Sharif Digital Repository

Topology is an important network attribute that greatly affects the power, performance, cost, and design time/effort of NoCs. In this paper, we propose a novel NoC architecture that can exploit the benefits of both application-specific and regular NoC topologies. To this end, a subset of NoC links bypass the router pipeline stages and directly connect remotely located nodes. This results in an NoC which holds both fixed connections between adjacent nodes and long connections virtually connecting non-adjacent nodes. These shortcut paths are constructed at run-time by employing a simple and fast mechanism composed of two processes: on-chip traffic monitoring and path reconfiguration. The... 

We propose a novel multi-parallel layer DCN architecture based on low radix FOS. Results show cost and power consumption reduction of 4 parallel-layer architecture is 60% and 26.7% with respect to the FatTree. © 2019 The Institute of Electronics, Information and Communication Engineers (IEICE)  

A fault tolerant routing algorithm is presented in this paper. proposed routing algorithm is based on making a redundant of each packet as well as sending the redundant packets the paths with low traffic loads. Since two copies of each reach the destination node, the erroneous packets are and replaced with the correct ones. To effectively use paths with lower traffic loads, the redundant packets are according to YX routing while the original packets are according to Duato's routing algorithm. Minimizing the of sent redundant packets and exploiting different paths sending the original and redundant packets enable the algorithm to improve the reliability of NoCs with power and performance... 

Network design problem has been, and is, an important problem in transportation. Following an earlier effort in designing a meta-heuristic search technique by an ant system, this paper attempts to hybridize this concept with other meta-heuristic concepts such as genetic algorithm, simulated annealing, and tabu search. Seven hybrids have been devised and tested on the network of Sioux Falls. It has been observed that the hybrids are more effective to solve the network design problem than the base ant system. Application of the hybrid containing all four concepts on a real network of a city with over 2 million population has also proved to be more effective than the base network, in the sense... 

Modern SoC architectures use NoCs for high-speed inter-IP communication. For NoC architectures, high-performance efficient routing algorithms with low power consumption are essential for real-time applications. NoCs with mesh and torus interconnection topologies are now popular due to their simple structures. A torus NoC is very similar to the mesh NoC, but has rather smaller diameter. For a routing algorithm to be deadlock-free in a torus, at least two virtual channels per physical channel must be used to avoid cyclic channel dependencies due to the warp-around links; however, in a mesh network deadlock freedom can be insured using only one virtual channel. The employed number of virtual... 

Cut-set bounds are not, in general, tight for all classes of network communication problems. In this paper, we introduce a new technique for proving converses for the problem of transmission of correlated sources in networks, which results in bounds that are tighter than the corresponding cut-set bounds. We also define the concept of 'uncertainty region' which might be of independent interest. We provide a full characterization of this region for the case of two correlated random variables. The bounding technique works as follows: on one hand, we show that if the communication problem is solvable, the uncertainty of certain random variables in the network with respect to imaginary parties... 

Cut-set bounds on achievable rates for network communication protocols are not in general tight. In this paper we introduce a new technique for proving converses for the problem of transmission of correlated sources in networks, that results in bounds that are tighter than the corresponding cut-set bounds. We also define the concept of "uncertainty region" which might be of independent interest. We provide a full characterization of this region for the case of two correlated random variables. The bounding technique works as follows: on one hand we show that if the communication problem is solvable, the uncertainty of certain random variables in the network with respect to imaginary parties... 

We consider a content delivery network where multiple servers are connected to multiple cache-enabled clients. Clients request their corresponding contents from the servers and servers collaboratively transmit packets to fulfill all the requests. It is assumed that some contents are stored in the caches in off-peak time of the network without knowing the actual requests, the so called cache content placement phase. The goal is to minimize the worst case delay in the content delivery phase. Considering a random linear network, we propose a coding strategy which exploits servers' multiplexing gains as well as caches' global and local coding gains. The main idea in our coding scheme is to... 

Core specialization is a promising solution to the dark silicon challenge. This approach trades off the cheaper silicon area with energy-efficiency by integrating a selection of many diverse application-specific cores into a single billion-transistor multicore chip. Each application then activates the subset of cores that best matches its processing requirements. These cores act as a customized application-specific CMP for the application. Such an arrangement of cores requires some special on-chip inter-core communication treatment to efficiently connect active cores. In this paper, we propose a reconfigurable network-on-chip that leverages the routers of the dark portion of the chip to... 

Visual attributes are great means of describing images or scenes, in a way both humans and computers understand. In order to establish a correspondence between images and to be able to compare the strength of each property between images, relative attributes were introduced. However, since their introduction, hand-crafted and engineered features were used to learn increasingly complex models for the problem of relative attributes. This limits the applicability of those methods for more realistic cases. We introduce a deep neural network architecture for the task of relative attribute prediction. A convolutional neural network (ConvNet) is adopted to learn the features by including an... 

Due to high aggregate idle time of Networks-on-Chip (NoCs) routers in practical applications, power-gating techniques have been proposed to combat the ever-increasing ratio of static power. Nevertheless, the sporadic packet arrivals compromise the effectiveness of power-gating by incurring significant latency and energy overhead. In this paper, we propose a Scalable Pivot-based On/Off Gating Engine (SPONGE) which efficiently manages power-gating decisions and routing mechanism by adaptively selecting a small set of powered-on columns of routers and keeping the others in power-gated state. To this end, a router architecture augmented with a novel routing algorithm is proposed in which a... 

Because of their various desirable properties, OTIS networks are subject to active research in the field of parallel processing. In this paper, the OTIS network architecture is presented, emphasizing two important varieties, the OTIS-mesh and the OTIS-hypercube. Next, a randomized sorting algorithm presented in the literature for OTIS-mesh structures will be reviewed. Finally, a discussion of sorting on OTIS-hypercubes will be offered, and a suggested algorithm for sorting on these networks, with an MPI implementation will be introduced. © 2008 Springer-Verlag  

The necklace hypercube has recently been introduced as an attractive alternative to the well-known hypercube. Previous research on this network topology has mainly focused on topological properties, VLSI and algorithmic aspects of this network Several analytical models have been proposed in the literature for different interconnection networks, as the most cost-effective tools to evaluate the performance merits of such systems. This paper proposes an analytical performance model to predict message latency in wormhole-switched necklace hypercube interconnection networks with fully adaptive routing. The analysis focuses on a fully adaptive routing algorithm which has been shown to be the most... 

The purpose of this investigation is to develop a model for prediction of topologically closed-packed (TCP) phases formation in superalloys. In this study, artificial neural networks (ANN), using several different network architectures, were used to investigate the complex relationships between TCP phases and chemical composition of superalloys. In order to develop an optimum ANN structure, more than 200 experimental data were used to train and test the neural network. The results of this investigation shows that a multilayer perceptron (MLP) form of the neural networks with one hidden layer and 10 nodes in the hidden layer has the lowest mean absolute error (MAE) and can be accurately used... 

In this paper, we propose a new architecture for nanophotonic Networks on Chip (NoC), named 2D-HERT, which consists of optical data and control planes. The proposed data plane is built upon a new topology and all-optical switches that passively route optical data streams based on their wavelengths. Utilizing wavelength routing method, the proposed deterministic routing algorithm, and Wavelength Division Multiplexing (WDM) technique, the proposed data plane eliminates the need for optical resource reservation at the intermediate nodes. For resolving end-point contention, we propose an all-optical request-grant arbitration architecture which reduces optical losses compared to the alternative... 

In this study, accurate estimation of nuclear power plant (NPP) parameters is done using the new and simple technique. The proposed technique using the genetic algorithm (GA) in combination with the Bayesian regularization (BR) and Levenberg- Marquardt (LM) learning algorithms identifies the appropriate architecture for estimation of the target parameters. In the first step, the input patterns features are selected using the features selection (FS) technique. In the second step, the appropriate number of hidden neurons and hidden layers are investigated to provide a more efficient initial population of the architectures. In the third step, the estimation of the target parameter is done using... 

This article presented an efficient deep learning technique to predict the compressive strength of high-performance concrete (HPC). This technique combined the convolutional neural network (CNN) and genetic algorithm (GA). Six CNN architectures were considered with different hyper-parameters. GA was employed to determine the optimum number of filters in each convolutional layer of the CNN architectures. The resulted CNN architectures were then compared to each other to find the best architecture in terms of accuracy and capability of generalization. It was shown that all of the proposed CNN models are capable of predicting the HPC compressive strength with high accuracy. Finally, the best of...