|February 29, 2016, 09:48||
Solving lots of equations in parallel on the same mesh
Join Date: Oct 2013
Posts: 371Rep Power: 7
I'd like to use parallelization to solve lots of equations on a small mesh. The equations are independent of each other, so I would like to solve them separately on different cores.
This is somewhat possible by copying the complete mesh in the processor subdirectories. However, doing this requires an equal number of iterations on each core because there are some communications happening during the solve function. This also limits the performance because the equations may take a different amount of time to solve and waiting for the other cores introduces unnecessary delays. Basically I would like to have only one communication at the end to combine the results.
Any ideas how to disable the inter process communication during the solving of an equation? I'm wondering if one could disable mpi temporarily and then later on resume it?
I think http://www.cfd-online.com/Forums/ope...tml#post584556 might be a possible solution, but I have yet to try it.
Edit: Tried and works.
Last edited by chriss85; March 2, 2016 at 11:40.
|Thread||Thread Starter||Forum||Replies||Last Post|
|Micro Scale Pore, icoFoam||gooya_kabir||OpenFOAM Running, Solving & CFD||2||November 2, 2013 14:58|
|pimpleFoam: turbulence->correct(); is not executed when using residualControl||hfs||OpenFOAM Running, Solving & CFD||3||October 29, 2013 09:35|
|calculation stops after few time steps||sivakumar||OpenFOAM Running, Solving & CFD||7||March 17, 2013 07:37|
|pisoFoam with k-epsilon turb blows up - Some questions||Heroic||OpenFOAM Running, Solving & CFD||26||December 17, 2012 04:34|
|Orifice Plate with a fully developed flow - Problems with convergence||jonmec||OpenFOAM Running, Solving & CFD||3||July 28, 2011 05:24|