SSA-based Register Allocation
AnySL: Efficient and Portable Multi-Language Shading
In cooperation with the Computer Graphics Group we develop a unified shading system that is independent of source language, target architecture and rendering engine without sacrificing runtime performance.
Whole-Function Vectorization enables data-parallel languages to use SIMD instruction sets to exploit more data-level parallelism in addition to multi-threading.
Sierra is a SIMD extension for C++.
Simple and Efficient Construction of Static Single Assignment Form
Noise: User-Defined Optimization Strategies
Specify custom optimization strategies for specific parts of your code. Implemented in Clang/LLVM: source-language and target-platform independent. Example realization in C/C++ via attributes.