C++ Lua HLSL CMake C Objective-C Other
Switch branches/tags
Clone or download

README.md

Diligent Engine - A Modern Cross-Platform 3D Graphics Library

Diligent Engine is a lightweight cross-platform abstraction layer between the application and the platform-specific graphics API designed to take advantages of next-generation APIs such as Direct3D12 and Vulkan, while providing support for older platforms via Direct3D11, OpenGL and OpenGLES. Diligent Engine exposes common front-end for all supported platforms and provides interoperability with underlying native API.
Shader source code converter allows HLSL shaders to be used on all supported platforms and rendering backends. The engine is intended to be used as a graphics subsystem in a game engine or any other 3D application, and supports integration with Unity. Diligent Engine is distributed under Apache 2.0 license and is free to use.

Features

  • Cross-platform
    • Exact same client code for all supported platforms and rendering backends
      • No #if defined(_WIN32) ... #elif defined(LINUX) ... #elif defined(ANDROID) ...
      • No #if defined(D3D11) ... #elif defined(D3D12) ... #elif defined(OPENGL) ...
    • Exact same HLSL shaders run on all platforms and all backends
  • Modular design
    • Components are clearly separated logically and physically and can be used as needed
      • Only take what you need for your project (do not want to keep samples and tutorials in your codebase? Simply remove Samples submodule. Only need core functionality? Use only Core submodule)
      • No 15000 lines-of-code files
  • Clear object-based interface
    • No global states
  • Key graphics features:
    • Automatic shader resource binding designed to leverage the next-generation rendering APIs
    • Multithreaded command buffer generation
    • Descriptor, memory and resource state management
  • Modern c++ features to make code fast and reliable

Supported Plaforms and Low-Level Graphics APIs

Platform APIs
Win32 (Windows desktop) Direct3D11, Direct3D12, OpenGL4.2+, Vulkan
Universal Windows Platform Direct3D11, Direct3D12
Android OpenGLES3.0+
Linux OpenGL4.2+
MacOS OpenGL4.1 (No compute shaders)
iOS OpenGLES3.0 (vertex and fragment shaders only)

Build Status

Platform Status
Win32/Universal Windows Build Status
Linux/MacOS/iOS Build Status

Last Stable Release - v2.3

Clonning the Repository

This is the master repository that contains three submodules. To get the repository and all submodules, use the following command:

git clone --recursive https://github.com/DiligentGraphics/DiligentEngine.git

Alternatively, you can get master repository fisrt, and then individually clone all submodules into the engine's root folder.

To checkout the last stable release, run the following commands:

  • git checkout tags/v2.3

  • git submodule update --init --recursive

Repository Structure

Master repository includes the following submodules:

Build and Run Instructions

Diligent Engine uses CMake as a cross-platform build tool. To start using cmake, download the latest release (3.10 or later is required).

Win32

To generate build files for Windows desktop platform, use either CMake GUI or command line tool. For example, to generate Visual Studio 2017 64-bit solution and project files in cmk_build/Win64 folder, navigate to the engine's root folder and run the following command:

cmake -H. -B./cmk_build/Win64 -G "Visual Studio 15 2017 Win64"

You can generate Win32 solution that targets Win8.1 SDK using the following command:

cmake -D CMAKE_SYSTEM_VERSION=8.1 -H. -B./cmk_build/Win64 -G "Visual Studio 15 2017 Win64"

WARNING! In current implementation, full path to cmake build folder must not contain white spaces. (If anybody knows a way to add quotes to CMake's custom commands, please let me know!)

To enable Vulkan validation layers, you will need to download Vulkan SDK and add environemt variable VK_LAYER_PATH that contains path to the Bin directory in VulkanSDK installation folder.

Make sure that Visual C++ ATL Support is installed via Visual Studio Installer.

Open DiligentEngine.sln file in cmk_build/Win64 folder, select configuration and build the engine. Set the desired project as startup project (by default, Asteroids demo will be selected) and run it.

By default, appplications will run in D3D11 mode. To select D3D12, OpenGL, or Vulkan use the following command line options: mode=D3D11, mode=D3D12, mode=GL, or mode=Vk (do not use spaces!). If you want to run an application outside of Visual Studio environment, the application's assets folder must be selected as a working directory. (For Visual Studio, this is automatically configured by CMake).

