Miroslav Dobrucký

Parallel Flood Modeling Systems

Flood modeling is a complex problem that requires cooperation of many scientists in different areas. In this paper, the architecture and results of the ANFAS (Data Fusion for Flood Analysis and Decision Support) project is presented. This paper also focuses on the parallel numerical solutions of flood modeling module that are the most computational-intensive part of the whole ANFAS architecture.

Flood Forecasting in a Grid Computing Environment

This paper presents a prototype of a flood forecasting sys- tem based on Grid computing. Flood forecasting is a complex problem that requires cooperation and coupling of many simulation models. To enable this coupling in an effective way, a part of the CrossGrid project is aimed towards developing a Grid Computing Environment (GCE) for flood forecasting, as described here. The Grid computing environment consists of a cascade of simulation models, a storage system for com- puted and measured data and other used datasets, a web-based portal with collaboration tools and a powerful computation facility. The whole system is tied together by Grid technology and is used to support a virtual organization of experts, developers and users.

Hybrid Approach to Task Allocation in Distributed Systems

This paper describes the static and dynamic task allocation tools in PVM environment for distributed memory parallel systems. For the static mapping the objective function is used to evaluate the optimality of the allocation of a task graph onto a processor graph. Together with our optimization method also augmented simulated annealing and heuristic move exchange methods in distributed form are implemented. For dynamic task allocation the semidistributed approach was designed based on the division of processor network topology into independent and symmetric spheres. Distributed static mapping (DSM) and dynamic load balancing (DLB) tools are controlled by user window interface. DSM and DLB tools are integrated together with software monitor (PG-PVM) in the graphical GRAPNEL environment.

Parallel Numerical Solution for Flood Modeling Systems

Simulation of the water flood problems often leads to solving large sparse systems of partial differential equations. For such systems, numerical methods can be very CPU-time consuming. Therefore, parallel simulation is beneficial for water flood studies, and provides satisfactory accuracy. In this paper, we present an approach to parallelizing water flow modeling systems and some experimental results obtained on Linux clusters.

Unified nanoscale gateway to HPC and grid environments

Nanotechnology and advanced materials research is offered as a transformative technology with the potential to improve every aspect of social, physical, and economic well-being. Presently, the excellent research is running at Slovak Academy of Sciences (SAS) in various subfields and research disciplines in nanotechnology. As many other fields, nanotechnology research also requires computational and infrastructure support, especially for modelling and simulations. From this point of view, the nanotechnology area represents a big challenge where different scales meet. The high performance computing (HPC) and the Grid computing gives a good chance to simulate real physical nano-structural systems even on atomic level under various external conditions. This paper presents the computational support approach in such environments: a scientific user-centric gateway into the high computational capacity for nanoscale simulations.

Grid Computing with Relativistic Quantum Chemistry Software

We describe the adaptation of the scientific software DIRAC for the grid environment and provide the study of the code’s performance. The most effective approach for effective utilizing of the grid computing power is not relying upon program’s specific dependencies installed on grid servers, but rather in precise, local preparation of all necessary static executables and related scripting launchers. Running few functionality and performance tests before carrying out any production calculations with the code is recommended.

Ground Water Flow Modelling in PVM

In this paper we present the ground water flow modelling in PVM. It requires the solution of a partial differential equation representing a 3-D diffusion process. Its numerical solution by a finite difference method leads to a very large sparse, well-structured system of linear algebraic equations. The Strongly Implicit Procedure (SIP) algorithm is chosen to solve the matrix. We also present our parallel version of SIP on PC cluster with PVM. The experiments show that our parallel SIP is very effective.