Spike-Engine

We are proud to announce that Spike-Engine beta is now out! Please visit our website to download your free copy: http://www.spike-engine.com/[^]

What is Spike-Engine?

Back in 2009, we've noticed that the modern software and web was becoming more and more interactive. Being independent video game developers and a passion for online, massively multiplayer games pushed us to create a technology that would suit all our needs. We wanted to build reliable web services without sacrificing any productivity or performance. Moreover, we wanted a single server to handle thousands of concurrent clients at the same time. This ambition led us to create Spike-Engine. Spike-Engine is a software platform that facilitates building client-server applications for .NET developers. In particular, it focuses on real-time and duplex communication, performance and productivity. The core idea behind Spike-Engine is quite simple. There are 3 core principles underpinning the architecture, duplex communication, performance and code generation. Duplex Communication. We needed duplex communication, and a connected, stateful mode. Duplex communication simply means that client and the server can both initiate the communication and exchange messages easily, at any time (unlike HTTP). After all, we wanted to build games and interactive applications! Performance. We wanted our technology to be performant and handle thousands of clients, and hundreds of thousands of packets per second. This proved to be very challenging to achieve and took 3 years to polish and to find a balance between flexibility, API and performance. Even better, good performance also saves cost, since we did not need to buy large servers with massive amount of bandwith and we could cut our costs. Code Generation. As most of you, we are a small and agile team. This means we needed to do things smart and be able to stay productive while building high quality, heteregeneous software. From the very beginning, we designed our technology to help us by generating automatically most of the networking code, while we could concentrate on actually implementing the business logic.

-- Just think in 10 dimensions!

Thank you for your reply! Ok so I figured out how to do the mirroring within same Window, something like this:

<Canvas x:Name="RendererCanvas" Background="White" Grid.Column="1" Grid.Row="1" >
<v:RenderView x:Name="Renderer" Width="300" Height="300" />
</Canvas>

    <Grid Grid.Column="0" Grid.Row="1" Background="DarkGray">
        <Viewbox Stretch="Uniform" DataContext="{Binding ElementName=Renderer}">
            <Grid Width="{Binding Width}" Height="{Binding Height}">
                <Grid.Background>
                    <VisualBrush Stretch="Uniform" Visual="{Binding}" />
                </Grid.Background>
            </Grid>
        </Viewbox>
    </Grid>

Now, would it be possible to do exact same thing but in 2 different windows? So 2 XAML files one binding another. Maybe with some static fields?

-- Just think in 10 dimensions!

Hello everyone, For small research project of my AI programme, I'm building a visualization for a multi-screen application, and 2 zones of which must be mirrored. So basically: - On one side I have a WPF layout with controls, data, images and animations happeling - This content should be replicated on another screen (simply another Window) Do you have any ideas how to achieve this? Something that could grab a frame and replicate it, like a texture brush maybe? Thank you in advance for your answers! Best regards, Kel

-- Everything is possible, even the impossible! ^_^

Thanks again. So basically I started the C# Wrapper to be able to compare it, after a few days and going to battle vs COM objets and performant InterOp, I finally managed to wrap a part of it. For now I'm doing just the D3DDevice creation and a simple window clear. I compared the exactly same function calls with Managed DirectX 1.1 ones (actually everything is the same, I compared every method parameter and call with PIX debugger). So creating D3D factory, creating a device and showing. (Just a first draft)

DirectX.Device.Clear(0, null, ClearFlags.D3DCLEAR_TARGET | ClearFlags.D3DCLEAR_ZBUFFER, 0, 1, 0);
DirectX.Device.BeginScene();
// todo
DirectX.Device.EndScene();
DirectX.Device.Present(null, null, IntPtr.Zero, null);

And guess what, the performance are almost exactly the same between MDX and my C# Wrapper. I'm getting around 2180 FPS with both of them. I still would like to check IL code generated for both.

