Predictable Latency
summary
Covers algorithms for high-throughput, low-and-predictable-latency concurrent systems without locks, focusing on relaxed memory ordering and non-blocking telemetry. Compares multiple concurrency approaches and measures their impact on latency, and discusses how Java 8 features affect some implementations.
AI-written summary · about
chapters (18)
- 0:00Intro
- 1:55Blocking
- 3:25Concurrent Algorithms Locks
- 4:50Queuing Theory
- 6:58Amdahl's & Universal Scalability Laws
- 8:33Condition Variable RIT
- 10:14Context for the Benchmarks
- 10:25Burst Length = 1: RTT (ns)
- 14:50Disruptor 1.0 - 2.0
- 16:46Disruptor 3.0
- 18:37ManyToOneConcurrentArray Queue
- 19:59Burst Length = 100: RTT (ns)
- 21:59MPSC Ring Buffer Message Header
- 25:46Aeron Data Message Header
- 27:35Burst Length = 100: 99% RTT (ns)
- 28:51C vs Java
- 34:052TB RAM and 100+ vcores!!!
- 34:34Where can I find the code?
related talks
45mThe Hidden Art of Thread-Safe Programming: Exploring java.util.concurrent
50mThe Hidden Art of Thread-Safe Programming: Exploring java.util.con...
41mEverything You Don’t Want to Know About Concurrency on the JVM
01h 03mIntro to Concurrent Programming 3/3: Relaxed Data Structures for Parallel Algorithms
39mLet's talk concurrency
48mConcurrency: It’s Harder (and Easier) Than You Think
48m“Is Parallel Programming still hard? PART 2 of 2”
49mUnconventional memory model extensions in practice