Star-CCM+/en: Difference between revisions

From Alliance Doc
Jump to navigation Jump to search
(Updating to match new version of source page)
 
(Updating to match new version of source page)
Tags: Mobile edit Mobile web edit
Line 2: Line 2:


[[Category:Software]]
[[Category:Software]]
STAR-CCM+ is an all-in-one solution that delivers accurate and efficient multidisciplinary technologies in a single integrated user interface. It is developed by [https://mdx.plm.automation.siemens.com/star-ccm-plus Siemens]
[https://mdx.plm.automation.siemens.com/star-ccm-plus STAR-CCM+] is an all-in-one solution that delivers accurate and efficient multidisciplinary technologies in a single integrated user interface; it is developed by Siemens.


= License limitations =
= License limitations =
Compute Canada has the authorization to host STAR-CCM+ binaries on its servers, but does not provide licenses to users. You will need to have your own license in order to use this software.
Compute Canada has the authorization to host STAR-CCM+ binaries on its servers, but does not provide licenses to users. You will need to have your own license in order to use this software.


= Running Star-CCM+ on Compute Canada servers =  
= Running Star-CCM+ on Compute Canada servers =
We have two flavours of Star-CCM+ installed on our servers: a double precision one, and a mixed precision one. You will first need to figure out which version is best suited for your computations. The <tt>starccm</tt> module is the double-precision flavour, while the <tt>starccm-mixed</tt> module is the mixed precision flavour. Star-CCM+ comes bundled with two different distributions of MPI: [https://www.ibm.com/developerworks/downloads/im/mpi/index.html IBM Platform MPI] and [https://software.intel.com/en-us/intel-mpi-library Intel MPI]. The default distribution is IBM Platform MPI. However, this distribution does not work on [[Cedar]]'s Intel Omnipath network fabric. One must therefore tell <tt>starccm+</tt> to use <tt>intel</tt> as the MPI distribution. This is done with the options <tt>-mpi intel</tt>.  
Select one of the available modules:
* <tt>starccm</tt> for the double-precision flavour,  
* <tt>starccm-mixed</tt> for the mixed precision flavour.


Moreover, since neither IBM Platform MPI nor Intel MPI are tightly coupled with our scheduler, you must tell <tt>starccm+</tt> what hosts to use by means of a file containing the list of hosts. To produce that file, we provide a script called <tt>slurm_hl2hl.py</tt> which, when called with the option <tt>--format STAR-CCM+</tt>, will output the list of hosts. This list can then be written to a file and read by Star-CCM+. Also, because these distributions of MPI are not tightly integrated with our scheduler, you should use the options <tt>--ntasks-per-node=1</tt> and <tt>--cpus-per-task=32</tt> for your job submission.  
Star-CCM+ comes bundled with two different distributions of MPI:
*[https://www.ibm.com/developerworks/downloads/im/mpi/index.html IBM Platform MPI] is the default distribution, but does not work on [[Cedar]]'s Intel OmniPath network fabric;
*[https://software.intel.com/en-us/intel-mpi-library Intel MPI] is specified with option <tt>-mpi intel</tt>.  


Finally, you will need to setup your job environment to use your license. If you are using Adapco's online "pay-on-usage" server, it is rather simple to configure. If you are using an internal license server, please [mailto:support@computecanada.ca contact us] so that we can help you setup the access to it.  
Neither IBM Platform MPI nor Intel MPI are tightly coupled with our scheduler; you must therefore tell <tt>starccm+</tt> which host to use by means of a file containing the list of available hosts. To produce this file, we provide the <tt>slurm_hl2hl.py</tt> script, which will output the list of hosts when called with the option <tt>--format STAR-CCM+</tt>. This list can then be written to a file and read by Star-CCM+. Also, because these distributions of MPI are not tightly integrated with our scheduler, you should use options <tt>--ntasks-per-node=1</tt> and <tt>--cpus-per-task=32</tt> when submitting a job.
When all is done, your submit script should look like this (this example uses 2 nodes for 1 hour, adjust these numbers based on the needs of your job):
 
You will also need to set up your job environment to use your license. If you are using Adapco's online "pay-on-usage" server, the configuration is rather simple. If you are using an internal license server, please [mailto:support@computecanada.ca contact us] so that we can help you setup the access to it.  
When all is done, your submit script should look like this, where 2 nodes are used for 1 hour; you can adjust these numbers to fit your needs.
{{File
{{File
|name=mysub.sh
|name=mysub.sh

Revision as of 23:58, 15 December 2017

Other languages:

STAR-CCM+ is an all-in-one solution that delivers accurate and efficient multidisciplinary technologies in a single integrated user interface; it is developed by Siemens.

License limitations

Compute Canada has the authorization to host STAR-CCM+ binaries on its servers, but does not provide licenses to users. You will need to have your own license in order to use this software.

Running Star-CCM+ on Compute Canada servers

Select one of the available modules:

  • starccm for the double-precision flavour,
  • starccm-mixed for the mixed precision flavour.

Star-CCM+ comes bundled with two different distributions of MPI:

  • IBM Platform MPI is the default distribution, but does not work on Cedar's Intel OmniPath network fabric;
  • Intel MPI is specified with option -mpi intel.

Neither IBM Platform MPI nor Intel MPI are tightly coupled with our scheduler; you must therefore tell starccm+ which host to use by means of a file containing the list of available hosts. To produce this file, we provide the slurm_hl2hl.py script, which will output the list of hosts when called with the option --format STAR-CCM+. This list can then be written to a file and read by Star-CCM+. Also, because these distributions of MPI are not tightly integrated with our scheduler, you should use options --ntasks-per-node=1 and --cpus-per-task=32 when submitting a job.

You will also need to set up your job environment to use your license. If you are using Adapco's online "pay-on-usage" server, the configuration is rather simple. If you are using an internal license server, please contact us so that we can help you setup the access to it. When all is done, your submit script should look like this, where 2 nodes are used for 1 hour; you can adjust these numbers to fit your needs.

File : mysub.sh

#!/bin/bash
#SBATCH --time=01:00:00
#SBATCH --nodes=2
#SBATCH --ntasks-per-node=1
#SBATCH --cpus-per-task=32
module load starccm/12.04.011-R8
export LM_PROJECT='YOUR ADAPCO PROJECT ID GOES HERE'
export CDLMD_LICENSE_FILE="1999@flex.cd-adapco.com"

slurm_hl2hl.py --format STAR-CCM+ > machinefile

NCORE=$((SLURM_NTASKS * SLURM_CPUS_PER_TASK))

starccm+ -power -np $NCORE -machinefile machinefile -batch -mpi intel /path/to/your/simulation/file