KNIME unexpectedly closes

KNIME Analytics Platform
1

Dear KNIME,

I am attempting to use the WEKA LinearRegression node or the KNIME Linear Regression (Learner) to perform a regression on a 500 column by 1M row matrix, but every time my phyical memory usage reaches about 6gigs KNIME unexpectedly closes.  My system memory is 24gigs and here is my knime.ini file:

-startup
plugins/org.eclipse.equinox.launcher_1.2.0.v20110502.jar
--launcher.library
plugins/org.eclipse.equinox.launcher.win32.win32.x86_64_1.1.100.v20110502
-vmargs
-Xmx4096m
-Dknime.expert.mode=true
-XX:MaxPermSize=10240m
-server
-Dsun.java2d.d3d=false
-Dosgi.classloader.lock=classname
-XX:+UnlockDiagnosticVMOptions
-XX:+UnsyncloadClass
-Dknime.enable.fastload=true
 
Please let me know how I can get this to work?
 
All the best,
 
Mark
2

Hi Mark,

maybe your MaxPermSize is a little bit too high?
You could try to set

-Xmx10G
-XX:MaxPermSize=512m

Hope this helps,

Christian

3

Hi Christian,

I made the suggested changes to my .ini file and I still experience the same issue.

All the best,

Mark

4

Hi Mark,

Problem still when your memory reaches 6GB? One thing you could try is to even increase your MaxMemory (-Xmx10G) to 20G if you have 24 available.

Cheers,

Christian

 

5

Update,

I found this Java error that is linked to event I am experiencing:

# A fatal error has been detected by the Java Runtime Environment:
#
#  EXCEPTION_ACCESS_VIOLATION (0xc0000005) at pc=0x00000000576853b0, pid=924, tid=2388
#
# JRE version: 7.0_07-b11
# Java VM: Java HotSpot(TM) 64-Bit Server VM (23.3-b01 mixed mode windows-amd64 compressed oops)
# Problematic frame:
# V  [jvm.dll+0x253b0]
#

# Failed to write core dump. Minidumps are not enabled by default on client versions of Windows

Any idea how to address this issue?

All the best,

Mark