Yingjian Ling, Kan Zhong, Yunsong Wu, Duo Liu, Jinting Ren, Renping Liu, Moming Duan, Weichen Liu, Liang Liang
{"title":"TaiJiNet: Towards Partial Binarized Convolutional Neural Network for Embedded Systems","authors":"Yingjian Ling, Kan Zhong, Yunsong Wu, Duo Liu, Jinting Ren, Renping Liu, Moming Duan, Weichen Liu, Liang Liang","doi":"10.1109/ISVLSI.2018.00034","DOIUrl":"https://doi.org/10.1109/ISVLSI.2018.00034","url":null,"abstract":"We have witnessed the tremendous success of deep neural networks. However, this success comes with the considerable computation and storage costs which make it difficult to deploy these networks directly on resource-constrained embedded systems. To address this problem, we propose TaiJiNet, a binary-network-based framework that combines binary convolutions and pointwise convolutions, to reduce the computation and storage overhead while maintaining a comparable accuracy. Furthermore, in order to provide TaiJiNet with more flexibility, we introduce a strategy called partial binarized convolution to efficiently balance network performance and accuracy. We evaluate TaiJiNet with the CIFAR-10 and ImageNet datasets. The experimental results show that with the proposed TaiJiNet framework, the binary version of AlexNet can achieve 26x compression rate with a negligible 0.8% accuracy drop when compared with the full-precision AlexNet.","PeriodicalId":114330,"journal":{"name":"2018 IEEE Computer Society Annual Symposium on VLSI (ISVLSI)","volume":"20 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133091534","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}
{"title":"Functional Obfuscation of DSP Cores Using Robust Logic Locking and Encryption","authors":"A. Sengupta, S. Mohanty","doi":"10.1109/ISVLSI.2018.00133","DOIUrl":"https://doi.org/10.1109/ISVLSI.2018.00133","url":null,"abstract":"Obfuscation plays a key role in thwarting attacks launched through reverse engineering process. This work presents a new obfuscation process for DSP cores using improved logic locking and encryption that incurs minimum design overhead and achieves reduced design cost compared to state of the art approaches. The proposed approach integrates particle swarm optimization driven design space exploration system (PSO-DSE) for obtaining reduced design cost of obfuscated DSP designs. Enhanced security of locking is provided through locking blocks that are capable of locking each output data bit of functional resources with 8 key bits. The presented approach includes countermeasures against key sensitization attacks, SAT attacks and removal attacks. Results indicate that the proposed approach has been capable of achieving enhanced obfuscation security by at least 4.29 e+9 times and a design cost reduction ~ 6.5 % compared to a recent approach.","PeriodicalId":114330,"journal":{"name":"2018 IEEE Computer Society Annual Symposium on VLSI (ISVLSI)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130919517","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}
Manoj Kumar Jyv, Ayass Kant Swain, K. SudeendraKumar, S. Sahoo, K. Mahapatra
{"title":"Run Time Mitigation of Performance Degradation Hardware Trojan Attacks in Network on Chip","authors":"Manoj Kumar Jyv, Ayass Kant Swain, K. SudeendraKumar, S. Sahoo, K. Mahapatra","doi":"10.1109/ISVLSI.2018.00139","DOIUrl":"https://doi.org/10.1109/ISVLSI.2018.00139","url":null,"abstract":"Globalization of semiconductor design and manufacturing has led to several hardware security issues. The problem of Hardware Trojans (HT) is one such security issue discussed widely in industry and academia. Adversary design engineer can insert the HT to leak confidential data, cause a denial of service attack or any other intention specific to the design. HT in cryptographic modules and processors are widely discussed. HT in Multi-Processor System on Chips (MPSoC) are also catastrophic, as most of the military applications use MPSoCs. Network on Chips (NoC) are standard communication infrastructure in modern day MPSoC. In this paper, we present a novel hardware Trojan which is capable of inducing performance degradation and denial of service attacks in a NoC. The presence of the Hardware Trojan in a NoC can compromise the crucial details of packets communicated through NoC. The proposed Trojan is triggered by a particular complex bit pattern from input messages and tries to mislead the packets away from the destined addresses. A mitigation method based on bit shuffling mechanism inside the router with a key directly extracted from input message is proposed to limit the adverse effects of the Trojan. The performance of a 4x4 NoC is evaluated under uniform traffic with the proposed Trojan and mitigation method. Simulation results show that the proposed mitigation scheme is useful in limiting the malicious effect of hardware Trojan.","PeriodicalId":114330,"journal":{"name":"2018 IEEE Computer Society Annual Symposium on VLSI (ISVLSI)","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131911457","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}
K. SudeendraKumar, S. Seth, S. Sahoo, Abhishek Mahapatra, Ayass Kant Swain, K. Mahapatra
{"title":"PUF-Based Secure Test Wrapper for SoC Testing","authors":"K. SudeendraKumar, S. Seth, S. Sahoo, Abhishek Mahapatra, Ayass Kant Swain, K. Mahapatra","doi":"10.1109/ISVLSI.2018.00127","DOIUrl":"https://doi.org/10.1109/ISVLSI.2018.00127","url":null,"abstract":"The increased testability and observability due to test structures make chips vulnerable to side channel attacks. The intention of side channel attack are leaking secret keys used in cryptographic cores and getting access to trade related sensitive information stored in chips. Several countermeasures against test based side-channel attacks are available in research literature. One such countermeasure scheme is password based access protection to IEEE 1500 test wrapper, such that only an authentic user with valid password is allowed to access the test structures. IEEE 1500 is a core test standard for enabling the streamlined test integration and test reuse. The trust model of existing schemes assume outsourced assembly and test (OSAT) centre are completely trusted and design house will share secret keys to unlock the IEEE 1500 wrapper during testing. In this paper, we propose a Physical Unclonable Function (PUF) based technique incorporating challenge-response to support comprehensive test security in which there is no need for design house to share secret keys with untrusted OSAT centre to unlock the scan chains. The proposed scheme comes at the cost of reasonable area and performance overhead.","PeriodicalId":114330,"journal":{"name":"2018 IEEE Computer Society Annual Symposium on VLSI (ISVLSI)","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116786359","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}
{"title":"Robust Timing Attack Countermeasure on Virtual Hardware","authors":"Kai Yang, Jungmin Park, M. Tehranipoor, S. Bhunia","doi":"10.1109/ISVLSI.2018.00036","DOIUrl":"https://doi.org/10.1109/ISVLSI.2018.00036","url":null,"abstract":"Field programmable gate arrays (FPGAs) are being increasingly used in Internet of Things (IoT) applications, as they usually provide lower power, lower latency and higher performance compared with their processor counterparts. However, security has emerged as a critical concern for FPGA-based systems as they are vulnerable to different form of physical attacks, such as side-channel attacks (SCAs). Existing protection methods, which primarily rely on bitstream encryption, are computationally expensive and more importantly, cannot protect against run-time attacks. Hardware virtualization, where instead of traditional direct mapping to FPGA, an application is mapped upon an application-specific virtual layer, called overlay, has been well-studied in past decades for productivity benefit, while its security implication has not been investigated at all. In this paper, for the first time to our knowledge, we present a novel usage of virtualization that limits damage from timing attacks and improves performance for RSA decryption by employing unique reconfigurable hardware architectures on FPGAs. Specific masking methods are implemented onto this architecture, and extensive security and performance analysis are done that demonstrates significant side-channel attack resistance under performance constraint.","PeriodicalId":114330,"journal":{"name":"2018 IEEE Computer Society Annual Symposium on VLSI (ISVLSI)","volume":"63 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124694789","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}
{"title":"Achieving Low Power Classification with Classifier Ensemble","authors":"Fanglei Hu, Min Zhang, Hailong Jiao","doi":"10.1109/ISVLSI.2018.00014","DOIUrl":"https://doi.org/10.1109/ISVLSI.2018.00014","url":null,"abstract":"Machine learning algorithms, such as Support Vector Machine (SVM) and Artificial Neural Networks, have been widely applied in many aspects of daily life. Low power/energy integrated circuit implementation of machine learning algorithms with high accuracy is however still a great challenge, which is critical for portable and wearable devices. In this paper, classifier ensemble is investigated for achieving low power classification. The classifier ensemble algorithm Adaboost is employed to combine multiple SVM classifiers with linear kernel to achieve high classification accuracy while reducing the hardware complexity. The proposed classifier ensemble is evaluated on the MNIST dataset by using a 45-nm CMOS technology. Compared to the traditional SVM classifier with second-order polynomial kernel, the proposed classifier ensemble achieves up to 45.7%, 20.3%, and 20.3% savings in total energy consumption, leakage power consumption, and area, respectively, while providing similar classification accuracy.","PeriodicalId":114330,"journal":{"name":"2018 IEEE Computer Society Annual Symposium on VLSI (ISVLSI)","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129819348","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}
