区分并发和并行
并发(Concurrency).
由于Python 的解释器并不是线程安全的，为了解决由此带来的 race condition 等问题，Python 便引入了全局解释器锁，也就是同一时刻，只允许一个线程执行。当然，在执行 I/O 操作时，如果一个线程被 block 了，全局解释器锁便会被释放，从而让另一个线程能够继续执行。所以在Python中，并发并不是指同一时刻有多个操作(thread、task)同时进行，而是同一时刻，只允许有一个线程或任务执行。
并行(Parallelism)
指多个进程完全同步同时的执行。
并发编程之 Futures
单线程与多线程性能比较
假设我们有一个任务，是下载一些网站的内容并打印。如果用单线程的方式，它的代码实现如下所示
import requests import time def download_one(url): resp = requests.get(url) print('Read {} from {}'.format(len(resp.content), url)) def download_all(sites): for site in sites: download_one(site) def main(): sites = [ 'https://en.wikipedia.org/wiki/Portal:Arts', 'https://en.wikipedia.org/wiki/Portal:History', 'https://en.wikipedia.org/wiki/Portal:Society', 'https://en.wikipedia.org/wiki/Portal:Biography', 'https://en.wikipedia.org/wiki/Portal:Mathematics', 'https://en.wikipedia.org/wiki/Portal:Technology', 'https://en.wikipedia.org/wiki/Portal:Geography', 'https://en.wikipedia.org/wiki/Portal:Science', 'https://en.wikipedia.org/wiki/Computer_science', 'https://en.wikipedia.org/wiki/Python_(programming_language)', 'https://en.wikipedia.org/wiki/Java_(programming_language)', 'https://en.wikipedia.org/wiki/PHP', 'https://en.wikipedia.org/wiki/Node.js', 'https://en.wikipedia.org/wiki/The_C_Programming_Language', 'https://en.wikipedia.org/wiki/Go_(programming_language)' ] start_time = time.perf_counter() download_all(sites) end_time = time.perf_counter() print('Download {} sites in {} seconds'.format(len(sites), end_time - start_time)) if __name__ == '__main__': main() # 输出 Read 129196 from https://en.wikipedia.org/wiki/Portal:Arts Read 183867 from https://en.wikipedia.org/wiki/Portal:History Read 224161 from https://en.wikipedia.org/wiki/Portal:Society Read 114387 from https://en.wikipedia.org/wiki/Portal:Biography Read 152871 from https://en.wikipedia.org/wiki/Portal:Mathematics Read 156339 from https://en.wikipedia.org/wiki/Portal:Technology Read 162872 from https://en.wikipedia.org/wiki/Portal:Geography Read 91504 from https://en.wikipedia.org/wiki/Portal:Science Read 323262 from https://en.wikipedia.org/wiki/Computer_science Read 391073 from https://en.wikipedia.org/wiki/Python_(programming_language) Read 319710 from https://en.wikipedia.org/wiki/Java_(programming_language) Read 470754 from https://en.wikipedia.org/wiki/PHP Read 180774 from https://en.wikipedia.org/wiki/Node.js Read 56799 from https://en.wikipedia.org/wiki/The_C_Programming_Language Read 325451 from https://en.wikipedia.org/wiki/Go_(programming_language) Download 15 sites in 67.349395015 seconds
以上代码的流程：先是遍历存储网站的列表； 然后对当前网站执行下载操作；等到当前操作完成后，再对下一个网站进行同样的操作，一直到结束。
接下来看多线程版本
import concurrent.futures import requests import threading import time def download_one(url): try: resp = requests.get(url) print('Read {} from {}'.format(len(resp.content), url)) except Exception as ex: print(ex) def download_all(sites): with concurrent.futures.ThreadPoolExecutor(max_workers=5) as executor: results = executor.map(download_one, sites) # with concurrent.futures.ProcessPoolExecutor() as executor: # results = executor.map(download_one,sites) def main(): sites = [ 'https://en.wikipedia.org/wiki/Portal:Arts', 'https://en.wikipedia.org/wiki/Portal:History', 'https://en.wikipedia.org/wiki/Portal:Society', 'https://en.wikipedia.org/wiki/Portal:Biography', 'https://en.wikipedia.org/wiki/Portal:Mathematics', 'https://en.wikipedia.org/wiki/Portal:Technology', 'https://en.wikipedia.org/wiki/Portal:Geography', 'https://en.wikipedia.org/wiki/Portal:Science', 'https://en.wikipedia.org/wiki/Computer_science', 'https://en.wikipedia.org/wiki/Python_(programming_language)', 'https://en.wikipedia.org/wiki/Java_(programming_language)', 'https://en.wikipedia.org/wiki/PHP', 'https://en.wikipedia.org/wiki/Node.js', 'https://en.wikipedia.org/wiki/The_C_Programming_Language', 'https://en.wikipedia.org/wiki/Go_(programming_language)' ] start_time = time.perf_counter() download_all(sites) end_time = time.perf_counter() print('Download {} sites in {} seconds'.format(len(sites), end_time - start_time)) if __name__ == '__main__': main() # 输出 Read 114387 from https://en.wikipedia.org/wiki/Portal:Biography Read 129196 from https://en.wikipedia.org/wiki/Portal:Arts Read 183867 from