diff options
| author | Ginger Bill <bill@gingerbill.org> | 2017-06-08 12:03:40 +0100 |
|---|---|---|
| committer | Ginger Bill <bill@gingerbill.org> | 2017-06-08 12:03:40 +0100 |
| commit | 9b61adb97dd78e1cf04ad410e72166f684f97925 (patch) | |
| tree | ccb50b757f31c36dcd2bac161d191e2d23dcb6d1 /src/murmurhash3.cpp | |
| parent | 333924cce15e10e941ee63d6fcdc19d5cb95bb3c (diff) | |
Build as C++
Diffstat (limited to 'src/murmurhash3.cpp')
| -rw-r--r-- | src/murmurhash3.cpp | 220 |
1 files changed, 220 insertions, 0 deletions
diff --git a/src/murmurhash3.cpp b/src/murmurhash3.cpp new file mode 100644 index 000000000..23c9ac454 --- /dev/null +++ b/src/murmurhash3.cpp @@ -0,0 +1,220 @@ +//----------------------------------------------------------------------------- +// MurmurHash3 was written by Austin Appleby, and is placed in the public +// domain. The author hereby disclaims copyright to this source code. + +// Note - The x86 and x64 versions do _not_ produce the same results, as the +// algorithms are optimized for their respective platforms. You can still +// compile and run any of them on any platform, but your performance with the +// non-native version will be less than optimal. + +#if defined(_MSC_VER) +#define ROTL32(x,y) _rotl(x,y) +#define ROTL64(x,y) _rotl64(x,y) +#else + +gb_inline u32 rotl32(u32 x, i8 r) { + return (x << r) | (x >> (32-r)); +} +gb_inline u64 rotl64(u64 x, i8 r) { + return (x << r) | (x >> (64-r)); +} + +#define ROTL32(x,y) rotl32(x,y) +#define ROTL64(x,y) rotl64(x,y) +#endif + +gb_inline u32 fmix32(u32 h) { + h ^= h >> 16; + h *= 0x85ebca6b; + h ^= h >> 13; + h *= 0xc2b2ae35; + h ^= h >> 16; + return h; +} + +gb_inline u64 fmix64(u64 k) { + k ^= k >> 33; + k *= 0xff51afd7ed558ccdULL; + k ^= k >> 33; + k *= 0xc4ceb9fe1a85ec53ULL; + k ^= k >> 33; + return k; +} + +gb_inline u32 mm3_getblock32(u32 *const p, isize i) { + return p[i]; +} +gb_inline u64 mm3_getblock64(u64 *const p, isize i) { + return p[i]; +} + +u128 MurmurHash3_x64_128(void *const key, isize len, u32 seed) { + u8 *const data = cast(u8 *const)key; + isize nblocks = len / 16; + + u64 h1 = seed; + u64 h2 = seed; + + u64 const c1 = 0x87c37b91114253d5ULL; + u64 const c2 = 0x4cf5ad432745937fULL; + + u64 *const blocks = cast(u64 *const)data; + + for (isize i = 0; i < nblocks; i++) { + u64 k1 = mm3_getblock64(blocks, i*2 + 0); + u64 k2 = mm3_getblock64(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; + } + + u8 *const tail = cast(u8 *const)(data + nblocks*16); + + u64 k1 = 0; + u64 k2 = 0; + + switch(len & 15) { + case 15: k2 ^= ((u64)tail[14]) << 48; + case 14: k2 ^= ((u64)tail[13]) << 40; + case 13: k2 ^= ((u64)tail[12]) << 32; + case 12: k2 ^= ((u64)tail[11]) << 24; + case 11: k2 ^= ((u64)tail[10]) << 16; + case 10: k2 ^= ((u64)tail[ 9]) << 8; + case 9: k2 ^= ((u64)tail[ 8]) << 0; + k2 *= c2; k2 = ROTL64(k2,33); k2 *= c1; h2 ^= k2; + + case 8: k1 ^= ((u64)tail[ 7]) << 56; + case 7: k1 ^= ((u64)tail[ 6]) << 48; + case 6: k1 ^= ((u64)tail[ 