Bulkyflip: A NAND-SPIN-Based Last-Level Cache With Bandwidth-Oriented Write Management Policy
文献类型：期刊论文
| 作者 | Wu, Bi2,3; Dai, Pengcheng2,3; Wang, Zhaohao2; Wang, Chao2,3; Wang, Ying4; Yang, Jianlei5; Cheng, Yuanqing2; Liu, Dijun6; Zhang, Youguang1; Zhao, Weisheng2 |
| 刊名 | IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS I-REGULAR PAPERS
 |
| 出版日期 | 2020 |
| 卷号 | 67期号:1页码:108-120 |
| 关键词 | NAND-SPIN spin orbit torque (SOT) MRAM last level cache write throughput high performance |
| ISSN号 | 1549-8328 |
| DOI | 10.1109/TCSI.2019.2947242 |
| 英文摘要 | High capacity last-level caches (LLCs) are being used to help alleviate the growing speed gap between the processor and main memory. However, traditional CMOS based memory technologies (SRAM, DRAM, et al.) for such LLCs consume high static power. Non-volatile memory such as STT-MRAM has been proposed as a low power solution for LLCs. Nevertheless, the high write current induces a so-called "supply current threshold" issue and limits the maximum number of bit-cells that can be written concurrently in one cycle in an STT-MRAM cache. This drawback significantly decreases the bandwidth of the STT-MRAM cache compared with SRAM. In this work, we present a hardware implementation of NAND-like spintronic memory (NAND-SPIN) LLC for the first time. By exploiting the unique erase-then-program operation for writing NAND-SPIN, we propose an adaptive buffer entry (ABE) write policy for each cache write access. Instead of writing a fixed number of bits sequentially, our method adaptively extends the write data length under a fixed maximum cache supply current. Compared to existing STT-MRAM caches, 'ABE' can achieve 70 performance improvements on average. Compared with the conventional early write terminate (EWT) policy, 'ABE' can save 33 write energy on average with negligible hardware overhead. |
| 资助项目 | Asian Research Grant by the University of Notre Dame |
| WOS研究方向 | Engineering |
| 语种 | 英语 |
| WOS记录号 | WOS:000508385000010 |
| 出版者 | IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC |
| 源URL | [http://119.78.100.204/handle/2XEOYT63/15008]  |
| 专题 | 中国科学院计算技术研究所期刊论文_英文 |
| 通讯作者 | Wang, Zhaohao; Wang, Ying; Zhao, Weisheng |
| 作者单位 | 1.Beihang Univ, Sch Elect & Informat Engn, Fert Beijing Res Inst, Beijing 100191, Peoples R China 2.Beihang Univ, Beijing Adv Innovat Ctr Big Data & Brain Comp, Sch Microelect, Fert Beijing Res Inst, Beijing 100191, Peoples R China 3.Beihang Univ, Sch Elect & Informat Engn, Beijing 100191, Peoples R China 4.Chinese Acad Sci, Inst Comp Technol, Beijing 100190, Peoples R China 5.Beihang Univ, Beijing Adv Innovat Ctr Big Data & Brain Comp, Fert Beijing Res Inst, Sch Comp Sci & Engn, Beijing 100191, Peoples R China 6.CAICT, Beijing 100191, Peoples R China 7.Univ Notre Dame, Dept Comp Sci & Engn, Notre Dame, IN 46656 USA |
| 推荐引用方式 GB/T 7714 | Wu, Bi,Dai, Pengcheng,Wang, Zhaohao,et al. Bulkyflip: A NAND-SPIN-Based Last-Level Cache With Bandwidth-Oriented Write Management Policy[J]. IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS I-REGULAR PAPERS,2020,67(1):108-120. |
| APA | Wu, Bi.,Dai, Pengcheng.,Wang, Zhaohao.,Wang, Chao.,Wang, Ying.,...&Hu, Xiaobo Sharon.(2020).Bulkyflip: A NAND-SPIN-Based Last-Level Cache With Bandwidth-Oriented Write Management Policy.IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS I-REGULAR PAPERS,67(1),108-120. |
| MLA | Wu, Bi,et al."Bulkyflip: A NAND-SPIN-Based Last-Level Cache With Bandwidth-Oriented Write Management Policy".IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS I-REGULAR PAPERS 67.1(2020):108-120. |
入库方式： OAI收割
来源：计算技术研究所
浏览0
下载0
收藏0
其他版本
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。