Corpus ID: 15427862

Refreshing Thoughts on DRAM : Power Saving vs . Data Integrity

@inproceedings{Rahmati2014RefreshingTO,
  title={Refreshing Thoughts on DRAM : Power Saving vs . Data Integrity},
  author={Amir Rahmati and Matthew Hicks and Daniel E. Holcomb and Kevin Fu},
  year={2014}
}
To head-off the trend of increasing power consumption and throughput overheads due to refresh in DRAM, researchers are exploring ways to fine-tune refresh rate. Refresh management proposals range from temperature-aware refresh to partitioning data cells based on volatility, with each partition having its own refresh rate. One hurdle in this area is the lack of precise description of the evaluation setups used in current proposals. The incomplete description makes it difficult to reproduce… Expand

Figures and Tables from this paper

Sparkk : Quality-Scalable Approximate Storage in DRAM
DRAM memory stores its contents in leaky cells that require periodic refresh to prevent data loss. The refresh operation does not only degrade system performance, but also consumes significantExpand
Quality Configurable Approximate DRAM
TLDR
This paper proposes four novel strategies for partitioning the DRAM in a system into a number of quality bins based on the frequency, location, and nature of bit errors in each of the physical pages, while also taking into account the property of variable retention time exhibited by DRAM cells. Expand
Quality-aware data allocation in approximate DRAM*
TLDR
Experimental results demonstrate a significant improvement in the energy-quality trade-off compared to previous work and show a reduction in DRAM refresh power of up to 73% with minimal loss in output quality. Expand
Probable cause: The deanonymizing effects of approximate DRAM
TLDR
This paper builds an approximate memory platform and performs experiments that show that the fingerprint due to approximation is device dependent and resilient to changes in environment and level of approximation, and develops a distance metric that yields a two-orders-of-magnitude difference between approximate results produced by the same DRAM chip and those produced by other DRAM chips. Expand
A Platform to Analyze DDR3 DRAM’s Power and Retention Time
TLDR
The authors explore the intrinsic trade-off in a DRAM between the power consumption (due to refresh) and the reliability and their unique platform allows tailoring to the design constraints depending on whether power consumption, performance or reliability has the highest design priority. Expand
Memory-Aware Scheduling for Fixed Priority Hard Real-Time Computing Systems
TLDR
This dissertation focuses on studying how to schedule high-priority hard real-time tasks with memory impacts taken into considerations, and develops an approach that takes into consideration the increasing transistor density and surging access demands from a rapidly growing number of processing cores. Expand
Mitigating Wordline Crosstalk Using Adaptive Trees of Counters
TLDR
A Counter-based Adaptive Tree (CAT) approach to mitigate wordline crosstalk using adaptive trees of counters to guide appropriate refreshing of vulnerable rows by tuning the distribution of the counters to the rows in a bank based on the memory reference patterns. Expand
Processor/memory Co-Scheduling using periodic resource server for real-time systems under peak temperature constraints
TLDR
The approach takes advantage of the periodic resource model for its hard deadline guarantee capability and in the meantime, by periodically (deterministically) throttling the accesses of the CPU and memory resources, can effectively guarantee the thermal constraints for both theCPU and memory. Expand
Energy-efficient hardware design based on high-level synthesis
TLDR
This dissertation describes research activities broadly concerning the area of High-level synthesis, but more specifically, regarding the HLS-based design of energy-efficient hardware (HW) accelerators and the development of a methodology complementing HLS to automatically derive power optimization directives. Expand
Approximate Flash Storage: A Feasibility Study
Approximate storage is an essential part of any approximate computing architecture. In this work, we provide a platform for conducting experiments on flash storage targeted at approximate computingExpand
...
1
2
...

References

SHOWING 1-10 OF 21 REFERENCES
Understanding and mitigating refresh overheads in high-density DDR4 DRAM systems
TLDR
An analysis of DDR4 DRAM's FGR feature is conducted, and there is no one-size-fits-all option across a variety of applications, and Adaptive Refresh is presented, a simple yet effective mechanism that dynamically chooses the best FGR mode for each application and phase within the application. Expand
Flikker: saving DRAM refresh-power through critical data partitioning
TLDR
Flikker exposes and leverages an interesting trade-off between energy consumption and hardware correctness, and shows that many applications are naturally tolerant to errors in the non-critical data, and in the vast majority of cases, the errors have little or no impact on the application's final outcome. Expand
Block-based multi-period refresh for energy efficient dynamic memory
  • Joohee Kim, M. Papaefthymiou
  • Computer Science, Chemistry
  • Proceedings 14th Annual IEEE International ASIC/SOC Conference (IEEE Cat. No.01TH8558)
  • 2001
TLDR
A novel scheme that relies on multiple refresh periods and small refresh blocks to reduce DRAM dissipation by decreasing the number of cells refreshed too often and gives a novel polynomial-time algorithm for computing an optimal set of refresh periods for block-based multiperiod refresh. Expand
RAIDR: Retention-aware intelligent DRAM refresh
TLDR
This paper proposes RAIDR (Retention-Aware Intelligent DRAM Refresh), a low-cost mechanism that can identify and skip unnecessary refreshes using knowledge of cell retention times and group DRAM rows into retention time bins and apply a different refresh rate to each bin. Expand
RAMZzz: Rank-aware DRAM power management with dynamic migrations and demotions
TLDR
This paper proposes a novel memory system design named RAMZzz with rank-aware energy saving optimizations that relies on a memory controller to monitor the memory access locality, and group the pages with similar access locality into the same rank. Expand
MemScale: active low-power modes for main memory
TLDR
The results demonstrate that MemScale reduces energy consumption significantly compared to modern memory energy management approaches, and it is concluded that the potential benefits of the MemScale mechanisms and policy more than compensate for their small hardware cost. Expand
Retention-aware placement in DRAM (RAPID): software methods for quasi-non-volatile DRAM
TLDR
This work proposes retention-aware placement in DRAM (RAPID), novel software approaches that can exploit off-the-shelf DRAMs to reduce refresh power to vanishingly small levels approaching non-volatile memory. Expand
Hardware/software techniques for DRAM thermal management
TLDR
This work develops a thermal model to estimate the temperature of DRAM chips and proposes three hardware and software schemes to reduce peak temperatures, and introduces a new cache line replacement policy that reduces the number of accesses to the overheatingDRAM chips. Expand
PPT: joint performance/power/thermal management of DRAM memory for multi-core systems
TLDR
This paper proposes the first joint performance, power and thermal management framework (PPT) through orchestrating task execution and page allocation and adapts to system loading to maximize power saving and avoid memory hotspot at the same time whiling sustaining the system bandwidth demand. Expand
Memory power management via dynamic voltage/frequency scaling
TLDR
A large opportunity for memory power reduction is demonstrated with a simple control algorithm that adjusts memory voltage and frequency based on memory bandwidth utilization, and a simple algorithm is evaluated in a real system. Expand
...
1
2
3
...