Faster.Map 4.0.5

Faster.Map - A collection of Robin-hood hashmaps (FastMap, DenseMapSIMD and DenseMap)

The goal of Faster is to provide a collection of incredible fast hashmaps that integrates into the .net framework.

About

Faster.Map is a collection of hashmaps with minimal memory overhead and incredibly fast runtime speed. See benchmarks, or try it out yourself. Faster evolved from the fact that C# dictionaries in targetframework 4.0 are terribly slow. So i decided to create my own robinhood hashmap, turns out that this hashmap even performs better than the current dictionary written in .net7.

Get Started

1. Install nuget package Faster.Map to your project.

dotnet add package Faster.Map

How to use

Faster.Map provides 3 unique hashmaps:

1. FastMap<Tkey, TValue> is a hashmap which has incredible performance, will only work with numerical keys. Keys need to be unique. Won`t handle hashcollisions

2. DenseMap<Tkey, TValue> is a hashmap which can be used as a replacement to IDicionary. Default loadfactor is 0.5

3. DenseMapSIMD<Tkey, TValue> is a next level hashmap using simd intructions. The default loadfactor is 0.9. This allows us to store 15% more entries than a dictionary while maintaining incredible speed. This hashmap is fast and i mean mindblowing fast. DenseMapSIMD wont cache hashcodes, using types like strings actually need a wrapper that caches the hashcode. Hence using a slow hash fuction will result in a slow hashmap.

Examples

Default Example

private FastMap<uint, uint> _map = new FastMap<uint, uint>(16);     
  _map.Emplace(1, 50); 
  _map.Remove(1);
  _map.Get(1, out var result);
  _map.Update(1, 51); 
 var result = _map[1];    

DenseMap Example

private DenseMap<uint, uint> _map = new DenseMap<uint, uint>(16);
 _map.Emplace(1, 50); 
 _map.Remove(1);
 _map.Get(1, out var result);
 _map.Update(1, 51);

DenseMapSIMD

private DenseMapSIMD<uint, uint> _map = new DenseMapSIMD<uint, uint>(16);
 _map.Emplace(1, 50); 
 _map.Remove(1);
 _map.Get(1, out var result);
 _map.Update(1, 51);

Benchmark

Retrieving a million random generated keys

Adding a million keys

Updating a million keys

Removing a million keys