Search

The origin of poor and moderate globular cluster systems around galaxies against the background of a nonlinear non-stationary model of collapsing galaxies is studied. The stability of the model with respect to four perturbation modes is studied, and the degree of the mode determines on average the number of clusters in the system. A critical dependence of the initial virial relation on the degree of rotation is constructed for each mode. The dependences of the instability increments on the physical parameters of the model are found.

Представленная статья посвящена теоретико-правовым аспектам процесса кластеризации, экономико-правовой регламентации вопросов создания и функционирования экономических кластеров в России на примере Мурманской области. This article is devoted to the theoretical and legal aspects of cauterization, economic and legal regulation of the creation and functioning of economic clusters in Russia by the example of the Murmansk region.

The key points of modernization of the national economy are identified: widespread development of market infrastructure, minimization of direct government intervention; cooperation and integration of interested business entities and the creation of cooperative integration corporations (firms) and cluster systems; end-to-end export orientation of domestic agribusiness; a high level of labor stimulation and its direct connection with production results.

The effect of a static magnetic field on the redistribution of the stochastic energy in an ensemble of charged particles in an electrostatic trap has been studied numerically and analytically. The simulation has been performed for cluster systems of particles with the Coulomb interaction in a wide range of their parameters. Conditions of energy balance have been proposed for charged particles in systems with spatially inhomogeneous heat sources.

As the demand for high-performance computing (HPC) resources has increased in the field of computational science, an inevitable consideration is service availability in large cluster systems such as supercomputers. In particular, the factor that most affects availability in supercomputing services is the job scheduler utilized for allocating resources. Consequent to submitting user data through the job scheduler for data analysis, 25.6% of jobs failed because of program errors, scheduler errors, or I/O errors. Based on this analysis, we propose a K-hook method for scheduling to increase the success rate of job submissions and improve the availability of supercomputing services. By applying this method, the job-submission success rate was improved by 15% without negatively affecting users’ waiting time. We also achieved a mean time between interrupts (MTBI) of 24.3 days and maintained average system availability at 97%. As this research was verified on the Nurion supercomputer in a real service environment, the value of the research is expected to be found in significant service improvements.

Reducing energy consumption in processors has become a critical issue in cluster systems. The dynamic voltage and frequency scaling (DVFS) is one of the effective techniques to minimize energy consumption in processors when executing parallel applications. These applications may consist of many interrelated tasks that may be computed when their precedence constraints are satisfied. The scheduling of these precedence-constrained tasks on to processors, to optimize makespan or energy consumption or both, has been studied as an NP-complete problem. Thus, many algorithms have been given by the researchers and newer algorithms keep coming in. In the parallel applications, there exist some tasks which may have slacks due to the dependencies between tasks. The EASLA (energy aware service level agreement) is one of the well-known energy-aware task scheduling algorithms which utilize the underused slack of tasks. In this paper, we present an improved version of its sub-algorithm, NCM (Not Changing Makespan), for heterogeneous cluster systems that is used to downscale frequencies when schedule length does not change. The presented algorithm also makes use of a fast and low complexity algorithm called PEFT (Predict Earliest Finish Time) to compute the schedule length of the application. We performed experiments for randomly generated graphs, and the results illustrate that the improved version of the NCM algorithm achieves better energy saving compared to the NCM algorithm.

This article reviews the performance of supercomputers. The increase in computational performance is also provided through parallel programming, the development of which is represented by a set of high-level programming languages, the emergence of new architectures and distributed and cluster systems. The article presents a numerical simulation of the calculation of the physical characteristics of the composites using the OpenMX package. At the same time, the use of parallel programming methods led to a significant increase in computing performance.

