Threads vs. Tasks
Posted on Friday, October 11, 2013
.Net has three low-level mechanisms to run code in parallel: Thread
, ThreadPool
, and Task
. These three mechanism serve different purposes.
Thread
Thread
represents an actual OS-level thread, with its own stack and kernel resources. (technically, a CLR implementation could use fibers instead, but no existing CLR does this) Thread
allows the highest degree of control; you can Abort()
or Suspend()
or Resume()
a thread (though this is a very bad idea), you can observe its state, and you can set thread-level properties like the stack size, apartment state, or culture.
The problem with Thread
is that OS threads are costly. Each thread you have consumes a non-trivial amount of memory for its stack, and adds additional CPU overhead as the processor context-switch between threads. Instead, it is better to have a small pool of threads execute your code as work becomes available.
There are times when there is no alternative Thread
. If you need to specify the name (for debugging purposes) or the apartment state (to show a UI), you must create your own Thread
(note that having multiple UI threads is generally a bad idea). Also, if you want to maintain an object that is owned by a single thread and can only be used by that thread, it is much easier to explicitly create a Thread
instance for it so you can easily check whether code trying to use it is running on the correct thread.
ThreadPool
ThreadPool
is a wrapper around a pool of threads maintained by the CLR. ThreadPool
gives you no control at all; you can submit work to execute at some point, and you can control the size of the pool, but you can't set anything else. You can't even tell when the pool will start running the work you submit to it.
Using ThreadPool
avoids the overhead of creating too many threads. However, if you submit too many long-running tasks to the threadpool, it can get full, and later work that you submit can end up waiting for the earlier long-running items to finish. In addition, the ThreadPool
offers no way to find out when a work item has been completed (unlike Thread.Join()
), nor a way to get the result. Therefore, ThreadPool
is best used for short operations where the caller does not need the result.
Task
Finally, the Task
class from the Task Parallel Library offers the best of both worlds. Like the ThreadPool
, a task does not create its own OS thread. Instead, tasks are executed by a TaskScheduler
; the default scheduler simply runs on the ThreadPool.
Unlike the ThreadPool, Task
also allows you to find out when it finishes, and (via the generic Task<T>
) to return a result. You can call ContinueWith()
on an existing Task
to make it run more code once the task finishes (if it's already finished, it will run the callback immediately). If the task is generic, ContinueWith()
will pass you the task's result, allowing you to run more code that uses it.
You can also synchronously wait for a task to finish by calling Wait()
(or, for a generic task, by getting the Result
property). Like Thread.Join()
, this will block the calling thread until the task finishes. Synchronously waiting for a task is usually bad idea; it prevents the calling thread from doing any other work, and can also lead to deadlocks if the task ends up waiting (even asynchronously) for the current thread.
Since tasks still run on the ThreadPool, they should not be used for long-running operations, since they can still fill up the thread pool and block new work. Instead, Task
provides a LongRunning
option, which will tell the TaskScheduler
to spin up a new thread rather than running on the ThreadPool.
All newer high-level concurrency APIs, including the Parallel.For*()
methods, PLINQ, C# 5 await
, and modern async methods in the BCL, are all built on Task
.
Conclusion
The bottom line is that Task
is almost always the best option; it provides a much more powerful API and avoids wasting OS threads.
The only reasons to explicitly create your own Thread
s in modern code are setting per-thread options, or maintaining a persistent thread that needs to maintain its own identity.
http://blog.slaks.net/2013-10-11/threads-vs-tasks/
Threads vs. Tasks的更多相关文章
- java多线程系类:JUC线程池:03之线程池原理(二)(转)
概要 在前面一章"Java多线程系列--"JUC线程池"02之 线程池原理(一)"中介绍了线程池的数据结构,本章会通过分析线程池的源码,对线程池进行说明.内容包 ...
- Understanding the Internal Message Buffers of Storm
Understanding the Internal Message Buffers of Storm Jun 21st, 2013 Table of Contents Internal messag ...
- 进程物理内存远大于Xmx的问题分析
问题描述 最近经常被问到一个问题,”为什么我们系统进程占用的物理内存(Res/Rss)会远远大于设置的Xmx值”,比如Xmx设置1.7G,但是top看到的Res的值却达到了3.0G,随着进程的运行,R ...
