CSC 543 -
Multiprocessing & Concurrent Programming, Spring 2024,
Assignment 1 on intro to thread safety.
Assignment 1 due by 11:59 PM on Thursday February 15 via make
turnitin
Dataflow diagram added to the bottom of
this page on February 3, 2024
I added an example doc on using
control-\ to see stack traces for a hung Java process
on 2/6.
cd $HOME
# places you into your login directory
mkdir multip
# all of your csc543 projects go into this directory
cd ./multip
# makes multip your current working directory
cp
~parson/multip/CSC543s24DataflowAssn1.problem.zip
CSC543s24DataflowAssn1.problem.zip
unzip CSC543s24DataflowAssn1.problem.zip
# unzips your working copy of the project directory
cd ./CSC543s24DataflowAssn1
# your project working directory
Perform all test execution on mcgonagall to avoid any
platform-dependent output differences.
Also, large input and output files for your code reside in my file
system to avoid overloading yours.
Here are the files of interest in this project directory. The ones
through TupleExchange.java require
modification. You may need to understand the others.
CSC543s24DataflowAssn1.java
# Pipeline builder and main.
Do this first to see make test explode.
MakeUniform.java
# The three Make* classes makes a
uniform, normal, & exponential
MakeNormal.java
# distribution respectively
of pseudo-random numbers. They require
MakeExponential.java
# added
field modifiers to make them thread-safe + some annotations.
TenTuple.java
# A
simple, immutable data container.
FiveTuple.java
#
Slightly more complicated immutable data container.
DistributionGenerator.java
# 1st stage active class that
invokes one of the Make* generators above.
StatisticalAnalysisGenerator.java
# 2nd stage
active class that analyzes / reduces stage 1's output.
StatisticalAnalysisCSVSaver.java
# 3rd stage in
pipeline writes entries to a CSV file for diffing.
TupleExchange.java
# Implicit
& explicit locks to synchronize pipeline data handoff.
A mutex
(lock) with no condition variable and a mutex
with a condition variable.
IMakeDistribution.java
#
Interface for the above Make* classes.
Stats.java
# Basic statistics & CSV writing because Java
default libraries are not great.
makefile & makelib
# make files for make clean test
and make turnitin.
CSC223f23DataflowAssn4.py
#
Python from last semester is the reference
design for this project.
CSC223f23DataflowAssn4.ref
# Reference
files ending in .ref are the expected output CSV files.
CSC543s24multi.ref
CSC543s24single.ref
jcip-annotations.jar
# Annotations for the textbook denote
design intent.
$ jar tvf jcip-annotations.jar
0 Fri Dec 04 18:31:48 EST 2009 META-INF/
102 Fri Dec 04 18:31:46 EST 2009 META-INF/MANIFEST.MF
0 Fri Dec 04 18:31:48 EST 2009 net/
0 Fri Dec 04 18:31:48 EST 2009 net/jcip/
0 Fri Dec 04 18:31:48 EST 2009
net/jcip/annotations/
436 Fri Dec 04 18:31:48 EST 2009
net/jcip/annotations/GuardedBy.class
430 Fri Dec 04 18:31:48 EST 2009
net/jcip/annotations/Immutable.class
438 Fri Dec 04 18:31:48 EST 2009
net/jcip/annotations/NotThreadSafe.class
432 Fri Dec 04 18:31:48 EST 2009
net/jcip/annotations/ThreadSafe.class
Source files requiring your work contain STUDENT comments in upper
case.
$ grep -l STUDENT *.java
CSC543s24DataflowAssn1.java
DistributionGenerator.java
FiveTuple.java
IMakeDistribution.java
MakeExponential.java
MakeNormal.java
MakeUniform.java
StatisticalAnalysisCSVSaver.java
StatisticalAnalysisGenerator.java
Stats.java
TenTuple.java
TupleExchange.java
Stats.java does not require changes, but it has an example line of
code useful in
improving FiveTuple.java's thread safety.
Use make clean test to compile & test your code
changes.
Testing must occur on mcgonagall.
$ ssh YOURLOGIN@mcgonagall # You must
log in from acad.
You can edit on acad.
