Walkthrough: Capturing Graphics Information Programmatically

  • Artikel

Note

This article applies to Visual Studio 2015. If you're looking for the latest Visual Studio documentation, see Visual Studio documentation. We recommend upgrading to the latest version of Visual Studio. Download it here

You can use Visual Studio Graphics Diagnostics to programmatically capture graphics information from a Direct3D app.

Programmatic capture is useful in scenarios such as:

  • Begin capture programmatically when your graphics app doesn't use swapchain present, such as when it renders to a texture.

  • Begin capture programmatically when your app doesn't render at all, such as when it uses DirectCompute to perform calculations.

  • Call CaptureCurrentFramewhen a rendering problem is difficult to anticipate and capture in manual testing but can be predicted programmatically by using information about the state of the app at runtime.

Programmatic capture in Windows 8.1

This part of the walkthrough demonstrates programmatic capture in apps that use the DirectX 11.2 API on Windows 8.1, which uses the robust capture method. For information about how to use programmatic capture in apps that use earlier versions of DirectX on Windows 8.0, see Programmatic capture in Windows 8.0 and earlier later in this walkthrough.

This section shows how to do these tasks:

  • Preparing your app to use programmatic capture

  • Getting the IDXGraphicsAnalysis interface

  • Capturing graphics information

Note

Previous implementations of programmatic capture relied on Remote Tools for Visual Studio to provide capture functionality, Windows 8.1 supports capture directly through Direct3D 11.2. As a result, you no longer have to install the Remote Tools for programmatic capture on Windows 8.1.

Preparing your app to use programmatic capture

To use programmatic capture in your app, it must include the necessary headers. These headers are part of the Windows 8.1 SDK.

To include programmatic capture headers

  • Include these headers in the source file where you will define the IDXGraphicsAnalysis interface:

    #include <DXGItype.h>  
#include <dxgi1_2.h>  
#include <dxgi1_3.h>  
#include <DXProgrammableCapture.h>

    Important

    Do not include the header file vsgcapture.h—which supports programmatic capture on Windows 8.0 and earlier—to perform programmatic capture in your Windows 8.1 apps. This header is incompatible with DirectX 11.2. If this file is included after the d3d11_2.h header is included, the compiler issues a warning. If vsgcapture.h is included before d3d11_2.h, the app will not start.

    Note

    If the June 2010 DirectX SDK is installed on your machine and your project's include path contains %DXSDK_DIR%includex86, move it to the end of the include path. Do the same for your library path.

Windows Phone 8.1

Because the Windows Phone 8.1 SDK doesn't include the DXProgrammableCapture.h header, you'll need to define the IDXGraphicsAnalysis interface yourself so that you can use the BeginCapture() and EndCapture() methods. Include the other headers as described in the previous section.

To define the IDXGraphicsAnalysis interface

  • Define the IDXGraphicsAnalysis interface in the same file in which you included the header files.

    interface DECLSPEC_UUID("9f251514-9d4d-4902-9d60-18988ab7d4b5") DECLSPEC_NOVTABLE  
IDXGraphicsAnalysis : public IUnknown  
{  
    STDMETHOD_(void, BeginCapture)() PURE;  
    STDMETHOD_(void, EndCapture)() PURE;  
};

    For convenience, you can perform these steps in a new header file and then include it where it's needed in your app.

Getting the IDXGraphicsAnalysis interface

Before you can capture graphics information from DirectX 11.2, you have to get the DXGI debug interface.

Important

When using programmatic capture, you must still run your app under graphics diagnostics (Alt+F5 in Visual Studio) or under the Command-Line Capture Tool.

To get the IDXGraphicsAnalysis interface

  • Use the following code to hook up the IDXGraphicsAnalysis interface to the DXGI debug interface.

    IDXGraphicsAnalysis* pGraphicsAnalysis;  
HRESULT getAnalysis = DXGIGetDebugInterface1(0, __uuidof(pGraphicsAnalysis), reinterpret_cast<void**>(&pGraphicsAnalysis));

    Be sure to check the HRESULT returned by DXGIGetDebugInterface1 to ensure you get a valid interface before you use it:

    if (FAILED(getAnalysis))  
{  
    // Abort program or disable programmatic capture in your app.  
}

    Note

    If DXGIGetDebugInterface1 returns E_NOINTERFACE (error: E_NOINTERFACE No such interface supported), make sure the app is running under graphics diagnostics (Alt+F5 in Visual Studio).

Capturing graphics information

Now that you have a valid IDXGraphicsAnalysis interface, you can use BeginCapture and EndCapture to capture graphics information.

To capture graphics information

  • To start capturing graphics information, use BeginCapture:

    ...  
pGraphicsAnalysis->BeginCapture();  
...

