Large Scale High Performance Systems

Catalog description

This course introduces the concept of high performance computing for big data processing and analysis. This course covers high performance computing tools including distributed system concept, practical techniques and applications. Students will learn distributed machine learning system with tools in various structures and utilize them for data processing and analysis through lectures, hands-on laboratories, and projects. Topics that will be included in this course are high performance computing tools, distributed system structures, open source data tools and implementation, data analysis, and more. Typically offered Fall Spring Summer. Prerequisite(s): ITS 36500 FOR LEVEL UG WITH MIN. GRADE OF C Course Learning Outcomes 1. Describe the historical evolution of computer architecture including parallel paradigm. 2. Compare and contrast the Implicit Parallelism such as pipelining, multithreading, and etc. with the limitations of memory system performance. 3. Explain and implement the task dependency graph, task interaction graph, and task decomposition. 4. Explain and implement the communication models in parallel programming such as broadcast, reduction, circular shift and etc. 5. Explain and implement a parallel program (MPI) that performs I/O. 6. Understand the Distributed system architecture and discuss the characteristics. 7. Understand and explain Hadoop ecosystem. 8. Explain communication protocols, system architecture, node configuration, cluster management, and etc. 9. Understand and implement HDFS, Yarn, Apache Sqoop, and MapReduce. View Class Schedule