All issues

Distributed saddle point problems (SPPs) have numerous applications in optimization, matrix games and machine learning. For example, the training of generated adversarial networks is represented as a min-max optimization problem, and training regularized linear models can be reformulated as an SPP as well. This paper studies distributed nonsmooth SPPs with Lipschitz-continuous objective functions. The objective function is represented as a sum of several components that are distributed between groups of computational nodes. The nodes, or agents, exchange information through some communication network that may be centralized or decentralized. A centralized network has a universal information aggregator (a server, or master node) that directly communicates to each of the agents and therefore can coordinate the optimization process. In a decentralized network, all the nodes are equal, the server node is not present, and each agent only communicates to its immediate neighbors.

We assume that each of the nodes locally holds its objective and can compute its value at given points, i. e. has access to zero-order oracle. Zero-order information is used when the gradient of the function is costly, not possible to compute or when the function is not differentiable. For example, in reinforcement learning one needs to generate a trajectory to evaluate the current policy. This policy evaluation process can be interpreted as the computation of the function value. We propose an approach that uses a smoothing technique, i. e., applies a first-order method to the smoothed version of the initial function. It can be shown that the stochastic gradient of the smoothed function can be viewed as a random two-point gradient approximation of the initial function. Smoothing approaches have been studied for distributed zero-order minimization, and our paper generalizes the smoothing technique on SPPs.

Algorithms of wireless sensor networks operation based on modified ZigBee/IEEE 802.15.4 protocol stack and problems of energy saving with simultaneous decrease of network latency are studied. Theoretical computations are given. Roles distribution and routers schedule assignment algorithms are described. Both results of experiments carried out with real devices and results of simulations with ns-2 (open-source network simulator) are given and analyzed.

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.

This article discusses the contribution of computing resources monitoring system to the work of a distributed computing system. Method of evaluation of this contribution and performance monitoring system based on measures of certainty the state-controlled system is proposed. The application of this methodology in the design and development of local monitoring of the Central Information and Computing Complex, Joint Institute for Nuclear Research is listed.

Processing of large amounts of data often consists of several steps, e.g. pre- and post-processing stages, which are executed sequentially with data written to disk after each step, however, when pre-processing stage for each task is different the more efficient way of processing data is to construct a pipeline which streams data from one stage to another. In a more general case some processing stages can be factored into several parallel subordinate stages thus forming a distributed pipeline where each stage can have multiple inputs and multiple outputs. Such processing pattern emerges in a problem of classification of wave energy spectra based on analytic approximations which can extract different wave systems and their parameters (e.g. wave system type, mean wave direction) from spectrum. Distributed pipeline approach achieves good performance compared to conventional “sequential-stage” processing.

Getting a qualified medical examination can be difficult for people in remote areas because medical staff available can either be inaccessible or it might lack expert knowledge at proper level. Telemedicine technologies can help in such situations. On one hand, such technologies allow highly qualified doctors to consult remotely, thereby increasing the quality of diagnosis and plan treatment. On the other hand, computer-aided analysis of the research results, anamnesis and information on similar cases assist medical staff in their routine activities and decision-making.

Creating telemedicine system for a particular domain is a laborious process. It’s not sufficient to pick proper medical experts and to fill the knowledge base of the analytical module. It’s also necessary to organize the entire infrastructure of the system to meet the requirements in terms of reliability, fault tolerance, protection of personal data and so on. Tools with reusable infrastructure elements, which are common to such systems, are able to decrease the amount of work needed for the development of telemedicine systems.

An interactive tool for creating distributed telemedicine systems is described in the article. A list of requirements for the systems is presented; structural solutions for meeting the requirements are suggested. A composition of such elements applicable for distributed systems is described in the article. A cardiac telemedicine system is described as a foundation of the tool

Independent components communicating with each other due to complex control make the work of complex distributed computer systems poorly scalable within the framework of the existing communication middleware. Two major problems of such systems' scaling can be defined: overloading of unequal nodes due to proportional redistribution of workload and difficulties in the realization of continuous communication between several nodes of the system. This paper is focused on the developed solution enabling visualization of the work of such a dynamical system.

Although each task of recreating artifacts is truly unique, the modeling process for façades, foundations and building elements can be parametrized. This paper is focused on a complex of the existing programming libraries and solutions that need to be united into a single computer system to solve such a task. An algorithm of generating 3D filling of objects under reconstruction is presented. The solution architecture necessary for the system's adaptation for a cloud environment is studied.

We'd like to present a specific grid infrastructure and web application development and deployment. The purpose of infrastructure and web application is to solve particular geophysical problems that require heavy computational resources. Here we cover technology overview and connector framework internals. The connector framework links problem-specific routines with middleware in a manner that developer of application doesn't have to be aware of any particular grid software. That is, the web application built with this framework acts as an interface between the user 's web browser and Grid's (often very) own middleware.

Our distributed computing system is built around Gridway metascheduler. The metascheduler is connected to TORQUE resource managers of virtual compute nodes that are being run atop of compute cluster utilizing the virtualization technology. Such approach offers several notable features that are unavailable to bare-metal compute clusters.

The first application we've integrated with our framework is seismic anisotropic parameters determination by inversion of SKS and converted phases. We've used probabilistic approach to inverse problem solution based on a posteriory probability distribution function (APDF) formalism. To get the exact solution of the problem we have to compute the values of multidimensional function. Within our implementation we used brute-force APDF calculation on rectangular grid across parameter space.

The result of computation is stored in relational DBMS and then represented in familiar human-readable form. Application provides several instruments to allow analysis of function's shape by computational results: maximum value distribution, 2D cross-sections of APDF, 2D marginals and a few other tools. During the tests we've run the application against both synthetic and observed data.