Performance of a C# application built on AnyCPU vs x64 platform on a 64 bit machine

c# .net clr
Saurabh R S picture Saurabh R S · Jun 16, 2012 · Viewed 13.7k times · Source

I have to deploy a C# application on a 64 bit machine though there is a slight probability that it could also be deployed on a 32 bit machine. Should I build two separate executables targeting x86 and x64 platform or should I go for a single executable built targeting 'AnyCPU' platform (specified in the project property's Build option'. Would there be any performace difference between a C# assembly built targeting 'AnyCPU' is deployed on a 64 bit machine vs the same assembly built targeting specifically 'x64' platform ?

Answer

svick picture svick · Jun 16, 2012

No, there is no difference in performance between AnyCPU application running on a 64-bit Windows and an x64 application running on it. The only thing that flag changes are some flags in the header of the compiled assembly and the CLR uses it only to decide whether to use x86 or x64, nothing else

If you were asking whether there is a difference between x86 application running on a 64-bit Windows and an x64 (or AnyCPU), then the answer would be yes. The differences between the two are:

  • 64-bit obviously uses references that are twice as big as 32-bit, which means larger memory consumption, but it also means you can use more memory
  • 64-bit can use more registers that are available only in the 64-bit mode of CPUs
  • 64-bit JIT is different from 32-bit JIT, it has different set of optimizations: for example 64-bit JIT sometimes uses the tail call optimization even if you don't specifically request it using the tail. instruction (which for example C# never does)

Related questions

What are major differences between C# and Java?

I just want to clarify one thing. This is not a question on which one is better, that part I leave to someone else to discuss. I don't care about it. I've been asked this question on my job interview …

Read More
Try-catch speeding up my code?

I wrote some code for testing the impact of try-catch, but seeing some surprising results. static void Main(string[] args) { Thread.CurrentThread.Priority = ThreadPriority.Highest; Process.GetCurrentProcess().PriorityClass = ProcessPriorityClass.RealTime; long start = 0, stop = 0, elapsed = 0; double avg = 0.0; long temp = Fibo(1); for (…

Read More
Float vs Double Performance

I did some timing tests and also read some articles like this one (last comment), and it looks like in Release build, float and double values take the same amount of processing time. How is this possible? When float is …

Read More