An FETI-preconditioned conjuerate gradient method for large-scale stochastic finite element problems

In the spectral stochastic finite element method for analyzing an uncertain system. the uncertainty is represented by a set of random variables, and a quantity of Interest such as the system response is considered as a function of these random variables Consequently, the underlying Galerkin projection yields a block system of deterministic equations where the blocks are sparse but coupled. The solution of this algebraic system of equations becomes rapidly challenging when the size of the physical system and/or the level of uncertainty is increased This paper addresses this challenge by presenting a preconditioned conjugate gradient method for such block systems where the preconditioning step is based on the dual-primal finite element tearing and interconnecting method equipped with a Krylov subspace reusage technique for accelerating the iterative solution of systems with multiple and repeated right-hand sides. Preliminary performance results on a Linux Cluster suggest that the proposed Solution method is numerically scalable and demonstrate its potential for making the uncertainty quantification Of realistic systems tractable.

集群网络评测模型的新探索

传统集群网络(cluster area network,简称cLAN)的评测模型主要考虑了延迟、带宽、路由、拥塞、网络拓扑结构等因素.但这些因素是否足以描述实际应用程序在集群上的通信行为,或者对其在集群系统上的性能给出一个很好的预测呢?当对NAS Parallel Benchmark(2.4版本)在集群系统深腾1800(DeepComp 1800)上进行大量测试时发现,集群网络的通信性能可以被一种特殊的通信模式(LU模式)所严重影响.更深入的研究表明,这个影响LU模式的因素是独立于前面所述的如延迟、带宽、路由、拥塞、网络拓扑结构等因素的.因此有必要对集群网络的评测模型重新进行审视,并增加一个新的性能评测因子以反映这个新发现的现象.从研究结果来看,这个重新审视也将对集群系统上的并行算法设计以及实际大规模科学计算的应用程序性能的优化提供一些新的思路.

面向高性能数值计算的并行计算模型DRAM(h)

提出了一个基于存储层次的新并行计算模型DRAM（h），并在该模型下对两个经典并行数值计算算法的不同实现形式：四种形式并行了三角方程求解（PTRS）和六种形式无列选主元并行LU分解（PLU），进行了分析。模型分析表明，具有近乎相同时间和空间复杂性的同一算法不同实现形式在该模型下会有完全不同的存储复杂度。作者在日立公司SR2201 MPP并行机、曙光3000超级服务器和中国科学院科学与工程计算国家重点实验室（LSEC）的128节点Linux Cluster等三种并行计算平台上对模型分析结果进行了实验验证。结果表明，该模型分析在绝大数情况下都能较好地与不同实验枰台上的实验结果吻合。个别出现偏差的分析结果，在根据计算平台的存储层次特点修改模型分析的假定后，也能够进行解释。这说明了该模型以不同形式的算法实现进行存储访问模式区分的有效性。对在计算模型中加入指令/线程级并行的可行性和方法的研究是下一步的工作。

What is the right balance for performance and isolation with virtualization in HPC?

The use of virtualization in high-performance computing (HPC) has been suggested as a means to provide tailored services and added functionality that many users expect from full-featured Linux cluster environments. The use of virtual machines in HPC can offer several benefits, but maintaining performance is a crucial factor. In some instances the performance criteria are placed above the isolation properties. This selective relaxation of isolation for performance is an important characteristic when considering resilience for HPC environments that employ virtualization. In this paper we consider some of the factors associated with balancing performance and isolation in configurations that employ virtual machines. In this context, we propose a classification of errors based on the concept of “error zones”, as well as a detailed analysis of the trade-offs between resilience and performance based on the level of isolation provided by virtualization solutions. Finally, a set of experiments are performed using different virtualization solutions to elucidate the discussion.

