|
|
Signal processing using parallel mathematical operations
Polášek, Jaromír ; Ležák, Petr (referee) ; Mžourek, Zdeněk (advisor)
This Bachelor thesis deals with the acceleration of function calculations, using parallel computing mediated by NVDIA graphics cards via CUDA technology. The theoretical part describes the general principles of parallel computing and the basic characteristics and parameters of graphics cards NVDIA. The theoretical part also deals with basic principles of CUDA technology. End of the theoretical part focuses on FFTW and cuFFT libraries. The practical part deals with the comparison of the performance between GPU and CPU functions filter2D and Canny and practical possibilities of accelerating fast convolution calculation. The practical part also describes sample code that was used to compare the performance between GPU and CPU. The results of this program are then plotted and evaluated.
|
|
|
Semi - analytical computations and continuous systems simulation
Kopřiva, Jan ; Kubátová, Hana (referee) ; Novitzká,, Valerie (referee) ; Kunovský, Jiří (advisor)
The thesis deals with speedup and accuracy of numerical computation, especially when differential equations are solved. Algorithms, which are fulling these conditions are named semi-analytical. One posibility how to accelerate computation of differential equation is paralelization. Presented paralelization is based on transformation numerical solution into residue number system, which is extended to floating point computation. A new algorithm for modulo multiplication is also proposed. As application applications in solution of differential calculus are the main goal it is discussed numeric integration with modified Euler, Runge - Kutta and Taylor series method in residue number system. Next possibilities and extension for implemented residue number system are mentioned at the end.
|
|
|
Residue Number System Based Building Blocks for Applications in Digital Signal Processing
Younes, Dina ; Brzobohatý, Jaromír (referee) ; Vlček, Čestmír (referee) ; Šteffan, Pavel (advisor)
Předkládaná disertační práce se zabývá návrhem základních bloků v systému zbytkových tříd pro zvýšení výkonu aplikací určených pro digitální zpracování signálů (DSP). Systém zbytkových tříd (RNS) je neváhová číselná soustava, jež umožňuje provádět paralelizovatelné, vysokorychlostní, bezpečné a proti chybám odolné aritmetické operace, které jsou zpracovávány bez přenosu mezi řády. Tyto vlastnosti jej činí značně perspektivním pro použití v DSP aplikacích náročných na výpočetní výkon a odolných proti chybám. Typický RNS systém se skládá ze tří hlavních částí: převodníku z binárního kódu do RNS, který počítá ekvivalent vstupních binárních hodnot v systému zbytkových tříd, dále jsou to paralelně řazené RNS aritmetické jednotky, které provádějí aritmetické operace s operandy již převedenými do RNS. Poslední část pak tvoří převodník z RNS do binárního kódu, který převádí výsledek zpět do výchozího binárního kódu. Hlavním cílem této disertační práce bylo navrhnout nové struktury základních bloků výše zmiňovaného systému zbytkových tříd, které mohou být využity v aplikacích DSP. Tato disertační práce předkládá zlepšení a návrhy nových struktur komponent RNS, simulaci a také ověření jejich funkčnosti prostřednictvím implementace v obvodech FPGA. Kromě návrhů nové struktury základních komponentů RNS je prezentován také podrobný výzkum různých sad modulů, který je srovnává a determinuje nejefektivnější sadu pro různé dynamické rozsahy. Dalším z klíčových přínosů disertační práce je objevení a ověření podmínky určující výběr optimální sady modulů, která umožňuje zvýšit výkonnost aplikací DSP. Dále byla navržena aplikace pro zpracování obrazu využívající RNS, která má vůči klasické binární implementanci nižší spotřebu a vyšší maximální pracovní frekvenci. V závěru práce byla vyhodnocena hlavní kritéria při rozhodování, zda je vhodnější pro danou aplikaci využít binární číselnou soustavu nebo RNS.
|
|
|
Parallelization of complex tasks in reconstruction of dynamic magnetic resonance
Bijotová, Kateřina ; Rajmic, Pavel (referee) ; Mašek, Jan (advisor)
This thesis deals with parallelization of complex tasks in reconstruction of dynamic magnetic resonance. It describes the basic principle of magnetic resonance and its relation to Fourier transform. It deals with the difference between static and dynamic magnetic resonance image reconstruction. It analyzes SVD algorithm and its use in magnetic resonance image reconstruction. It presents the principles and the importance of parallel computing in magnetic resonance imaging and describes CUDA technology. The thesis also contains a description and execution of the implementation of the reconstruction model in MATLAB and Java programming language which were optimized by JCuda library for Java implementation and gpuArray function in case of MATLAB implementation.
