DEV Community

Cover image for 50 C# (Advanced) Optimization Performance Tips🔥
ByteHide
ByteHide

Posted on • Originally published at bytehide.com

50 C# (Advanced) Optimization Performance Tips🔥

As an experienced C# developer, you’re always looking for ways to improve your application’s performance. Good news! You’ve come to the right place.

In this article, we’ll explore 50 fantastic C# performance tips that will help you optimize your code and make sure your app runs as smoothly as possible. From memory management to parallel computing, we’ll cover everything you need to know about C# optimization.

So let’s dive right in and unlock the full potential of your C# applications!

Memory Management and Garbage Collection

In this section, we’ll introduce effective strategies for handling memory and reducing garbage collection overhead in your C# applications.

Memory management and garbage collection are essential aspects of performance tuning in C#, so these best practices will help you optimize your code for maximum efficiency.

Leverage the IDisposable interface

Utilizing the IDisposable interface is a crucial C# performance tip. It helps you properly manage unmanaged resources and ensures that your application’s memory usage is efficient.

Bad way:

public class ResourceHolder
{
    private Stream _stream;

    public ResourceHolder(string filePath)
    {
        _stream = File.OpenRead(filePath);
    }

    // Missing: IDisposable implementation
}
Enter fullscreen mode Exit fullscreen mode

In the bad way above, the ResourceHolder class doesn’t implement the IDisposable interface, which means the unmanaged resources might not be released, causing potential memory leaks.

Good way:

public class ResourceHolder : IDisposable
{
    private Stream _stream;

    public ResourceHolder(string filePath)
    {
        _stream = File.OpenRead(filePath);
    }

    public void Dispose()
    {
        _stream?.Dispose(); // Properly disposing the unmanaged resource.
    }
}
Enter fullscreen mode Exit fullscreen mode

By implementing the IDisposable interface, you ensure that unmanaged resources will be released when no longer needed, preventing memory leaks and reducing pressure on the garbage collector.

This is a fundamental code optimization technique in C# that developers should utilize.

Avoid premature optimizations

Premature optimizations can be counterproductive, making your C# code harder to read, maintain, and extend. It’s essential to first focus on writing clean, efficient code and only optimize when necessary after thoroughly profiling your application.

Bad way:

private void ProcessData()
{
    Stopwatch stopwatch = new Stopwatch();
    stopwatch.Start();

    // [Complex processing logic with unnecessary micro-optimizations]

    stopwatch.Stop();
    Console.WriteLine($"Processing time: {stopwatch.ElapsedMilliseconds} ms");
}
Enter fullscreen mode Exit fullscreen mode

The bad way above focuses too much on micro-optimizations, which can lead to complex, cluttered code that sacrifices maintainability for a negligible performance improvement.

Good way:

private void ProcessData()
{
    // [Straightforward processing logic without premature optimization]

    // Optimize only if necessary, and only after profiling and identifying bottlenecks.
}
Enter fullscreen mode Exit fullscreen mode

Premature optimizations can make your code harder to maintain and may not have a significant impact on overall performance. Instead, focus on writing clean and straightforward code, then optimize only when necessary after thorough profiling.

This approach will lead to more maintainable and higher-performing C# applications.

Asynchronous Programming with async/await

Asynchronous programming is a powerful technique for improving C# performance in I/O-bound operations, allowing you to enhance your app’s responsiveness and efficiency. Here, we’ll explore some best practices for async/await in C#.

Limit the number of concurrent operations

Managing concurrency is crucial for C# performance optimization. By limiting the number of concurrent operations in your application, you help to reduce the system’s overall load.

Bad way:

public async Task ProcessManyItems(List<string> items)
{
    var tasks = items.Select(async item => await ProcessItem(item));
    await Task.WhenAll(tasks);
}
Enter fullscreen mode Exit fullscreen mode

In the bad way, tasks are spawned concurrently for each item without a proper limit, potentially causing significant strain on the system.

Good way:

public async Task ProcessManyItems(List<string> items, int maxConcurrency = 10)
{
    using (var semaphore = new SemaphoreSlim(maxConcurrency))
    {
        var tasks = items.Select(async item =>
        {
            await semaphore.WaitAsync(); // Limit concurrency by waiting for the semaphore.
            try
            {
                await ProcessItem(item);
            }
            finally
            {
                semaphore.Release(); // Release the semaphore to allow other operations.
            }
        });

        await Task.WhenAll(tasks);
    }
}
Enter fullscreen mode Exit fullscreen mode

Without limiting concurrency, many tasks will run simultaneously, which can lead to heavy load and degraded overall performance. Instead, use a SemaphoreSlim to control the number of concurrent operations.

This is a great example of how to improve application performance in C# without sacrificing readability or maintainability.

UseConfigureAwait(false) when possible

ConfigureAwait(false) is a valuable C# performance trick that can help prevent deadlocks in your async code and improve efficiency by not forcing continuations to run on the original synchronization context.

Bad way:

public async Task<string> LoadDataAsync()
{
    var data = await ReadDataAsync();
    return ProcessData(data);
}
Enter fullscreen mode Exit fullscreen mode

The bad way above does not use ConfigureAwait(false), which carries a risk of potential deadlocks in certain cases.

Good way:

public async Task<string> LoadDataAsync()
{
    var data = await ReadDataAsync().ConfigureAwait(false); // Use ConfigureAwait(false) to avoid potential deadlocks.
    return ProcessData(data);
}
Enter fullscreen mode Exit fullscreen mode

ConfigureAwait(false) helps to avoid potential deadlocks in your async code and improves efficiency by not forcing continuations to run on the original context. Use it whenever it’s safe, typically in library code and non-UI applications.

