OpenSubdiv/README.md

219 lines
9.0 KiB
Markdown
Raw Normal View History

# OpenSubdiv
2012-06-08 17:05:35 +00:00
OpenSubdiv is a set of open source libraries that implement high performance subdivision surface (subdiv) evaluation on massively parallel CPU and GPU architectures. This codepath is optimized for drawing deforming subdivs with static topology at interactive framerates. The resulting limit surface matches Pixar's Renderman to numerical precision.
2012-06-08 17:05:35 +00:00
OpenSubdiv is covered by the [Microsoft Public License](http://www.microsoft.com/en-us/openness/licenses.aspx#MPL), and is free to use for commercial or non-commercial use. This is the same code that Pixar uses internally for animated film production. Our intent is to encourage high performance accurate subdiv drawing by giving away the "good stuff".
2012-06-08 17:05:35 +00:00
The current version is "Release Candidate 1.0" (12/05/2012) and we hope to have "Release 1.0" ready by February 2013. Feel free to use it and let us know what you think.
2012-06-08 17:05:35 +00:00
For more details about OpenSubdiv, see [Pixar Graphics Technologies](http://graphics.pixar.com).
2012-06-08 17:05:35 +00:00
## What's New in Release Candidate 1.0 ?
- [Feature Adaptive GPU Rendering of Catmull-Clark Surfaces](http://research.microsoft.com/en-us/um/people/cloop/tog2012.pdf).
- New API architecture : we are planning to lock on to this new framework as the basis for backward compatibility, which we will enforce from Release 1.0 onward. Subsequent releases of OpenSubdiv should not break client code.
- DirectX 11 support
2012-06-08 17:05:35 +00:00
## Quickstart
2012-06-08 17:05:35 +00:00
2012-06-08 18:30:50 +00:00
Basic instructions to get started with the code.
### Dependencies
2012-06-08 18:30:50 +00:00
Cmake will adapt the build based on which dependencies have been successfully discovered and will disable certain features and code examples accordingly.
2012-08-15 19:45:10 +00:00
Please refer to the documentation of each of the dependency packages for specific build and installation instructions.
Required:
2012-08-15 19:45:10 +00:00
* [cmake](http://www.cmake.org/cmake/resources/software.html)
* [GLEW](http://sourceforge.net/projects/glew/) (Windows/Linux only)
2012-06-08 18:30:50 +00:00
Optional:
2012-08-15 19:45:10 +00:00
* [CUDA](http://developer.nvidia.com/category/zone/cuda-zone)
* [OpenCL](http://www.khronos.org/opencl/)
2013-02-01 11:22:12 +00:00
* [GLFW](http://www.glfw.org/)
2012-08-15 19:45:10 +00:00
* [Ptex](https://github.com/wdas/ptex)
* [Zlib](http://www.zlib.net) (required for Ptex under Windows)
* [Maya SDK](http://www.autodesk.com/maya/) (sample code for Maya viewport 2.0 primitive)
* [DX11 SDK](http://www.microsoft.com/en-us/download/details.aspx?id=6812)
2012-08-15 19:45:10 +00:00
### Useful cmake options and environment variables
````
-DCMAKE_BUILD_TYPE=[Debug|Release]
-DCUDA_TOOLKIT_ROOT_DIR=[path to CUDA]
-DPTEX_LOCATION=[path to Ptex]
-DGLEW_LOCATION=[path to GLEW]
2013-02-01 11:22:12 +00:00
-DGLFW_LOCATION=[path to GLFW]
2012-08-15 19:45:10 +00:00
-DMAYA_LOCATION=[path to Maya]
-DNO_OMP=1 // disable OpenMP
-DNO_GCD=1 // disable GrandCentralDispatch on OSX
2012-08-15 19:45:10 +00:00
````
2013-02-01 11:22:12 +00:00
The paths to Maya, Ptex, GLFW, and GLEW can also be specified through the
following environment variables: `MAYA_LOCATION`, `PTEX_LOCATION`, `GLFW_LOCATION`,
2012-08-15 19:45:10 +00:00
and `GLEW_LOCATION`.
