Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
7 changes: 0 additions & 7 deletions Cargo.lock

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.

1 change: 0 additions & 1 deletion Cargo.toml
Original file line number Diff line number Diff line change
Expand Up @@ -36,7 +36,6 @@ rustdoc-args = ["--cfg", "docsrs"]

[dependencies]
byteorder = { version = "1.5.0" }
mur3 = { version = "0.1.0" }

[dev-dependencies]
googletest = { version = "0.14.2" }
Expand Down
20 changes: 20 additions & 0 deletions src/hash/mod.rs
Original file line number Diff line number Diff line change
@@ -0,0 +1,20 @@
// Licensed to the Apache Software Foundation (ASF) under one
// or more contributor license agreements. See the NOTICE file
// distributed with this work for additional information
// regarding copyright ownership. The ASF licenses this file
// to you under the Apache License, Version 2.0 (the
// "License"); you may not use this file except in compliance
// with the License. You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing,
// software distributed under the License is distributed on an
// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
// KIND, either express or implied. See the License for the
// specific language governing permissions and limitations
// under the License.

mod murmurhash;

pub use murmurhash::MurmurHash3X64128;
237 changes: 237 additions & 0 deletions src/hash/murmurhash.rs
Original file line number Diff line number Diff line change
@@ -0,0 +1,237 @@
// Licensed to the Apache Software Foundation (ASF) under one
// or more contributor license agreements. See the NOTICE file
// distributed with this work for additional information
// regarding copyright ownership. The ASF licenses this file
// to you under the Apache License, Version 2.0 (the
// "License"); you may not use this file except in compliance
// with the License. You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing,
// software distributed under the License is distributed on an
// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
// KIND, either express or implied. See the License for the
// specific language governing permissions and limitations
// under the License.

use byteorder::{ByteOrder, LE};
use std::hash::Hasher;

const DEFAULT_SEED: u64 = 9001;
const C1: u64 = 0x87c37b91114253d5;
const C2: u64 = 0x4cf5ad432745937f;

/// The MurmurHash3 is a fast, non-cryptographic, 128-bit hash function that has
/// excellent avalanche and 2-way bit independence properties.
#[derive(Debug)]
pub struct MurmurHash3X64128 {
Copy link
Member Author

@tisonkun tisonkun Dec 22, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Naming is still hard. Also we don't need a murmurhash3_x64_128 method based on our current usage.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Let's be explicit here. Name = Spec of the algorithm. Maybe let's add some small comment saying why we are specifically using the x64 variant.

h1: u64,
h2: u64,
total: u64,
buf: [u8; 16],
buf_len: usize,
}

impl MurmurHash3X64128 {
pub fn with_seed(seed: u64) -> Self {
MurmurHash3X64128 {
h1: seed,
h2: seed,
total: 0,
buf: [0; 16],
buf_len: 0,
}
}

pub fn finish128(&self) -> (u64, u64) {
let mut h1 = self.h1;
let mut h2 = self.h2;

let total = self.total + self.buf_len as u64;
let rem = self.buf_len;

// tail
if rem > 0 {
if rem > 8 {
// read k2 little endian
let mut buf = [0u8; 8];
let k2_len = rem - 8;
buf[..k2_len].copy_from_slice(&self.buf[8..rem]);
// mix k2
let mut k2 = u64::from_le_bytes(buf);
k2 = k2.wrapping_mul(C2);
k2 = k2.rotate_left(33);
k2 = k2.wrapping_mul(C1);
h2 ^= k2;
}

// read k1 little endian
let mut buf = [0u8; 8];
let k1_len = rem.min(8);
buf[..k1_len].copy_from_slice(&self.buf[..k1_len]);
// mix k1
let mut k1 = u64::from_le_bytes(buf);
k1 = k1.wrapping_mul(C1);
k1 = k1.rotate_left(31);
k1 = k1.wrapping_mul(C2);
h1 ^= k1;
}

h1 ^= total;
h2 ^= total;
h1 = h1.wrapping_add(h2);
h2 = h2.wrapping_add(h1);
h1 = fmix64(h1);
h2 = fmix64(h2);
h1 = h1.wrapping_add(h2);
h2 = h2.wrapping_add(h1);
(h1, h2)
}

#[inline]
fn update(&mut self, mut k1: u64, mut k2: u64) {
// k1 *= c1; k1 = MURMUR3_ROTL64(k1, 31); k1 *= c2; out.h1 ^= k1;
k1 = k1.wrapping_mul(C1);
k1 = k1.rotate_left(31);
k1 = k1.wrapping_mul(C2);
self.h1 ^= k1;

// out.h1 = MURMUR3_ROTL64(out.h1, 27); out.h1 += out.h2; out.h1 = out.h1*5+0x52dce729;
self.h1 = self.h1.rotate_left(27);
self.h1 = self.h1.wrapping_add(self.h2);
self.h1 = self.h1.wrapping_mul(5).wrapping_add(0x52dce729);

// k2 *= c2; k2 = MURMUR3_ROTL64(k2, 33); k2 *= c1; out.h2 ^= k2;
k2 = k2.wrapping_mul(C2);
k2 = k2.rotate_left(33);
k2 = k2.wrapping_mul(C1);
self.h2 ^= k2;

// out.h2 = MURMUR3_ROTL64(out.h2,31); out.h2 += out.h1; out.h2 = out.h2*5+c4;
self.h2 = self.h2.rotate_left(31);
self.h2 = self.h2.wrapping_add(self.h1);
self.h2 = self.h2.wrapping_mul(5).wrapping_add(0x38495ab5);

