一篇文章帶你了解Java中ThreadPool線程池
線程池做的工作主要是控制運(yùn)行的線程數(shù)量,處理過(guò)程中將任務(wù)放入隊(duì)列,然后在線程創(chuàng)建后啟動(dòng)這些任務(wù),如果線程數(shù)量超過(guò)了最大數(shù)量,超出的線程排隊(duì)等候,等待其他線程執(zhí)行完畢,再?gòu)年?duì)列中取出任務(wù)來(lái)執(zhí)行
線程池的特點(diǎn)線程復(fù)用、控制最大并發(fā)數(shù)、管理線程
降低資源消耗。重復(fù)利用已創(chuàng)建的線程,降低創(chuàng)建和銷毀線程的開(kāi)銷 提高響應(yīng)速度。當(dāng)任務(wù)到達(dá)時(shí),任務(wù)可以不需要等待線程創(chuàng)建就能立刻執(zhí)行 提高線程的可管理性。使用線程池可以對(duì)線程進(jìn)行統(tǒng)一的分配、調(diào)優(yōu)和監(jiān)控1 線程池的方法執(zhí)行長(zhǎng)期任務(wù)性能好,創(chuàng)建一個(gè)線程池,一池有N個(gè)固定的線程,可以控制線程最大并發(fā)數(shù),有固定線程數(shù)的線程池[
ExecutorService threadPool = Executors.newFixedThreadPool(N);
單個(gè)任務(wù)執(zhí)行,它只會(huì)使用單個(gè)工作線程,一池一線程
ExecutorService threadPool = Executors.newSingleThreadExecutor();
執(zhí)行短期異步任務(wù),可緩存線程池,線程池根據(jù)需要?jiǎng)?chuàng)建新線程,但在先前構(gòu)造的線程可以復(fù)用,也可靈活回收空閑的線程,可擴(kuò)容的池
ExecutorService threadPool = Executors.newCachedThreadPool();
周期性線程池;支持定時(shí)及周期性任務(wù)執(zhí)行
ExecutorService threadPool = Executors.newScheduledThreadPool();(1) newFixedThreadPool
可以控制線程最大并發(fā)數(shù)的線程池:
public class FixedThreadPool { private static AtomicInteger num = new AtomicInteger(0); private static ExecutorService executorService = Executors.newFixedThreadPool(2); public static void main(String[] args) {countSum c= new countSum();//將coutSum作為T(mén)ask,submit至線程池for (int i = 0; i < 2; i++) { executorService.submit(c);}//Task執(zhí)行完成后關(guān)閉executorService.shutdown(); } static class countSum implements Runnable{@Overridepublic void run() { for (int i = 0; i < 500; i++) {try{ System.out.println('Thread - '+Thread.currentThread().getName()+' count= '+ num.getAndIncrement()); Thread.sleep(100);}catch (Exception e){ e.printStackTrace();} }} }}
結(jié)果:
只會(huì)使用唯一的工作線程執(zhí)行任務(wù)的線程池:
public class SingleThreadExecutor { private static AtomicInteger num = new AtomicInteger(0); private static ExecutorService executorService = Executors.newSingleThreadExecutor(); public static void main(String[] args) {//將coutSum作為T(mén)ask,submit至線程池for (int i = 0; i < 2; i++) { executorService.submit(new countSum());}//Task執(zhí)行完成后關(guān)閉executorService.shutdown(); } static class countSum implements Runnable{@Overridepublic void run() { for (int i = 0; i < 500; i++) {try{ System.out.println('Thread - '+Thread.currentThread().getName()+' count= '+ num.getAndIncrement()); Thread.sleep(100);}catch (Exception e){ e.printStackTrace();} }} }}
結(jié)果:
傳參值為corePoolSize大小,支持定時(shí)及周期性任務(wù)執(zhí)行
延期執(zhí)行示例:調(diào)用schedule方法,三個(gè)參數(shù):Task,Delay,TimeUnit
