CS 272 Software Development

Table of Contents Requirements
Grading
Getting Started

For this project, you will extend your previous project to support multithreading. In addition to meeting the previous project requirements, your code must use a thread-safe inverted index using a reentrant conditional read/write lock if multithreading is enabled.

When multithreading is enabled, your code must use work queues to manage threads. The build should be multithreaded such that each worker thread processes a single file. The search should be multithreaded such that each worker thread processes a single multi-word query line.

To do this, you should create new classes (or extend existing classes) to support multithreading. DO NOT REMOVE YOUR SINGLE THREADING CODE, as you still need to support single threaded building and searching the index.

Do NOT use any of the classes in the java.util.concurrent package and do NOT use the Stream.parallel method for the multithreaded code.

Requirements

The following detail the functionality requirements that must be implemented for this project.

Input Requirements

Your main method must be placed in a class named Driver and must process the following command-line arguments:

• -threads [num] where the flag -threads enables multithreading, and indicates the next optional argument [num] is the number of worker threads to use. If the [num] argument is not provided, not a number, or less than 1, use 5 as the default number of worker threads.

  If the -threads flag is not provided, then multithreading should not be enabled. A thread-safe inverted index should not be initialized, a work queue should not be initialized, no worker threads should be created, and the project should execute with a single main thread exactly as previous projects.

The command-line flag/value pairs may be provided in any order or not at all. Do not convert paths to absolute form when processing command-line input!

Output user-friendly error messages in the case of exceptions or invalid input. Under no circumstance should your main() method output a stack trace to the user!

Your code should support all of the command-line arguments from the previous project as well.

Simple Read/Write Lock

If multithreading is enabled, the inverted index must be made thread-safe using a simple conditional read/write lock. The conditional read/write lock must support concurrent read operations, non-concurrent write and read/write operations, and allow an active writer to acquire additional read or write locks as necessary (i.e. reentrant write lock).

The thread-safe inverted index should not be used unless multithreading is enabled. The original version without locking must be used when single threading.

Simple Work Queue

If multithreading is enabled, your code must use a work queue to manage threads. The work queue must support the ability to automatically track unfinished (or pending) work (or tasks), and provide a method that waits until there is no more pending work.

The work queue should also support the ability to shutdown gracefully, and should be gracefully shutdown after all building and searching operations are complete.

The work queue should not be initialized unless multithreading is enabled. It is possible to reuse the same work queue for both building and searching if multithreading is enabled.

Multithreaded Building

If multithreading is enabled, your code must support multithreaded building of the thread-safe inverted index using a work queue. Each worker thread should process one file at a time (i.e. create one “task” per file).

Multithreaded Searching

If multithreading is enabled, your code must support multithreaded searching of the thread-safe inverted index using a work queue. Each worker thread should process one multi-word query line at a time (i.e. create one task per line).

Output Requirements

The output of your inverted index and search results should be the same from the previous project.

As before, your code should only generate output files if the necessary flags are provided. If the correct flags are provided, your code should perform the indexing and search operations even if file output is not being generated.

Run Examples

The following are a few examples (non-comprehensive) to illustrate the usage of the command-line arguments that can be passed to your Driver class via a “Run Configuration” in Eclipse, assuming you set the working directory to the project-tests directory.

Consider the following example:

-text "input/text/simple/" -query "input/query/simple.txt" -results actual/search-exact-simple.json -threads 3

The above arguments behave the same as project 2, except use 3 worker threads in a work queue to multithread both the building and search.

-text "input/text/simple/" -query "input/query/simple.txt" -results actual/search-exact-simple.json -threads

The above arguments are nearly the same, except use the default of 5 worker threads.

-text "input/text/simple/" -query "input/query/simple.txt" -threads

The above arguments are similar, except it does NOT produce any file output. The code should still build the index and calculate the search results! This is important for benchmarking your code.

-text "input/text/simple/" -query "input/query/simple.txt" -results actual/search-exact-simple.json

This should behave exactly the same as project 2, using single-threading without any worker threads.

Grading

This project grade is split into the following assignments:

¹ You should start on project 4 after receiving a grade for Project 3 Review 2—do NOT wait to finish this project to start the next one!

² It takes approximately 3 to 4 code reviews to pass the design of this project. Keep in mind you may have only 30 minutes worth of code review appointments every 5 days. Due to weekends, that works out to approximately 3 weeks or more of code reviews.