This is a practical example of C# performance tuning that can have a significant positive impact on your application’s overall responsiveness and stability.

Parallel Computing and Task Parallel Library (TPL)

Parallel computing can help harness the power of multicore processors and speed up CPU-bound operations, ultimately improving the performance of your C# applications. Let’s explore some tips to get the most out of parallel computing in C#.

Utilize parallel loops with Parallel.For() and Parallel.ForEach()

Bad way:

private void ProcessData(List<int> data)
{
    for (int i = 0; i < data.Count; i++)
    {
        PerformExpensiveOperation(data[i]);
    }
}
Enter fullscreen mode Exit fullscreen mode

In the bad way above, a standard for loop is used to process the data collection, resulting in sequential execution of the operations. This does not take advantage of the full potential of modern, multicore CPUs.

Good way:

private void ProcessData(List<int> data)
{
    Parallel.ForEach(data, item => PerformExpensiveOperation(item));
}
Enter fullscreen mode Exit fullscreen mode

Parallel loops can considerably accelerate processing of large collections by distributing the workload among multiple CPU cores. Switch from regular for and foreach loops to their parallel counterparts whenever it’s feasible and safe.
This is a solid example of how to radically speed up your C# code using parallel computing techniques.

Use Partitioner class for efficient workload distribution

Bad way:

private void ProcessData(IEnumerable<int> data)
{
    Parallel.ForEach(data, item => PerformExpensiveOperation(item));
}
Enter fullscreen mode Exit fullscreen mode

In the bad way above, no special consideration is taken to optimize the partitioning of the workload among the parallel tasks. This can lead to potential overhead and imbalanced load distribution.

Good way:

private void ProcessData(IEnumerable<int> data)
{
    var partitioner = Partitioner.Create(data);
    Parallel.ForEach(partitioner, item => PerformExpensiveOperation(item));
}
Enter fullscreen mode Exit fullscreen mode

By employing the Partitioner class, you can efficiently distribute workloads into chunks, reducing potential overhead and improving load balancing among parallel tasks.

The Partitioner creates optimal work chunks to minimize the overhead of task synchronization, resulting in better performance and workload distribution for your C# applications.

Importance of Caching Data

Caching can significantly improve application performance by reducing the time taken to fetch and process data. In this section, we’ll discuss some effective caching techniques and their proper implementation in C# code optimization.

Implement data caching with in-memory cache

Utilizing in-memory caching can drastically reduce time-consuming database fetches and speed up your application.

Bad way:

public Product GetProductById(int id)
{
    // Fetching product data from the database every time
    var product = _dbContext.Products.FirstOrDefault(p => p.Id == id);
    return product;
}
Enter fullscreen mode Exit fullscreen mode

In the bad way above, product data is fetched from the database every time the method is called. This can cause significant performance degradation, especially if the database is located remotely or is under heavy load.

Good way:

private static MemoryCache _cache = new MemoryCache(new MemoryCacheOptions());

public Product GetProductById(int id)
{
    // Fetching product data from the cache if available
    if (!_cache.TryGetValue(id, out Product product))
    {
        product = _dbContext.Products.FirstOrDefault(p => p.Id == id);
        _cache.Set(id, product, TimeSpan.FromMinutes(30));
    }
    return product;
}
Enter fullscreen mode Exit fullscreen mode

The good way demonstrates the use of in-memory caching to store product data and reduce time-consuming database fetches. Utilize MemoryCache to cache frequently requested data and improve performance.

This is a .NET performance optimization technique that helps to speed up data retrieval and reduce the load on your database server.

Implement caching with distributed cache systems (e.g., Redis)

Distributed cache systems, like Redis, can further enhance your application’s performance by caching data in a manner that scales across multiple servers and provides high availability.

Bad way:

public List<Product> GetPopularProducts()
{
    // Fetching popular product data from the database every time
    var popularProducts = _dbContext.Products.Where(p => p.IsPopular).ToList();
    return popularProducts;
}
Enter fullscreen mode Exit fullscreen mode

The bad way above retrieves popular product data from the database every time the method is called, resulting in unnecessary database fetch operations and diminished performance.

Good way:

private static IDistributedCache _distributedCache;

public List<Product> GetPopularProducts()
{
    // Fetching popular product data from the distributed cache if available
    string cacheKey = "popularProducts";
    string cachedProducts = _distributedCache.GetString(cacheKey);

    if (cachedProducts == null)
    {
        var popularProducts = _dbContext.Products.Where(p => p.IsPopular).ToList();
        _distributedCache.SetString(cacheKey, JsonConvert.SerializeObject(popularProducts), new DistributedCacheEntryOptions
        {
            AbsoluteExpirationRelativeToNow = TimeSpan.FromMinutes(30)
        });
        return popularProducts;
    }
    else
    {
        return JsonConvert.DeserializeObject<List<Product>>(cachedProducts);
    }
}
Enter fullscreen mode Exit fullscreen mode

The good way showcases implementing distributed caching with Redis to store popular product data, again reducing database fetches. Employ distributed cache systems like Redis for caching across multiple servers and improving application scalability.

By using Redis, you can optimize your C# code and ensure fast data access even when your application runs on multiple servers.

Concurrency and Thread Safety

Managing concurrency is a fundamental aspect of developing high-quality C# applications. Ensuring thread-safety can prevent undesirable bugs and performance issues, so let’s consider some best practices.

Use lock-free data structures when possible

Opting for lock-free data structures, such as ConcurrentBag, ConcurrentQueue, or ConcurrentDictionary, can help you maintain thread safety in multi-threaded scenarios without sacrificing performance.

