[PicklER]

Hello

I declared two volScalarfields in createFields.H, namely:

volScalarField r
(
    IOobject
    (
        "r",
        runTime.timeName(),
        mesh,
        IOobject::NO_READ,
        IOobject::NO_WRITE
    ),
    mesh,
    dimensionedScalar("0", dimLength, 1.0)
)

and

volScalarField faceCellDist
    (
        IOobject
        (
            "faceCellDist", 
            runTime.timeName(), 
            mesh, 
            IOobject::NO_READ, 
            IOobject::NO_WRITE
        ), 
        mesh, 
        dimensionedScalar
        ("0", dimLength, 1.0)
    );

Then I determine distance from cell centre to cell face and store it in "faceCellDist". This is to get delta x, y and z by multiply "faceCellDist" by 2. And assign a constant (initial value) to "r":

const surfaceVectorField& Cf  = mesh.Cf();
const volVectorField& C = mesh.C();
const labelList& own = mesh.owner();
forAll (own, I)
    {
        faceCellDist[I] = Foam::mag(C[own[I]] - Cf[I]);
        faceCellDist[I+1]=faceCellDist[I];
    }

The second line "faceCellDist[I+1]=faceCellDist[I];" is to set the last element, otherwise it stays 0 (zero).

and the initial value for "r":

r=r*id.value()/2;

where "id.value()" is just a constant from test case.

My printStack (multiply) error comes when I do:

r*faceCellDist

Any suggestions?
Vrede

[marupio]

printStack is not the error. What is the error? Was it a segmentation fault? (SigSEGV) or a floating point error? (SigFPE) The forAll(own, I) loop will cause a segmentation fault when it reaches the final index (i = faceCellDist.size()-1)... because I+1 will be out of range.

[PicklER]

Good morning David.

Thank you for the quick reply. The error is as follows:

#0 Foam::error::printStack(Foam::Ostream&) at ??:?
#1 Foam::sigSegv::sigHandler(int) at ??:?
#2 in "/lib/x86_64-linux-gnu/libc.so.6"
#3 Foam::multiply(Foam::Field<double>&, Foam::UList<double> const&, Foam::UList<double> const&) at ??:?
#4 void Foam::multiply<Foam::fvPatchField>(Foam::FieldFiel d<Foam::fvPatchField, double>&, Foam::FieldField<Foam::fvPatchField, double> const&, Foam::FieldField<Foam::fvPatchField, double> const&) at ??:?
#5 Foam::tmp<Foam::GeometricField<double, Foam::fvPatchField, Foam::volMesh> > Foam::operator*<Foam::fvPatchField, Foam::volMesh>(Foam::GeometricField<double, Foam::fvPatchField, Foam::volMesh> const&, Foam::GeometricField<double, Foam::fvPatchField, Foam::volMesh> const&) at ??:?
#6
at ??:?
#7 __libc_start_main in "/lib/x86_64-linux-gnu/libc.so.6"
#8
at ??:?
Segmentation fault (core dumped)

Weirdly enough, the forAll loop does not give me an error, since it run successfully through the loop.

Note: When I did this for a "1D" test case (shock tube tutorial of rhoCentralFoam), I did not have this error. Only when I went to "2D" (30 cells in x-dir, 3 in y-dir and 1 cell in z-dir) that this error showed.

Kind regards

[marupio]

Ah, I see. It's a segmentation fault. Opt doesn't always throw an error exactly where the bug is. You could try running in debug mode. And you should fix your forall loop.

[sabago]

Dear all,

what if it's a floating point error?

such as

#0  Foam::error::printStack(Foam::Ostream&) at ??:?
#1  Foam::sigFpe::sigHandler(int) at ??:?
#2   in "/lib/x86_64-linux-gnu/libc.so.6"
#3  Foam::multiply(Foam::Field<double>&, Foam::UList<double> const&, Foam::UList<double> const&) at ??:?
#4  
 at ??:?
#5  Foam::tmp<Foam::GeometricField<double, Foam::fvPatchField, Foam::volMesh> > Foam::operator*<Foam::fvPatchField, Foam::volMesh>(Foam::GeometricField<double, Foam::fvPatchField, Foam::volMesh> const&, Foam::tmp<Foam::GeometricField<double, Foam::fvPatchField, Foam::volMesh> > const&) at ??:?
#6  
 at ??:?
#7  __libc_start_main in "/lib/x86_64-linux-gnu/libc.so.6"
#8  
 at ??:?
Floating point exception (core dumped)

Best,
Sandra

[wyldckat]

Quick answer:

Originally Posted by sabago

what if it's a floating point error?
such as

Correct, it's a signal for Floating Point Error: "sigFpe"

Originally Posted by sabago

#1  Foam::sigFpe::sigHandler(int) at ??:?

Occurred while multiplying two lists of double values:

Originally Posted by sabago

#3  Foam::multiply(Foam::Field<double>&, Foam::UList<double> const&, Foam::UList<double> const&) at ??:?

The two lists are actually two volume fields, identified with the type "GeometricField< ... volMesh>":

Originally Posted by sabago

#5  Foam::tmp<Foam::GeometricField<double, Foam::fvPatchField, Foam::volMesh> > Foam::operator*<Foam::fvPatchField, Foam::volMesh>(Foam::GeometricField<double, Foam::fvPatchField, Foam::volMesh> const&, Foam::tmp<Foam::GeometricField<double, Foam::fvPatchField, Foam::volMesh> > const&) at ??:?

In order for this to occur, either one of the values was Not-a-Number (NaN) or Infinite (Inf).

If you have access to the source code for the equation where this product occurs, you can try using old school debugging, as... I've already explained here: http://www.cfd-online.com/Forums/ope...tml#post577363 post #3

[sabago]

Forgot to update that I fixed this.

Issue was that the fields being multiplied were of different sizes so I fixed that.

Best,
Sandra