Fluids

Fluids

Scheme heap Space error on cluster, -h does not work?

Tagged: ,

    • Yannice
      Subscriber

      Dear Ansys,

      I got an Heap Space error on my simulation on the cluster (see ERROR 1). The solution proposed by you in other Discussions does not work (to add -h500000 or something similar, see ERROR 2).


      My script to get the cluster running looks as following:

      echo "Starting Fluent"

      fluent 2ddp -h500000 -rsh -t $NP -g -cnf=$PBS_NODEFILE -i *.jou>& output.$PBS_JOBID


      Would you know how to solve this error?


      Thank you in advance!

      Bests,

      Yannice

      -------------------------------------

      ERROR 1:

      Error: Out of Scheme heap space.

      Error: Out of Scheme heap space.

      Error: Out of Scheme heap space.

      Error: Out of Scheme heap space.

      Rerun with -h flag set greater than (multiples of 100000) current value of -h100000.

      Error: Out of Scheme heap space.

      Rerun with -h flag set greater than (multiples of 100000) current value of -h100000.

      Rerun with -h flag set greater than (multiples of 100000) current value of -h100000.

      Rerun with -h flag set greater than (multiples of 100000) current value of -h100000.

      Rerun with -h flag set greater than (multiples of 100000) current value of -h100000.

      ---------------------------------------------

      ERROR 2:

      /opt/ud/ansys-2019r3/v195/fluent/fluent19.5.0/bin/fluent -r19.5.0 2ddp -h500000 -rsh -t 5 -g -cnf=/var/opt/ud/torque-4.2.10/aux//2732265.hpc11.hpc -i iiiE_2D_1_v1.jou

      /opt/ud/ansys-2019r3/v195/fluent/fluent19.5.0/cortex/lnamd64/cortex.19.5.0 -f fluent -h500000 -g -i iiiE_2D_1_v1.jou (fluent "2ddp -pmpi-auto-selected -host -r19.5.0 -t5 -mpi=ibmmpi -cnf=/var/opt/ud/torque-4.2.10/aux//2732265.hpc11.hpc -path/opt/ud/ansys-2019r3/v195/fluent")


      Error: Unable to undump. Dump is larger than heap

      ()

      Error encountered in critical code section

    • Rob
      Ansys Employee
      Can you open the file using the GUI? I want to eliminate the cluster and journal commands from the cause.
Viewing 1 reply thread
  • You must be logged in to reply to this topic.