Machine architectures define interfaces - specifications of what can be done and how to make those things happen. Architectures are typically thought of as the lowest level interface in a system, the one defined by the hardware. Complicated software behavior is obtained by layering many interfaces on top of this lowest level one - e.g., the operating system, the language runtime, a set of libraries/packages, and the application code.This course typically looks down from the architecture interface, concentrating on how to build hardware systems that implement it efficiently. We will do that as well, following the book to concentrate on critical components of the hardware organization, in particular the central processing unit (CPU) and memory subsystem.
More than in most instances of 378, though, I'd like to spend a little time also looking up from the architecture interface, to explain the role of other system components (e.g., compilers, linkers, and the operating system), how their jobs are affected by the architecture, and what the architecture must do to support them.
By the end of the quarter, you should understand not only the fundamentals of how machines are built (and why they're built that way), but also the fundamentals of how computer systems operate. (And you may have learned a bit of C along the way as well.)
CSE 370 - Especially computer organization and some of the material in the finite state machines topic. In contrast to CSE 370, though, phrases like "high impedance" will never come up in this course.Java / C - I'll assume that you know Java. I don't expect we'll be doing any Java programming, though.
On the other hand, although I'll assume you don't know C, we will be doing some programming in a derivative language, C--. Most of C-- is roughly a subset of Java (as is a lot of C), so the meaning of simple programs should be clear the first time you see them. The most important difference between C--/C and Java is that the former include pointer variables. The meaning of pointers, and how to use them, will be made clear by the content of this course - you don't have to know anything about them now.
Cygwin - The project infrastructure requires a Window system to run, but relies on a set of GNU tools, primarily make. It will be useful, but not absolutely necessary, to have had some familiarity with Unix.
There will be homework. In all cases the goal of the homework is to help motivate you to learn the material, and to help verify that you've done so. The majority of the homeworks will be done usingSMOK/Cebollita
, and are cumulative: each successive homework builds upon the previous one, either directly, as a modification of an earlier machine design, or more indirectly, making use of the key concepts from the earlier homework.Most of the homeworks will be done in (randomly assigned) teams.
I'd like this to be an interactive class. Interaction with the instructor and the TAs in class and during office hours is highly encouraged. "Appropriate" interaction with each other, say through the class mail list, is also encouraged. (See the Policies section for some guidance on what "appropriate" might mean.)
There will be two mideterms. The dates on which they will be held could change (e.g., if we get off schedule in the lectures).
- Midterm I: Monday, April 20 (Date subject to change.)
- Midterm II: Monday, May 18 (Date subject to change.)
- Final Exam: 8:30-10:20am, Monday, June 8
Grades will be assigned roughly as follows:
- Homeworks: 40%
- Midterms: 10/15%
- Final: 30%
- Section quizzes & course participation: 5%
- Late Policy: unless otherwise indicated, assignments and projects are due by the beginning of lecture on their due date. If you hand in an assignment late, we will take off 20% for each day (or portion thereof) it is late. However, you have two extension days that you can use as you please to avoid the late penalty. If you want to use one, make sure that it is clear to the TA you are doing so.
- Cheating vs. Collaboration: Let me just state up front that this is an unresolvable issue. Collaboration, as an aid to learning, is highly encouraged. Cheating is just as highly discouraged.
The university more or less requires that each course present you (students) with a clear definition of unacceptable behavior. Here it is for this course: cheating is any effort made to earn points without actually doing the work those points represent or understanding the material to the degree those points indicate. Not cheating is anything you do that helps you understand the course material (not just the solution to a single problem, but the underlying information on which the solution rests). If you're unsure about any particular activity, please ask.