Memoization.Net

A high-performance memoization library for .NET that provides automatic function result caching to improve performance by avoiding redundant computations.

Overview

Memoization is an optimization technique that stores the results of expensive function calls and returns the cached result when the same inputs occur again. This library provides a simple and efficient way to memoize functions in C# using Microsoft's IMemoryCache.

Features

• High Performance: Built on top of Microsoft.Extensions.Caching.Abstractions for optimal performance
• Generic Support: Supports functions with up to 16 parameters
• Flexible Caching: Use global default cache or specify per-function cache instances
• Cache Options: Full support for MemoryCacheEntryOptions (TTL, size limits, etc.)
• Exception Safety: Failed computations don't pollute the cache
• Thread Safe: Operations are thread-safe when using thread-safe IMemoryCache implementations
• Multi-Target: Supports .NET Standard 2.0/2.1, .NET Framework 4.6.2, .NET 8.0, and .NET 9.0

Installation

dotnet add package Memoization.Net

Quick Start

1. Initialize Default Cache

using Microsoft.Extensions.Caching.Memory; using Memoization; // Set up the default cache (typically done once at application startup) Memoization.DefaultCache = new MemoryCache(new MemoryCacheOptions { SizeLimit = 1000, CompactionPercentage = 0.25 });

2. Create Memoized Functions

// Example: Expensive recursive function static int Fibonacci(int n) { Console.WriteLine($"Computing Fibonacci({n})"); return n < 2 ? 1 : Fibonacci(n - 1) + Fibonacci(n - 2); } // Create memoized version var memoizedFib = Memoization.Create<int, int>(Fibonacci); // First call - computes and caches result var result1 = memoizedFib(10); // Prints computation messages // Second call - returns cached result instantly var result2 = memoizedFib(10); // No computation, returns cached result

Advanced Usage

Custom Cache Instance

// Use a specific cache instance var customCache = new MemoryCache(new MemoryCacheOptions()); var memoizedFunc = Memoization.Create(expensiveFunction, customCache);

Cache Options

// Configure cache entry options var options = new MemoryCacheEntryOptions { AbsoluteExpirationRelativeToNow = TimeSpan.FromMinutes(30), SlidingExpiration = TimeSpan.FromMinutes(5), Priority = CacheItemPriority.High }; var memoizedFunc = Memoization.Create(expensiveFunction, options);

Multi-Parameter Functions

// Function with multiple parameters static double ExpensiveCalculation(double x, double y, int iterations) { // Simulate expensive computation return Math.Pow(x + y, iterations % 10); } var memoized = Memoization.Create<double, double, int, double>(ExpensiveCalculation); var result = memoized(3.14, 2.71, 1000);

Performance Considerations

Cache Key Generation

The library uses tuples of input parameters as cache keys. For best performance:

• Ensure parameter types implement efficient GetHashCode() and Equals() methods
• Avoid using mutable objects as parameters
• Consider the memory overhead of storing keys and values

Memory Management

// Configure cache with size limits to prevent memory issues var options = new MemoryCacheOptions { SizeLimit = 1000, CompactionPercentage = 0.20 }; var cache = new MemoryCache(options);

Exception Handling

Failed function calls do not pollute the cache:

var memoized = Memoization.Create<int, string>(x => { if (x < 0) throw new ArgumentException("Negative input"); return x.ToString(); }); try { memoized(-1); } catch { /* Exception thrown, nothing cached */ } var result = memoized(5); // Will compute and cache successfully

Thread Safety

All memoization operations depend on the thread safety of the underlying IMemoryCache implementation. Since this library uses Microsoft.Extensions.Caching.Abstractions, the thread safety depends on the specific IMemoryCache implementation being used. The default Microsoft.Extensions.Caching.Memory implementation is thread-safe, so multiple threads can safely call memoized functions concurrently:

var memoized = Memoization.Create(expensiveFunction); // Safe to call from multiple threads (with thread-safe IMemoryCache implementation) Parallel.For(0, 100, i => { var result = memoized(i % 10); // Concurrent access is safe });

Note: While cache operations are typically thread-safe, the memoized function itself should be thread-safe (or pure) to avoid issues with concurrent execution. Additionally, when the same inputs are being processed concurrently by multiple threads, the expensive function may be executed multiple times before the result is cached.

Best Practices

1. Initialize Once: Set up DefaultCache once during application startup
2. Pure Functions: Only memoize pure functions (same input always produces same output)
3. Immutable Parameters: Use immutable types as function parameters when possible
4. Cache Sizing: Configure appropriate cache size limits to prevent memory leaks
5. Monitoring: Monitor cache hit ratios and memory usage in production

Examples

Database Query Caching

// Cache expensive database queries var cachedQuery = Memoization.Create<int, User>(userId => { return database.Users.FirstOrDefault(u => u.Id == userId); }, new MemoryCacheEntryOptions { AbsoluteExpirationRelativeToNow = TimeSpan.FromMinutes(10) });

Computational Results

// Cache complex mathematical computations var cachedCompute = Memoization.Create<double[], double>(weights => { return weights.Select((w, i) => w * Math.Pow(i, 2)).Sum(); });

License

LGPL-3.0-or-later WITH LGPL-3.0-linking-exception. See LICENSE for details.

Related Projects

• Microsoft.Extensions.Caching.Memory - The underlying caching implementation
• Memoization (Wikipedia) - Background on the memoization technique