Universal Windows Platform

To generate build files for Universal Windows platform, you need to define the following two cmake variables:

  • CMAKE_SYSTEM_NAME=WindowsStore
  • CMAKE_SYSTEM_VERSION=< Windows SDK Version >

For example, to generate Visual Studio 2017 64-bit solution and project files in cmk_build/UWP64 folder, run the following command from the engine's root folder:

cmake -D CMAKE_SYSTEM_NAME=WindowsStore -D CMAKE_SYSTEM_VERSION=10.0 -H. -B./cmk_build/UWP64 -G "Visual Studio 15 2017 Win64"

You can target specific SDK version by refining CMAKE_SYSTEM_VERSION, for instance:

cmake -D CMAKE_SYSTEM_NAME=WindowsStore -D CMAKE_SYSTEM_VERSION=10.0.15063.0 -H. -B./cmk_build/UWP64 -G "Visual Studio 15 2017 Win64"

Set the desired project as startup project (by default, Atmosphere sample will be selected) and run it.

By default, appplications will run in D3D11 mode. You can select D3D11 or D3D12 using the following command line options: mode=D3D11, mode=D3D12 (do not use spaces!).

Note: you can generate solution that targets Windows 8.1 by defining CMAKE_SYSTEM_VERSION=8.1 cmake variable, but the solution will fail to build as it will use Visual Studio 2013 (v120) toolset that lacks proper c++11 support.

Linux

Your Linux environment needs to be set up for c++ development. If it already is, make sure your c++ tools are up to date as Diligent Engine uses modern c++ features (gcc/g++ 7 or later is recommended). To configure my fresh Ubuntu 17.10, I installed the following packages:

  1. gcc, make and other essential c/c++ tools:
  • sudo apt-get update
  • sudo apt-get upgrade
  • sudo apt-get install build-essential
  1. cmake
  • sudo apt-get install cmake
  1. Other required packages:
  • sudo apt-get install libx11-dev
  • sudo apt-get install mesa-common-dev
  • sudo apt-get install mesa-utils
  • sudo apt-get install libgl-dev

To generate make files for debug configuration, run the following CMake command from the engine's root folder:

cmake -H. -B./cmk_build/Linux64 -G "Unix Makefiles" -DCMAKE_BUILD_TYPE="Debug"

To build the engine, run the following command:

cmake --build ./cmk_build/Linux64

The engine's root folder contains Visual Studio Code settings files that configure the IDE to build the engine. You can run applications directly from the IDE. To run an application from the command line, the app's assets folder must be current directory.

Android

Please make sure that your machine is set up for Android development. Download Android Studio, Android NDK and other required tools. To verify that your environment is properly set up, try building teapots sample.

Open DiligentSamples/Android or UnityPlugin/Android folders with Android Studio to build and run the engine samples and Unity emulator on Android.

MacOS

After you clone the repo, run the following command from the engine's root folder to generate Xcode project (you need to have CMake installed on the system):

cmake -H. -B./cmk_build/MacOS -G "Xcode"

Open Xcode project file in cmk_build/MacOS folder to build the engine and run the applications.

iOS

Run the command below from the engine's root folder to generate Xcode project configured for iOS build (you need to have CMake installed on your Mac):

cmake -DCMAKE_TOOLCHAIN_FILE=DiligentCore/ios.toolchain.cmake -H. -Bcmk_build/IOS -GXcode

Open Xcode project file in cmk_build/IOS folder and build the engine. To run the applications on an iOS device, you will need to set the appropriate development team in the project settings.

Customizing Build

Diligent Engine allows clients to customize build settings by providing configuration script file that defines two optional functions:

  • custom_configure_build() - defines global build properties such as build configurations, c/c++ compile flags, link flags etc.
  • custom_configure_target() - defines custom settings for every target in the build.

The path to the configuration script should be provided through BUILD_CONFIGURATION_FILE variable when running cmake and must be relative to the cmake root folder, for example:

cmake -D BUILD_CONFIGURATION_FILE=BuildConfig.cmake -H. -B./cmk_build/Win64 -G "Visual Studio 15 2017 Win64"

Customizing global build settings with custom_configure_build() function

If defined, custom_configure_build() function is called before any build target is added. By default, cmake defines the following four configurations: Debug, Release, RelWithDebInfo, MinSizeRel. If you want, you can define your own build configurations by setting CMAKE_CONFIGURATION_TYPES variable. For instance, if you want to have only two configuration: Debug and ReleaseMT, add the following line to the custom_configure_build() function:

set(CMAKE_CONFIGURATION_TYPES Debug ReleaseMT CACHE STRING "Configuration types: Debug, ReleaseMT" FORCE)

The build system needs to know the list of debug and release (optimized) configurations, so the following two variables must also be set when CMAKE_CONFIGURATION_TYPES variable is defined:

set(DEBUG_CONFIGURATIONS DEBUG CACHE INTERNAL "" FORCE)
set(RELEASE_CONFIGURATIONS RELEASEMT CACHE INTERNAL "" FORCE)

Note that due to cmake specifics, configuration names listed in DEBUG_CONFIGURATIONS and RELEASE_CONFIGURATIONS must be capitalized.

