Related
What are the "best practices" for creating (and releasing) millions of small objects?
I am writing a chess program in Java and the search algorithm generates a single "Move" object for each possible move, and a nominal search can easily generate over a million move objects per second. The JVM GC has been able to handle the load on my development system, but I'm interested in exploring alternative approaches that would:
Minimize the overhead of garbage collection, and
reduce the peak memory footprint for lower-end systems.
A vast majority of the objects are very short-lived, but about 1% of the moves generated are persisted and returned as the persisted value, so any pooling or caching technique would have to provide the ability to exclude specific objects from being re-used.
I don't expect fully-fleshed out example code, but I would appreciate suggestions for further reading/research, or open source examples of a similar nature.
Run the application with verbose garbage collection:
java -verbose:gc
And it will tell you when it collects. There would be two types of sweeps, a fast and a full sweep.
[GC 325407K->83000K(776768K), 0.2300771 secs]
[GC 325816K->83372K(776768K), 0.2454258 secs]
[Full GC 267628K->83769K(776768K), 1.8479984 secs]
The arrow is before and after size.
As long as it is just doing GC and not a full GC you are home safe. The regular GC is a copy collector in the 'young generation', so objects that are no longer referenced are simply just forgotten about, which is exactly what you would want.
Reading Java SE 6 HotSpot Virtual Machine Garbage Collection Tuning is probably helpful.
Since version 6, the server mode of JVM employs an escape analysis technique. Using it you can avoid GC all together.
Well, there are several questions in one here !
1 - How are short-lived objects managed ?
As previously stated, the JVM can perfectly deal with a huge amount of short lived object, since it follows the Weak Generational Hypothesis.
Note that we are speaking of objects that reached the main memory (heap). This is not always the case. A lot of objects you create does not even leave a CPU register. For instance, consider this for-loop
for(int i=0, i<max, i++) {
// stuff that implies i
}
Let's not think about loop unrolling (an optimisations that the JVM heavily performs on your code). If max is equal to Integer.MAX_VALUE, you loop might take some time to execute. However, the i variable will never escape the loop-block. Therefore the JVM will put that variable in a CPU register, regularly increment it but will never send it back to the main memory.
So, creating millions of objects are not a big deal if they are used only locally. They will be dead before being stored in Eden, so the GC won't even notice them.
2 - Is it useful to reduce the overhead of the GC ?
As usual, it depends.
First, you should enable GC logging to have a clear view about what is going on. You can enable it with -Xloggc:gc.log -XX:+PrintGCDetails.
If your application is spending a lot of time in a GC cycle, then, yes, tune the GC, otherwise, it might not be really worth it.
For instance, if you have a young GC every 100ms that takes 10ms, you spend 10% of your time in the GC, and you have 10 collections per second (which is huuuuuge). In such a case, I would not spend any time in GC tuning, since those 10 GC/s would still be there.
3 - Some experience
I had a similar problem on an application that was creating a huge amount of a given class. In the GC logs, I noticed that the creation rate of the application was around 3 GB/s, which is way too much (come on... 3 gigabytes of data every second ?!).
The problem : Too many frequent GC caused by too many objects being created.
In my case, I attached a memory profiler and noticed that a class represented a huge percentage of all my objects. I tracked down the instantiations to find out that this class was basically a pair of booleans wrapped in an object. In that case, two solutions were available :
Rework the algorithm so that I do not return a pair of booleans but instead I have two methods that return each boolean separately
Cache the objects, knowing that there were only 4 different instances
I chose the second one, as it had the least impact on the application and was easy to introduce. It took me minutes to put a factory with a not-thread-safe cache (I did not need thread safety since I would eventually have only 4 different instances).
The allocation rate went down to 1 GB/s, and so did the frequency of young GC (divided by 3).
Hope that helps !
If you have just value objects (that is, no references to other objects) and really but I mean really tons and tons of them, you can use direct ByteBuffers with native byte ordering [the latter is important] and you need some few hundred lines of code to allocate/reuse + getter/setters. Getters look similar to long getQuantity(int tupleIndex){return buffer.getLong(tupleInex+QUANTITY_OFFSSET);}
That would solve the GC problem almost entirely as long as you do allocate once only, that is, a huge chunk and then manage the objects yourself. Instead of references you'd have only index (that is, int) into the ByteBuffer that has to be passed along. You may need to do the memory align yourself as well.
The technique would feel like using C and void*, but with some wrapping it's bearable. A performance downside could be bounds checking if the compiler fails to eliminate it. A major upside is the locality if you process the tuples like vectors, the lack of the object header reduces the memory footprint as well.
Other than that, it's likely you'd not need such an approach as the young generation of virtually all JVM dies trivially and the allocation cost is just a pointer bump. Allocation cost can be a bit higher if you use final fields as they require memory fence on some platforms (namely ARM/Power), on x86 it is free, though.
