coded FO, cellzone and new volScalarField: problem with mpirun

elden · March 13, 2020, 08:49

Hi,
I'm trying to run a few tests before converting the following coded FO in an actual OpenFoam FO.

This FO initialises a zero volScalarField in all the grid, then it changes its values ONLY in a previously generated cellZone (by means of topoSet). Each cell value of the volScalarField is now 0 or 1. The output consists of cells number where 0 becomes 1, printed in the logfile.

I noticed differences in the output running the case with mpirun. Looking at the results in paraview it seems that there is an error in forAll cycle: the coded FO changes the values of the new volScalarField outside the defined cellzone.

I thought the problem was related to the definition of the scalar "sum" and that the answer I was looking for was associated with the use of the reduce() command, but I didn't solve it.

olesen · March 13, 2020, 11:33

An indexing error in your code.
You want something like this instead:

Code:

forAll(cz, i)
{
    const label celli = cz[i];
    ....
}

Unless you need the loop index, it is often safer and clearer to use a range-for:

Code:

for (const label celli : cz)
{
    ....
}

/mark

@Bruno - relocate to programming?

elden · March 13, 2020, 11:47

Thanks a lot.
And sorry for the silly question.

March 13, 2020, 11:33		#2
olesen Senior Member Mark Olesen Join Date: Mar 2009 Location: https://olesenm.github.io/ Posts: 1,686 Rep Power: 40	An indexing error in your code. You want something like this instead: Code: forAll(cz, i) { const label celli = cz[i]; .... } Unless you need the loop index, it is often safer and clearer to use a range-for: Code: for (const label celli : cz) { .... } /mark @Bruno - relocate to programming? elden likes this.

March 13, 2020, 11:47		#3
elden New Member Pietro Tavazzi Join Date: Aug 2019 Location: Milan, Italy Posts: 8 Rep Power: 6	Thanks a lot. And sorry for the silly question.