Author Topic: Maximum Number of Kerlnels ans Size of System of Eq.  (Read 5048 times)

carina.nisters

  • New Member
  • *
  • Posts: 1
Maximum Number of Kerlnels ans Size of System of Eq.
« on: October 13, 2014, 02:52:01 AM »
Hi everybody

I'm completely new to ParFEAP and do have some general questions:

- Do you have any experiences on how many Kerlnels you can use ParFEAP?
So is there a limitation to a certain number of parallelizations?

- Can you give me an idea of how large the systems are, you are computing with ParFEAP?

Thanks a lot
Best
Carina

luc

  • Full Member
  • ***
  • Posts: 53
Re: Maximum Number of Kerlnels ans Size of System of Eq.
« Reply #1 on: October 13, 2014, 08:36:25 AM »
I am not too sure what you mean by Kerlnels but if you refer to processors or compute nodes (=multiple processors and dedicated memory on a card), then there is really no limitation. But you should pay attention to the solver that you are using, parfeap requires the use of PETSc or some other parallel solver library that can have their own limitations.

Personally I run simulations with PETSc and there is no hard limitations, that being said, depending on the size of my problem I might want to decide to switch from a direct to an iterative solver and consider sometimes using preconditioniners such as multigrid for example.

Prof. S. Govindjee

  • Administrator
  • FEAP Guru
  • *****
  • Posts: 1165
Re: Maximum Number of Kerlnels ans Size of System of Eq.
« Reply #2 on: October 13, 2014, 10:26:39 AM »
The primary limitation in parFEAP is that the entire problem has to be read into FEAP for partitioning.  This limits the size of problems to those where arrays like node numbers need to be able to fit within arrays with lengths less than the magnitude of a signed 32-bit integer.  The same hold for the number of processors.  Remaining limits are those inherent to PETSc.