Integration scheme for retention-aware DRAM refresh

Cheng, Wei-Chen; Li, Xin-Lun; Chen, Jian-Kai

doi:10.1109/edssc.2017.8126536

Cited by 2 publications

(4 citation statements)

References 3 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In research [5], in addition to all-bank refresh and per-bank refresh, two new fine-granularity refresh schemes, namely half-bank refresh, and quarter-bank refresh were proposed to further reduce unnecessary refresh. Although this integration refresh scheme combined both the advantage of all-bank refresh and per-bank refresh, provided fine-granularity refresh modes selection to resolve the problem of too much unnecessary refresh in all-bank refresh and high refresh overhead in per-bank refresh, it did not integrate the 2x granularity auto-refresh and hence the performance improvement was less than 10%.…”

Section: Motivation and Examplementioning

confidence: 99%

“…There has been some research on reducing the overhead of DRAM by retention-aware refresh or refresh scheduling [3,4,5,6,7,8,9,10,11,12]. Research [4] proposed a RAAR technique to reduce refresh energy and performance degradation, only part of refresh bundles need a refresh in each refresh interval, while other refresh bundles can skip at least 50% of refresh operations.…”

Section: Introductionmentioning

confidence: 99%

“…Although this methodology could reduce a lot of redundant refreshes, because it chose a unique refresh scheme all over the RAAR period, it still wastes quite a number of refresh time on normal cells. Research [5] improved this phenomenon by integrating all the RAAR refresh schemes and chose the optimal refresh scheme for different refresh bundles to minimize refresh time and refresh energy. Research [6] integrated the RAAR scheme with the bank reordering technique to gather up weak rows in adjacent banks, and hence less refresh bundles need refresh in each refresh interval.…”

Section: Introductionmentioning

confidence: 99%

“…In summary of the related works that addressed the RAAR technique, although research [5] had proposed an integration refresh scheme, it did not integrate the 2x granularity auto-refresh and hence the performance improvement was less than 10%. Research [6] and research [8] also reveal the fact that although they integrated RAAR scheme with bank ordering and BISR technique, respectively, there was still very little performance improvement achieved.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Retention-Aware DRAM Auto-Refresh Scheme for Energy and Performance Efficiency

Cheng

Shen

2019

Micromachines

Self Cite

View full text Add to dashboard Cite

Dynamic random access memory (DRAM) circuits require periodic refresh operations to prevent data loss. As DRAM density increases, DRAM refresh overhead is even worse due to the increase of the refresh cycle time. However, because of few the cells in memory that have lower retention time, DRAM has to raise the refresh frequency to keep the data integrity, and hence produce unnecessary refreshes for the other normal cells, which results in a large refresh energy and performance delay of memory access. In this paper, we propose an integration scheme for DRAM refresh based on the retention-aware auto-refresh (RAAR) method and 2x granularity auto-refresh simultaneously. We also explain the corresponding modification need on memory controllers to support the proposed integration refresh scheme. With the given profile of weak cells distribution in memory banks, our integration scheme can choose the most appropriate refresh technique in each refresh time. Experimental results on different refresh cycle times show that the retention-aware refresh scheme can properly improve the system performance and have a great reduction in refresh energy. Especially when the number of weak cells increased due to the thermal effect of 3D-stacked architecture, our methodology still keeps the same performance and energy efficiency.

show abstract

Section: Motivation and Examplementioning

confidence: 99%