Use make turnitin and hit
the Enter key when prompted by the project deadline.
Late penalty is 10% per day, and I cannot accept a late
assignment after I
go over the solution. At that point it is worth 0%.
We will walk through the spring
2023 solution to their Assignment 2 as example code.
Their solution is under
~parson/multip/prisonerd2j2023.solution.zip and prisonerd2j2023/.
A note from last year's Assignment 1:
Some basic guidelines for making Assignment 1's classes thread
safe:
1. Make private any data field that can be private.
2. Inert data containers like TenTuple can have public
fields i.f.f. those fields
are immutable, for example primitive data
types, or references to immutable
objects such as String or most (all?) java.lang.Number
objects.
Mutable objects such as arrays should be private
so as not to escape their objects.
3. Fields should be final when possible. Making a
reference to a mutable object final
does not make it immutable, but it does
guarantee flushing to main memory when the
constructor returns. A mutable field should be
either volatile, or have every access
guarded by an implicit or explicit lock,
or have a final reference to a thread-safe
object from package
java.util.concurrent.atomic. We will cover atomic later.
4. Use the textbook @Immutable, @ThreadSafe, @NotThreadSafe, and
@GuardedBy annotations as required in STUDENT
comment to communicate
design intent.
Here is what make clean test looks like from the handout
directory:
$ make clean test
/bin/rm -f *.o *.class .jar core *.exe *.obj *.pyc
__pycache__/*.pyc
/bin/rm -f junk* *.pyc
/bin/rm -f *.tmp *.o *.dif *.out __pycache__/*
/bin/rm -f DEBUG*arff* tmp*
/bin/rm -f *.class *.csv
PYTHONPATH=/home/kutztown.edu/parson:.:..:/usr/local/lib/python3.7
time /usr/local/bin/python3.7 CSC223f23DataflowAssn4.py 543343458
CSC223f23DataflowAssn4.csv
4.15user 1.78system 0:02.83elapsed 209%CPU (0avgtext+0avgdata
25672maxresident)k
0inputs+16outputs (1major+15215minor)pagefaults 0swaps
diff --ignore-trailing-space --strip-trailing-cr
CSC223f23DataflowAssn4.csv CSC223f23DataflowAssn4.ref >
CSC223f23DataflowAssn4.dif
/bin/bash -c "CLASSPATH=..:./jcip-annotations.jar
/usr/bin/javac MakeUniform.java"
/bin/bash -c "CLASSPATH=..:./jcip-annotations.jar
/usr/bin/javac MakeNormal.java"
/bin/bash -c "CLASSPATH=..:./jcip-annotations.jar
/usr/bin/javac TupleExchange.java"
/bin/bash -c "CLASSPATH=..:./jcip-annotations.jar
/usr/bin/javac MakeExponential.java"
/bin/bash -c "CLASSPATH=..:./jcip-annotations.jar
/usr/bin/javac CSC543s24DataflowAssn1.java"
bash -c "CLASSPATH=..:./jcip-annotations.jar time /usr/bin/java
CSC543s24DataflowAssn1.CSC543s24DataflowAssn1 343458543 test1.csv
0"
2.08user 0.21system 0:01.11elapsed 206%CPU (0avgtext+0avgdata
300160maxresident)k
0inputs+96outputs (0major+9793minor)pagefaults 0swaps
head -1 test1.csv > CSC543s24single.csv
grep uniform test1.csv >> CSC543s24single.csv
grep normal test1.csv >> CSC543s24single.csv
grep exponential test1.csv >> CSC543s24single.csv
diff --ignore-trailing-space --strip-trailing-cr
CSC543s24single.csv CSC543s24single.ref > CSC543s24single.dif
bash -c "CLASSPATH=..:./jcip-annotations.jar time /usr/bin/java
CSC543s24DataflowAssn1.CSC543s24DataflowAssn1 343458543 test2.csv
1"
0.12user 0.04system 0:00.14elapsed 115%CPU (0avgtext+0avgdata
33972maxresident)k
0inputs+72outputs (0major+6104minor)pagefaults 0swaps
head -1 test2.csv > CSC543s24multi.csv
grep uniform test2.csv >> CSC543s24multi.csv
make: *** [testjava] Error 1
My handout single-main-thread test passes, but the multi-threaded
solution won't pass
until you complete your work. A successful test does not guarantee
that all requirements
are met because a race condition that allows tests to pass
sometimes is possible.