Assuming you find GC is an issue (as others point out it might not be) you will be implementing your own memory management for you special case i.e. a class which suffers massive churn. Give object pooling a go, I've seen cases where it works quite well. Implementing object pools is a well trodden path so no need to re-visit here, look out for:
multi-threading: using thread local pools might work for your case
backing data structure: consider using ArrayDeque as it performs well on remove and has no allocation overhead
limit the size of your pool :)
Measure before/after etc,etc
I've met a similar problem. First of all, try to reduce the size of the small objects. We introduced some default field values referencing them in each object instance.
For example, MouseEvent has a reference to Point class. We cached Points and referenced them instead of creating new instances. The same for, for example, empty strings.
Another source was multiple booleans which were replaced with one int and for each boolean we use just one byte of the int.
I dealt with this scenario with some XML processing code some time ago. I found myself creating millions of XML tag objects which were very small (usually just a string) and extremely short-lived (failure of an XPath check meant no-match so discard).
I did some serious testing and came to the conclusion that I could only achieve about a 7% improvement on speed using a list of discarded tags instead of making new ones. However, once implemented I found that the free queue needed a mechanism added to prune it if it got too big - this completely nullified my optimisation so I switched it to an option.
In summary - probably not worth it - but I'm glad to see you are thinking about it, it shows you care.
Given that you are writing a chess program there are some special techniques you can use for decent performance. One simple approach is to create a large array of longs (or bytes) and treat it as a stack. Each time your move generator creates moves it pushes a couple of numbers onto the stack, e.g. move from square and move to square. As you evaluate the search tree you will be popping off moves and updating a board representation.
If you want expressive power use objects. If you want speed (in this case) go native.
One solution I've used for such search algorithms is to create just one Move object, mutate it with new move, and then undo the move before leaving the scope. You are probably analyzing just one move at a time, and then just storing the best move somewhere.
If that's not feasible for some reason, and you want to decrease peak memory usage, a good article about memory efficiency is here: http://www.cs.virginia.edu/kim/publicity/pldi09tutorials/memory-efficient-java-tutorial.pdf
Just create your millions of objects and write your code in the proper way: don't keep unnecessary references to these objects. GC will do the dirty job for you. You can play around with verbose GC as mentioned to see if they are really GC'd. Java IS about creating and releasing objects. :)
I think you should read about stack allocation in Java and escape analysis.
Because if you go deeper into this topic you may find that your objects are not even allocated on the heap, and they are not collected by GC the way that objects on the heap are.
There is a wikipedia explanation of escape analysis, with example of how this works in Java:
http://en.wikipedia.org/wiki/Escape_analysis
I am not a big fan of GC, so I always try finding ways around it. In this case I would suggest using Object Pool pattern:
The idea is to avoid creating new objects by store them in a stack so you can reuse it later.
Class MyPool
{
LinkedList<Objects> stack;
Object getObject(); // takes from stack, if it's empty creates new one
Object returnObject(); // adds to stack
}
Object pools provide tremendous (sometime 10x) improvements over object allocation on the heap. But the above implementation using a linked list is both naive and wrong! The linked list creates objects to manage its internal structure nullifying the effort.
A Ringbuffer using an array of objects work well. In the example give (a chess programm managing moves) the Ringbuffer should be wrapped into a holder object for the list of all computed moves. Only the moves holder object references would then be passed around.
Is it possible to mark java objects non-collectable from gc perspective to save on gc-sweep time?
Something along the lines of http://wwwasd.web.cern.ch/wwwasd/lhc++/Objectivity/V5.2/Java/guide/jgdStorage.fm.html and specifically non-garbage-collectible containers there (non-garbage-collectable?).
The problem is that I have lots of ordinary temporary objects, but I have even bigger (several Gigs) of objects that are stored for Cache purposes. For no reason should the Java GC traverse all those Cache gigabytes trying to find anything to collect, because they contain cached data which have their own timeouts.
This way I could partition my data in a custom way into infinite-lived and normal-lived objects, and hopefully GC would be quite fast because normal objects don't live so long and amount to smaller amounts.
There are some workarounds to this problem, such as Apache DirectMemory and Commercial Terracotta BigMemory(http://terracotta.org/products/bigmemory), but a java-native solution would be nicer (I mean free and probably more reliable?). Also I want to avoid serialization overhead which means it should happen within same jvm. To my understanding DirectMemory and BigMemory operate mainly off heap which means that the objects must be serialized/deserialized to/from memory outside jvm. Simply marking non-gc regions within the jvm would seem a better solution. Using Files for cache is not an option either, it has the same unaffordable serialization/deserialization overhead - use case is a HA server with lots of data used in random (human) order and low latency needed.