// accumulate total length
self.total += 16;
}
}

impl Default for MurmurHash3X64128 {
fn default() -> Self {
Self::with_seed(DEFAULT_SEED)
}
}

impl Hasher for MurmurHash3X64128 {
fn finish(&self) -> u64 {
self.finish128().0
}

fn write(&mut self, mut bytes: &[u8]) {
Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

No need unsafe as mur3.

if self.buf_len + bytes.len() < 16 {
self.buf[self.buf_len..self.buf_len + bytes.len()].copy_from_slice(bytes);
self.buf_len += bytes.len();
return;
}

if self.buf_len != 0 {
let wanted = 16 - self.buf_len;
self.buf[self.buf_len..].copy_from_slice(&bytes[..wanted]);

let k1 = LE::read_u64(&self.buf[0..8]);
let k2 = LE::read_u64(&self.buf[8..16]);
self.update(k1, k2);

bytes = &bytes[wanted..];
self.buf_len = 0;
}

// Number of full 128-bit blocks of 16 bytes.
// Possible exclusion of a remainder of up to 15 bytes.
let blocks = bytes.len() >> 4; // bytes / 16

// Process the 128-bit blocks (the body) into the hash
for i in 0..blocks {
let lo = i << 4;
let mi = lo + 8;
let hi = mi + 8;
let k1 = LE::read_u64(&bytes[lo..mi]);
let k2 = LE::read_u64(&bytes[mi..hi]);
self.update(k1, k2);
}

// remain bytes
let len = bytes.len() % 16;
if len > 0 {
self.buf[0..len].copy_from_slice(&bytes[blocks << 4..]);
self.buf_len = len;
}
}
}

/// Finalization mix: force all bits of a hash block to avalanche.
#[inline]
fn fmix64(mut k: u64) -> u64 {
k ^= k >> 33;
k = k.wrapping_mul(0xff51afd7ed558ccd);
k ^= k >> 33;
k = k.wrapping_mul(0xc4ceb9fe1a85ec53);
k ^ (k >> 33)
}

#[cfg(test)]
mod tests {
use super::*;

fn murmurhash3_x64_128(key: &[u8], seed: u64) -> (u64, u64) {
let mut hasher = MurmurHash3X64128::with_seed(seed);
hasher.write(key);
Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Note that this is different from key.hash(&mut hasher), which will hash the slice length first.

hasher.finish128()
}

#[test]
fn test_remainder() {
// remainder > 8
let key = "The quick brown fox jumps over the lazy dog";
let (h1, h2) = murmurhash3_x64_128(key.as_bytes(), 0);
assert_eq!(h1, 0xe34bbc7bbc071b6c);
assert_eq!(h2, 0x7a433ca9c49a9347);

// change one bit
let key = "The quick brown fox jumps over the lazy eog";
let (h1, h2) = murmurhash3_x64_128(key.as_bytes(), 0);
assert_eq!(h1, 0x362108102c62d1c9);
assert_eq!(h2, 0x3285cd100292b305);

// test a remainder < 8
let key = "The quick brown fox jumps over the lazy dogdogdog";
let (h1, h2) = murmurhash3_x64_128(key.as_bytes(), 0);
assert_eq!(h1, 0x9c8205300e612fc4);
assert_eq!(h2, 0xcbc0af6136aa3df9);

// test a remainder = 8
let key = "The quick brown fox jumps over the lazy1";
let (h1, h2) = murmurhash3_x64_128(key.as_bytes(), 0);
assert_eq!(h1, 0xe3301a827e5cdfe3);
assert_eq!(h2, 0xbdbf05f8da0f0392);

// test a remainder = 0
let key = "The quick brown fox jumps over t";
let (h1, h2) = murmurhash3_x64_128(key.as_bytes(), 0);
assert_eq!(h1, 0xdf6af91bb29bdacf);
assert_eq!(h2, 0x91a341c58df1f3a6);

// test a ones byte and a zeros byte
let key = [
0x54, 0x68, 0x65, 0x20, 0x71, 0x75, 0x69, 0x63, 0x6b, 0x20, 0x62, 0x72, 0x6f, 0x77,
0x6e, 0x20, 0x66, 0x6f, 0x78, 0x20, 0x6a, 0x75, 0x6d, 0x70, 0x73, 0x20, 0x6f, 0x76,
0x65, 0x72, 0x20, 0x74, 0x68, 0x65, 0x20, 0x6c, 0x61, 0x7a, 0x79, 0x20, 0x64, 0x6f,
0x67, 0xff, 0x64, 0x6f, 0x67, 0x00,
];
let (h1, h2) = murmurhash3_x64_128(&key, 0);
assert_eq!(h1, 0xe88abda785929c9e);
assert_eq!(h2, 0x96b98587cacc83d6);
}
}
6 changes: 3 additions & 3 deletions src/hll/mod.rs
Original file line number Diff line number Diff line change
Expand Up @@ -74,6 +74,8 @@

use std::hash::Hash;

use crate::hash::MurmurHash3X64128;

mod array4;
mod array6;
mod array8;
Expand Down Expand Up @@ -147,9 +149,7 @@ fn pack_coupon(slot: u32, value: u8) -> u32 {

/// Generate a coupon from a hashable value.
fn coupon<H: Hash>(v: H) -> u32 {
const DEFAULT_SEED: u32 = 9001;

let mut hasher = mur3::Hasher128::with_seed(DEFAULT_SEED);
let mut hasher = MurmurHash3X64128::default();
v.hash(&mut hasher);
let (lo, hi) = hasher.finish128();

Expand Down
2 changes: 2 additions & 0 deletions src/lib.rs
Original file line number Diff line number Diff line change
Expand Up @@ -33,3 +33,5 @@ compile_error!("datasketches does not support big-endian targets");
pub mod error;
pub mod hll;
pub mod tdigest;

mod hash;