[Lilly]

Dear all,
I want to run my simulations as a batch job and want to use 8 cores for this, but when I set up the case file I cannot use 8 cores. What is the best way to proceed in this situation:
open and set up the case in serial;auto-partition or manual-partition of the grid? do I have to set the number of parallel cores at "parallel" at the GUI or is it ok, if fluent does this automatically when the batch job runs?

Another thing is: is it ok if I use another number of cores for more expensive simulations or can I not compare the simulations calculated with a different number of cores then (will the results be dependent on the number of cores?)?

If would be really nice, if one could give me a hint!

Thanks a million in advance!
Lilly

[shk12345]

Originally Posted by Lilly

Dear all,
I want to run my simulations as a batch job and want to use 8 cores for this, but when I set up the case file I cannot use 8 cores. What is the best way to proceed in this situation:
open and set up the case in serial;auto-partition or manual-partition of the grid? do I have to set the number of parallel cores at "parallel" at the GUI or is it ok, if fluent does this automatically when the batch job runs?

Another thing is: is it ok if I use another number of cores for more expensive simulations or can I not compare the simulations calculated with a different number of cores then (will the results be dependent on the number of cores?)?

If would be really nice, if one could give me a hint!

Thanks a million in advance!
Lilly

Just go to usage >> time .
i guess the second last tab in fluent.
Then calculate the time of the simulation with 8 (cpu time) with time of
simulation when (clock time) and see if it is comming to 8 times .
If it is so then your fluent is working fine otherwise there can be some problem with the hardware.

Regards
shk

[gfoam]

Hello:
First of all, why can't you use parallel processing to do the set up of the case? Are you using Workbench? If so, then when you create a new standalone system for FLUENT and then edit the setup cell, there you can select the number of cores to do the setup, or RMBC==>Properties and select Parallel processing and the number of cores. In this case you will do an automatic partition. Regarding of manual partition, I never do this so i can't help you in this case.
When you finish to do the setup, you can save the case and close FLUENT, then back in Workbench you can Update your solution cell through RMBC ==> Update, and you can change the numbers of cores in the Properties panel of the solution cell.
And the last question: yes you can compare the solution and the time to reach it with different numbers of cores. The solution doesn't depend on the number of cores used to calculate it. But the maximun numbers of cores you can use depends on the license you have. Hope this helps you.
Gonzalo