Energy Efficiency of a Multi-Core Processor by Tag Reduction

We consider the energy saving problem for caches on a multi-core processor.In the previous research on low power processors,there are various methods to reduce power dissipation.Tag reduction is one of them.This paper extends the tag reduction technique on a single-core processor to a multi-core processor and investigates the potential of energy saving for multi-core processors.We formulate our approach as an equivalent problem which is to find an assignment of the whole instruction pages in the physical memory to a set of cores such that the tag-reduction conflicts for each core can be mostly avoided or reduced.We then propose three algorithms using different heuristics for this assignment problem.We provide convincing experimental results by collecting experimental data from a real operating system instead of the traditional way using a processor simulator that cannot simulate operating system functions and the full memory hierarchy.Experimental results show that our proposed algorithms can save total energy up to 83.93% on an 8-core processor and 76.16% on a 4-core processor in average compared to the one that the tag-reduction is not used for.They also significantly outperform the tag reduction based algorithm on a single-core processor.     

Subsurface structures in initial stage of FeSi2 growth studied by high-resolution Rutherford backscattering spectroscopy

The initial stage of iron silicide formation is investigated by high-resolution Rutherford backscattering spectroscopy. During the Fe deposition on Si(001) at 470 °C, the formation of FeSi₂ is confirmed by the surface peak analysis. Initially, FeSi₂ grows epitaxially so that one of the major crystallographic axes is parallel to the <111> axis of the Si substrate. With increasing Fe deposition, the deviation between the major crystallographic axis of the silicide region and Si<111> increases although the electron diffraction pattern is independent of the amount of Fe deposition. Therefore, the subsurface crystallographic structure of iron silicide is transformed from a cubic-like to a low-symmetry structure.     

A nonlinear model to rank association rules based on semantic similarity and genetic network programing

Many methods have been studied for mining association rules efficiently. However, because these methods usually generate a large number of rules, it is still a heavy burden for the users to find the most interesting ones. In this paper, we propose a novel method for finding what the user is interested in by assigning several keywords, like searching documents on the Web using search engines. By considering both the semantic similarity between the rules and keywords, and the statistical information like support, confidence, chi-squared value, etc. we could rank the rules by a new method named RuleRank, where evolutionary methods are applied to find the optimal ranking model. Experiments show that our approach is effective for the users to find what they want. Copyright © 2009 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.     

Notch effect on creep–fatigue life for Sn–3.5Ag solder

This paper studies the creep–fatigue crack initiation and failure lives of Sn–3.5Ag solder notched specimens focused on the multiaxial strain at the notch root. Push–pull creep–fatigue tests were performed using three circumferential notched specimens using four kinds of creep–fatigue strain waveforms. Multiaxial strains at the notched section were calculated by finite element (FE) analysis under four kinds of creep–fatigue loading. Creep–fatigue damage laws were applied for evaluating the crack initiation and failure lives using the multiaxial strains obtained by the FE analysis. von Mises equivalent strain at the notch root estimated the crack initiation lives with a large scatter as well as the failure lives. Instead, the mean value of von Mises equivalent strain over the cross section of the notch root estimated the crack initiation and failure lives with a small scatter.     

A new detection method for the primary cosmic heavy nuclei by the plastic track detector CR-39

A new multiple scanning method is developed and applied to detect cosmic heavy nuclei. The plastic track detectors CR-39 exposed to cosmic heavy nuclei are over-etched in 8.0N NaOH for 94 h at 80°C. We found that the etched cones can easily be scanned on the plastic plates without using an optical microscope. The trajectories of heavy nuclei can be reconstructed with the naked eyes by superposing four or more over-etched plates. The charge detection threshold is obtained by this method, and appears to be 10. The charge resolution is estimated to be 1.0 charge units for iron nuclei. By adopting this method, the scanning time is much shortened and CR-39 plastic chamber size can easily be scaled up. The method is very useful for detecting particles with low flux intensity, like for searching magnetic monopoles or for measuring the high energy spectra of cosmic heavy nuclei.     

Prism-pair interferometry by homodyne interferometers with a common light source for high-accuracy measurement of the absolute refractive index of glasses

A prism-pair interferometer comprising two homodyne interferometers with a common light source was developed for high-precision measurements of the refractive index of optical glasses with an uncertainty of the order of 10(-6). The two interferometers measure changes in the optical path length in the glass sample and in air, respectively. Uncertainties in the absolute wavelength of the common light source are cancelled out by calculating a ratio between the results from the interferometers. Uncertainties in phase measurement are suppressed by a quadrature detection system. The combined standard uncertainty of the developed system is evaluated as 1.1×10(-6).     

First-principles study of structure,vacancy formation,and strength of bcc Fe/V<Subscript>4</Subscript>C<Subscript>3</Subscript> interface

Voids are representative of the damage process in both creep and ductile fractures. Although the matrix/precipitate interface has been considered the preferential nucleation site for voids, the relationship between the atomic structure of this interface and the nucleation mechanism of a void has never been sufficiently investigated. In this study, the bcc Fe/V₄C₃ interface is selected as a model interface between a matrix and precipitate. The vacancy formation energy and intrinsic mechanical strength at this interface are investigated using a first-principles calculation because they should be related with the nucleation of creep and ductile voids, respectively. Within the considered interface, the Fe vacancy is found to be dominant. When the Baker–Nutting orientation relationship is satisfied at the interface, the calculated intrinsic mechanical strength of the interface is 23.8 GPa. However, when the geometric coherence at the interface is low as compared to that of the Baker–Nutting orientation relationship, it is found that the interfacial mechanical strength is significantly weakened. At each interface, it is found that the back-bond of the interface determined the interfacial strength because of the strongly bonded Fe–C on the interface. The nucleation mechanism of a void at the matrix/precipitate interface is discussed based on the present findings. It is suggested that local decohesion at the matrix/precipitate interface should be the origin of the nucleation of a ductile void.     

面向动态场景去模糊的对偶学习生成对抗网络

针对动态场景下的图像去模糊问题,提出一种对偶学习生成对抗网络(dual learning generative adversarial network, DLGAN),该网络可以在对偶学习的训练模式下使用非成对的模糊图像和清晰图像进行图像去模糊计算,不再要求训练图像集合必须由模糊图像与其对应的清晰图像成对组合而成. DLGAN利用去模糊任务与重模糊任务之间的对偶性建立反馈信号,并使用这个信号约束去模糊任务和重模糊任务从两个不同的方向互相学习和更新,直到收敛.实验结果表明,在结构相似度和可视化评估方面, DLGAN与9种使用成对数据集训练的图像去模糊方法相比具有更好的性能.