    Capture begins immediately when BeginCapture is called; it doesn't wait for the next frame to begin. Capture stops when the current frame is presented, or when you call EndCapture:

    ...  
pGraphicsAnalysis->EndCapture();  
...

Programmatic capture in Windows 8.0 and earlier

This part of the walkthrough demonstrates programmatic capture in apps for Windows 8.0 and earlier that use the DirectX 11.1 API, which uses the legacy capture method. For information about how to use programmatic capture in apps that use DirectX 11.2 on Windows 8.1, see Programmatic capture in Windows 8.1 earlier in this walkthrough.

This part shows these tasks:

  • Preparing your computer to use programmatic capture

  • Preparing your app to use programmatic capture

  • Configuring the name and location of the graphics log file

  • Using the CaptureCurrentFrame API

Preparing your computer to use programmatic capture

The programmatic capture API uses Remote Tools for Visual Studio to provide capture functionality. The computer where the app will run must have the remote tools installed, even when you are using programmatic capture on your local computer. Visual Studio doesn't have to be running when you perform programmatic capture on a local computer.

To use the remote capture APIs in an app that's running on a computer, you first have to install the Remote Tools for Visual Studio on that computer. Different versions of the remote tools support different hardware platforms. For information about how to install the remote tools, see the Remote Tools download page on the Microsoft downloads website.

Alternatively, Visual Studio installs the necessary components to perform remote capture for 32-bit apps.

Note

Because most Windows desktop apps—including Visual Studio—are not supported on Windows 8 for ARM devices, using Remote Tools for Visual Studio together with the programmatic capture API is the only way to capture graphics diagnostics on ARM devices.

Preparing your app to use programmatic capture

To use the Graphics Diagnostics tools, you first have to capture the graphics information that it relies on. You can programmatically capture the information by using the CaptureCurrentFrame API.

To prepare your app to capture graphics information programmatically

  1. Make sure that the vsgcapture.h header is included in the source code for the app. It can be included in just one location—for example, in the source code file where you will call the programmatic capture API—or in a precompiled header file to call the API from multiple source code files.

  2. In the source code for the app, whenever you want to capture the remainder of the current frame, call g_pVsgDbg->CaptureCurrentFrame(). This method takes no parameters and doesn't return a value.

Configuring the name and location of the graphics log file

The graphics log is created in the location that's defined by the DONT_SAVE_VSGLOG_TO_TEMP and VSG_DEFAULT_RUN_FILENAME macros.

To configure the name and location of the graphics log file

  • To prevent the graphics log from being written to the temp directory, before the #include <vsgcapture.h> line, add this:

    #define DONT_SAVE_VSGLOG_TO_TEMP

    You can define this value to write the graphics log to a location that's relative to the working directory, or to an absolute path if the definition of VSG_DEFAULT_RUN_FILENAME is an absolute path.

  • To save the graphics log to a different location, or give it a different file name, before the #include <vsgcapture.h> line, add this:

    #define VSG_DEFAULT_RUN_FILENAME <filename>

    If you don't perform this step, the file name is default.vsglog. If you didn't define DONT_SAVE_VSGLOG_TO_TEMP, then the location of the file is relative to the temp directory; otherwise, it's relative to the working directory or in another location if you specified an absolute file name.

    For Windows 8.x Store apps, the location of the temp directory is specific to each user and app, and is typically found in a location such as C:\users\username\AppData\Local\Packages\package family name\TempState\. For desktop apps, the location of the temp directory is specific to each user and is typically found in a location such as C:\Users\username\AppData\Local\Temp\.

Note

To write to a specific location, you must have permissions to write to that location; otherwise, an error occurs. Keep in mind that Windows 8.x Store apps are more restricted than desktop apps about where they can write data, and might require additional configuration to write to certain locations.

Capturing the graphics information

After you have prepared the app for programmatic capture and optionally configured the location and name of the graphics log file, build the app and then run or debug it to capture data; do not start graphics diagnostics from Visual Studio when you use the programmatic capture API. The graphics log is written to the location that you specified. If you want to keep this version of the log, move it to another location; otherwise, it will be overwritten when you run the app again.

Tip

You can still capture graphics information manually while you're using programmatic capture— with the app in focus, just press Print Screen. You can use this technique to capture additional graphics information that's not captured by the programmatic capture API.

Next Steps

This walkthrough demonstrated how to capture graphics information programmatically. As a next step, consider this option:

  • Learn how to analyze captured graphics information by using the Graphics Diagnostics tools. See Overview.

See Also

Walkthrough: Capturing Graphics Information
Capturing Graphics Information
Command-Line Capture Tool