Current Status and Future Plans of DAMON
summary
DAMON provides a centralized memory access monitoring framework that enforces bounded overhead by default and conceptually scans all system memory every 5 ms while using under 1% CPU. It identifies hot regions — found 7 GB working set among 70 GB — and runs configurable schemes to apply actions (reclaim, migrate, THP promotion/demotion). Reproduced prior research: one demo reduces THP memory bloat 76% while keeping 51% of THP speed, another cuts 93% of resident set and 23% system memory with ~1.3–2% runtime overhead. Mainline since 5.16; plans include IBS, error-info sampling, auto‑tuning and allow/deny lists.
AI-written summary · about
chapters (18)
- 0:00Introduction and motivation
- 1:24DAMON overview and challenges
- 3:50Current status of DAMON
- 6:43DAMON in production
- 8:08Demos: automated optimizations
- 10:57Demos safety and quotas
- 13:08DAMON and Demos in Linux
- 15:16Demos modules and tools
- 17:25Future plans for DAMON
- 20:37Extending DAMON use cases
- 23:03Improving DAMON overhead control
- 25:14DAMON allow/deny lists
- 26:19New Demos actions and modules
- 27:42Unifying Demos modules
- 28:03DAMON auto-tuning
- 30:30Community and future goals
- 35:51Community handover and dreams
- 38:49Summary and Q&A
related talks
46mDAMON Current Status and Future Plans
42mDAMON Recipes: Ways to Save Memory Using a Linux Kernel Subsystem in the Real World- 42mDAMON Recipes: Ways to Save Memory Using a Linux Kernel Subsystem in the Real World
52mDAMON, DAMOS, and DAMO: Kernel Subsystems and User-Space Tools for Data Access...
47mDAMO[N,S]?: Implementing Self-Driven Data Access-Aware Efficient Linux System
48mData Access Monitoring Operator (DAMO): User-Space Tool/Python Library for Access-Aware Profiling and Optimization of Your Linux Systems
43mHybrid THP Mechanism. Selective Use of Huge Pages by Hot Applications
26mLet's demystify memory