anthibo / advanced-python-topics Goto Github PK

• separate memory space -> memory not shared between processes
• Heavyweight
• One GIL for each process
• IPC is more complicated

• leightweight
• great for I/O bound tasks
• one thread at a time
• threads share memory
• limited by GIL: one thread at a time
• not interruptable/killable
• careful with race conditions -> change one memory space at the same time

• Global Interpreter Lock
• a lock that allows only on thread at a time to execute in py
• avoid multiprocessing
• reference counting

• Daemon thread: a thread that dies after the main thread dies
• By default, threads are not daemon threads

thread.daemon = True  # demon thread -> background thread that will die when the main thread diethread.daemon = True  # demon thread -> background thread that will die when the main thread die

• Race condition: two threads accessing the same resource at the same time
• To deal with race conditions: lock and release resources or using context managers (with)

def increase(lock: Lock):
    global database_value

    lock.acquire()
    local_cp = database_value

    # processing
    local_cp += 1
    
    time.sleep(0.1)
    database_value = local_cp
    lock.release()
# Or with
def increase(lock: Lock):
    global database_value
    
    with lock:
        local_cp = database_value

        # processing
        local_cp += 1
        
        time.sleep(0.1)
        database_value = local_cp

Queues:

• thread-safe env
• q.join(): block main thread until all queue items are done and processed
• q.task_done(): marks a thread as done

Share data between processes:

• Processes do not share same memory