Turnerj / Quickenshtein

Making the quickest and most memory efficient implementation of Levenshtein Distance with SIMD and Threading support
MIT License
284 stars 14 forks source link
edit-distance hardware-intrinsics levenshtein levenshtein-distance simd string-distance threading
![Icon](images/icon.png) # Quickenshtein A quick and memory efficient Levenshtein Distance calculator for .NET [![AppVeyor](https://img.shields.io/appveyor/ci/Turnerj/Quickenshtein/master.svg)](https://ci.appveyor.com/project/Turnerj/Quickenshtein) [![Codecov](https://img.shields.io/codecov/c/github/Turnerj/Quickenshtein/master.svg)](https://codecov.io/gh/Turnerj/Quickenshtein) [![NuGet](https://img.shields.io/nuget/v/Quickenshtein.svg)](https://www.nuget.org/packages/Quickenshtein/)

Performance

Quickenshtein gets its speed and memory efficiency from a number of different optimizations. To get the most performance out of the library, you will need .NET Core 3 or higher as this has support for hardware intrinsics.

Quickenshtein takes advantage of the following hardware intrinsics. On any recent x86 system, you will likely have these available.

If your computer doesn't have one of the hardware intrinsics available, Quickenshtein will still work - just slower than optimal.

Multi-Threading

By default, Quickenshtein performs in single-threaded mode as this mode performs best for small to medium size strings while having no memory allocations. When dealing with huge strings of 8000 characters or more, it may be useful to switch to multi-threaded mode. In this mode, the calculation is broken up and shared between multiple cores in a system.

Multi-threading is especially useful for systems without hardware intrinsics or for .NET Framework as shown in the table below where it provided a 3x performance improvement.

Method Runtime NumberOfCharacters Mean Error StdDev Gen 0 Gen 1 Gen 2 Allocated
Quickenshtein .NET 4.7.2 8000 110.686 ms 10.118 ms 0.554 ms - - - -
Quickenshtein_Threaded .NET 4.7.2 8000 36.601 ms 16.121 ms 0.883 ms - - - 1260 B

To enable threading, you can pass in CalculationOptions.DefaultWithThreading to Levenshtein.GetDistance() or configure your own CalculationOptions with settings that work best for you.

Note: Multi-threading is not allocation free (unlike single-threading mode) and will allocate a small amount depending on the number of threads used.

Benchmarking

There are a number of benchmarks in the repository that you can run on your system to see how well Quickenshtein performs.

Most of these benchmarks...

You can view results to these benchmarks at the links below:

Example Usage

using Quickenshtein;

// Common usage (uses default CalculationOptions with threading disabled)
var distance1 = Levenshtein.GetDistance("Saturday", "Sunday");

// Enable threading
var distance2 = Levenshtein.GetDistance("Saturday", "Sunday", CalculationOptions.DefaultWithThreading);

// Custom calculation options (helps with tuning for your specific workload and environment - you should benchmark your configurations on your system)
var distance3 = Levenshtein.GetDistance("Saturday", "Sunday", new CalculationOptions {
    EnableThreadingAfterXCharacters = 10000,
    MinimumCharactersPerThread = 25000
});

Learning Resources

I've written quite a bit about Levenshtein Distance and various ways you can extract performance from it:

If you prefer video: