3 December 2019 in IEEE Journal of Selected Topics in Quantum Electronics
Neural Networks (NNs) have become the mainstream technology in the artificial intelligence (AI) renaissance over the past decade. Among different types of neural networks, convolutional neural networks (CNNs) have been widely adopted as they have achieved leading results in many fields such as computer vision and speech recognition. This success in part is due to the widespread availability of capable underlying hardware platforms. In parallel, hardware specialization can expose us to novel architectural solutions, which can outperform general purpose computers for the tasks at hand. Although different applications demand for different performance measures, they all share speed and energy efficiency as high priorities. Meanwhile, photonics processing has seen a resurgence due to its inherited high speed and low power nature. Here, we investigate the potential of using photonics in CNNs by proposing a CNN accelerator design based on Winograd filtering algorithm. Our evaluation results show that while a photonic accelerator can compete with current state-of-the-art electronic platforms in terms of both speed and power, it has the potential to improve the energy efficiency by up to three orders of magnitude.
1 July 2014 in Journal of Advanced Research
The problem of hardware Trojans is becoming more serious especially with the widespread of fabless design houses and design reuse. Hardware Trojans can be embedded on chip during manufacturing or in third party intellectual property cores (IPs) during the design process. Recent research is performed to detect Trojans embedded at manufacturing time by comparing the suspected chip with a golden chip that is fully trusted. However, Trojan detection in third party IP cores is more challenging than other logic modules especially that there is no golden chip. This paper proposes a new methodology to detect/prevent hardware Trojans in third party IP cores. The method works by gradually building trust in suspected IP cores by comparing the outputs of different untrusted implementations of the same IP core. Simulation results show that our method achieves higher probability of Trojan detection over a naive implementation of simple voting on the output of different IP cores. In addition, experimental results show that the proposed method requires less hardware overhead when compared with a simple voting technique achieving the same degree of security.
15 January 2020 in IEEE Photonics Journal
A content addressable memory (CAM) is a type of memory that implements a parallel search engine at its core. A CAM takes as an input a value and outputs the address where this value is stored in case of a match. CAMs are used in a wide range of applications including networking, cashing, neuromorphic associative memories, multimedia, and data analytics. Here, we introduce a novel opto-electronic CAM (OE-CAM) utilizing the integrated silicon photonic platform. In our approach, we explore the performance of an experimental OE-CAM and verify the efficiency of the device at 25Gbit/s while maintaining the bit integrity under noise conditions. We show that OE-CAM enables a) two orders of magnitude faster search functionality resulting in b) a five orders of magnitude lower power-delay-product compared to CAMs implementations based on other emerging technologies. This remarkable performance potential is achieved by utilizing i) a high parallelism of wavelength-division- multiplexing in the optical domain, combined with ii) 10's of GHz-fast opto-electronic components, packaged in iii) integrated photonics for 10-100's ps-short communication delays. We further verify the upper optical input power limit of this OE-CAM to be given by parasitic nonlinearities inside the silicon waveguides, and the minimal detectable optical power at the back-end photoreceiver's responsivity given channel noise. Such energy-efficient and short-delay OE-CAMs could become a key component of functional photonic-augmented ASICS, co-processors, or smart sensors.
13 October 2022 in ACM Journal on Emerging Technologies in Computing Systems
The acceleration of Deep Neural Networks (DNNs) has attracted much attention in research. Many critical real-time applications benefit from DNN accelerators but are limited by their compute-intensive nature. This work introduces an accelerator for Convolutional Neural Network (CNN) , based on a hybrid optoelectronic computing architecture and residue number system (RNS) . The RNS reduces the optical critical path and lowers the power requirements. In addition, the wavelength division multiplexing (WDM) allows high-speed operation at the system level by enabling high-level parallelism. The proposed RNS compute modules use one-hot encoding, and thus enable fast switching between the electrical and optical domains. We propose a new architecture that combines residue electrical adders and optical multipliers as the matrix-vector multiplication unit. Moreover, we enhance the implementation of different CNN computational kernels using WDM-enabled RNS based integrated photonics. The area and power efficiency of the proposed accelerator are 0.39 TOPS/s/mm 2 and 3.22 TOPS/s/W, respectively. In terms of computation capability, the proposed chip is 12.7× and 4.02× better than other optical implementation and memristor implementation, respectively. Our experimental evaluation using DNN benchmarks illustrates that our architecture can perform on average more than 72 times faster than GPU under the same power budget.