README.md

Task Scheduler

A multithreaded task scheduler implemented in C++23, featuring a thread pool and a lock-based concurrent queue. The project includes benchmarking examples for simple summation, Monte Carlo estimation of Pi, and matrix multiplication, with speedup visualization.

Features

• Thread Pool: Efficiently manages a pool of worker threads for parallel task execution
• Concurrent Queue: Thread-safe priority queue for task scheduling
• Speedup Visualization: Prints tables comparing execution time and speedup for different thread counts
• Examples:
  • Benchmarking Simple summation (examples/simple/main.cpp)
  • Benchmarking Monte Carlo Pi estimation (examples/monte_carlo_pi/main.cpp)
  • Benchmarking Matrix multiplication (examples/matrix_multiplication/main.cpp)
  • Priority Tasks (examples/priority_tasks/main.cpp)

Building

This project uses CMake (version 3.25 or newer) and requires a C++23-compatible compiler.

git clone https://github.com/codingwithmagga/task_scheduler
cd task_scheduler
cmake -B build -S .
cmake --build build

Executables for the examples will be built in the build/examples/ subdirectories.

Running Examples

After building, run any of the example executables:

# Simple summation example
./build/examples/simple/simple_example

# Monte Carlo Pi estimation
./build/examples/monte_carlo_pi/monte_carlo_pi

# Matrix multiplication
./build/examples/matrix_multiplication/matrix_multiplication

# Priority Tasks
./build/examples/priority_tasks/priority_tasks

Each of the first example prints a table showing the speedup achieved with different thread counts which looks like this:

   Threads      Time (ms)       Speedup
----------------------------------------
         1           1642           1.00
         2            887           1.85
         4            516           3.18
         6            417           3.94
         8            350           4.69
        10            314           5.23
        12            300           5.47

Note that the matrix multiplication example does not show the expected speedup. This issue could be caused by different reasons, but I assume inefficient memory use and caching issues is a problem in this example.

The last example shows which tasks are executed first, based on their priority

Run task with priority: HIGH
Run task with priority: HIGH
Run task with priority: HIGH
Run task with priority: HIGH
Run task with priority: HIGH
Run task with priority: MEDIUM
Run task with priority: MEDIUM
Run task with priority: MEDIUM
Run task with priority: LOW
Run task with priority: LOW

License

This project is licensed under the GNU General Public License v3.0. See LICENSE for details.