This article explains techniques for shrinking the size of a Rust static library intended for use with Go. It details the process of removing unused code, optimizing linked sections, and converting LLVM bitcode to achieve a significantly smaller library file. The author shares practical steps and results of the optimization efforts.

The article details a case study where a payment service at TikTok, originally built in Go, faced CPU bottlenecks due to increased traffic. By selectively rewriting the most CPU-intensive endpoints in Rust, the team achieved double the performance and projected annual savings of $300,000 in cloud costs.

hyperpb is a new high-performance Protobuf library for Go, designed to leverage optimizations from UPB while addressing the challenges of Go's C FFI. It features a dynamic, runtime-based parser that outperforms existing Go Protobuf parsers in benchmarks. The library aims to provide an efficient and flexible solution for handling Protobuf messages in Go applications.

hyperpb is a new high-performance Protobuf library for Go that implements many optimizations from UPB while addressing the unique challenges of Go's runtime. It leverages a dynamic table-driven parsing approach to improve performance and reduce instruction cache issues associated with traditional Protobuf parsers. The library's API allows for efficient message handling and compilation, making it faster than existing Go Protobuf solutions.

The article explores the inefficiencies in Go related to handling io.Reader interfaces, particularly when decoding images with libraries like libavif and libheif. It discusses the challenges of type inspection and the need for optimization to avoid unnecessary data copying, ultimately leading to a workaround that allows for efficient byte extraction. The author critiques the design of Go's standard library concerning structural typing and the hidden requirements for certain functionalities.

Understanding software performance often requires profiling to determine where code execution time is spent. Go offers built-in profiling tools and the article explains how to use flame graphs to visualize profiling data, helping developers identify performance bottlenecks effectively.