Bad way:

private readonly object _syncRoot = new object();
private readonly List<int> _list = new List<int>();

public void Add(int item)
{
    lock (_syncRoot)
    {
        _list.Add(item);
    }
}
Enter fullscreen mode Exit fullscreen mode

In the bad way above, the lock keyword is used to synchronize access to the list, which can lead to contention and degraded performance.

Good way:

private readonly ConcurrentBag<int> _bag = new ConcurrentBag<int>();

public void Add(int item)
{
    _bag.Add(item);
}
Enter fullscreen mode Exit fullscreen mode

By using lock-free data structures such as ConcurrentBag, ConcurrentQueue, or ConcurrentDictionary, you can minimize contention, improve performance, and ensure thread-safety in multi-threaded scenarios.

Use efficient synchronization constructs

Utilizing efficient synchronization constructs, such as SemaphoreSlim, ReaderWriterLockSlim, or Monitor, can help you protect shared resources and maintain thread safety while minimizing contention and performance impact.

Bad way:

private readonly object _syncRoot = new object();
private readonly List<int> _list = new List<int>();

public void Add(int item)
{
    lock (_syncRoot)
    {
        _list.Add(item);
    }
}
Enter fullscreen mode Exit fullscreen mode

In the bad way above, the lock keyword is used again for synchronization. This can lead to contention and negatively impact performance.

Good way:

private readonly SemaphoreSlim _semaphoreSlim = new SemaphoreSlim(1, 1);
private readonly List<int> _list = new List<int>();

public async Task AddAsync(int item)
{
    await _semaphoreSlim.WaitAsync();
    try
    {
        _list.Add(item);
    }
    finally
    {
        _semaphoreSlim.Release();
    }
}
Enter fullscreen mode Exit fullscreen mode

Efficient synchronization constructs like SemaphoreSlim, ReaderWriterLockSlim, or Monitor allow you to protect shared resources and ensure thread safety while minimizing contention and performance overhead.

Choose the most suitable synchronization construct based on your application’s requirements and use them judiciously to avoid potential performance bottlenecks.

Employ the Interlocked class for atomic operations

Using the Interlocked class, you can perform simple atomic operations without relying on locks, reducing contention and improving performance.

Bad way:

private int _counter;
private readonly object _syncRoot = new object();

public void IncrementCounter()
{
    lock (_syncRoot)
    {
        _counter++;
    }
}
Enter fullscreen mode Exit fullscreen mode

In the bad way above, the lock keyword is utilized to ensure thread-safety during the counter increment. However, this can result in contention and performance degradation.

Good way:

private int _counter;

public void IncrementCounter()
{
    Interlocked.Increment(ref _counter);
}
Enter fullscreen mode Exit fullscreen mode

The Interlocked class lets you perform simple atomic operations without using locks, resulting in increased performance and reduced contention. Use it whenever possible for operations like incrementing, decrementing, or addition.

Understanding and Optimizing LINQ Performance

LINQ is a powerful tool, but it can impact performance if used improperly. In this section, we’ll explore tips and tricks to optimize LINQ usage in your C# applications.

Know the difference between deferred and immediate execution

Understanding deferred and immediate execution permits you to have better control over when your LINQ queries execute while avoiding potential performance issues.

Bad way:

public IEnumerable<int> GetEvenNumbers(IEnumerable<int> numbers)
{
    var evens = numbers.Where(n => n % 2 == 0);

    // Do some other work

    return evens;
}
Enter fullscreen mode Exit fullscreen mode

In the bad way above, the LINQ query’s execution is deferred, which can lead to redundant query executions if the returned IEnumerable is enumerated multiple times.

Good way:

public IReadOnlyList<int> GetEvenNumbers(IEnumerable<int> numbers)
{
    var evens = numbers.Where(n => n % 2 == 0).ToList();

    // Do some other work

    return evens;
}
Enter fullscreen mode Exit fullscreen mode

Understanding deferred and immediate execution helps you control when your LINQ queries execute and avoid potential performance problems. Force immediate execution using ToList() or ToArray() when needed.

Opt for query syntax over method syntax when possible

Choosing query syntax over method syntax can result in more readable and maintainable code, especially for complex queries.

Bad way:

var query = items.Select(item => new { PropertyName = item.SomeProperty });
Enter fullscreen mode Exit fullscreen mode

In the bad way above, method syntax is used to express the query, which can become unreadable if the query is more complex.

Good way:

var query = from item in items
            select new { PropertyName = item.SomeProperty };
Enter fullscreen mode Exit fullscreen mode

Using query syntax over method syntax can result in more readable and maintainable code, especially for complex queries. Make use of query syntax whenever feasible.

Be aware of potential pitfalls when using LINQ in a multithreaded environment

Using LINQ in parallel scenarios requires caution to avoid potential issues related to thread safety and performance bottlenecks.

Bad way:

Parallel.ForEach(items, item =>
{
    var matchingItems = items.Where(i => i.Name == item.Name);
    Process(matchingItems);
});
Enter fullscreen mode Exit fullscreen mode

In the bad way above, multiple threads enumerate the same IEnumerable resulting from the LINQ query, which can lead to unpredictable behavior.

Good way:

Parallel.ForEach(items, item =>
{
    // Create a separate LINQ query for each parallel operation
    var matchingItems = items.AsParallel().Where(i => i.Name == item.Name);
    Process(matchingItems);
});
Enter fullscreen mode Exit fullscreen mode

Using LINQ in parallel scenarios requires special attention to avoid potential issues, such as thread safety and performance bottlenecks. Employ the AsParallel() extension method to ensure safety and parallelism.

