|
[Sponsors] |
July 12, 2018, 04:11 |
Load balancing question
|
#1 |
New Member
Join Date: Nov 2017
Posts: 5
Rep Power: 8 |
Should the blocks per core also be similar for faster run time in addition to cells per core among all cores?
I have a case where the cells/core are 1e6 and well distributed across all cores (max 1e6 min 80k per core) however the blocks/core aren't (max 2k min 1 per core)! Kindly advise. |
|
July 13, 2018, 01:47 |
|
#2 |
Senior Member
Tobias
Join Date: May 2016
Location: Germany
Posts: 268
Rep Power: 11 |
I would say the blocks/core doesnt matter.
However this will change anyway with CONVERGE 3.0 |
|
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Run-Time Parallel Load Balancing | tgvosk | OpenFOAM Programming & Development | 22 | January 8, 2024 05:53 |
[OpenFOAM] How to load all <case>/surfaces/*/*.vtk at once into Paraview | holodeck10 | ParaView | 8 | November 22, 2019 17:52 |
Dymanic load balancing in dynamicRefineFvMesh | Aleksey_R | OpenFOAM Programming & Development | 3 | January 29, 2016 17:29 |
Surface Load Line Can not Transfer | Mohamed.gaber | ANSYS | 0 | June 30, 2015 03:42 |
mpi on vmware-no load balancing between cores | Helium-3 Man | OpenFOAM Running, Solving & CFD | 1 | February 26, 2015 18:46 |