Modeling and simulation of a synthetic genetic circuit that implements a multicelled behavior in a growing microcolony of E. coli

Synthetic Biology is a relatively new discipline, born at the beginning of the New Millennium, that brings the typical engineering approach (abstraction, modularity and standardization) to biotechnology. These principles aim to tame the extreme complexity of the various components and aid the construction of artificial biological systems with specific functions, usually by means of synthetic genetic circuits implemented in bacteria or simple eukaryotes like yeast. The cell becomes a programmable machine and its low-level programming language is made of strings of DNA. This work was performed in collaboration with researchers of the Department of Electrical Engineering of the University of Washington in Seattle and also with a student of the Corso di Laurea Magistrale in Ingegneria Biomedica at the University of Bologna: Marilisa Cortesi. During the collaboration I contributed to a Synthetic Biology project already started in the Klavins Laboratory. In particular, I modeled and subsequently simulated a synthetic genetic circuit that was ideated for the implementation of a multicelled behavior in a growing bacterial microcolony. In the first chapter the foundations of molecular biology are introduced: structure of the nucleic acids, transcription, translation and methods to regulate gene expression. An introduction to Synthetic Biology completes the section. In the second chapter is described the synthetic genetic circuit that was conceived to make spontaneously emerge, from an isogenic microcolony of bacteria, two different groups of cells, termed leaders and followers. The circuit exploits the intrinsic stochasticity of gene expression and intercellular communication via small molecules to break the symmetry in the phenotype of the microcolony. The four modules of the circuit (coin flipper, sender, receiver and follower) and their interactions are then illustrated. In the third chapter is derived the mathematical representation of the various components of the circuit and the several simplifying assumptions are made explicit. Transcription and translation are modeled as a single step and gene expression is function of the intracellular concentration of the various transcription factors that act on the different promoters of the circuit. A list of the various parameters and a justification for their value closes the chapter. In the fourth chapter are described the main characteristics of the gro simulation environment, developed by the Self Organizing Systems Laboratory of the University of Washington. Then, a sensitivity analysis performed to pinpoint the desirable characteristics of the various genetic components is detailed. The sensitivity analysis makes use of a cost function that is based on the fraction of cells in each one of the different possible states at the end of the simulation and the wanted outcome. Thanks to a particular kind of scatter plot, the parameters are ranked. Starting from an initial condition in which all the parameters assume their nominal value, the ranking suggest which parameter to tune in order to reach the goal. Obtaining a microcolony in which almost all the cells are in the follower state and only a few in the leader state seems to be the most difficult task. A small number of leader cells struggle to produce enough signal to turn the rest of the microcolony in the follower state. It is possible to obtain a microcolony in which the majority of cells are followers by increasing as much as possible the production of signal. Reaching the goal of a microcolony that is split in half between leaders and followers is comparatively easy. The best strategy seems to be increasing slightly the production of the enzyme. To end up with a majority of leaders, instead, it is advisable to increase the basal expression of the coin flipper module. At the end of the chapter, a possible future application of the leader election circuit, the spontaneous formation of spatial patterns in a microcolony, is modeled with the finite state machine formalism. The gro simulations provide insights into the genetic components that are needed to implement the behavior. In particular, since both the examples of pattern formation rely on a local version of Leader Election, a short-range communication system is essential. Moreover, new synthetic components that allow to reliably downregulate the growth rate in specific cells without side effects need to be developed. In the appendix are listed the gro code utilized to simulate the model of the circuit, a script in the Python programming language that was used to split the simulations on a Linux cluster and the Matlab code developed to analyze the data.