-- Everything is possible, even the impossible! ^_^

Thank you for the answers. Yeah, so basically I think I'm just going to wrap the whole DirectX in C# and then test it against the managed C++ implementation. Think it will worth a good codeproject article :)

-- Everything is possible, even the impossible! ^_^

Hello, I'm interested in writing a managed Direct3D 9/10 wrapper on my own for learning purpose. I pretty much aware about existing MDX (which is deprecate) and XNA (which is multiplatform, but less performant). While searching for something already existing, I found a MDX 9/10/10.1 wrapper written by Ralf Kornmann http://www.codeplex.com/MD3D10[^] This wrapper uses Managed C++ to wrap the D3D Api, but a lot of other OpenGL wrappers, uses a standard C# InterOp syntax, decorating and Marshalling. I have concern about performance issues so my main question is: what is actually faster, wrap Direct3D with Managed C++ or in C# using standard C# InterOp ? Is there some performance issues? What do you think? What I mean by C# Style wrapper:

public static class DirectX
{
/* IDirect3D9* Direct3DCreate9(
UINT SDKVersion
);*/
[DllImport("d3d9.dll", EntryPoint = "Direct3DCreate9", CallingConvention = CallingConvention.Winapi), SuppressUnmanagedCodeSecurity]
[return: MarshalAs(UnmanagedType.Interface)]
public static extern IDirect3D9 Direct3DCreate9(UInt32 SDKVersion);

}

/* DECLARE_INTERFACE_(IDirect3D9, IUnknown)*/
[ComVisible(true), ComImport]
[Guid("81BDCBCA-64D4-426d-AE8D-AD0147F4275C"), InterfaceType(ComInterfaceType.InterfaceIsIUnknown)]
public interface IDirect3D9 // http://msdn.microsoft.com/en-us/library/bb174300(VS.85).aspx
{
/* UINT GetAdapterCount(
);*/
[PreserveSig]
UInt32 GetAdapterCount();

    /\*  HRESULT GetAdapterDisplayMode(
           UINT Adapter,
           D3DDISPLAYMODE\* pMode
        );\*/
    \[PreserveSig\]
    int GetAdapterDisplayMode(UInt32 Adapter, \[In, Out\] D3DDISPLAYMODE pMode);

    ...
}

-- Everything is possible, even the impossible! ^_^

Yeah actually single(float) and double are working the same way, it's not precise, because it's floating-point numbers representation. So it's a lot better and precise to use fixed-point number representation = Decimal in .Net :)

-- Everything is possible, even the impossible! ^_^

Hey, Not sure, because I haven't deployed 3.5 stuff yet, but I think there's one with Windows SDK 6 (the one installed automatically with VS 2008). Should be somethere here if you have VS 2008 installed: C:\Program Files\Microsoft SDKs\Windows\v6.0A\Bootstrapper\Packages\DotNetFX35

-- Everything is possible, even the impossible! ^_^

Thank you! This is helpful :) So I should look for creating a driver, I'm interested in some examples... For the work and knowledge, I have my time and I like to learn. Just for the context: my application is XNA flight simulator for a real world helicopter.

-- Everything is possible, even the impossible! ^_^

Yeah, but it should work in windows. I'm sure there's something like nanosleep in win32 kernel mode.

-- Everything is possible, even the impossible! ^_^

Yeah, it's a good solution for measuring time, but for waiting (sleeping) I wasn't able to find a way to use it. Thread.Sleep in combination with winmm.dll calls give 1ms precision only :/

-- Everything is possible, even the impossible! ^_^

Already read and tested, with this I can only get 1 millisecond resolution (Multimedia Timer), but I need 1 microsecond...

-- Everything is possible, even the impossible! ^_^

Hello, In my application, I have to get a microsecond resolution for a timer. Is it possible to do (on WinXP or Vista) ? There should be high precision timers in kernel mode, but what is it and how could I get to them from .Net code? All help will be appreciated. Roman

-- Everything is possible, even the impossible! ^_^

I'm writing a differed tasks framework for an application, intercepting computer shutdown and launching a form to execute planified tasks. ^^ Security sandbox is framework 3 only feature?

-- Everything is possible, even the impossible! ^_^

Hello, I'm trying to figure out how I could make the same "effect" like in Windows CardSpace (just blocking all screens, change their luminosity and show one WinForm to execute my procedure). Any ideas? ps: you can view this "effect" after installing framework 3 and going to Control panel -> Windows CardSpace

-- Everything is possible, even the impossible! ^_^

would be nice !

-- Everything is possible, even the impossible! ^_^

*hugs* He's really sooo bad ? Can he use a notepad ? :p

-- Everything is possible, even the impossible! ^_^