2012-06-08 18:30:50 +00:00
### Build instructions (Linux/OSX/Windows):
2012-06-05 21:07:27 +00:00
__Clone the repository:__
2012-06-05 21:07:27 +00:00
From the GitShell, Cygwin or the CLI :
````
git clone git://github.com/PixarAnimationStudios/OpenSubdiv.git
````
2012-06-08 17:05:35 +00:00
Alternatively, on Windows, GIT also provides a GUI to perform this operation.
__Generate Makefiles:__
2012-06-08 18:30:50 +00:00
Assuming that we want the binaries installed into a "build" directory at the root of the OpenSubdiv tree :
````
cd OpenSubdiv
mkdir build
cd build
````
2012-06-08 18:30:50 +00:00
Here is an example cmake configuration script for a full typical windows-based build that can be run in GitShell :
2012-06-08 18:30:50 +00:00
````
#/bin/tcsh
# Replace the ".." with a full path to the root of the OpenSubdiv source tree if necessary
"c:/Program Files (x86)/CMake 2.8/bin/cmake.exe" \
-G "Visual Studio 10 Win64" \
-D "GLEW_LOCATION:string=c:/Program Files/glew-1.9.0" \
2013-02-01 11:22:12 +00:00
-D "GLFW_LOCATION:string=c:/Program Files/glfw-2.7.7.bin.WIN64" \
-D "OPENCL_INCLUDE_DIRS:string=c:/ProgramData/NVIDIA Corporation/NVIDIA GPU Computing SDK 4.2/OpenCL/common/inc" \
-D "_OPENCL_CPP_INCLUDE_DIRS:string=c:/ProgramData/NVIDIA Corporation/NVIDIA GPU Computing SDK 4.2/OpenCL/common/inc" \
-D "OPENCL_LIBRARIES:string=c:/ProgramData/NVIDIA Corporation/NVIDIA GPU Computing SDK 4.2/OpenCL/common/lib/x64/OpenCL.lib" \
-D "MAYA_LOCATION:string=c:/Program Files/Autodesk/Maya2013.5" \
-D "PTEX_LOCATION:string=c:/Users/opensubdiv/demo/src/ptex/x64" \
..
# copy Ptex dependencies (Windows only)
mkdir -p bin/{Debug,Release}
\cp -f c:/Users/opensubdiv/demo/src/zlib-1.2.7/contrib/vstudio/vc10/x64/ZlibDllRelease/zlibwapi.dll bin/Debug/
\cp -f c:/Users/opensubdiv/demo/src/zlib-1.2.7/contrib/vstudio/vc10/x64/ZlibDllRelease/zlibwapi.dll bin/Release/
\cp -f c:/Users/opensubdiv/demo/src/ptex/x64/lib/Ptex.dll bin/Debug/
\cp -f c:/Users/opensubdiv/demo/src/ptex/x64/lib/Ptex.dll bin/Release/
````
2012-06-08 18:30:50 +00:00
Alternatively, you can use the cmake GUI or run the commands from the CLI.
2012-08-15 19:45:10 +00:00
2013-02-04 19:49:30 +00:00
Note : the OSX builder in cmake for Xcode is -G "Xcode"
__Build the project:__
2012-08-15 19:45:10 +00:00
Windows : launch VC++ with the solution generated by cmake in your build directory.
2013-02-01 18:21:10 +00:00
OSX : run xcodebuild in your build directory
*Nix : run make in your build directory
2012-06-08 18:30:50 +00:00
## Standalone viewers
OpenSubdiv builds a number of standalone viewers that demonstrate various aspects of the software.
2012-06-08 23:15:08 +00:00
__Common Keyboard Shortcuts:__
2012-06-08 23:15:08 +00:00
````
Left mouse button drag : orbit camera
Middle mouse button drag : pan camera
Right mouse button : dolly camera
n, p : next/prev model
1, 2, 3, 4, 5, 6, 7 : specify adaptive isolation or uniform refinment level
+, - : increase / decrease tessellation
w : switch display mode
q : quit
````
## Build instructions (iOS/Android)
OpenSubdiv may also be used for mobile app development.
