What This Document Is
This document comprises lecture materials from CDA 6938, a graduate-level course at the University of Central Florida focusing on Multi-Core and Many-Core Architectures and Programming. It provides a foundational overview of parallel computing, exploring the shift from traditional single-core processors to the increasingly prevalent multi-core and many-core systems used in high-performance computing. The material delves into the principles behind leveraging these architectures for significant performance gains.
Why This Document Matters
This resource is ideal for graduate students in computer science, electrical engineering, or related fields who are seeking a deep understanding of parallel processing. It’s particularly valuable for those interested in high-performance computing, GPU programming, and the challenges of modern processor design. Students preparing for advanced work in areas like computer architecture, compiler design, or parallel algorithms will find this material highly relevant. It’s best utilized as a core component of a course on parallel architectures or as a focused study aid for related topics.
Topics Covered
* The evolution and motivations behind multi-core and many-core processors.
* Comparative analysis of CPU and GPU architectures.
* Overview of key GPU technologies from AMD/ATI and NVIDIA.
* Exploration of programming models for GPGPU (General-Purpose computing on Graphics Processing Units).
* Data-level and thread-level parallelism concepts and their application.
* Architectures like the IBM Cell BE processor.
* Future trends in multi-core/many-core architectures and programming.
What This Document Provides
* A comprehensive course outline detailing the scope of study.
* Discussion of the course prerequisites and grading policies.
* Insights into the research areas of the instructor.
* A list of suggested (optional) textbooks and supplementary materials.
* Information regarding available laboratory resources and access procedures.
* Acknowledgement of source materials and related courses.
* A historical perspective on the challenges driving the shift towards concurrency in computing.