CMSC 411: Computer Architecture

Building a Better Reference Base---NOT Current for Fall-2008

Currently, the 411 webpage is aesthetically lovely, but rather short on content. Your job is to remedy the situation by producing potential technical and practical resources for other students to draw upon. Topic submissions can focus on an architecture not covered in the book, or in providing more detail on an architecture covered in the book. A reasonable topic would also be a concept in the book which is difficult to understand, such as...

• normalizing the frequency of instruction execution, or
• understanding the differences among register-register, register-memory, and memory-memory ISA's.

Now, as stated previously, you may work on these in a group of from two to four people. However, the scope of the accepted proposal, and the expectations from the particular project will be adjusted according to the size of the group. However, all group members must be listed on the proposal, with preferred email address, and SSN last 4 digits. While single-person proposals will be accepted under special circumstances, the desired minimum group size is two (2). Why? Because you have to produce several kinds of information for a minimal project, as described below. And, it will help each of you to write questions and have the other person answer and evaluate them. The topic submission is worth 30% of your grade.

Dates:

Topic submissions, in the form of electronic (email text preferred) or neatly formatted documents from a good quality printer, of at most 2 pages in length will be accepted beginning immediately. All initial submissions must be received no later than the beginning of class, Monday, August 7, 2000. Unless you hear back from me within 1 week, you may assume that your proposal has been accepted. Late topic submissions garner an automatic 10% deduction from the 30 points out of 100 that the submission is worth, based on their arriving at some arbitrary time after 2PM -- where I decide when late is too late on that day and post it. Each 24 hours late garners another 10 percent (aka 3pt) deduction. I will provide feedback on accepted proposals as I get to them.

Goal:

The general idea is to produce short documents of a tutorial nature that would provide information not found in Hennessey & Patterson or that would further explain difficult material, provide detailed work-ed examples, and provide a worksheet with solutions for the reader to practice a new skill. Alternatively, the document could add more current information to that found in HP, or explore prior architectures, whether successful commercially or not, and discuss their contribution to current computing. In all cases, questions and answers tying the document to the material in the course must be produced, and a self-test section for the user, with answers, must be provided. Check out the ones that are there already. (The list will be updated within the next two weeks; so, I'll let you know when new ones appear.)

Format:

The general format of the web-page to-be, including required sections, etc. will be released later, or choose your own, as long as the required info is there somewhere. If you don't know html, or don't know any html authoring tools, either find a partner who does, learn it, or state in your proposal that you will be turning in only a paper document (no deduction for paper only, so long as I know in advance). I'm looking for webpages that are heavier on content than looks. Any group of more than 2 is required to produce an on-line product. (Note that the most recent version of Netscape has an editor in it that even I managed to use to make the class Syllabus. [editor: She used Netscape 3.0. More recent versions have an even more robust editor.] )

Make sure to list the members of your group. Full names, email, and last 6 Student ID digits. Then, indicate the main contact and email address for communication from me.

First, the proposal should include a title for your topic that captures the content and intent of the document, along with four other key words, not in the title, that could be used to characterize what the reader will learn from reading and performing the attached exercises. Snappy or catchy titles can replace the pedestrian title you'll use now. I just need enough to get an idea of what you think you're doing.

Next should be a description of your topic, what you're going to focus on, or perhaps what questions you're trying to ask. Sort of like an extended thesis statement, except you tell the assumptions and what the result is, as in an abstract.

Third is a few sentences explaining what you want the reader to take away with them either from reading your text, or doing your problems. That is, tie what you're doing to the book, or even to some other sub-field of computer science, or to several of each -- this is important. That is, don't go for too narrow a focus here, unless you are teaching a skill needed everywhere, like Amdahl's law.

At least one reference in the form of a web address (current URL) and another document or paper-based source besides H&P must be included to demonstrate that you have sufficient documentation to draw on, unless your topic tries to expand on points not well explained, or provide how-to's for doing problems in H&P.

Fourth should be an outline of the document you intend to write. This can be modified prior to submission of the final product (and will not be graded here -- it's just for me to check that you understand what I want), and can be the format I give you (by the end of the week too, I hope) or one of your own choosing (if you choose not to wait for me).

Note that there are several existing pages from prior student projects. Unless you plan to substantially improve on one of the topics already on the class webpage, the topics currently dealt with there are considered "taken".

Topics:

Topics are on a first come-first served basis. However, if you are number 2, I'll help modify your topic to make it different, or suggest a grouping to improve the final product. I know a lot of the places where students have problems, and will include them in the list. What I write will be what I expect from a two-person group unless otherwise stated.

I'll be brainstorming and have a list of suggested topics available upon request. In the interim, feel free to surf and select your own.