Browsing by Subject "parallel computing"

Now showing 1 - 17 of 17

Access status: Open Access ,
A parallel algorithm of icsym forcomplexsymmetric linear systems in quantum chemistry
(Wydawnictwa AGH, 2018) Zhang, Yingchun; Lv, Quanyi; Xiao, Manyu; Xie, Gongnan; Breitkopf, Piotr
Computational effort is a common issue for solving large-scale complex symmetric linear systems, particularly in quantum chemistry applications. In order to alleviate this problem, we propose a parallel algorithm of improved conjugate gradient-type iterative (ICSYM). Using three-term recurrence relation and or- thogonal properties of residual vectors to replace the tridiagonalization process of classical CSYM, which allows to decrease the degree of the reduce-operator from two to one communication at each iteration and to reduce the amount of vector updates and vector multiplications. Several numerical examples are implemented to show that high performance of proposed improved version is obtained both in convergent rate and in parallel efficiency.
Access status: Open Access ,
A parallel approach for metaheuristics solving the labs problem using CPU and GPU
(Wydawnictwa AGH, 2025) Żurek, Dominik; Piętak, Kamil; Pietroń, Marcin; Kisiel-Dorohinicki, Marek
This paper contributes to solving the low autocorrelation binary sequence (LABS) problem that remains an open hard-optimization problem with many applications. The current direction of research is focused on developing algorithms dedicated to parallel architectures such as GPGPU or multi-core CPUs. The paper follows this direction and proposes new heuristics developed from the steepest-descent local search algorithm that extends the notion of a neighborhood of a given sequence. The introduced algorithms utilize the parallel nature of multicore CPUs and provide an effective method for solving the LABS problem. The efficiency levels of SDSL and the new algorithm are presented; to ensure an effective comparison, they were both implemented in the same manner. The comparison shows that exploring the larger neighborhood improves the efficiency of the search method.
Access status: Open Access ,
Akceleracja algorytmów przetwarzania obrazów z wykorzystaniem zasobów karty graficznej
(Wydawnictwa AGH, 2009) Jabłoński, Mirosław; Bubliński, Zbigniew
The paper presents an attempt to utilize the graphics card resources for acceleration of selected image processing operations. The use of tools obtained from NVIDIA for GeForce graphics cards series makes it possible to distribute a computational task on threads executed by GPU floating-point cores. Execution times of parallelized operations were measured and compared with execution time of single-threaded program ran on CPU. Also, the speed-up factors (acceleration factors) were calculated.
Access status: Open Access ,
Comparing parallel programming environments for the joint inversion of geoelectrical data
(Wydawnictwa AGH, 2009) Pięta, Anna; Bała, Justyna
The article presents the comparison of the implementation of the inverse problem in geoelectrical methods in two different parallel computational environments. Combination of Monte Carlo method and Multistart algorithm was applied in the inversion process. Parallelization was done by fine grain decomposition. Execution time, speed-up and efficiency received for parallel algorithms in both computational environments were presented and analyzed.
Access status: Open Access ,
Dynamic tile free scheduling for code with acyclic inter-tile dependence graphs
(Wydawnictwa AGH, 2017) Bielecki, Włodzimierz; Skotnicki, Piotr
Free scheduling is a task ordering technique under which instructions are executed as soon as their operands become available. Coarsening the grain of computations under the free schedule, by means of using groups of loop nest statement instances (tiles) in place of single statement instances, increases the locality of data accesses and reduces the number of synchronization events, and as a consequence improves program performance. The paper presents an approach for code generation that allows for the free schedule for tiles of arbitrarily nested affine loops at run-time. The scope of the applicability of the introduced algorithms is limited to tiled loop nests whose inter-tile dependence graphs are cycle-free. The approach is based on the polyhedral model. Results of experiments with the PolyBench benchmark suite, demonstrating significant tiled code speed-up, are discussed.
Access status: Open Access ,
FPGA-based multi-core processor
(Wydawnictwa AGH, 2013) Wójcik, Wojciech; Długopolski, Jacek
The paper presents the results of investigations concerning the possibility of using programm able logic devices (FPGA) to build virtual multi-core processors dedicated specifically towards particular applications.The paper shows the designed architecture of a multi-core processor specialized to perform a particular task, and it discusses its computational efficiency depending on the number of cores used. An evaluation of the results is also discussed.