Any memory the JVM manages is also garbage-collected by the JVM. And any “live” objects which are directly available to Java methods without deserialization have to live in JVM memory. Therefore in my understanding you cannot have live objects which are immune to garbage collection.
On the other hand, the usage you describe should make the generational approach to garbage collection quite efficient. If your big objects stay around for a while, they will be checked for reclamation less often. So I doubt there is much to be gained from avoiding those checks.
Is it possible to mark java objects non-collectable from gc perspective to save on gc-sweep time?
No it is not possible.
You can prevent objects from being garbage collected by keeping them reachable, but the GC will still need to trace them to check reachability on each full; GC (at least).
Is simply my assumption, that when the jvm is starving it begins scanning all those unnecessary objects too.
Yes. That is correct. However, unless you've got LOTS of objects that you want to be treated this way, the overhead is likely to be insignificant. (And anyway, a better idea is to give the JVM more memory ... if that is possible.)
Quite simply, for you to be able to do this, the garbage collection algorithm would need to be aware of such a flag, and take it into account when doing its work.
I'm not aware of any of the standard GC algorithms having such a flag, so for this to work you would need to write your own GC algorithm (after deciding on some feasible way to communicate this information to it).
In principle, in fact, you've already started down this track - you're deciding how garbage collection should be done rather than being happy to leaving it to the JVM's GC algo. Is the situation you describe a measurable problem for you; something for which the existing garbage collection is insufficient, but your plan would work? Garbage collectors are extremely well-tuned, so I wouldn't be surprised if the "inefficient" default strategy is actually faster than your naively-optimal one.
(Doing manual memory management is tricky and error-prone at the best of times; managing some memory yourself while using a stock garbage collector to handle the rest seems even worse. I expect you'd run into a lot of edge cases where the GC assumes it "knows" what's happening with the whole heap, which would no longer be true. Steer clear if you can...)
The recommended approaches would be to use either a commerical RTSJ implementation to avoid GC, or to use off heap memory. One could also look into soft references for caches as well (they do get collected).
This is not recommended:
If for some reason you do not believe these options are sufficient, you could look into direct memory access which is UNSAFE (part of sun.misc.Unsafe). You can use the 'theUnsafe' field to get the 'Unsafe' instance. Unsafe allows to allocation/deallocate memory via 'allocateMemory' and 'freeMemory'. This is not under GC control nor limited by JVM heap size. The impact on GC/application, once you go down this route, is not guaranteed - which is why using byte buffers might be the way to go (if you're not using a RTSJ like implementation).
Hope this helps.
Living Java objects will always be part of the GC life cycle. Or said another way, marking an object to be non-gc is the same order of overhead than having your object referenced by a root reference (a static final map for instance).
But thinking a bit further, data put in a cache are most likely to be temporary, and would eventually be evicted. At that point you will start again to like the JVM and the GC.
If you have 100's of GBs of permanent data, you may want to rethink the architecture of your application, and try to shard and distribute your data (horizontally scalability).
Last but not least, lots of work has been done around serialization, and the overhead of serialization (I'm not speaking about the poor reputation of ObjectInputStream and ObjectOutputStream) is not that big.
More than that, if your data is mainly composed of primitive types (including bytes array), there is efficient way to readInt() or readBytes() from off heap buffers (for instannce netty.io's ChannelBuffer). This could be a way to go.
I'm working with a program that runs lengthy SQL queries and stores the processed results in a HashMap. Currently, to get around the slow execution time of each of the 20-200 queries, I am using a fixed thread pool and a custom callable to do the searching. As a result, each callable is creating a local copy of the data which it then returns to the main program to be included in the report.
I've noticed that 100 query reports, which used to run without issue, now cause me to run out of memory. My speculation is that because these callables are creating their own copy of the data, I'm doubling memory usage when I join them into another large HashMap. I realize I could try to coax the garbage collector to run by attempting to reduce the scope of the callable's table, but that level of restructuring is not really what I want to do if it's possible to avoid.
Could I improve memory usage by replacing the callables with runnables that instead of storing the data, write it to a concurrent HashMap? Or does it sound like I have some other problem here?
Don't create copy of data, just pass references around, ensuring thread safety if needed. If without data copying you still have OOM, consider increasing max available heap for application.
Drawback of above approach not using copy of data is that thread safety is harder to achieve, though.
Do you really need all 100-200 reports at the same time?
May be it's worth to limit the 1st level of caching by just 50 reports and introduce a 2nd level based on WeakHashMap?
When 1st level exceeds its size LRU will be pushed to the 2nd level which will depend on the amount of available memory (with use of WeakHashMap).
Then to search for reports you will first need to query 1st level, if value is not there query 2nd level and if value is not there then report was reclaimed by GC when there was not enough memory and you have to query DB again for this report.