Accelerated clusters, which are cluster systems equipped with accelerators, are one of the most common systems in parallel computing. In order to exploit the performance of such systems, it is important to reduce communication latency between accelerator memories. In addition, there is also a need for a programming language that facilitates the maintenance of high performance by such systems. The goal of the present article is to evaluate XcalableACC (XACC), a parallel programming language, with tightly coupled accelerators/InfiniBand (TCAs/IB) hybrid communication on an accelerated cluster. TCA/IB hybrid communication is a combination of low-latency communication with TCA and high bandwidth with IB. The XACC language, which is a directive-based language for accelerated clusters, enables programmers to use TCA/IB hybrid communication with ease. In order to evaluate the performance of XACC with TCA/IB hybrid communication, we implemented the lattice quantum chromodynamics (LQCD) mini-application and evaluated the application on our accelerated cluster using up to 64 compute nodes. We also implemented the LQCD mini-application using a combination of CUDA and MPI (CUDA + MPI) and that of OpenACC and MPI (OpenACC + MPI) for comparison with XACC. Performance evaluation revealed that the performance of XACC with TCA/IB hybrid communication is 9% better than that of CUDA + MPI and 18% better than that of OpenACC + MPI. Furthermore, the performance of XACC was found to further increase by 7% by new expansion to XACC. Productivity evaluation revealed that XACC requires much less change from the serial LQCD code to implement the parallel LQCD code than CUDA + MPI and OpenACC + MPI. Moreover, since XACC can perform parallelization while maintaining the sequential code image, XACC is highly readable and shows excellent portability due to its directive-based approach.

Paramagnetic polyoxometalates [RE30Co8Ge12W108O408(OH)42(OH2)30]56− (Rare Earth (RE): Gd, Dy, Eu, and Y) are of special interest with regard to their application as alternative contrast agents in non-human magnetic resonance imaging which is increasingly used in materials science and process engineering. This class of new paramagnetic materials promises detailed findings in the magnetic resonance images due to their rather large total electron spin on the one hand, i.e. large, field-dependent relaxivities up to the highest magnetic fields, and due to their relatively large cluster sizes with an impact on adsorption and penetration on the other hand. Apart from the magnetic field dependence, the sensitivity of relaxivities to motional correlation times will be shown for these polyoxometalates which is a prerequisite for modelling and understanding the physical behaviour of this new class of polyoxometalates in MRI. Also for the qualitative and quantitative interpretation of MR images, the knowledge of transverse and longitudinal relaxivities of the paramagnetic clusters in a given environment is mandatory. Examples considered in this publication are proteins in milk fractionation, the deposit of which was measured by MRI.

There is increasing demand for High-Performance Computing (HPC) cluster systems because more and more computational resources are necessary to carry out large-scale simulations and machine learning which are a crucial part of research and business today. There are two choices of HPC cluster systems: academic HPC cluster systems and public cloud services. Academic HPC cluster systems can be used with reasonable usage fees, although they take a long time until the cluster system is ready to use, because of its user registration procedure. On the other hand, public cloud services, such as Amazon Web Services, can be used immediately, but the utilization fees are expensive compared with those for academic systems. To fill the gaps between them, we propose an academic HPC cluster system with automated user registration that allows users to utilize the HPC cluster on demand.

Electrostatic polarization effect plays a crucial role in determining the charge-carrier energy levels in organic electronics devices. The polarizable force field (PFF) is one of the popular methods to explicitly calculate charge carrier’s energy levels in condensed phase. It is a challenge to evaluate the performance of PFFs on estimation the carrier’s site energies due to the lack of accurate photoelectron spectroscopy (PES) data. As such, developing the first-principle based method is an alternative option to estimate the performance of PFFs on electronic polarization effect. In this paper, we applied the block-localized wavefunction (BLW) coupled with DFT to solve the charge localized energy in some oligoacene clusters, which provide a nice reference to evaluate the performance of AMOEBA PFF. Our results demonstrate that the AMOEBA PFF can well describe the electronic polarization in some small cluster systems.

As the scales of cluster systems increase, failures become normal and have made reliability management be a major concern for system administrators. Failure prediction is a proactive measure through mining failure patterns and predicting when the systems will fail. In general, it is helpful to improve the accuracy of failure prediction by mining true failure patterns. And currently, the statistical and data mining driven methods are often used for mining failure patterns. However, since the overwhelming volumes and complicated interleaving of logs, the efficient and accurate failure pattern detection and automated failure prediction are still challenging.

As the scales of cluster systems increase, failures become normal and make reliability be a major concern. System logs contain rich semantic information of all components, and they are always used to detect abnormal behaviors and failures by mining methods.

