Make big things run again! Insight on Eclipse Platform’s performance
summary
Eclipse Platform profiling with YourKit Java Profiler drives concrete fixes: a regex-heavy replaceAll call becomes String.replace and drops from ~6 seconds to 3 milliseconds; adding a UI throttle (100 ms) stops thousands of redundant updates; swapping duplicated classpath rules into a cache cuts ~200 MB to ~300 KB of heap but briefly regresses auto-builds (later fixed); project build parallelization and tiny memory fixes shrink typical large-workspace usage (1,600 projects, 60,000 Java files) from >1 GB heap and long startup to noticeably lower footprints and faster UI responsiveness.
AI-written summary · about
chapters (18)
- 0:00Intro
- 1:24Where performance matters
- 4:49Startup Time
- 5:52Single Thread Build
- 6:23Parallel Project Build
- 8:48Import Projects from Git Repository
- 9:24Throttler API
- 12:05Check for Updates
- 14:40Zero Length Arrays
- 17:24Optimize Data Structures
- 20:15With introduced Cache
- 20:49Regression
- 23:49Performance Tests
- 24:17Challenges
- 25:28Slow scrolling on macOS
- 26:15Text Search with many fast results
- 27:48Java Project Wizard
- 30:57WE NEED YOUR HELP!
related talks
48mYour first Application based on Eclipse FAQ
49mThe Java Council
39mCreate an Eco-Friendly World with Green Software Engineering
50mEnergy Efficient Design Principles of Software Engineering
12mThe State of the Eclipse Community
47mMigrating to Modules
15mDivining Profiler Accuracy: An Approach to Approximate Profiler Accuracy Through(…)
23mMigrate your RCP software now!