All issues

The method for mapping of intermediate representations (IR) set of C, C++ programs to vector embedding space is considered to create an empirical estimation framework for static performance prediction using LLVM compiler infrastructure. The usage of embeddings makes programs easier to compare due to avoiding Control Flow Graphs (CFG) and Data Flow Graphs (DFG) direct comparison. This method is based on transformation series of the initial IR such as: instrumentation — injection of artificial instructions in an instrumentation compiler’s pass depending on load offset delta in the current instruction compared to the previous one, mapping of instrumented IR into multidimensional vector with IR2Vec and dimension reduction with t-SNE (t-distributed stochastic neighbor embedding) method. The D1 cache miss ratio measured with perf stat tool is considered as performance metric. A heuristic criterion of programs having more or less cache miss ratio is given. This criterion is based on embeddings of programs in 2D-space. The instrumentation compiler’s pass developed in this work is described: how it generates and injects artificial instructions into IR within the used memory model. The software pipeline that implements the performance estimation based on LLVM compiler infrastructure is given. Computational experiments are performed on synthetic tests which are the sets of programs with the same CFGs but with different sequences of offsets used when accessing the one-dimensional array of a given size. The correlation coefficient between performance metric and distance to the worst program’s embedding is measured and proved to be negative regardless of t-SNE initialization. This fact proves the heuristic criterion to be true. The process of such synthetic tests generation is also considered. Moreover, the variety of performance metric in programs set in such a test is proposed as a metric to be improved with exploration of more tests generators.

Virtual machines are usually associated with an ability to create them on demand by calling web services, then these machines are used to deliver resident services to their clients; however, providing clients with an ability to run an arbitrary programme on the newly created machines is beyond their power. Such kind of usage is useful in a high performance computing environment where most of the resources are consumed by batch programmes and not by daemons or services. In this case a cluster of virtual machines is created on demand to run a distributed or parallel programme and to save its output to a network attached storage. Upon completion this cluster is destroyed and resources are released. With certain modifications this approach can be extended to interactively deliver computational resources to the user thus providing virtual desktop as a service. Experiments show that the process of creating virtual clusters on demand can be made efficient in both cases.

Dynamic scenarios leading to multistability in the form of continuous families of stable solutions are studied for a system of autonomous differential equations. The approach is based on determining the cosymmetries of the problem, calculating stationary solutions, and numerically-analytically studying their stability. The analysis is carried out for equations of the Lotka –Volterra type, describing the interaction of two predators feeding on two related prey species. For a system of ordinary differential equations of the 4th order with 11 real parameters, a numerical-analytical study of possible interaction scenarios was carried out. Relationships are found analytically between the control parameters under which the cosymmetry linear in the variables of the problem is realized and families of stationary solutions (equilibria) arise. The case of multicosymmetry is established and explicit formulas for a two-parameter family of equilibria are presented. The analysis of the stability of these solutions made it possible to reveal the division of the family into regions of stable and unstable equilibria. In a computational experiment, the limit cycles branching off from unstable stationary solutions are determined and their multipliers corresponding to multistability are calculated. Examples of the coexistence of families of stable stationary and non-stationary solutions are presented. The analysis is carried out for the growth functions of logistic and “hyperbolic” types. Depending on the parameters, scenarios can be obtained when only stationary solutions (coexistence of prey without predators and mixed combinations), as well as families of limit cycles, are realized in the phase space. The multistability scenarios considered in the work allow one to analyze the situations that arise in the presence of several related species in the range. These results are the basis for subsequent analysis when the parameters deviate from cosymmetric relationships.

The article discusses the problem of the influence of the research goals on the structure of the multivariate model of regression analysis (in particular, on the implementation of the procedure for reducing the dimension of the model). It is shown how bringing the specification of the multiple regression model in line with the research objectives affects the choice of modeling methods. Two schemes for constructing a model are compared: the first does not allow taking into account the typology of primary predictors and the nature of their influence on the performance characteristics, the second scheme implies a stage of preliminary division of the initial predictors into groups, in accordance with the objectives of the study. Using the example of solving the problem of analyzing the causes of burnout of creative workers, the importance of the stage of qualitative analysis and systematization of a priori selected factors is shown, which is implemented not by computing means, but by attracting the knowledge and experience of specialists in the studied subject area. The presented example of the implementation of the approach to determining the specification of the regression model combines formalized mathematical and statistical procedures and the preceding stage of the classification of primary factors. The presence of this stage makes it possible to explain the scheme of managing (corrective) actions (softening the leadership style and increasing approval lead to a decrease in the manifestations of anxiety and stress, which, in turn, reduces the severity of the emotional exhaustion of the team members). Preclassification also allows avoiding the combination in one main component of controlled and uncontrolled, regulatory and controlled feature factors, which could worsen the interpretability of the synthesized predictors. On the example of a specific problem, it is shown that the selection of factors-regressors is a process that requires an individual solution. In the case under consideration, the following were consistently used: systematization of features, correlation analysis, principal component analysis, regression analysis. The first three methods made it possible to significantly reduce the dimension of the problem, which did not affect the achievement of the goal for which this task was posed: significant measures of controlling influence on the team were shown. allowing to reduce the degree of emotional burnout of its participants.

