Summary: The author discusses their performance work with serde and serde_json, highlighting discrepancies in error path speed and proposing optimizations using SIMD and SWAR techniques. They detail the process of optimizing code for faster results, including a discussion on two-pass vs. single-pass algorithms and the intricacies of parsing Unicode characters. The author’s experience with contributing to serde_json and the decision-making process behind optimizations are shared, showcasing a unique perspective on code optimization and efficiency. Throughout the text, the author provides valuable insights and lessons learned from their optimization journey, making it an informative and engaging read.
https://purplesyringa.moe/blog/i-sped-up-serde-json-strings-by-20-percent/