A collection of header-only C++ libraries implementing common parallel processing design patterns, including process pools, thread pools, and pipelines.

---

ProcessPool

A lightweight, robust C++ library for Linux designed to manage high-volume process execution with strict resource throttling. This ProcessPool handles spawning, monitoring, and cleaning up child processes while ensuring your system isn't overwhelmed by "fork bombs" or orphaned zombies.

---

🚀 Features

• Fixed-Size Worker Pool: Spawn $N$ total tasks while ensuring no more than $M$ are active at any given time.
• Real-time Monitoring: Uses poll() and Unix pipes to detect worker completion or crashes instantly.
• Fail-Fast Architecture: Automatically triggers a global shutdown (SIGTERM) of all active workers if a single process fails, preventing inconsistent state.
• Zombie Prevention: Proactively reaps finished processes using waitid and O_CLOEXEC to keep the process table clean.
• Parent-Death Suicide Pact: Utilizes PR_SET_PDEATHSIG so that if the parent manager dies unexpectedly, all child processes are automatically terminated.

---

🛠 How It Works

The ProcessPool acts as a governor for your application's subprocesses:

1. Throttling: It iterates through the total child processes count, but blocks if the active child count reaches maxSimultaneous.
2. Communication: Each child is born with a pipe. It reports a ChildStatus struct (exit code and health bit) just before exiting.
3. Cleanup: The ProcessPool continuously monitors these pipes. If a pipe closes without a "healthy" report, the ProcessPool a crash and halts the entire pool.

---

💻 Usage Example

Integrating the ProcessPool into your project is straightforward. Define a worker function and tell the ProcessPool how many times to run it.

#include"ProcessPool.hpp"// Define the task for the child processesintMyHeavyTask(int id)
{
std::cout << "Worker " << id << " is processing..." << std::endl;
sleep(2); // Simulate workreturn0; // Child process exit code
}
intmain()
{
// Total Tasks: 10, Max Simultaneous: 3
TaskPool pool;
pool.Create(MyHeavyTask, 10, 3);
return0;
}

---

Pipe

A thread-safe C++ implementation of a pipe allowing multiple threads to concurrently write data to one end and multiple threads to concurrently read data from the other. It provides a mechanism for ordered, asynchronous data transfer in high-throughput parallel applications.

---

ThreadPool

Simple and efficient c++ implementation of thread pool design pattern. Check example.cpp for detailed example.

Check my another repository, dircpy, to see ThreadPool in action.