|
|
Praktické datové struktury
Pokorný, Michael ; Mareš, Martin (advisor) ; Babka, Martin (referee)
In this thesis, we implement several data structures for ordered and unordered dictionaries and we benchmark their performance in main memory on synthetic and practical workloads. Our survey includes both well-known data structures (B-trees, red-black trees, splay trees and hashing) and more exotic approaches (k-splay trees and k-forests). Powered by TCPDF (www.tcpdf.org)
|
|
|
Applications of Gray codes in cache-oblivious algorithms
Mička, Ondřej ; Fink, Jiří (advisor) ; Gregor, Petr (referee)
Modern computers employ a sophisticated hierarchy of caches to decrease the latency of memory accesses. This led to the development of cache-oblivious algorithms that strive to achieve the best possible performance on such memory hierarchies with minimal knowledge of the exact parameters of the hierarchy. A common technique used in the design of cache-oblivious algorithms is a recursion-based divide-and-conquer method. In this work, we show an alternative technique based on the Gray codes. We use the binary reflected Gray code to traverse arrays in the cache-oblivious way, allowing us to design algorithms for problems such as matrix transposition, naive matrix multiplication or naive convolution that match the asymptotic performance of their recursion-based counterparts. The advantage is that our algorithms can be implemented without recursion (or a stack that simulates it) by using a loopless algorithm. We also introduce a variant of the binary reflected Gray code tuned to certain applications of our technique and an almost loopless algorithm to generate it. Apart from the theoretical analysis of our technique's performance, we also examine its practical performance on the problem of matrix transposition.
|
|
|
Compact I/O-Efficient Graph Representations
Tětek, Jakub ; Gavenčiak, Tomáš (advisor) ; Mareš, Martin (referee)
The objective of this thesis is to develop a fast memory-efficient representa- tion of some graphs that occur in real-world applications. We consider separable graph classes (e.g. planar graphs or graphs of bounded genus) and show how to represent them in a way that (1) makes accessing vertices in a walk cache-efficient on average and (2) is highly memory-efficient. In particular, we show a compact representation of separable graph classes with the I/O cost of a random walk of length k being O(K/(Bw)1−c ) w.h.p. In the second part of the thesis, we consider layout of trees with optimal worst-case I/O cost for root-to-leaf traversal, show an additive (+1)-approximation of I/O optimal compact layout and contrast this with a proof of NP-hardness of exact solution. In this thesis, we also prove generalisations of the recursive separator theo- rem. The first one generalises the theorem for weighted graphs and the second one replaces minimum region size by average region size in the bound. 1
|
|
|
Praktické datové struktury
Pokorný, Michael ; Mareš, Martin (advisor) ; Babka, Martin (referee)
In this thesis, we implement several data structures for ordered and unordered dictionaries and we benchmark their performance in main memory on synthetic and practical workloads. Our survey includes both well-known data structures (B-trees, red-black trees, splay trees and hashing) and more exotic approaches (k-splay trees and k-forests). Powered by TCPDF (www.tcpdf.org)
|
guest ::
