Anchor-NDP: An importance-based prefix KV management framework for CXL-enabled NDP
Authors:
- Hansol KimHKHansol Kim
- Woohyun KimWKWoohyun Kim
- Taehee HanTaehee Han
- 17 October 2026
- journal article
- Published by Elsevier in Journal of Systems Architecture
- Vol. 181
Abstract
No abstract availableFunding Information
- Ministry of Trade, Industry and Energy
- Samsung
- Institute for Information and Communications Technology Promotion
This publication has 22 references indexed in Scilit:
- 🧜Siren’s Song in the AI Ocean: A Survey on Hallucination in Large Language ModelsComputational Linguistics, 2025
- RAGCache: Efficient Knowledge Caching for Retrieval-Augmented GenerationACM Transactions on Computer Systems, 2025
- PNM Meets Sparse Attention: Enabling Multi-Million Tokens Inference at ScaleIEEE Computer Architecture Letters, 2025
- PIM Is All You Need: A CXL-Enabled GPU-Free System for Large Language Model Inference2025
- CacheGen: KV Cache Compression and Streaming for Fast Large Language Model Serving2024
- PCI-Express: Evolution of a Ubiquitous Load-Store Interconnect Over Two Decades and the Path Forward for the Next Two DecadesIEEE Circuits and Systems Magazine, 2024
- Efficient Memory Management for Large Language Model Serving with PagedAttention2023
- AccelWattch: A Power Modeling Framework for Modern GPUs2021
- The NarrativeQA Reading Comprehension ChallengeTransactions of the Association for Computational Linguistics, 2018
- Ramulator: A Fast and Extensible DRAM SimulatorIEEE Computer Architecture Letters, 2015