Micro-optimizations and JIT Compilation

Micro-optimizations in your C# code may appear minor but can lead to significant performance improvements. Now, we will discuss some techniques to fine-tune your code.

Perform loop unrolling for better performance

Loop unrolling can accelerate your code execution by reducing the overhead of loop control structures. However, apply it cautiously, as excessive loop unrolling can negatively affect code readability and maintenance.

Bad way:

for (int i = 0; i < array.Length; i++)
{
    array[i] = i * 2;
}
Enter fullscreen mode Exit fullscreen mode

In the bad way above, a simple loop iterates through each element of an array, causing high loop control structure overhead.

Good way:

int len = array.Length;
for (int i = 0; i < len; i += 4)
{
    array[i] = i * 2;
    array[i + 1] = (i + 1) * 2;
    array[i + 2] = (i + 2) * 2;
    array[i + 3] = (i + 3) * 2;
}
Enter fullscreen mode Exit fullscreen mode

Loop unrolling can lead to faster execution of your code by reducing the overhead of loop control structures. Apply it cautiously, though, as excessive loop unrolling can impact readability and maintenance.

Utilize the aggressive inlining attribute for critical methods

By marking critical methods with the AggressiveInlining attribute, you can instruct the JIT compiler to inline them, potentially improving performance by reducing the overhead of method calls.

Bad way:

private int MultiplyByTwo(int value)
{
    return value * 2;
}
Enter fullscreen mode Exit fullscreen mode

In the bad way above, the method is not marked with the AggressiveInlining attribute, so it may not be inlined during JIT compilation, resulting in potentially slower execution.

Good way:

[MethodImpl(MethodImplOptions.AggressiveInlining)]
private int MultiplyByTwo(int value)
{
    return value * 2;
}
Enter fullscreen mode Exit fullscreen mode

By marking critical methods with the AggressiveInlining attribute, you can instruct the JIT compiler to inline them, potentially improving performance by reducing the overhead of method calls.

Stack vs. Heap Allocation

Understanding the difference between stack and heap allocation is essential for C# performance optimization. Let’s explore some tips for efficient allocation that can help you radically speed up your code.

Limit the use of heap-allocated objects when possible

Bad way:

private string GetUserName(int index)
{
    return new string($"User{index}".ToCharArray());
}
Enter fullscreen mode Exit fullscreen mode

Using the new keyword to create a string object introduces heap allocation and contributes towards garbage collection overhead, negatively impacting the overall performance of your application.

Good way:

private string GetUserName(int index)
{
    return $"User{index}";
}
Enter fullscreen mode Exit fullscreen mode

By simply returning the interpolated string, we avoid heap allocation and reduce the overhead provided by garbage collection, which accelerates the performance of your C# code.

Know when to use stackalloc keyword for memory allocation

Bad way:

private double CalculateSum(double[] values)
{
    double sum = 0;
    for (int i = 0; i < values.Length; i++)
    {
        sum += values[i];
    }
    return sum;
}
Enter fullscreen mode Exit fullscreen mode

The bad way code example here uses a double array parameter which might be allocated on the heap, increasing the overhead from garbage collection and impacting .NET performance.

Good way:

private unsafe double CalculateSum(int count)
{
    double sum = 0;
    double* values = stackalloc double[count]; // Allocate memory on the stack
    for (int i = 0; i < count; i++)
    {
        sum += values[i];
    }
    return sum;
}
Enter fullscreen mode Exit fullscreen mode

Using stackalloc, we efficiently allocate memory on the stack, reducing our dependency on the heap and garbage collector. This leads to better .NET performance and potentially faster code execution.

Efficient Data Structures and Algorithms

Choosing the right data structures and algorithms directly impacts your C# performance. Let’s examine some techniques to make better choices with high performance coding with .NET Core and C#.

Choose the right data structure for your needs

Bad way:

List<int> userList = new List<int>();
Enter fullscreen mode Exit fullscreen mode

Using a List to store user identifiers introduces performance bottlenecks, particularly when frequent look-ups are required.

Good way:

HashSet<int> userList = new HashSet<int>();
Enter fullscreen mode Exit fullscreen mode

Selecting a HashSet instead of a List offers faster look-up times and greater performance. Recognizing the suitable data structure is vital for efficient C# coding and solving .NET performance issues.

Employ custom sorting algorithms for specific use cases

Bad way:

double[] values = { 3.14, 1.61, 2.72 };
Array.Sort(values);
Enter fullscreen mode Exit fullscreen mode

Relying on default sorting algorithms may not always be the best choice for specific performance-centric use cases.

Good way:

double[] values = { 3.14, 1.61, 2.72 };
CustomSortAlgorithm.Sort(values);
Enter fullscreen mode Exit fullscreen mode

Employing a custom sorting algorithm can significantly improve your C# performance as it allows you to optimize for your specific needs. This way, you can develop high-performance code that is better suited to your scenarios.

Reflection and Code Generation

Reflection and code generation are powerful tools in C#, but improper usage can slow down your applications. Let’s delve into some best practices to optimize their use and evade .NET performance issues.

Avoid excessive use of Reflection APIs

Bad way:

Type userType = typeof(User);
object user = Activator.CreateInstance(userType);
Enter fullscreen mode Exit fullscreen mode

The bad way code example leverages Reflection APIs for object instantiation, which incurs a notable performance cost.

Good way:

By directly creating a new object using the constructor, you can reduce the burden of runtime overhead associated with reflection APIs, thus enhancing the C# performance.

