The design of today's System-on-Chip (SoC) architectures faces many challenges in respect to the involved complexity and heterogeneity. An early and systematic exploration of alternatives is mandatory to find a solution that meets all design requirements. Therefore, the experience of system architects has to be supplemented with efficient performance evaluation methods and tools that help in the broad exploration of the solution space. This article describes TAPES (Trace-based Architecture Performance Evaluation with SystemC), an approach that supports system designers in the performance evaluation of SoC architectures. The concept captures the functionality of the architecture in the form of traces for each resource. The trace primitives making up a trace are translated at simulation run-time into transactions and superposed on the system architecture. The method uses SystemC as modeling language, requires low modeling effort and yet provides accurate results within reasonable turnaround times. A concluding application example for the exploration of a network processor architecture demonstrates the effectiveness of the TAPES approach.

No abstract

No abstract

In signal processing, the zero-error probability (ZEP) criterion and related algorithm (MZEP) outperforms MSE-based algorithms and yields superior and stable convergence in impulsive noise environment. In this paper, the analysis of the relationship with MSE criterion proves that ZEP criterion has equivalent optimum solution of MSE criterion. Also this work reveals that the magnitude controlled input of MZEP algorithm plays the role in keeping the optimum solution undisturbed from impulsive noise.°First and Corresponding Author : Division of Electronic, Information and Commun. Eng., Kangwon National Univ., namyong@kangwon.ac.kr, 정회원 * Division of Electronic, Information and Commun. Eng., Kangwon National Univ.논문번호：KICS2015 -08-268, Received August 27, 2015; Revised October 20, 2015; Accepted October 20, 2015 Ⅰ. IntroductionBesides the harsh problems such as multipath propagation and severe fading in wireless network communication environment, impulsive noise from a variety of sources affects the links [1][2][3] . Many signal processing algorithms designed on the basis of MSE criterion may fail when impulsive noise is present [4] .As an alternative to the MSE, the zero-error probability (ZEP) criterion has been introduced in [5]. By maximization of ZEP (MZEP) and steepest descent method, the MZEP algorithm has been developed for communication systems with impulsive noise and channel distortions. For application for underwater communication channels, the nonlinear MZEP has been proposed to compensate for ISI without error propagation [6] .One drawback of MZEP in which weights are calculated based on block processing method is a heavy computational burden. In the work in [7] a method utilizing the current gradient in estimation of the next gradient has been proposed and shown that its computational complexity can be significantly reduced.

We investigated the resonance frequency distributions of carbon nanotube (CNT) oscillators with an intertube gap using molecular dynamics simulations. The resonance frequency distributions could be regressed by the use of Gaussian distributions. The maximum resonance frequencies linearly decreased with increasing initial velocity of the coretube. For the same initial velocity of the coretube, the maximum resonance frequencies were constant regardless of gap spacing changes, and the center positions of the Gaussian distributions were shifted by changing the outertube length whereas their standard deviations were not changed. Although their standard deviations were influenced by the initial velocity of the coretube, the normalized resonance frequencies could be regressed into single Gaussian distribution within some error. So considering the difficulty of exactly controlling of the initial velocity, the resonance frequency distribution predicted by the gap spacing can give important revolution for CNT-oscillators to be utilized as components of nanoelectromechanical system.