Volatile-by-Default VbD
A volatile-by-default semantics for Java in which every memory access is treated as volatile (sequentially consistent) by default. Bare VbD carries a considerable penalty (28% average / 81% maximum on x86; 57% / 157% on Armv8), so the authors add a just-in-time technique that speculatively treats each object as thread-local and compiles its accesses without fences, recompiling to insert fences once concurrent access is detected — bringing the Armv8 overhead down to 37% average / 73% maximum. It admits exactly the SC behaviours.
Properties
Property vector survey-sourced — from Tables 1–2 of the Moiseenko et al. survey, except cells that cite a specific source.
- Compilation optimal mapping to
- no x86
- no POWER
- no Armv7
- no Armv8
- Reordering sound
- no Store→Load
- no Store→Store
- no Load→Load
- no Load→Store
- Elimination sound
- yes Store/Load
- yes Store/Store
- yes Load/Load
- yes Load/Store
- Other local transformations
- yes Irrelevant load elim.
- yes Speculative load intro.
- yes Roach motel
- yes Inverse roach motel Poetzl & Kroening 2015
- yes Strengthening
- yes Trace preserving
- no Common subexpr. elim.
- Global transformations
- yes Register promotion
- yes Thread inlining
- Reasoning guarantees
- yes External DRF
- yes Coherence
- yes No undefined behaviour
- yes In-order execution
- yes No out-of-thin-air
Ordering relationships
- Equivalent to
- Sequential Consistency (SC) — Volatile-by-default makes every Java access sequentially consistent, so it admits exactly the SC outcomes; it differs from SC only in enforcement mechanism (speculative JIT recompilation) and performance.
References
- Lun Liu, Todd Millstein, Madanlal Musuvathi. A Volatile-by-Default JVM for Server Applications. OOPSLA 2017 (PACMPL vol. 1), 2017. doi:10.1145/3133873
- Lun Liu, Todd Millstein, Madanlal Musuvathi. Accelerating Sequential Consistency for Java with Speculative Compilation. PLDI 2019, 2019. doi:10.1145/3314221.3314611
- Evgenii Moiseenko, Anton Podkopaev, Dmitrii Koznov. A Survey of Programming Language Memory Models. Programming and Computer Software 47(6), pp. 439–456, 2021. doi:10.1134/S0361768821060050
- Daniel Poetzl, Daniel Kroening. Formalizing and Checking Thread Refinement for Data-Race-Free Execution Models. arXiv:1510.07171, 2015. arxiv.org/abs/1510.07171