Runtime Hot-Path Optimization

Learn about the high-performance runtime optimization system introduced in WeaveDI v3.1.0.

Overview

Runtime hot-path optimization is an advanced optimization system designed to eliminate performance bottlenecks in dependency injection.

Core Optimization Techniques

1. TypeID + Index Access

   • ObjectIdentifier → Int slot mapping
   • O(1) array index access instead of dictionary lookup
   • Memory access pattern optimization

2. Snapshot/Lock-Free Reads

   • Immutable Storage class-based snapshot approach
   • Complete elimination of read contention
   • Locking only during writes

3. Inline Optimization

   • Applied @inlinable + @inline(__always)
   • Cross-module optimization with @_alwaysEmitIntoClient
   • Reduced function call overhead

4. Factory Chain Elimination

   • Direct call paths without intermediate factory steps
   • Dependency chain flattening
   • Removal of multi-stage factory costs

5. Scope-Specific Static Storage

   • Separation of singleton/session/request scopes
   • Atomic once initialization
   • Race condition elimination

Usage

Enable Optimization

import WeaveDI

// Enable optimization mode
await UnifiedRegistry.shared.enableOptimization()

// Existing code gets performance improvements without changes
let service = await UnifiedDI.resolve(UserService.self)

Check Optimization

// Check optimization status
let isOptimized = await UnifiedRegistry.shared.isOptimizationEnabled
print("Optimization enabled: \(isOptimized)")

// Disable optimization
await UnifiedRegistry.shared.disableOptimization()

Performance Improvements

Scenario	Improvement	Description
Single-threaded resolve	50-80% faster	TypeID + direct access
Multi-threaded reads	2-3x throughput	Lock-free snapshots
Complex dependencies	20-40% faster	Chain flattening

Benchmarks

Run the included benchmarks to measure performance improvements:

swift run -c release Benchmarks --count 100k --quick

Compatibility

• 100% API compatibility: No changes to existing code
• Opt-in optimization: Enable/disable anytime
• Gradual migration: Phased adoption support
• Zero breaking changes: Complete preservation of existing behavior

Internal Implementation

Optimizations are implemented in the following files:

• OptimizedTypeRegistry.swift - TypeID system
• AtomicStorage.swift - Lock-free storage
• DirectCallRegistry.swift - Direct call paths
• OptimizedScopeStorage.swift - Scope optimization

See Also

• Performance Optimization Guide
• Benchmarks Documentation
• UnifiedDI API