An official website of the United States government
Many recent works have shown substantial efficiency boosts from performing inference tasks on Internet of Things (IoT) nodes rather than merely transmitting raw sensor data. However, such tasks, e.g., convolutional neural networks (CNNs), are very compute intensive. They are therefore challenging to complete at sensing-matched latencies in ultra-low-power and energy-harvesting IoT nodes. ReRAM crossbar-based accelerators (RCAs) are an ideal candidate to perform the dominant multiplication-and-accumulation (MAC) operations in CNNs efficiently, but conventional, performance-oriented RCAs, while energy-efficient, are power hungry and ill-optimized for the intermittent and unstable power supply of energy-harvesting IoT nodes. This paper presents the ResiRCA architecture that integrates a new, lightweight, and configurable RCA suitable for energy harvesting environments as an opportunistically executing augmentation to a baseline sense-and-transmit battery-powered IoT node. To maximize ResiRCA throughput under different power levels, we develop the ResiSchedule approach for dynamic RCA reconfiguration. The proposed approach uses loop tiling-based computation decomposition, model duplication within the RCA, and inter-layer pipelining to reduce RCA activation thresholds and more closely track execution costs with dynamic power income. Experimental results show that ResiRCA together with ResiSchedule achieve average speedups and energy efficiency improvements of 8× and 14× respectively compared to a baseline RCA with intermittency-unaware scheduling.
more »
« less
Constraint Based Compiler Optimization for Energy Harvesting Applications
We propose a method for optimizing the energy efficiency of software code running on small computing devices in the Internet of Things (IoT) that are powered exclusively by electricity harvested from ambient energy in the environment. Due to the weak and unstable nature of the energy source, it is challenging for developers to manually optimize the software code to deal with mismatch between the intermittent power supply and the computation demand. Our method overcomes the challenge by using a combination of three techniques. First, we use static program analysis to automatically identify opportunities for precomputation, i.e., computation that may be performed ahead of time as opposed to just in time. Second, we optimize the precomputation policy, i.e., a way to split and reorder steps of a computation task in the original software to match the intermittent power supply while satisfying a variety of system requirements; this is accomplished by formulating energy optimization as a constraint satisfiability problem and then solving the problem using an off-the-shelf SMT solver. Third, we use a state-of-the-art compiler platform (LLVM) to automate the program transformation to ensure that the optimized software code is correct by construction. We have evaluated our method on a large number of benchmark programs, which are C programs implementing secure communication protocols that are popular for energy-harvesting IoT devices. Our experimental results show that the method is efficient in optimizing all benchmark programs. Furthermore, the optimized programs significantly outperform the original programs in terms of energy efficiency and latency, and the overall improvement ranges from 2.3X to 36.7X.
more »
« less
- Award ID(s):
- 2220345
- PAR ID:
- 10467191
- Publisher / Repository:
- Schloss Dagstuhl -- Leibniz-Zentrum f{\"u}r Informatik
- Date Published:
- Format(s):
- Medium: X
- Location:
- Seattle, WA
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Recent advancements in energy-harvesting techniques provide an alternative to batteries for resource-constrained IoT devices and lead to a new computing paradigm, the intermittent computing model. In this model, a software module continues its execution from where it left off when an energy shortage occurred. Enforcing security of an intermittent software module is challenging because its power-off state has to be protected from a malicious adversary in addition to its power-on state, while the security mechanisms put in place must have a low overhead on the performance, resource consumption, and cost of a device. In this paper, we propose SIA (Secure Intermittent Architecture), a security architecture for resource-constrained IoT devices. SIA leverages low-cost security features available in commercial off-the-shelf microcontrollers to protect both the power-on and power-off state of an intermittent software module. Therefore, SIA enables a host of secure intermittent computing applications such as self-attestation, remote attestation, and secure communication. Moreover, our architecture provides confidentiality and integrity guarantees to an intermittent computing module at no cost compared to previous approaches in the literature that impose significant overheads. The salient characteristic of SIA is that it does not require any hardware modifications, and hence, it can be directly applied to existing IoT devices. We implemented and evaluated SIA on a resource-constrained IoT device based on an MSP430 processor. Besides being secure, SIA is simple and efficient. We confirm the feasibility of SIA for resource-constrained IoT devices with experimental results of several intermittent computing applications. Our prototype implementation outperforms by two to three orders of magnitude the secure intermittent computing solution of Suslowicz et al. presented at IGSC 2018.more » « less
