Packages

Erlang collection of hash algorithms

Current section

Files

Jump to
erl_hash c_src murmur3.cc
Raw

c_src/murmur3.cc

#include "murmur3.h"
// Based on:
// https://raw.githubusercontent.com/VoltDB/voltdb/master/third_party/cpp/murmur3/MurmurHash3.cpp
// Preserves the legacy 32-bit API while exposing the full MurmurHash3 x64 128-bit hash.
#define FORCE_INLINE inline __attribute__((always_inline))
#define ROTL64(x,y) rotl64(x,y)
#define BIG_CONSTANT(x) (x##LLU)
namespace {
FORCE_INLINE uint64_t rotl64(uint64_t x, int8_t r)
{
return (x << r) | (x >> (64 - r));
}
FORCE_INLINE uint64_t getblock(const uint64_t* p, int i)
{
return p[i];
}
FORCE_INLINE uint64_t fmix(uint64_t k)
{
k ^= k >> 33;
k *= BIG_CONSTANT(0xff51afd7ed558ccd);
k ^= k >> 33;
k *= BIG_CONSTANT(0xc4ceb9fe1a85ec53);
k ^= k >> 33;
return k;
}
FORCE_INLINE void murmur3_x64_128_core(const void* key, int len, uint32_t seed, uint64_t* h1_out, uint64_t* h2_out)
{
const uint8_t* data = static_cast<const uint8_t*>(key);
const int nblocks = len / 16;
uint64_t h1 = seed;
uint64_t h2 = seed;
const uint64_t c1 = BIG_CONSTANT(0x87c37b91114253d5);
const uint64_t c2 = BIG_CONSTANT(0x4cf5ad432745937f);
const uint64_t* blocks = reinterpret_cast<const uint64_t*>(data);
for (int i = 0; i < nblocks; i++)
{
uint64_t k1 = getblock(blocks, i * 2 + 0);
uint64_t k2 = getblock(blocks, i * 2 + 1);
k1 *= c1; k1 = ROTL64(k1, 31); k1 *= c2; h1 ^= k1;
h1 = ROTL64(h1, 27); h1 += h2; h1 = h1 * 5 + 0x52dce729;
k2 *= c2; k2 = ROTL64(k2, 33); k2 *= c1; h2 ^= k2;
h2 = ROTL64(h2, 31); h2 += h1; h2 = h2 * 5 + 0x38495ab5;
}
const uint8_t* tail = data + nblocks * 16;
uint64_t k1 = 0;
uint64_t k2 = 0;
switch (len & 15)
{
case 15: k2 ^= static_cast<uint64_t>(tail[14]) << 48;
case 14: k2 ^= static_cast<uint64_t>(tail[13]) << 40;
case 13: k2 ^= static_cast<uint64_t>(tail[12]) << 32;
case 12: k2 ^= static_cast<uint64_t>(tail[11]) << 24;
case 11: k2 ^= static_cast<uint64_t>(tail[10]) << 16;
case 10: k2 ^= static_cast<uint64_t>(tail[9]) << 8;
case 9: k2 ^= static_cast<uint64_t>(tail[8]);
k2 *= c2; k2 = ROTL64(k2, 33); k2 *= c1; h2 ^= k2;
case 8: k1 ^= static_cast<uint64_t>(tail[7]) << 56;
case 7: k1 ^= static_cast<uint64_t>(tail[6]) << 48;
case 6: k1 ^= static_cast<uint64_t>(tail[5]) << 40;
case 5: k1 ^= static_cast<uint64_t>(tail[4]) << 32;
case 4: k1 ^= static_cast<uint64_t>(tail[3]) << 24;
case 3: k1 ^= static_cast<uint64_t>(tail[2]) << 16;
case 2: k1 ^= static_cast<uint64_t>(tail[1]) << 8;
case 1: k1 ^= static_cast<uint64_t>(tail[0]);
k1 *= c1; k1 = ROTL64(k1, 31); k1 *= c2; h1 ^= k1;
}
h1 ^= len;
h2 ^= len;
h1 += h2;
h2 += h1;
h1 = fmix(h1);
h2 = fmix(h2);
h1 += h2;
h2 += h1;
*h1_out = h1;
*h2_out = h2;
}
}
// Legacy API — Returns top 32 bits of h1
int32_t hash_murmur3_x64_128_legacy(const void* key, int len, uint32_t seed)
{
uint64_t h1, h2;
murmur3_x64_128_core(key, len, seed, &h1, &h2);
return static_cast<int32_t>(h1 >> 32);
}
// New API — full 128-bit hash
murmur3_128_t hash_murmur3_x64_128(const void* key, int len, uint32_t seed)
{
murmur3_128_t out;
murmur3_x64_128_core(key, len, seed, &out.high, &out.low);
return out;
}