https://en.wikipedia.org/wiki/Portal:History Read 152871 from https://en.wikipedia.org/wiki/Portal:Mathematics Read 224161 from https://en.wikipedia.org/wiki/Portal:Society Read 156339 from https://en.wikipedia.org/wiki/Portal:Technology Read 91504 from https://en.wikipedia.org/wiki/Portal:Science Read 391073 from https://en.wikipedia.org/wiki/Python_(programming_language) Read 162872 from https://en.wikipedia.org/wiki/Portal:Geography Read 323262 from https://en.wikipedia.org/wiki/Computer_science Read 56799 from https://en.wikipedia.org/wiki/The_C_Programming_Language Read 319710 from https://en.wikipedia.org/wiki/Java_(programming_language) Read 325451 from https://en.wikipedia.org/wiki/Go_(programming_language) Read 180774 from https://en.wikipedia.org/wiki/Node.js Read 470754 from https://en.wikipedia.org/wiki/PHP Download 15 sites in 10.022916933 seconds
以上代码效率提高了6倍。使用ThreadPoolExecutor创建了一个线程池，max_workers分配了5个线程，executor.map(download_one, sites)对sites的元素并发的调用download_one函数。其中requests.get()方法是线程安全的(thread-safe)，在多线程环境中可以安全地使用。线程的数量虽可以自定，但过多的线程会造成系统的开销增大。可以根据实际需求做测试，寻找最优线程数量。
以上代码也可以用并行的方法来实现。在download_all()函数中：
with futures.ThreadPoolExecutor(workers) as executor => with futures.ProcessPoolExecutor() as executor:
对于这种IO场景，用并行的方式并不会比并发的方式效率高.
到底什么是 Futures ？
Python 中的 Futures 模块，位于 concurrent.futures 和 asyncio 中，它们都表示带有延迟的操作。Futures 会将处于等待状态的操作包裹起来放到队列中，这些操作的状态随时可以查询，当然，它们的结果或是异常，也能够在操作完成后被获取。
import concurrent.futures import requests import time def download_one(url): resp = requests.get(url) print('Read {} from {}'.format(len(resp.content), url)) return f'download {len(resp.content)} ok' # def over(arg): # print(arg) # print('over') def download_all(sites): #future列表中每个future完成的顺序，和它在列表中的顺序并不一定完全一致。 #到底哪个先完成、哪个后完成，取决于系统的调度和每个future的执行时间 with concurrent.futures.ThreadPoolExecutor(max_workers=5) as executor: to_do = [] for site in sites: #executor.submit返回future实例 future = executor.submit(download_one, site) to_do.append(future) #future.add_done_callback(over) #在futures完成后打印结果 for future in concurrent.futures.as_completed(to_do): print(future.result()) def main(): sites = [ 'https://en.wikipedia.org/wiki/Portal:Arts', 'https://en.wikipedia.org/wiki/Portal:History', 'https://en.wikipedia.org/wiki/Portal:Society', 'https://en.wikipedia.org/wiki/Portal:Biography', 'https://en.wikipedia.org/wiki/Portal:Mathematics', 'https://en.wikipedia.org/wiki/Portal:Technology', 'https://en.wikipedia.org/wiki/Portal:Geography', 'https://en.wikipedia.org/wiki/Portal:Science', 'https://en.wikipedia.org/wiki/Computer_science', 'https://en.wikipedia.org/wiki/Python_(programming_language)', 'https://en.wikipedia.org/wiki/Java_(programming_language)', 'https://en.wikipedia.org/wiki/PHP', 'https://en.wikipedia.org/wiki/Node.js', 'https://en.wikipedia.org/wiki/The_C_Programming_Language', 'https://en.wikipedia.org/wiki/Go_(programming_language)' ] start_time = time.perf_counter() download_all(sites) end_time = time.perf_counter() print('Download {} sites in {} seconds'.format(len(sites), end_time - start_time)) if __name__ == '__main__': main() # 输出 Read 129886 from https://en.wikipedia.org/wiki/Portal:Arts Read 107634 from https://en.wikipedia.org/wiki/Portal:Biography Read 224118 from https://en.wikipedia.org/wiki/Portal:Society Read 158984 from https://en.wikipedia.org/wiki/Portal:Mathematics Read 184343 from https://en.wikipedia.org/wiki/Portal:History Read 157949 from https://en.wikipedia.org/wiki/Portal:Technology Read 167923 from https://en.wikipedia.org/wiki/Portal:Geography Read 94228 from https://en.wikipedia.org/wiki/Portal:Science Read 391905 from https://en.wikipedia.org/wiki/Python_(programming_language) Read 321352 from https://en.wikipedia.org/wiki/Computer_science Read 180298 from https://en.wikipedia.org/wiki/Node.js Read 321417 from https://en.wikipedia.org/wiki/Java_(programming_language) Read 468421 from https://en.wikipedia.org/wiki/PHP Read 56765 from https://en.wikipedia.org/wiki/The_C_Programming_Language Read 324039 from https://en.wikipedia.org/wiki/Go_(programming_language) Download 15 sites in 0.21698231499976828 seconds
future列表中每个future完成的顺序，和它在列表中的顺序并不一定完全一致。到底哪个先完成、哪个后完成，取决于系统的调度和每个future的执行时间。
并发通常用于 I/O 操作频繁的场景，而并行则适用于 CPU heavy 的场景。