Java Performance Techniques - The Cost of HotSpot Runtime Optimizations

Explore advanced Java performance techniques and the impact of HotSpot runtime optimizations in this 54-minute conference talk. Delve into how HotSpot internally handles performance optimizations, including static, speculative, and inter-procedural techniques that can make Java code run faster than comparable C/C++ code. Examine the penalties of code deoptimizations, such as uncommon traps, which return the running thread to the Interpreter. Gain a deeper understanding of crucial optimizations and potential pitfalls that may unexpectedly hinder performance. Cover key topics including uncommon traps, null sanity checks, loop unrolling, string deduplication, devirtualization, inter-procedural optimizations, branch predictions, concurrency implications, and methods that do not JIT. Learn about tools like JITWatch, JMH (Java Measurement Harness), and HPjmeter used to reveal these optimizations. Benefit from the expertise of Ionut Balosin, a Software Architect with over 10 years of experience in various business applications and a regular speaker at technology conferences.