Microsoft DXC is the new shader compiler stack, but the FXC compiler is still the dominant HLSL compiler for a number of reasons:
- Performance and correctness regressions of DXIL shaders compared to DXBC
- Many cross compilers and custom toolchains still rely on DXBC
- IHV drivers are still being adapted to consume DXIL, which is more low-level compared to DXBC
- DXC is a complex codebase, as it is based on LLVM - difficult to build, and many components
- DXIL is Direct3D 12 only, which makes it Windows 10 only
Therefore, it is still important to support shader compilation with FXC in some situations.
Microsoft’s DirectX shader compiler now compiles on Linux, and can generate both SPIR-V and DXIL from HLSL. However, in order to create shaders from DXIL in a running application without using development or experimental mode, the DXIL must be validated with the official
dxil.dll binary releases from Microsoft.
The actual validation logic isn’t a secret, as it is public in the GitHub repository, but the official binaries are built at a specific revision so that hardware drivers have a well-known and deterministic set of rules to rely on.
When compilation is performed, there is a sneaky
LoadLibrary call for
dxil.dll, and if found, the shader byte code result will be officially signed. If the library is not found, then the shader byte code will be unsigned. In both situations, all validation rules are still evaluated.
Microsoft’s new DirectX shader compiler (DXC) is based on LLVM and Clang, which has been traditionally a cross platform codebase, but became Windows-centric through COM, SAL, etc. used to support DirectX 12 shaders. Recently, Google listened (thank you!) to a number of requests from myself and others (here and here) to refactor the codebase to support Linux and macOS compilation.
I have been pursuing cloud based shader compilation for a while, in order to scale our very slow shader compilation pipelines to greatly improve developer iteration time. Running content pipelines on Windows-based virtual machines is not a feasible approach due to concerns of cost, maintainability, and robustness.
With Linux compilation support, it is now possible to run the DXC compiler within a Docker container, and scale it out in a Kubernetes cluster (like GKE).
Through many discussions about pitfalls and gotchas with Tobias Hector about Vulkan synchronization, it was pretty clear that there was an opportunity to simplify synchronization patterns into something more approachable and less error-prone. Tobias created a C++ implementation based on our discussions, and this library was used successfully on some of our internal projects at SEED (i.e. PICA PICA).
Available at on crates.io.
In an effort to make Vulkan synchronization more accessible, this library provides an efficient simplification of core synchronization mechanisms such as pipeline barriers and events.
Rather than the complex maze of enums and bit flags in Vulkan - many combinations of which are invalid or nonsensical - this library collapses this to a much shorter list of ~40 distinct usage types, and a couple of options for handling image layouts.
Additionally, these usage types provide an easier mapping to other graphics APIs like DirectX 12.