Do the results of the queries depend on other query results? If not, whenever you discover the results in another thread, just use a ConcurrentHashMap like you are implying. Do you really need to ask if creating several unnecessary copies of data is causing your program to run out of memory? This should almost be obvious.
I have a web app that serializes a java bean into xml or json according to the user request.
I am facing a mind bending problem when I put a little bit of load on it, it quickly uses all allocated memory, and reach max capacity. I then observe full GC working really hard every 20-40 seconds.
Doesnt look like a memory leak issue... but I am not quite sure how to trouble shoot this?
The bean that is serialized to xml/json has reference to other beans and those to others. I use json-lib and jaxb to serialize the beans.
yourkit memory profiler is telling me that a char[] is the most memory consuming live object...
any insight is appreciated.
There are two possibilities: you've got a memory leak, or your webapp is just generating lots of garbage.
The brute-force way to tell if you've got a memory leak is to run it for a long time and see if it falls over with an OOME. Or turn on GC logging, and see if the average space left after garbage collection continually trends upwards over time.
Whether or not you have a memory leak, you can probably improve performance (reduce the percentage GC time) by increasing the max heap size. The fact that your webapp is seeing lots of full GCs suggests to me that it needs more heap. (This is just a bandaid solution if you have a memory leak.)
If it turns out that you are not suffering from a memory leak, then you should take a look at why your application is generating so much garbage. It could be down to the way that you are doing the XML and JSON serialization.
Why do you think you have a problem? GC is a natural and normal thing to happen. We have customers that GC every second (for less than 100ms duration), and that's fine as long as memory keeps getting reclaimed.
GCing every 20-40 seconds isn't a problem IMO - as long as it doesn't take a large % of that 20-40s. Most major commercial JVMs aim to keep GC in the 5-10% of time range (so 1-4 seconds of that 20-40s). Posting more data in the form of the GC logs might help, and I'd also suggest tools like GCMV would help you visualize and get recommendations on what your GC profile looks like.
It's impossible to diagnose this without a lot more information - code and GC logs - but my guess would be that you're reading data in as large strings, then chopping out little bits with substring(). When you do that, the substring string is made using the same underlying character array as the parent string, and so as long as it's alive, will keep that array in memory. That means code like this:
String big = a string of one million characters;
String small = big.substring(0, 1);
big = null;
Will still keep the huge string's character data in memory. If this is the case, then you can address it by forcing the small strings to use fresh, smaller, character arrays by constructing new instances:
small = new String(small);
But like i said, this is just a guess.
I'm not sure how much of it is in your code and how much might be in the tools you are using, but there are some key things to watch for.
One of the worst is if you constantly add to strings in loops. A simple "hello"+"world" is no problem at all, it's actually very smart about that, but if you do it in a loop it will constantly reallocate the string. Use StringBuilder where you can.
There are profilers for Java that should quickly point you to where the allocations are taking place. Just fool around with a profiler for a while while your java app is running and you will probably be able to reduce your GCs to virtually nothing unless the problem is inside your libraries--and even then you may figure out some way to fix it.
Things you allocate and then free quickly don't require time in the GC phase--it's pretty much free. Be sure you aren't keeping Strings around longer than you need them. Bring them in, process them and return to your previous state before returning from your request handler.
You should attach yourkit and record allocations (e.g., every 10th allocation; including all large ones). They have a step by step guide on diagnosing excessive gc:
http://www.yourkit.com/docs/90/help/excessive_gc.jsp
To me that sounds like you are trying to serialize a recursive object by some encoder which is not prepared for it.
(or at least: very deep/almost recursive)
Java's native XML API is really "noisy" and generally wasteful in terms of resources which means that if your requests and XML/JSON generation cycles are short-lived, the GC will have lots to clean up for.
I have debugged a very similar case and found out this the hard way, only way I could at least somewhat improve the situation without major refactorings was implicitly calling GC with the appropriate VM flags which actually turn System.gc(); from a non-op call to maybe-op call.
I would start by inspecting my running application to see what was being created on the heap.
HPROF can collect this information for you, which you can then analyse using HAT.
To debug issues with memory allocations, InMemProfiler can be used at the command line. Collected object allocations can be tracked and collected objects can be split into buckets based on their lifetimes.
In trace mode this tool can be used to identify the source of memory allocations.
After answering a question about how to force-free objects in Java (the guy was clearing a 1.5GB HashMap) with System.gc(), I was told it's bad practice to call System.gc() manually, but the comments were not entirely convincing. In addition, no one seemed to dare to upvote, nor downvote my answer.
I was told there that it's bad practice, but then I was also told that garbage collector runs don't systematically stop the world anymore, and that it could also effectively be used by the JVM only as a hint, so I'm kind of at loss.
