I want to know how AsyncTask works internally.
I know it uses the Java Executor to perform the operations but still some of the questions I am not understanding. Like:
How many AsyncTask can be started at a time in an Android app?
When I start 10 AsyncTask, will all tasks will run simultaneously or one by one?
I have tried with 75000 AsyncTask to test the same. I don't get any problem and seems like all the tasks will be pushed to stack and will run one by one.
Also when I start 100000 AsyncTasks, I start getting OutOfMemoryError.
So is there any limit of no of AsyncTask which can be run at a time?
Note: I have tested these on SDK 4.0
AsyncTask has a rather long story.
When it first appeared in Cupcake (1.5) it handled background operations with a single additional thread (one by one). In Donut (1.6) it was changed, so that a pool of thread had begun to be used. And operations could be processed simultaneously until the pool had been exhausted. In such case operations were enqueued.
Since Honeycomb default behavior is switched back to use of a single worker thread (one by one processing). But the new method (executeOnExecutor) is introduced to give you a possibility to run simultaneous tasks if you wish (there two different standard executors: SERIAL_EXECUTOR and THREAD_POOL_EXECUTOR).
The way how tasks are enqueued also depends on what executor you use. In case of a parallel one you are restricted with a limit of 10 (new LinkedBlockingQueue<Runnable>(10)). In case of a serial one you are not limited (new ArrayDeque<Runnable>()).
So the way your tasks are processed depends on how you run them and what SDK version you run them on.
As for thread limits, we are not guaranteed with any, yet looking at the ICS source code we can say that number of threads in the pool can vary in range 5..128.
When you start 100000 with default execute method serial executor is used.
Since tasks that cannot be processed immediately are enqueued you get OutOfMemoryError (thousands of tasks are added to the array backed queue).
Exact number of tasks you can start at once depends on the memory class of the device you are running on and, again, on executor you use.
let us dive deep into the Android’s Asynctask.java file to understand it from a designer’s perspective and how it has nicely implemented Half Sync-Half Async design pattern in it.
In the beginning of the class few lines of codes are as follows:
private static final ThreadFactory sThreadFactory = new ThreadFactory() {
private final AtomicInteger mCount = new AtomicInteger(1);
public Thread newThread(Runnable r) {
return new Thread(r, "AsyncTask #" + mCount.getAndIncrement());
}
};
private static final BlockingQueue<Runnable> sPoolWorkQueue =
new LinkedBlockingQueue<Runnable>(10);
/**
* An {#link Executor} that can be used to execute tasks in parallel.
*/
public static final Executor THREAD_POOL_EXECUTOR
= new ThreadPoolExecutor(CORE_POOL_SIZE, MAXIMUM_POOL_SIZE, KEEP_ALIVE,
TimeUnit.SECONDS, sPoolWorkQueue, sThreadFactory);
The first is a ThreadFactory which is responsible for creating worker threads. The member variable of this class is the number of threads created so far. The moment it creates a worker thread, this number gets increased by 1.
The next is the BlockingQueue. As you know from the Java blockingqueue documentation, it actually provides a thread safe synchronized queue implementing FIFO logic.
The next is a thread pool executor which is responsible for creating a pool of worker threads which can be taken as and when needed to execute different tasks.
If we look at the first few lines we will know that Android has limited the maximum number of threads to be 128 (as evident from private static final int MAXIMUM_POOL_SIZE = 128).
