Python并发之Asyncio_python asyncio

import asyncio
import aiohttp
import time
import concurrent.futures
import multiprocessing
 
 
# 异步网页下载
async def download_one(url):
	async with aiohttp.ClientSession() as session:
		async with session.get(url) as resp:
			print("read {} from {}".format(resp.content_length, url))
			
			
async def download_all(sites):
	tasks = [asyncio.create_task(download_one(site)) for site in sites]
	await asyncio.gather(*tasks)
	
	
# 求列表中元素的整数平方和 常规版本
def cpu_bound(number):
	print("number={}, result={}".format(number, sum(i*i for i in range(number))))
	
def calculate_sums(numbers):
	for number in numbers:
		cpu_bound(number)
		
def calcuter_normal(numbers):
	start_time = time.perf_counter()
	calculate_sums(numbers)
	end_time = time.perf_counter()
	print("普通版本，耗时{}秒".format(end_time-start_time))
	
# 并行版本
def calculate_sums_future(numbers):
	with concurrent.futures.ThreadPoolExecutor() as executor:
		executor.map(cpu_bound, numbers)
		
def calcuter_future(numbers):
	start_time = time.perf_counter()
	calculate_sums_future(numbers)
	end_time = time.perf_counter()
	print("多进程版本，耗时{}秒".format(end_time-start_time))
	
# 动态规划版本
squ = {} # 用来储存中间结果
def cpu_dp(number):
	result = 0
	for i in range(number):
		if i not in squ.keys():
			squ[i] = i*i
		result += squ[i]
	print("number={}, result={}".format(number, result))
		
def calculate_sums_dp(numbers):
	for number in numbers:
		cpu_dp(number)
		
def calcuter_dp(numbers):
	start_time = time.perf_counter()
	calculate_sums_dp(numbers)
	end_time = time.perf_counter()
	print("动态规划版本，耗时{}秒".format(end_time-start_time))
	
# multiprocessing版本
def calculate_sums_multiprocessing(numbers):
	with multiprocessing.Pool() as pool:
		pool.map(cpu_bound, numbers)
		
def calcuter_multiprocessing(numbers):
	start_time = time.perf_counter()
	calculate_sums_multiprocessing(numbers)
	end_time = time.perf_counter()
	print("multiprocessing版本，耗时{}秒".format(end_time-start_time))
 
 
if __name__ == "__main__":
	sites = [
		'http://www.dapenti.com/blog/readforwx.asp?name=xilei&id=143655',
		'http://www.dapenti.com/blog/readforwx.asp?name=xilei&id=143656',
		'http://www.dapenti.com/blog/readforwx.asp?name=xilei&id=143657',
		'http://www.dapenti.com/blog/readforwx.asp?name=xilei&id=143658',
		'http://www.dapenti.com/blog/readforwx.asp?name=xilei&id=143659',
		'http://www.dapenti.com/blog/readforwx.asp?name=xilei&id=143660',
		'http://www.dapenti.com/blog/readforwx.asp?name=xilei&id=143661',
		'http://www.dapenti.com/blog/readforwx.asp?name=xilei&id=143662'
	]
	start_time = time.perf_counter()
	asyncio.run(download_all(sites))
	end_time = time.perf_counter()
	print("异步版下载了{}个网站，耗时{}".format(len(sites), end_time-start_time))
	
	# 思考题 
	numbers = [1000000+x for x in range(10)]
	calcuter_normal(numbers)
	calcuter_future(numbers)
	calcuter_dp(numbers)
	calcuter_multiprocessing(numbers)

END

----------------------------------------------------------------

欢迎关注我的公众号！