2017 IEEE/ACM International Symposium on Low Power Electronics and Design (ISLPED)最新文献_第6页

Approximate memory compression for energy-efficiency 近似内存压缩能源效率

2017 IEEE/ACM International Symposium on Low Power Electronics and Design (ISLPED) Pub Date : 2017-07-01 DOI: 10.1109/ISLPED.2017.8009173

Ashish Ranjan, Arnab Raha, V. Raghunathan, A. Raghunathan

{"title":"Approximate memory compression for energy-efficiency","authors":"Ashish Ranjan, Arnab Raha, V. Raghunathan, A. Raghunathan","doi":"10.1109/ISLPED.2017.8009173","DOIUrl":"https://doi.org/10.1109/ISLPED.2017.8009173","url":null,"abstract":"Memory subsystems are a major energy bottleneck in computing platforms due to frequent transfers between processors and off-chip memory. We propose approximate memory compression, a technique that leverages the intrinsic resilience of emerging workloads such as machine learning and data analytics to reduce off-chip memory traffic and energy. To realize approximate memory compression, we enhance the memory controller to be aware of memory regions that contain approximation-resilient data, and to transparently compress/decompress the data written to/read from these regions. To provide control over approximations, the quality-aware memory controller conforms to a specified error constraint for each approximate memory region. We design a software interface that programmers can use to identify data structures that are resilient to approximations. We also propose a runtime quality control framework that automatically determines the error constraints for the identified data structures such that a given target application-level quality is maintained. We evaluate our proposal by implementing a hardware prototype using the Intel UniPHY-DDR3 memory controller and NIOS-II processor, a Hynix DDR3 DRAM module, and a Stratix-IV FPGA development board. Across a suite of 8 machine learning benchmarks, approximate memory compression obtains a 1.28× benefit in DRAM energy and a simultaneous 11.5% improvement in execution time for a small (< 1.5%) loss in output quality.","PeriodicalId":385714,"journal":{"name":"2017 IEEE/ACM International Symposium on Low Power Electronics and Design (ISLPED)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2017-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122204334","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 24

Power optimizations in MTJ-based Neural Networks through Stochastic Computing 基于随机计算的mtj神经网络功率优化

2017 IEEE/ACM International Symposium on Low Power Electronics and Design (ISLPED) Pub Date : 2017-07-01 DOI: 10.1109/ISLPED.2017.8009167

Ankit Mondal, Ankur Srivastava

{"title":"Power optimizations in MTJ-based Neural Networks through Stochastic Computing","authors":"Ankit Mondal, Ankur Srivastava","doi":"10.1109/ISLPED.2017.8009167","DOIUrl":"https://doi.org/10.1109/ISLPED.2017.8009167","url":null,"abstract":"Artificial Neural Networks (ANNs) have found widespread applications in tasks such as pattern recognition and image classification. However, hardware implementations of ANNs using conventional binary arithmetic units are computationally expensive, energy-intensive and have large area overheads. Stochastic Computing (SC) is an emerging paradigm which replaces these conventional units with simple logic circuits and is particularly suitable for fault-tolerant applications. Spintronic devices, such as Magnetic Tunnel Junctions (MTJs), are capable of replacing CMOS in memory and logic circuits. In this work, we propose an energy-efficient use of MTJs, which exhibit probabilistic switching behavior, as Stochastic Number Generators (SNGs), which forms the basis of our NN implementation in the SC domain. Further, error resilient target applications of NNs allow us to introduce Approximate Computing, a framework wherein accuracy of computations is traded-off for substantial reductions in power consumption. We propose approximating the synaptic weights in our MTJ-based NN implementation, in ways brought about by properties of our MTJ-SNG, to achieve energy-efficiency. We design an algorithm that can perform such approximations within a given error tolerance in a single-layer NN in an optimal way owing to the convexity of the problem formulation. We then use this algorithm and develop a heuristic approach for approximating multi-layer NNs. To give a perspective of the effectiveness of our approach, a 43% reduction in power consumption was obtained with less than 1% accuracy loss on a standard classification problem, with 26% being brought about by the proposed algorithm.","PeriodicalId":385714,"journal":{"name":"2017 IEEE/ACM International Symposium on Low Power Electronics and Design (ISLPED)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2017-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116047467","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 6

Signal strength-aware adaptive offloading for energy efficient mobile devices 高效节能移动设备的信号强度感知自适应卸载

2017 IEEE/ACM International Symposium on Low Power Electronics and Design (ISLPED) Pub Date : 2017-07-01 DOI: 10.1109/ISLPED.2017.8009182

Young Geun Kim, S. Chung

引用次数: 7

Placement mitigation techniques for power grid electromigration 电网电迁移的安置缓解技术

2017 IEEE/ACM International Symposium on Low Power Electronics and Design (ISLPED) Pub Date : 2017-07-01 DOI: 10.1109/ISLPED.2017.8009178

Wei Ye, Yibo Lin, Xiaoqing Xu, Wuxi Li, Yiwei Fu, Yongsheng Sun, Canhui Zhan, D. Pan

