Micro-Optimization and Meatballs

January 15, 2005

In my previous entry on the real cost of performance, there were some complaints that my code's slow and it sucks. If I had a nickel every time someone told me that, I could have retired years ago. Let's take a look at the specific complaint that the s <> "" comparison is inefficient, using low-level windows API timing in the Stopwatch class:

Const iterations As Integer = 1000000
Dim s As String = "sample string"
Dim sw As New Stopwatch
Dim n As Integer

n = 0
sw.Start()
For i As Integer = 1 To iterations
    If s.Length = 0 Then
        n += 1
    End If
Next
sw.Stop()
Console.WriteLine(sw.ElapsedMs)

n = 0
sw.Start()
For i As Integer = 1 To iterations
    If s = String.Empty Then
        n += 1
    End If
Next
sw.Stop()
Console.WriteLine(sw.ElapsedMs)

n = 0
sw.Start()
For i As Integer = 1 To iterations
    If s = "" Then
        n += 1
    End If
Next
sw.Stop()
Console.WriteLine(sw.ElapsedMs)

n = 0
sw.Start()
For i As Integer = 1 To iterations
    If s.Equals(String.Empty) Then
        n += 1
    End If
Next
sw.Stop()
Console.WriteLine(sw.ElapsedMs)

Here are the results:

So, yes, String.Length is five (or more) times faster. And yes, using String.Equals is twice as fast. However, neither of those will work when the string is Nothing, and we're still talking about a difference of 30 milliseconds, on the slowest computer I own, over a MILLION string comparisons! This brings to mind a Bill Murray quote from Meatballs: It just doesn't matter! It just doesn't matter!

Arguments about which method results in code that is easier to read and easier to maintain will be gladly entertained. Arguments about speed will not. Stop micro-optimizing and start macro-optimizing: per Lippert, code that makes sense is code which can be analyzed and maintained, and that makes it performant.

If you'd like to time this yourself, here's a stopwatch class which uses the high resolution API counters. Good luck-- you're gonna need it. The resolution, I mean.