100 lines of code showcasing the trade-offs between process-based vs thread-based concurrency in Python.

ThreadPoolExecutor is limited by Python's GIL (Global Interpreter Lock) and can only utilize a single CPU-core at a time.

ProcessPoolExecutor spawns a new process, and can effectively use all cores concurrently, but there is an initial overhead.

TL;DR:

• ThreadPoolExecutor works best for I/O-bound tasks

• ProcessPoolExecutor works best for CPU-bound tasks

Options:

# run a CPU-bound workload using thread-concurrency
python demo.py cpu threads 

# run a I/O-bound workload using thread-concurrency
python demo.py io threads 

# run a CPU-bound workload using thread-concurrency
python demo.py cpu procs 

# run a I/O-bound workload using thread-concurrency
python demo.py io procs