Runtime Resource Management -- Paper Review Project

For this final project, you need to read all four of the following papers:

1. Automatic vs. Explicit Memory Management: Settling the Performance Debate, by Hertz and Berger.

2. Page-Level Cooperative Garbage Collection, by Hertz, Feng, and Berger.

3. MC²: High-Performance Garbage Collection for Memory Constrained Environments, by Sachindran, Moss, and Berger.

4. Autonomic Heap Sizing: Taking Real Memory Into Account, by Yang, Berger, Hertz, Kaplan, and Moss.

Note that these papers were all produced by the same group of researchers, so there may be some similaries that you would not expect to see in a broader sample of writing. However, the papers are quite different, and the text is largely written by different people.

Each of you has also been assigned a specific task with regards to one of these papers: You must argue for or against it. The two longer papers have been assigned an additional ``freelance'' reader who may take either side.

For the paper to which you are assigned, you must also write a review of that paper. Specifically, you should provide the following in your review:

1. A brief summary
2. Does the paper address an important topic?
3. Does the paper appropriate address prior work?
4. Is the experimental evaluation correct and thorough?
5. Is the presentation well structured and the writing clear?

For more information on how to write a review, you should read Smith's The Task of The Referee. This piece covers somewhat more than you need to know, but it will provide you with some useful perspective on how you should approach the review-writing task.

You should place a copy of your task in the directory below by Wednesday, 5-May-2004, at 11:59 pm. (Note that a plain text file is just fine -- no need for word processing, LaTeX, PDF files, etc.). Make sure that the file is group-readable (e.g. chmod g+r my-review.txt). The directories should be placed in:

/scratch/cluster0-d/cs40/cs40/paper-reviews

After everyone has posted their reviews, you should read all the other reviews. Specifically, you should come to class on Friday, 7-May-2004 ready to discuss all four papers. DO NOT LEAVE THE TASK OF DISCUSSING A PAPER ONLY TO ITS ASSIGNED REVIEWERS!

Note, finally, that I will not be participating in the discussions. My role will be only to guide the discussions or, if needed, to answer a few questions if I have special knowledge (about related work, etc.) Our goal is to provide feedback to the authors, particularly for papers that we reject as being not ready for publication. We must be ready to give specific criticism of writing, content, and experimental evaluation.