25 May 2015

feedAndroid Developers Blog

Game Performance: Geometry Instancing

Posted by Shanee Nishry, Games Developer Advocate

Imagine a beautiful virtual forest with countless trees, plants and vegetation, or a stadium with countless people in the crowd cheering. If you are heroic you might like the idea of an epic battle between armies.

Rendering a lot of meshes is desired to create a beautiful scene like a forest, a cheering crowd or an army, but doing so is quite costly and reduces the frame rate. Fortunately this is possible using a simple technique called Geometry Instancing.

Geometry instancing can be used in 2D games for rendering a large number of sprites, or in 3D for things like particles, characters and environment.

The NDK code sample More Teapots demoing the content of this article can be found with the ndk inside the samples folder and in the git repository.

Support and Extensions

Geometry instancing is available from OpenGL ES 3.0 and to OpenGL 2.0 devices which support the GL_NV_draw_instanced or GL_EXT_draw_instanced extensions. More information on how to using the extensions is shown in the More Teapots demo.

Overview

Submitting draw calls causes OpenGL to queue commands to be sent to the GPU, this has an expensive overhead which may affect performance. This overhead grows when changing states such as alpha blending function, active shader, textures and buffers.

Geometry Instancing is a technique that combines multiple draws of the same mesh into a single draw call, resulting in reduced overhead and potentially increased performance. This works even when different transformations are required.

The algorithm

To explain how Geometry Instancing works let's quickly overview traditional drawing.

Traditional Drawing

To a draw a mesh you'd usually prepare a vertex buffer and an index buffer, bind your shader and buffers, set your uniforms such as a World View Projection matrix and make a draw call.

To draw multiple instances using the same mesh you set new uniform values for the transformations and other data and call draw again. This is repeated for every instance.

Drawing with Geometry Instancing

Geometry Instancing reduces CPU overhead by reducing the sequence described above into a single buffer and draw call.

It works by using an additional buffer which contains custom per-instance data needed by your shader, such as transformations, color, light data.

The first change to your workflow is to create the additional buffer on initialization stage.

To put it into code let's define an example per-instance data that includes a world view projection matrix and a color:

C++

struct PerInstanceData
{
 Mat4x4 WorldViewProj;
 Vector4 Color;
};

You also need to the structure to your shader. The easiest way is by creating a Uniform Block with an array:

GLSL

#define MAX_INSTANCES 512

layout(std140) uniform PerInstanceData {
    struct
    {
        mat4      uMVP;
        vec4      uColor;
    } Data[ MAX_INSTANCES ];
};

Note that uniform blocks have limited sizes. You can find the maximum number of bytes you can use by querying for GL_MAX_UNIFORM_BLOCK_SIZE using glGetIntegerv.

Example:

GLint max_block_size = 0;
glGetIntegerv( GL_MAX_UNIFORM_BLOCK_SIZE, &max_block_size );

Bind the uniform block on the CPU in your program's initialization stage:

C++

#define MAX_INSTANCES 512
#define BINDING_POINT 1
GLuint shaderProgram; // Compiled shader program

// Bind Uniform Block
GLuint blockIndex = glGetUniformBlockIndex( shaderProgram, "PerInstanceData" );
glUniformBlockBinding( shaderProgram, blockIndex, BINDING_POINT );

And create a corresponding uniform buffer object:

C++

// Create Instance Buffer
GLuint instanceBuffer;

glGenBuffers( 1, &instanceBuffer );
glBindBuffer( GL_UNIFORM_BUFFER, instanceBuffer );
glBindBufferBase( GL_UNIFORM_BUFFER, BINDING_POINT, instanceBuffer );

// Initialize buffer size
glBufferData( GL_UNIFORM_BUFFER, MAX_INSTANCES * sizeof( PerInstanceData ), NULL, GL_DYNAMIC_DRAW );

The next step is to update the instance data every frame to reflect changes to the visible objects you are going to draw. Once you have your new instance buffer you can draw everything with a single call to glDrawElementsInstanced.