I do understand that the JVM usually knows better than you when it needs to reclaim memory. I also understand that worrying about a few kilobytes of data is silly. I also understand that even megabytes of data isn't what it was a few years back. But still, 1.5 gigabytes? And you know there's like 1.5 GB of data hanging around in memory; it's not like it's a shot in the dark. Is System.gc() systematically bad, or is there some point at which it becomes okay?
So the question is actually double:
Why is or isn't it bad practice to call System.gc()? Is it really merely a hint to the JVM under certain implementations, or is it always a full collection cycle? Are there really garbage collector implementations that can do their work without stopping the world? Please shed some light over the various assertions people have made in the comments to my answer.
Where's the threshold? Is it never a good idea to call System.gc(), or are there times when it's acceptable? If so, what are those times?
The reason everyone always says to avoid System.gc() is that it is a pretty good indicator of fundamentally broken code. Any code that depends on it for correctness is certainly broken; any that rely on it for performance are most likely broken.
You don't know what sort of garbage collector you are running under. There are certainly some that do not "stop the world" as you assert, but some JVMs aren't that smart or for various reasons (perhaps they are on a phone?) don't do it. You don't know what it's going to do.
Also, it's not guaranteed to do anything. The JVM may just entirely ignore your request.
The combination of "you don't know what it will do," "you don't know if it will even help," and "you shouldn't need to call it anyway" are why people are so forceful in saying that generally you shouldn't call it. I think it's a case of "if you need to ask whether you should be using this, you shouldn't"
EDIT to address a few concerns from the other thread:
After reading the thread you linked, there's a few more things I'd like to point out.
First, someone suggested that calling gc() may return memory to the system. That's certainly not necessarily true - the Java heap itself grows independently of Java allocations.
As in, the JVM will hold memory (many tens of megabytes) and grow the heap as necessary. It doesn't necessarily return that memory to the system even when you free Java objects; it is perfectly free to hold on to the allocated memory to use for future Java allocations.
To show that it's possible that System.gc() does nothing, view
JDK bug 6668279
and in particular that there's a -XX:DisableExplicitGC VM option:
By default calls to System.gc() are enabled (-XX:-DisableExplicitGC). Use -XX:+DisableExplicitGC to disable calls to System.gc(). Note that the JVM still performs garbage collection when necessary.
It has already been explained that calling system.gc() may do nothing, and that any code that "needs" the garbage collector to run is broken.
However, the pragmatic reason that it is bad practice to call System.gc() is that it is inefficient. And in the worst case, it is horribly inefficient! Let me explain.
A typical GC algorithm identifies garbage by traversing all non-garbage objects in the heap, and inferring that any object not visited must be garbage. From this, we can model the total work of a garbage collection consists of one part that is proportional to the amount of live data, and another part that is proportional to the amount of garbage; i.e. work = (live * W1 + garbage * W2).
Now suppose that you do the following in a single-threaded application.
System.gc(); System.gc();
The first call will (we predict) do (live * W1 + garbage * W2) work, and get rid of the outstanding garbage.
The second call will do (live* W1 + 0 * W2) work and reclaim nothing. In other words we have done (live * W1) work and achieved absolutely nothing.
We can model the efficiency of the collector as the amount of work needed to collect a unit of garbage; i.e. efficiency = (live * W1 + garbage * W2) / garbage. So to make the GC as efficient as possible, we need to maximize the value of garbage when we run the GC; i.e. wait until the heap is full. (And also, make the heap as big as possible. But that is a separate topic.)
If the application does not interfere (by calling System.gc()), the GC will wait until the heap is full before running, resulting in efficient collection of garbage1. But if the application forces the GC to run, the chances are that the heap won't be full, and the result will be that garbage is collected inefficiently. And the more often the application forces GC, the more inefficient the GC becomes.
Note: the above explanation glosses over the fact that a typical modern GC partitions the heap into "spaces", the GC may dynamically expand the heap, the application's working set of non-garbage objects may vary and so on. Even so, the same basic principal applies across the board to all true garbage collectors2. It is inefficient to force the GC to run.
1 - This is how the "throughput" collector works. Concurrent collectors such as CMS and G1 use different criteria to decide when to start the garbage collector.
2 - I'm also excluding memory managers that use reference counting exclusively, but no current Java implementation uses that approach ... for good reason.
Lots of people seem to be telling you not to do this. I disagree. If, after a large loading process like loading a level, you believe that:
You have a lot of objects that are unreachable and may not have been gc'ed. and
You think the user could put up with a small slowdown at this point
there is no harm in calling System.gc(). I look at it like the c/c++ inline keyword. It's just a hint to the gc that you, the developer, have decided that time/performance is not as important as it usually is and that some of it could be used reclaiming memory.
Advice to not rely on it doing anything is correct. Don't rely on it working, but giving the hint that now is an acceptable time to collect is perfectly fine. I'd rather waste time at a point in the code where it doesn't matter (loading screen) than when the user is actively interacting with the program (like during a level of a game.)
