Alexander B. Degtyarev

Constructing Virtual Private Supercomputer Using Virtualization and Cloud Technologies

One of efficient ways to conduct experiments on HPC platforms is to create custom virtual computing environments tailored to the requirements of users and their applications. In this paper we investigate virtual private supercomputer, an approach based on virtualization, data consolidation, and cloud technologies. Virtualization is used to abstract applications from underlying hardware and operating system while data consolidation is applied to store data in a distributed storage system. Both virtualization and data consolidation layers offer APIs for distributed computations and data processing. Combined, these APIs shift the focus from supercomputing technologies to problems being solved. Based on these concepts, we propose an approach to construct virtual clusters with help of cloud computing technologies to be used as on-demand private supercomputers and evaluate performance of this solution.

Synoptic and short-term modeling of ocean waves

The problem of wave climate description and modeling are considered on both a short-term and synoptic basis. The wave climate is considered as an ensemble of conditions of spatio-temporal wave fields characterized by frequency-directional spectra. Using an expanded set of wave and wind characteristics makes it possible to correctly introduce the concept of a wave weather “scenario” and to use it to evaluate vessel safety. The mathematical basis of the short-term wave description is by the representation of the stochastic processes by a generalized autoregressive model (ARM) and related models of periodically correlated random process. An ARM is used to generate space-time realizations of the wave surface of any short-term duration. Application of the aforesaid model takes into account the true stochastic process (lack of repetition in quasi-steady implementation of the process inherent in other models of wind waves). Ways of treating synoptic and seasonal variability of wave forecasts for a specific geographic region are also described. Here, for the first time using the general positions of stationary wave processes and non-stationary synoptic processes of average wave height variation are combined. In this case mathematical models of the same type are nested.

Simulation of Space Charge Dynamics in High Intensive Beams on Hybrid Systems

The method for construction of analytical expressions for electric and magnetic fields for some set of the distributions of the charge density is described. These expressions are used for symbolic computation of the corresponding electric and magnetic fields generated by the beam during the evolution in accelerators. Here we focus on the use of the matrix form for Lie algebraic methods for calculating the beam dynamics in the presence of self-field of the beam. In particular, the corresponding calculations are based on the predictor-corrector method. The suggested approach allows not only to carry out numerical experiments, but also to provide accurate analytical analysis of the impact of different effects with the use of ready-made modules in accordance with the concept of Virtual Accelerator Laboratory. To simulate the large number of particle distributed resources for computations are used. Pros and cons of using described approach on hybrid systems are discussed. In particular, the investigation of overall performance of the predictor-corrector method is made.

Application of Optimization of Parallel Algorithms to Queries in Relational Databases

All known approaches to parallel data processing in relational client-server database management systems are based only on inter-query parallelism. Nevertheless, it’s possible to achieve intra-query parallelism by consideration of a request structure and implementation of mathematical methods of parallel calculations for its equivalent transformation. This article presents an example of complex query parallelization and describes applicability of the graph theory and methods of parallel computing both for query parallelization and optimization.

Novel Approaches for Distributing Workload on Commodity Computer Systems

Efficient management of a distributed system is a common problem for universitys and commercial computer centres, and handling node failures is a major aspect of it. Failures which are rare in a small commodity cluster, at large scale become common, and there should be a way to overcome them without restarting all parallel processes of an application. The efficiency of existing methods can be improved by forming a hierarchy of distributed processes. That way only lower levels of the hierarchy need to be restarted in case of a leaf node failure, and only root node needs special treatment. Process hierarchy changes in real time and the workload is dynamically rebalanced across online nodes. This approach makes it possible to implement efficient partial restart of a parallel application, and transactional behaviour for computer centre service tasks.

Agent system service for supporting river boats navigation

Abstract In this paper agent system service is used for solving the troubles of river boats navigation and creating the cheapest and efficient solution for boats without onboard radiolocation system, with describing of system’s main modules construction.

Problem-Solving Environment for Beam Dynamics Analysis in Particle Accelerators

In particle accelerator physics the problem is that we can not see what is going on inside the working machine. There are a lot of packages for modelling the behaviour of the particles in numerical or analytical way. But for most physicists it is better to see the picture in motion to say exactly what is happening and how to influence on this. The goal of this work is to provide scientists with such a problem-solving environment, which can not only do some numerical calculations, but show the dynamics of changes as a motion 3D picture. To do this we use the power of graphical processors from both sides: for general purpose calculations and for there direct appointment – drawing 3D motion. Besides, this environment should analyse the behaviour of the system to provide the user with all necessary information about the problem and how to deal with it.

Analog-Digital Approach in Human Brain Modeling

Many companies and institutions in their attempts construct decision-making system, face a bottleneck in performance of their systems. Training neural networks can take from several days to several weeks. The traditional approach suggests modification of modern systems and microcircuits as long as their performance reaches a permissible limit. A different approach, unconventional, looks for opportunities in computing inspired by the human brain, neuromorphic computing. The idea was proposed by the engineer Carver Mead in the 80s and suggests combining artificial neural networks with specialized microcircuits. The architecture of the microchip needs to reproduce the mechanisms of the human brain and to be a kind of hardware support for neural networks. Last decade is characterized by a sharp growth of interest in neuromorphic computing, human brain modeling and peculiarities of how it works during making decisions. This is evidenced by the launch of a large-scale research programs like DARPA SyNAPSE (USA) and the Human Brain Project (EU), the purpose of which is to build a microprocessor system, which resembles the human brain in functionality, size and energy consumption. Existing models of the brain even on powerful supercomputers require significant computation time and are not yet able to solve problems in real time. Since the human brain consists of two parts with different functions and different data processing principles, there is a very promising approach which suggests combining digital and analog systems into single one. In current collaboration we incorporate some results of study of activity of human brain as a base of building of hybrid computational system and foundation to the approach of running it.

Desktop supercomputer: what can it do?

The paper addresses the issues of solving complex problems that require using supercomputers or multiprocessor clusters available for most researchers nowadays. Efficient distribution of high performance computing resources according to actual application needs has been a major research topic since high-performance computing (HPC) technologies became widely introduced. At the same time, comfortable and transparent access to these resources was a key user requirement. In this paper we discuss approaches to build a virtual private supercomputer available at user’s desktop: a virtual computing environment tailored specifically for a target user with a particular target application. We describe and evaluate possibilities to create the virtual supercomputer based on light-weight virtualization technologies, and analyze the efficiency of our approach compared to traditional methods of HPC resource management.

Distributed Computing Infrastructure Based on Dynamic Container Clusters

Modern scientific and business applications often require fast provisioning of an infrastructure tailored to particular application needs. In turn, actual physical infrastructure contains resources that might be underutilized by applications if allocated in dedicated mode (e.g., a process does not utilize provided CPU or network connection fully). Traditional virtualization technologies can solve the problem partially, however, overheads on bootstrapping a virtual infrastructure for each application and sharing physical resources might be significant. In this paper we propose and evaluate an approach to create and configure dedicated computing environment tailored to the needs of particular applications, which is based on light-weight virtualization also known as containers. We investigate available capabilities to model and create dynamic container-based virtual infrastructures sharing a common set of physical resources, and evaluate their performance on a set of test applications with different requirements.