If you define any configuration other than four standard cmake ones, you also need to set the following variables, for every new configuration:

  • CMAKE_C_FLAGS_<Config> - c compile flags
  • CMAKE_CXX_FLAGS_<Config> - c++ compile flags
  • CMAKE_EXE_LINKER_FLAGS_<Config> - executable link flags
  • CMAKE_SHARED_LINKER_FLAGS_<Config> - shared library link flags

For instance:

set(CMAKE_C_FLAGS_RELEASEMT "/MT" CACHE INTERNAL "" FORCE)
set(CMAKE_CXX_FLAGS_RELEASEMT "/MT" CACHE INTERNAL "" FORCE)
set(CMAKE_EXE_LINKER_FLAGS_RELEASEMT "/OPT:REF" CACHE INTERNAL "" FORCE)
set(CMAKE_SHARED_LINKER_FLAGS_RELEASEMT "/OPT:REF" CACHE INTERNAL "" FORCE)

Below is an example of custom_configure_build() function:

function(custom_configure_build)
    if(CMAKE_CONFIGURATION_TYPES)
        # Debug configurations
        set(DEBUG_CONFIGURATIONS DEBUG CACHE INTERNAL "" FORCE)
        # Release (optimized) configurations
        set(RELEASE_CONFIGURATIONS RELEASEMT CACHE INTERNAL "" FORCE)
        # CMAKE_CONFIGURATION_TYPES variable defines build configurations generated by cmake
        set(CMAKE_CONFIGURATION_TYPES Debug ReleaseMT CACHE STRING "Configuration types: Debug, ReleaseMT" FORCE)

        set(CMAKE_CXX_FLAGS_RELEASEMT "/MT" CACHE INTERNAL "" FORCE)
        set(CMAKE_C_FLAGS_RELEASEMT "/MT" CACHE INTERNAL "" FORCE)
        set(CMAKE_EXE_LINKER_FLAGS_RELEASEMT "/OPT:REF" CACHE INTERNAL "" FORCE)
        set(CMAKE_SHARED_LINKER_FLAGS_RELEASEMT "/OPT:REF" CACHE INTERNAL "" FORCE)
    endif()
endfunction()

Customizing individual target build settings with custom_configure_target() function

If defined, custom_configure_target() is called for every target created by the build system and allows configuring target-specific properties.

By default, the build system sets some target properties. If custom_configure_target() sets all required properties, it can tell the build system that no further processing is required by setting TARGET_CONFIGURATION_COMPLETE parent scope variable to TRUE:

set(TARGET_CONFIGURATION_COMPLETE TRUE PARENT_SCOPE)

The following is an example of custom_configure_target() function:

function(custom_configure_target TARGET)
    set_target_properties(${TARGET} PROPERTIES
        STATIC_LIBRARY_FLAGS_RELEASEMT /LTCG
    )
    set(TARGET_CONFIGURATION_COMPLETE TRUE PARENT_SCOPE)   
endfunction()

Tutorials

Tutorial 01 - Hello Triangle

This tutorial shows how to render a simple triangle using Diligent Engine API.

Tutorial 02 - Cube

This tutorial demonstrates how to render an actual 3D object, a cube. It shows how to load shaders from files, create and use vertex, index and uniform buffers.

Tutorial 03 - Texturing

This tutorial demonstrates how to apply a texture to a 3D object. It shows how to load a texture from file, create shader resource binding object and how to sample a texture in the shader.

Tutorial 04 - Instancing

This tutorial demonstrates how to use instancing to render multiple copies of one object using unique transformation matrix for every copy.

Tutorial 05 - Texture Array

This tutorial demonstrates how to combine instancing with texture arrays to use unique texture for every instance.

Tutorial 06 - Multithreading

This tutorial shows how to generate command lists in parallel from multiple threads.

Tutorial 07 - Geometry Shader

This tutorial shows how to use geometry shader to render smooth wireframe.

Tutorial 08 - Tessellation

This tutorial shows how to use hardware tessellation to implement simple adaptive terrain rendering algorithm.

Tutorial 09 - Quads

This tutorial shows how to render multiple 2D quads, frequently swithcing textures and blend modes.

Tutorial 10 - Data Streaming

This tutorial shows dynamic buffer mapping strategy using MAP_FLAG_DISCARD and MAP_FLAG_DO_NOT_SYNCHRONIZE flags to efficiently stream varying amounts of data to GPU.