There is one time when i will force collection: when attempting to find out is a particular object leaks (either native code or large, complex callback interaction. Oh and any UI component that so much as glances at Matlab.) This should never be used in production code.
People have been doing a good job explaining why NOT to use, so I will tell you a couple situations where you should use it:
(The following comments apply to Hotspot running on Linux with the CMS collector, where I feel confident saying that System.gc() does in fact always invoke a full garbage collection).
After the initial work of starting up your application, you may be a terrible state of memory usage. Half your tenured generation could be full of garbage, meaning that you are that much closer to your first CMS. In applications where that matters, it is not a bad idea to call System.gc() to "reset" your heap to the starting state of live data.
Along the same lines as #1, if you monitor your heap usage closely, you want to have an accurate reading of what your baseline memory usage is. If the first 2 minutes of your application's uptime is all initialization, your data is going to be messed up unless you force (ahem... "suggest") the full gc up front.
You may have an application that is designed to never promote anything to the tenured generation while it is running. But maybe you need to initialize some data up-front that is not-so-huge as to automatically get moved to the tenured generation. Unless you call System.gc() after everything is set up, your data could sit in the new generation until the time comes for it to get promoted. All of a sudden your super-duper low-latency, low-GC application gets hit with a HUGE (relatively speaking, of course) latency penalty for promoting those objects during normal operations.
It is sometimes useful to have a System.gc call available in a production application for verifying the existence of a memory leak. If you know that the set of live data at time X should exist in a certain ratio to the set of live data at time Y, then it could be useful to call System.gc() a time X and time Y and compare memory usage.
This is a very bothersome question, and I feel contributes to many being opposed to Java despite how useful of a language it is.
The fact that you can't trust "System.gc" to do anything is incredibly daunting and can easily invoke "Fear, Uncertainty, Doubt" feel to the language.
In many cases, it is nice to deal with memory spikes that you cause on purpose before an important event occurs, which would cause users to think your program is badly designed/unresponsive.
Having ability to control the garbage collection would be very a great education tool, in turn improving people's understanding how the garbage collection works and how to make programs exploit it's default behavior as well as controlled behavior.
Let me review the arguments of this thread.
It is inefficient:
Often, the program may not be doing anything and you know it's not doing anything because of the way it was designed. For instance, it might be doing some kind of long wait with a large wait message box, and at the end it may as well add a call to collect garbage because the time to run it will take a really small fraction of the time of the long wait but will avoid gc from acting up in the middle of a more important operation.
It is always a bad practice and indicates broken code.
I disagree, it doesn't matter what garbage collector you have. Its' job is to track garbage and clean it.
By calling the gc during times where usage is less critical, you reduce odds of it running when your life relies on the specific code being run but instead it decides to collect garbage.
Sure, it might not behave the way you want or expect, but when you do want to call it, you know nothing is happening, and user is willing to tolerate slowness/downtime. If the System.gc works, great! If it doesn't, at least you tried. There's simply no down side unless the garbage collector has inherent side effects that do something horribly unexpected to how a garbage collector is suppose to behave if invoked manually, and this by itself causes distrust.
It is not a common use case:
It is a use case that cannot be achieved reliably, but could be if the system was designed that way. It's like making a traffic light and making it so that some/all of the traffic lights' buttons don't do anything, it makes you question why the button is there to begin with, javascript doesn't have garbage collection function so we don't scrutinize it as much for it.
The spec says that System.gc() is a hint that GC should run and the VM is free to ignore it.
what is a "hint"? what is "ignore"? a computer cannot simply take hints or ignore something, there are strict behavior paths it takes that may be dynamic that are guided by the intent of the system. A proper answer would include what the garbage collector is actually doing, at implementation level, that causes it to not perform collection when you request it. Is the feature simply a nop? Is there some kind of conditions that must me met? What are these conditions?
As it stands, Java's GC often seems like a monster that you just don't trust. You don't know when it's going to come or go, you don't know what it's going to do, how it's going to do it. I can imagine some experts having better idea of how their Garbage Collection works on per-instruction basis, but vast majority simply hopes it "just works", and having to trust an opaque-seeming algorithm to do work for you is frustrating.
There is a big gap between reading about something or being taught something, and actually seeing the implementation of it, the differences across systems, and being able to play with it without having to look at the source code. This creates confidence and feeling of mastery/understanding/control.
To summarize, there is an inherent problem with the answers "this feature might not do anything, and I won't go into details how to tell when it does do something and when it doesn't and why it won't or will, often implying that it is simply against the philosophy to try to do it, even if the intent behind it is reasonable".
