Class xxHash3

A feature complete hashing API based on xxHash3 (https://github.com/Cyan4973/xxHash)

Inheritance

object

xxHash3

Namespace: Unity.Collections

Assembly: Unity.Collections.dll

Syntax

public static class xxHash3

Remarks

Features:

• Compute 64bits or 128bits hash keys, based on a private key, with an optional given seed value.
• Hash on buffer (with or without a ulong based seed value)
• Hash on buffer while copying the data to a destination
• Use instances of xxHash3.StreamingState to accumulate data to hash in multiple calls, suited for small data, then retrieve the hash key at the end.
• xxHash3 has several implementation based on the size to hash to ensure best performances
• We currently have two implementations:
  • A generic one based on Unity.Mathematics, that should always be executed compiled with Burst.
  • An AVX2 based implementation for platforms supporting it, using Burst intrinsics.
• Whether or not the call site is compiled with burst, the hashing function will be executed by Burst() to ensure optimal performance. () Only when the hashing size justifies such transition.

Methods

Hash128(void*, long)

Compute a 128bits hash of a memory region

Declaration

public static uint4 Hash128(void* input, long length)

Parameters

Type	Name	Description
void*	input	The memory buffer, can't be null
long	length	The length of the memory buffer, can be zero

Returns

Type	Description
uint4	The hash result

Hash128(void*, long, ulong)

Compute a 128bits hash of a memory region using a given seed value

Declaration

public static uint4 Hash128(void* input, long length, ulong seed)

Parameters

Type	Name	Description
void*	input	The memory buffer, can't be null
long	length	The length of the memory buffer, can be zero
ulong	seed	The seed value to alter the hash computation from

Returns

Type	Description
uint4	The hash result

Hash128(void*, void*, long)

Compute a 128bits hash while copying the data to a destination buffer

Declaration

public static uint4 Hash128(void* input, void* destination, long length)

Parameters

Type	Name	Description
void*	input	The memory buffer to compute the hash and copy from, can't be null
void*	destination	The destination buffer, can't be null and must be at least big enough to match the input's length
long	length	The length of the memory buffer, can be zero

Returns

Type	Description
uint4	The hash result

Remarks

Use this API to avoid a double memory scan in situations where the hash as to be compute and the data copied at the same time. Performances improvements vary between 30-50% on big data.

Hash128(void*, void*, long, ulong)

Compute a 128bits hash while copying the data to a destination buffer using a given seed value

Declaration

public static uint4 Hash128(void* input, void* destination, long length, ulong seed)

Parameters

Type	Name	Description
void*	input	The memory buffer to compute the hash and copy from, can't be null
void*	destination	The destination buffer, can't be null and must be at least big enough to match the input's length
long	length	The length of the memory buffer, can be zero
ulong	seed	The seed value to alter the hash computation from

Returns

Type	Description
uint4	The hash result

Hash128<T>(in T)

Compute a 128bits hash from the contents of the input struct

Declaration

public static uint4 Hash128<T>(in T input) where T : unmanaged

Parameters

Type	Name	Description
T	input	The input struct that will be hashed

Returns

Type	Description
uint4	The hash result

Type Parameters

Name	Description
T	The input type.

Hash64(void*, long)

Compute a 64bits hash of a memory region

Declaration

public static uint2 Hash64(void* input, long length)

Parameters

Type	Name	Description
void*	input	The memory buffer, can't be null
long	length	The length of the memory buffer, can be zero

Returns

Type	Description
uint2	The hash result

Hash64(void*, long, ulong)

Compute a 64bits hash of a memory region using a given seed value

Declaration

public static uint2 Hash64(void* input, long length, ulong seed)

Parameters

Type	Name	Description
void*	input	The memory buffer, can't be null
long	length	The length of the memory buffer, can be zero
ulong	seed	The seed value to alter the hash computation from

Returns

Type	Description
uint2	The hash result

Hash64<T>(in T)

Compute a 64bits hash from the contents of the input struct

Declaration

public static uint2 Hash64<T>(in T input) where T : unmanaged

Parameters

Type	Name	Description
T	input	The input struct that will be hashed

Returns

Type	Description
uint2	The hash result

Type Parameters

Name	Description
T	The input type.