Access status: Open Access ,
Heterogeneous GPU&CPU cluster for High Performance Computing in cryptography
(Wydawnictwa AGH, 2012) Marks, Michał; Jantura, Jarosław; Niewiadomska-Szynkiewicz, Ewa; Strzelczyk, Przemysław; Góźdź, Krzysztof
This paper addresses issues associated with distributed computing systems and the application of mixed GPU&CPU technology to data encryption and decryption algorithms. We describe a heterogenous cluster HGCC formed by two types of nodes: Intel processor with NVIDIA graphics processing unit and AMD processor with AMD graphics processing unit (formerly ATI), and a novel software framework that hides the heterogeneity of our cluster and provides tools for solving complex scientific and engineering problems. Finally, we present the results of numerical experiments. The considered case study is concerned with parallel implementations of selected cryptanalysis algorithms. The main goal of the paper is to show the wide applicability of the GPU&CPU technology to large scale computation and data processing.
Access status: Open Access ,
Implementacja równoległa, zastosowania i wyniki metody SPH w przypadku symulacji płynów nieściśliwych
(Wydawnictwa AGH, 2008) Wróblewski, Paweł; Boryczko, Krzysztof
The SPH method for simulating incompressible fluids is presented in the article. The background and principles of the SPH method are explained and its application to incompressible fluids simulations is discussed. Next, two methods of neighbor search for the SPH algorithm based on a constant number of neighbors and a constant cut-off radius are presented. Firts, feasible methods of comparison are analyzed. Then the methods are compared visually and computationally. The obtained results suggest that the method with a constant cut-off radius is better than that with a constant number of neighbors. Subsequently, adaptations of the SPH method for simulating mcompressible fluids which focus on surface tension are presented. The modification for surface tension simulation, which relies on incorporating additional forces into the model, as well as the methodology are suggested. The results obtained with the method are presented and discussed. Finally, the parallel implementation of the SPH simulation with OpenMP environment is demonstrated and analyzed.
Access status: Open Access ,
Komputer równoległy ASGARD przeznaczony do obliczeń w geofizyce
(Wydawnictwa AGH, 2009) Mirek, Janusz
The new parallel computer ASGARD has been designed and built in Department of Geophysics of Faculty of Geology, Geophysics and Environmental Protection AGH University, which contains twenty dual-core processors. The computer is designed for numerical calculation and mathematical modeling. It works under Linux operating system with OpenSSI clustering system. User can find there installed software like compilators, Matlab, Octave, DORIS and other programs written for special purposes.
Access status: Restricted ,
Możliwości implementacji czasochłonnych procedur dla środowiska Matlab
(Data obrony: 2017-01-25) Kopacz, Barbara
Wydział Geologii, Geofizyki i Ochrony Środowiska
W pracy przedstawiono sposoby implementacji czasochłonnych procedur numerycznych dla środowiska Matlab. Rozważono możliwość skompilowania kodu napisanego w C do pliku wykonywalnego mex, skompilowanie kodu źródłowego w C# do biblioteki dll i wywołanie z niej funkcji w Matlabie, użycie programu ImageJ i wtyczki do niego napisanej w Javie, a także zaimplementowanie algorytmu w kodzie skryptowym Matlaba. Dla każdego z wymienionych sposobów przetestowano dwa warianty: dla obliczeń jednowątkowych i równoległych. Badania przeprowadzono implementując algorytm statystycznej dominacji służący do przetwarzania obrazów i porównując uzyskane wyniki.
Access status: Restricted ,
Możliwości implementacji środowisk rozproszonych bazujących na przeglądarkach internetowych
(Data obrony: 2014-10-03) Szemla, Przemysław Witold
Wydział Geologii, Geofizyki i Ochrony Środowiska
The goal of the master’s thesis is a presentation of possibilities of implementation of development environments which allows on creation of distributed applications. The environments should provide the DistributedFor functionality. This mechanism lets on implementation of loops which are executed on many nodes simultaneously. Very important assumption is usage of web browsers, on a client-side, and high-level programming languages such as Java or C#. Moreover, created environments should be characterized by maximal ease of use. The thesis includes also accurate description of algorithms and detailed comparisons of all the aspects of construction of the applications which use technologies mentioned above. Additionally, the author presents, for the .NET based implemention, a solution which allows on parallel execution of code on each node of the system.
Access status: Restricted ,
Możliwości równoległego przetwarzania po stronie przeglądarki internetowej
(Data obrony: 2017-10-06) Rumanek, Rafał
Wydział Geologii, Geofizyki i Ochrony Środowiska
The research work presents an analysis of possibilities of parallel computing in web browser. Despite relatively short history of existance of parallelisation techniques in web applications, the dynamic environment evolution provides increasing number of solutions year on year. The main goal of the work was to explore technologies allowing to reduce the time of algorithm execution by using parallel architecture, to investigate constraints of possible ways of parallelisation and to point out the best suitable mechanism basing on problem specification. Additionaly, the research included performance tests of parallelisation in a web application.