It might be okay for Java GC to behave the way it does, or it might not, but to understand it, it is difficult to truly follow in which direction to go to get a comprehensive overview of what you can trust the GC to do and not to do, so it's too easy simply distrust the language, because the purpose of a language is to have controlled behavior up to philosophical extent(it's easy for a programmer, especially novices to fall into existential crisis from certain system/language behaviors) you are capable of tolerating(and if you can't, you just won't use the language until you have to), and more things you can't control for no known reason why you can't control them is inherently harmful.
Sometimes (not often!) you do truly know more about past, current and future memory usage then the run time does. This does not happen very often, and I would claim never in a web application while normal pages are being served.
Many year ago I work on a report generator, that
Had a single thread
Read the “report request” from a queue
Loaded the data needed for the report from the database
Generated the report and emailed it out.
Repeated forever, sleeping when there were no outstanding requests.
It did not reuse any data between reports and did not do any cashing.
Firstly as it was not real time and the users expected to wait for a report, a delay while the GC run was not an issue, but we needed to produce reports at a rate that was faster than they were requested.
Looking at the above outline of the process, it is clear that.
We know there would be very few live objects just after a report had been emailed out, as the next request had not started being processed yet.
It is well known that the cost of running a garbage collection cycle is depending on the number of live objects, the amount of garbage has little effect on the cost of a GC run.
That when the queue is empty there is nothing better to do then run the GC.
Therefore clearly it was well worth while doing a GC run whenever the request queue was empty; there was no downside to this.
It may be worth doing a GC run after each report is emailed, as we know this is a good time for a GC run. However if the computer had enough ram, better results would be obtained by delaying the GC run.
This behaviour was configured on a per installation bases, for some customers enabling a forced GC after each report greatly speeded up the production of reports. (I expect this was due to low memory on their server and it running lots of other processes, so hence a well time forced GC reduced paging.)
We never detected an installation that did not benefit from a forced GC run every time the work queue was empty.
But, let be clear, the above is not a common case.
These days I would be more inclined to run each report in a seperate process leaving the operating system to clear up memory rather then the garbage collector and having the custom queue manager service use mulple working processes on large servers.
GC efficiency relies on a number of heuristics. For instance, a common heuristic is that write accesses to objects usually occur on objects which were created not long ago. Another is that many objects are very short-lived (some objects will be used for a long time, but many will be discarded a few microseconds after their creation).
Calling System.gc() is like kicking the GC. It means: "all those carefully tuned parameters, those smart organizations, all the effort you just put into allocating and managing the objects such that things go smoothly, well, just drop the whole lot, and start from scratch". It may improve performance, but most of the time it just degrades performance.
To use System.gc() reliably(*) you need to know how the GC operates in all its fine details. Such details tend to change quite a bit if you use a JVM from another vendor, or the next version from the same vendor, or the same JVM but with slightly different command-line options. So it is rarely a good idea, unless you want to address a specific issue in which you control all those parameters. Hence the notion of "bad practice": that's not forbidden, the method exists, but it rarely pays off.
(*) I am talking about efficiency here. System.gc() will never break a correct Java program. It will neither conjure extra memory that the JVM could not have obtained otherwise: before throwing an OutOfMemoryError, the JVM does the job of System.gc(), even if as a last resort.
Maybe I write crappy code, but I've come to realize that clicking the trash-can icon on eclipse and netbeans IDEs is a 'good practice'.
Some of what I am about to write is simply a summarization of what has already been written in other answers, and some is new.
The question "Why is it bad practice to call System.gc()?" does not compute. It assumes that it is bad practice, while it is not. It greatly depends on what you are trying to accomplish.
The vast majority of programmers out there have no need for System.gc(), and it will never do anything useful to them in the vast majority of use cases. So, for the majority, calling it is bad practice because it will not do whatever it is that they think it will do, it will only add overhead.
However, there are a few rare cases where invoking System.gc() is actually beneficial:
When you are absolutely sure that you have some CPU time to spare now, and you want to improve the throughput of code that will run later. For example, a web server that discovers that there are no pending web requests at the moment can initiate garbage collection now, so as to reduce the chances that garbage collection will be needed during the processing of a barrage of web requests later on. (Of course this can hurt if a web request arrives during collection, but the web server could be smart about it and abandon collection if a request comes in.) Desktop GUIs are another example: on the idle event (or, more broadly, after a period of inactivity,) you can give the JVM a hint that if it has any garbage collection to do, now is better than later.
