Learn about ConcurrentHashMap vs. SynchronizedMap in Java concurrency. Code example & performance analysis. Boost your Java programming skills with Crunchify!
HashMap is a very powerful data structure in Java. We use it everyday and almost in all applications. There are quite a few examples which I have written before on How to Implement Threadsafe cache, How to convert Hashmap to Arraylist?
We used Hashmap in both above examples but those are pretty simple use cases of Hashmap. HashMap is a non-synchronized collection class.
Do you have any of below questions?
- What’s the difference between ConcurrentHashMap and Collections.synchronizedMap(Map)?
- What’s the difference between ConcurrentHashMap and Collections.synchronizedMap(Map) in term of performance?
- ConcurrentHashMap vs Collections.synchronizedMap()
- Popular HashMap and ConcurrentHashMap interview questions
In this tutorial we will go over all above queries and reason why and how we could Synchronize Hashmap?
Why?
The Map object is an associative containers that store elements, formed by a combination of a uniquely identify key and a mapped value. If you have very highly concurrent application in which you may want to modify or read key value in different threads then it’s ideal to use Concurrent Hashmap. Best example is Producer Consumer which handles concurrent read/write.
So what does the thread-safe Map means? If multiple threads access a hash map concurrently, and at least one of the threads modifies the map structurally, it must be synchronized externally to avoid an inconsistent view of the contents.
How?
There are two ways we could synchronized HashMap
- Java Collections synchronizedMap() method
- Use ConcurrentHashMap
//Hashtable Map<String, String> normalMap = new Hashtable<String, String>(); //synchronizedMap synchronizedHashMap = Collections.synchronizedMap(new HashMap<String, String>()); //ConcurrentHashMap concurrentHashMap = new ConcurrentHashMap<String, String>();
ConcurrentHashMap
- You should use ConcurrentHashMap when you need very high concurrency in your project.
- It is thread safe without synchronizing the
whole map. - Reads can happen very fast while write is done with a lock.
- There is no locking at the object level.
- The locking is at a much finer granularity at a hashmap bucket level.
- ConcurrentHashMap doesn’t throw a
ConcurrentModificationExceptionif one thread tries to modify it while another is iterating over it. - ConcurrentHashMap uses multitude of locks.
SynchronizedHashMap
- Synchronization at Object level.
- Every read/write operation needs to acquire lock.
- Locking the entire collection is a performance overhead.
- This essentially gives access to only one thread to the entire map & blocks all the other threads.
- It may cause contention.
- SynchronizedHashMap returns
Iterator, which fails-fast on concurrent modification.
Detailed Explanation:
Code demonstrates a performance comparison between different types of thread-safe maps: HashTable, synchronizedMap, and ConcurrentHashMap. It measures the time it takes to perform a set of operations (addition and retrieval of entries) on each of these map implementations using multiple threads.
Breakdown of the code:
- Import Statements: The code includes necessary import statements for various Java classes and interfaces.
- Class and Field Definitions:
CrunchifyConcurrentHashMapVsSynchronizedMapis the main class that contains the performance testing logic.THREAD_POOL_SIZEdefines the number of threads to be used in the thread pool for testing.
- Main Method (
main):- The
mainmethod is where the program starts execution. - Three different types of maps are tested: Hashtable, synchronizedMap, and ConcurrentHashMap.
- For each map type, the
crunchifyPerformTestmethod is called to measure its performance.
- The
crunchifyPerformTestMethod:- This method takes a thread-safe map (passed as a parameter) and performs a series of tests on it.
- It runs the tests five times to get an average performance result.
- Inside the loop, an
ExecutorServiceis created with a fixed thread pool size (THREAD_POOL_SIZE). - For each thread, a
Runnableis created that adds and retrieves entries from the map.
RunnableImplementation:- The
runmethod of theRunnableperforms the following steps:- Generates a random number (
crunchifyRandomNumber). - Retrieves a value from the map (although the value is not used).
- Puts the random number as both the key and value in the map.
- Generates a random number (
- The
- Timing and Results:
- The start time is recorded using
System.nanoTime()before the threads are started. - After all threads have completed their tasks, the end time is recorded.
- The total time taken to complete the operations is calculated and printed for each test iteration.
- The average time taken for five iterations is calculated and printed for each map type.
- The start time is recorded using
- Console Output:
- The program prints the time it takes to add and retrieve 500,000 entries from each map type.
