[AliVali]

I am running an OpenFOAM test using interFoam, teh strange thing is that after applying renumberMesh, my model gets about 30% slower. I am using 288 cpus, checked with OF8 and Of/2006, both show the same defficiency after renumberMesh.
here is teh output of renumberMesh:
"
/*---------------------------------------------------------------------------*\
========= |
\\ / F ield | OpenFOAM: The Open Source CFD Toolbox
\\ / O peration | Website: https://openfoam.org
\\ / A nd | Version: 8
\\/ M anipulation |
\*---------------------------------------------------------------------------*/
Build : 8-c67e2b482b0f
Exec : renumberMesh -overwrite -parallel
Date : May 19 2022
Time : 14:17:19
Host : "gadi-cpu-clx-2906.gadi.nci.org.au"
PID : 2191322
I/O : uncollated
Case : /g/data/rr38/from_short_rr38/rr38/aliv/43803371/parametricModel/Surflakes_v2_layout16_runs/Island_v2_Layout16_St3.0
nProcs : 288
Slaves :
287
(
"gadi-cpu-clx-2906.gadi.nci.org.au.2191323"
"gadi-cpu-clx-2906.gadi.nci.org.au.2191324"
"gadi-cpu-clx-2906.gadi.nci.org.au.2191325"
...
"gadi-cpu-clx-2914.gadi.nci.org.au.733733"
)

Pstream initialized with:
floatTransfer : 0
nProcsSimpleSum : 0
commsType : nonBlocking
polling iterations : 0
fileModificationChecking : Monitoring run-time modified files using timeStampMaster (fileModificationSkew 10)
allowSystemOperations : Allowing user-supplied system call operations

// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * //
Create time

Create mesh for time = 0

// using new solver syntax:
pcorrFinal
{
solver GAMG;
tolerance 1e-07;
relTol 0;
smoother DIC;
nPreSweeps 0;
nPostSweeps 2;
nFinestSweeps 2;
cacheAgglomeration true;
nCellsInCoarsestLevel 10;
agglomerator faceAreaPair;
mergeLevels 1;
}

Mesh size: 38486181
Before renumbering :
band : 134955
profile : 1.80501e+11

Using default renumberMethod.

Selecting renumberMethod CuthillMcKee

Reading geometric fields

Reading volScalarField nut
Reading volScalarField p_rgh
Reading volScalarField k
Reading volScalarField epsilon
Reading volScalarField omega
Reading volScalarField alpha.water
Reading volScalarField nuTilda
Reading volVectorField U
Reading pointVectorField pointDisplacement
Selecting solid-body motion function tabulated6DoFMotion

Renumbering processor cell decomposition map cellProcAddressing
Renumbering processor face decomposition map faceProcAddressing
Renumbering processor point decomposition map pointProcAddressing
After renumbering :
band : 5357
profile : 8.42177e+10



Writing mesh to "constant"

End

Finalising parallel run "


Any help would be highly appreciated.

[sbaffini]

Not an expert in OF or its numbering but, it seems that you actually achieved a band reduction. So, the question is why a band reduction generated worst performance?

I can't be of much help as this requires specific OF knowledge but, out of blue, I would point to something related to GAMG