What This Document Is
This document contains the introductory lecture materials (Lectures 1 & 2) for CSE 502, a graduate-level Computer Architecture course at Stony Brook University. It serves as a foundational overview of the field, setting the stage for more in-depth exploration of modern computer systems. The lectures, delivered by Larry Wittie, draw upon and adapt materials from David Patterson’s renowned work at UC-Berkeley. This resource is designed to provide a high-level understanding of the core principles and historical context of computer architecture.
Why This Document Matters
This material is essential for graduate students beginning their study of computer architecture, as well as professionals seeking a refresher on the fundamental concepts driving modern computing. It’s particularly valuable at the start of a course or specialization, providing a crucial framework for understanding subsequent topics. Individuals interested in processor design, system performance, and the evolution of computing technology will find this a useful starting point. Accessing the full content will allow for a comprehensive grasp of the challenges and opportunities in the field.
Topics Covered
* The evolving landscape of computer science and its impact on architectural design.
* Distinctions between Computer Architecture and Instruction Set Architecture.
* Historical trends and shifts in conventional wisdom regarding performance bottlenecks.
* The impact of power constraints and transistor density on architectural choices.
* The emergence of multi-core processors and the challenges of parallel programming.
* Analysis of uniprocessor performance trends over time.
* A look at the historical context of multiprocessor systems.
What This Document Provides
* An outline of the course’s introductory themes and objectives.
* A discussion of key performance metrics and their historical evolution.
* Visual representations of performance trends, illustrating the changing priorities in computer design.
* A comparative analysis of different processor generations and their characteristics.
* Insights into the challenges associated with scaling performance in modern computer systems.
* A foundational understanding of the shift from Instruction Level Parallelism to Thread Level Parallelism.
