How to increase jvm heap?

LuckyTran · April 8, 2020, 18:06

I have the infamous java.lang.OutOfMemoryError when trying to export a plot containing a lot of monitors as a .csv.

Anybody know how to send the -Xmx option to the jvm when the Star server launches?

LuckyTran · April 8, 2020, 20:10

Solved.

pass -jvmargs -Xmx[#] as a command line argument when you launch Star-CCM.

Also I learned Star abuses LD_PRELOAD and uses -Xmx768m ( max heap space of 768 MB). It ignores the system settings... which for my machine was already configured at 20 GB. =(

wcj1n18 · November 25, 2021, 07:29

Hello,

How would I do this when submitting in batch? I tried simply adding
" -jvmargs -Xmx16G " to my batch script in the final line...

" starccm+ -jvmargs -Xmx16G -np $SLURM_NTASKS -machinefile $NODEFILE -load FILENAME.sim -batch > OUTPUT "

I have a large number of monitors like you and I keep getting the OOM error. The job consist of 80 million cells which should work out to require 200GB of RAM. I have tried up to 7 nodes each of which have usable RAM of 162GB. This works out as as over 1 TB of RAM. Hence, I don't think that this is the issue and think it's heap space.

Any ideas?

Thanks

Will

wcj1n18 · November 26, 2021, 09:38

OK, i've worked out that this was because I was initialising with pressure and velocity data taken from a steady simulation with the same mesh size (80 million cells). The CSV file was 2.5 GB and Star-CCM+ apparently couldn't handle it. I dont know the best way to initalise...

April 8, 2020, 18:06	How to increase jvm heap?	#1
LuckyTran Senior Member Lucky Join Date: Apr 2011 Location: Orlando, FL USA Posts: 5,677 Rep Power: 66	I have the infamous java.lang.OutOfMemoryError when trying to export a plot containing a lot of monitors as a .csv. Anybody know how to send the -Xmx option to the jvm when the Star server launches?

April 8, 2020, 20:10		#2
LuckyTran Senior Member Lucky Join Date: Apr 2011 Location: Orlando, FL USA Posts: 5,677 Rep Power: 66	Solved. pass -jvmargs -Xmx[#] as a command line argument when you launch Star-CCM. Also I learned Star abuses LD_PRELOAD and uses -Xmx768m ( max heap space of 768 MB). It ignores the system settings... which for my machine was already configured at 20 GB. =( bluebase likes this.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Temperature increase causes divergence in a chemical reaction system	nfurkan	FLUENT	0	June 27, 2018 07:22
Increase in flow rate decrease in Outlet pressure.	trovilb1	System Analysis	1	December 5, 2017 07:22
Extremely Small Constant Increase of Heat Transfer Coefficient With Concentration	Shomaz ul Haq	CFX	6	July 9, 2017 08:06
Unexpected species increase in discontinuous phase	CeesH	Fluent Multiphase	0	August 6, 2014 05:41
How can I increase Heat Transfer at Domain Interf?	B.Simon	CFX	3	October 28, 2008 18:53

November 25, 2021, 07:29		#3
wcj1n18 Member Will Crawford-Jones Join Date: Jan 2021 Posts: 43 Rep Power: 5	Hello, How would I do this when submitting in batch? I tried simply adding " -jvmargs -Xmx16G " to my batch script in the final line... " starccm+ -jvmargs -Xmx16G -np $SLURM_NTASKS -machinefile $NODEFILE -load FILENAME.sim -batch > OUTPUT " I have a large number of monitors like you and I keep getting the OOM error. The job consist of 80 million cells which should work out to require 200GB of RAM. I have tried up to 7 nodes each of which have usable RAM of 162GB. This works out as as over 1 TB of RAM. Hence, I don't think that this is the issue and think it's heap space. Any ideas? Thanks Will

November 26, 2021, 09:38		#4
wcj1n18 Member Will Crawford-Jones Join Date: Jan 2021 Posts: 43 Rep Power: 5	OK, i've worked out that this was because I was initialising with pressure and velocity data taken from a steady simulation with the same mesh size (80 million cells). The CSV file was 2.5 GB and Star-CCM+ apparently couldn't handle it. I dont know the best way to initalise...