When my XNA application is inactive or when I'm draggin it or clicked
other folder or even the desktop XNA application framerates reduced
tremendously, assuming your XNA active application is running at 150
frames per second then try to click other folder or make your application
inactive. the framerates per second reduce from 150 to 35 FPS.

:::

This is not good if your testing or simulating a network logic on a
single machine. I'm trying to load at least three XNA application to
test my network library a server and two client running on a single PC.

:::

For me to make this feasible, the three application should be running
at the same speed at least on the frame per second update. I need to
make them  running 60 frames per second even each one of them is not
an active application.

:::

//To make the application running on it's full or specified speed even
//when the application is inactive, I need to adjust the Game.InactiveSleepTime
//to zero on span time.
//
Game.InactiveSleepTime = System.TimeSpan.Zero;

//Now to make the three application running at 60 frames per second
//setting fixed time step to true and specifying the target elapsed time
//will do the trick.
//
//
Game.FixedTimeStep       = true;
Game.TargetElapsedTime = TimeSpan.FromMilliseconds(  17  );  // = 1000/60FPS

:::

That's it, I can now continue to test my network library on a single
machine and I made a simple video feeds for this. syncronization of
1000 network objects rendered by host and the client on same machine.
tho, the client may received garbage data from time to time, it should
still run and the game should not be interrupted, Maybe tonight
when I got home I can test it using my other PC on lan set-up to adjust
the client's interpolation of network objects and some latency fine-tuning.

it I'm now playing RF On-Line Episode 2 Pioneers of Novus and I'm still on Lvl 40+  I know that

much aware of this   and as a game developer hobbyist a built-in linking up features would be a

and use their GUIDE features  but now there are two application running in the background my

networking support is my choice, at least for now . Most of the functionality or methods I want to be
included from other network library are mostly already available in .NET framework we'll since this is

just my first attempt in writing my own network service in managed code, let's see how far I can go

punching, crunching, munching and NAT holing  just to make a remote peer connection. but I really think I

My network support, How easy I want it to be? I'm planning to implement a no receiving
approach  meaning I dont have to expose the actual receiving of packets coming
from the port I'm listening, rather a game logic just need to scan the Z.Network.Messenger
if there's a new message throwned by the system. The messages in Z.Network.Messenger
is process internally so I don't need to implement any receiving routine in my game logic,

EDIT : 03/12/2009 : Hmmm... most of my visitors are looking into MY NETWORK DATA FLOW DIAGRAM, to those interested in my C# Network library project, please go to my other live spaces, specifically for my network game library project under managed code,  check this link  http://dexterz.spaces.live.com/

I  remember  when  I  was  still  working on  my  older  project,  though  it’s  not  using  XNA API,  and  I  did
not  implement  any  shading  stuff  then,  I usually  disable  my  skybox when I'm debugging  or  testing  my
application, due  to  the  fact  that  rendering  a  skybox  really  eats up much of  my frame rates.  and I think 
the  largest  triangle  to  be  rendered   and  filling  the  screen  is a skybox,  specially if the camera is looking 
to one of it's four  corners, I’m rendering at least 8 triangles  or  more  in  some  angle,  Ok  I know rendering
8 triangles is nothing,  but hey,  rendering the skybox  is filling the entire  screen that  makes it slow.  Sky box
consist of 6 sides multiply by 2 then  I  got  12  triangle  primitives and  each side uses a separate texture so I
got  6 separate  textures to  handle the  materials  for each side, Oh yes, each primitive needs it's position and
texture coordinates.  Well, at least  I'm doing some culling  stuff back then,  I'm using triangle strip not triangle
list for  me to check each of s kybox triangle primitive against  the camera  bounding  frustum view. Rendering
a skybox with 12 primitives, 6 textures, transforming it's position from camera and triangle culling test, are the
ways I implemented my skybox before. Whew!.

I am just wondering, how can I eliminate the  culling  process  for each triangle strip and minimize the number
of  primitives  and it's  textures  handle   for  the  final render,  Well,  this technique is not new and it's already
been there from directx  sdk, and I think  this  is a cool  stuff  to  implement  under XNA.  Rendering a skybox
by using only 2 primitives, a single texture  cube map and a custom vertex position. If  it  is using a cube map
HDR and mirror reflection can be implemented in a breeze.


Here are some of the benefits I can think of, by using this techniques:

* Loads only a single texture file.
     - This will minimize the I/O handling, since I only need to load a single file,
        rather than loading 6 separate texture files.

* Handles only single texture.
     -  This skybox only needs a single cube map texture material,  rather than handling
        6 separate texture materials for each side of the box.

* Using only the vertex Position element.
     - Only vertex Position element is needed in this implementation,
        additional elements for texture coordinates and vertex color are not needed.

* Drawing with only 2 primitives.
     -  Best of all no need to check for the visibility of  the primitives against camera frustum view,
        since it’s only need to draw 2 primitives.

* Vertex buffer and mapping via shader.
     - The 4 vertices  position are  fixed  base  on the  size  of the back buffer height and width. 
       Naturally, it uses a vertex buffer because the vertices position are not changing and since
       the mapping is done in the shader by a normalize skybox position multiply by the camera
       view and projection, more or less the processing is done in the GPU.

* Enables to implement HDR: High Dynamic Range.
     - By changing the Surface Format to Floating-Point 16bit RGB or higher equivalent  the
       cube map is already capable for the HDR implementation. Check the tools below .


Here are some cool tools for cube map creation:
        
     - DxTex  :       
 
             MS DirectX Texture Tool, is already included in SDK, and can be found under
             directx sdk utilities, very easy to use in making a cube map and the usage is self explanatory.

     - ATI CubeMapGen :  http://www.ati.com/developer/cubemapgen/index.html

            Another one from ATI, I prefer to use this  over  DxTex,  this tool  Is freaking cool! in
            terms of cube map creation, you can actually see the Skybox you are making and it's
            reflection to object, plus a filtering  and Mipchain  Generation tools you can output and
            save it, as cube map DDS texture file.

     - NVIDIA texture tools:   http://developer.nvidia.com/object/nv_texture_tools.html
             
              NVIDIA also offer lots of DDS texture plug-ins and tools, but yet, I haven't tried
              any of these tools.


Here's my simple implementation :
  
       I already  added  this technique to Project-Z GDK, but  for the sake of simplicity, I
       created a  generic  SkyBox  class  using  this technique and added it to MS sample 
       Chase Camera, you  may grab the project file from the link below, and I hope you
       enjoy the cube map I made

       SkyCube Project  file  Download  includes SkyBox.cs
       SkyBoxCubeMap  file  Download  copy this to content directory, I forgot to include this, Lol.  

Watch : Clicky