|
|
|
GPU-Accelerated Design of Optically Generated Ultrasound Using Binary Amplitude Holograms
Knotek, Martin ; Vaverka, Filip (referee) ; Jaroš, Jiří (advisor)
In this thesis, we deal with the possibilities of the acceleration of scientific computations using the graphical processing unit. The term scientific computation in this context means an algorithm, which computes binary holograms that are used to generate ultrasound. We will concentrate specifically on the design of the hologram, focusing at the speed we can achieve when computing the surface of the hologram. For this purpose, we will use two popular parallel data processing platforms - CUDA and OpenMP. The surface design pattern of the hologram is important due to the fact, that it determines the hologram’s specific physical characteristics.
|
|
|
Parallelization of complex tasks in reconstruction of dynamic magnetic resonance
Bijotová, Kateřina ; Rajmic, Pavel (referee) ; Mašek, Jan (advisor)
This thesis deals with parallelization of complex tasks in reconstruction of dynamic magnetic resonance. It describes the basic principle of magnetic resonance and its relation to Fourier transform. It deals with the difference between static and dynamic magnetic resonance image reconstruction. It analyzes SVD algorithm and its use in magnetic resonance image reconstruction. It presents the principles and the importance of parallel computing in magnetic resonance imaging and describes CUDA technology. The thesis also contains a description and execution of the implementation of the reconstruction model in MATLAB and Java programming language which were optimized by JCuda library for Java implementation and gpuArray function in case of MATLAB implementation.
|
|
|
Techniques for parallel computing
Vodák, René ; Hasmanda, Martin (referee) ; Lattenberg, Ivo (advisor)
The text of this thesis deals with techniques of parallel processing calculations. It is an analysis of the most important libraries for parallelization including libraries for parallelization on GPU graphics cards and computing speed by comparing these libraries in Visual Studio 2010 based on a simple application searching primes on three different computer hardware configurations. With OpenCL library, that achieved the best result, there are formed two applications – an improved program for searching prime numbers using the sieve of Eratosthenes and a program for calculating the integral with the trapezoidal rule.
|
|
|
Multicriteria graph partitioning
Houška, Ondřej ; Tůma, Miroslav (advisor) ; Hnětynková, Iveta (referee)
The thesis is about graph partitioning and applications of graph partitioning in paral- lel algorithms for solving big sparse linear equations. The problem of graph partitioning is thorougly described and standard graph partitioning algorithms are explained. The appli- cation part is focusing on the Conjugate Gradient method preconditioned by a variant of incomplete Cholesky factorization based on drop tolerance. The role of graph partitioning in the problem decomposition is described and a load balancing problem is studied. 1
|
|
|
Reduced communication algoritms: theory and practice
Slevínský, Rostislav ; Tůma, Miroslav (advisor) ; Rozložník, Miroslav (referee)
Development in the parallel computing environment in the last decade comes with the need of being able to use these in solving large algebraic systems. In this thesis, we focus on the Krylov subspace methods (namely the conjugate gradient method) as one of the most powerful tools and the possibilities of their parallelization. We discuss the communication avoiding Krylov subspace methods and various problems introduced by the parallelization e.g. loss of orthogonality or delay of convergence. Application of the Krylov subspace methods comes usually with some preconditioner, therefore part of this thesis is dedicated to the preconditioning in parallel computing environments.
|
|
|
Parallelization of complex tasks in reconstruction of dynamic magnetic resonance
Bijotová, Kateřina ; Rajmic, Pavel (referee) ; Mašek, Jan (advisor)
This thesis deals with parallelization of complex tasks in reconstruction of dynamic magnetic resonance. It describes the basic principle of magnetic resonance and its relation to Fourier transform. It deals with the difference between static and dynamic magnetic resonance image reconstruction. It analyzes SVD algorithm and its use in magnetic resonance image reconstruction. It presents the principles and the importance of parallel computing in magnetic resonance imaging and describes CUDA technology. The thesis also contains a description and execution of the implementation of the reconstruction model in MATLAB and Java programming language which were optimized by JCuda library for Java implementation and gpuArray function in case of MATLAB implementation.
|
guest ::
