added benchmarks

2026-09-14 23:09:25 +02:00
parent f1aca261ae
commit 3dde5e5acf
11 changed files with 823 additions and 18 deletions
@@ -0,0 +1,85 @@
+using System.IO;
+using System.Text.RegularExpressions;
+using BenchmarkDotNet.Attributes;
+using BenchmarkDotNet.Engines;
+using ParadoxSaveParser.Lib;
+using PCRE;
+
+namespace ParadoxSaveParser.Benchmarks;
+
+/// <summary>
+///     In-process engines, all extracting <c>countries.*.technology</c> from the same 110 MB save.
+///     Input is preloaded into memory, so disk I/O is not part of any measurement.
+/// </summary>
+[MemoryDiagnoser]
+[SimpleJob(RunStrategy.Monitoring, launchCount: 1, warmupCount: 1, iterationCount: 5)]
+public class ExtractionBenchmarks
+{
+    private byte[] _pdx = null!;
+    private string _pdxText = null!;
+    private byte[] _json = null!;
+    private ISearchExpression _query = null!;
+
+    private Regex _flatInterpreted = null!;
+    private Regex _flatCompiled = null!;
+    private Regex _flatSourceGen = null!;
+    private Regex _pathAwareCompiled = null!;
+    private Regex _pathAwareNonBacktracking = null!;
+    private Regex _countriesBlock = null!;
+    private PcreRegex _pcreFlat = null!;
+    private PcreRegex _pcrePathAware = null!;
+
+    [GlobalSetup]
+    public void Setup()
+    {
+        _pdx = BenchData.PdxBytes;
+        _pdxText = BenchData.PdxText;
+        _json = File.ReadAllBytes(BenchData.JsonPath);
+        _query = SearchExpressionCompiler.Compile(BenchData.PdxQuery);
+
+        _flatInterpreted = new Regex(Extractors.FlatPattern, RegexOptions.None);
+        _flatCompiled = new Regex(Extractors.FlatPattern, RegexOptions.Compiled);
+        _flatSourceGen = Extractors.FlatSourceGen();
+        _pathAwareCompiled = new Regex(Extractors.PathAwarePattern, RegexOptions.Compiled);
+        _pathAwareNonBacktracking = new Regex(Extractors.PathAwarePattern, RegexOptions.NonBacktracking);
+        _countriesBlock = new Regex(Extractors.CountriesBlockPattern, RegexOptions.Compiled);
+        _pcreFlat = new PcreRegex(Extractors.FlatPattern, PcreOptions.Compiled);
+        _pcrePathAware = new PcreRegex(Extractors.PathAwarePattern, PcreOptions.Compiled);
+    }
+
+    [Benchmark(Baseline = true, Description = "SearchExpression (this project)")]
+    public long SearchExpression() => Extractors.SearchExpression(_pdx, _query);
+
+    [Benchmark(Description = "Full parse, then select")]
+    public long FullParse() => Extractors.FullParseThenSelect(_pdx);
+
+    [Benchmark(Description = ".NET Regex flat, interpreted")]
+    public long RegexFlatInterpreted() => Extractors.RegexScan(_flatInterpreted, _pdxText);
+
+    [Benchmark(Description = ".NET Regex flat, compiled")]
+    public long RegexFlatCompiled() => Extractors.RegexScan(_flatCompiled, _pdxText);
+
+    [Benchmark(Description = ".NET Regex flat, source generated")]
+    public long RegexFlatSourceGen() => Extractors.RegexScan(_flatSourceGen, _pdxText);
+
+    [Benchmark(Description = ".NET Regex path aware, compiled")]
+    public long RegexPathAwareCompiled() => Extractors.RegexScan(_pathAwareCompiled, _pdxText);
+
+    [Benchmark(Description = ".NET Regex path aware, NonBacktracking")]
+    public long RegexPathAwareNonBacktracking() => Extractors.RegexScan(_pathAwareNonBacktracking, _pdxText);
+
+    [Benchmark(Description = ".NET Regex balanced block + flat")]
+    public long RegexBalancedTwoStage() => Extractors.RegexTwoStage(_countriesBlock, _flatCompiled, _pdxText);
+
+    [Benchmark(Description = "PCRE.NET flat, JIT compiled")]
+    public long PcreFlat() => Extractors.PcreScan(_pcreFlat, _pdxText);
+
+    [Benchmark(Description = "PCRE.NET path aware, JIT compiled")]
+    public long PcrePathAware() => Extractors.PcreScan(_pcrePathAware, _pdxText);
+
+    [Benchmark(Description = "Utf8JsonReader over JSON twin")]
+    public long Utf8JsonReader() => Extractors.Utf8JsonReaderScan(_json);
+
+    [Benchmark(Description = "JsonDocument over JSON twin")]
+    public long JsonDocument() => Extractors.JsonDocumentScan(_json);
+}