Optimizing .NET Performance with Copilot Profiler Agent in Visual Studio 2026

Nik Karpinsky showcases the new Copilot Profiler Agent in Visual Studio 2026, guiding developers on leveraging AI-powered profiling to analyze and optimize performance bottlenecks in .NET applications.

Optimizing .NET Performance with Copilot Profiler Agent in Visual Studio 2026

Nik Karpinsky introduces Copilot Profiler Agent, a new AI-powered assistant for Visual Studio 2026 that leverages the capabilities of GitHub Copilot and the Visual Studio Profiler to help developers identify and resolve performance issues in their codebases.

Copilot Profiler Agent Overview

• Integrates GitHub Copilot’s natural language processing with Visual Studio’s performance profiling tools
• Enables direct queries about code performance, hot paths, and optimization opportunities via Copilot Chat

Real-World Example: Optimizing CsvHelper

• Uses the open-source CsvHelper library to illustrate the workflow
• Provides step-by-step instructions to clone the CSV Helper repository and prepare the benchmarking setup

Setting Up Benchmarks

• Copilot Profiler Agent assists in generating benchmarks for the WriteRecords method using natural language queries (@Profiler Help me write a benchmark for the #WriteRecords method)
• Agent installs relevant NuGet packages (Microsoft.VisualStudio.DiagnosticsHub.BenchmarkDotNetDiagnosers) and adapts benchmarks to fit the repository style
• Builds and validates the benchmark setup

public class BenchmarkWriteCsv {
    private const int entryCount = 10000;
    private readonly List<Simple> records = new(entryCount);
    public class Simple {
        public int Id1 { get; set; }
        public int Id2 { get; set; }
        public string Name1 { get; set; }
        public string Name2 { get; set; }
    }
    [GlobalSetup]
    public void GlobalSetup() {
        var random = new Random(42);
        var chars = new char[10];
        string getRandomString() {
            for (int i = 0; i < 10; ++i) chars[i] = (char)random.Next('a', 'z' + 1);
            return new string(chars);
        }
        for (int i = 0; i < entryCount; ++i) {
            records.Add(new Simple {
                Id1 = random.Next(),
                Id2 = random.Next(),
                Name1 = getRandomString(),
                Name2 = getRandomString(),
            });
        }
    }
    [Benchmark]
    public void WriteRecords() {
        using var stream = new MemoryStream();
        using var streamWriter = new StreamWriter(stream);
        using var writer = new CsvHelper.CsvWriter(streamWriter, CultureInfo.InvariantCulture);
        writer.WriteRecords(records);
        streamWriter.Flush();
    }
}

Analyzing and Optimizing with Profiler Agent

• Agent runs the benchmark, applies BenchmarkSwitcher for flexible test selection
• Presents diagnostic session results and identifies performance bottlenecks (e.g., delegate compilation and invocation overhead)
• Profiler Agent explains issues and suggests optimizations via chat:
  • Recognizes frequent delegate invocation as a hot path due to high frequency (40,000 invocations over 10,000 records and 4 fields)
  • Recommends combining field writes into a single delegate using Expression.Block to reduce multicast delegate overhead

var expressions = new List<Expression>(members.Count);
foreach (var memberMap in members) {
    // ... field writing logic ...
    expressions.Add(writeFieldMethodCall);
}
if (expressions.Count == 0) {
    return new Action<T>((T parameter) => { });
}
var block = Expression.Block(expressions);
return Expression.Lambda<Action<T>>(block, recordParameter).Compile();

Measuring the Impact

• Copilot Profiler Agent reruns benchmarks post-optimization
• Reports approximately 24% performance improvement in current tests; confirmed by CPU profiler traces
• Staged pull request in the CsvHelper repo shows ~15% performance gain in broader usage
• Optimization benefits applications writing large CSV files, speeding up record processing and reducing CPU time

Key Takeaways

• Copilot Profiler Agent streamlines performance analysis and remediation using AI-assisted code review and profiling
• Offers actionable insights and clear explanations to guide developers on optimal performance strategies
• Automation of benchmark setup, analysis, and result measurement accelerates the feedback cycle and improves developer productivity

---

Feedback: The article invites readers to share their experiences and suggestions for further improvement. Share your feedback.

This post appeared first on “Microsoft VisualStudio Blog”. Read the entire article here