- Java多线程系列--“JUC线程池”03之 线程池原理(二)
概要 在前面一章"Java多线程系列--“JUC线程池”02之 线程池原理(一)"中介绍了线程池的数据结构,本章会通过分析线程池的源码,对线程池进行说明.内容包括:线程池示例参考代 ...
- ThreadPoolExecutor机制探索-我们到底能走多远系列(41)
我们到底能走多远系列(41) 扯淡: 这一年过的不匆忙,也颇多感受,成长的路上难免弯路,这个世界上没人关心你有没有变强,只有自己时刻提醒自己,不要忘记最初出发的原因. 其实这个世界上比我们聪明的人无数 ...
- 怎么通过activity里面的一个按钮跳转到另一个fragment(android FragmentTransaction.replace的用法介绍)
即:android FragmentTransaction.replace的用法介绍 Fragment的生命周期和它的宿主Activity密切相关,几乎和宿主Activity的生命周期一致,他们之间最 ...
- OpenMP初步(英文)
Beginning OpenMP OpenMP provides a straight-forward interface to write software that can use multipl ...
- ThreadPoolExecutor 分析
一.从用法入手 Creates a thread pool that creates new threads as needed, but will reuse previously construc ...
- Threading in C#
http://www.albahari.com/threading/ PART 1: GETTING STARTED Introduction and Concepts C# supports par ...
随机推荐
- window10 还原精灵 破解版 冰点
韩梦飞沙 韩亚飞 313134555@qq.com yue31313 han_meng_fei_sha=======
- BZOJ.5120.[清华集训2017]无限之环(费用流zkw 黑白染色)
题目链接 LOJ 洛谷 容易想到最小费用最大流分配度数. 因为水管形态固定,每个点还是要拆成4个点,分别当前格子表示向上右下左方向. 然后能比较容易地得到每种状态向其它状态转移的费用(比如原向上的可以 ...
- swoole深入学习 4. process
swoole-1.7.2增加了一个进程管理模块,用来替代PHP的pcntl扩展.pcntl是php新增的一个多进程扩展,用来实现多进程,但是有很多不完善的地方,swoole 就完善了这些地方,而且使得 ...
- Windows7的MySQL数据库的安装
碰巧重装了系统,需要重新安装MySQL. 1.进入官网下载:https://dev.mysql.com/ 2.severonle 3.alt+n 4.alt+x
- 20172302 《Java软件结构与数据结构》第二周学习总结
2018年学习总结博客总目录:第一周 第二周 教材学习内容总结 第三章 集合概述-栈 3.1 集合 (1) 集合是一种聚集,组织了其他对象的对象.它定义一张破那个特定的方式,可以访问.管理所包含的对象 ...
- SQL数据库连接到服务器出错——无法连接到XXX
问题:Windows或者SQL Server身份验证下,出现连接到服务器出错 解决方式: 1.检查实例名称是否正确,如下 根据自己电脑的实际名称修改修改如下: 1)Microsoft SQL Serv ...
- python图片和分形树
链接: 这10个Python项目很有趣! Python 绘制分形图(曼德勃罗集.分形树叶.科赫曲线.分形龙.谢尔宾斯基三角等)附代码 使用Python生成树形图案 神奇的代码:用 Python 生成分 ...
- 最基本的区块链hello world(python3实现)
源自 用不到 50 行的 Python 代码构建最小的区块链 (英文原文:Let’s Build the Tiniest Blockchain ) ,但是文中的代码是基于python2的,python ...
- cached-query 将缓存和查询数据库高速连接起来的轻类库
介绍 我们经常有这种需求:当我们把memcached增加到项目后我还还要写一个 cacheUtils 或者 cacheManager 之类的类来操作memcached. 而且一般的操作不外乎是这种操作 ...
- java监控指定路径下文件及文件夹变化
之前用jdk7的WatchService API(java.nio.file包)来做目录下的子文件监控,后改为使用commons-io包.主要有下面几点不同:1. WatchService是采用扫描式 ...