7.4 A 256b-wordlength ReRAM-based TCAM with 1ns search-time and 14× improvement in wordlength-energyefficiency-density product using 2.5T1R cell

2016 IEEE International Solid-State Circuits Conference (ISSCC)(2016)

引用 68|浏览181
暂无评分
摘要
Ternary content-addressable memory (TCAM) is used in search engines for network and big-data processing [1]–[6]. Nonvolatile TCAM (nvTCAM) was developed to reduce cell area (A), search energy (ES), and standby power beyond what can be achieved using SRAM-based TCAM (sTCAM) [1]–[2]: particularly in applications with long idle times and frequent-search-few-write operations. nvTCAMs were previously designed using diode-4T2R (D4T2R) with STT-MTJ [3], 2T2R with phase-change memory [4], 4T2R and 3T1R with ReRAM [5,6]. However, these NV devices suffer from the following issues: 1) High ES requirements due to cell-DC-current (IDC-CELL) as well as large match-line (ML) parasitic load (CML), particularly when word-length (WDL) is long; 2) Large A due to the use of two NVM (2R) devices [3]–[5] or in-cell control logic [6]; 3) Limited WDL caused by small ML current-ratio (IML-ratio ≅ IML-MIS/N*IML-M) between mismatch current (IML-MIS) and the leakage-current (IML-M) from cells on a ML, particularly when NVM resistance (R)-ratio (=RHRS /RLRS) between high-R (HRS, RHRS) and low-R (LRS, RLRS) states is small due to process variation; 4) Long search delays (TSD) due to large CML and small IML-ratio. This work proposes 1) a 2.5T1R cell to reduce A, CML, and ES as well as increase IML-ratio; and 2) a region-splitter (RS) sense amplifier (SA) to achieve robust sensing with a smaller ML-Voltage (VML) swing (VMLS) to reduce TSD and ES.
更多
查看译文
关键词
ternary content-addressable memory,search engines,robust sensing,region-splitter sense amplifier,TSD,search delays,NVM resistance-ratio,leakage-current,mismatch current,ML current-ratio,in-cell control logic,WDL,word-length,CML,match-line parasitic load,cell-DC-current,ReRAM,phase-change memory,STT-MTJ,D4T2R,diode-4T2R,sTCAM,SRAM-based TCAM,nvTCAM,nonvolatile TCAM
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要