Streaming system scheduling for Xeon Phi
Název práce v češtině: | Plánování v proudových systemech na Xeonu Phi |
---|---|
Název v anglickém jazyce: | Streaming system scheduling for Xeon Phi |
Klíčová slova: | plánování, proudové systémy, Bobox, Xeon Phi |
Klíčová slova anglicky: | scheduling, streaming systems, Bobox, Xeon Phi |
Akademický rok vypsání: | 2015/2016 |
Typ práce: | diplomová práce |
Jazyk práce: | angličtina |
Ústav: | Katedra softwarového inženýrství (32-KSI) |
Vedoucí / školitel: | RNDr. Jakub Yaghob, Ph.D. |
Řešitel: | skrytý - zadáno a potvrzeno stud. odd. |
Datum přihlášení: | 27.10.2015 |
Datum zadání: | 04.11.2015 |
Datum potvrzení stud. oddělením: | 17.07.2017 |
Datum a čas obhajoby: | 06.09.2017 09:00 |
Datum odevzdání elektronické podoby: | 20.07.2017 |
Datum odevzdání tištěné podoby: | 21.07.2017 |
Datum proběhlé obhajoby: | 06.09.2017 |
Oponenti: | doc. RNDr. Martin Kruliš, Ph.D. |
Zásady pro vypracování |
Task scheduling in operating system area is well known and well understood problem on traditional system architectures (SMP, NUMA). Unfortunately, current scheduling techniques and algorithms do not perform well on emerging many-core architectures (Intel Xeon Phi).
The first goal of the thesis is a research about a real many-core architecture (Intel Xeon Phi). Although there are books and papers about it, the preliminary experiments show, the internal architecture does not match the presented architecture in a literature. The internal architecture should be more clarified by a set of experiments. The second goal follows the first one. We have an experimental streaming system Bobox. It performs very well on traditional system architectures, unfortunately its performance on many-core architecture is terrible. The second goal is an improvement/evolution of Bobox task scheduling algorithm for many-core architecture (Intel Xeon Phi). |
Seznam odborné literatury |
Falt Z., Kruliš M., Bednárek D., Yaghob J., Zavoral F.: Towards Efficient Locality Aware Parallel Data Stream Processing, in Journal of Universal Computer Science, 20151
Falt Z., Bednárek D., Kruliš M., Yaghob J., Zavoral F.: Bobolang - A Language for Parallel Streaming Applications, in Proceedings of the 23rd International ACM Symposium on High-Performance Parallel and Distributed Computing, Vancouver, ACM, ISBN: 978-1-4503-2749-7, pp. 311-314, 20141 Bednárek D., Kruliš M., Malý P., Yaghob J., Zavoral F., Pokorný J.: Combining Distributed Computing and Massively Parallel Devices to Accelerate Stream Data Processing, accepted for publication in Proceedings of the The Seventh International Conference on Advances in Databases, Knowledge, and Data Applications, Řím, IARIA, pp. 1-8, 20151 Bednárek D., Dokulil J., Yaghob J., Zavoral F.: Bobox: Parallelization Framework for Data Processing, accepted for publication in Advances in Information Technology and Applied Computing, Denpasar, AITAC, ISSN: 2251-3418, pp. 189-194, 20121 Jeffers J., Reinders J.: Intel Xeon Phi Coprocessor High-Performance Programming, ISBN 978-0124104143, Morgan Kaufmann, 2013 Wang E,, Zhang Q., Shen B., Zhang G., Lu X., Wu Q., Wang Y.: High-Performance Computing on the Intel® Xeon Phi(TM): How to Fully Exploit MIC Architectures, ISBN 978-3319064857, Springer, 2014 |