HyperStudy Tips and Tricks: #31 - Enabling Flux Distributed Computing In HyperStudy
Altair EmployeeHello HyperStudy and Flux Users,
You know that Multi-Execution in HyperStudy allows running concurrent Flux simulations to speed up your DOE or Optimization studies.
But you may not know that in some case you can combine Multi-Execution in HyperStudy with Flux Distributed computing to further speed up your evaluations !?
If you deal with Flux 3D Magnetostatic applications, this tip can be of interest for you !
Let’s consider an example of electromagnetic actuator designed in Flux3D Magnetostatic. For a given geometrical design, the performance is analyzed by solving a parametric scenario with several Force (Position) steps. Within Flux, the Parametric Distributed computing can be used to speed up the runtime.
The parametric distributed computing can also be enabled when running Flux in batch from HyperStudy. To achieve this, you would need to add the following solver arguments in your HyperStudy-Flux setup:
Note: MAXJOBS value depends on the number of steps to solve and/or on the hardware resources available.
As an example, if Multi-Execution=2 and MAXJOBS=4, there will two concurrent Flux with 4 concurrent jobs each one.
To learn more on Flux Distributed computing :
https://2023.help.altair.com/2023/flux/Flux/Help/english/HowTo/English/HowTo_LaunchFluxViaCommandLine/topics/launchfluxviacommandline_r.htm
Does this help?
Let us know your feedback!