-
Recent advancements in energy-harvesting techniques provide an alternative to batteries for resource constrained IoT devices and lead to a new computing paradigm, the intermittent computing model. In this model, a software module continues its execution from where it left off when an energy shortage occurred. Enforcing security of an intermittent software module is challenging because its power-off state has to be protected from a malicious adversary in addition to its power-on state, while the security mechanisms put in place must have a low overhead on the performance, resource consumption, and cost of a device. In this paper, we propose SIA (Secure Intermittent Architecture), a security architecture for resource-constrained IoT devices. SIA leverages low-cost security features available in commercial off-the-shelf microcontrollers to protect both the power-on and power-off state of an intermittent software module. Therefore, SIA enables a host of secure intermittent computing applications such as self-attestation, remote attestation, and secure communication. Moreover, our architecture provides confidentiality and integrity guarantees to an intermittent computing module at no cost compared to previous approaches in the literature that impose significant overheads. The salient characteristic of SIA is that it does not require any hardware modifications, and hence, it can be directly applied to existing IoT devices. We implemented and evaluated SIA on a resource-constrained IoT device based on an MSP430 processor. Besides being secure, SIA is simple and efficient. We confirm the feasibility of SIA for resource-constrained IoT devices with experimental results of several intermittent computing applications. Our prototype implementation outperforms by two to three orders of magnitude the secure intermittent computing solution of Suslowicz et al. presented at IGSC 2018.more » « less
-
Intermittent computing is gaining traction in application domains such as Energy Harvesting Devices (EHDs) that experience arbitrary power failures during program execution. To make progress, programs require system support to checkpoint state and re-execute after power failure by restoring the last saved state. This re-execution should be correct, i.e., simulated by a continuously-powered execution. We study the logical underpinning of intermittent computing and model checkpoint, crash, restore, and re-execution operations as computation on Crash types. We draw inspiration from adjoint logic and define Crash types by introducing two adjoint modality operators to model persistent and transient memory values of partial (re-)executions and the transitions between them caused by checkpoints and restoration. We define a Crash type system for a core calculus. We prove the correctness of intermittent systems by defining a novel logical relation for Crash types.more » « less
-
A Reduction – an accumulation over a set of values, using an associative and commutative operator – is a common computation in many numerical computations, including scientific computations, machine learning, computer vision, and financial analytics. Contemporary polyhedral-based compilation techniques make it possible to optimize reductions, such as prefix sums, in which each component of the reduction’s output potentially shares computation with another component in the reduction. Therefore an optimizing compiler can identify the computation shared between multiple components and generate code that computes the shared computation only once. These techniques, however, do not support reductions that – when phrased in the language of the polyhedral model – span multiple dependent statements. In such cases, existing approaches can generate incorrect code that violates the data dependences of the original, unoptimized program. In this work, we identify and formalize the optimization of dependent reductions as an integer bilinear program. We present a heuristic optimization algorithm that uses an affine sequential schedule of the program to determine how to simplfy reductions yet still preserve the program’s dependences. We demonstrate that the algorithm provides optimal complexity for a set of benchmark programs from the literature on probabilistic inference algorithms, whose performance critically relies on simplifying these reductions. The complexities for 10 of the 11 programs improve siginifcantly by factors at least of the sizes of the input data, which are in the range of 10 4 to 10 6 for typical real application inputs. We also confirm the significance of the improvement by showing speedups in wall-clock time that range from 1.1x to over 10 6 x.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- The DOI is not currently available.
