Duration: 3 Days
Location: On Site / Instil HQ
Price: On Request
About The Trainer
Martin Thompson is one of the world’s foremost experts on concurrency and low-latency software.
Having spent many years designing and developing systems for some of the busiest exchanges in the world, Martin now works as an independent consultant and trainer.
Martin is in huge demand both as a conference speaker and as a trainer, and has been packing in the crowds at JavaOne, QCon, GOTO, and much more besides.
Backed by many years of hard-won experience, Martin has created a truly inspiring and challenging course for Java and C++ programmers interested in building low-latency, high-throughput systems.
Taking a unique hardware-up approach, Martin will teach you how to build code that reaches world-class performance levels.
By the end of this course candidates will have acquired a new in-depth knowledge of concurrent programming that allows them to develop data structures such as queues, executors, and shared memory IPC transports that significantly out perform any products currently available.
In addition to learning the basic techniques of building high-performance data structures, a performance testing and profiling discipline will be introduced that is fundamental to working at the extremes of systems performance.
Understanding Mechanical Sympathy
- Performance considerations on modern hardware
- How modern x86-64 hardware informs concurrent and high performance programming techniques
- Memory models for software and hardware
- Why contention of any type is the enemy and how to avoid it
- How we manage contention when we absolutely must
- How the JVM provides support for concurrent programs
- Profiling and performance testing
Concurrent and High-performance Algorithm Design
- Lock-free concurrency primitives
- Signalling state change
- Ensuring order
- Concurrent data structures – Queues, Dynamic Lists, Executors, Shared memory IPC, Locks
- Preventing speculative execution during critical data exchanges
- Managing contended state
- Wait-free techniques
- API design to avoid the latency J-curve
- Efficient back-off strategies
- Discovering hidden contention and how to avoid it
Ideally you should bring a Linux-based laptop capable of running at least 4 concurrent threads. That is, a quad core or dual core with hyperthreading. You should also have at least Java 7 and your favourite IDE installed.