[Bug 520] Crash with big mutli dimesnional arrays works in Sun(Oracle)JRE

[bugzilla-daemon at icedtea.classpath.org]

http://icedtea.classpath.org/bugzilla/show_bug.cgi?id=520


jon.vanalten at redhat.com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID




------- Comment #1 from jon.vanalten at redhat.com  2010-07-12 18:28 -------
Hi,

I find that on my system I get what I believe is the same thing, running your
jar:

$ java -jar rucksack.jar 
Segmentation fault (core dumped)

But if the jar is extracted and I recompile the source, I get:

$ java main.Rucksack
Exception in thread "main" java.lang.OutOfMemoryError: Java heap space
        at main.Rucksack.main(Rucksack.java:110)

This is using:

$ java -version
java version "1.6.0_18"
OpenJDK Runtime Environment (IcedTea6 1.8) (fedora-39.b18.fc12-x86_64)
OpenJDK 64-Bit Server VM (build 14.0-b16, mixed mode)

I also have a b20 closed JDK:

$ /opt/jdk1.6.0_20/jre/bin/java -version
java version "1.6.0_20"
Java(TM) SE Runtime Environment (build 1.6.0_20-b02)
Java HotSpot(TM) 64-Bit Server VM (build 16.3-b01, mixed mode)


Now, in my case I can run with a lot more than 20 elements with icedtea, but
there is a point when I run out of heap.  Also, I run out of heap with closed
java at some higher number of elements than with icedtea.  FWIW I found 1000
elements to be too large for either jdk.

So I guess there is some difference between how default max heap size is
determined with closed vs open jdk.  I am not expert in such matters, but I had
the impression that JVM uses some fraction of detected system RAM as max heap
size (in openjdk/icedtea at least; I have no idea how closed java decides this
default).  One can specify "-XX:MaxHeapSize=<SOME MEMORY SIZE>" as java
argument, I was able to run icedtea java over 1000 elements by setting some
high value for this.  You might try the same as workaround.

Another possibility is difference in garbage collection strategy; perhaps
closed java is more aggressive about this so it is less likely for heap space
to run out.  I did not examine your program closely to determine if it leaves
many objects available for GC.

Either way it looks to me like the problem is a matter of the massive amounts
of memory required, which goes beyond the maximum heap size available at
runtime, which as noted above can be specified.


-- 
Configure bugmail: http://icedtea.classpath.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.