What is multiprocessing.shared_memory and when to use it?

Question

Accepted Answer

**shared_memory** allows **memory sharing** between processes without serialization.

```python
from multiprocessing import Process, shared_memory
import numpy as np

def worker(shm_name, shape, dtype):
    existing_shm = shared_memory.SharedMemory(name=shm_name)
    arr = np.ndarray(shape, dtype=dtype, buffer=existing_shm.buf)
    arr *= 2  # Modify shared data
    existing_shm.close()

if __name__ == "__main__":
    # Create shared array
    arr = np.array([1, 2, 3, 4, 5], dtype=np.int32)
    shm = shared_memory.SharedMemory(create=True, size=arr.nbytes)
    shared_arr = np.ndarray(arr.shape, dtype=arr.dtype, buffer=shm.buf)
    shared_arr[:] = arr[:]  # Copy data

# Process modifies shared memory
    p = Process(target=worker, args=(shm.name, arr.shape, arr.dtype))
    p.start()
    p.join()

print(shared_arr)  # [2, 4, 6, 8, 10]
    shm.close()
    shm.unlink()
```

**Benefits:**
- Zero-copy data sharing.
- Much faster than Queue/Pipe for large data.
- Works with NumPy arrays.

**Use cases:**
- Shared state between workers.
- Large dataset processing.

Python: What is multiprocessing.shared_memory and when to use it?