Awesome
ZstdSharp
ZstdSharp is a port of zstd compression library to С#
Based on Zstandard v1.5.6
Supports .NET Core 3.1, .NET 5+, .NET Standard 2.0+, .NET Framework 4.6.1+
Usage
ZstdSharp has an unsafe API much the same as zstd.
There are also safe wrappers.
Compress data:
var src = File.ReadAllBytes("dickens");
using var compressor = new Compressor(level);
var compressed = compressor.Wrap(src);
Decompress data:
var src = File.ReadAllBytes("dickens.zst");
using var decompressor = new Decompressor();
var decompressed = decompressor.Unwrap(src);
Streaming compression:
using var input = File.OpenRead("dickens");
using var output = File.OpenWrite("dickens.zst");
using var compressionStream = new CompressionStream(output, level);
input.CopyTo(compressionStream);
Streaming decompression:
using var input = File.OpenRead("dickens.zst");
using var output = File.OpenWrite("dickens");
using var decompressionStream = new DecompressionStream(input);
decompressionStream.CopyTo(output);
Multi-threaded compression:
using var input = File.OpenRead("dickens");
using var output = File.OpenWrite("dickens.zst");
using var compressionStream = new CompressionStream(output, level);
compressionStream.SetParameter(ZSTD_cParameter.ZSTD_c_nbWorkers, Environment.ProcessorCount);
input.CopyTo(compressionStream);
Benchmark
Best performance is achieved on .NET
. System.Runtime.Intrinsics
namespace is required for hardware accelerated bit and vector operations. .NET Standard
and .NET Framework
will use software implementation
Comparision zstd
(native) and ZstdSharp
BenchmarkDotNet v0.14.0, Windows 10 (10.0.19045.5011/22H2/2022Update)
12th Gen Intel Core i7-12700, 1 CPU, 20 logical and 12 physical cores
.NET SDK 9.0.100
[Host] : .NET 8.0.11 (8.0.1124.51707), X64 RyuJIT AVX2
.NET 9.0 : .NET 9.0.0 (9.0.24.52809), X64 RyuJIT AVX2
Compression level 1
Method | Mean | Error | StdDev | Ratio |
---|---|---|---|---|
CompressNative | 24.596 ms | 0.0341 ms | 0.0285 ms | 1.00 |
CompressSharp | 25.318 ms | 0.0329 ms | 0.0275 ms | 1.03 |
DecompressNative | 5.642 ms | 0.0145 ms | 0.0128 ms | 1.00 |
DecompressSharp | 5.896 ms | 0.0053 ms | 0.0045 ms | 1.05 |
Compression level 5
Method | Mean | Error | StdDev | Ratio |
---|---|---|---|---|
CompressNative | 69.903 ms | 0.3729 ms | 0.3306 ms | 1.00 |
CompressSharp | 78.997 ms | 0.1981 ms | 0.1756 ms | 1.13 |
DecompressNative | 6.214 ms | 0.0154 ms | 0.0137 ms | 1.00 |
DecompressSharp | 6.746 ms | 0.0083 ms | 0.0069 ms | 1.09 |
Compression level 15
Method | Mean | Error | StdDev | Ratio |
---|---|---|---|---|
CompressNative | 2,237.101 ms | 10.1419 ms | 9.4868 ms | 1.00 |
CompressSharp | 2,008.744 ms | 6.2452 ms | 5.5362 ms | 0.90 |
DecompressNative | 5.385 ms | 0.0088 ms | 0.0074 ms | 1.00 |
DecompressSharp | 5.687 ms | 0.0203 ms | 0.0180 ms | 1.06 |