See the Project Grading guide for details on how projects are graded.

Multithreading Tests

The project tests primarily check the file output of your code. Therefore, it is difficult to detect whether your code is actually multithreading as required. The tests can only detect is whether your code initializes a work queue, not whether that work queue is properly used.

Submitting code for a test grade that is not multithreading using a work queue and conditional read/write lock will result in a one-time 10 to 20 point deduction to the test score. It is up to you to manually verify that:

• Your code is creating tasks as required (one per file for building, one per line for searching) and adding those tasks to a work queue to be executed.

• Your code is utilizing a conditional read/write lock to protect access to your inverted index data structure.

You are strongly encouraged to use logging to verify tasks are being created and run as intended.

Benchmarking Tests

Unlike previous projects, the tests for this project are broken in two: Project3aTest.java and Project3bTest.java.

For the test grade and first code review, your code only needs to pass the tests in the Project3aTest.java group of JUnit tests. These tests verify your code still produces the correct file output.

For the following code reviews and the final release of this project, your code must also pass the Project3bTest.java group of JUnit tests. These tests benchmark your single-threaded versus multithreaded code and make sure your multithreaded version is faster.

For code reviews, this includes getting a speedup of at least 1.1x faster on average using multithreading. For the final release, the speedup should be at least 1.5x faster.

Getting Started

The following sections may be useful for getting started on this project.

Related Homework

The following homework assignments may be useful for this project:

• The LoggerSetup homework is useful for learning how to set up and configure log4j2, which will be helpful when it comes to debugging multithreaded code.

• The ReadWriteLock homework is useful for creating a simple conditional read/write lock. The conditional lock class can be used directly for your project. It also illustrates how to use a conditional lock to make a data structure class thread-safe.

• The PrimeFinder homework is useful for creating a work queue that tracks pending work. This work queue class can be used directly for your project. It also illustrates how to use this work queue and create tasks for non-recursive problems.

You can modify homework assignments as necessary for this project. However, make sure your code is passing all of the tests before using.

You should NOT wait until you have completed all of the associated homework assignments to start the project. You should develop the project iteratively as you progress throughout the semester, integrating assignments one at a time into your project code.

Related Content

The following lecture content may be useful for this project:

• The synchronization lecture code illustrates how to use a conditional read/write lock to make a data structure class thread-safe.

• The work queues lecture code illustrates how to use a work queue and create tasks for recursive problems. If your approach is not recursive, this example might not be helpful for this project.

You can use and modify lecture code as necessary for this project. However, make sure you understand the concepts before using the code.

You should NOT wait until you have covered all of the associated lecture content to start the project. You should develop the project iteratively as you progress throughout the semester, integrating concepts one at a time into your project code.

Suggestions

Your goal should be to get to testable code as quickly as possible first, and then developing iteratively to pass the functionality tests.

One possible breakdown of tasks are:

• Configure log4j2 add debug messages in your code. Once you are certain a class is working, disable debug messages for that class in your log4j2.xml configuration file.

• Extend your previous inverted index to create a thread-safe version using a conditional read/write lock.

• Create new code to build the index using a work queue (creating one task per file). Make sure your code still passes the tests.

• Create new code to process the query file using a work queue (creating one task per query line). Make sure your code still passes the tests.

• Test your code in a multitude of settings and with different numbers of threads. Some issues will only come up occasionally, or only on certain systems.

• Test your code with logging enabled. Then, test your code with logging completely disabled. Your code will run faster without logging, which sometimes causes some concurrency problems.

Do not worry about efficiency until after your first code review. However, if you have had one code review and are running into issues, look for the following common issues:

• Make sure the code does not over-notify (e.g. does not wake up threads more often than necessary).

• Make sure the code does not over-finish (e.g. does not call finish more often than necessary).

• Make sure the code does not over-block (e.g. does not perform blocking operations within a loop).

• Make sure the code does not over-synchronize (e.g. does not prevent operations that could occur at the same time).

These issues are best detected using logging.

It is important to get started early so you have plenty of time to think about how you want to approach the project and start coding iteratively. Planning to complete the code in too large of a chunk is a recipe to get stuck and fall behind!

 These hints may or may not be useful depending on your approach. Do not be overly concerned if you do not find these hints helpful for your approach for this project.