Samples

Sample source code

AntTweakBar Sample

This sample demonstrates how to use AntTweakBar library to create simple user interface.

Atmosphere Sample

The sample implements physically-based atmospheric light scattering model and demonstrates how Diligent Engine can be used to accomplish various rendering tasks: loading textures from files, using complex shaders, rendering to textures, using compute shaders and unordered access views, etc.

Projects

Asteroids Performance Benchmark

This sample is designed to be a performance benchmark and is based on this demo developed by Intel. It renders 50,000 unique textured asteroids. Every asteroid is a combination of one of 1000 unique meshes and one of 10 unique textures. The sample uses original D3D11 and D3D12 native implementations, and adds implementation using Diligent Engine API to allow comparing performance of different rendering modes.

Unity Integration Demo

This project demonstrates integration of Diligent Engine with Unity

References

Diligent Engine on the Web

Diligent Engine on Twitter

Diligent Engine on Facebook

API Reference

Version History

Current Progress

  • Core

    • Added IFence interface and IDeviceContext::SignalFence() method to enable CPU-GPU synchronization
  • Samples:

    • Added Tutorial 10 - Data Streaming

v2.3

  • Core:
    • Implemented Vulkan backend
    • Implemented hardware adapter & display mode enumeration in D3D11 and D3D12 modes
    • Implemented initialization in fullscreen mode as well as toggling between fullscreen and windowed modes in run time
    • Added sync interval parameter to ISwapChain::Present()
    • API Changes
      • Added NumViewports member to GraphicsPipelineDesc struct
      • Removed PRIMITIVE_TOPOLOGY_TYPE type
      • Replaced PRIMITIVE_TOPOLOGY_TYPE GraphicsPipelineDesc::PrimitiveTopologyType with PRIMITIVE_TOPOLOGY GraphicsPipelineDesc::PrimitiveTopology
      • Removed DrawAttribs::Topology
      • Removed pStrides prarameter from IDeviceContext::SetVertexBuffers(). Strides are now defined through vertex layout.
  • API Changes:
    • Math library functions SetNearFarClipPlanes(), GetNearFarPlaneFromProjMatrix(), Projection(), OrthoOffCenter(), and Ortho() take bIsGL flag instead of bIsDirectX
    • Vertex buffer strides are now defined by the pipeline state as part of the input layout description (LayoutElement::Stride)
    • Added COMMIT_SHADER_RESOURCES_FLAG_VERIFY_STATES flag
    • Added NumViewports member to GraphicsPipelineDesc structure
  • Samples:
    • Added fullscreen mode selection dialog box
    • Implemented fullscreen mode toggle on UWP with shift + enter
    • Implemented fullscreen window toggle on Win32 with alt + enter
    • Added tutorial 09 - Quads
  • Fixed the following issues:

v2.2.a

v2.2

  • Added MacOS and iOS support

v2.1.b

  • Removed legacy Visual Studio solution and project files
  • Added API reference
  • Added tutorials 1-8

v2.1.a

  • Refactored build system to use CMake and Gradle for Android
  • Added support for Linux platform

v2.1

New Features

Core

  • Interoperability with native API
    • Accessing internal objects and handles
    • Createing diligent engine buffers/textures from native resources
    • Attaching to existing D3D11/D3D12 device or GL context
    • Resource state and command queue synchronization for D3D12
  • Integraion with Unity
  • Geometry shader support
  • Tessellation support
  • Performance optimizations

HLSL->GLSL converter

  • Support for structured buffers
  • HLSL->GLSL conversion is now a two-stage process:
    • Creating conversion stream
    • Creating GLSL source from the stream
  • Geometry shader support
  • Tessellation control and tessellation evaluation shader support
  • Support for non-void shader functions
  • Allowing structs as input parameters for shader functions

v2.0 (alpha)

Alpha release of Diligent Engine 2.0. The engine has been updated to take advantages of Direct3D12:

  • Pipeline State Object encompasses all coarse-grain state objects like Depth-Stencil State, Blend State, Rasterizer State, shader states etc.

  • New shader resource binding model implemented to leverage Direct3D12

  • OpenGL and Direct3D11 backends

  • Alpha release is only available on Windows platform

  • Direct3D11 backend is very thoroughly optimized and has very low overhead compared to native D3D11 implementation

  • Direct3D12 implementation is preliminary and not yet optimized

v1.0.0

Initial release

License

Licensed under the Apache License, Version 2.0

Copyright 2015-2018 Egor Yusov

diligentgraphics.com

Diligent Engine on Twitter Diligent Engine on Facebook