The One Pass (OP) Compactor (an intellectual abstract)
summary
The One Pass Compactor claims full mark-compact heap compaction using a single heap pass by combining pointer threading with an auxiliary address mapping so forwards pointers are fixed via threading and backwards via mapping, keeping the same space overhead as Abadi’s IBM-style compactor and preserving object order. A parallel extension shifts spatial to temporal reasoning, limits threading critical sections to a single CAS, addresses ABA risks, and is published as an intellectual abstract with no implementation or benchmarks yet.
AI-written summary · about
chapters (10)
- 0:00One Pass Compactor intro
- 1:09Compaction motivation
- 3:41Compaction algorithms evolution
- 5:27One Pass Compactor
- 6:09Threaded algorithm
- 9:18IBM address mapping
- 11:30Sequential One Pass Compactor
- 12:57Parallel One Pass Compactor
- 15:29Conclusion
- 15:51Q&A
related talks
23mCorrect Compilation of Concurrent C Code
23mTyped, Concise, Nanopass (pick 3) (Extended Abstract)
34mAutomatic GC Compaction in MRI
46mAhead Of Time (AOT) Internals
21mNavigating the Frontier of EDDO-Based Accelerators: Programming and Code(…)
40mType-safe Nanopasses: How to write a safe and modern compiler front-end
18mParallelism in a Region Inference Context
24mMapping Sparse Applications to Accelerated Computing Systems