views:

476

answers:

5

What is it that determines when the garbage collector actually collects? Does it happen after a certain time or after a certain amount of memory have been used up? Or are there other factors?

+4  A: 

It depends a lot on what garbage collector you're actually using, how its tuned, and a whole lot of inputs.

For a run down of the HotSpot Garbage Collector (the common one that comes with Java) and how it's tuned, you can check out this link

tschaible
Sounds interesting, will read it later. :)
nalo
+4  A: 
  • It depends on way program JIT compiled.
  • From outside we cannot definitely tell when it will run.
  • It follows some algorithm which depends on that particular GC.
  • Java virtual machine runs on the client machine with some virtual memory in case of windows default is 4GB. It also depends on that free virtual memory at that particular time.

You can try this small program to check behavior of GC

public class GCTest {

   final int NELEMS = 50000;

   void eatMemory() {

      int[] intArray = new int[NELEMS];

      for (int i=0; i<NELEMS; i++) {
        intArray[i] = i;
      }

   }

   public static void main (String[] args) {

      GCTest gct = new GCTest();

      // Step 1: get a Runtime object
      Runtime r = Runtime.getRuntime();

      // Step 2: determine the current amount of free memory
      long freeMem = r.freeMemory();
      System.out.println("free memory before creating array: " + freeMem);

      // Step 3: consume some memory
      gct.eatMemory();

      // Step 4: determine amount of memory left after consumption
      freeMem = r.freeMemory();
      System.out.println("free memory after creating array:  " + freeMem);

      // Step 5: run the garbage collector, then check freeMemory
      r.gc();
      freeMem = r.freeMemory();
      System.out.println("free memory after running gc():    " + freeMem);
   }
}

possible output -- May be different in your case

free memory before creating array: 4054912
free memory after creating array:  3852496
free memory after running gc():    4064184

Check this link http://www.devdaily.com/java/edu/pj/pj010008/

Xinus
A: 

The garbage collector runs when it needs resources and on a regular basis that you able to influence by telling when is a good time to spend CPU on collecting, using System.gc()

You can help the garbage collector by nulling references explicitly, for instance by giving your objects init() methods that allocate resources and cleanup() methods that explicitly clean up those resources and nulling their references. By nulling references yourself you prevent tye garbage collector from having to find clusters of obects that have to more paths to a root.

rsp
For a modern gc, setting variables to null or not will make no difference and calling System.gc() is really almost always just stupid.
Fredrik
@Fredrik, please don't use terms like "stupid". Setting variables to null is a good practice and it helps the garbage collect to execute in smalls steps instead of bigger chunks. (Mind you I talk about class attributes not temporary method variables on the stack.) As for calling System.gc() trivial applications need not bother with it, but large applications that get massive concurrent requests sometimes need it so smooth the collecting times keeping the application responsive at all times.
rsp
@rsp: The way GCs work it is almost always stupid or at least damaging for gc performance to call System.gc() and it doesn't help at all which is why it is a) not mandatory for the gc to even bother b) possible to turn it off to not be hit by code written by someone who think he is smarter than the gc algoritm. The reason it is still possible is because it is as I said "almost always" stupid, meaning there are a few places where you really might want to do it.Setting variables to null gives you nothing but more code, it will not affect the gc at all.
Fredrik
@rsp: For better understanding about why setting to null doesn't matter in most cases, take a look at the following blog posts. They are not perfect but good enough to explain it for a layman: http://chaoticjava.com/posts/how-does-garbage-collection-work/ and http://chaoticjava.com/posts/parallel-and-concurrent-garbage-collectors/ The only time nulling out something help is if the variable keeping the reference is going to stay alive whereas the referenced data is not needed anymore but that is sort of obvious and not at all like "null every reference yourself".
Fredrik
@FRerik, it's not a matter of thinking to be smarter, it's a matter of hinting the garbage collector "now is a good time to spend some CPU because I don't need it". The garbage collector can use the hint or ignore it, in my experience it can be the difference between an application server that runs smoothly and one that is unresponsive when it does get to do garbage collecting. I agree with you that most applications don't need it but it is good to know. Cleaning up your resources _DOES_ help. I spend months refactoring code from people who say it doesn't halving CPU usage in the process.
rsp
@rsp: well, it doesn't hurt you to null out references for no use so if you are happy doing it please continue doing so (I never said that part was stupid). Regarding System.gc() I think you need to either read up on how concurrent garbage collectors work or just look at the frequency of their work, doing so might give you a hint on how easily you will be able to hit the right millisecond with your "this is a good time to call the gc". The following link is what JRockit say about it and both Sun and IBM share that view: http://tinyurl.com/yfh32sm
Fredrik
I do know how garbage collectors work, its just that sometimes with high volumes of traffic they tend to postpone collecting until they need to do too much in one step getting in the way of business. Funny you mentioned IBM. I recall one particular instance involving their code on their application server, JVM and OS that would hickup if the garbage collector wasn't hinted.
rsp
@rsp: Ok. I can assure you that even IBM have decent garbage collectors nowadays. You shouldn't be too afraid of its JVMs just because you were once hit.
Fredrik
+3  A: 

