Optimizing the steady-state throughput of scatter and reduce operations on heterogeneous platforms

In this paper, we consider the communications involved by the execution of a complex application, deployed on a heterogeneous large-scale distributed platform. Such applications intensively use collective macro-communication schemes, such as scatters, personalized all-to-alls or gather/reduce operations. Rather than aiming at minimizing the execution time of a single macro-communication, we focus on the steady-state operation. We assume that there is a large number of macro-communications to perform in pipeline fashion, and we aim at maximizing the throughput, i.e., the (rational) number of macro-communications which can be initiated every time-step. We target heterogeneous platforms, modeled by a graph where resources have different communication and computation speeds. The situation is simpler for series of scatters or personalized all-to-alls than for series of reduces operations, because of the possibility of combining various partial reductions of the local values, and of interleaving computations with communications. In all cases, we show how to determine the optimal throughput, and how to exhibit a concrete periodic schedule that achieves this throughput.     

A multivariate statistical framework for the analysis of software effort phase distribution

ContextIn software project management, the distribution of resources to various project activities is one of the most challenging problems since it affects team productivity, product quality and project constraints related to budget and scheduling.ObjectiveThe study aims to (a) reveal the high complexity of modelling the effort usage proportion in different phases as well as the divergence from various rules-of-thumb in related literature, and (b) present a systematic data analysis framework, able to offer better interpretations and visualisation of the effort distributed in specific phases.MethodThe basis for the proposed multivariate statistical framework is Compositional Data Analysis, a methodology appropriate for proportions, along with other methods like the deviation from rules-of-thumb, the cluster analysis and the analysis of variance. The effort allocations to phases, as reported in around 1500 software projects of the ISBSG R11 repository, were transformed to vectors of proportions of the total effort and were analysed with respect to prime project attributes.ResultsThe proposed statistical framework was able to detect high dispersion among data, distribution inequality and various interesting correlations and trends, groupings and outliers, especially with respect to other categorical and continuous project attributes. Only a very small number of projects were found close to the rules-of-thumb from the related literature. Significant differences in the proportion of effort spent in different phrases for different types of projects were found.ConclusionThere is no simple model for the effort allocated to phases of software projects. The data from previous projects can provide valuable information regarding the distribution of the effort for various types of projects, through analysis with multivariate statistical methodologies. The proposed statistical framework is generic and can be easily applied in a similar sense to any dataset containing effort allocation to phases.     

Theoretical Investigation of Phase Equilibria for Metal-Hydrogen Alloy

Cluster variation method (CVM) was applied to calculate phase equilibria of metal-hydrogen systems. Two subjects are introduced in the present report. One is the summary of previous studies on the Pd-H system, and it is demonstrated that a single CVM free energy formula can systematically derive information of phase equilibria, intrinsic stability, and short range order diffuse intensities. The second subject is the theoretical calculations of superabundant vacancy (SAV) formation. Within the square approximation of the CVM, it is shown that abundant vacancies are introduced with the absorption of hydrogen when the interaction between vacancy and hydrogen is considered. This article was presented at the Multi-Component Alloy Thermodynamics Symposium sponsored by the Alloy Phase Committee of the joint EMPMD/SMD of The Minerals, Metals, and Materials Society (TMS), held in San Antonio, Texas, March 12-16, 2006, to honor the 2006 William Hume-Rothery Award recipient, Professor W. Alan Oates of the University of Salford, UK. The symposium was organized by Y. Austin Chang of the University of Wisconsin, Madison, WI, Patrice Turchi of the Lawrence Livermore National Laboratory, Livermore, CA, and Rainer Schmid-Fetzer of the Technische Universitat Clausthal, Clauthal-Zellerfeld, Germany.     

