View source code
Display the source code in std/digest/murmurhash.d from which this page was generated on github.
Report a bug
If you spot a problem with this page, click here to create a Bugzilla issue.
Improve this page
Quickly fork, edit online, and submit a pull request for this page. Requires a signed-in GitHub account. This works well for small changes. If you'd like to make larger changes you may want to consider using local clone.

Module std.digest.murmurhash

Computes MurmurHash hashes of arbitrary data. MurmurHash is a non-cryptographic hash function suitable for general hash-based lookup. It is optimized for x86 but can be used on all architectures.

The current version is MurmurHash3, which yields a 32-bit or 128-bit hash value. The older MurmurHash 1 and 2 are currently not supported.

MurmurHash3 comes in three flavors, listed in increasing order of throughput:

  • MurmurHash3!32 produces a 32-bit value and is optimized for 32-bit architectures
  • MurmurHash3!(128, 32) produces a 128-bit value and is optimized for 32-bit architectures
  • MurmurHash3!(128, 64) produces a 128-bit value and is optimized for 64-bit architectures

Note

  • MurmurHash3!(128, 32) and MurmurHash3!(128, 64) produce different values.
  • The current implementation is optimized for little endian architectures. It will exhibit different results on big endian architectures and a slightly less uniform distribution.

This module conforms to the APIs defined in std.digest.

This module publicly imports std.digest and can be used as a stand-alone module.

References

Reference implementation
Wikipedia

Example

// MurmurHash3!32, MurmurHash3!(128, 32) and MurmurHash3!(128, 64) implement
// the std.digest Template API.
static assert(isDigest!(MurmurHash3!32));
// The convenient digest template allows for quick hashing of any data.
ubyte[4] hashed = digest!(MurmurHash3!32)([1, 2, 3, 4]);
writeln(hashed); // [0, 173, 69, 68]

Example

// One can also hash ubyte data piecewise by instanciating a hasher and call
// the 'put' method.
const(ubyte)[] data1 = [1, 2, 3];
const(ubyte)[] data2 = [4, 5, 6, 7];
// The incoming data will be buffered and hashed element by element.
MurmurHash3!32 hasher;
hasher.put(data1);
hasher.put(data2);
// The call to 'finish' ensures:
// - the remaining bits are processed
// - the hash gets finalized
auto hashed = hasher.finish();
writeln(hashed); // [181, 151, 88, 252]

Example

// Using `putElements`, `putRemainder` and `finalize` you gain full
// control over which part of the algorithm to run.
// This allows for maximum throughput but needs extra care.

// Data type must be the same as the hasher's element type:
// - uint for MurmurHash3!32
// - uint[4] for MurmurHash3!(128, 32)
// - ulong[2] for MurmurHash3!(128, 64)
const(uint)[] data = [1, 2, 3, 4];
// Note the hasher starts with 'Fast'.
MurmurHash3!32 hasher;
// Push as many array of elements as you need. The less calls the better.
hasher.putElements(data);
// Put remainder bytes if needed. This method can be called only once.
hasher.putRemainder(ubyte(1), ubyte(1), ubyte(1));
// Call finalize to incorporate data length in the hash.
hasher.finalize();
// Finally get the hashed value.
auto hashed = hasher.getBytes();
writeln(hashed); // [188, 165, 108, 2]

Structs

Name Description
MurmurHash3 Implements the MurmurHash3 functions. You can specify the size of the hash in bit. For 128 bit hashes you can specify whether to optimize for 32 or 64 bit architectures. If you don't specify the opt value it will select the fastest version of the host platform.

Authors

Guillaume Chatelet

License

Boost License 1.0.

Comments