You update the instance buffer using glMapBufferRange. This function locks the buffer and retrieves a pointer to the byte data allowing you to copy your per-instance data. Unlock your buffer using glUnmapBuffer when you are done.

Here is a simple example for updating the instance data:

const int NUM_SCENE_OBJECTS = …; // number of objects visible in your scene which share the same mesh

// Bind the buffer
glBindBuffer( GL_UNIFORM_BUFFER, instanceBuffer );

// Retrieve pointer to map the data
PerInstanceData* pBuffer = (PerInstanceData*) glMapBufferRange( GL_UNIFORM_BUFFER, 0,
                NUM_SCENE_OBJECTS * sizeof( PerInstanceData ),
                GL_MAP_WRITE_BIT | GL_MAP_INVALIDATE_RANGE_BIT );

// Iterate the scene objects
for ( int i = 0; i < NUM_SCENE_OBJECTS; ++i )
{
    pBuffer[ i ].WorldViewProj = ... // Copy World View Projection matrix
    pBuffer[ i ].Color = …               // Copy color
}

glUnmapBuffer( GL_UNIFORM_BUFFER ); // Unmap the buffer

And finally you can draw everything with a single call to glDrawElementsInstanced or glDrawArraysInstanced (depending if you are using an index buffer):

glDrawElementsInstanced( GL_TRIANGLES, NUM_INDICES, GL_UNSIGNED_SHORT, 0,
                NUM_SCENE_OBJECTS );

You are almost done! There is just one more step to do. In your shader you need to make use of the new uniform buffer object for your transformations and colors. In your shader main program:

void main()
{
    …
    gl_Position = PerInstanceData.Data[ gl_InstanceID ].uMVP * inPosition;
    outColor = PerInstanceData.Data[ gl_InstanceID ].uColor;
}

You might have noticed the use gl_InstanceID. This is a predefined OpenGL vertex shader variable that tells your program which instance it is currently drawing. Using this variable your shader can properly iterate the instance data and match the correct transformation and color for every vertex.

That's it! You are now ready to use Geometry Instancing. If you are drawing the same mesh multiple times in a frame make sure to implement Geometry Instancing in your pipeline! This can greatly reduce overhead and improve performance.

Join the discussion on

+Android Developers

25 May 2015 10:00pm GMT

21 May 2015

feedAndroid Developers Blog

Always-on and Wi-Fi with the latest Android Wear update

Posted by Wayne Piekarski, Developer Advocate

A new update to Android Wear is rolling out with lots of new features like always-on apps, Wi-Fi connectivity, media browsing, emoji input, and more. Let's discuss some of the great new capabilities that are available in this release.

Always-on apps

Above all, a watch should make it easy to tell the time. That's why most Android Wear watches have always-on displays, so you can see the time without having to shake your wrist or lift your arm to wake up the display. In this release, we're making it possible for apps to be always-on as well.

With always-on functionality, your app can display dynamic data on the device, even when the app is in ambient mode. This is useful if your app displays information that is continuously updated. For example, running apps like Endomondo, MapMyRun, and Runtastic use the always-on screen to let you keep track of how long and far you've been running. Zillow keeps you posted about the median price of homes nearby when you're house-hunting.

Always-on functionality is also useful for apps that may not update data very frequently, but present information that's useful for reference over a longer period of time. For example, Bring! lets you keep your shopping list right on your wrist, and Golfshot gives you accurate distances from tee to pin. If you're at the airport and making your way to your gate, American Airlines, Delta, and KLM let you keep all of your flight info a glance away on your watch.

Note: the above apps will not display always-on functionality on your watch until you receive the update for the latest version of Android Wear.

Always-on functionality works similar to watch faces, in that the power usage of the display and processor is kept to a minimum by reducing the colors and refresh rate of the display. To implement an always-on Activity, you need to make a few small changes to your app's AndroidManifest.xml, your app's build.gradle, and the Activity to declare that it supports ambient mode. A code sample and documentation are available to show you how it works. Be sure to tune in to the livestream at Google I/O next week for Android Wear: Your app and the always-on screen.

