Hi Matt,
Is this reproducible on different machines?
Have you tried -Xmx3192m? I'm having to push
it this far for some JAS3 SiD work. When I was
running the conversion at JLAB, the jobs would
run for several 10 of thousands of events before
dying on the mentioned error. Using -Xmx2048
resolved that problem but there is clearly something
that slowly consumes memory. This is somewhat normal
and it may not be a leak. Recent code changes may
have slightly increased the memory consumption rate.
How many events do you succeed in processing before
the crash?
Cheers,
Homer
On Fri, 9 Nov 2012, McCormick, Jeremy I. wrote:
> Hi, Matt.
>
> Do you know how to look at memory usage in Java?
>
> You'll might want to debug with something like this:
>
> ----
>
> Runtime runtime = Runtime.getRuntime();
>
> NumberFormat format = NumberFormat.getInstance();
>
> StringBuilder sb = new StringBuilder();
> long maxMemory = runtime.maxMemory();
> long allocatedMemory = runtime.totalMemory();
> long freeMemory = runtime.freeMemory();
>
> sb.append("free memory: " + format.format(freeMemory / 1024) + "<br/>");
> sb.append("allocated memory: " + format.format(allocatedMemory / 1024) + "<br/>");
> sb.append("max memory: " + format.format(maxMemory / 1024) + "<br/>");
> sb.append("total free memory: " + format.format((freeMemory + (maxMemory - allocatedMemory)) / 1024) + "<br/>");
>
> ----
>
> I took the above idea from here.
>
> http://stackoverflow.com/questions/74674/how-to-do-i-check-cpu-and-memory-usage-in-java
>
> You could make that a static method and call it at various points in TestRunEvioToLcio or within one of your drivers in the .lcsim file you're using.
>
> If memory is steadily increasing every event then this indicates some systematic memory leak. If it looks fine and then explodes on one event,
> that indicates some problem with the data of that event. For instance, there could be a bogus (very large) value that gets into an array size setting
> due to corrupted data (e.g. when size of something is read from a data block), which is often what happens to cause these kinds of problems.
>
> --Jeremy
>
> -----Original Message-----
> From: Graham, Mathew Thomas
> Sent: Friday, November 09, 2012 12:38 PM
> To: Uemura, Sho; Omar Moreno; Per Hansson
> Cc: McCormick, Jeremy I.
> Subject: OutOfMemory error in jevio?
>
> I'm trying to run over data (from a pretty pristine hps-java build) like this:
> java -Xmx2048m -cp hps-java/target/hps-java-1.2-SNAPSHOT-bin.jar org.lcsim.hps.evio.TestRunEvioToLcio -x steering/TestRunOfflineRecon.lcsim /nfs/slac/g/hps/mgraham/DarkPhoton/testrun_data/hps_001349.evio.0 -DoutputFile=recon.slcio -d HPS-TestRun-v3
>
>
> and I get this...
>
> Exception in thread "main" java.lang.RuntimeException: java.io.IOException: Map failed
> at org.lcsim.hps.evio.TestRunEvioToLcio.main(TestRunEvioToLcio.java:188)
> Caused by: java.io.IOException: Map failed
> at sun.nio.ch.FileChannelImpl.map(FileChannelImpl.java:748)
> at org.jlab.coda.jevio.EvioReader.mapFile(EvioReader.java:323)
> at org.jlab.coda.jevio.EvioReader.<init>(EvioReader.java:178)
> at org.lcsim.hps.evio.TestRunEvioToLcio.main(TestRunEvioToLcio.java:186)
> Caused by: java.lang.OutOfMemoryError: Map failed
> at sun.nio.ch.FileChannelImpl.map0(Native Method)
> at sun.nio.ch.FileChannelImpl.map(FileChannelImpl.java:745)
> ... 3 more
>
>
> ...
>
> ########################################################################
> Use REPLY-ALL to reply to list
>
> To unsubscribe from the HPS-SOFTWARE list, click the following link:
> https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=HPS-SOFTWARE&A=1
>
########################################################################
Use REPLY-ALL to reply to list
To unsubscribe from the HPS-SOFTWARE list, click the following link:
https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=HPS-SOFTWARE&A=1
