EE7720       Advanced Computer Architecture

 

Instructor: Dr. Lu Peng

Office: EE 233, Phone: 578-5535, Email: lpeng@lsu.edu

 

 

Course Objectives and Outline

 

This course teaches students fundamental knowledge in Computer Architecture and Microarchitecture. Though titled with ‘advanced’, it is an entry-level graduate course and best suitable for the first year Ph.D. or Master students who are interested in computer architecture, compilers, operating systems, programming languages, VLSI designs, application specific architectures. Many exciting topics will be covered:

 

  • Fundamentals of Computer Design
    • Performance, Power, Reliability, Security etc.
  • Instruction Level Parallelism (ILP) and Its Exploitation
    • Pipelining
    • Compiler Techniques for Exposing ILP
    • Hardware-Based Speculation
    • Out-of-order execution, Multiple issue
    • Dynamic branch prediction
  • Memory Hierarchy Design
    • Advanced Optimizations of Cache Performance
    • Virtual Memory and Virtual Machines
  • Multiprocessors and Thread-Level Parallelism
    • Symmetric Shared-Memory Architectures (SMP)
    • Distributed Shared Memory and Directory-Based Coherence (DSM)
    • Synchronization
  • Case studies
    • Intel Pentium 4, Core 2 Duo
    • Sun Niagara
    • AMD Opteron
    • IBM Cell processor

 

Required Text

Computer Architecture: A Quantitative Approach. 4th Ed.
        John L. Hennessy, David Patterson
        ISBN: 978-0-12-370490-0
        ISBN10: 0-12-370490-1  
        Publisher: Morgan Kaufmann Publishers
        Pub. Date: 2006

 

Prerequisite

            Computer organization EE 3755 or equivalent

 

Recommended Reading

Computer Organization and Design, 3rd Ed.

                 David A. Patterson, John L. Hennessy

                 ISBN: 1558606041, Pub. Date: 2004
                 Publisher: Morgan Kaufmann Publishers

            Modern Processor Design: Fundamentals of Superscalar Processors

                 John Paul Shen, Mikko H. Lipasti
                 ISBN 0-07-057064-7

                 Publisher: McGraw-Hill Publishers, 2005

Grading Policy

Homework and Project (45%)   Midterm exam (25%)   Final exam (30%)

Bonus (10%): I will randomly give 10 simple quizzes at the end of 10 classes. The major purpose of quizzes is to check your attendance.

Final grades will be given by curve. A: top 55%. B: students meeting the minimal requirements but not in top 55%. C, D or F: not meeting the minimal requirements.

 

To be fair to all students, Homework and Projects should be handed in on time. You will lose 20 points (based on 100 points for each homework or project) for each day late until lose all points after five days. You must make sure to join the midterm and final exam. If you have any necessary reason to reschedule your exams, let me know and show me your evidences.