Quality Configurable Approximate DRAM

Raha, Arnab; Sutar, Soubhagya; Jayakumar, Hrishikesh; Raghunathan, Vijay

doi:10.1109/tc.2016.2640296

Cited by 65 publications

(53 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In the hypothesis of a circuit consisting of true-cells, different error rates and looseness levels are explored. For example, an error rate of about 10 −3 is obtained in case of a 60x increase in refresh period [5]. The figure also shows that the looseness level can be used in order to rise SNR orthogonally to error rate, at the cost of keeping some bits exact.…”

Section: Resultsmentioning

confidence: 91%

“…These error models take into account the techniques proposed and the circuital implementations for approximate DRAM and SRAM. Reducing refresh rate in DRAMs and allowing errors is a strategy for reducing power consumption that has produced many research works in the field of approximate memories [1], [5], [10]. The effects of these techniques at bit level can vary depending on DRAM architectures and will be discussed in Section III-A.…”

Section: Memoriesmentioning

confidence: 99%

“…In approximate memories, errors are allowed by design and are non-negligible for the software application. Their presence is the result, in general, of design implementations introduced to significantly reduce energy consumption [4], [5], [6]. Different strategies can be actively used in order to reduce energy consumption introducing approximation: in SRAM, for example, supply voltage scaling can be applied [7], [8]; while in DRAM refresh rate can be reduced or completely disabled [9], [10], [4], since the refresh operation degrades performance and wastes energy (e.g.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

AppropinQuo: A Platform Emulator for Exploring the Approximate Memory Design Space

Stazi

Mastrandrea

Olivieri

et al. 2018

2018 New Generation of CAS (NGCAS)

View full text Add to dashboard Cite

In this work we present AppropinQuo, a flexible and configurable emulator for embedded platforms with approximate memory. The emulator includes models of the effects of approximate memory circuits and architectures, that depend on the internal structure and organization of the cells. The ability to emulate a complete platform, including CPU, peripherals and hardware-software interactions, is particularly important since it allows to execute the application as on the real board, reproducing the effects of errors on output. In fact, output quality is related not only to error rate but it also depends on the application, implementation and its data representation.AppropinQuo allows to run actual applications and operating system as on the physical platform, to analyze the behavior and to expose the effects of specific approximate memory circuits and architectures on output quality. By exploring the design space regarding approximate memories, a complete characterization of the application is possible, as a step toward the determination of the trade-off between saved energy and output quality (energyquality tradeoff).

show abstract

Section: Resultsmentioning

confidence: 91%

Section: Memoriesmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

AppropinQuo: A Platform Emulator for Exploring the Approximate Memory Design Space

Stazi

Mastrandrea

Olivieri

et al. 2018

2018 New Generation of CAS (NGCAS)

View full text Add to dashboard Cite

show abstract

“…As regards DRAMs, different techniques to reduce refresh rate have been proposed. 1,[8][9][10][11] In fact, in exact DRAMs refresh time interval is set according to the worst case access-statistics of the most leaky cells. Commercial DRAM modules, for example, have a worst case retention time of 64 ms determined by the leakiest cells in the entire array.…”

Section: Related Work and Contributionmentioning

confidence: 99%

“…In Ref. [10], after an experimental characterization of memory errors as a function of the DRAM refresh-rate, the authors propose a methodology for constructing a quality configurable approximate DRAM system. Experiments were performed on a FPGA board where a soft-processor (Nios II) and a DDR3 memory controller (UniPHY) are programmed; they revealed a reduction in DRAM refresh power of up to 73% on average.…”

Section: Related Work and Contributionmentioning

confidence: 99%

Full System Emulation of Approximate Memory Platforms with AppropinQuo

Stazi¹,

Mastrandrea²,

Olivieri³

et al. 2019

Journal of Low Power Electronics

View full text Add to dashboard Cite

In this work we present an emulation framework for hardware platforms provided with approximate memory units, called AppropinQuo. The specific characteristic of AppropinQuo is to reveal the effects, on the hardware platform and on software, of errors introduced by approximate memory circuits and architectures. The emulator allows to execute software code without any modification with respect to the target physical board, since it includes the CPU, the memory hierarchy and the peripherals, capturing as well software-hardware interactions and faults due to approximate memory units. The final scope is reproducing the effects of errors generated by approximate memory circuits, allowing to evaluate the impact (quality degradation) on the output produced by the software. In fact, output quality is related to error rate, but their relationship strongly depends on the application, the implementation and its data representation on physical memory. The idea behind approximate memory circuits and approximate computing in general is to trade off energy consumption at the expense of computational accuracy and degradation of output quality. Memory is accounted for a large part of total power consumption in advanced architectures and it is supposed to increase as new memory hungry applications migrate toward the implementation on embedded systems (embedded machine learning, high definition video codecs, etc.). By relaxing design constraints regarding error probability on bit cells, researchers have proposed techniques that significantly reduce memory energy consumption. These techniques, which can be accounted in the general topic of approximate memory design, are implemented at circuit or architecture level, and are specific to the memory technology (i.e., SRAM or DRAM memories). However, the level of acceptable output degradation is the final metric that must be used to assess if, and to what extent, an approximate memory technique can be introduced. Our emulator allows to run actual applications as on the physical platform, to expose the effects of specific approximate memory circuits and architectures on output quality and to vary their parameters (e.g., error rate, number of affected bits, etc.). By exploring the approximate memory design space and its effects on the output of a software application, it is possible to characterize the application behavior, as a step toward the determination of the trade-off between saved energy and output quality (energy-quality tradeoff).

show abstract

Impact of Approximate Memory Data Allocation on a H.264 Software Video Encoder

Stazi

Adani

Mastrandrea

et al. 2018

Lecture Notes in Computer Science

View full text Add to dashboard Cite

This paper describes the analysis, in terms of tolerance to errors on data, of a H.264 software video encoder; proposes a strategy to select data structures for approximate memory allocation and reports the impact on output video quality. Applications that tolerate errors on their data structures are known as ETA (Error Tolerant Applications) and have an important part in pushing interest on approximate computing research. We centered our study on H.264 video encoding, a video compression format developed for use in high definition systems, and today one of the most widespread video compression standard, used for broadcast, consumer and mobile applications. While data fault resilience of H.264 has already been studied considering unwanted and random faults due to unreliable hardware platforms, an analysis, considering controlled hardware faults and the corresponding energy quality tradeoff, has never been proposed.

show abstract

Quality Configurable Approximate DRAM

Cited by 65 publications

References 14 publications

AppropinQuo: A Platform Emulator for Exploring the Approximate Memory Design Space

AppropinQuo: A Platform Emulator for Exploring the Approximate Memory Design Space

Full System Emulation of Approximate Memory Platforms with AppropinQuo

Impact of Approximate Memory Data Allocation on a H.264 Software Video Encoder

Contact Info

Product

Resources

About