I'm currently working on a project where I need to call a rather large jar file from Python code. For this I'm using jpype and with a smaller stub jar the code worked perfectly. However, the larger jar requires a lot of memory, so I set -Xmx in the arguments that are passed to the startJVM call. However, this crashes with a could not reserve enough space for object heap error in Python.
The value I'm using for -Xmx is 1099m, if I decrement this value by one and use 1098m then enough memory can be reserved and the JVM starts. However, to successfully use the Jar I need more than 1098.
When calling startJVM enough memory is available on my machine, usually at least 3GB. Furthermore, if I run the jar directly from the command line using java -Xmx2000m -jar then it starts without any problems. Which leads me to believe that I might be doing something wrong.
I'm running the code on a 64 bit Windows installation with Python 2.7 32 bit and a 32 bit Java 8 JDK. The following line is the line I use to start the JVM from Python:
startJVM("PATH", "-Xmx1099m", "-Djava.class.path=PATH_TO_JAR")
Does anyone know a possible solution to start the JVM with jpype with more than 1089 MB of heap space?
