2011 IEEE/ACM International Conference on Computer-Aided Design (ICCAD) 2011
DOI: 10.1109/iccad.2011.6105304
|View full text |Cite
|
Sign up to set email alerts
|

Bandwidth-aware reconfigurable cache design with hybrid memory technologies

Abstract: Abstract-In chip-multiprocessor (CMP) designs, limited memory bandwidth is a potential bottleneck of the system performance. New memory technologies, such as spin-transfer torque memory (STT-RAM), resistive memory (RRAM), and embedded DRAM (eDRAM), are promising on-chip memory solutions for CMPs. In this paper, we propose a bandwidth-aware reconfigurable cache hierarchy (BARCH) with hybrid memory technologies. BARCH consists of a hybrid cache hierarchy, a reconfiguration mechanism, and a statistical prediction… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

0
15
0

Year Published

2015
2015
2020
2020

Publication Types

Select...
5
2
1

Relationship

1
7

Authors

Journals

citations
Cited by 18 publications
(15 citation statements)
references
References 22 publications
0
15
0
Order By: Relevance
“…Zhao et al [14] propose a hybrid cache hierarchy where each level is designed with a memory technology such that the bandwidth provided by the overall hierarchy is optimized. They study the write endurance, latency, energy and bandwidth of different technologies.…”
Section: Sram+sttram Hybrid Cachesmentioning
confidence: 99%
See 2 more Smart Citations
“…Zhao et al [14] propose a hybrid cache hierarchy where each level is designed with a memory technology such that the bandwidth provided by the overall hierarchy is optimized. They study the write endurance, latency, energy and bandwidth of different technologies.…”
Section: Sram+sttram Hybrid Cachesmentioning
confidence: 99%
“…Also, in hybrid caches, early eviction of dead blocks from SRAM improves its utilization [55]. The dynamic cache reconfiguration based techniques [12,14] work on the principle that there exists large variation in intraapplication and inter-application cache requirement of different programs. Thus, by only allocating suitable amount of cache to an application in each phase, significant amount of energy can be saved.…”
Section: Classification and Overviewmentioning
confidence: 99%
See 1 more Smart Citation
“…Consequently, memory bandwidth becomes one of the most important factors that influence high performance system design. To address the bandwidth issue, Zhao et al [61] proposed a bandwidth-aware reconfigurable cache hierarchy to enhance system performance of multicore processors with hybrid memory technologies, by leveraging NVM's bandwidth benefits at large capacities. The hybrid cache hierarchy maximizes the provided bandwidth of processor caches and minimizes the bandwidth demand to the off-chip main memory.…”
Section: Processor Cache Design With Nvmsmentioning
confidence: 99%
“…In recent years, several proposals, either at the circuit or architecture level [Wu et al 2009;Zhao et al 2010;Zhou et al 2009;Rasquinha et al 2010], have been made to address the energy and latency challenges of STT-RAM writes. Another approach for mitigating write problem of STT-RAM cells is device-level optimizations.…”
Section: Introductionmentioning
confidence: 99%