Now the next important class is SerialExecutor which has been defined as follows:
private static class SerialExecutor implements Executor {
final ArrayDeque<Runnable> mTasks = new ArrayDeque<Runnable>();
Runnable mActive;
public synchronized void execute(final Runnable r) {
mTasks.offer(new Runnable() {
public void run() {
try {
r.run();
} finally {
scheduleNext();
}
}
});
if (mActive == null) {
scheduleNext();
}
}
protected synchronized void scheduleNext() {
if ((mActive = mTasks.poll()) != null) {
THREAD_POOL_EXECUTOR.execute(mActive);
}
}
}
The next important two functions in the Asynctask is
public final AsyncTask<Params, Progress, Result> execute(Params... params) {
return executeOnExecutor(sDefaultExecutor, params);
}
and
public final AsyncTask<Params, Progress, Result> executeOnExecutor(Executor exec,
Params... params) {
if (mStatus != Status.PENDING) {
switch (mStatus) {
case RUNNING:
throw new IllegalStateException("Cannot execute task:"
+ " the task is already running.");
case FINISHED:
throw new IllegalStateException("Cannot execute task:"
+ " the task has already been executed "
+ "(a task can be executed only once)");
}
}
mStatus = Status.RUNNING;
onPreExecute();
mWorker.mParams = params;
exec.execute(mFuture);
return this;
}
AS it becomes clear from the above code we can call the executeOnExecutor from exec function of Asynctask and in that case it takes a default executor. If we dig into the sourcecode of Asynctask, we will find that this default executor is nothing but a serial executor, the code of which has been given above.
Now lets delve into the SerialExecutor class. In this class we have final ArrayDeque<Runnable> mTasks = new ArrayDeque<Runnable>();.
This actually works as a serializer of the different requests at different threads. This is an example of Half Sync Half Async pattern.
Now lets examine how the serial executor does this. Please have a look at the portion of the code of the SerialExecutor which is written as
if (mActive == null) {
scheduleNext();
}
So when the execute is first called on the Asynctask, this code is executed on the main thread (as mActive will be initialized to NULL) and hence it will take us to the scheduleNext() function.
The ScheduleNext() function has been written as follows:
protected synchronized void scheduleNext() {
if ((mActive = mTasks.poll()) != null) {
THREAD_POOL_EXECUTOR.execute(mActive);
}
}
So in the schedulenext() function we initialize the mActive with the Runnable object which we have already inserted at the end of the dequeue. This Runnable object (which is nothing but the mActive) then is executed on a thread taken from the threadpool. In that thread, then "finally "block gets executed.
Now there are two scenarios.
another Asynctask instance has been created and we call the execute method on it when the first task is being executed.
execute method is called for the second time on a same instance of the Asynctask when the first task is getting executed.
Scenario I : if we look at the execute function of the Serial Executor, we will find that we actually create a new runnable thread (Say thread t) for processing the background task.
Look at the following code snippet-
public synchronized void execute(final Runnable r) {
mTasks.offer(new Runnable() {
public void run() {
try {
r.run();
} finally {
scheduleNext();
}
}
});
As it becomes clear from the line mTasks.offer(new Runnable), every call to the execute function creates a new worker thread. Now probably you are able to find out the similarity between the Half Sync - Half Async pattern and the functioning of SerialExecutor. Let me, however, clarify the doubts. Just like the Half Sync - Half Async pattern's Asynchronous layer, the
mTasks.offer(new Runnable() {
....
}
part of the code creates a new thread the moment execute function is called and push it to the queue (the mTasks). It is done absolutely asynchronously, as the moment it inserts the task in the queue, the function returns. And then background thread executes the task in a synchronous manner. So its similar to the Half Sync - Half Async pattern. Right?
Then inside that thread t, we run the run function of the mActive. But as it is in the try block, the finally will be executed only after the background task is finished in that thread. (Remember both try and finally are happening inside the context of t). Inside finally block, when we call the scheduleNext function, the mActive becomes NULL because we have already emptied the queue. However, if another instance of the same Asynctask is created and we call execute on them, the execute function of these Asynctask won’t be executed because of the synchronization keyword before execute and also because the SERIAL_EXECUTOR is a static instance (hence all the objects of the same class will share the same instance… its an example of class level locking) i mean no instance of the same Async class can preempt the background task that is running in thread t. and even if the thread is interrupted by some events, the finally block which again calls the scheduleNext() function will take care of it.what it all means that there will be only one active thread running the task. this thread may not be the same for different tasks, but only one thread at a time will execute the task. hence the later tasks will be executed one after another only when the first task complets. thats why it is called SerialExecutor.
