Basic Course Information:

• Instructor: Dean Tullsen
• CSE 3216
• tullsen at cs dot ucsd dot edu
• 534-6181
• office hours: M 3-4, W 2-3, or by appt
• TA: Craig Disselkoen
• cdisselk at eng dot ucsd dot edu
• office hours Tu 10-11, W 10-11
• office hours location, CSE B250A
• TA: Andreas Prodromeau
• aprodrom at eng dot ucsd dot edu
• office hours TBD (after Nov 1)
• office hours location, TBD
• Meeting times and place:
• Center Hall 105
• lecture: Tu Th 3:30-4:50 p.m.
• Course textbook
• Hennessy & Patterson, "Computer Architecture: A Quantitative Approach", Fifth Edition, Morgan Kaufmann
• We'll also be reading some research papers as well as referencing some books from the outstanding series Synthesis Lectures on Computer Architecture. Those books will be free to access while you are on the UCSD domain. If you're interested in computer architecture, you should be browsing these books.
• Other background reading
  • Patterson & Hennessy, "Computer Organizatin and Design:  The Hardware/Software Interface" gives a lower-level treatment of the material (more from the design standpoint).  This represents the assumed background/prerequisite for this class.
• Let me emphasize that the prerequisite material for this course is an undergraduate class that covered ISA design, pipelining, caches, etc. If you do not have the prerequisites, you should not be in this class.
• Assignments will be of three types:
• traditional homeworks, e.g., problems from the book, typically with 1-week notice.
• small projects, typically involving simulation and analysis tools
• reading summaries for papers (I will try to give one week notice on reading assignments. History indicates I don't always succeed.)
• Everyone in the class is expected to be on Piazza. It will be the repository for lecture slides, homework and project assignments, etc. Please sign up (if not already) here: sign up, and access the course page here: course page.

I. Introduction

II. Computer System Performance

III. Instruction Set Architecture

IV. Pipelining

V. The Memory/Cache Hierarchy

VI. Instruction-Level Parallelism

VII. Parallel Machines

• The grade for 240 will be based on homeworks, projects, one midterm, and a final, as follows (this is subject to change, based on the number of projects):
• pretest: 5%
• homeworks (no more than 20% total), projects, reading summaries: 30%
• midterm: 25%
• final: 40%
• subjective influences like class participation will have an impact in the margins.
• The final will be inclusive of all course material.
• Late assignments are not encouraged. Late assignments will receive partial credit -- they may or may not be graded. We will make every effort to return assignments to you in a timely manner -- limiting your ability to turn things in late is, unfortunately, critical to that goal.
• You have the right of appeal for grading on all tests. You have one week from the time the midterms are returned to make appeals, including addition errors on your score. Check it over carefully when you get it.
• There is no appeal on homeworks, except for addition errors. No single problem will have a significant impact on your grade.

• I expect we have the highest standards of integrity in a graduate class. However, we will check aggressively for cheating, including copying off of other people or from online resources, etc. If cheating is found, of any kind, severe consequences will result.