Please re-read all STUDENT requirements before turning it in. When
make test
does not report an error, testing has succeeded.
$ cat CSC543s24multi.csv
Distribution,Param1,Param2,Count,Mean,Median,Mode,Pstdev,Min,Max
The multi-threaded output is missing its data. That is only a
header line.
Use make turnitin and hit the Enter key when prompted by
the project deadline.
Late penalty is 10% per day, and I cannot accept a late assignment
after I
go over the solution. At that point it is worth 0%.
Here is a successful test run:
$ make clean test
/bin/rm -f *.o *.class .jar core *.exe *.obj *.pyc
__pycache__/*.pyc
/bin/rm -f junk* *.pyc
/bin/rm -f *.tmp *.o *.dif *.out __pycache__/*
/bin/rm -f DEBUG*arff* tmp*
/bin/rm -f *.class *.csv
PYTHONPATH=/home/kutztown.edu/parson:.:..:/usr/local/lib/python3.7
time /usr/local/bin/python3.7 CSC223f23DataflowAssn4.py 543343458
CSC223f23DataflowAssn4.csv
4.33user 1.82system 0:03.16elapsed 194%CPU (0avgtext+0avgdata
25648maxresident)k
0inputs+16outputs (1major+14870minor)pagefaults 0swaps
diff --ignore-trailing-space --strip-trailing-cr
CSC223f23DataflowAssn4.csv CSC223f23DataflowAssn4.ref >
CSC223f23DataflowAssn4.dif
/bin/bash -c "CLASSPATH=..:./jcip-annotations.jar
/usr/bin/javac MakeUniform.java"
/bin/bash -c "CLASSPATH=..:./jcip-annotations.jar
/usr/bin/javac MakeNormal.java"
/bin/bash -c "CLASSPATH=..:./jcip-annotations.jar
/usr/bin/javac TupleExchange.java"
/bin/bash -c "CLASSPATH=..:./jcip-annotations.jar
/usr/bin/javac MakeExponential.java"
/bin/bash -c "CLASSPATH=..:./jcip-annotations.jar
/usr/bin/javac CSC543s24DataflowAssn1.java"
bash -c "CLASSPATH=..:./jcip-annotations.jar time /usr/bin/java
CSC543s24DataflowAssn1.CSC543s24DataflowAssn1 343458543 test1.csv
0"
2.13user 0.19system 0:01.10elapsed 210%CPU (0avgtext+0avgdata
300208maxresident)k (single threaded)
0inputs+96outputs (0major+10107minor)pagefaults 0swaps
head -1 test1.csv > CSC543s24single.csv
grep uniform test1.csv >> CSC543s24single.csv
grep normal test1.csv >> CSC543s24single.csv
grep exponential test1.csv >> CSC543s24single.csv
diff --ignore-trailing-space --strip-trailing-cr
CSC543s24single.csv CSC543s24single.ref > CSC543s24single.dif
bash -c "CLASSPATH=..:./jcip-annotations.jar time /usr/bin/java
CSC543s24DataflowAssn1.CSC543s24DataflowAssn1 343458543 test2.csv
1"
2.25user 0.22system 0:00.91elapsed 271%CPU (0avgtext+0avgdata
302388maxresident)k (multi-threaded)
0inputs+104outputs (0major+9917minor)pagefaults 0swaps
head -1 test2.csv > CSC543s24multi.csv
grep uniform test2.csv >> CSC543s24multi.csv
grep normal test2.csv >> CSC543s24multi.csv
grep exponential test2.csv >> CSC543s24multi.csv
diff --ignore-trailing-space --strip-trailing-cr
CSC543s24multi.csv CSC543s24multi.ref > CSC543s24multi.dif
When that last diff does not report an error, it is working.
Multi-threading does not always accelerate this test run. We will
work with acceleration later.
For now we are working on the basics of thread safety and using
locks.
Dataflow
Diagram for this Assignment 1:
