Showing results for 
Search instead for 
Do you mean 
Solved! Go to solution

NX Nastran solver on LAN?



I am not sure, but if I have 3 or 5 NX Nastran workstation. Could I put together on LAN somehow, and reduce analysis time with NX Nastran?


For example, if I have multiple cases, 50 or more cases.




Re: NX Nastran solver on LAN?

Yes you can create your own local area network and use a job queuing software to schedule jobs to different machines.


Depending on the operating system, you can also try to run DMP (distributed memory parallel) analysis. Not all solutions currently support DMP. Also, the local area connection should be high speed such as GigE, Infiniband, etc.


Please do let us know what analysis type you will be performing and what the operating system is.



Re: NX Nastran solver on LAN?

  Hi Easwaran, could you tell us any software for job queuing. The workstations have windows 8.1 and NX9.0.3.4 and NX Nastran 9. Thanks.

Accepted by topic author Igor
‎08-26-2015 04:32 AM

Re: NX Nastran solver on LAN?

There are commercial as well as public domain software. Not all software support windows.


1. LSF from IBM

2. PBS Pro from Altair


are popular commerical job scheduling software. The public domain versions are


1. OpenPBS

2. Torque/Moab/Maui

3. Sun Grid Engine

4. HTCondor from University of Wisconsin


Windows has its own version of job scheduling software in the Microsoft HPC Pack.

Re: NX Nastran solver on LAN?

You can also use VisQ, which ships with FEMAP, or QMan that ships with I-DEAS...

Re: NX Nastran solver on LAN?



As long as you are familiar with the NX Nastran batch command syntax, it should be possible to send such a solving job into a HTCondor grid of NX machines. If you are running the NX Nastran Desktop license version, take care that there exists the restriction that the solving has to be launched on the same machine, where the solver input file was generated by the pre-processing step.

Nevertheless the HTCondor system can be used similar to the Teamcenter TSTK system for running tasks in a background manner on various executing computer nodes. It offers a high flexibility by its ClassAD system to define various complex or even simpler criterias and rules for the matchmaking between computer resources and waiting jobs. Furthermore by its capability to detect user operations resp. absence on the machines, you can link also regular user workstations into such a HTCondor pool and address jobs on those machine when they became idle (user is out of office for lunch).