[dominik_christ]

Hello everyone,
the following bug is present in both OF1.5 and OF1.6:

You would expect the following functions to return the same value

Info << pow(turbulence->epsilon(), (1./2.) );
Info << pow(turbulence->epsilon(), (1/2) );

Both compile without error or warning. However, the second line only returns the value 1 in all fields. The first line returns the square root of the epsilon field just as expected.

NB: This bug does not occur with plain integers, so
Info << pow(turbulence->epsilon(), 2. );
Info << pow(turbulence->epsilon(), 2 );
will give you the same (correct) result.

Regards,
Dominik Christ

P.S.: As this is such a fundamental function, I wonder why it passed unnoticed for so long ...

[mattijs]

The compiler will evaluate 1/2 as integer division. Hence it will be doing a pow(.., 0) which returns 1 as it should. Apart from this you could also use sqrt or sqr.

[dominik_christ]

Oh, of course, that's the reason.
Thanks Mattijs for pointing that out to me!

[wadekar]

Hello

I am also struggling with same problem, but i want to do the power (1/3).

scalar rad = Foam::pow(Vol,1.0/3.0);

Vol is a scalar.

My main problem is that- when i using

scalar rad = Foam::pow(Vol,1.0/3.0);

then it gives right result without any problem, but its not running in parallel.

And, when i used

 scalar rad = Foam::pow(Vol,1/3);

then i am getting 1, but code is running in parallel.