Visualização simultânea e multi-nível de informações de monitoramento de cluster

Clusters de computadores são geralmente utilizados para se obter alto desempenho na execução de aplicações paralelas. Sua utilização tem aumentado significativamente ao longo dos anos e resulta hoje em uma presença de quase 60% entre as 500 máquinas mais rápidas do mundo. Embora a utilização de clusters seja bastante difundida, a tarefa de monitoramento de recursos dessas máquinas é considerada complexa. Essa complexidade advém do fato de existirem diferentes configurações de software e hardware que podem ser caracterizadas como cluster. Diferentes configurações acabam por fazer com que o administrador de um cluster necessite de mais de uma ferramenta de monitoramento para conseguir obter informações suficientes para uma tomada de decisão acerca de eventuais problemas que possam estar acontecendo no seu cluster. Outra situação que demonstra a complexidade da tarefa de monitoramento acontece quando o desenvolvedor de aplicações paralelas necessita de informações relativas ao ambiente de execução da sua aplicação para entender melhor o seu comportamento. A execução de aplicações paralelas em ambientes multi-cluster e grid juntamente com a necessidade de informações externas à aplicação é outra situação que necessita da tarefa de monitoramento. Em todas essas situações, verifica-se a existência de múltiplas fontes de dados independentes e que podem ter informações relacionadas ou complementares. O objetivo deste trabalho é propor um modelo de integração de dados que pode se adaptar a diferentes fontes de informação e gerar como resultado informações integradas que sejam passíveis de uma visualização conjunta por alguma ferramenta. Esse modelo é baseado na depuração offline de aplicações paralelas e é dividido em duas etapas: a coleta de dados e uma posterior integração das informações. Um protótipo baseado nesse modelo de integração é descrito neste trabalho Esse protótipo utiliza como fontes de informação as ferramentas de monitoramento de cluster Ganglia e Performance Co-Pilot, bibliotecas de rastreamento de aplicações DECK e MPI e uma instrumentação do Sistema operacional Linux para registrar as trocas de contexto de um conjunto de processos. Pajé é a ferramenta escolhida para a visualização integrada das informações. Os resultados do processo de integração de dados pelo protótipo apresentado neste trabalho são caracterizados em três tipos: depuração de aplicações DECK, depuração de aplicações MPI e monitoramento de cluster. Ao final do texto, são delineadas algumas conclusões e contribuições desse trabalho, assim como algumas sugestões de trabalhos futuros.

Analisi di un'architettura Multi-Master ad alta affidabilità basata su MariaDB Galera Cluster

Progettazione, realizzazione ed analisi prestazionale e di robustezza di un sistema di cluster ad alta affidabilità basato su MariaDB Galera Cluster. Affiancamento al cluster di un sistema di proxy MaxScale. Studio e realizzazione di una procedura di migrazione da DBMS MySQL a MariaDB Galera Cluster.

Cluster scheduling and load balancing via TCP options

This paper describes an experiment in designing, implementing and testing a Transport layer cluster scheduling and dispatching architecture. The motivation for the experiment was the hypothesis that a Transport layer clustering solution may offer advantantages over the existing industry-standard Network layer and Data Link Layer approaches. The critical success factors initially established to guide and evaluate the experiment were reduced dispatcher work load, reduced dispatcher internal state memory requirements, distributed denial of service resilience, and cluster software design simplicity. The functional design stage of the experiment produced a Transport layer strategy for scheduling and load balancing based on the specification of two new TCP options. Implementation required the introduction of the newly specified TCP options into the Linux (2.4) kernel. The implementation produced an extended Linux Socket API to facilitate user-process access to the additional TCP capability. The testing stage of the experiment confirmed the operational efficiency of the solution.

A cluster-randomised intervention trial against Schistosoma japonicum in the Peoples' Republic of China: bovine and human transmission

Background Zoonotic schistosomiasis japonica is a major public health problem in China. Bovines, particularly water buffaloes, are thought to play a major role in the transmission of schistosomiasis to humans in China. Preliminary results (1998–2003) of a praziquantel (PZQ)-based pilot intervention study we undertook provided proof of principle that water buffaloes are major reservoir hosts for S. japonicum in the Poyang Lake region, Jiangxi Province. Methods and Findings Here we present the results of a cluster-randomised intervention trial (2004–2007) undertaken in Hunan and Jiangxi Provinces, with increased power and more general applicability to the lake and marshlands regions of southern China. The trial involved four matched pairs of villages with one village within each pair randomly selected as a control (human PZQ treatment only), leaving the other as the intervention (human and bovine PZQ treatment). A sentinel cohort of people to be monitored for new infections for the duration of the study was selected from each village. Results showed that combined human and bovine chemotherapy with PZQ had a greater effect on human incidence than human PZQ treatment alone. Conclusions The results from this study, supported by previous experimental evidence, confirms that bovines are the major reservoir host of human schistosomiasis in the lake and marshland regions of southern China, and reinforce the rationale for the development and deployment of a transmission blocking anti-S. japonicum vaccine targeting bovines.