How (Not) to Write Java Pointer Analyses after 2020
summary
Argues a concrete checklist of precise-yet-efficient design rules for Java pointer analyses, backed by multi-year implementation experience and a runnable prototype analyzer. Describes the actual trade-offs measured on standard Java benchmarks: precision losses that cut runtime by orders of magnitude, surprising cases where common optimizations harm recall, and which heap and context models produce the best cost/benefit. Includes code-level techniques, evaluation numbers, and practical pitfalls for new researchers.
AI-written summary · about
related talks
29mHow (Not) To Write Java Pointer Analyses After 2020
42mAn Opinionated Overview on Static Analysis for Java
15mMaking Pointer Analysis More Precise by Unleashing the Power of Selective Context Sensitivity
55mThe Performance Engineer's Guide To Java HotSpot
59mMove Deliberately and Don’t Break Anything: Lessons from...Java
43mRolling with New Java Enhancements: Practical Design Patterns
19mPointer Analysis for Database-Backed Applications
45mHow difficult can it be to write efficient code?