Benchmarks: Performance Analysis of AutoMappic v0.1.0

Abstract

AutoMappic achieves high-performance object mapping by shifting resolution and execution paths from runtime reflection to compile-time static analysis. This document details the comparative performance of AutoMappic against legacy mappers, explicit source generators, and manual assignment.

1. Test Environment

All benchmarks were executed using BenchmarkDotNet v0.14+ on the following configuration:

• Runtime: .NET 9.0.2
• Processor: Apple M2 Pro (12 cores)
• OS: macOS 15.1.0
• Compilation: Release mode with EmitCompilerGeneratedFiles=true

2. Methodology

We evaluate performance across two distinct axes: Runtime Mapping Throughput (operations per nanosecond) and Startup Latency/Cold Start (time to first successful map).

Case Study: Nested Flat-Mapping

We measured mapping a complex User object graph to a flattened UserDto.

• Source: User { int Id, string Username, string Email, Address { string City } }
• Target: UserDto { int Id, string Username, string Email, string AddressCity }

3. Results: Runtime Execution

Method	Engine	Mean	Ratio	Allocated
Manual HandWritten	Static Assignment	0.81 ns	1.00	0 B
AutoMappic_Intercepted	Source Gen + Interceptors	0.82 ns	1.01	0 B
Mapperly_Explicit	Source Generation	0.84 ns	1.04	0 B
AutoMapper_Legacy	Reflection / IL Emit	14.20 ns	17.53	120 B

Analysis

AutoMappic performs within the margin of error of manual, hand-written C#. By using Roslyn Interceptors, we eliminate the virtual dispatch overhead of an IMapper interface, allowing the JIT compiler to inline the mapping logic directly into the call site.

4. Results: Startup Performance (Cold Starts)

One of the primary goals of AutoMappic is to eliminate the Startup Scanning Phase. Legacy mappers crawl the AppDomain at runtime, causing CPU spikes during container cold starts.

Implementation	Discovery Method	Startup Latency	Impact
Legacy AutoMapper	Runtime Reflection Scanning	~450ms	Baseline
AutoMappic	Chained Static Registration	~125ms	-325ms

The Sales Angle: In serverless environments like Azure Functions or AWS Lambda, this 300ms+ reduction represents a significant improvement in user-perceived responsiveness and a reduction in provisioned concurrency costs.

5. Reproducibility

To regenerate these benchmarks locally, execute the following command from the repository root:

dotnet run -c Release --project tests/AutoMappic.Benchmarks/AutoMappic.Benchmarks.csproj

The results will be generated in ./BenchmarkDotNet.Artifacts/results/.