Software performance and complexity are related. It’s common for refactoring to introduce unanticipated regressions, and for performance optimisations to attract scrutiny in code review; how much performance improvement is worth a perceived loss of readability?
This presentation explores the relationship between complexity and performance through the lens of recent improvements to the Java String class, visiting continuous profiling and some assembly code along the way.
Interview:
What's the focus of your work these days?
I am currently working on a continuous profiling product that helps our users pinpoint and analyze performance problems and cost optimization opportunities.
What's the motivation for your talk at QCon London 2024?
To help the audience understand the different phases of code optimization, which tools to use to analyze problems in each phase, and the complex implications of taking action in each phase.
How would you describe your main persona and target audience for this session?
There will be something in this talk for a range of personas, so long as they are interested in Java/the JVM; for example, it will include an analysis of instruction profiles later on in the talk, which some may find enlightening, but this talk may not appeal to hardcore performance engineers. I would like to reach an audience that is suspicious of performance optimization or lacks knowledge and put different levels of optimization into context.
Is there anything specific that you'd like people to walk away with after watching your session?
Be careful specifying public interfaces, which may constrain performance later, because solutions to constrained performance problems can be much more complex and require you to be aware of where your bottlenecks are at all times.
Speaker
Richard Startin
Senior Software Engineer @Datadog
Richard has over ten years experience working as a software engineer, primarily on big data and observability. Richard is an avid open source contributor and blogger. Richard currently works as a Senior Software Engineer at Datadog, where he has been working on tracing and profiling of applications running on the JVM.