Most optimal use of resources - parallel computing

Szymon Adamus

Hello!

I have a question regarding performance of solving in Altair Flux software.

In our company we have got highly efficient PC with 12 cores and 128GB of RAM memory. We would like to improve the time of solving.

Our main applications are "Magneto Static 3D". Is there any optmal way to alocate resources? For know i found out that I can change number of cores used in parallel computing, as well as amount of memory alocated to the application.

 

Thank You in advance!

Mohammed Elamin_22169

Hi Szymon, 

You can check the attached document. Beside what you did if you are running parametric solving, it make sense to use CDE and distribute the computation among the different cores. 

Regarding the memory, there is also the dynamic memory option. which means Flux will manage the memory needed during solving without you allocating specific amount. 

Thanks,  

Szymon Adamus

Hi Mohammed,

 

Thank You for quick response.

Unfortunately we use Linux based system. Is it any way to distribute computation on different cores?

 

Thanks

Best Regards

Mohammed Elamin_22169

Szymon Adamus said:

Hi Mohammed,

 

Thank You for quick response.

Unfortunately we use Linux based system. Is it any way to distribute computation on different cores?

 

Thanks

Best Regards

Not with the current version. But the good news is the upcoming version of Flux 2022.1 (expected during August) should allow the computation distribution on Linux. 

 

Otherwise, you can only distribute the computation on Linux through PBS or Altair access. 

Szymon Adamus

OK, so I awaite for new version of Flux.

Thank You for your time.