fop performance with big big files

[fs...@clearstream.com]

Hi,

I am using fop for 3 months now to generate many kinds of PDF in a Java
Application. It works great.
My only issue is when the XML file is huge then the memory usage is too high
and the system crashes, also it takes several minutes to generate the PDF
file. My XML files are several hundred of K.

Is there some trivial ways of optimizing FOP (say using a SAX parser, I am
already using the Xerces parser)?
I am also thinking in cutting the XML into chunk and generate several output
PDF files (if no I do not find any other solution).

Thanks for your ideas.

Frederic.