When processing large data sets, KNIME seems to be running out of memory. What can I do?

By default KNIME uses a maximum of 1024 MB of heap space for the Java Virtual Machine (JVM). When processing very large data sets this amount might not be sufficient, and you may see messages similar to the following:

java.lang.OutOfMemoryError: Java heap space
FATAL node-name Fatal error
FATAL node-name Java heap space

You can change the settings by editing the file:

$SCHRODINGER/knime-vversion/bin/arch/knime/knime.ini

where arch is Linux-x86, Linux-x86_64, WIN32-x86, or Windows-x64. The file should contain the following settings:

-clean
-vmargs
-Xmx1024m
-XX:MaxPermSize=128m

In order to increase the amount of heap memory you should change the -Xmx setting. For example, to allocate 2048 MB of heap space to the JVM you should change -Xmx1024m to -Xmx2048m.

Keywords: KNIME, crash, error

Back to Search Results