public class ScheduledThreadPool { // corePoolSize = 2 private static ScheduledExecutorService service = Executors.newScheduledThreadPool(2); public static void main(String[] args) {System.out.println('Thread - '+Thread.currentThread().getName()+' BEGIN '+ new Date());service.schedule(new print(),5, TimeUnit.SECONDS);service.shutdown(); } static class print implements Runnable{@Overridepublic void run() { for (int i = 0; i < 10; i++) {try{ System.out.println('Thread - '+Thread.currentThread().getName()+' Delay 5 second and sleep 2 second '+ new Date()); Thread.sleep(2000);}catch (Exception e){ e.printStackTrace();} }} }}
結(jié)果:
定時(shí)執(zhí)行示例:調(diào)用scheduleAtFixedRate方法,四個(gè)參數(shù):Task,initialDelay,Period,TimeUnit
public class ScheduledThreadPool { // corePoolSize = 1 private static ScheduledExecutorService service = Executors.newScheduledThreadPool(1); public static void main(String[] args) {System.out.println('Thread - '+Thread.currentThread().getName()+' BEGIN '+ new Date());service.scheduleAtFixedRate(new print(),5,3,TimeUnit.SECONDS); } static class print implements Runnable{@Overridepublic void run() { System.out.println('Thread - '+Thread.currentThread().getName()+' Delay 5 second and period 3 second '+ new Date());} }}
結(jié)果:
可緩存線程池,如果線程池長(zhǎng)度超過(guò)處理需要,回收空閑線程,若無(wú)可回收,則新建線程。即若前一個(gè)任務(wù)已完成,則會(huì)接著復(fù)用該線程:
public class CachedThreadPool { private static AtomicInteger num = new AtomicInteger(0); private static ExecutorService service = Executors.newCachedThreadPool(); public static void main(String[] args) {countSum c = new countSum();for (int i = 0; i < 3; i++) { try {service.submit(c);Thread.sleep(1000); }catch (Exception e){e.printStackTrace(); }}service.shutdown(); } static class countSum implements Runnable{@Overridepublic void run() { for (int i = 0; i < 1000; i++) {System.out.println('Thread - '+Thread.currentThread().getName()+' countSum= '+num.getAndIncrement()); }} }}
結(jié)果:Thread.sleep(1000)即sleep一秒,上個(gè)任務(wù)完成可繼續(xù)復(fù)用該線程,不需要?jiǎng)?chuàng)建新的線程
若將Tread.sleep(1000)注釋掉,你會(huì)發(fā)現(xiàn)有3個(gè)線程在跑
若感興趣可以去了解一下它們的底層源碼,對(duì)于CachedThreadPool而言,可新建線程最大數(shù)量為INTEGER.MAXIMUM
2 線程池底層原理以newFixedThreadPool為例
public static ExecutorService newFixedThreadPool(int nThreads) { return new ThreadPoolExecutor(nThreads, nThreads,0L, TimeUnit.MILLISECONDS,new LinkedBlockingQueue<Runnable>());}
public ThreadPoolExecutor(int corePoolSize, int maximumPoolSize, long keepAliveTime, TimeUnit unit, BlockingQueue<Runnable> workQueue) { this(corePoolSize, maximumPoolSize, keepAliveTime, unit, workQueue,Executors.defaultThreadFactory(), defaultHandler); }
