Logo image
VRL-DRAM: Improving DRAM Performance via Variable Refresh Latency
Conference proceeding

VRL-DRAM: Improving DRAM Performance via Variable Refresh Latency

Anup Das, Hasan Hassan, Onur Mutlu and IEEE
2018 55TH ACM/ESDA/IEEE DESIGN AUTOMATION CONFERENCE (DAC), v 137710, pp 1-6
01 Jan 2018

Abstract

Computer Science Computer Science, Artificial Intelligence Computer Science, Hardware & Architecture Science & Technology Technology
A DRAM chip requires periodic refresh operations to prevent data loss due to charge leakage in DRAM cells. Refresh operations incur significant performance overhead as a DRAM bank/rank becomes unavailable to service access requests while being refreshed. In this work, our goal is to reduce the performance overhead of DRAM refresh by reducing the latency of a refresh operation. We observe that a significant number of DRAM cells can retain their data for longer than the worst-case refresh period of 64ms. Such cells do not always need to be fully refreshed; a low-latency partial refresh is sufficient for them. We propose Variable Refresh Latency DRAM (VRL-DRAM), a mechanism that fully refreshes a DRAM cell only when necessary, and otherwise ensures data integrity by issuing low-latency partial refresh operations. We develop a new detailed analytical model to estimate the minimum latency of a refresh operation that ensures data integrity of a cell with a given retention time profile. We evaluate VRL-DRAM with memory traces from real workloads, and show that it reduces the average refresh performance overhead by 34% compared to the state-of-the-art approach.

Metrics

15 Record Views
27 citations in Scopus

Details

InCites Highlights

Data related to this publication, from InCites Benchmarking & Analytics tool:

Collaboration types
Domestic collaboration
International collaboration
Web of Science research areas
Computer Science, Artificial Intelligence
Computer Science, Hardware & Architecture
Logo image