CODEWITHSUNDEEP
c#performanceloopscount
Sturm
Sturm

Reputation: 4125

Performance: List.Count vs checking a stored variable

I wonder if this makes any difference:

 for (int i = 0; i < values.Count; i++)
        {
            //
        }

vs

int num = values.Count;

for(int=0; i<num; i++)
{

}

I think the second approach is better because you don't need to count all the items in each iteration. But I may be wrong. Could someone illuminate me?

Upvotes: 21

Views: 11552

Answers (4)

user3029478
user3029478

Reputation: 196

I tried it both ways, and using "int num=values.Count;" won by a large margin. My test used Performance Analyzer and ran the approaches back to back. Then I switched which one runs first, and again storing it to an integer won. I guess accessing List.Count is slower and the compiler doesn't optimize it to only run once.

Upvotes: 0

ProVega
ProVega

Reputation: 5914

Well you can look at the .NET Source code here http://referencesource.microsoft.com/#mscorlib/system/collections/generic/list.cs#aa7e01fcb80a917e

  public int Count {
        get {
            Contract.Ensures(Contract.Result<int>() >= 0);
            return _size; 
        }
    }

It would appear that .Count property on list does a quick internal check and then returns _size. So it should be pretty close to the performance of you storing the value yourself.

Upvotes: 18

Dave Bish
Dave Bish

Reputation: 19646

This entirely depends on what values is. Count could be implemented completely differently, depending on what the type of this object is.

If you are talking about a generic List<T> - then Count is implemented as an internal property that isn't re-evaluated - and therefore is the better choice.

Upvotes: 3

Bas
Bas

Reputation: 27105

The list already stores its Count internally. The comparison you are doing is related to code style, not performance. Since the compiler will optimize the retrieval of 'Count'

Upvotes: 19

Related Questions