Wi-Fi connectivity and cloud sync

Many existing Android Wear devices already contain hardware support for Wi-Fi, and this release enables software support for Wi-Fi. The saved Wi-Fi networks on your phone are copied to your watch during setup, and your watch automatically connects to those Wi-Fi networks when it loses Bluetooth connection to your phone. Your watch can then connect to your phone over the Internet, even if they're not on the same Wi-Fi network.

You should continue to use the Data Layer API for all communications between the watch and phone. By using this standard API, your app will always work, no matter what kind of connectivity the user's wearable supports. Cloud sync also introduces a new virtual node in the Data Layer called the cloud node, which may be returned in calls to getConnectedNodes(). Learn more in the Multi-wearable support section below.

Multi-wearable support

The release of Google Play services 7.3 now allows support for multiple wearable devices to be paired simultaneously to a single phone or tablet, so you can have a wearable for fitness, and another for dressing up. While DataItems will continue to work in the same way, since they are synchronized to all devices, working with the MessageApi is a little different. When you update your build.gradle to use version 7.3 or higher, getConnectedNodes() from the NodeApi will usually return multiple nodes. There is an extra virtual node added to represent the cloud node used to communicate over Wi-Fi, so all developers need to deal with this situation in their code.

To help simplify finding the right node among many devices, we have added a CapabilityApi, allowing your nodes to announce features they provide, for example downloading images or music. You can also now use the ChannelApi to open up a connection to a specific device to transfer large resources such as images or audio streams, without having to send them to all devices like you would when embedding assets into data items. We have updated our Android Wear samples and documentation to show the best practices in implementing this.

MediaBrowser support

The Android 5.0 release added the ability for apps to browse the media content of another app, via the android.media.browse API. With the latest Android Wear update, if your media playback app supports this API, then you will be able to browse to find the next song directly from your watch. This is the same browse capability used in Android Auto. You implement the API once, and it will work across a variety of platforms. To do so, you just need to allow Android Wear to browse your app in the onGetRoot() method validator. You can also add custom actions to the MediaSession that will appear as controls on the watch. We have a Universal Media Player sample that shows you how to implement this functionality.

Updates to existing devices

The latest version of Android Wear will roll out via an over-the-air (OTA) update to all Android Wear watches over the coming weeks. To take advantage of these new features, you will need to use targetSdkVersion 22 and add the necessary dependencies for always-on support. We have also expanded the collection of emulators available via the SDK Manager, to simulate the experience on all the currently available devices, resolutions, and shapes, including insets like the Moto 360.

In this update, we have also disabled support for apps that use the unofficial, activity-based approach for displaying watch faces, as announced in December. These watch faces will no longer work and should be updated to use the new watch face API.

Since the launch of Android Wear last summer, Android Wear has grown into a platform that gives users many possibilities to personalize their watches, with a variety of shapes and styles, a range of watch bands, and thousands of apps and watch faces. Features such as always-on apps and Wi-Fi allow developers even more flexibility to give users amazing experiences with Android Wear.

21 May 2015 5:19pm GMT

20 May 2015

feedAndroid Developers Blog

Android Developer Story: Wooga’s fast iterations on Android and Google Play

Posted by Leticia Lago, Google Play team

In order to make the best possible games, Wooga works on roughly 40 concepts and prototypes per year, out of which 10 go into production, around seven soft launch, and only two make it to global launch. It's what they call "the hit filter." For their latest title, Agent Alice, they follow up with new episodes every week to maintain player interest and engagement over time.

The ability to quickly iterate both live and under development games is therefore key to Wooga's business model - Android and Google Play provide them the tools they need and mean that new features and updates are made on Android first, before they get to other platforms.

Find out more from Sebastian Kriese, Head of Partnerships, and Pal Tamas Feher, Head of Engineering, and learn how the iteration features of Android and Google Play have contributed to successes such as Diamond Dash, Jelly Splash, and Agent Alice.

You can find out more about building successful games businesses on Android and Google Play at Google I/O 2015: in person, on the live stream, or session recordings after the event. Check out the following:

20 May 2015 6:02pm GMT