Scenario II: In this case we will get an exception error. To understand why the execute function cannot be called more than once on the same Asynctask object, please have a look at the below code snippet taken from executorOnExecute function of Asynctask.java especially in the below mentioned portion:
if (mStatus != Status.PENDING) {
switch (mStatus) {
case RUNNING:
throw new IllegalStateException("Cannot execute task:"
+ " the task is already running.");
case FINISHED:
throw new IllegalStateException("Cannot execute task:"
+ " the task has already been executed "
+ "(a task can be executed only once)");
}
}
AS from the above code snippet it becomes clear that if we call execute function twice when a task is in the running status it throws an IllegalStateException saying “Cannot execute task: the task is already running.”.
if we want multiple tasks to be executed parallely, we need to call the execOnExecutor passing Asynctask.THREAD_POOL_EXECUTOR (or maybe an user defined THREAD_POOL as the exec parameter.
You can read my discussion on Asynctask internals here.
AsyncTasks has a fixed size queue internally for storing delayed tasks. The queue size by default is 10. For example if you start 15 your tasks in a row, then first 5 will enter their doInBackground(), but the rest will wait in the queue for free worker thread. As one of the first 5 finishes, and thus releases the worker thread, a task from the queue will start execution. In this case at most 5 tasks will run together.
Yes, there is limit of how many tasks can be run run at a time. So AsyncTask uses thread pool executor with limited max number of the worker threads and the delayed tasks queue use fixed size 10. Max number of worker threads is 128. If you try to execute more than 138 custom tasks your application will throw the RejectedExecutionException.
How many AsyncTask can be started at a time in an Android app?
AsyncTask is backed by a LinkedBlockingQueue with a capacity of 10 (in ICS and gingerbread). So it really depends on how many tasks you are trying to start & how long they take to finish - but it's definitely possible to exhaust the queue's capacity.
When I start 10 AsyncTask, will all tasks will run simultaneously or one by one?
Again, this depends on the platform. The maximum pool size is 128 in both gingerbread and ICS - but the *default behavior* changed between 2.3 and 4.0 - from parallel by default to serial. If you want to execute in parallel on ICS you need to call [executeOnExecutor][1] in conjunction with THREAD_POOL_EXECUTOR
Try switching to the parallel executor and spam it with 75 000 tasks - the serial impl. has an internal ArrayDeque that has no upper capacity bound (except OutOfMemoryExceptions ofc).
Related
I have been reading about Executor in Android documentations. If I understood it correctly, it is used for multiple thread management and it does some of the work for you like spawning new threads when needed. Or you may choose to manage stuff yourself.
In the example below, a group of executors are used instead of one. So it is something like a pool of pool of threads (?).
/**
* Global executor pools for the whole application.
*
* Grouping tasks like this avoids the effects of task starvation (e.g. disk
reads don't wait behind
* webservice requests).
*/
#Singleton
open class AppExecutors(
private val diskIO: Executor,
private val networkIO: Executor,
private val mainThread: Executor
) {
#Inject
constructor() : this(
Executors.newSingleThreadExecutor(),
Executors.newFixedThreadPool(3),
MainThreadExecutor()
)
fun diskIO(): Executor {
return diskIO
}
fun networkIO(): Executor {
return networkIO
}
fun mainThread(): Executor {
return mainThread
}
private class MainThreadExecutor : Executor {
private val mainThreadHandler = Handler(Looper.getMainLooper())
override fun execute(command: Runnable) {
mainThreadHandler.post(command)
}
}
}
Why would one choose to use a group of executors? What do you achieve with it which you can't with just one executor?
That's just structuring and assigning the right executor for the right jobs they might execute:
It's nicely put in a single class for easy reuse.
