In this post, the author emphasizes the importance of small optimizations in improving performance. They categorize optimizations into three levels: architecture, algorithms, and data structures, and source code. The author argues that higher-level optimizations have a greater impact on applications than lower-level ones. They also debunk the misconception that optimizations always require complex algorithms or hard-to-read code. The author suggests techniques for finding optimization opportunities, such as profiling the application and exploring all possibilities. They provide examples of optimizations for algorithms, data structures, and source code. Additionally, the author shares a specific optimization they made in ClickHouse that resulted in significant performance improvement for string deserialization. They conclude by highlighting that performance improvement can be achieved through both high-level and small optimizations.
https://maksimkita.com/blog/power-of-small-optimizations.html