5]) << 40; + case 5: k1 ^= ((u64)tail[ 4]) << 32; + case 4: k1 ^= ((u64)tail[ 3]) << 24; + case 3: k1 ^= ((u64)tail[ 2]) << 16; + case 2: k1 ^= ((u64)tail[ 1]) << 8; + case 1: k1 ^= ((u64)tail[ 0]) << 0; + k1 *= c1; k1 = ROTL64(k1,31); k1 *= c2; h1 ^= k1; + } + + h1 ^= len; + h2 ^= len; + + h1 += h2; + h2 += h1; + + h1 = fmix64(h1); + h2 = fmix64(h2); + + h1 += h2; + h2 += h1; + + return u128_lo_hi(h1, h2); +} + +u128 MurmurHash3_x86_128(void *const key, isize len, u32 seed) { + u8 *const data = cast(u8 * const)key; + isize nblocks = len / 16; + + u32 h1 = seed; + u32 h2 = seed; + u32 h3 = seed; + u32 h4 = seed; + + u32 const c1 = 0x239b961b; + u32 const c2 = 0xab0e9789; + u32 const c3 = 0x38b34ae5; + u32 const c4 = 0xa1e38b93; + + //---------- + // body + + u32 *const blocks = cast(u32 *const)(data + nblocks*16); + + for (isize i = -nblocks; i != 0; i++) { + u32 k1 = mm3_getblock32(blocks, i*4 + 0); + u32 k2 = mm3_getblock32(blocks, i*4 + 1); + u32 k3 = mm3_getblock32(blocks, i*4 + 2); + u32 k4 = mm3_getblock32(blocks, i*4 + 3); + + k1 *= c1; k1 = ROTL32(k1,15); k1 *= c2; h1 ^= k1; + + h1 = ROTL32(h1,19); h1 += h2; h1 = h1*5+0x561ccd1b; + + k2 *= c2; k2 = ROTL32(k2,16); k2 *= c3; h2 ^= k2; + + h2 = ROTL32(h2,17); h2 += h3; h2 = h2*5+0x0bcaa747; + + k3 *= c3; k3 = ROTL32(k3,17); k3 *= c4; h3 ^= k3; + + h3 = ROTL32(h3,15); h3 += h4; h3 = h3*5+0x96cd1c35; + + k4 *= c4; k4 = ROTL32(k4,18); k4 *= c1; h4 ^= k4; + + h4 = ROTL32(h4,13); h4 += h1; h4 = h4*5+0x32ac3b17; + } + + //---------- + // tail + + u8 *const tail = cast(u8 *const)(data + nblocks*16); + + u32 k1 = 0; + u32 k2 = 0; + u32 k3 = 0; + u32 k4 = 0; + + switch(len & 15) { + case 15: k4 ^= tail[14] << 16; + case 14: k4 ^= tail[13] << 8; + case 13: k4 ^= tail[12] << 0; + k4 *= c4; k4 = ROTL32(k4,18); k4 *= c1; h4 ^= k4; + + case 12: k3 ^= tail[11] << 24; + case 11: k3 ^= tail[10] << 16; + case 10: k3 ^= tail[ 9] << 8; + case 9: k3 ^= tail[ 8] << 0; + k3 *= c3; k3 = ROTL32(k3,17); k3 *= c4; h3 ^= k3; + + case 8: k2 ^= tail[ 7] << 24; + case 7: k2 ^= tail[ 6] << 16; + case 6: k2 ^= tail[ 5] << 8; + case 5: k2 ^= tail[ 4] << 0; + k2 *= c2; k2 = ROTL32(k2,16); k2 *= c3; h2 ^= k2; + + case 4: k1 ^= tail[ 3] << 24; + case 3: k1 ^= tail[ 2] << 16; + case 2: k1 ^= tail[ 1] << 8; + case 1: k1 ^= tail[ 0] << 0; + k1 *= c1; k1 = ROTL32(k1,15); k1 *= c2; h1 ^= k1; + }; + + //---------- + // finalization + + h1 ^= len; h2 ^= len; h3 ^= len; h4 ^= len; + + h1 += h2; h1 += h3; h1 += h4; + h2 += h1; h3 += h1; h4 += h1; + + h1 = fmix32(h1); + h2 = fmix32(h2); + h3 = fmix32(h3); + h4 = fmix32(h4); + + h1 += h2; h1 += h3; h1 += h4; + h2 += h1; h3 += h1; h4 += h1; + + u64 lo = (u64)h1 | ((u64)h2 << 32); + u64 hi = (u64)h3 | ((u64)h4 << 32); + return u128_lo_hi(lo, hi); +} + +gb_inline u128 MurmurHash3_128(void *const key, isize len, u32 seed) { +#if defined(GB_ARCH_64_BIT) + return MurmurHash3_x64_128(key, len, seed); +#else + return MurmurHash3_x86_128(key, len, seed); +#endif +} + + |