Use dynamically generated lambda expressions instead of reflection

Bad way:

private static void SetPropertyViaReflection(object obj, PropertyInfo property, object value)
{
    property.SetValue(obj, value);
}
Enter fullscreen mode Exit fullscreen mode

Performance C# suffers when using reflection to set property values due to the additional overhead required to process the operation.

Good way:

private static void SetPropertyViaExpression(object obj, PropertyInfo property, object value)
{
    var setter = property.SetMethod.CreateDelegate(typeof(Action<,>).MakeGenericType(property.DeclaringType, property.PropertyType));
    ((dynamic)setter)(obj, value);
}
Enter fullscreen mode Exit fullscreen mode

Instead of using reflection, dynamically generating lambda expressions can substantially improve performance. By employing the Just-In-Time (JIT) compiler optimization, you can achieve high-performance coding with .NET Core and C#.

SIMD (Single Instruction Multiple Data) using System.Numerics

SIMD can significantly improve performance by processing multiple data elements in parallel. Let’s explore how to utilize SIMD in your C# applications for high-performance coding with .NET Core and C#.

Harness the power of SIMD instructions with Vector

Bad way:

private void Normalize(float[] data)
{
    for (int i = 0; i < data.Length; i++)
    {
        data[i] = data[i] / 2f;
    }
}
Enter fullscreen mode Exit fullscreen mode

The bad way code example processes data elements one by one, which can be slow and limit the C# performance potential.

Good way:

private void Normalize(float[] data)
{
    Vector<float> factor = new Vector<float>(0.5f);
    for (int i = 0; i < data.Length; i += Vector<float>.Count)
    {
        Vector<float> vector = new Vector<float>(data, i);
        (vector * factor).CopyTo(data, i); // Process multiple elements in parallel
    }
}
Enter fullscreen mode Exit fullscreen mode

By employing Vector<T>, you can harness SIMD instructions to process multiple data elements simultaneously. This can lead to substantial performance improvements in your C# code.

Ensure compatibility with hardware-accelerated SIMD

Bad way:

private void Normalize(float[] data)
{
    Vector<float> factor = new Vector<float>(0.5f);
    for (int i = 0; i < data.Length; i += Vector<float>.Count)
    {
        Vector<float> vector = new Vector<float>(data, i);
        (vector * factor).CopyTo(data, i);
    }
}
Enter fullscreen mode Exit fullscreen mode

The bad way code example may not work with some configurations that lack SIMD support, limiting the code’s compatibility across different hardware.

Good way:

private void Normalize(float[] data)
{
    if (Vector.IsHardwareAccelerated) // Check for SIMD support
    {
        Vector<float> factor = new Vector<float>(0.5f);
        for (int i = 0; i < data.Length; i += Vector<float>.Count)
        {
            Vector<float> vector = new Vector<float>(data, i);
            (vector * factor).CopyTo(data, i);
        }
    }
    else
    {
        for (int i = 0; i < data.Length; i++)
        {
            data[i] = data[i] / 2f;
        }
    }
}
Enter fullscreen mode Exit fullscreen mode

By checking for hardware acceleration support with Vector.IsHardwareAccelerated before using SIMD instructions, you ensure your code stays portable and works correctly on different platforms, even on those without SIMD support.

Providing a fallback to regular code when SIMD isn’t available ensures better compatibility across various hardware configurations.

Task and ValueTask for reusing asynchronous code

Leveraging ValueTask<TResult> can help reduce heap allocations and improve performance in asynchronous scenarios. Let’s see how to use it effectively for optimizing .NET code.

Use ValueTask to reduce heap allocations

Bad way:

public async Task<string> ReadDataAsync()
{
    var data = await ReadFromStreamAsync(_stream);
    return ProcessData(data);
}
Enter fullscreen mode Exit fullscreen mode

The bad way code example depends on heap-allocated Task<TResult> objects, which contribute to garbage collection overhead and slow down C# performance.

Good way:

public async ValueTask<string> ReadDataAsync()
{
    var data = await ReadFromStreamAsync(_stream);
    return ProcessData(data);
}
Enter fullscreen mode Exit fullscreen mode

By switching from Task<TResult> to ValueTask<TResult>, you can reduce heap allocations and ultimately improve your C# performance. It’s particularly helpful for high-frequency async operations.

Optimize performance with appropriate async operations

Bad way:

public async Task<int> CalculateAsync(int x, int y)
{
    return await Task.FromResult(x + y);
}
Enter fullscreen mode Exit fullscreen mode

The bad way code example uses unnecessary heap allocations through Task.FromResult that can hamper performance.

Good way:

public ValueTask<int> CalculateAsync(int x, int y)
{
    return new ValueTask<int>(x + y); // Reduce heap allocations
}
Enter fullscreen mode Exit fullscreen mode

Using the appropriate async operation can significantly optimize your C# performance. In cases where a method is likely to complete synchronously or its asynchronous paths can be merged, using ValueTask<TResult> instead of Task<TResult> can help reduce heap allocations and improve performance.

Be mindful of when to choose ValueTask over Task based on specific scenarios and the nature of the asynchronous operations involved.

Detecting and Reducing Boxing and Unboxing

Reducing boxing and unboxing overhead can significantly contribute to your C# performance. Let’s explore some techniques to avoid these costly operations for optimizing .NET code.

Understand the cost of boxing and unboxing

Bad way:

int number = 42;
object boxedNumber = number; // Boxing
int unboxedNumber = (int)boxedNumber; // Unboxing
Enter fullscreen mode Exit fullscreen mode

Boxing and unboxing introduce additional overhead that can have a negative impact on C# performance.

