|
|
A Tool for Transformation of PDF to Text
Bujok, Jonáš ; Raab, Jan (advisor) ; Falt, Zbyněk (referee)
Title: A Tool for Transformation of PDF to Text Author: Jonáš Bujok Department: Institute of Formal and Applied Linguistics (32-UFAL) Supervisor: Mgr. Jan Raab, Institute of Formal and Applied Linguistics (32-UFAL) Abstract: In this thesis we described an extraction procedure of text information from PDF (Portable Document Format) files. Thesis is focused mainly on middle-Europe languages. We designed, described and implemented program for this purpose. Besides the program and it's description the thesis contains information about PDF format object structure, it's syntax and logic necessary for proper understanding of text searching principles in PDF file. We also discussed filters, fonts and all other PDF Objects that the program need to process. This thesis also deals with methods and possibilities of improving program's functionality, speed, memory usage, reliability an universality of usage.
|
|
|
Persistent data structures in C
Hájek, Jakub ; Kopecký, Michal (advisor) ; Falt, Zbyněk (referee)
Persistent data structures in C Abstract: This work describes the design and implementation of persistent data structures framework for C programming language. It discusses effective imple- mentation of persistent structures and providing transparent manipulation of data. This work defines services which should be provided by persistent layer and tries to define some general persistent layer interface for good extensibility. In the end it analyses performance issues of various data storages and tries to specify class of applications, which could use persistent data structures framework. Keywords: persistent tree structures, C, data structures
|
|
|
Suitable methods of data prepration for mutistream compression
Jelínek, Jan ; Žemlička, Michal (advisor) ; Falt, Zbyněk (referee)
This work focuses on compression algorithm called multistream compression. We have tried to improve this algorithm to be competitive with other commonly used compression algorithms. We have also tried to combine it with well-known transformations like BWT or MTF to efficiently handle text files. Another part of the research was compose characters into bigger parts and transform the result by multistream compression. We have tried blocks of constant length like bigrams and blocks of variable length like syllables or words. Last experiments was with small changes of MSC algorithm and compare them with the original algorithm.
|
|
|
Cache-oblivious Algorithms
Vaner, Michal ; Mareš, Martin (advisor) ; Falt, Zbyněk (referee)
In this work, we study the cache-oblivious computation model, which is inspired by the behaviour of the memory hierarchy of current computers. We study several graph algorithms and techniques of their design in this model. We consider graph searching, identifying connected components and computing maximal matching. We also study sorting and matrix multiplication as subproblems of many graph algorithms. In ad- dition to previously known algorithms, we present several new ones. We study their efficiency both by the means of asymptotic complexity and by benchmarking them on real hardware and we compare them with classical algorithms.
|
|
|
Improving efficiency of HEP applications
Horký, Jiří ; Zavoral, Filip (advisor) ; Falt, Zbyněk (referee)
The Large Hadron Collider (LHC) located at CERN, Geneva has finally been put in production, generating unprecedented amount of data. These data are distributed across many computing centers all over the world that form the Worldwide LHC Computing Grid (WLCG). One of the main issues since the beginning of the WLCG project is an effective file access on the site level in order to fully exploit huge computing farms. The aim of this thesis is to explore existing data distribution work flows, standards, methods and protocols. An integral part of the work is the analysis of jobs of physicists to understand input/output workloads and to discover possible inefficiencies. Then, new upcoming solutions are evaluated in terms of performance, sustainability and integration into existing frameworks. It is expected that these solutions will be based on distributed file systems such as NFS 4.1, Lustre and HDFS.
|
|
|
Scheduler and memory allocator for the Bobox system
Falt, Zbyněk ; Yaghob, Jakub (advisor) ; Kruliš, Martin (referee)
The content of this work is the description of the Bobox system, which is system for parallel data processing developed by the Department of Software Engineering of the Charles University, the analysis of its original scheduler and the design of a new, more efficient task scheduler and memory allocator of the system. The work also contains an experimental comparison of different aproaches to the problematic.
|
|
|
Displaying images in JPEG format
Falt, Zbyněk ; Hubička, Jan (referee) ; Mareš, Martin (advisor)
The content of this work is the description of techniques and algorithms, which accelerate loading of sequential JPEG images. To reach this, special methods, such as the elimination of arithmetic computation, the reduction of conditional jumps in the most critical sections, the use of the vector instruction and the multithreaded preloading of images, are used. These techniques are platform independent, except the usa of SSE2 vector instructions, which are supported only on the IA-32 platform. The attached computer program demonstrates their real implementation and their efficiency.
|
guest ::