Three types of executors are employed, each for a specific type of task it could run. Remember that executors have threads to execute jobs or Runnables and each thread the executor creates can run one job at a time:
diskIO is (from the constrcutor) a Executors.newSingleThreadExecutor() since the tasks are best queued and executed one at a time to reduce write and read locks or race conditions for example. Hence a SingleThreadExecutor would run only one task at a time no matter how many are queued to ensure that design. Being a single thread could also mean that it's being used for writing app logs to a file for example which allows for the logs to be written in the proper order as being submitted to the executor. Hence single thread is best at maintaining output as in the order of jobs queued.
networkIO is a Executors.newFixedThreadPool(3) since the tasks are usually network related like connecting to a server on the internet and performing requests or getting data. These tasks usually make the user wait (could be between seconds to minutes) and need to be executed in parallel and fast to make the wait shorter in case many requests need be performed together. Hence the reason there are 3 threads employed with this executor is to assign the tasks among them and execute together. Order of jobs is not a concern here since jobs take different amount of time to execute but what matters the most is that they're running in parallel.
mainThread is a MainThreadExecutor() which in an Android app handles the UI and drawing it. The UI should function smoothly and not lag and hence the reason to use the above two executors is to let any heavy task (like writing a file or performing requests) to run in the background or separately from the mainThread of the app. This executor keeps performing tasks even if the app didn't submit any to it. The tasks it keeps performing is drawing the UI continuously on the screen which constantly repeats. Tasks executed by the mainThread need to lightweight and fast (time they take are in the order of milliseconds), and so any task that slows it down will be noticed as the UI will lag or glitch with it because the mainThread is busy finishing that task instead of drawing and updating the UI. The mainThread here simply uses a Handler which is part of the Android SDK/architecture, is of a single thread type and behaves like an executor (with some differences) that queues tasks to create/update the UI. Only a Handler can perform UI tasks, none of the other executors can.
I am using priority job queue , there are number of jobs running in parallel, so that their result populates on UI at same time which takes application to ANR, is there any way , so that i can run asynchronous calls and populate ui synchronously?
UI is always populated synchronously, if it is done in correct way. The correct way is to call activity.runOnUiThread(Runnable), directly or indirectly. Seems that your problem is that your jobs post to UI thread in a too high rate.
First, check if the Runnables to update UI does only UI work. Any calculations should be done outside the UI thread. If it is so, create an intermediate object which makes pauses between UI updates from the parallel jobs and so lets the UI thread to respond to updates from user. It can look as follows:
public class PauseMaker {
Semaphore sem = new Semaphore(1);
public void runOnUiThread(Runnable r) {
sem.aquire();
Thread.sleep(1);
activity.runOnUiThread(new Runnable(){
try {
r();
} finally {
sem.release();
}
});
}
}
You can use the zip operator of rxjava2 to merge the responses together and when the combined response comes you can populate the UI synchronously .. for reference you can check..
http://www.codexpedia.com/android/rxjava-2-zip-operator-example-in-android/
Note The zipper will the return merged response after all the responses are received
I have a few questions about the the usages of a generic thread framework(for a specific amount of jobs of course) vs the usage of many asyntasks.
I would like to know if it is better to have many asyntasks for small jobs, a handler thread when the job takes a bit longer or it is better to have a generic thread by yourself having generic jobs with a notification system built on top of the the running thread(or a subclass, same story).
My idea is for creating thread that handles different jobs without knowing beforehand which are the jobs. This goes in the direction of creating a sort of small framework for handling different generic jobs.
For example my approach goes in the direction of the code below:
public (abstract if you want to extend and add something on top) class WorkerThread extends Thread {
private static final String TAG = WorkerThread();
private List<WorkTask> syncQueue = new ArrayList< WorkTask >();
private boolean clearQueue = false;
public WorkerThread() {
}
public void stop(boolean clear) {
clearQueue = clear;
this.stopWorker = true;
}
public void addTask(WorkerTask task) {
synchronized (syncQueue) {
if (task != null && !getSynQueue().contains(task)) {
getSynQueue().add(task);
}
}
}
#Override
public void run() {
while (!stopWorker) {
WorkerTask task = null;
synchronized (syncQueue) {
if (!getSynQueue().isEmpty()) {
task = getSynQueue().get(0);
}
}
if (task != null) {
try {
task.run();
synchronized (syncQueue) {
if (!getSynQueue().isEmpty()) {
getSynQueue().remove(task);
//notify something/someone
}
}
} catch (Exception e) {
Log.e(TAG, "Error in running the task." + e.getMessage());
synchronized (syncQueue) {
//again u can notify someone
}
} finally {
//here you can actually notify someone of success
}
}
}
if(clearQueue){
getSynQueue().clear();
}
}
private List<WorkerTask> getSynQueue() {
return this.syncQueue;
}
}
Here the task is the abstract base class that all the jobs extend.