- After all iterations are completed, the average time for each map type is displayed.
Here is a Java code:
package crunchify.com.tutorials;
import java.util.Collections;
import java.util.HashMap;
import java.util.Hashtable;
import java.util.Map;
import java.util.concurrent.ConcurrentHashMap;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;
import java.util.concurrent.TimeUnit;
/**
* @author Crunchify.com
*
*/
public class CrunchifyConcurrentHashMapVsSynchronizedMap {
public final static int THREAD_POOL_SIZE = 5;
public static Map<String, Integer> crunchifyHashTableObject = null;
public static Map<String, Integer> crunchifySynchronizedMapObject = null;
public static Map<String, Integer> crunchifyConcurrentHashMapObject = null;
public static void main(String[] args) throws InterruptedException {
// Test with Hashtable Object
crunchifyHashTableObject = new Hashtable<String, Integer>();
crunchifyPerformTest(crunchifyHashTableObject);
// Test with synchronizedMap Object
crunchifySynchronizedMapObject = Collections.synchronizedMap(new HashMap<String, Integer>());
crunchifyPerformTest(crunchifySynchronizedMapObject);
// Test with ConcurrentHashMap Object
crunchifyConcurrentHashMapObject = new ConcurrentHashMap<String, Integer>();
crunchifyPerformTest(crunchifyConcurrentHashMapObject);
}
public static void crunchifyPerformTest(final Map<String, Integer> crunchifyThreads) throws InterruptedException {
System.out.println("Test started for: " + crunchifyThreads.getClass());
long averageTime = 0;
for (int i = 0; i < 5; i++) {
long startTime = System.nanoTime();
ExecutorService crunchifyExServer = Executors.newFixedThreadPool(THREAD_POOL_SIZE);
for (int j = 0; j < THREAD_POOL_SIZE; j++) {
crunchifyExServer.execute(new Runnable() {
@SuppressWarnings("unused")
@Override
public void run() {
for (int i = 0; i < 500000; i++) {
Integer crunchifyRandomNumber = (int) Math.ceil(Math.random() * 550000);
// Retrieve value. We are not using it anywhere
Integer crunchifyValue = crunchifyThreads.get(String.valueOf(crunchifyRandomNumber));
// Put value
crunchifyThreads.put(String.valueOf(crunchifyRandomNumber), crunchifyRandomNumber);
}
}
});
}
// Initiates an orderly shutdown in which previously submitted tasks are executed, but no new tasks will be accepted. Invocation
// has no additional effect if already shut down.
// This method does not wait for previously submitted tasks to complete execution. Use awaitTermination to do that.
crunchifyExServer.shutdown();
// Blocks until all tasks have completed execution after a shutdown request, or the timeout occurs, or the current thread is
// interrupted, whichever happens first.
crunchifyExServer.awaitTermination(Long.MAX_VALUE, TimeUnit.DAYS);
long entTime = System.nanoTime();
long totalTime = (entTime - startTime) / 1000000L;
averageTime += totalTime;
System.out.println("500K entried added/retrieved in " + totalTime + " ms");
}
System.out.println("For " + crunchifyThreads.getClass() + " the average time is " + averageTime / 5 + " ms\n");
}
}
shutdown()means the executor service takes no more incoming tasks.awaitTermination()is invoked after a shutdown request.
And hence, you need to first shutdown the serviceExecutor and then block and wait for threads to finish.
Eclipse Console Result:
Test started for: class java.util.Hashtable 500K entried added/retrieved in 1832 ms 500K entried added/retrieved in 1723 ms 500K entried added/retrieved in 1782 ms 500K entried added/retrieved in 1607 ms 500K entried added/retrieved in 1851 ms For class java.util.Hashtable the average time is 1759 ms Test started for: class java.util.Collections$SynchronizedMap 500K entried added/retrieved in 1923 ms 500K entried added/retrieved in 2032 ms 500K entried added/retrieved in 1621 ms 500K entried added/retrieved in 1833 ms 500K entried added/retrieved in 2048 ms For class java.util.Collections$SynchronizedMap the average time is 1891 ms Test started for: class java.util.concurrent.ConcurrentHashMap 500K entried added/retrieved in 1068 ms 500K entried added/retrieved in 1029 ms 500K entried added/retrieved in 1165 ms 500K entried added/retrieved in 840 ms 500K entried added/retrieved in 1017 ms For class java.util.concurrent.ConcurrentHashMap the average time is 1023 ms