End the Senseless Killing: Improving Memory Management for Mobile Operating Systems
Niel Lebeck, Arvind Krishnamurthy, Henry M. Levy, Irene Zhang
Abstract
To ensure low-latency memory allocation, mobile operating systems kill applications instead of swapping memory to disk. This design choice shifts the burden of managing over-utilized memory to application programmers, requiring them to constantly checkpoint their application state to disk. This paper presents Marvin, a new memory manager for mobile platforms that efficiently supports swapping while meeting the strict performance requirements of mobile apps. Marvin's swap-enabled language runtime is co-designed with OS-level memory management to avoid common pitfalls of traditional swap mechanisms. Its key features are: (1) a new swap mechanism, called ahead-of-time (AOT) swap, which pre-writes memory to disk, then harvests it quickly when needed, (2) a modified bookmarking garbage collector that avoids swapping in unused memory, and (3) an object-granularity working set estimator. Our experiments show that Marvin can run more than 2x as many concurrent apps as Android, and that Marvin can reclaim memory over 60x faster than Android with a Linux swap file can allocate memory under memory pressure.
Ask about this paper
Your agent reads all of it.
Lune indexed this paper to the last equation, along with the top-tier papers that cite it. Ask a question and the answer quotes them.
Your agent calls
Luneget_paper_fulltext
Free to start. No credit card required.
Terminal
Install the CLIlune papers fulltext 32dcbaf3-8eba-40aa-88b6-5919bf62260cCited by top-tier papers17
- FwdLLM: Efficient Federated Finetuning of Large Language Models with Perturbed InferencesMengwei Xu, Dongqi Cai, Yaozong Wu, Xiang Li et al.USENIX ATC 2024 · 78 citations
- Acclaim: Adaptive Memory Reclaim to Improve User Experience in Android SystemsYu Liang, Jinheng Li, Rachata Ausavarungnirun, Riwei Pan et al.USENIX ATC 2020 · 49 citations
- ASAP: Fast Mobile Application Switch via Adaptive PrepagingSam Son, Seung Yul Lee, Yunho Jin, Jonghyun Bae et al.USENIX ATC 2021 · 31 citations
- STI: Turbocharge NLP Inference at the Edge via Elastic PipeliningLiwei Guo, Wonkyo Choe, Felix Xiaozhu LinASPLOS 2023 · 25 citations
- CacheSifter: Sifting Cache Files for Boosted Mobile Performance and LifetimeYu Liang, Riwei Pan, Tianyu Ren, Yufei Cui et al.FAST 2022 · 24 citations
Builds on1
Related papers
- More Apps, Faster Hot-Launch on Mobile Devices via Fore/Background-aware GC-Swap Co-designJiacheng Huang, Yunmo Zhang, Junqiao Qiu, Yu Liang et al.ASPLOS 2024 · 12 citations
- MobileSwap: Cross-Device Memory Swapping for Mobile DevicesChanglong Li, Liang Shi, Chun Jason XueDAC 2021 · 10 citations
- Ariadne: A Hotness-Aware and Size-Adaptive Compressed Swap Technique for Fast Application Relaunch and Reduced CPU Usage on Mobile DevicesYu Liang, Aofeng Shen, Chun Jason Xue, Riwei Pan et al.HPCA 2025 · 8 citations
- ElasticZRAM: Revisiting ZRAM for Swapping on Mobile DevicesWentong Li, Dingcui Yu, Yunpeng Song, Longfei Luo et al.DAC 2024 · 2 citations
- PMR: Fast Application Response via Parallel Memory Reclaim on Mobile DevicesWentong Li, Li-Pin Chang, Yu Mao, Liang ShiUSENIX ATC 2025 · 4 citations
