Feature Description
- Concept: Continuous feature-guided optimization (CFGO) is a continuous feedback optimization method that integrates multimodal information of source code and binary and covers the entire software lifecycle (compilation, linking, post-linking optimization, runtime, OS, and libraries).
- Core idea: Assuming that the program functions remain unchanged, program runtime information is collected to leverage various compilation optimization technologies to make more accurate optimization decisions, resulting in a more refined program with better performance.
- Major optimizations include:
- Code layout optimization: Techniques such as basic block reordering, function rearrangement, and hot/cold partitioning are used to optimize the binary layout of the target program, improving I-cache and I-TLB hit rates.
- Advanced compiler optimization: Compilation optimization techniques, such as inlining, loop unrolling, vectorization, and indirect calls, enable the compiler to make more accurate optimization decisions based on feedback information.
Parent topic: CFGO