Performance of calling delegates vs methods

Following this question - https://stackoverflow.com/questions/2082615/pass-method-as-parameter-using-c and some of my personal experience I'd like to know a little more about the performance of calling a delegate vs just calling a method in C#.

Although delegates are extremely convenient, I had an app that did lots of callbacks via delegates and when we rewrote this to use callback interfaces we got an order of magnitude speed improvement. This was with .NET 2.0 so I'm not sure how things have changed with 3 and 4.

How are calls to delegates handled internally in the compiler/CLR and how does this affect performance of method calls?

---

EDIT - To clarify what I mean by delegates vs callback interfaces.

For asynchronous calls my class could provide an OnComplete event and associated delegate which the caller could subscribe to.

Alternatively I could create an ICallback interface with an OnComplete method that the caller implements and then registers itself with the class that will then call that method on completion (i.e. the way Java handles these things).

I haven't seen that effect - I've certainly never encountered it being a bottleneck.

Here's a very rough-and-ready benchmark which shows (on my box anyway) delegates actually being faster than interfaces:

using System;
using System.Diagnostics;

interface IFoo
{
    int Foo(int x);
}

class Program : IFoo
{
    const int Iterations = 1000000000;
    
    public int Foo(int x)
    {
        return x * 3;
    }
    
    static void Main(string[] args)
    {
        int x = 3;
        IFoo ifoo = new Program();
        Func<int, int> del = ifoo.Foo;
        // Make sure everything's JITted:
        ifoo.Foo(3);
        del(3);

        Stopwatch sw = Stopwatch.StartNew();        
        for (int i = 0; i < Iterations; i++)
        {
            x = ifoo.Foo(x);
        }
        sw.Stop();
        Console.WriteLine("Interface: {0}", sw.ElapsedMilliseconds);

        x = 3;
        sw = Stopwatch.StartNew();        
        for (int i = 0; i < Iterations; i++)
        {
            x = del(x);
        }
        sw.Stop();
        Console.WriteLine("Delegate: {0}", sw.ElapsedMilliseconds);
    }
}

Results (.NET 3.5; .NET 4.0b2 is about the same):

Interface: 5068
Delegate: 4404

Now I don't have particular faith that that means delegates are really faster than interfaces... but it makes me fairly convinced that they're not an order of magnitude slower. Additionally, this is doing almost nothing within the delegate/interface method. Obviously the invocation cost is going to make less and less difference as you do more and more work per call.

One thing to be careful of is that you're not creating a new delegate several times where you'd only use a single interface instance. This could cause an issue as it would provoke garbage collection etc. If you're using an instance method as a delegate within a loop, you will find it more efficient to declare the delegate variable outside the loop, create a single delegate instance and reuse it. For example:

Func<int, int> del = myInstance.MyMethod;
for (int i = 0; i < 100000; i++)
{
    MethodTakingFunc(del);
}

is more efficient than:

for (int i = 0; i < 100000; i++)
{
    MethodTakingFunc(myInstance.MyMethod);
}

Could this have been the problem you were seeing?

I find it completely implausible that a delegate is substantially faster or slower than a virtual method. If anything the delegate should be negligibly faster. At a lower level, delegates are usually implemented something like (using C-style notation, but please forgive any minor syntax errors as this is just an illustration):

struct Delegate {
    void* contextPointer;   // What class instance does this reference?
    void* functionPointer;  // What method does this reference?
}

Calling a delegate works something like:

struct Delegate myDelegate = somethingThatReturnsDelegate();
// Call the delegate in de-sugared C-style notation.
ReturnType returnValue = 
    (*((FunctionType) *myDelegate.functionPointer))(myDelegate.contextPointer);

A class, translated to C, would be something like:

struct SomeClass {
    void** vtable;        // Array of pointers to functions.
    SomeType someMember;  // Member variables.
}

To call a vritual function, you would do the following:

struct SomeClass *myClass = someFunctionThatReturnsMyClassPointer();
// Call the virtual function residing in the second slot of the vtable.
void* funcPtr = (myClass -> vtbl)[1];
ReturnType returnValue = (*((FunctionType) funcPtr))(myClass);

They're basically the same, except that when using virtual functions you go through an extra layer of indirection to get the function pointer. However, this extra indirection layer is often free because modern CPU branch predictors will guess the address of the function pointer and speculatively execute its target in parallel with looking up the address of the function. I've found (albeit in D, not C#) that virtual function calls in a tight loop are not any slower than non-inlined direct calls, provided that for any given run of the loop they're always resolving to the same real function.

Since CLR v 2, the cost of delegate invocation is very close to that of virtual method invocation, which is used for interface methods.

See Joel Pobar's blog.

I did some tests (in .Net 3.5... later I will check at home using .Net 4). The fact is: Getting an object as an interface and then executing the method is faster than getting a delegate from a method then calling the delegate.

Considering the variable is already in the right type (interface or delegate) and simple invoking it makes the delegate win.

For some reason, getting a delegate over an interface method (maybe over any virtual method) is MUCH slower.

And, considering there are cases when we simple can't pre-store the delegate (like in Dispatches, for example), that may justify why interfaces are faster.