引用次数: 5

Secure Human-Internet using dynamic Human Body Communication 使用动态人体通信安全的人类互联网

2017 IEEE/ACM International Symposium on Low Power Electronics and Design (ISLPED) Pub Date : 2017-07-01 DOI: 10.1109/ISLPED.2017.8009190

Shovan Maity, D. Das, X. Jiang, Shreyas Sen

{"title":"Secure Human-Internet using dynamic Human Body Communication","authors":"Shovan Maity, D. Das, X. Jiang, Shreyas Sen","doi":"10.1109/ISLPED.2017.8009190","DOIUrl":"https://doi.org/10.1109/ISLPED.2017.8009190","url":null,"abstract":"Continuous miniaturization and cost reduction of unit computing has led to the prolific growth of smart wearable devices. These devices, present on and around the human body, form a complex network known as the Human-Intranet. The Human-Intranet is typically connected through Wireless Body Area Network (WBAN). However, Human Body Communication (HBC) has recently emerged as an energy-efficient and secure alternative that uses the human body as the communication medium. Human-human, human-machine interaction creates dynamic HBC channels, which allow these Human-Intranets to interact with each other forming a Human-Internet. In this paper, we present the concept and demonstration of Secure Human-Internet using dynamic HBC. We highlight important applications of Human-Internet and discuss the architecture of a wearable Human-Internet device capable of communicating through inter-body dynamic HBC. A custom-built hardware prototype is used to demonstrate for the first time information exchange (e.g. business card) during handshaking. Dynamic signal transfer characteristics during inter-body communication through handshake between two individuals wearing such devices are measured and analyzed. The effects of data transmission rate, handshake posture on the HBC based inter-body communication is explored to demonstrate its effectiveness and limitations under varying realistic scenarios. The specific COTS based HBC implementation shows > 8× better energy efficiency compared to the Bluetooth implementation.","PeriodicalId":385714,"journal":{"name":"2017 IEEE/ACM International Symposium on Low Power Electronics and Design (ISLPED)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2017-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128962161","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 25

Transistor-level monolithic 3D standard cell layout optimization for full-chip static power integrity 晶体管级单片3D标准单元布局优化，实现全芯片静态电源完整性

2017 IEEE/ACM International Symposium on Low Power Electronics and Design (ISLPED) Pub Date : 2017-07-01 DOI: 10.1109/ISLPED.2017.8009189

B. W. Ku, Taigon Song, A. Nieuwoudt, S. Lim

引用次数: 7

An energy-efficient and high-throughput bitwise CNN on sneak-path-free digital ReRAM crossbar 一种基于无隐路径数字ReRAM交叉棒的高能效高吞吐量位CNN

2017 IEEE/ACM International Symposium on Low Power Electronics and Design (ISLPED) Pub Date : 2017-07-01 DOI: 10.1109/ISLPED.2017.8009177

Leibin Ni, Zichuan Liu, Wenhao Song, J. Yang, Hao Yu, Kanwen Wang, Yuangang Wang

引用次数: 22

Gabor filter assisted energy efficient fast learning Convolutional Neural Networks Gabor滤波器辅助高效快速学习卷积神经网络

2017 IEEE/ACM International Symposium on Low Power Electronics and Design (ISLPED) Pub Date : 2017-05-12 DOI: 10.1109/ISLPED.2017.8009202

Syed Shakib Sarwar, P. Panda, K. Roy

{"title":"Gabor filter assisted energy efficient fast learning Convolutional Neural Networks","authors":"Syed Shakib Sarwar, P. Panda, K. Roy","doi":"10.1109/ISLPED.2017.8009202","DOIUrl":"https://doi.org/10.1109/ISLPED.2017.8009202","url":null,"abstract":"Convolutional Neural Networks (CNN) are being increasingly used in computer vision for a wide range of classification and recognition problems. However, training these large networks demands high computational time and energy requirements; hence, their energy-efficient implementation is of great interest. In this work, we reduce the training complexity of CNNs by replacing certain weight kernels of a CNN with Gabor filters. The convolutional layers use the Gabor filters as fixed weight kernels, which extracts intrinsic features, with regular trainable weight kernels. This combination creates a balanced system that gives better training performance in terms of energy and time, compared to the standalone CNN (without any Gabor kernels), in exchange for tolerable accuracy degradation. We show that the accuracy degradation can be mitigated by partially training the Gabor kernels, for a small fraction of the total training cycles. We evaluated the proposed approach on 4 benchmark applications. Simple tasks like face detection and character recognition (MNIST and TiCH), were implemented using LeNet architecture. While a more complex task of objet recognition (CIFAR10) was implemented on a state-of-the-art deep CNN (Network in Network) architecture. The proposed approach yields 1.31–1.53× improvement in training energy in comparison to conventional CNN implementation. We also obtain improvement up to 1.4× in training time, up to 2.23× in storage requirements, and up to 2.2× in memory access energy. The accuracy degradation suffered by the approximate implementations is within 0– 3% of the baseline.","PeriodicalId":385714,"journal":{"name":"2017 IEEE/ACM International Symposium on Low Power Electronics and Design (ISLPED)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2017-05-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114963562","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 92