It runs when it determines that it is time to run. A common strategy in generational garbage collectors is to run the collector when an allocation of generation-0 memory fails. That is, every time you allocate a small block of memory (big blocks are typically placed directly into "older" generations), the system checks whether there's enough free space in the gen-0 heap, and if there isn't, it runs the GC to free up space for the allocation to succeed. Old data is then moved to the gen-1 heap, and when space runs out there, the GC runs a collection on that, upgrading the data which has been there longest to the gen-2 heap, and so on. So the GC doesn't just "run". It might run on the gen-0 heap only (and most collections will do just that), or it might check every generation if it really has to free up a lot of memory (which is only necessary fairly rarely).

But this is far from the only strategy. A concurrent GC runs in the background, cleaning up while the program is running. Some GC's might run as part of every memory allocation. An incremental collector might do that, scanning a few objects at every memory allocation.

The entire point in a garbage collector is that it should just do its thing without requiring any input from the user. So in general, you can't, and shouldn't, predict when it'll run.

I believe Suns JVM gained a generational GC not too long ago (v1.6 maybe? I haven't coded Java for ages, so not sure on this, but I remember being surprised not too long ago, when one of the selling points for the new version was "a generational GC". Not least because .NET has had one since day 1.)

Other JVM's are of course free to pick whichever strategy they like.

EDIT: The above part about Java and generational GC is not true. See below for more details:

The 1.0 and 1.1 Virtual Machines used a mark-sweep collector, which could fragment the heap after a garbage collection. Starting with Java 1.2, the Virtual Machines switched to a generational collector, which has a much better defragmentation behavior (see Java theory and practice: Garbage collection and performance).

So Java actually has a generational GC for ages. What's new in Java 6 is the Garbage-First garbage collector (G1) that is available in Java 6u14. According to the article claiming the release in 1.6.0_14: It is not enabled by default. The parallel collector is still the default GC and is the most efficient GC for common household usage. G1 is meant to be an alternative for the concurrent collector. It is designed to be more predictable and enable fast allocation with memory regions design.

jalf
Java has a generational GC since JDK 1.3 (see e.g. here: http://www.javaworld.com/javaworld/jw-01-2002/jw-0111-hotspotgc.html )
Johannes Weiß
@Johannes Weiß No, it's since JDK 1.2. The 1.0 and 1.1 Virtual Machines used a mark-sweep collector, which could fragment the heap after a garbage collection. Starting with Java 1.2, the Virtual Machines switched to a generational collector, which has a much better defragmentation behaviour. See http://www.ibm.com/developerworks/library/j-jtp01274.html.
Pascal Thivent
@Pascal Thivent: Ok, the article I linked to said that JDK1.2 had an 'exact garbage collector' which was optimized and renamed to 'generational GC' in JDK 1.3 but you may be right...
Johannes Weiß
thanks for correcting me then. :) As I said, I haven't really followed Java development for ages.
jalf
A: 

This completely depends on the actual JVM and what it chooses to do, and is basically out of your hands as a programmer. Greybearded die-hard experts may want to tell the JVM they know better, but for mere mortals this should be considered black magic better left alone.

What should concern you is if it can keep up with the rate that your programs creates and discards objects. If not your whole program is halted while global cleaning up occurs. That turns out to very bad response times, but happens rarely for modern JVM's on modern computers.

If you are curious about what happens in your program and when, then investigate the "jvisualvm" tool in the recent versions of the Java 6 JDK. It is really great for peeking inside.

Thorbjørn Ravn Andersen