It’s a mechanism used in the CPython implementation of Python to ensure that only one thread executes Python bytecode at a time, even on multi-core systems.
⸻
• CPython is not thread-safe by default.
• The GIL simplifies memory management in CPython by preventing concurrent access to Python objects.
⸻
Even if you create multiple threads for CPU-bound tasks: • Only one thread can execute Python code at any moment. • Other threads are just waiting, switching in and out of the lock. • You don’t get real parallelism—just concurrent context switching.
✅ I/O-bound tasks (like reading files, waiting for a network) benefit from threading. ❌ CPU-bound tasks (like computing primes) don’t benefit because of the GIL.
go mod init runtimeviz
go get gonum.org/v1/plot/...| Feature | Python (CPython) | Go |
|---|---|---|
| Concurrency model | Threads (limited by GIL) | Goroutines (green threads) |
| Parallelism | Requires multiprocessing |
Achieved naturally via goroutines |
| Thread cost | OS threads (heavy) | Goroutines (lightweight, managed by Go runtime) |
| Memory sharing | Shared, but needs locks or queues | Shared by default, channels preferred |
| GIL | Yes (CPython only) | No GIL in Go |