Abstract
Phase-change memory (PCM) has garnered attention as a next-generation memory owing to its non-volatility and scalability. However, PCM wears out under excessive write accesses; hence, it must be supported by wear-leveling algorithms to uniformly distribute the number of accesses across the entire address space. Table-based wear leveling is one of the representative algorithms that stores a write counter for each address region for remapping frequently accessed addresses with lower overhead; however, write counters consume resources in a PCM system. In this study, a bit-width reduction method in write counters for wear leveling is proposed, where the method utilizes a stochastic finite-state machine to probabilistically count the number of write accesses. The proposed method shows only a 1.2% lifetime degradation using six bits for each counter, with 40% fewer resources spent on write counters when the endurance of a 4KB block is 1E+06.
| Original language | English |
|---|---|
| Pages (from-to) | 413-419 |
| Number of pages | 7 |
| Journal | IEIE Transactions on Smart Processing and Computing |
| Volume | 9 |
| Issue number | 5 |
| DOIs | |
| State | Published - 30 Oct 2020 |
Keywords
- Finite-state machine
- Non-volatile memory
- Phase-change memory
- Wear leveling
Fingerprint
Dive into the research topics of 'Bit-width reduction in write counters for wear leveling in a phase-change memory system'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver