Project CS3103 Operating Systems Version 2.1 (20201021)
1 CS3103 - Operating Systems
Project: Parallel Zip
Due Date: Tuesday, November 24, 2020 at 8 PM HKT. Late submissions will be penalized as per syllabus.
I. Project Instructions
Overview
In the earlier programming assignment, you implemented a simple compression tool based on run-length
encoding, known simply as czip. For this project, you'll implement something similar, except you'll use
threads to make a parallel version of czip. We'll call this version pzip.
There are three specific objectives to this project:
• To familiarize yourself with the Linux pthreads library for writing multi-threaded programs.
• To learn how to parallelize a program.
• To learn how to program for performance.
Project Organization
Each group should do the following pieces to complete the project. Each piece is explained below:
• Design
• Implementation
• Evaluation
30 points
30 points
40 points
You’re required to submit a project report which concisely describes your design, implementation, and
experimental results.
Design
The design space of building an effective compression tool is large. A practical zip that achieve a better
performance will require you to address (at least) the following issues (see part II. Project Description for
more details):
• How to parallelize the compression.
• How to determine the number of threads to create.
• How to efficiently compress multiple files in parallel.
• How to access the input file efficiently.
In your project report, please describe the main techniques or mechanisms proposed to parallelize the
compression. List and describe all the functions used in this project.
2 CS3103 - Operating Systems
Implementation
Your code should be nicely formatted with plenty of comments. Each function should be preceded by a
header comment that describes what the function does. The code should be easy to read, properly indented,
employ good naming standards, good structure, and should correctly implement the design. Your code
should match your design.
Evaluation
We provide 10 test cases for you to test your code. Before submitting your work, please run your pzip on the
test cases and check if your pzip zips input files correctly. Time limitation is set for each test case, and if this
limit is exceeded, your test will fail. For each test case, your grade will be calculated as follows:
• Correctness. Your code will first be measured for correctness, ensuring that it zips input files correctly.
You will receive full points if your solution passes the correctness tests performed by the test script. You
will receive zero points for this test case if your code is buggy.
• Performance. If you pass the correctness tests, your code will be tested for performance; Test script will
record the running time of your program for performance evaluation. Shorter time/higher performance
will lead to better scores.
In your project report, summary and analyze the results. You can also compare your solution with the provided
baseline implementation.
Tips: Keep a log of work you have done. You may wish to list optimizations you tried, what failed, etc.
Keeping a good log will make it easy to put together your final writeup.
Bonus
You’re encouraged to be creative and innovative, and this project award bonus points (up to 10 points) for
additional and/or exemplary work.
• New ideas/designs are welcome to fully explore the parallelism of the process of zipping files.
Additional test cases will be used to test your solution for performance evaluation.
• To encourage healthy competition and desire to improve, we'll provide a scoreboard that shows scores
for each group. The latest scores are displayed, rank ordered, on the scoreboard. We will award bonus
points for top 10 groups. Further details will be posted on Canvas soon.
Language/Platform
The project should be written in ANSI standard C.
This project can be done on Linux (recommended), MacOS, or Windows using Cygwin. Since grading of
this project will be done using gateway Linux server, students who choose to develop their code on any
other machine are strongly encouraged to run their programs on the gateway Linux server before turning it
in. There will be no points for programs that do not compile and run on gateway Linux server, even if they
run somewhere else.
3 CS3103 - Operating Systems
Handing In
The project can be done individually, in pairs, or in groups, where each group can have a maximum of three
members. All students are required to join one project group in Canvas: "People" section > "Groups" tab >
“Project” Group > Group 01–Group 40 or Individual 41–Individual 50. Contact TA to add additional groups
if necessary. Self sign-up is enabled for these groups. Instead of all students having to submit a solution to the
project, Canvas allows one person from each group to submit on behalf of their team. If you work with
partner(s), both you and your partner(s) will receive the same grade for the entire project unless you explicitly
specify each team member's contribution in your report. Please be sure to indicate who are in your group when
submitting the project report.
Before you hand in, make sure to add the requested identifying information about your project group, which
contains project group number, full name and e-mail address for each group member.
When you're ready to hand in your solution, go to the course site in Canvas, choose the "Assignments" section >
"Project" group > "Project" item > "Submit Assignment" button and upload your files, including the following:
1) A PDF document which concisely describes your design, implementation, and experimental results;
If you are working in a team, please also describe each team member's contribution.
2) The source file, i.e., pzip.c;
Academic Honesty
All work must be developed by each group separately. Please write your own code. All submitted source
code will be scanned by anti-plagiarism software. If the code does not work, please indicate in the report
clearly.
Questions?
If you have questions, please first post them on Canvas so others can get the benefit of the TA's answer.
Avoid posting code that will give away your solution or allow others to cheat. If this does not resolve your
issue, contact the TA (Mr. LI Jinheng