|
|
Traveling Salesman Problem: Parallel Methods Using SMP (Threads)
Weigner, Martin ; Jaroš, Jiří (referee) ; Kašpárek, Tomáš (advisor)
This thesis is focused on solving the problem of traveling salesman. At first, the problem is solved by serial access at four algorithms. There are two of them chosen and transferred to parallel access. In the end there are summarized observations about different parameters of both access. This thesis also introduces questions of programming parallel applications with threads to the reader.
|
|
|
Longest Prefix Match Algorithms
Weigner, Martin ; Puš, Viktor (referee) ; Tobola, Jiří (advisor)
The speed of computer network is increasing. One of the basic tasks which has to be solved by the network devices is longest prefix match. Many algorithms are able to solve this task but it's necessary to solve it very fast because of increasing transfer speed and the routing standard IPv6. This standard has longer addresses and it's necessary to search prefixes in much bigger sets. The thesis describes present algorithms that solve the problem. These are confronted with new algorithm HashTreeBitmap which is also described in the work. On the tests is documented that this algorithm is able to cope with high demands which are placed on it.
|
|
|
Traveling Salesman Problem: Parallel Methods Using SMP (Threads)
Weigner, Martin ; Jaroš, Jiří (referee) ; Kašpárek, Tomáš (advisor)
This thesis is focused on solving the problem of traveling salesman. At first, the problem is solved by serial access at four algorithms. There are two of them chosen and transferred to parallel access. In the end there are summarized observations about different parameters of both access. This thesis also introduces questions of programming parallel applications with threads to the reader.
|
|
|
Longest Prefix Match Algorithms
Weigner, Martin ; Puš, Viktor (referee) ; Tobola, Jiří (advisor)
The speed of computer network is increasing. One of the basic tasks which has to be solved by the network devices is longest prefix match. Many algorithms are able to solve this task but it's necessary to solve it very fast because of increasing transfer speed and the routing standard IPv6. This standard has longer addresses and it's necessary to search prefixes in much bigger sets. The thesis describes present algorithms that solve the problem. These are confronted with new algorithm HashTreeBitmap which is also described in the work. On the tests is documented that this algorithm is able to cope with high demands which are placed on it.
|
guest ::
