Scale Up with Lock-Free Algorithms - Java Performance Optimization

Discover how to implement scalable lock-free algorithms in Java without advanced academic expertise in this 45-minute talk by Roman Elizarov, Manager at JetBrains. Explore the advantages of the Java Memory Model and modern garbage collection for building vertically scalable systems that efficiently handle large datasets across multiple threads. Learn about atomic updates, lock-free data structures, and performance optimization techniques. Examine practical challenges, compare lock-free and synchronized approaches, and gain insights into benchmarking and scaling strategies for back-end systems utilizing multiple CPU cores and substantial memory resources.

Intro
What is shareit
Why shareit
Stack
Does it work
Does it scale
Synchronization
Dead Locks
Atomic Update
Atomic Reference
Lock for Algorithms
Performance Problem
Two Solutions
Atomic Reference Updater
War Handle
Atomic Phone
Was it worth it
Results
Problem
Simulation
Simulation Whisperer
Scaling
Books
Benchmarking
Operations Per Second
Virus Techniques
Practical Challenges
Lock Free vs Synchronized