Employing Parallel Computing in Data-Intensive Tasks
| Thesis title in Czech: | Použití paralelních výpočtů v datově-intenzivních úlohách |
|---|---|
| Thesis title in English: | Employing Parallel Computing in Data-Intensive Tasks |
| Key words: | paralelní výpočty, CUDA, GPU, datová analýza, data mining, vysoce výkonné počítání, strojové učení |
| English key words: | parallel computing, CUDA, GPU, data analytics, data mining, high performance computing, machine learning |
| Academic year of topic announcement: | 2019/2020 |
| Thesis type: | dissertation |
| Thesis language: | angličtina |
| Department: | Department of Software Engineering (32-KSI) |
| Supervisor: | doc. RNDr. Martin Kruliš, Ph.D. |
| Author: | hidden - assigned and confirmed by the Study Dept. |
| Date of registration: | 08.09.2020 |
| Date of assignment: | 08.09.2020 |
| Confirmed by Study dept. on: | 05.10.2020 |
| Guidelines |
| In the past two decades, mainstream hardware has experienced a significant shift towards parallelism.
Multicore CPUs as well as manycore GPUs are present in commodity PCs and laptops and specialized massively parallel accelerators (such as Xeon Phi devices) have emerged in the high performance computing domain. Unfortunately, most applications and algorithms of the day are not ready for such a radical change and cannot be easily ported to parallel platforms. The main objective of this thesis is to investigate algorithms, data structures, and methods employed in the domains of data analytics, intensive real-time (big) data processing, optimization, and machine learning in order to adopt them for parallel platforms of the day, especially GPUs. Particular attention will be payed to computationally intensive problems, which are solved solely by fully automated processing, and which will benefit the most from the potential speedup of parallel hardware. The expected outcome of this research will be twofold: - Selected algorithms will be implemented and measured for speedup. The evaluation prototypes could be subsequently used as basis for libraries and software tools which will help speeding up data processing in various empirical scientific domains. - Lessons learned in these algorithms will provide better understanding of the behavior of massively parallel hardware and optimization guidelines. They will be used in computer science education and will form a basis for generalized methodology in massively parallel computing and fine-grained optimizations. |
| References |
| Tan, Pang-Ning, Michael Steinbach, and Vipin Kumar. Introduction to data mining. Vol. 1. Boston: Pearson Addison Wesley, 2006.
Russom, Philip. "Big data analytics." TDWI Best Practices Report, Fourth Quarter (2011). Wu, Ren, Bin Zhang, and Meichun Hsu. "GPU-accelerated large scale analytics." IACM UCHPC (2009). |