Good way:

int number = 42; // Avoid boxing and unboxing
Enter fullscreen mode Exit fullscreen mode

By being aware of the performance implications of boxing and unboxing, you can make better decisions in your code to avoid unnecessary overhead. Optimize your code by minimizing these operations when possible.

Utilize generics and custom interfaces to avoid boxing

Bad way:

public interface INumber
{
    object Value { get; set; }
}

public class Number : INumber
{
    public object Value { get; set; }
}
Enter fullscreen mode Exit fullscreen mode

The bad way code example incurs boxing overhead due to the use of object types, which can impact C# performance.

Good way:

public interface INumber<T>
{
    T Value { get; set; }
}

public class Number<T> : INumber<T> // Utilize generics to avoid boxing
{
    public T Value { get; set; }
}
Enter fullscreen mode Exit fullscreen mode

Using generics and custom interfaces can help prevent boxing and improve performance. By employing type parameters, you can write more efficient code that avoids boxing overhead for value types and maintains flexibility for reference types.

This results in better C# performance and helps address .NET performance issues.

Network Programming Optimization

Optimizing network communication is crucial for responsive and high-performing C# applications. Let’s learn how to enhance your network programming with expert tips.

Choose efficient serialization methods

C# developers often need to serialize and deserialize data to communicate with external systems or file storage. Choosing an efficient serialization method can significantly impact your C# application’s performance, impacting the productivity of the C# optimizer as well.

Bad way:

// Using XmlSerializer for data serialization
private string SerializeObjectToXml<T>(T obj)
{
    var serializer = new XmlSerializer(typeof(T));
    using (var writer = new StringWriter())
    {
        serializer.Serialize(writer, obj);
        return writer.ToString();
    }
}
Enter fullscreen mode Exit fullscreen mode

XML serialization is a slow and outdated method for data serialization due to its verbose nature. The XmlSerializer generates a large amount of temporary objects and may affect the .NET code optimization techniques in use, resulting in slow performance, increased memory usage, and the risk of blocking the GC.

Good way:

// Using Newtonsoft.Json for data serialization
private string SerializeObjectToJson<T>(T obj)
{
    return JsonConvert.SerializeObject(obj);
}
Enter fullscreen mode Exit fullscreen mode

The good way example shows the use of Newtonsoft.Json – a faster, more efficient library for serialization compared to XmlSerializer.
This library ensures better performance and provides additional features that help optimize code in Visual Studio, allowing C# compiler optimizations to work more effectively.

Use HttpClientFactory to manage HttpClient instances

Not properly reusing HttpClient instances may lead to an exhaustion of available sockets as well as other performance issues. HttpClientFactory enables proper management and reuse of HttpClient instances, reducing the chances of such problems.

Bad way:

// Creating a new HttpClient instance for every request
var httpClient = new HttpClient();
var response = await httpClient.GetAsync("http://api.example.com/data");
Enter fullscreen mode Exit fullscreen mode

Good way:

// Injecting HttpClient into the class using dependency injection and using the HttpClient provided by HttpClientFactory
private readonly HttpClient _httpClient;

public MyClass(HttpClient httpClient)
{
    _httpClient = httpClient;
}

public async Task GetDataAsync()
{
    var response = await _httpClient.GetAsync("http://api.example.com/data");
}
Enter fullscreen mode Exit fullscreen mode

The good way example demonstrates using HttpClientFactory to provide HttpClient instances to your classes via dependency injection. This approach manages the lifetimes of your HttpClient instances more efficiently, preventing socket exhaustion and performance issues that may arise due to improper handling.

Optimizing Exception Handling

Exception handling is a crucial aspect of C# programming, but improper use can result in performance bottlenecks. Let’s see how to handle exceptions efficiently and responsibly.

Avoid using exceptions for flow control

Treating exceptions as a part of the normal application flow can significantly impact C# performance, by generating unnecessary work for the C# optimizer and creating potential performance hiccups in the runtime.

Bad way:

try
{
    int.Parse(input);
}
catch (FormatException)
{
    // Handle the invalid input
}
Enter fullscreen mode Exit fullscreen mode

In the bad way example, trying to parse an invalid input string would throw an exception. Throwing an exception here is not ideal for performance and forces us to handle the FormatException as control flow.

Good way:

if (int.TryParse(input, out int result))
{
    // Use the parsed value
}
else
{
    // Handle the invalid input
}
Enter fullscreen mode Exit fullscreen mode

The good way example leverages the TryParse method to avoid relying in exception for control flow. This approach ensures better C# performance and cleaner code.

Use exception filters to minimize catch blocks

Exception filters help in writing efficient exception handling code that keeps catch blocks more concise and easier to maintain.

Bad way:

try
{
    // Perform an operation
}
catch (Exception ex)
{
    if (ex is InvalidOperationException || ex is ArgumentNullException)
    {
        // Handle the specific exceptions
    }
    else
    {
        throw;
    }
}
Enter fullscreen mode Exit fullscreen mode

In the bad way example, multiple exceptions are caught in a single catch block, with nested if statements used to differentiate between them. This may lead to a more complex and harder to maintain code.

Good way:

try
{
    // Perform an operation
}
catch (Exception ex) when (ex is InvalidOperationException || ex is ArgumentNullException)
{
    // Handle the specific exceptions
}
Enter fullscreen mode Exit fullscreen mode

The good way example demonstrates the use of exception filters. These allow you to catch exceptions only when a certain condition is met, which simplifies your catch blocks and eliminates the need for multiple catch blocks or rethrowing unhandled exceptions.

