Databases
StackOverflow2013
Postsdbo
POStatic constructor performance and why we can't specify beforefieldinit
 

Note that there are some explanatory texts on larger screens.

plurals
  1. POStatic constructor performance and why we can't specify beforefieldinit
    primarykey
    data
    text
    <p>I came across a difference in speed using the following two structs:</p> <pre><code>public struct NoStaticCtor { private static int _myValue = 3; public static int GetMyValue() { return _myValue; } } public struct StaticCtor { private static int _myValue; public static int GetMyValue() { return _myValue; } static StaticCtor() { _myValue = 3; } } class Program { static void Main(string[] args) { long numTimes = 5000000000; // yup, 5 billion Stopwatch sw = new Stopwatch(); sw.Start(); for (long i = 0; i &lt; numTimes; i++) { NoStaticCtor.GetMyValue(); } sw.Stop(); Console.WriteLine("No static ctor: {0}", sw.Elapsed); sw.Restart(); for (long i = 0; i &lt; numTimes; i++) { StaticCtor.GetMyValue(); } sw.Stop(); Console.WriteLine("with static ctor: {0}", sw.Elapsed); } } </code></pre> <p>Which produces the results:</p> <pre><code>Release (x86), no debugger attached: No static ctor: 00:00:05.1111786 with static ctor: 00:00:09.9502592 Release (x64), no debugger attached: No static ctor: 00:00:03.2595979 with static ctor: 00:00:14.5922220 </code></pre> <p>The compiler produces a static constructor for the <code>NoStaticCtor</code> that is identical to the one explicitly declared in <code>StaticCtor</code>. I understand that the compiler will only emit <code>beforefieldinit</code> when a static constructor is not explicitly defined.</p> <p>They produce <em>almost</em> identical il code, except for one difference, declaring the struct with <code>beforefieldinit</code>, which is where I feel the difference lies since I know it determines when the type constructor is called, though I can't quite figure out why there's such a difference. It assume it isn't <em>calling</em> the type constructor every iteration, since a type constructor can only be called once.<sup>1</sup></p> <p>So,</p> <p>1) Why the time difference between the struct with <code>beforefieldinit</code> and the one without? (I imagine the JITer is doing something extra in the for loop, however, I have no clue how to view the output of the JITer to see what.</p> <p>2) Why did the compiler designers a) not make all structs <code>beforefieldinit</code> be default and b) not give developers the ability to explicitly specify that behavior? Of course, this <em>is</em> assuming you can't, as I haven't been able to find a way.</p> <hr> <p>Edit:</p> <p><sup>1</sup>. <a href="http://pastebin.com/w1AHfCiF" rel="nofollow">I modified the code</a>, essentially running each loop a second time, expecting a improvement, but it wasn't much:</p> <pre><code>No static ctor: 00:00:03.3342359 with static ctor: 00:00:14.6139917 No static ctor: 00:00:03.2229995 with static ctor: 00:00:12.9524860 Press any key to continue . . . </code></pre> <p>I did this because I though, well, <em>maybe</em>, however unlikely it is, the JITer was actually calling the type constructor every iteration. It seems to me the JITer would know the type constructor has already been called and not emit code to do that when the second loop was compiled.</p> <p>In addition to Motti's answer: <a href="http://pastebin.com/BGJYEBHN" rel="nofollow">This code</a> produces better results, because of difference in JITing, the JITing of <code>DoSecondLoop</code> doesn't emit the static ctor check, because it detected it was done previously in <code>DoFirstLoop</code>, causing each loop to perform at the same speed. (~3 seconds)</p>
    singulars
    1. PO
      singulars
      1. PTAnswer
    1. This table or related slice is empty.
    1. PTQuestion
    1. USMotti Shaked
    1. USChristopher Currens
    plurals
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. PO
      singulars
      1. PTAnswer
    1. VO
      singulars
      1. POStatic constructor performance and why we can't specify beforefieldinit
      1. This table or related slice is empty.
      1. VTUpMod
    2. VO
      singulars
      1. POStatic constructor performance and why we can't specify beforefieldinit
      1. This table or related slice is empty.
      1. VTUpMod
    3. VO
      singulars
      1. POStatic constructor performance and why we can't specify beforefieldinit
      1. USIgor Korkhov
      1. VTFavorite
    1. COSome perspective on large numbers is required here. The overhead you measured is *one nanosecond*. Yes, that's about what a test + jump instruction takes.
      singulars
      1. POStatic constructor performance and why we can't specify beforefieldinit
      1. USHans Passant
    2. CO@Hans I know the overhead is small. I wouldn't ever actually write code like this used in production. I'm currently on a "how does the CLR work" stint and I've been messing around with unusual things. I'm really just trying to understand why the JITer makes the decisions it does, based on the attributes the compiler emits. I should probably buy a book.
      singulars
      1. POStatic constructor performance and why we can't specify beforefieldinit
      1. USChristopher Currens
 

Querying!

 
Guidance
A row detail

Detail views are divided into sections. All the information in the data section comes from columns in the selected row. The other sections display data from other, related rows.

Related data can be related in a to-one or a to-many fashion. Captions of data related in a to-many fashion link to a list view showing a filtered view of the table.

Try moving around until you find a non-empty to-many entry and click on the label to get to one. You can move back to the root by clicking on the database name in the header.

SQuiL has stopped working due to an internal error.

If you are curious you may find further information in the browser console, which is accessible through the devtools (F12).

Reload