Different ways of null checking in C#

It is my understanding that the second one does have a slight performance improvement. Something about only having to see if the second object is null or not rather than the first way it has to load the actual object and then compare against null.

I do not know the technical explanation real well but I had also learned once that it is slightly faster. I rarely write it that way because it does not read as well to me so it just depends on if saving a few microseconds is important to you or not.

Quote:

few says there are performance improvements when we use 2

Yes here is the proof
randomly ran the following code for 5 times and the results as below


obj == null

object obj = null;
     System.Diagnostics.Stopwatch watch = new System.Diagnostics.Stopwatch();
     watch.Start();
     if (obj == null) { }
     watch.Stop();
     var a = watch.Elapsed.Ticks;

null == obj1

object obj1 = null;
            System.Diagnostics.Stopwatch watch1 = new System.Diagnostics.Stopwatch();
            watch1.Start();
            if (null == obj1) { }
            watch1.Stop();
            var b = watch1.Elapsed.Ticks;

Result

Elapsed Ticks
obj == null	null == obj1
37	0
32	2
27	0
135	18
37	9

Basicly I agree with Richard and Jochen.
Because of the results of Karthik I made my own test because I couldn't believe those results.

I tested it with VB.Net and also with C#.Net with this Code-Blocks :

Dim myObject As Object = Nothing
Dim x As Boolean

Dim d1 As Date = Now

For i As Integer = 1 To 100000000
    If (myObject Is Nothing) Then x = True
Next

Dim d2 As Date = Now

For i As Integer = 1 To 100000000
    If (Nothing Is myObject) Then x = True
Next

Dim d3 As Date = Now

Dim ts1 As TimeSpan = d2 - d1
Dim ts2 As TimeSpan = d3 - d2

object myObject = null ;
bool x = false;

System.DateTime d1 = System.DateTime.Now;

for (int i = 1; i <= 100000000; i++)
{
    if ((myObject == null))
        x = true;
}

System.DateTime d2 = System.DateTime.Now;

for (int i = 1; i <= 100000000; i++)
{
    if ((null == myObject))
        x = true;
}

System.DateTime d3 = System.DateTime.Now;

TimeSpan ts1 = d2 - d1;
TimeSpan ts2 = d3 - d2;

These are my results (by doing the compare 100 Million times :

Compare Timespan with VB Timespan with C#
myObj (assigned) to Nothing / Null 0,2390137 seconds 0,2808005 seconds
Nothing / Null to myObj (assigned) 0,3090174 seconds 0,2808006 seconds
myObj (not assigned) to Nothing / Null 0,3588006 seconds 0,3900007 seconds
Nothing / Null to myObj (not assigned) 0,2962005 seconds 0,2744006 seconds

Int-Constant to Integer-Variable 0,2652005 seconds 0,2652005 seconds
Int-Variable to Integer-Constant 0,2652005 seconds 0,2652005 seconds

I think that an average over 100 Million cycles says much more that an only 1 cycle-test (sorry Karthik) which shows different values with a big range.
But for me it was also very interesting that there is a difference (by handling Objects) between VB and C# ...

  object myObject = null;


        System.Diagnostics.Stopwatch watch1 = new System.Diagnostics.Stopwatch();
        watch1.Start();


        for (int i = 1; i <= 100000000; i++)
        {
            if (null == myObject) { }

        }
        watch1.Stop();


        System.Diagnostics.Stopwatch watch = new System.Diagnostics.Stopwatch();
        watch.Start();

        for (int i = 1; i <= 100000000; i++)
        {
            if (myObject == null){}
               
        }
        watch.Stop();

      
        var a = watch.Elapsed.TotalMilliseconds;
        var b = watch1.Elapsed.TotalMilliseconds;