For formation control and obstacle avoidance of multi-agent cluster systems, this paper proposes a new formation control and obstacle avoidance algorithm based on artificial potential field method, the advantage of which is that it can enhance the robustness of multi-agent systems and make the overall performance as no central control. Formation control adopts the follower-leader method. Leaders are not artificially designated, but selected by autonomous decision-making of multi-agent systems. When the selected leader fails to continue to perform the task, the remaining followers can independently choose a new leader and form a new formation to continue the task according to the number of existing agents. Finally, the effectiveness of the proposed method is verified by a numerical simulation. The simulation results show that the algorithm proposed in this paper can perform no central formation control and obstacle avoidance for multi-agent systems.

The article is devoted to the study of the possibilities of taking into account the processes of formation of cluster systems in the structure of liquids and the influence of these processes on combination scatter of light. Within the framework of the classical theory of the Raman effect, a relation is obtained for the Stokes and anti-Stokes light scatters, taking into account the influence of clusters on scattered frequencies, and it is shown that the scattered frequencies are proportional to the square root of the number of particles in the cluster. The calculated frequencies of the Raman spectrum are compared according to the proposed model and experimental data in the frequency range 20–300 cm −1 , which showed the adequacy of the proposed model.

Esta tesis se centra en la aplicación de técnicas de computación de altas prestaciones a dos problemas bioinformáticos. La primera parte del trabajo ha consistido en el desarrollo de un software eficiente para el alineamiento de secuencias ADN/ARN sobre un genoma de referencia capaz de superar tanto la sensibilidad como la especificidad de los alineadores actuales y de reducir el tiempo de procesamiento. Como resultado final de este apartado se ha implementado un framework capaz de ejecutar cualquier alineador de ADN/ARN en un sistema clúster de forma transparente para el usuario. La segunda parte del trabajo ha consistido en el desarrollo de una versión optimizada del módulo de la aplicación FaST-LMM que se utiliza para estudiar la epistasis. También se ha implementado una versión paralela para sistemas clúster dotados de aceleradores gráficos de tipo GPU que permite descargar sobre ellos los cálculos matriciales más pesados. This thesis is focused on the application of high performance computing techniques to two problems in the field of bioinformatics. The main goal of the first part of this work is the development of an efficient software for the alignment of sequences ADN/ARN in a reference genome capable of overcoming both the sensitivity and the specificity of the current aligners, while reducing the processing time. As a final result, a framework has been carried out, capable of executing any DNA/RNA aligner in a cluster in a transparent way for the user. The second part has focused on the optimization of the FaST-LMM application which is used to detect epistasis. Specifically, the work has developed a new version with a better performance. A parallel version has also been implemented for cluster systems equipped with GPU-type graphic accelerators that allow downloading the matrix calculations on them. Programa de Doctorat en Informàtica

