Databases
StackOverflow2013
Postsdbo
POCallback with PInvoke is very slow
 

Note that there are some explanatory texts on larger screens.

plurals
  1. POCallback with PInvoke is very slow
    primarykey
    data
    text
    <p>I'm using a native/unmanaged C++ DLL in my C# application. I call a native function to register a callback method in C# using PInvoke:</p> <pre><code>[DllImport("MyHook.dll", CallingConvention = CallingConvention.Cdecl)] [return: MarshalAs(UnmanagedType.U1)] // necessary due to http://dotnet.dzone.com/articles/pinvoke-c-bool-return-values private static extern bool InstallHook(CallbackPrototype callback); private delegate int CallbackPrototype(byte* bytes, int size, int max); // ... var callback = new CallbackPrototype(MyCallback); _callbackHandle = GCHandle.Alloc(callback); var result = InstallHook(callback); // somewhere when I'm removing the hook, I call _callbackHandle.Free(); </code></pre> <p>Then, when an event happens, the native C++ DLL calls my callback.</p> <pre><code>private int MyCallback(byte* bytes, int size, int max) { //var buffer = new byte[size]; // A //Marshal.Copy(new IntPtr(bytes), buffer, 0, size); // B //WrapperInstance.ParseBytes(buffer, 0, size); // C var x = 1 + 1; // D return size; } </code></pre> <p>When everything is commented out, it works well. If I start uncommenting line A in the callback, it gets veeery slow. I notice it because the callback is called in a critical moment. It isn't acceptable that the callback runtime duration is so long that I notice it as a human being.</p> <p>How could this happen? The <em>size</em> parameter is only from one to 100. Thats only 100 bytes at maximum that need to be allocated - very few for todays computers. I need improvements to make the runtime of the callback much faster. Any ideas?</p> <p>To go further in testing, I've added line D. This does not have any impact.</p> <p>Edit: For more detailed testing I've allocated a big buffer when the class is created. So line A is not in the callback anymore. If line B is then commented out, it suddenly starts taking a huge amount of time again. This must have something to do with the managed code / virtual machine. I hope there are possibilities to make it faster.</p>
    singulars
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. PTQuestion
    1. USThe Wavelength
    1. USThe Wavelength
    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. This table or related slice is empty.
    1. COSeeing code being slowed-down by *removing* an expensive method call is of course entirely unintuitive. Clearly you'll need to stop the trial-and-error approach and gather *evidence* by using a profiler.
      singulars
      1. POCallback with PInvoke is very slow
      1. USHans Passant
    2. COCould your callback be getting invoked an enormous number of times?
      singulars
      1. POCallback with PInvoke is very slow
      1. USChaseMedallion
    3. COIt is called approx. once in two seconds. Hans Passant, if I get you correctly you misunderstand something. The call slows down when I UNcomment line A. Sorry if I got you wrong, my english isn't very good at all. I can't see why I need to use a profiler when I notice performance issues with my eyes.
      singulars
      1. POCallback with PInvoke is very slow
      1. USThe Wavelength
 

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