Yes it does consuming 100% CPU usage when you try to run it continously but not more than 5 times using Hashmap instance. But when I changed it to ConcurrentHashMap it works well. Hmmm..

import java.util.HashMap;
import java.util.Map;
import java.util.concurrent.ConcurrentHashMap;

/**
* @author henry banzon
* @version 1
* @since 08/17/2010
* @mailto: henry.banzon@pioneer.com.ph
*/

public class HashMapTest{
// occurs 100% cpu usage
Map map = new HashMap();

// runs smoothly on this instance
// Map map = new ConcurrentHashMap();

public class HenryInner1 implements Runnable{
// This thread does the inputs
public void run() {
for(int i = 1; i < 1000; i++){
System.out.println("HenryInner1 : " + 1);
if(i % 5 == 0) map.put("sample" + i, null);
else map.put("sample" + i, "value" + i);
}
}
}
public class HenryInner2 implements Runnable{
// This thread does the retrieval
public void run() {
for(int i = 1; i < 1000; i++){
System.out.println("HenryInner2 : " + 1);
System.out.println(map.get("sample" + i));
map.put("henry" + i, "banzon" + i);
}
}
}
public class HenryInner3 implements Runnable{
// This thread does the retrieval, removal and inputs
public void run() {
for(int i = 1; i < 2000; i++){
System.out.println("HenryInner3 : " + 1);
System.out.println(map.get("henry" + i));
System.out.println(map.remove("sample" + i));
map.put("sample" + i, "value" + i);
}
}
}

public static void main(String[] args) throws InterruptedException {
HashMapTest o = new HashMapTest();
Thread thread1 = new Thread(o.new HenryInner1);
Thread thread2 = new Thread(o.new HenryInner2);
Thread thread3 = new Thread(o.new HenryInner3);
thread1.start();

Thread.sleep(5);
thread2.start();

Thread.sleep(5);
thread3.start();
}

}

Thanks for that article. We faced a similar situation in a job which spawned multiple threads, each of which were doing look-up in a HashMap for certain keys. If not found, the key-value was added in the Map for future lookup (Obviously, it should have been synchronized but it was not). Once in a while we would see the threads getting struck at the call to HashMap.get(). After analysing the code, we figured that making it a Hastable was obviously required but we were not sure this could be causing the threads to be struck. (Data consistency, ofcourse, we knew was compromised). Since the conversion to Hashtable, we have not observed the issue. This article helps explain a bit clearly what can lead to the infitine looping.

From my experience fixing threading problems in the OS code for VMS, Irix, SunOS and VxWorks, no matter how good you are you can never anticipate all the ways threaded code will break. With increasingly complex and common multicore machines, this becomes more true. When documentation says “must” take it as gospel.

This thread was a great read!

For example : #friendfeed .bottom {background:none!important;}

But i agree with you, it’s still difficult to custom this widget …

You can also find advices over here :
http://groups.google.com/group/friendfeed/browse_thread/thread/0e63ed90af1cc018

http://learnbythedrop.com/drop/127

Thanks,
Ashish