Структура i обсяг роботи: магістерська дисертація викладена на 128 cторiнкаx, складається зі вступу, 3 розділів, висновку, містить 20 рисунків, 4 таблиці, 13 формул, список використаних джерел із 33 найменувань на 5 cторiнкаx. Актуальність роботи. Станом на сьогодні спостерігається поява все більшої кількості комп’ютерних систем які мають гетерогенну структуру, засновану на використанні центрального процесора та акселераторів з відмінною від нього архітектурою. В той же час, для достатнього рівня швидкодії складних програмних комплексів стає не достатньо застосування лише паралельних систем, стандартною практикою для високонавантажених обчислень стає їх розгортка в розподілених комп’ютерних системах, для яких паралельні системи є лише складовими елементами. Для всіх розподілених систем одним з ключових факторів, що визначають ефективність роботи системи, є вміле балансування навантаження на всіх рівнях цієї системи. Вирішенню цієї NP-повної задачі наразі присвячено цілий розділ комп’ютерних наук, але вона досі лишається актуальною проблемою, оскільки абсолютно ідеального рішення за прийнятний час в ній знайти наразі не можливо. А в випадку, коли кінцеві вузли розподіленої системи являють собою гетерогенні комп’ютерні системи, то задача ускладняється ще більше, оскільки необхідно провести балансування обчислень між різнорідними елементами із зазвичай значно відмінними між собою характеристиками продуктивності та комунікабельності. Наразі існує великий набір стандартних методів балансування навантаження в розподілених системах, але більшість із них орієнтовані на конкретний тип задач, лише невелика кількість не прив’язані до конкретних задач. сучасні засоби організації паралельних обчислень в розподілених комп’ютерних системах з акселераторами приймають низький рівень абстракції, що веде до рядку складнощів та обмежень. Вбудовані в них методи балансування навантаження орієнтовані на гомогенні і здебільшого повнозв’язні системи. Мета роботи: підвищення ефективності проведення паралельних обчислень в гетерогенних розподілених комп’ютерних системах шляхом оптимального розподілу та балансування навантаження на її складові елементи. Завдання дослідження: 1. Виконати аналіз сучасних підходів до організації паралельних обчислень в РГКС, а також основні методи оптимізації таких обчислень. 2. Виділити основні проблеми які виникають при організації РГКС та паралельних обчислень в них, дослідити та систематизувати існуючі пропозиції їх вирішення. 3. Запропонувати методи та підходи до вирішення виявлених проблем, які дозволять підвищити ефективність паралельних обчислень в РГКС. 4. Провести експериментальні дослідження ефективності запропонованих методів та підходів шляхом розробки та тестування пакету програм для реальних РГКС. 5. Провести порівняльний аналіз створених програмних комплексів для РГКС із вже існуючими реалізаціями альтернативних підходів за показниками максимально досягнутих показників прискорення та ефективності. Об’єкт дослідження: паралельні та розподілені обчислення, високопродуктивні обчислення. Предмет дослідження: методи організації паралельних обчислень в розподілених гетерогенних комп’ютерних системах, методи планування обчислень та балансування навантаження в РГКС. Методи дослідження: методи статистичного опрацювання даних, теорія паралельних та розподілених обчислень, теорія планування, теорія оптимізації, теорія компіляторів, теорія алгоритмів, теорія графів. Публікації: 1. «Дослідження ефективності дрібнозернистого паралелізму в багатоядерних комп'ютерних системах», «Вісник НТУУ «КПІ. Інформатика, управління та обчислювальна техніка: зб. наук. праць», 2020, № 66, с. 56 – 61. Також результати та розвиток цього дослідження були представлені на міжнародних конференціях CSNT-2017, CSNT-2018, ICSFTI-2018 та опубліковані у відповідних збірниках праць даних конференцій. 2. «Neural network acceleration method in the two-component CPU-GPU computer systems», Збірник тез доповідей VI Міжнародної конференції «High Performance Computing» (HPC-UA 2020) [на рев’ю]. Також матеріали цього дослідження були представлені на міжнародних конференціях CSNT-2019, ICSFTI-2019 та опубліковані у відповідних збірниках праць даних конференцій. 3. «Застосування технології WCF для підвищення ефективності обчислень в сучасних розподілених комп’ютерних системах», Збірник тез доповідей XIII Міжнародної науково-технічної конференції «Комп’ютерні системи та мережні технології» (CSNT-2020) [очікує публікації]. 