MOUSE: Inference In Non-volatile Memory for Energy Harvesting Applications

Resch, Salonik; Khatamifard, S. Karen; Chowdhury, Zamshed I.; Zabihi, Masoud; Zhao, Zhengyang; Cılasun, Hüsrev; Wang, Jianping; Sapatnekar, Sachin S.; Karpuzcu, Ulya R.

doi:10.1109/micro50266.2020.00042

Cited by 26 publications

(23 citation statements)

References 77 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…3) Computation: Computation takes the most time and energy, therefore it is critical to make it resilient to intermittent operation and energy efficient. Prior work has demonstrated that in-memory logic performed with MTJs is inherently intermittent safe [77]. We exploit this same property to enable high frequency and light-weight checkpointing during computation.…”

Section: B Operating Semanticsmentioning

confidence: 98%

“…RODENT must perform homomorphic computation while both remaining within a low power budget and being resilient to intermittent operation. We design an accelerator using a non-volatile processing-in-memory (PIM) substrate which is shown to be both highly energy efficient and inherently intermittent resistant [77]. We augment the PIM capability with simple, intermittent safe protocols and hardware for reception, transmission, and encoding.…”

Section: Architecture Designmentioning

confidence: 99%

“…For every operation in each array, we need to specify which logic gate is being performed (what voltage to apply) and the addresses of the input(s) and the output. Just as in prior work [77], the input and output addresses are specified with each instruction. Additionally, as noted in the previous paragraph, we must known which rows (or columns) are currently active.…”

Section: Architecture Designmentioning

confidence: 99%

“…This process repeats until the program has finished or RODENT runs out of power. Write and logic operations with MTJs [15], [52], [95] remain correct when interrupted or performed multiple times [77]. This means that if the driver triggers a logic operation in the computation arrays, it can be interrupted i.…”

Section: B Operating Semanticsmentioning

confidence: 99%

“…Checkpointing and guaranteeing correctness has significant overhead for energy efficiency and complexity [53]. We design the accelerator to make use of non-volatile processing-inmemory (PIM), which not only provides high energy efficiency, but inherent resilience to interruptions and low-cost checkpointing mechanisms [77].…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Towards Homomorphic Inference Beyond the Edge

Resch¹,

Chowdhury²,

Cılasun³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Beyond edge devices can function off the power grid and without batteries, enabling them to operate in difficult to access regions. However, energy costly long-distance communication required for reporting results or offloading computation becomes a limitation. Here, we reduce this overhead by developing a beyond edge device which can effectively act as a nearby server to offload computation. For security reasons, this device must operate on encrypted data, which incurs a high overhead. We use energy-efficient and intermittent-safe in-memory computation to enable this encrypted computation, allowing it to provide a speedup for beyond edge applications within a power budget of a few milliWatts.

show abstract

Section: B Operating Semanticsmentioning

confidence: 98%

Section: Architecture Designmentioning

confidence: 99%

Section: Architecture Designmentioning

confidence: 99%

Section: B Operating Semanticsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Towards Homomorphic Inference Beyond the Edge

Resch¹,

Chowdhury²,

Cılasun³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

Experimental demonstration of magnetic tunnel junction-based computational random-access memory

Lv,

Zink,

Bloom

et al. 2024

npj Unconv. Comput.

View full text Add to dashboard Cite

The conventional computing paradigm struggles to fulfill the rapidly growing demands from emerging applications, especially those for machine intelligence because much of the power and energy is consumed by constant data transfers between logic and memory modules. A new paradigm, called “computational random-access memory (CRAM),” has emerged to address this fundamental limitation. CRAM performs logic operations directly using the memory cells themselves, without having the data ever leave the memory. The energy and performance benefits of CRAM for both conventional and emerging applications have been well established by prior numerical studies. However, there is a lack of experimental demonstration and study of CRAM to evaluate its computational accuracy, which is a realistic and application-critical metric for its technological feasibility and competitiveness. In this work, a CRAM array based on magnetic tunnel junctions (MTJs) is experimentally demonstrated. First, basic memory operations, as well as 2-, 3-, and 5-input logic operations, are studied. Then, a 1-bit full adder with two different designs is demonstrated. Based on the experimental results, a suite of models has been developed to characterize the accuracy of CRAM computation. Scalar addition, multiplication, and matrix multiplication, which are essential building blocks for many conventional and machine intelligence applications, are evaluated and show promising accuracy performance. With the confirmation of MTJ-based CRAM’s accuracy, there is a strong case that this technology will have a significant impact on power- and energy-demanding applications of machine intelligence.

show abstract

Deep Learning on Energy Harvesting IoT Devices: Survey and Future Challenges

2022

IEEE Access

View full text Add to dashboard Cite

Internet-of-Things (IoT) devices are becoming both intelligent and green. On the one hand, Deep Neural Network (DNN) compression techniques make it possible to run deep learning applications on devices equipped with low-end microcontrollers (MCUs). By performing deep learning on IoT devices, in-situ decision-making can be made, which can improve the responsiveness of such devices to the environment and reduce data uploading to edge servers or clouds to save valuable network bandwidth. On the other hand, many IoT devices in the future will be powered by energy harvesters instead of batteries to reduce environmental pollution and achieve permanent service free of battery maintenance. As the energy output of energy harvesters is tiny and unstable, energy harvesting IoT (EH-IoT) devices will experience frequent power failures during their execution, making the software task hard to progress. The deep learning tasks running on such devices must face this challenge and, at the same time, ensure satisfactory execution efficiency. We believe deploying deep learning on EH-IoT devices that execute intermittently will be a challenging yet promising research direction. To motivate research in this direction, this paper summarizes existing solutions and provides an in-depth discussion of future challenges that deserve further investigation. With IoT devices becoming more intelligent and green, DNN inference on EH-IoT devices will generate a much more significant impact in the future in academia and industry.

show abstract

MOUSE: Inference In Non-volatile Memory for Energy Harvesting Applications

Cited by 26 publications

References 77 publications

Towards Homomorphic Inference Beyond the Edge

Towards Homomorphic Inference Beyond the Edge

Experimental demonstration of magnetic tunnel junction-based computational random-access memory

Deep Learning on Energy Harvesting IoT Devices: Survey and Future Challenges

Contact Info

Product

Resources

About