{"title":"91dB Dynamic Range 9.5nW Low Pass Filter for Bio-Medical Applications","authors":"M. Reddy, Sreenivasulu Polineni, Laxminidhi Tonse","doi":"10.1109/ISVLSI.2018.00088","DOIUrl":"https://doi.org/10.1109/ISVLSI.2018.00088","url":null,"abstract":"This paper presents a second order, fully differential, low pass filter. The filter has a tunable bandwidth in the range 4 Hz to 100 Hz and offers a dynamic range of 91 dB. The filter is based on the source-follower biquad operating in the sub-threshold region. The main idea is to exploit the strengths of sub-threshold source follower circuit, like low noise, low output impedance, high linearity and low power. The filter design has been validated in UMC 0.18 um CMOS process. The filter consumes only 9.5 nW of power at 1.8 V supply, making it suitable for bio-medical applications. In terms of noise and dynamic range the reported filter is better than previous works found from the literature.","PeriodicalId":114330,"journal":{"name":"2018 IEEE Computer Society Annual Symposium on VLSI (ISVLSI)","volume":"962 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127038181","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}
Ankit Rehani, Sujay Deb, P. Bahubalindruni, Bhavin Odedara, S. Bojja
{"title":"A High-Efficient Current-Mode PWM DC-DC Buck Converter Using Dynamic Frequency Scaling","authors":"Ankit Rehani, Sujay Deb, P. Bahubalindruni, Bhavin Odedara, S. Bojja","doi":"10.1109/ISVLSI.2018.00090","DOIUrl":"https://doi.org/10.1109/ISVLSI.2018.00090","url":null,"abstract":"Improving efficiency of power management solutions for battery-operated devices is an important issue. This paper presents design and implementation of a novel high-efficiency current-mode pulse width modulated (PWM) buck converter with dynamic frequency scaling. This circuit is capable of ensuring high efficiency under different load conditions, by dynamically changing the frequency of operation with respect to the load current with the help of a novel frequency decision circuit. In addition, soft start operation protects the circuit from large in-rush current during start up of the converter. The proposed circuit is implemented in TSMC 16nm FinFET CMOS technology. Simulation outcome has shown an output voltage of 0.8V, when the input voltage is ranging from 2.3 to 3.6V, with a maximum conversion efficiency of 92.25% (heavy loads) and 92.11% (light loads), making this circuit quite useful in power management ICs.","PeriodicalId":114330,"journal":{"name":"2018 IEEE Computer Society Annual Symposium on VLSI (ISVLSI)","volume":"222 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127157153","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}
{"title":"Enhancing Lifetime of PCM-Based Main Memory with Efficient Recovery of Stuck-at Faults","authors":"Marjan Asadinia, C. Bobda","doi":"10.1109/ISVLSI.2018.00072","DOIUrl":"https://doi.org/10.1109/ISVLSI.2018.00072","url":null,"abstract":"Among several nonvolatile memory (NVM) candidates, PCM is selected as an attractive replacement to DRAM and comes with both challenges and opportunities. It has beautiful characteristics like non-volatility, better scalability, and lower leakage power. However, limited write endurance is the main burden toward its adoption in practice. It means that after a certain number of writes, some memory cells permanently stuck at either '0' or '1'. To solve this problem and provide strong fault tolerant system, some recovery techniques with minimal storage overhead are required. In this work, we propose a recovery mechanism that relies on static partitioning of a data block into some small number of groups and spreading out faults across the groups uniformly. We then exploit inversion mechanism along with shifting mechanism to continue the use of the failed cell with stuck-at value. So, our proposed method can recover multi bit stuck at faults per partition. Compare to the existing mechanisms, our experimental results for multi-threaded workloads reveal considerable improvement in lifetime and the number of recoverable failures per data block","PeriodicalId":114330,"journal":{"name":"2018 IEEE Computer Society Annual Symposium on VLSI (ISVLSI)","volume":"41 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127342178","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}
{"title":"Floorplanning in Graphene Nanoribbon (GNR) Based Circuits","authors":"Subrata Das, D. K. Das","doi":"10.1109/ISVLSI.2018.00061","DOIUrl":"https://doi.org/10.1109/ISVLSI.2018.00061","url":null,"abstract":"In this paper, we propose a technique for floorplanning in case of graphene nanoribbon (GNR) based circuits. Graphene nanoribbon based devices and interconnects are now found to be better alternatives over traditional CMOS based devices and interconnects. Logic blocks of GNR can be assumed to be hexagonal in shape. Due to special geometric structure in GNR, the interconnects can be bent only in 0°, 60° and 120° angles. Hence routing grids are aligned to these angles only. The concept of floorplanning in traditional VLSI design is extended first time for GNR based circuits in this paper.","PeriodicalId":114330,"journal":{"name":"2018 IEEE Computer Society Annual Symposium on VLSI (ISVLSI)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129932752","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}