4. «Застосування технології WCF для підвищення ефективності паралельних обчислень в хмарних розподілених комп’ютерних системах», Безпека. Відмовостійкість. Інтелект: збірник праць міжнародної науково-практичної конференції ICSFTI2020 [очікує публікації]. Structure and scope of master's thesis: Master's thesis is presented on 128 pages, consists of introduction, 3 sections, conclusion, contains 20 drawings, 4 tables, 13 formulas, a bibliography of 33 titles on 5 pages. Relevance of work. Today, an increasing number of computer systems have a heterogeneous structure based on the use of central processing unit and accelerators with a different architecture. At the same time, it is not enough to use only parallel systems for a sufficient level of performance of complex software systems; it is standard practice for high-load computations to deploy them in distributed computer systems, in which parallel systems are only constituent elements. For all distributed systems, one of the key characteristics of the system performance is the ability to balance workloads at all levels of the system. A whole section of computer science is currently devoted to solving NP-complete problem, but it still remains an actual problem, since it is not possible to find the perfect solution in an acceptable time for now. And if the end nodes of a distributed system are heterogeneous computer systems, the task becomes even more complicated as it is necessary to balance the computations between heterogeneous elements with typically significantly different performance and communication characteristics. Currently, there is a large set of standard load balancing methods in distributed systems, but most of them are focused on specific types of tasks, and only a small number are not tied to them. Modern parallel computing tools in distributed computer systems with accelerators accept a low level of abstraction, leading to a number of difficulties and limitations. The load balancing techniques built into them are oriented on homogeneous and fully connected systems. Purpose of work: increasing the efficiency of parallel computing in distributed heterogeneous computer systems by optimally distributing and balancing the load on its elements. Research tasks: 1. To analyze modern approaches to the organization of parallel calculations in DHCS, as well as the basic methods of optimization of such computations; 2. To identify the main problems that arise in the organization of DHCS and parallel calculations in them, to investigate and systematize existing proposals for their solution; 3. To highlight the main problems and disadvantages of existing approaches and technologies, propose own solutions of the identified problems and ideas for improving the work of DHCS; 4. Conduct experimental studies of the effectiveness of the proposed methods and approaches by developing and testing a software systems for real DHCS; 5. To carry out a comparative analysis of the created software systems for DHCS with already existing implementations of alternative approaches based on the achieved acceleration and efficiency. Object of research: parallel and distributed computing, high-performance computing. Subject of research: methods of organization of parallel calculations in distributed heterogeneous computer systems, methods of calculations planning and load balancing in DHCS. Research methods: methods of statistical data processing, theory of parallel and distributed calculations, theory of planning, theory of optimization, theory of compilers, theory of algorithms, theory of graphs. Publications: 1. “Doslidzhennia efektyvnosti dribnozernystoho paralelizmu v bahatoiadernykh kompiuternykh systemakh” [The investigation of effectiveness in using of finegrained parallelism in multicore computer systems], Visnyk NTUU “KPI”. Informatyka, upravlinnia ta obchysliuvalna tekhnika: zbirnyk naukovykh prats [Herald of NTUU “KPI”. Information technology, management and computing technics: a collection of scientific papers], vol. 66, pp. 56 – 61., in press. Also, the results of this study were presented at international conferences CSNT-2017, CSNT-2018, ICSFTI-2018 and published in the relevant proceedings of these conferences. 2. «Neural network acceleration method in the two-component CPU-GPU computer systems», Proceedings of the 6 th International Conference “High Performance Computing HPC-UA 2020” [on review]. Also, the results of this study were presented at international conferences CSNT-2019, ICSFTI-2019 and published in the relevant proceedings of these conferences. 3. «Zastosuvannia tekhnolohii WCF dlia pidvyshchennia efektyvnosti obchyslen v suchasnykh rozpodilenykh komp’iuternykh systemakh» [The application of WCF technology to increase the efficiency of computing in modern distributed computer systems], Proceedings of the XIII International Conference CSNT2020 [in press]. 4. «Zastosuvannia tekhnolohii WCF dlia pidvyshchennia efektyvnosti paralelnykh obchyslen v khmarnykh rozpodilenykh komp’iuetrnykh systemakh» [The application of WCF technology to increase the efficiency of parallel computing in cloud distributed computer systems], Proceedings of the The International Conference on Security, Fault Tolerance, Intelligence (ICSFTI2020) [in press].

