Concurrent programming
content
- Operating system history
- Multi-channel technology
- Process theory
- Two ways to start the process
- Join method of process object
- Data isolation between processes
- Other methods of process objects
- Zombie process and orphan process
- Daemon
- Mutex
Essential knowledge review
-
The computer is also called the computer, that is, the brain with electricity. The computer was invented to enable him to work like a human after being powered on, and it is more efficient than human because it can be uninterrupted for 24 hours.
-
The five major components of a computer
Controller
Operator
Memory
input device
Output device
The core of the computer really works is the CPU (controller + arithmetic unit = central processing unit)
-
For a program to be run by a computer, its code must be read from the hard disk to memory, and then the CPU fetches instructions before executing
Operating system history
Just refer to the blog: https://www.cnblogs.com/Dominic-Ji/articles/10929381.html
Multi-channel technology
Single core achieves concurrent effect
Required knowledge
-
Concurrency
It looks like running concurrently can be called concurrent
-
parallel
Real simultaneous execution
ps:
- Parallel must be considered concurrent
- Single-core computers certainly cannot achieve parallelism, but they can achieve concurrency! ! !
Supplement: We directly assume that a single core is a core, and only one person works, don't consider the number of cores in the CPU
Multi-channel technical illustration
Save the total time spent running multiple programs
Reference group screenshot
Multi-channel technology key knowledge
Taking in space and taking in time
-
Spatial multiplexing
Multiple programs share a set of computer hardware
-
Multiplexing in time
Example: Washing clothes for 30s, cooking for 50s, boiling water for 30s
Single channel needs 110s, multiple channels only need the long switch of task to save time
Example: Playing a game while eating and saving state
Switch + save state
"""
切换(CPU)分为两种情况
1.当一个程序遇到IO操作的时候,操作系统会剥夺该程序的CPU执行权限
作用:提高了CPU的利用率 并且也不影响程序的执行效率
2.当一个程序长时间占用CPU的时候,操作吸引也会剥夺该程序的CPU执行权限
弊端:降低了程序的执行效率(原本时间+切换时间)
"""
Process theory
Required knowledge
The difference between program and process
"""
程序就是一堆躺在硬盘上的代码,是“死”的
进程则表示程序正在执行的过程,是“活”的
"""
Process scheduling
-
First come first served scheduling algorithm
"""对长作业有利,对短作业无益"""
-
Short job priority scheduling algorithm
"""对短作业有利,多长作业无益"""
-
Time slice rotation method + multi-level feedback queue
Reference illustration
Three state diagram of process running
Refer to the illustration to understand
Two pairs of important concepts
-
Synchronous and asynchronous
"""描述的是任务的提交方式""" 同步:任务提交之后,原地等待任务的返回结果,等待的过程中不做任何事(干等) 程序层面上表现出来的感觉就是卡住了 异步:任务提交之后,不原地等待任务的返回结果,直接去做其他事情 我提交的任务结果如何获取? 任务的返回结果会有一个异步回调机制自动处理
-
Blocking non-blocking
"""描述的程序的运行状态""" 阻塞:阻塞态 非阻塞:就绪态、运行态 理想状态:我们应该让我们的写的代码永远处于就绪态和运行态之间切换
Combination of the above concepts: the most efficient combination is asynchronous non-blocking
Two ways to start the process
# 方式一:直接调用函数
from multiprocessing import Process
import time
def task(name):
print(f'子进程{name}开始')
time.sleep(3)
print(f'子进程运行{name}结束')
if __name__ == '__main__':
# Process是一个类,调用类生成对象p,需要传参
print(f'主进程开始')
p=Process(target=task, args=('1',))
# 所有容器类型哪怕只有一个参数后面也要加逗号,防止是元组
p.start()
print(f'主进程结束')
# 方式2:调用类
class MyProcess(Process):
def __init__(self,name):
super().__init__()
self.name=name
def run(self):
print(f'子进程{self.name}开始')
time.sleep(3)
print(f'子进程{self.name}运行结束')
if __name__ == '__main__':
print(f'主进程开始')
p = MyProcess('1')
p.start()
print(f'主进程结束')
to sum up
"""
创建进程就是在内存中申请一块内存空间将需要运行的代码丢进去
一个进程对应在内存中就是一块独立的内存空间
多个进程对应在内存中就是多块独立的内存空间
进程与进程之间数据默认情况下是无法直接交互,如果想交互可以借助于第三方工具、模块
"""
join method
Join is to let the main process wait for the sub-process code to finish running before continuing. Does not affect the execution of other child processes
# join是让主进程等待子进程的结束
from multiprocessing import Process
import time
def task(name, sleep_time):
print(f'子进程{name}开始')
time.sleep(sleep_time)
print(f'子进程运行{name}结束')
if __name__ == '__main__':
print(f'主进程开始')
start_time=time.time()
tup=('egon', 'tank', 'json',)
p_list=[]
for i,name in enumerate(tup):
p = Process(target=task, args=(name, i,))
p.start()
p_list.append(p)
for p in p_list:
p.join()
print(f'主进程结束',time.time()-start_time)
Data between processes is isolated from each other
from multiprocessing import Process
money = 100
def task():
global money
money = 666
print(money)
if __name__ == '__main__':
p = Process(target=task)
p.start()
p.join()
print(money)