(Latest Revision: 11/02/2014)
Parallel Bubble Sort
- Read the file HiHo.cpp (actually HiHo.cpp.html) in the ThreadInfo
directory. It illustrates everything you need to know about using
function calls from the pthreads facility on the CS Lab Mac's, and about
using the queuing semaphores implemented by the sem.cpp and sem.h files.
You may make copies of ThreadInfo/sem.cpp and ThreadInfo/sem.h
for your own use, or the identical copies of sem.h and sem.cpp
available in the set of resources accompanying
this assignment. (I'll try to explain a bit more about this in class.)
- Working by yourself, use queuing semaphores (as provided by sem.h and
sem.cpp) to implement a "parallel bubble sort" program that reads
a list of 32 integers from standard input, puts them into an array,
sorts them, and writes the sorted array to standard output. If you
have difficulty, ask me for help in office hours. Don't look at
other people's solutions to this problem.
You must use this program skeleton
as a starter, and solve the problem by adding semaphore variables
and semaphore function calls to the skeleton.
You should make a separate directory on a CS Department Lab
Macintosh computer running OS X, place copies of Makefile,
bubble.cpp, sem.h, and sem.cpp in the directory, and use the
single command make to compile the program.
Don't make any changes at all to the Makefile, to sem.h, or to sem.cpp.
Just make changes to bubble.cpp as required. (Read the rest of
these directions for more details about what you need to do to
the bubble.cpp file.)
Code in bubble.cpp causes the main thread (the mother)
to spawns 31 "child" threads to do interchanges in the data
array called cell[].
The child threads are assigned serial numbers from 1 to
31. These serial numbers are stored in a particular field of a
data structure associated with each thread (idPtr->id). Thread #i
is in charge of interchanges between cell #i-1 and cell #i in the
array.
You will need to add a declaration of an array of semaphores
to bubble.cpp, one semaphore for each data cell. Then you will have to
add other code that satisfies these two goals:
- enforce mutually exclusive access to each cell
in the data array, and
- promote as much parallel processing as possible
of the cells in the data array
Your program may designate one thread
(it can be the mother) to check to determine when the sorting has
been finished, and to print out the sorted list afterwards.
The skeleton bubble.cpp program is set up that way. (See the second
half of the function called mother()).
If you can figure out a way to make the child threads do the
checking and/or printing more symmetrically, you may alter the
code so it does that instead.
All threads must be synchronized appropriately to insure correct
operation and results, mutual exclusion where required, progress,
and bounded waiting. At the same time, the code must maximize the
amount of parallel processing that can go on during the run of the
algorithm.
It's very important that you understand the ideas in the requirements
just described above. Make sure to ask questions about things you
don't understand. I'll try to bring some things up in class and
explain them, but it's your responsibility to try to understand
and ask questions about whatever does not seem clear.
In the finished program, the mother task must not exit until after
the list has been sorted and printed. (However, you don't have to
guarantee that the mother waits for all the child threads to exit first
before she exits.)
- At the beginning of class on the due date, place in front of me a
printout of a complete, compilable, listing of your modified
bubble.cpp file, properly commented, formatted, and easy to read.
Do not give me copies of the Makefile, sem.h, or sem.cpp.
I want only the code you write. Make sure your name is on the
printout.
- Before midnight on the due date, e-mail me a copy of your bubble.cpp
source code. Use this subject line: "CS3750ProgOne"
As I explained above, I want just the C++ code you wrote.
Please do not e-mail me copies of the Makefile, sem.h, or sem.cpp.
Also don't send me any compiled code
- only the source code, bubble.cpp.
Don't send me anything as an attachment.
I won't accept it. Send the code inline with
your e-mail message. From one of the CS Lab Macs,
you can use a command like this:
mail -s "CS3750ProgOne" john@ishi.csustan.edu < bubble.cpp
My e-mail address is:
john@ishi.csustan.edu.
I will compile the program you send me on one of the CS Lab Macs, with my
own copies of sem.h and sem.cpp. Your program is required to work
(correctly) under those testing conditions.
- Beware: since some errors in concurrent programs are timing-dependent,
you will need to think of some novel ways to test your program. The file
HiHo.cpp illustrates this. I will compile and run your program as a part
of my grading procedure. I may vary the timing of the threads by
inserting code in your program. The code will delay threads for random
or non-random amounts of time. If this "breaks" your code, it means your
code is not a correct solution to the problem. You may ask for more
details on this subject in class.
DUE DATES:
For the program due date, see
the class schedule.