); + + /*-------------------------------------------- report the result */ + state[0]=a; state[1]=b; state[2]=c; state[3]=d; + state[4]=e; state[5]=f; state[6]=g; state[7]=h; +} -- cgit v1.2.3