openfoam/applications/test/limits/Test-limits.C
2019-12-05 11:47:19 +00:00

102 lines
3.9 KiB
C

/*---------------------------------------------------------------------------*\
========= |
\\ / F ield | OpenFOAM: The Open Source CFD Toolbox
\\ / O peration |
\\ / A nd | www.openfoam.com
\\/ M anipulation |
-------------------------------------------------------------------------------
Copyright (C) 2018 OpenCFD Ltd.
-------------------------------------------------------------------------------
License
This file is part of OpenFOAM.
OpenFOAM is free software: you can redistribute it and/or modify it
under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
OpenFOAM is distributed in the hope that it will be useful, but WITHOUT
ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
for more details.
You should have received a copy of the GNU General Public License
along with OpenFOAM. If not, see <http://www.gnu.org/licenses/>.
Description
Print max limits.
\*---------------------------------------------------------------------------*/
#include <limits>
#include "int.H"
#include "uint.H"
#include "string.H"
#include "scalar.H"
#include "IOstreams.H"
using namespace Foam;
// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * //
// Main program:
int main(int argc, char *argv[])
{
//NONE Info<<"int16:" << pTraits<int16_t>::max << nl;
Info<< "=max=" << nl;
Info<< "uint8:" << std::numeric_limits<uint8_t>::max() << nl;
Info<< "int16:" << std::numeric_limits<int16_t>::max() << nl;
Info<< "int32:" << pTraits<int32_t>::max << nl;
Info<< "uint32:" << pTraits<uint32_t>::max << nl;
Info<< "int64:" << pTraits<int64_t>::max << nl;
Info<< "uint64:" << pTraits<uint64_t>::max << nl;
Info<< nl;
cout<< "int16:" << std::numeric_limits<int16_t>::max() << nl;
cout<< "int32:" << pTraits<int32_t>::max << nl;
cout<< "uint32:" << pTraits<uint32_t>::max << nl;
cout<< "int64:" << pTraits<int64_t>::max << nl;
cout<< "uint64:" << pTraits<uint64_t>::max << nl;
Info<< nl << "=digits=" << nl;
cout<< "int16:" << std::numeric_limits<int16_t>::digits << nl;
cout<< "int32:" << std::numeric_limits<int32_t>::digits << nl;
cout<< "uint32:" << std::numeric_limits<uint32_t>::digits << nl;
cout<< "int64:" << std::numeric_limits<int64_t>::digits << nl;
cout<< "uint64:" << std::numeric_limits<uint64_t>::digits << nl;
cout<< "float:" << std::numeric_limits<float>::digits << nl;
cout<< "double:" << std::numeric_limits<double>::digits << nl;
cout<< "long double:" << std::numeric_limits<long double>::digits << nl;
Info<< nl << "=float=" << nl;
cout<< "max:" << std::numeric_limits<float>::max()
<< " VGREAT:" << floatScalarVGREAT << nl;
cout<< "min:" << std::numeric_limits<float>::min()
<< " VSMALL:" << floatScalarVSMALL << nl;
cout<< "epsilon:" << std::numeric_limits<float>::epsilon()
<< " SMALL:" << floatScalarSMALL << nl;
cout<< "1/epsilon:" << 1.0f/std::numeric_limits<float>::epsilon()
<< " GREAT:" << floatScalarGREAT << nl;
Info<< nl << "=double=" << nl;
cout<< "max:" << std::numeric_limits<double>::max()
<< " VGREAT:" << doubleScalarVGREAT << nl;
cout<< "min:" << std::numeric_limits<double>::min()
<< " VSMALL:" << doubleScalarVSMALL << nl;
cout<< "epsilon:" << std::numeric_limits<double>::epsilon()
<< " SMALL:" << doubleScalarSMALL << nl;
cout<< "1/epsilon:" << 1.0f/std::numeric_limits<double>::epsilon()
<< " GREAT:" << doubleScalarGREAT << nl;
Info << "---\nEnd\n" << endl;
return 0;
}
// ************************************************************************* //