Then on top of this thread or a subclass of this class can be an observer that notifies when something went wrong with the jobs/tasks.
So far, as I know, the pros and cons for my approach are like that:
Thread:
Pros:
1. Long time operations.
2. Centralized.
3. Scalable.
4. Once you have it properly tested it will work smoothly.
Cons
1. Complex architecture.
2. Hard to maintain.
3. Over-engineering for small jobs.
AsynTask:
Pros
1. Easy to be used.
2. Good for short-time operation jobs.
3. Easy to maintain/understand.
Cons
1. Cannot scale that much, you need to stick to doInBackground and onPostExecute.
2. Not good for long-time operation jobs.
If I missed something please correct me.
Final question would be, when the architecture gets a bit big with a lot of requests, short-time, long-time, isn't it better to try and make a generic framework that can handle it both rather than do it with asynctasks there, maybe handlerthread in other parts, etc?
AsyncTask should be used for short operations. From the official documentation:
AsyncTask
AsyncTask is designed to be a helper class around Thread and Handler
and does not constitute a generic threading framework. AsyncTasks
should ideally be used for short operations (a few seconds at the
most.) If you need to keep threads running for long periods of time,
it is highly recommended you use the various APIs provided by the
java.util.concurrent package such as Executor, ThreadPoolExecutor and
FutureTask.
Therefore, if the operation is designed to take longer than a few seconds, it is recommended to use the concurrent package. E.g of operations are, storing a file, a picture, a login, etc.
However, Asynctask has also a few limitations as it can be seen in the link below:
AsyncTask limitations
There is a limit of how many tasks can be run simultaneously. Since AsyncTask uses a thread pool executor with max number of worker threads (128) and the delayed tasks queue has fixed size 10. If you try to execute more than 138 AsyncTasks the app will crash with java.util.concurrent.RejectedExecutionException.
Also, between Api 1.6 and 3.0 there is no way to customize the AsyncTask. It can run tasks in parallel but no customization possible. A
Between API 3.0 and API 4.3.1 there is a default fixed delayed queue size of 10, a minimum number of tasks, 5, and a maximum number of tasks 128. However, after API 3.0, one can define its own executor.
Therefore, if you have a number of minimum 16 tasks to run, the first 5 will start, the next 10 will go into the queue, but from the 16th a new worker thread will be allocated. After version 4.4(kitkat), the number of parallel asynctasks depends on the amount of processors that a device has:
private static final int CPU_COUNT = Runtime.getRuntime().availableProcessors();
private static final int CORE_POOL_SIZE = CPU_COUNT + 1;
private static final int MAXIMUM_POOL_SIZE = CPU_COUNT * 2 + 1;
private static final BlockingQueue<Runnable> sPoolWorkQueue = new LinkedBlockingQueue<Runnable>(128);
The aforementioned details, make the AsyncTask usable only in very specific cases, like loading an image or a file from or to the storage.
On the hand a generic threading framework(or library) can be easily extendable and customizable. Example of libraries is the RxJava that is easily customizable. Using RxJava one can specify task to run on either UI thread and background. Also, for loading/saving pictures asynchronously to/on the sdcard, one can use different libraries like Picasso or Glide.
I have a small Android application in which I need to do some FTP stuff every couple of seconds.