When you want to detect memory leaks. This is often done in combination with a debug-mode-only finalizer, or with the java.lang.ref.Cleaner class from Java 9 onwards. The idea is that by forcing garbage collection now, and thus discovering memory leaks now as opposed to some random point in time in the future, you can detect the memory leaks as soon as possible after they have happened, and therefore be in a better position to tell precisely which piece of code has leaked memory and why. (Incidentally, this is also one of, or perhaps the only, legitimate use cases for finalizers or the Cleaner. The practice of using finalization for recycling of unmanaged resources is flawed, despite being very widespread and even officially recommended, because it is non-deterministic. For more on this topic, read this: https://blog.michael.gr/2021/01/object-lifetime-awareness.html)
When you are measuring the performance of code, (benchmarking,) in order to reduce/minimize the chances of garbage collection occurring during the benchmark, or in order to guarantee that whatever overhead is suffered due to garbage collection during the benchmark is due to garbage generated by the code under benchmark, and not by unrelated code. A good benchmark always starts with an as thorough as possible garbage collection.
When you are measuring the memory consumption of code, in order to determine how much garbage is generated by a piece of code. The idea is to perform a full garbage collection so as to start in a clean state, run the code under measurement, obtain the heap size, then do another full garbage collection, obtain the heap size again, and take the difference. (Incidentally, the ability to temporarily suppress garbage collection while running the code under measurement would be useful here, alas, the JVM does not support that. This is deplorable.)
Note that of the above use cases, only one is in a production scenario; the rest are in testing / diagnostics scenarios.
This means that System.gc() can be quite useful under some circumstances, which in turn means that it being "only a hint" is problematic.
(For as long as the JVM is not offering some deterministic and guaranteed means of controlling garbage collection, the JVM is broken in this respect.)
Here is how you can turn System.gc() into a bit less of a hint:
private static void runGarbageCollection()
{
for( WeakReference<Object> ref = new WeakReference<>( new Object() ); ; )
{
System.gc(); //optional
Runtime.getRuntime().runFinalization(); //optional
if( ref.get() == null )
break;
Thread.yield();
}
}
This still does not guarantee that you will get a full GC, but it gets a lot closer. Specifically, it will give you some amount of garbage collection even if the -XX:DisableExplicitGC VM option has been used. (So, it truly uses System.gc() as a hint; it does not rely on it.)
Yes, calling System.gc() doesn't guarantee that it will run, it's a request to the JVM that may be ignored. From the docs:
Calling the gc method suggests that the Java Virtual Machine expend effort toward recycling unused objects
It's almost always a bad idea to call it because the automatic memory management usually knows better than you when to gc. It will do so when its internal pool of free memory is low, or if the OS requests some memory be handed back.
It might be acceptable to call System.gc() if you know that it helps. By that I mean you've thoroughly tested and measured the behaviour of both scenarios on the deployment platform, and you can show it helps. Be aware though that the gc isn't easily predictable - it may help on one run and hurt on another.
First, there is a difference between spec and reality. The spec says that System.gc() is a hint that GC should run and the VM is free to ignore it. The reality is, the VM will never ignore a call to System.gc().
Calling GC comes with a non-trivial overhead to the call and if you do this at some random point in time it's likely you'll see no reward for your efforts. On the other hand, a naturally triggered collection is very likely to recoup the costs of the call. If you have information that indicates that a GC should be run than you can make the call to System.gc() and you should see benefits. However, it's my experience that this happens only in a few edge cases as it's very unlikely that you'll have enough information to understand if and when System.gc() should be called.
One example listed here, hitting the garbage can in your IDE. If you're off to a meeting why not hit it. The overhead isn't going to affect you and heap might be cleaned up for when you get back. Do this in a production system and frequent calls to collect will bring it to a grinding halt! Even occasional calls such as those made by RMI can be disruptive to performance.
In my experience, using System.gc() is effectively a platform-specific form of optimization (where "platform" is the combination of hardware architecture, OS, JVM version and possible more runtime parameters such as RAM available), because its behaviour, while roughly predictable on a specific platform, can (and will) vary considerably between platforms.
Yes, there are situations where System.gc() will improve (perceived) performance. On example is if delays are tolerable in some parts of your app, but not in others (the game example cited above, where you want GC to happen at the start of a level, not during the level).
However, whether it will help or hurt (or do nothing) is highly dependent on the platform (as defined above).
So I think it is valid as a last-resort platform-specific optimization (i.e. if other performance optimizations are not enough). But you should never call it just because you believe it might help(without specific benchmarks), because chances are it will not.
Since objects are dynamically allocated by using the new operator,
you might be wondering how such objects are destroyed and their
memory released for later reallocation.
In some languages, such as C++, dynamically allocated objects must
be manually released by use of a delete operator.
Java takes a different approach; it handles deallocation for you
automatically.
The technique that accomplishes this is called garbage collection.
It works like this: when no references to an object exist, that object is assumed to be no longer needed, and the memory occupied by the object can be reclaimed. There is no explicit need to destroy objects as in C++.
Garbage collection only occurs sporadically (if at all) during the
execution of your program.
It will not occur simply because one or more objects exist that are
no longer used.
Furthermore, different Java run-time implementations will take
varying approaches to garbage collection, but for the most part, you
should not have to think about it while writing your programs.