Nullability and Nullable Reference Types

Handling nullable reference types is a critical part of C# programming, especially for avoiding null reference exceptions. Let’s take a look at some expert tips to safely work with nullable types without hurting performance.

Leverage null-coalescing operators (??, ??=)

Null-coalescing operators help you to write concise and performant code when working with nullable types, ensuring that null values are replaced with a default value.

Bad way:

string input = GetNullableString();
if (input == null)
{
    input = "default";
}
Enter fullscreen mode Exit fullscreen mode

The bad way example demonstrates a verbose and less performant code when dealing with null values.

Good way:

string input = GetNullableString() ?? "default";
Enter fullscreen mode Exit fullscreen mode

The good way example uses a null-coalescing operator, which provides a more concise and efficient way of handling null values in C#. This ensures better C sharp performance and more maintainable code.

Use nullable reference types to avoid runtime null reference exceptions

Nullable reference types, introduced in C# 8.0, help catch potential null reference exceptions at compile-time rather than runtime.

Bad way:

string name = GetName();
int length = name.Length; // Potential NullReferenceException
Enter fullscreen mode Exit fullscreen mode

In the bad way example, we have a potential NullReferenceException that would only be caught at runtime, which can lead to unexpected crashes.

Good way:

string? name = GetName();
int length = name?.Length ?? 0; // No NullReferenceException
Enter fullscreen mode Exit fullscreen mode

By using nullable reference types and null-conditional access in the good way example, you can remove potential null reference exceptions in your code.

This helps create safer and more performant code that’s easier to reason about during both development and execution.

Using Span and Memory for efficient buffer management

Managing memory and buffers efficiently play a crucial role in enhancing C# performance. Here we will examine how Span and Memory can aid in optimizing your code for better efficiency.

Know when to use Span over arrays

Span presents a more performant alternative to arrays in certain situations, enabling manipulation of contiguous memory regions without the need for additional memory allocation or copying.

Bad way:

// Using arrays may lead to unnecessary memory allocations and copying
byte[] data = GetData();
ProcessData(data);
Enter fullscreen mode Exit fullscreen mode

The bad way may result in additional memory allocation and copying which negatively impact performance.

Good way:

// Using Span<T> avoids additional memory allocation and copying
byte[] data = GetData();
Span<byte> dataSpan = data.AsSpan();
ProcessData(dataSpan);
Enter fullscreen mode Exit fullscreen mode

By employing Span<T> in place of arrays, you circumvent unnecessary memory allocations and copying, leading to faster and more efficient code execution.

Use ArrayPool to recycle temporary buffers

ArrayPool is a shared collection of arrays that helps reduce the frequency of allocating and garbage collecting large buffers.

Bad way:

// Allocating a new large buffer
byte[] buffer = new byte[4096];
Enter fullscreen mode Exit fullscreen mode

Creating new buffers this way may cause frequent garbage collection, and thus, lower performance.

Good way:

// Using ArrayPool<T> recycles previously allocated large buffers
var pool = ArrayPool<byte>.Shared;
byte[] buffer = pool.Rent(4096);
try
{
    // Work with the buffer
}
finally
{
    pool.Return(buffer);
}
Enter fullscreen mode Exit fullscreen mode

Using ArrayPool<T> enables your application to reuse previously allocated large buffers, minimizing garbage collection occurrences and improving overall performance.

Lazy and Eager Loading Techniques

Grasping the distinction between lazy and eager loading techniques allows you to create high-performing C# applications. Let’s investigate how to make the correct choices according to your application’s requirements.

Understand the trade-offs between lazy and eager loading

Lazy loading signifies that data is only loaded when required, whereas eager loading fetches all data upfront. Deciding on the appropriate method for your application entails balancing performance and memory consumption.

Bad way:

// Eagerly loading all entities from the database
var customers = dbContext.Customers.Include(c => c.Orders);
Enter fullscreen mode Exit fullscreen mode
// Lazily loading when not required, leading to performance issues
var customer = dbContext.Customers.Find(customerId);
var orderCount = customer.Orders.Count;
Enter fullscreen mode Exit fullscreen mode

The highlighted issues in the bad way can cause poor performance and unnecessary data loading.

Good way:

// Eagerly loading specific data for an operation
var user = dbContext.Users.Include(u => u.Profile).SingleOrDefault(u => u.Id == userId);
Enter fullscreen mode Exit fullscreen mode
// Lazily loading when appropriate for smaller object graphs
var product = dbContext.Products.Find(productId);
var price = product.Price;
Enter fullscreen mode Exit fullscreen mode

By comprehending the trade-offs between lazy and eager loading, you can make informed decisions concerning when to utilize each method, resulting in a more efficient application.

Implement lazy properties with the Lazy class

The Lazy<T> class allows you to create properties that are only initialized when accessed for the first time, which could potentially enhance performance by preventing unnecessary initializations.

Bad way:

// Initializing expensive resources upfront
private readonly ExpensiveObject _expensiveObject = new ExpensiveObject();
public ExpensiveObject ExpensiveObject => _expensiveObject;
Enter fullscreen mode Exit fullscreen mode

An approach like this may lead to a waste of resources and lower performance due to expensive resources being initialized even when not required.

Good way:

// Using Lazy<T> to initialize resources only when needed
private readonly Lazy<ExpensiveObject> _expensiveObject = new Lazy<ExpensiveObject>();
public ExpensiveObject ExpensiveObject => _expensiveObject.Value;
Enter fullscreen mode Exit fullscreen mode

By implementing lazy properties with the Lazy<T> class, you ensure that costly resources are initialized only when essential, resulting in a more efficient and responsive application.