Access status: Open Access ,
One-dimensional fully automatic h-adaptive isogeometric finite element method package
(Wydawnictwa AGH, 2016) Lipski, Paweł; Paszyński, Maciej
This paper deals with an adaptive finite element method originally developed by Prof. Leszek Demkowicz for hierarchical basis functions. In this paper, we investigate the extension of the adaptive algorithm for isogeometric analysis performed with $B$-spline basis functions. We restrict ourselves to $h$-adaptivity, since the polynomial order of approximation must be fixed in the isogeometric case. The classical variant of the adaptive FEM algorithm, as delivered by the group of Prof. Demkowicz, is based on a two-grid paradigm, with coarse and fine grids (the latter utilized as a reference solution). The problem is solved independently over a coarse mesh and a fine mesh. The fine-mesh solution is then utilized as a reference to estimate the relative error of the coarse-mesh solution and to decide which elements to refine. Prof. Demkowicz uses hierarchical basis functions, which (though locally providing $C^{p−1}$ continuity) ensure only $C^0$ on the interfaces between elements. The CUDA C library described in this paper switches the basis to $B$-spline functions and proposes a one-dimensional isogeometric version of the $h$-adaptive FEM algorithm to achieve global $C^{p−1}$ continuity of the solution.
Access status: Open Access ,
Parallel algorithm for sorting animal pedigrees
(Wydawnictwa AGH, 2009) Gierdziewicz, Maciej
In many analyses of animal genotype with the methods of quantitative genetics there is a need to account for relationships among individuals. Incorrectly calculated relationship coefficients may lead to biased estimates. The number of software packages exist which deal with that problem, however, in many of them it is assumed that pedigrees of the individuals are sorted chronologically, but in real data sets - containing information on traits and pedigrees - birth dates are often missing. In extreme cases, when (almost) no birth dates are present, the ordering must be made by comparing - at least once - each pair of individuals separately, since it is not sufficient to compare adjacent elements in order to check whether the data set is sorted. Two versions of parallel computer programs were compared, with constant or variable distance between elements of compared pairs. The results indicate that the second algorithm is more efficient.
Access status: Restricted ,
Profilowanie kodu wykorzystującego OpenCL
(Data obrony: 2017-01-24) Zdziechowski, Adam
Wydział Geologii, Geofizyki i Ochrony Środowiska
W pracy omówiono temat profilowania kodu źródłowego wykorzystującego technologię OpenCL. Opisana została ogólna architektura i nazewnictwo. Następnie przedstawiono niektóre z technik profilowania kodu na przykładzie operacji mnożenia macierzy. Dodatkowo opisane zostały gotowe narzędzia programistyczne usprawniające korzystanie z technologii OpenCL.
Access status: Open Access ,
Równoległy algorytm tabu z elementami inspirowanymi naturą dla problemu planowania tras
(Wydawnictwa AGH, 2011) Pempera, Jarosław
W pracy przedstawiono nową koncepcję konstruowania równoległego algorytmu tabu dla problemu planowania przydziału zadań transportowych i marszrutyzacji pojazdów. Główna idea tej koncepcji polega na wykorzystaniu mechanizmów inspirowanych naturą do zarządzania niezależnymi przebiegami algorytmu tabu. W pracy rozważany jest problem marszrutyzacji z ograniczeniami pracy kierowców. Przeprowadzono badania eksperymentalne mające na celu porównanie algorytmu z algorytmami opartymi na klasycznych metodach lokalnych przeszukiwań.
Access status: Restricted ,
Wykorzystanie systemów równoległych do badania pola falowego Ziemi w ośrodku 3D
(Data obrony: 2010-01-22) Kurcoń, Jakub
Wydział Geologii, Geofizyki i Ochrony Środowiska
The main goal of the work is to create a parallel application to calculate a wave field estabilished in a 3D model and thus to visualize wave propagation in this model. The work consists of a parallel algorithm used for the calculations and Flash Actionscript 3.0 application which is to depict the results. Theoretical overview introduces basic terms which include: a parallel calculation, MPI library, Actionscript 3.0 and Flash Player program and which constitute rudiments of the following deliberations. The work comprises visualizations of wave propagation in the estabilished model and both the discussed results as well as drawn conclusions.