In this talk, we will give examples of code that contains performance-reducing patterns, which cannot be identified [yet] with the help of modern static analysis utilities, and which are not always evident when executing the code review.
All examples are based on the code of applications that work in production. The focus is on improving performance by examining and changing the progress of code execution without the use of complex concurrency techniques or any secret data structures. Most part of examples is not connected to any framework. Moreover, they can be used with other programming languages (which, in Sergey’s opinion, is the most valuable part of this talk). The conclusions drawn are backed up by benchmarks.