npcdoom / vkspecgen Goto Github PK

VkSpecGen is a toolkit for generating code from the Vulkan specification XML. While VkSpecGen can be used by it's own, to generate flatbuffers schema for example, it also supports Jinja2 template rendering engine. Support for other or custom template rendering engines can be added and to VkSpecGen conveniently.

Please note that VkSpecGen is not an officially supported Google product.

1. Install Jinja2

   pip install Jinja2

2. Download Vulkan API registry

3. Generate the C++ code with

   python3 generator_script.py

Multiple examples are provided in examples directory. The code will be generated in the root folder. You need python3 to run the examples. You can pass the path to the Vulkan API registry file (vk.xml) using -s or --spec. If the path is not provided, the generator scripts expect to find the registry file in the root directory of the VkSpecGen. You can pass the target platform for the layer using -p or --platform. The default platform is the core platform. For the list of valid Vulkan platforms please see <platforms> node in vk.xml file.

examples/flatbuffer/vkflat.py generates flatbuffers schema for Vulkan types in C++. The schema is printed to stdout. Note that this script is not accompanied with a Jinja2 template and uses VkSpecGen to directly generate the output code.

examples/print_layer/vklayerpint.py generates the code for a Vulkan layer that intercepts all the Vulkan calls and prints the parameters to stdout. The generator script uses multiple Jinja2 templates in templates directory to generate the layer dispatch tables and the interceptors.

You can use CMake to build the layer from the example folder. However note that for the build to be successful, you need to have the proper version of the Vulkan headers installed. The headers should match the version of the vk.xml that you used to generate the layer.

The tests currently require a real vk.xml, which makes them version-dependent. The supported version is v1.2.202.

Run tests with:

./run_tests.sh <path to vk.xml>

This project uses yapf Python formatter. The settings for VSCode is as follows:

"python.formatting.provider": "yapf",
"python.formatting.yapfArgs": [
  "--style",
  "{based_on_style: google, indent_width: 2}"
],