Variational inequalities constitute a broad class of problems with applications in a number of fields, including game theory, economics, and machine learning. Today’s practical applications of VIs are becoming increasingly computationally demanding. It is therefore necessary to employ distributed computations to solve such problems in a reasonable time. In this context, workers have to exchange data with each other, which creates a communication bottleneck. There are three main techniques to reduce the cost and the number of communications: the similarity of local operators, the compression of messages and the use of local steps on devices. There is an algorithm that uses all of these techniques to solve the VI problem and outperforms all previous methods in terms of communication complexity. However, this algorithm is limited to unbiased compression. Meanwhile, biased (contractive) compression leads to better results in practice, but it requires additional modifications within an algorithm and more effort to prove the convergence. In this work, we develop a new algorithm that solves distributed VI problems using data similarity, contractive compression and local steps on devices, derive the theoretical convergence of such an algorithm, and perform some experiments to show the applicability of the method.

We propose a family of Gauss –Newton methods for solving optimization problems and systems of nonlinear equations based on the ideas of using the upper estimate of the norm of the residual of the system of nonlinear equations and quadratic regularization. The paper presents a development of the «Three Squares Method» scheme with the addition of a momentum term to the update rule of the sought parameters in the problem to be solved. The resulting scheme has several remarkable properties. First, the paper algorithmically describes a whole parametric family of methods that minimize functionals of a special kind: compositions of the residual of a nonlinear equation and an unimodal functional. Such a functional, entirely consistent with the «gray box» paradigm in the problem description, combines a large number of solvable problems related to applications in machine learning, with the regression problems. Secondly, the obtained family of methods is described as a generalization of several forms of the Levenberg –Marquardt algorithm, allowing implementation in non-Euclidean spaces as well. The algorithm describing the parametric family of Gauss –Newton methods uses an iterative procedure that performs an inexact parametrized proximal mapping and shift using a momentum term. The paper contains a detailed analysis of the efficiency of the proposed family of Gauss – Newton methods; the derived estimates take into account the number of external iterations of the algorithm for solving the main problem, the accuracy and computational complexity of the local model representation and oracle computation. Sublinear and linear convergence conditions based on the Polak – Lojasiewicz inequality are derived for the family of methods. In both observed convergence regimes, the Lipschitz property of the residual of the nonlinear system of equations is locally assumed. In addition to the theoretical analysis of the scheme, the paper studies the issues of its practical implementation. In particular, in the experiments conducted for the suboptimal step, the schemes of effective calculation of the approximation of the best step are given, which makes it possible to improve the convergence of the method in practice in comparison with the original «Three Square Method». The proposed scheme combines several existing and frequently used in practice modifications of the Gauss –Newton method, in addition, the paper proposes a monotone momentum modification of the family of developed methods, which does not slow down the search for a solution in the worst case and demonstrates in practice an improvement in the convergence of the method.

Parameter sweep applications are a very important class of applications, which are typically defined as a set of computational experiments over a set of input parameters, each of which is executed with its own parameter combination. These computations arise in many scientific contexts. This article introduces the Parameter Sweep web service that runs such applications in distributed computing environment. Also discussed is the Everest cloud platform, on which this service is built.

This work describes the idea of the system of informational support for the scientific projects and the development of computational task tracking complex. Due to large requirements for computational experiments the problem of presentation of the information about HPC tasks becomes one of the most important. Nonstandard usage of the service desk system as a basis of the computational task tracking and support system can be the solution of this problem. Particular attention is paid to the analysis and the satisfaction of the conflicting requirements to the task tracking complex from the different user groups. Besides the web service kit used for the integration of the task tracking complex and the datacenter environment is considered. This service kit became the main interconnect between the parts of the scientific project support system and also this kit allows to reconfigure the whole system quickly and safely.

The review of the distributed computing infrastructure of the Tier-1 sites for the Alice, ATLAS, LHCb experiments at the LHC is given. The special emphasis is placed on the main tasks and services of the Tier-1 site, which operates in the Kurchatov Institute in Moscow.

Efficiency and completeness of the numerical simulation in oceanography and hydrometeorology are entirely determined by algorithmic features of the construction of an interactive computer simulations in the scale of the oceans with adaptive coated closed seas and coastal waters refined mathematical models, with the possibility of specifying software parallelization calculations near the concrete — the protected areas of the sea coast. An important component of the research is continuous graphical visualization techniques in the course of calculations, including those undertaken in parallel processes with shared RAM or test points on the external media. The results of computational experiments are used in the description of hydrodynamic processes near the coast, which is important in keeping the organization of sea control services and forecasting marine hazards.