In order to solve many-body states in a wide energy region including continuum states, it is strongly desiredto develop a useful method. In nuclear cluster systems [1, 2,], we must treat both bound and unbound statessimultaneously. For such calculations, we propose to apply symplectic coherent-state bases (SCSB). In thisreport, the basic idea of the SCSB method will be explained. Especially the Pauli exclusion principle playsan important role in nuclear systems [3, 4], and the orthogonality condition model (OSM) [5] has beenstudied successfully for various kinds of nuclear cluster systems. We here discuss an application of theSCSB to the OCM. The SCSB method can be easily applied to two-cluster systems. However, for multiclustersystems, the applicability of the SCSB is not self-evident. The multi-cluster system is described bythe Pauli-allowed states (PAS) [6] under the Pauli exclusion principle. In this report, the multi-cluster SCSBwill be discussed on the basis of the PAS of the multi-cluster systems.

The selection of surface organic ligands is very important for the preparation of coinage metal clusters with precise structures, which can dictate the related properties of metal clusters. Alkynyl ligands, as a class of protecting ligands in metal cluster systems, have been gaining attentions because of the characteristics of both π and σ coordination modes. Specially, alkynyl-protected coinage metal clusters possess rich properties. However, there is no review to systemically summarize the functional properties and structure-functionality correlations of alkynyl-protected coinage metal clusters. In this review, the development of alkynyl ligands for the preparation of coinage metal clusters was reviewed. Additionally, the synthetic approaches of the alkynyl-protected coinage metal clusters were provided. Then, the functional properties of the alkynyl-protected coinage metal clusters were comprehensively summarized. In the end, the prospects and opportunities for the future development were discussed. We believed that the review will motivate the development of functionality-oriented alkynyl-protected coinage metal clusters.

Within the framework of the cluster model of the structure of disordered condensed media, using the pair potential of Mie interaction with the effective depth of the potential well, the relation was obtained for calculating the heat of sublimation of small cluster systems containing up to 500 particles. It is shown that the heat of sublimation of small clusters, referred to the energy of pair interaction between particles, is a universal function of the number of particles in the cluster system, which can be mathematically represented as the square of the hyperbolic tangent. The proposed model makes it possible to estimate the potential energy of the global minimum of the cluster system.

Within the framework of the cluster model, a relation is obtained for calculating the effective diameter of particles and the atomic packing coefficient in the cluster structure, which makes it possible to trace the dependence of the effective particle diameter on the temperature and the quantitative composition of the cluster system. The results of calculating the diameter of particles in a cluster model and classical models of hard and soft spheres are compared. The greatest discrepancy between the predicted particle diameters and the packing coefficients in the cluster structure within the framework of various models is observed near the melting point of the substance, where the difference can reach 16%.

Within the framework of the cluster model using the Mie pair power-law interaction potential with the effective depth of the potential well, whose depth is determined by the number of particles in the cluster, an equation for the calculation of the heat of melting of cluster systems is obtained. The resulting mathematical expression does not contain empirical constants and shows that the ratio of the heat of melting of a cluster to the heat of melting of a macroscopic sample is a universal function of the number of particles in the cluster and mathematically described by the square of the hyperbolic tangent.

We report the successful isolation and structural elucidation of two bimetallic doped [Co2 @Ge16 ]4- clusters (α and β form), which were synthesized through the reaction of [{(ArN)2 CtBu}Co(η6 -toluene)] (Ar=2,6-diisopropylphenyl) and K4 Ge9 in ethylenediamine (en) solution and co-crystallized together in [K(2,2,2-crypt)]4 [Co2 @Ge16 ]⋅en. The α-[Co2 @Ge16 ]4- isomer prefers a distinct D2h 3-connected architecture, whereas the deltahedral isomeric β-[Co2 @Ge16 ]4- isomer adopts a quasi-C2h geometry and can be seen as coupling of two distorted arachno-[Co@Ge10 ] units. Chemical bonding analyses indicate that the skeleton of the α isomer is mainly composed of localized bonds, whereas only multicenter bonding interactions govern the geometry of the β isomer, which was further found to exhibit a fluxional behavior. The coexistence of both isomers within one unit cell links the 3-connected clusters with their deltahedral congeners, thus highlighting the structural and electronic flexibility of such discreet cluster systems.

The article considers trends and dynamics of the regional development based on cluster initiatives. It should be noted that the current structural changes have a big impact on the economic development of regional and national markets. The object of the research has been selected by the elements such as regional cluster systems operating at the Russian and international markets, tendencies and features of the development. The authors are determined that in the context of the dynamic development of the economics of most countries, that have retained a high industrial potential, is needed an organizational support of industrial enterprises and complexes contributing to increase they competitiveness. Nowadays one of the most effective forms of such support is the increasing of the efficiency of industrial production including the integration and formation of industrial clusters. The implementation of cluster initiatives contributes to the development of business competitiveness by the realization of effective potential based on effective cooperation of the cluster’s participants. It is linked to the geographical position, including the expanding access to the innovations, technologies, “know-how”, specialized services and high-qualified staff, and also decrease of transaction costs, providing the formation of preconditions for the realization of join cooperation projects and productive competition.