GPU support (#88)

* Format

* GPU support

* Add description of parameters

* Apply suggestions from review

* Use OnnxRuntime with GPU

Author: Parzival-05

VSharp.API/VSharp.cs

@@ -111,7 +111,7 @@ public IEnumerable<FileInfo> Results()
 
     public static class TestGenerator
     {
-        private class Reporter: IReporter
+        private class Reporter : IReporter
         {
             private readonly UnitTests _unitTests;
             private readonly bool _isQuiet;
@@ -124,7 +124,7 @@ public Reporter(UnitTests unitTests, bool isQuiet)
 
             public void ReportFinished(UnitTest unitTest) => _unitTests.GenerateTest(unitTest);
             public void ReportException(UnitTest unitTest) => _unitTests.GenerateError(unitTest);
-            public void ReportIIE(InsufficientInformationException iie) {}
+            public void ReportIIE(InsufficientInformationException iie) { }
 
             public void ReportInternalFail(Method? method, Exception exn)
             {
@@ -179,7 +179,7 @@ private static Statistics StartExploration(
                     explorationMode: explorationMode.NewTestCoverageMode(
                         coverageZone,
                         options.Timeout > 0 ? searchMode.NewFairMode(baseSearchMode) : baseSearchMode
-                        
+
                     ),
                     recThreshold: options.RecursionThreshold,
                     solverTimeout: options.SolverTimeout,
@@ -191,8 +191,11 @@ private static Statistics StartExploration(
                     stopOnCoverageAchieved: 100,
                     randomSeed: options.RandomSeed,
                     stepsLimit: options.StepsLimit,
-                    aiAgentTrainingOptions: options.AIAgentTrainingOptions == null ? FSharpOption<AIAgentTrainingOptions>.None :FSharpOption<AIAgentTrainingOptions>.Some(options.AIAgentTrainingOptions),
-                    pathToModel: options.PathToModel == null ? FSharpOption<string>.None : FSharpOption<string>.Some(options.PathToModel));
+                    aiAgentTrainingOptions: options.AIAgentTrainingOptions == null ? FSharpOption<AIAgentTrainingOptions>.None : FSharpOption<AIAgentTrainingOptions>.Some(options.AIAgentTrainingOptions),
+                    pathToModel: options.PathToModel == null ? FSharpOption<string>.None : FSharpOption<string>.Some(options.PathToModel),
+                    useGPU: options.UseGPU == null ? FSharpOption<bool>.None : FSharpOption<bool>.Some(options.UseGPU),
+                    optimize: options.Optimize == null ? FSharpOption<bool>.None : FSharpOption<bool>.Some(options.Optimize)
+                    );
 
             var fuzzerOptions =
                 new FuzzerOptions(
@@ -326,7 +329,7 @@ private static int CheckCoverage(
         public static Statistics Cover(MethodBase method, VSharpOptions options = new())
         {
             AssemblyManager.LoadCopy(method.Module.Assembly);
-            var methods = new List<MethodBase> {method};
+            var methods = new List<MethodBase> { method };
             var statistics = StartExploration(methods, coverageZone.MethodZone, options);
 
             if (options.RenderTests)

VSharp.API/VSharpOptions.cs

@@ -115,6 +115,8 @@ public readonly record struct VSharpOptions
     public readonly uint StepsLimit = DefaultStepsLimit;
     public readonly AIAgentTrainingOptions AIAgentTrainingOptions = null;
     public readonly string PathToModel = DefaultPathToModel;
+    public readonly bool UseGPU = false;
+    public readonly bool Optimize = false;
 
     /// <summary>
     /// Symbolic virtual machine options.
@@ -133,6 +135,8 @@ public readonly record struct VSharpOptions
     /// <param name="stepsLimit">Number of symbolic machine steps to stop execution after. Zero value means no limit.</param>
     /// <param name="aiAgentTrainingOptions">Settings for AI searcher training.</param>
     /// <param name="pathToModel">Path to ONNX file with model to use in AI searcher.</param>
+    /// <param name="useGPU">Specifies whether the ONNX execution session should use a CUDA-enabled GPU.</param>
+    /// <param name="optimize">Enabling options like parallel execution and various graph transformations to enhance performance of ONNX.</param>
     public VSharpOptions(
         int timeout = DefaultTimeout,
         int solverTimeout = DefaultSolverTimeout,
@@ -147,7 +151,9 @@ public VSharpOptions(
         int randomSeed = DefaultRandomSeed,
         uint stepsLimit = DefaultStepsLimit,
         AIAgentTrainingOptions aiAgentTrainingOptions = null,
-        string pathToModel = DefaultPathToModel)
+        string pathToModel = DefaultPathToModel,
+        bool useGPU = false,
+        bool optimize = false)
     {
         Timeout = timeout;
         SolverTimeout = solverTimeout;
@@ -163,6 +169,8 @@ public VSharpOptions(
         StepsLimit = stepsLimit;
         AIAgentTrainingOptions = aiAgentTrainingOptions;
         PathToModel = pathToModel;
+        UseGPU = useGPU;
+        Optimize = optimize;
     }
 
     /// <summary>