Hybrid2: Combining Caching and Migration in Hybrid Memory Systems
Evangelos Vasilakis, Vassilis Papaefstathiou, Pedro Trancoso, Ioannis Sourdis
Abstract
This paper considers a hybrid memory system composed of memory technologies with different characteristics; in particular a small, near memory exhibiting high bandwidth, i.e., 3D-stacked DRAM, and a larger, far memory offering capacity at lower bandwidth, i.e., off-chip DRAM. In the past, the near memory of such a system has been used either as a DRAM cache or as part of a flat address space combined with a migration mechanism. Caches and migration offer different tradeoffs (between performance, main memory capacity, data transfer costs, etc.) and share similar challenges related to data-transfer granularity and metadata management. This paper proposes Hybrid 2 , a new hybrid memory system architecture that combines a DRAM cache with a migration scheme. Hybrid 2 does not deny valuable capacity from the memory system because it uses only a small fraction of the near memory as a DRAM cache; 64MB in our experiments. It further leverages the DRAM cache as a staging area to select the data most suitable for migration. Finally, Hybrid 2 alleviates the metadata overheads of both DRAM caches and migration using a common mechanism. Using near to far memory ratios of 1:16, 1:8 and 1:4 in our experiments, Hybrid 2 on average outperforms current state-of-the-art migration schemes by 7.9%, 9.1% and 6.4%, respectively. In the same system configurations, compared to DRAM caches Hybrid 2 gives away on average only 0.3%, 1.2%, and 5.3% of performance offering 5.9%, 12.1%, and 24.6% more main memory capacity, respectively.
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 644b0d6f-26c5-4b5a-8d1e-325cadc0e5f6Cited by top-tier papers5
- Demystifying CXL Memory with Genuine CXL-Ready Systems and DevicesYan Sun, Yifan Yuan, Zeduo Yu, Reese Kuper et al.MICRO 2023 · 133 citations
- Sibyl: adaptive and extensible data placement in hybrid storage systems using online reinforcement learningGagandeep Singh, Rakesh Nadig, Jisung Park, Rahul Bera et al.ISCA 2022 · 36 citations
- NOMAD: Enabling Non-blocking OS-managed DRAM Cache via Tag-Data DecouplingYoungin Kim, Hyeonjin Kim, William J. SongHPCA 2023 · 10 citations
- Stream-Based Data Placement for Near-Data Processing with Extended MemoryYiwei Li, Boyu Tian, Yi Ren, Mingyu GaoMICRO 2024 · 5 citations
- Bumblebee: A MemCache Design for Die-stacked and Off-chip Heterogeneous Memory SystemsYifan Hua, Shengan Zheng, Ji Yin, Weidong Chen et al.DAC 2023 · 4 citations
Related papers
- Baryon: Efficient Hybrid Memory Management with Compression and Sub-BlockingYiwei Li, Mingyu GaoHPCA 2023 · 14 citations
- Blenda: Dynamically-Reconfigurable Stacked DRAMMohammad Bakhshalipour, Hamidreza Zare, Farid Samandi, Fatemeh Golshan et al.MICRO 2024 · 1 citation
- OpenMem: Hardware/Software Cooperative Management for Mobile Memory SystemFei Wen, Mian Qin, Paul Gratz, A. L. Narasimha ReddyDAC 2021 · 9 citations
- ABNDP: Co-optimizing Data Access and Load Balance in Near-Data ProcessingBoyu Tian, Qihang Chen, Mingyu GaoASPLOS 2023 · 31 citations
- Stealth-Persist: Architectural Support for Persistent Applications in Hybrid Memory SystemsMazen Al-Wadi, Vamsee Reddy Kommareddy, Clayton Hughes, Simon David Hammond et al.HPCA 2021 · 9 citations