基于聚类分析原理建立化学反应速率方程

基于建立化学反应速率方程是对化学反应进行分类的观点， 提出了应用聚类分析技术中的相似系数法和距离法确定化学反应级数和速率常数的方法。该方法具有可靠性高、易于编程的特点， 并可用于确定非整数型反应级数。     

卫星系统新技术--应用机遇及发展策略考虑(下)

结合市场导向下的技术驱动重要性，重点论述现代卫星系统新技术及其在中国的应用机遇和发展策略。     

高可用性软件的设计与实现

从高可用性系统的类型和功能出发，在分析系统失效因素和高可用系统结构的基础上，设计了一个高可用性的软件结构系统，它可以运行于主备、双机和多机3种工作模式，并以有限状态机的状态转移图，具体实现了链路管理和业务接管等关键技术，最后将所设计的软件应用到实际的系统之中，能够很好地满足需求。     

Fundamental Mass Flow Measurement of Solid Particles

A generic mass flow measurement device was developed as a variation on the theme of counting. In a hypothetical infinitely sparse mass flow, the number of passing particles could be counted in a time frame and multiplied by the mean mass per particle to obtain a mass flow per time unit. In a mass flow of realistic density, however, particles travel in cluster formation and direct counting of individual particles is impossible. If a method could be available that reconstructs the original number of particles in a cluster, the mass flow can be computed for realistic clustered mass flows. This reconstruction algorithm was developed in this research; it uses the measured cluster lengths to reconstruct the total number of particles in each passing cluster. The lengths of the clusters were measured with an optoelectronic device. The reconstruction algorithm was developed using simulation, augmented by clustering theory. For identical diameter particle flow, simulation results showed that the number of particles in a cluster could be reconstructed using a very simple reconstruction formula. This formula uses only the total number of clusters per time frame and the total number of individual particles measured in the same time frame. However, identical diameter flow is not realistic, since even identical particles are measured with a certain error. Reconstruction of the realistic distributed diameter particle flow was approximated using the identical particle method. The optical mass flow sensor has major advantages over traditional methods. It is virtually insensitive to vibrations, contamination, temperature drift, and misalignment and the underlying measurement concept is well understood. But most importantly, the sensor does not require calibration. The mass flow of identical particles (4.5 mm air gun pellets) was measured with an error smaller than 3% even for high density flow rates.     

Clustering of variables, application in consumer and sensory studies

We consider two dissimilarity measures between variables that take account of the variances of the variables as well as of their correlations. When variables are standardised, we retrieve widely used dissimilarity measures. The first dissimilarity measure is Euclidean distance and is suitable in studies where negative correlation between variables implies disagreement. The second dissimilarity measure is a Procrustean distance and is suitable in situations where both positive and negative correlations imply agreement. We also discuss aggregation strategies in order to carry out hierarchical clustering and find groups of variables. Applications in consumer and sensory studies are outlined.     

Desorption of gold nanoclusters (2–150 nm) by 1 GeV Pb ions

Nanodispersed targets of gold (grains sized at 2–150 nm) were irradiated with 956 MeV ions of Pb⁵⁴⁺ ((dE/dx)_e in gold 87 keV/nm). Ejected gold was gathered on collectors. Desorbed nanoclusters of gold were detected by means of TEM while the total matter transfer was measured by neutron activation analysis. For all the targets a part of ejected gold presents nanoclusters in the same size range as that of the grains on the corresponding targets. Desorption of nanoclusters with the size up to 90 nm was observed for the first time for atomic primary ions in the electronic stopping regime. The yield of the desorbed nanoclusters decreases from 22 to 1.4 cluster/ion with increasing the mean grain size from 6 to 30 nm. The total matter transfer measured for the target with the grain size 6–10 nm has a great value – 5 × 10⁵ at./ion. Results are discussed.     

大型券商集中交易系统实现架构的研究

针对中国券商经纪业务正在进行经营模式转型的现实需要，该文提出采用低成本的Intel架构建设大型券商集中交易系统的方案，并从可靠性、高效性、扩充性、经济性的角度分析了系统的部署结构、关键技术和业务架构，揭示了基于低成本Intel架构的集中交易系统已达到券商进行经纪业务转型的要求。