線程池七大參數(shù)
corePoolSize:線程池中的常駐核心線程數(shù) maximumPoolSize:線程池中能夠容納同時(shí)執(zhí)行的最大線程數(shù),必須大于1 keepAliveTime:多余的空閑線程的存活時(shí)間;當(dāng)前池中線程數(shù)量超過(guò)corePoolSize時(shí),當(dāng)空閑時(shí)間達(dá)到keepAliveTime時(shí),多余線程會(huì)被銷毀 unit:keepAliveTime的單位 workQueue:任務(wù)隊(duì)列,被提交但尚未執(zhí)行的任務(wù) threadFactory:表示生成線程池中工作線程的線程工廠,用于創(chuàng)建線程,一般默認(rèn) handler:拒絕策略,表示當(dāng)隊(duì)列滿了,并且工作線程大于等于線程池的最大線程數(shù)時(shí)如何來(lái)拒絕請(qǐng)求執(zhí)行的runnable的策略線程池四大流程
1)創(chuàng)建線程池后,開(kāi)始等待請(qǐng)求
2)當(dāng)調(diào)用execute()方法添加一個(gè)請(qǐng)求任務(wù)時(shí),線程池會(huì)做以下判斷:
如果正在運(yùn)行的線程數(shù)量小于corePoolSize,馬上創(chuàng)建線程執(zhí)行任務(wù) 如果正在運(yùn)行的線程數(shù)量大于等于corePoolSize,將該任務(wù)放入等待隊(duì)列 如果等待隊(duì)列已滿,但正在運(yùn)行線程數(shù)量小于max,創(chuàng)建非核心線程執(zhí)行任務(wù) 如果隊(duì)列滿了且正在運(yùn)行的線程數(shù)量大于max,線程池會(huì)啟動(dòng)飽和拒絕策略3)當(dāng)一個(gè)線程完成任務(wù)時(shí),會(huì)從等待隊(duì)列中取下一個(gè)任務(wù)來(lái)執(zhí)行
4)當(dāng)空閑線程超過(guò)keepAliveTime定義時(shí)間,會(huì)判斷:
如果當(dāng)前運(yùn)行線程大于corePoolSize,該線程銷毀 所有線程執(zhí)行完任務(wù)后,線程個(gè)數(shù)恢復(fù)到corePoolSize大小3 線程池策略及分析Note:阿里巴巴JAVA開(kāi)發(fā)手冊(cè):線程池不允許使用Executors去創(chuàng)建線程池,而是通過(guò)使用ThreadPoolExecutor的方式自定義線程池,規(guī)避資源耗盡的風(fēng)險(xiǎn)
Executors返回的線程池對(duì)象的弊端:
1)FixedThreadPool和SingleThreadPool:
允許請(qǐng)求隊(duì)列長(zhǎng)度為Integer.MAX_VALUE,可能會(huì)堆積大量請(qǐng)求導(dǎo)致OOM
2)CachedThreadPool和ScheduledThreadPool:
允許創(chuàng)建線程數(shù)量為Integer.MAX_VALUE,可能會(huì)創(chuàng)建大量的線程導(dǎo)致OOM
拒絕策略1)AbortPolicy
直接拋出RejectedExecutionException異常阻止系統(tǒng)正常運(yùn)行
2)CallerRunsPolicy
'調(diào)用者運(yùn)行'的調(diào)節(jié)機(jī)制,該策略既不會(huì)拋棄任務(wù),也不會(huì)拋出異常,而是將某些任務(wù)回退到調(diào)用者,從而降低新任務(wù)的流量
3)DiscardPolicy
該策略拋棄無(wú)法處理的任務(wù),不予任何處理也不拋出異常。如果允許任務(wù)丟失,這是最好的一種策略
4)DiscardOldestPolicy
拋棄隊(duì)列中等待最久的任務(wù),然后把當(dāng)前任務(wù)加入隊(duì)列中嘗試再次提交當(dāng)前任務(wù)
如何設(shè)置maximumPoolSize大小Runtime.getRuntime().availableProcessors()方法獲取核數(shù)
CPU密集型
maximumPoolSize設(shè)為核數(shù)+1
IO密集型
maximumPoolSize設(shè)為核數(shù)/阻塞系數(shù)
以上就是一篇文章-帶你了解ThreadPool線程池的詳細(xì)內(nèi)容,更多關(guān)于ThreadPool線程池的資料請(qǐng)關(guān)注好吧啦網(wǎng)其它相關(guān)文章!
相關(guān)文章:
1. CSS清除浮動(dòng)方法匯總2. js開(kāi)發(fā)中的頁(yè)面、屏幕、瀏覽器的位置原理(高度寬度)說(shuō)明講解(附圖)3. XML 非法字符(轉(zhuǎn)義字符)4. CSS百分比padding制作圖片自適應(yīng)布局5. vue跳轉(zhuǎn)頁(yè)面常用的幾種方法匯總6. 不要在HTML中濫用div7. XML入門(mén)的常見(jiàn)問(wèn)題(三)8. CSS3實(shí)例分享之多重背景的實(shí)現(xiàn)(Multiple backgrounds)9. 深入了解React中的合成事件10. TypeScript實(shí)現(xiàn)十大排序算法之歸并排序示例詳解
