JVM crash with Molconvert 3D option

Hi,





I try to clean the following smiles code and dunno.smi contains:





C\C1=C/CC\C(C)=C\C2C(CC\C(C)=C\CC1)C2(C)C





using molconvert sdf -3:E dunno.smi -o crash.smi





the Windows JVM crashes:


#


# HotSpot Virtual Machine Error, Internal Error


# Please report this error at


# http://java.sun.com/cgi-bin/bugreport.cgi


#


# Java VM: Java HotSpot(TM) Client VM (1.4.2_06-b03 mixed mode)


#


# Error ID: 43113F2652414D452D41503F491418160E435050005C


#


# Problematic Thread: prio=5 tid=0x0003ee18 nid=0x27bc runnable


#





Heap at VM Abort:


Heap


def new generation total 576K, used 77K [0x10010000, 0x100b0000, 0x104f0000)


eden space 512K, 10% used [0x10010000, 0x1001d3b0, 0x10090000)


from space 64K, 38% used [0x100a0000, 0x100a6158, 0x100b0000)


to space 64K, 0% used [0x10090000, 0x10090000, 0x100a0000)


tenured generation total 3288K, used 2600K [0x104f0000, 0x10826000, 0x1401000


0)


the space 3288K, 79% used [0x104f0000, 0x1077a3e0, 0x1077a400, 0x10826000)


compacting perm gen total 4096K, used 3614K [0x14010000, 0x14410000, 0x1801000


0)


the space 4096K, 88% used [0x14010000, 0x143979a0, 0x14397a00, 0x14410000)








I have 1GByte RAM.


See also http://www.google.com/search?num=100&hl=en&lr=&safe=off&q=43113F2652414D452D41503F491418160E435050005C&btnG=Search





Thanks


Tobias

Dear Tobias,


Thanks for the bug report.


Actually, it is a Java 1.4.2_06 Hotspot bug. We have checked the problem and modifed the code of 3D cleaning to avoid this potential JVM crashing.


We are creating a bugfix release which will contain this workaround. I will notify you when it is available.

Thank you!


Tobias

The bugfix release (Marvin 3.5.4) is already available.


http://www.chemaxon.com/marvin/do-download.html

Hi,





no it still crashes with the same error codes. A very bad workaround would be to use the interpreter mode, which is very slow.





-Xint


Operate in interpreted-only mode. Compilation to native code is disabled, and all bytecodes are executed by the interpreter. The performance benefits offered by the Java HotSpot VMs' adaptive compiler will not be present in this mode.





Then MView, molconvert will run, but very slllooowww.





Tobias

Are you sure that you use the new Marvin?


We have not managed to reproduce the JVM crash with the new Marvin (3.5.4).


Please check your Marvin version by the following way:


Open MarvinView or MarvinSketch application and select "About Java VM" menu item from the Help menu.


The popup dialog display information about your software environment.


Please send us the displayed information.

Hi,





no 3d crash in the (Marvin 3.5.4) version, but in the JCHEM package.





Anyway the Help\About shows always version 3.5.3 in both versions (non crash and crash version).





So when will be the JCHEM package updated?





Thank you for bugfixing. :-)


Tobias

We plan to release the new JChem in a few days.

If "Help/About Java VM" shows 3.5.3 as Marvin version, it means that your Marvin uses the old jars.


If it is an application from the 3.5.4 version of Marvin Beans package, you should check the CLASSPATH. The old MarvinBeans.jar may be in the CLASSPATH.


If it is an applet from Marvin Applets 3.5.4 package, you should clear the Java Plug-in's cache to delete old jars from there.


If the "About Java VM" dialog in JChem's Marvin shows 3.5.3 as version number, it's OK, since the current JChem is distributed with Marvin 3.5.3.

I've also noticed the same error when calling the MolHandler.align(Molecule, int[]). I get Java HotSpot error 43113F2652414D452D41503F491418160E435050005C. This is with using jchem 3.0.9, which does include the newer version of Marvin. Unfortunately, the problem is only reproducible when I call the method at certain points in my code, and it causes no problems at other points, and thus I don't have a simple example to share in the forum.





But since you understand the root cause of the problem, would it be possible for you to examine this method and apply the same fix as before?

Hi,





Unfortunately we cannot write a similar workaround this time.





I suggest you to upgrade your Java version.


(this JVM bug only occurs in Java 1.4.2_06)





Best regards,





Szilard

This bug is also present with the latest JVM 1.4.2_07.

Hi,





I have just downloaded 1.4.2_08 from Sun, and it works fine for me:





http://java.sun.com/j2se/1.4.2/download.html





Best regards,





Szilard

Hi,





I have just downloaded Version 1.5.0 (Build 1.5.0_02-b09) from Sun, and it worked fine for me. No JVM crashes anymore.





Tobias