Support for the CPU and GPU APIs used by OpenSubdiv is more limited on today's mobile operating systems. For example, the most widely support graphics API is OpenGL ES 2.0 which doesn't yet provide the support for tessellation shaders needed to fully implement GPU accellerated Feature Adaptive Subdivision.
OpenSubdiv can still be used to compute uniform refinement of subdivision surfaces for display on these platforms, realizing all of the benefits of a consistent interpretation of subdivision schemes and tags.
The easiest way to get started using OpenSubdiv for mobile is to use CMake's support for cross-compiling:
* [CMake Cross Compiling](http://www.cmake.org/Wiki/CMake_Cross_Compiling)
### iOS
You will need a current version of Apple's Xcode and iOS SDK (tested with iOS 6.0.1 and Xcode 4.5.2):
* [Xcode](https://developer.apple.com/xcode/)
and a CMake toolchain for iOS:
* [iOS CMake](https://code.google.com/p/ios-cmake/)
You can then use CMake to configure and generate an Xcode project:
2012-12-21 19:53:08 +00:00
````
mkdir build-ios
cd build-ios
cmake -DCMAKE_TOOLCHAIN_FILE=[path to iOS.cmake] -GXcode ..
xcodebuild -target install -configuration Debug
2012-12-21 19:53:08 +00:00
````
You can open the resulting Xcode project directly, or include as a sub-project in the Xcode project for your app.
### Android NDK
You will need a current version of the Android NDK (tested with Android 4.2.1 and android-ndk-r8b):
* [Android NDK](http://developer.android.com/tools/sdk/ndk/index.html)
and a CMake toolchain for Android:
* [Android CMake](https://code.google.com/p/android-cmake/)
You can then use CMake to configure and build OpenSubdiv:
2012-12-21 19:53:08 +00:00
````
mkdir build-ndk
cd build-ndk
cmake -DCMAKE_TOOLCHAIN_FILE=[path to android.cmake] -DLIBRARY_OUTPUT_PATH_ROOT=`pwd`/modules/OpenSubdiv ..
make install
2012-12-21 19:53:08 +00:00
````
The resulting NDK module can be imported by other NDK modules by including it in your module search path:
2012-12-21 19:53:08 +00:00
````
export NDK_MODULE_PATH=[path to build-ndk/modules]
2012-12-21 19:53:08 +00:00
````
## Regression tests
OpenSubdiv builds a number of regression test executables for testing:
* hbr_regression: Regression testing matching HBR (low-level hierarchical boundary rep) to a pre-generated data set.
* far_regression: Matching FAR (feature-adaptive rep using tables) against HBR results.
* osd_regression: Matching full OSD subdivision against HBR results. Currently checks single threaded CPU kernel only.
2013-02-21 19:54:20 +00:00
## Contributing
2013-02-21 19:54:20 +00:00
In order for us to accept code submissions (merge git pull-requests), contributors need to sign the "Contributor License Agreement" (found in the code repository or [here](https://github.com/PixarAnimationStudios/OpenSubdiv/blob/master/OpenSubdivCLA.pdf)) and you can either email or fax it to Pixar.
## Wish List
There are many things we'd love to do to improve support for subdivs but don't have the resources to. In particular, we would welcome contributions for the following items :
2012-06-08 17:05:35 +00:00
* The maya plugins don't integrate with Maya shading. That would be cool.
* John Lasseter loves looking at film assets in progress on an iPad. If anyone were to get this working on iOS he'd be looking at your code, and the apple geeks in all of us would smile.
2012-06-08 17:05:35 +00:00
* Alembic support would be wonderful, but we don't use Alembic enough internally to do the work.
* The precomputation step with hbr can be slow. Does anyone have thoughts on higher performance with topology rich data structures needed for feature adaptive subdivision? Maybe a class that packs adjacency into blocks of indices efficiently, or supports multithreading ?