Here are the results:

To get real results, compile this in Release mode and run it outside Visual Studio.

> Checking direct calls twice
> 00:00:00.5834988
> 00:00:00.5997071

> Checking interface calls, getting the interface at every call
> 00:00:05.8998212 > > Checking interface calls, getting the interface once
> 00:00:05.3163224 > > Checking Action (delegate) calls, getting the action at every call
> 00:00:17.1807980 > > Checking Action (delegate) calls, getting the Action once
> 00:00:05.3163224 > > Checking Action (delegate) over an interface method, getting both at > every call
> 00:03:50.7326056 > > Checking Action (delegate) over an interface method, getting the > interface once, the delegate at every call
> 00:03:48.9141438 > > Checking Action (delegate) over an interface method, getting both once
> 00:00:04.0036530

As you can see, the direct calls are really fast. Storing the interface or delegate before, and then only calling it is really fast. But having to get a delegate is slower than having to get an interface. Having to get a delegate over an interface method (or virtual method, not sure) is really slow (compare the 5 seconds of getting an object as an interface to the almost 4 minutes of doing the same to get the action).

The code that generated those results is here:

using System;

namespace ActionVersusInterface
{
	public interface IRunnable
	{
		void Run();
	}
	public sealed class Runnable:
		IRunnable
	{
		public void Run()
		{
		}
	}
	
	class Program
	{
		private const int COUNT = 1700000000;
		static void Main(string[] args)
		{
			var r = new Runnable();

			Console.WriteLine("To get real results, compile this in Release mode and");
			Console.WriteLine("run it outside Visual Studio.");
			
			Console.WriteLine();
			Console.WriteLine("Checking direct calls twice");
			{
				DateTime begin = DateTime.Now;
				for (int i = 0; i < COUNT; i++)
				{
					r.Run();
				}
				DateTime end = DateTime.Now;
				Console.WriteLine(end - begin);
			}
			{
				DateTime begin = DateTime.Now;
				for (int i = 0; i < COUNT; i++)
				{
					r.Run();
				}
				DateTime end = DateTime.Now;
				Console.WriteLine(end - begin);
			}

			Console.WriteLine();
			Console.WriteLine("Checking interface calls, getting the interface at every call");
			{
				DateTime begin = DateTime.Now;
				for (int i = 0; i < COUNT; i++)
				{
					IRunnable interf = r;
					interf.Run();
				}
				DateTime end = DateTime.Now;
				Console.WriteLine(end - begin);
			}

			Console.WriteLine();
			Console.WriteLine("Checking interface calls, getting the interface once");
			{
				DateTime begin = DateTime.Now;
				IRunnable interf = r;
				for (int i = 0; i < COUNT; i++)
				{
					interf.Run();
				}
				DateTime end = DateTime.Now;
				Console.WriteLine(end - begin);
			}

			Console.WriteLine();
			Console.WriteLine("Checking Action (delegate) calls, getting the action at every call");
			{
				DateTime begin = DateTime.Now;
				for (int i = 0; i < COUNT; i++)
				{
					Action a = r.Run;
					a();
				}
				DateTime end = DateTime.Now;
				Console.WriteLine(end - begin);
			}

			Console.WriteLine();
			Console.WriteLine("Checking Action (delegate) calls, getting the Action once");
			{
				DateTime begin = DateTime.Now;
				Action a = r.Run;
				for (int i = 0; i < COUNT; i++)
				{
					a();
				}
				DateTime end = DateTime.Now;
				Console.WriteLine(end - begin);
			}


			Console.WriteLine();
			Console.WriteLine("Checking Action (delegate) over an interface method, getting both at every call");
			{
				DateTime begin = DateTime.Now;
				for (int i = 0; i < COUNT; i++)
				{
					IRunnable interf = r;
					Action a = interf.Run;
					a();
				}
				DateTime end = DateTime.Now;
				Console.WriteLine(end - begin);
			}

			Console.WriteLine();
			Console.WriteLine("Checking Action (delegate) over an interface method, getting the interface once, the delegate at every call");
			{
				DateTime begin = DateTime.Now;
				IRunnable interf = r;
				for (int i = 0; i < COUNT; i++)
				{
					Action a = interf.Run;
					a();
				}
				DateTime end = DateTime.Now;
				Console.WriteLine(end - begin);
			}

			Console.WriteLine();
			Console.WriteLine("Checking Action (delegate) over an interface method, getting both once");
			{
				DateTime begin = DateTime.Now;
				IRunnable interf = r;
				Action a = interf.Run;
				for (int i = 0; i < COUNT; i++)
				{
					a();
				}
				DateTime end = DateTime.Now;
				Console.WriteLine(end - begin);
			}
			Console.ReadLine();
		}
	}
	
}

What about the fact that delegates are containers? Doesn't the multicast ability add overhead? While we are on the subject, what if we push this container aspect a little further? Nothing forbids us, if d is a delegate, from executing d += d; or from building an arbitrarily complex directed graph of (context pointer, method pointer) pairs. Where can I find the documentation describing how this graph is traversed when the delegate is called?