Summary  
This chapter covers process-based parallelism in Python using the multiprocessing module: creating, starting, and joining processes, along with interprocess communication through queues.  

General domain of usage  
CPU-bound tasks such as data crunching

Picture your Python process as a restaurant franchise, not just a single kitchen. In multiprocessing, you don't just hire more chefs for one kitchen—you actually open several kitchens in different buildings. Each kitchen (process) has its own chef, fridge, stove, and even its own 'Master Knife'—there's no sharing. This means Chef A in Kitchen A and Chef B in Kitchen B can both chop, cook, and serve at the same time, truly in parallel, because they're not waiting for a shared resource like the GIL in threading.

This is the core advantage of multiprocessing: each process runs in its own memory space, so they never bump into each other. It's perfect for heavy tasks such as mathematical calculations, image processing, or data crunching, where real parallelism can speed things up dramatically.

However, if Chef A in Kitchen A needs to pass an ingredient to Chef B in Kitchen B, it's not as easy as just handing it over the counter. They need a delivery truck—this is Inter-Process Communication (IPC). IPC is slow and expensive compared to sharing ingredients in the same kitchen, so you should only use it when necessary.

Let's see how this looks in Python using the multiprocessing module. First, you'll see how to launch two processes that each perform a task in parallel:

```python
from multiprocessing import Process
import time

def print_numbers(name):
    for i in range(1, 4):
        print(f"Process {name}: {i}")
        time.sleep(0.5)

# Create process objects
p1 = Process(target=print_numbers, args=("A",))
p2 = Process(target=print_numbers, args=("B",))
```

Each process is like a chef in their own kitchen. Next, start and join the processes:

```python
p1.start()
p2.start()

p1.join()
p2.join()

print("Both processes have finished.")
```

The output will show both processes running in parallel, with their print statements interleaved. This is true parallelism—each process is independent and can use the CPU at the same time.

But what if you need to share data? You must use special tools, like queues or pipes, which act like delivery trucks between kitchens. Here's how to use a queue for IPC:

```python
from multiprocessing import Process, Queue

def worker(q):
    q.put('Ingredient from worker')

queue = Queue()
p = Process(target=worker, args=(queue,))
p.start()
print(queue.get())
p.join()
```

This approach is slower than simply sharing data between threads, so use multiprocessing when you need real parallelism for CPU-heavy tasks, and keep communication between processes to a minimum. 



Pythons **`multiprocessing`**-modul muliggør **ægte parallelisme** ved at oprette separate processer, hver med sin egen Python-fortolker og hukommelsesområde. I modsætning til tråde, som deler hukommelse og kører i samme proces, er **processer fuldstændig isolerede** fra hinanden. Denne isolation undgår begrænsningen fra **Global Interpreter Lock (GIL)**, hvilket gør det muligt for flere CPU-kerner at arbejde parallelt på CPU-tunge opgaver. Efter at have set videoen ovenfor, bør du forstå, at **procesbaseret parallelisme** især er nyttig, når der skal udføres tunge beregninger eller udnyttes flere kerner til opgaver, der ellers ville blive begrænset af GIL. **`multiprocessing`**-modulet tilbyder et velkendt API, der ligner **`threading`**-modulet, hvilket gør det nemt at skifte mellem tråde og processer afhængigt af behov.

Represents an independent process. Use this class to create and manage separate processes.

A process-safe FIFO queue for sharing data between processes. 

Provides a connection between two processes for two-way communication.

Manages a pool of worker processes for parallel execution of a function across data. 

A synchronization primitive for controlling access to shared resources between processes.

Allows sharing a single value of a specific type between processes.

Allows sharing an array of values between processes. 

Creates a server process that holds Python objects and allows other processes to manipulate them using proxies.

Returns a reference to the currently running process object.

Returns the number of CPU cores available on the system.

Brug disse klasser og funktioner til at opbygge robuste parallelle applikationer, håndtere datadeling og koordinere procesudførelse i Python.

import multiprocessing
import time

def print_message(message, delay):
    time.sleep(delay)
    print(message)

process1 = multiprocessing.Process(target=print_message, args=("Process 1 finished", 2))
process2 = multiprocessing.Process(target=print_message, args=("Process 2 finished", 1))

process1.start()
process2.start()

process1.join()
process2.join()


Dette kodeeksempel demonstrerer, hvordan man bruger Pythons **`multiprocessing`**-modul til at køre opgaver parallelt ved at oprette separate processer. Eksemplet anvender simple tidsforsinkelser og udskriver beskeder for at vise, hvordan processer kan køre uafhængigt og samtidigt.

#### Centrale begreber illustreret

- **Procesoprettelse:**
  - Nye processer oprettes med `multiprocessing.Process`-klassen;
  - Hver proces får en målfunktion (`print_message`) og argumenter, der skal videregives til funktionen.

- **Processtart:**
  - Metoden `start()` starter hver proces, så de kan køre samtidigt;
  - Når de er startet, kører hver proces sin målfunktion uafhængigt af de andre.

- **Parallel udførelse:**
  - Begge processer sover i forskellige tidsrum (`2` og `1` sekunder) ved hjælp af `time.sleep()`;
  - Da de kører parallelt, afslutter processen med den korteste forsinkelse først, selvom den blev startet efter den anden.

- **Proces-sammenkædning:**
  - Metoden `join()` sikrer, at hovedprogrammet venter på, at begge processer er færdige, før det fortsætter;
  - Uden `join()` kunne hovedprogrammet afslutte, før processerne er færdige.

Dette eksempel viser, at med multiprocessing kan opgaver køre samtidigt på flere CPU-kerner, hvilket gør det ideelt til CPU-tunge operationer, der drager fordel af parallel udførelse.

Hvilken påstand beskriver bedst forskellen mellem tråde og processer i Python?

Du vil opnå praktiske færdigheder i Pythons moduler og imports for effektiv kodestyring. Behersk avancerede teknikker til fejlhåndtering for mere pålidelige applikationer, automatiser filoperationer med filhåndtering, og lær robuste teststrategier med Pytest og Unittest for at sikre din kodes integritet.

Master Python's exception system for robust error handling and debugging.

Explore advanced techniques for file operations and resource management using context managers.

Compare and apply Python's multithreading and multiprocessing modules for concurrent programming.

Master asynchronous programming in Python using the asyncio library for scalable I/O-bound applications.

Explore how to write, organize, and run tests using Python's unittest and pytest frameworks.

Arbejde med processer

Centrale begreber illustreret