Influence of String Interpolation and Comparison on Performance

Handling strings is a common operation in C# that can have significant consequences on your application’s performance. We will delve into some expert guidance on optimizing string usage.

Use StringComparison options for efficient string comparison

String comparisons are frequent operations that can generate performance bottlenecks. Employing the suitable StringComparison option can enhance the efficiency of string comparison.

Bad way:

// Allocating additional memory for string conversion before comparison
bool equal = string1.ToLower() == string2.ToLower();
Enter fullscreen mode Exit fullscreen mode

This approach results in unnecessary string allocations before comparison, leading to performance degradation.

Good way:

// Comparing strings directly using StringComparison options
bool equal = string.Equals(string1, string2, StringComparison.OrdinalIgnoreCase);
Enter fullscreen mode Exit fullscreen mode

By leveraging StringComparison options, you can avoid needless string allocations (e.g., ToLower) and carry out more efficient string comparisons.

Opt for StringBuilder over string concatenation in loops

When concatenating strings within loops or performing multiple string manipulations, using StringBuilder is more efficient and can lead to significant performance improvements.

Bad way:

// Concatenating strings in a loop creates many intermediate strings
string result = string.Empty;
for (int i = 0; i < 1000; i++)
{
    result += "Iteration: " + i;
}
Enter fullscreen mode Exit fullscreen mode

The bad way generates many intermediate strings during concatenation, which can lead to a significant drop in performance.

Good way:

// StringBuilder minimizes string allocations and deallocations
StringBuilder sb = new StringBuilder();
for (int i = 0; i < 1000; i++)
{
    sb.AppendFormat("Iteration: {0}", i);
}
string result = sb.ToString();
Enter fullscreen mode Exit fullscreen mode

Utilizing StringBuilder helps minimize the number of new string allocations and deallocations, resulting in better performance and lower memory consumption.

By incorporating these expert tips into your C# programming, you can significantly improve the performance of your applications and write efficient, well-optimized code. Mastering advanced C# performance techniques is a key skill for senior developers who want to take their skills to the next level.

Top comments (24)

Collapse
 
deexter profile image
deexter

Top article

Collapse
 
bytehide profile image
ByteHide

Great that you like it!

Collapse
 
canro91 profile image
Cesar Aguirre

I think you actually meant Task vs ValueTask...

Collapse
 
bytehide profile image
ByteHide

Fixed! Thanks for the notice😅😁

Collapse
 
ezrankings profile image
Mansi Rana

Thanks for sharing This amazing article.

Collapse
 
bytehide profile image
ByteHide

A lot of thanks Mansi! :D

Collapse
 
j0nimost profile image
John Nyingi

you missed a really good one IAsyncEnumerable it's powerful

Collapse
 
bytehide profile image
ByteHide

Wow, you are so right😥! Noted and thank you very much for the comment!

Collapse
 
rezanouri profile image
reza-noori

Great
Thanks ❤️

Collapse
 
bytehide profile image
ByteHide

Thanks to you for taking the time to read it!

Collapse
 
michalispapamichael profile image
Michalis Papamichael

Great Article!

Collapse
 
bytehide profile image
ByteHide

Thanks!

Collapse
 
ant_f_dev profile image
Anthony Fung

This is a comprehensive list of great tips - thanks for sharing.

Collapse
 
bytehide profile image
ByteHide

Thank you! If you have any tips you would like to share, we'd love to hear from you! All this is open-source 😎

Collapse
 
slavius profile image
Slavius

I just want to warn everyone that many of the tips are very opinionated and some are just blatantly copy/pasted without really understanding the problem or explaining the context in which they are appropriate while making them risky in your code.
Few examples are ConfigureAwait(false) and use of stackalloc instead of e.g. Span<T>.

Collapse
 
lico profile image
SeongKuk Han

Awesome 🔥.
I'm not into C# now, I'm mainly working with js, but there are lots of things I could learn and can apply for me such as Avoid premature optimizations, importance of caching data, avoid using exceptions for flow control, utilize generics and custom interfaces to avoid boxing...
It was also interesting seeing what what they think as important things in C# and the way it works, especially about parallelism and features and effort to work with multi-threaded system.

Collapse
 
lidiaaadotnet profile image
lidiaaadotnet

Now that's a real top😍

Collapse
 
dmitryeremenko715 profile image
dmitryeremenko715

Hello, I'm interested in finding a partner to work on .NET projects with. I noticed that you're also interested in this space and wanted to reach out to see if you might be interested in discussing the possibility of working together.

I value clear communication and collaboration, and I'm committed to finding a good fit for both of us. Let me know if you're open to discussing this further, and we can set up a time to chat more.

Telegram: @krumfi_cto
Gmail: dmitryeremenko715@gmail.com

Looking forward to hearing from you soon.
Best regards,
Dmitry

Collapse
 
sbenzenko profile image
Sergey Benzenko • Edited

Could you please show the source of information that you used for the example where you used LINQ AsParallel() within Parallel.ForEach?
I could not find this kind of usages of AsParallel anywhere. And in the documentation AsParallel is used for the whole different purpose entirely (not for thread safety within Parallel.ForEach).

Thank you in advance.

Collapse
 
rafo profile image
Rafael Osipov

Thank you for sharing. It is really valuable information here.

Collapse
 
ghadzhigeorgiev profile image
Georgi Hadzhigeorgiev

Great one, cheers!

Collapse
 
trungnguyen1697 profile image
Trung Nguyen Thanh

Thanks for sharing

Collapse
 
chamacr profile image
ChamaCR

Awesome tips, thanks!