Concurrency

Question 1

How to create a thread?

Answer 1

1. Extends the Thread class, override run(). call start()
   2.Implement the Runnable interface, new Thread(new MyRunnable()).start();
   3.ExecutorService pool = Executors.
   newFixedThreadPool(n)
   used when have long running tasks
   newCachedThreadPool()
   used when have big amounts of short tasks
   adv:
2. less cost than create/destroy thread for each task, have ready to use thread, also keep them live
3. Have total thread number control

Question 2

Different types of ExecutorService

Answer 2

1. newCachedThreadPool()
   2.newFixedThreadPool()
   newly added task will be added to a queue when there is no avaliable thread
   3.newScheduledThreadPool()
   4.newSingleThreadExecutor()

Question 3

Different Status of a Thread

Answer 3

NEW
RUNNABLE
RUNNING
BLOCKED
WAITING
TIMED_WAITING

Question 4

Thread join(), yield(), isInterrupted()

Answer 4

t2.join() - When we invoke the join() method on a thread, the calling thread goes into a waiting state. It remains in a waiting state until the referenced thread terminates.

If the referenced thread was already terminated or hasn’t been started, the call to join() method returns immediately.
“Happens-before”: This means that when a thread t1 calls t2.join(), then all changes done by t2 are visible in t1 on return

yield() - no release lock
Thread. sleep() will cause currently executing thread to stop execution and relinquish the CPU to allow Thread scheduler ot allocate CPU to another thread or same thread depends upon Thread scheduler. Thread. yield() also used to relinquish CPU but behavior of sleep() is more determined than yield across platform.

t2.isInterrupted() - check if the flag is true
t2.interrupt() - interrupt t2
Thread.interruped() - check if current thread is interrupted, and clear the status(set to false)

CATCH(InterruptedException){
Thread.currentThread().interrupt();
}

Question 5

Different Status of a Thread

Answer 5

NEW - thread just created
RUNNABLE - cpu allocate stack, PC registers
RUNNING - start execution
BLOCKED - 
waiting for I/O to complete, 
waiting for a  monitor lock

WAITING - waiting indefinitly for another thread to perform an action and give a signal
Object.wait with no timeout
Thread.join with no timeout

TIMED_WAITING
Thread.sleep
Object.wait with no timeout
Thread.join with no timeout

TERMINATED

Question 6

What is Java demon thread?

Answer 6

1. Threads that not prevent JVM exit, mainly to provide foundational service for user thread.
2. GC thread.
3. setDaemon(true)

Question 7

synchronised keyword on object method/static method

Answer 7

object method - lock on the object access

static method - synchronised on all the thread calling the method, as static are stored in medaspace that is shared

Question 8

Lock Vs synchronzed

Answer 8

1. Lock have more unblocking features:
tryLock()
tryLock(time)
lockInterruptly()
newCondition();

2. but can forget to unlock, so always put in finally block

Question 9

Lock Vs synchronzed

Answer 9

1. Lock have more unblocking features:
tryLock()
tryLock(time)
lockInterruptly()
newCondition();  - await/signal/signalAll

2. but can forget to unlock, so always put in finally block

Question 10

Lock Vs synchronzed

Answer 10

1. Lock have more unblocking features:
   tryLock()
   tryLock(time)
   lockInterruptly()
   newCondition(); - await/signal/signalAll
   read/write lock more efficient.
2. but can forget to unlock, so always put in finally block

Question 11

How AtomicXX implemented?

Answer 11

CAS - compare and set operation:

1. There is a M
2. Set M to B only when M ‘s current value is A

Why it is better than lock?
Avoid thread suspend/resume, when casing is failed, just moving to next steps.

Java Atomic variables? - only used when concurrency is  not very high, thread keep retry to add the counter
AtomicInteger
AtomicLong
AtominBoolean
AtomicReference

Question 12

Concurrent Collections

Answer 12

ConcurrentHashMap

CopyOnWriteArrayList

Question 13

Other questions

Answer 13

How to stop executor service?
how executor service works internally?
Java memory model/happen before replation shop

Question 14

What is ThreadLocal ?

Answer 14

1. A class wrapper to wrap create expensive objects
2. Provide per thread per copy.
3. Compare with local variable, reduce the no. of objs created also have thread safety without synchronisation/immutability

Question 15

What is Java memory model?

Answer 15

A set of rule that JVM follow on memory operation/information sharing:

1. Volitine variable rule - every write happens before subsquent read.
2. Monitor lock rule - every unlock happens before lock on that monitor lock

Question 16

What is FutureTask<> ?

Answer 16

1. Wrapper of a runnable

2. can be submitted to executor to execute

Question 17

What happens when submit to a executor when it is already full?

Answer 17

TO_DO

Question 18

Best practice for multi-thread?

Answer 18

1. Using thread factory to give each thread a name
2. prefer concurrent collection over synchronised collection
3. minimize synchronzed block

Question 19

What is fork/join pool in Java?

How to implement?

Answer 19

Definition

1. a implementation of executor service
2. Mainly used for tasks can be break into smaller tasks recursively, to maximum usage of the CPU processing powers

How to implement
1. ForkJoinTask
ForkJoinTask fork() - Arranges to asynchronously execute this task in the pool the current task is running in, if applicable, or using the ForkJoinPool.commonPool() if not inForkJoinPool().

 V join()
Returns the result of the computation when it is done. This method differs from get() in that abnormal completion results in RuntimeException or Error, not ExecutionException, and that interrupts of the calling thread do not cause the method to abruptly return by throwing InterruptedException.

Collection invokeAll(Collection tasks)
Forks all tasks in the specified collection, returning when isDone holds for each task or an (unchecked) exception is encountered, in which case the exception is rethrown.
Need to call join() on each to fetch result

int result = forkJoinPool.invoke(customRecursiveTask); performing this task, awaits its completion if necessary, and returns its result, or throws an (unchecked) RuntimeException or Error if the underlying computation did so.

2. extends RecursiveAction(runnable, no return value)/RecursiveTask(callable return a V ){override compute(){ invokeAll(new task1, new task2)}}

https: //www.pluralsight.com/guides/introduction-to-the-fork-join-framework
https: //www.baeldung.com/java-fork-join

Question 20

Synchronized keyword Vs Lock

Answer 20

https://stackoverflow.com/questions/36371149/reentrantlock-vs-synchronized-on-cpu-level
1. If you’re simply locking an object, I’d prefer to use synchronized
in general consider that synchronized is just an easy-to-use and concise approach of locking.

2.Some time ago ReentrantLock was way faster under certain conditions (high contention for example), but now Java uses different optimizations techniques (like lock coarsening and adaptive locking) to make performance differences in many typical scenarios barely visible to the programmer.

3. Lock provides
4. lock polling, timed lock waits, and interruptible lock waits.
5. Read write lock to improve performance
6. more complex code structures than Synchronized

Why Synchronized is expensive comparing with normal method?

1. read everything from main memory
2. do the job
3. flush the cache to memory
4. prevents compiler as well as cpu reordering of instructions