After learning the hard way that running network stuff on the UI thread is something Android does not really like, I've come to this solution:
// This class gets declared inside my Activity
private class CheckFtpTask extends AsyncTask<Void, Void, Void> {
protected Void doInBackground(Void... dummy) {
Thread.currentThread().setName("CheckFtpTask");
// Here I'll do the FTP stuff
ftpStuff();
return null;
}
}
// Member variables inside my activity
private Handler checkFtpHandler;
private Runnable checkFtpRunnable;
// I set up the task later in some of my Activitiy's method:
checkFtpHandler = new Handler();
checkFtpRunnable = new Runnable() {
#Override
public void run() {
new CheckFtpTask().execute((Void[])null);
checkFtpHandler.postDelayed(checkFtpRunnable, 5000);
}
};
checkFtpRunnable.run();
Is this good practice to perform a recurring task that cannot run on the UI thread directly?
Furthermore, instead of creating a new AsyncTask object all the time by calling
new CheckFtpTask().execute((Void[])null);
would it be an option to create the CheckFtpTask object once and then reuse it?
Or will that give me side effects?
Thanks in advance,
Jens.
would it be an option to create the CheckFtpTask object once and then reuse it? Or will that give me side effects?
No, there will be side-effects. Quoting the docs Threading Rules:
The task can be executed only once (an exception will be thrown if a second execution is attempted.)
You will just need to create a separate instance of the task each time you want to run it.
And I'm not sure why you need the Runnable or Handler. AsyncTask has methods that run on the UI Thread (all but doInBackground(), actually) if you need to update the UI.
Check this answer if you need a callback to update the UI when the task has finished.
You should create a new Async task for every call.
See the Android Documentation: AsyncTask. According to this documentation:
The task can be executed only once (an exception will be thrown if a second execution is attempted.)
Specifically check out the threading rules section. There is a similar answer here, https://stackoverflow.com/a/18547396/2728623
Java ThreadPools and the ExecutorFramework allows you to execute threads as needed and reduces thread creation overhead. Check out the singleThreadExecutor. The thread pool usage is pretty easy too!
1.
I have main UI thread and it uses library functions.
MainUI {
library.funcA(1);
library.funcB(2,3);
}
I don't want Main-Thread to wait for the end of each call. Also creating new thread for each call is ineffective and not thread-safe so I want them to run on another same thread.
So I created another thread with a single message queue / handler.
MainUI {
handler.post(new Message(FUNC_A,[1]));
handler.post(new Message(FUNC_B,[2,3]));
}
but it seems having overhead that I need to create new Message object every time and decode it again in handleMessage(). Furthermore It looks ugly to handle the various number of arguments packing/unpacking. I thought making all functions as Runnable and post them to handler, but functions have different arguments and I need to make new classes extending Runnable class!
my current plan is to insert synchronized keywords into every function in library so threads run one by one,
but I dislike many threads creating/disappering so making system heavy (also not thread-safe). I want to hear opinions.
2.
if the library is called from many threads (not only Main-thread), what's the best way to convert calls to run on the library's own single thread?
In my opinion, your current approach is feasible option to make thread-safe. However, if you want to go for having Runnable passed to handler, you could consider overloading your Runnable class. Here is pseudo-code
private Class MyRunner extends Runnable{
public MyRunner(){
//This is without any argument
}
public MyRunner(arg1){
//This is with 1 argument
}
public MyRunner(arg1,arg2){
//This is with 2 arguments
}
}
Still, you will have many instances of MyRunner class created.
You can use a Runnable like this:
int x = 1;
int y = 2;
// "a" and "b" are here to show how variable values can be sent to funcB.
// You could skip this and just have "funcB( 1, 2 )", of course.
final int a = x;
final int b = y;
handler.post( new Runnable() {
public void run() { library.funcB( a, b ); }
} );
The handler bit is a bit trickier. It has to run a looping thread to run each Runnable from the queue. Then it has to have a post method to put things on the queue. Some form of blocking queue would be the simplest way to link the two, saving you from having to do your own Object.wait/Object.notify. Synchronizing the post method would be the easiest way to keep multiple threads from tripping over each other.