A high-level energy consumption model for heterogeneous data centers

Data centers consume anywhere between 1.7% and 2.2% of the United States’ power. A handful of studies focused on ways of predicting power consumption of computing platforms based on performance events counters. Most of existing power-consumption models retrieve performance counters from hardware, which offer accurate measurement of energy dissipation. Although these models were verified on several machines with specific CPU chips, it is difficult to deploy these models into data centers equipped by heterogeneous computing platforms. While models based on resource utilization via OS monitoring tools can be used in heterogeneous data centers, most of these models were linear model. In this paper, we analyze the accuracy of linear models with the SPECpower benchmark results, which is a widely adopted benchmark to evaluate the power and performance characteristics of servers. There are 392 published results until October 2012; these servers represent most servers in heterogeneous data centers. We use R-squared, RMSE (Root Mean Square Error) and average error to validate the accuracy of the linear model. The results show that not all servers fit the linear model very well. 6.5% of R-squared values are less than 0.95, which means linear regression does not fit the data very well. 12.5% of RMSE values are greater than 20, which means there is still big difference between modeled and real power consumption. We extend the linear model to high degree polynomial models. We found the cubic polynomial model can get better results than the linear model. We also apply the linear model and the cubic model to estimate real-time energy consumption on two different servers. The results show that linear model can get accurate prediction value when server energy consumption swing in a small range. The cubic model can get better results for servers with small and wide range.     

The costs and benefits of tonal centers for chord processing.

Harmonic priming studies have shown that a musical context, with its established tonal center, influences target chord processing. This study investigated costs and benefits of priming tonal centers for target processing by adding a baseline condition (sequences without a specific tonal center). Results confirmed harmonic priming, with faster processing for related than for unrelated and less related targets (tonic chord, out-of-key chord, subdominant chord). Comparing targets in baseline contexts with targets in sequences with well-established tonal centers revealed a benefit of processing for related targets but a cost of processing for unrelated and less related targets. Findings are discussed in terms of tonal knowledge activation and suggest that an activated tonal center gives rise to strong expectations for the tonic. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Leakage radiation microscopy of surface plasmons launched by a nanodiamond-based tip

Leakage-radiation microscopy of a thin gold film demonstrates the ability of an ensemble of fluorescent diamond nanoparticles attached onto the apex of an optical tip to serve as an efficient near-field surface-plasmon polariton launcher. The implementation of the nanodiamond-based tip in a near-field scanning optical microscope will allow for an accurate control on the launching position, thereby opening the way to scanning plasmonics.     

一种改进的势函数欠定盲源分离算法

针对原有的拉普拉斯混合模型势函数法复杂度高、随机选取部分观测数据点作为初始聚类中心的算法聚类结果不稳定、准确率低的问题,提出了一种改进的势函数欠定盲源分离算法．该算法在基于密度概念的基础上,以簇内距离小、簇间距离大为原则,选取部分高密度点作为势函数的初始聚类中心．理论分析与仿真实验表明,改进算法的复杂度大大降低,而估计准确度降低很少．在信噪比为10dB时,该算法仿真时间降为原始势函数法的5％;相对随机选取算法,在计算复杂度基本一致的前提下,该算法的估计准确度大大提高,源信号个数估计准确率由61％提高到85％,混合矩阵估计误差由0.47下降为0.27．     

初始聚类中心优化的K-均值项目聚类推荐算法

针对协同过滤推荐系统存在的数据稀疏性和扩展性差问题,提出了初始聚类中心优化的K-均值项目聚类推荐算法。该算法首先采用SlopeOne方法对评分矩阵预测填充来缓解数据稀疏性,然后采用初始聚类中心优化的K-均值算法对项目进行聚类,将相似度高的项目聚到同一个类中,最后根据目标项目所在的聚类搜索其最近邻并产生推荐。实验结果表明,该算法有效改善了数据的稀疏性和扩展性,提高了推荐质量。     

Effects of Different Ligands on Fluorescent Properties of Nd3+ Organic Complexes

合成了4种掺Nd3+的配合物分别是Nd(C2F5COO)3Tfpy,Nd(C3F7COO)3Tfpy,Nd(C2F5COO)2(C6F5COO)Tfpy和Nd(C2F5COO)3Phen,来研究不同的配体对配合物发光性能的影响。通过紫外吸收谱、红外光谱和核磁共振H谱表征了配合物的结构。在吸收谱、荧光谱和辐射寿命的基础上,应用Judd-Ofelt理论对4种配合物进行了分析。结构分析表明配体和中心的钕离子均成功配位。JO理论分析得到较小的?2表明钕离子和配体是通过离子键的形式配合的。配合物中的环状结构可以显著地提高钕离子的荧光量子效率。与常用的第二配体Phen相比,Tfpy由于自身较小的对称性和含较少的H原子而成为更好的中性第二配体。本研究合成的4种配合物都有较大的受激辐射面积,可以作为潜在的液体激光材料。     

聚类分析在储层综合评价中的应用

针对常规分析方法在储层分类中存在的一些问题，本文筛选不同储层特征的最有代表性参数作为变量，利用综合聚类分析方法，对研究区油层组各主力储层进行了详细的综合分类评价。划分Ⅰ类储层115个，Ⅱ类储层105个，分类结果经已开发区证实吻合性强，对有利目标区的优选具有一定的指导意义。     

Self-consistent Recombination Scheme in Porous Silicon Under Intense Laser Excitation

An experimental investigation of the general characteristics of nonradiative and radiative recombination of charge carriers in strongly excited porous silicon is presented. It is shown that photoconductivity, photomagnetoelectric effect, quantum yield, and intensity of visible radiation of porous silicon demonstrates strong nonlinearities against laser excitation intensity. It is suggested that the band-to-band Auger recombination is dominant similar to that in crystalline silicon, whereas the visible luminescence is determined by the bimolecular process. The nonequilibrium density of charge carriers n 10¹⁹ cm^–3, and the bimolecular radiative recombination coefficient B_rad 9 × 10^–14 cm₃/s have been found.     

Approximation and analysis of a call center with flexible and specialized servers

This paper describes a decomposition algorithm to estimate the performance of a call center with two types of customers and two server categories. In this system specialized servers can process only one customer type, while flexible servers handle both types. The algorithm divides the systems state space into regions, and simple approximate models find the conditional system performance within each region. While the procedure described here is tailored for a system with a priority queue discipline and two customer classes, it can be adapted for systems with FCFS queue disciplines and for systems with more than two customer types. Performance measures generated by the procedure are sufficiently accurate for many service system design decisions, such as setting telephone call center staffing levels and long-term capacity planning. The procedure is also extremely fast, and its computational requirements do not grow with system congestion. Numerical tests demonstrate that its running time is significantly lower than traditional numerical methods for generating approximations. As an example of its use, we employ the procedure to demonstrate the benefits of server flexibility in a particular telephone call center.The author thanks Gregory Dobson, who participated in the initial consulting project and problem formulation, as well as Paul Schweitzer, Harry Groenevelt, and two anonymous referees for their helpful suggestions.