Concurrency and Multiprocessing

Threading

Pyvorin-compiled functions release the GIL during pure computation, allowing true parallelism with threading:

from concurrent.futures import ThreadPoolExecutor

def worker(data):
    return sum(x * x for x in data)

with ThreadPoolExecutor(max_workers=4) as ex:
    results = list(ex.map(worker, chunks))

Multiprocessing

For CPU-bound workloads, use multiprocessing to bypass the GIL entirely:

from multiprocessing import Pool

def worker(data):
    return compiled_transform(data)

with Pool(4) as p:
    results = p.map(worker, chunks)

Async Fallback

Async/await is unsupported for native compilation but falls back gracefully to CPython. Use async for I/O boundaries and compiled sync functions for CPU work:

async def handle_request(data):
    # I/O: CPython async
    raw = await fetch_data(data['id'])
    # CPU: compiled native
    result = process_data(raw)
    return result