Skip to content
Snippets Groups Projects
  • Andreas Lauser's avatar
    4b898fa4
    introducing quad precision floating point support · 4b898fa4
    Andreas Lauser authored
    this is an aid for debugging precision problems. in order to use this,
    add --enable-quad to the CONFIGURE_FLAGS of your options file, re-run
    dunecontrol, include the header file "dumux/common/quad.hh" after
    config.h and *before* everything else. next, set the Scalar property
    to 'quad'. finally, you need to patch Dune::className() to not access
    the typeid method, for example by returning a constant string like
    "<unknown>" (so far, typeids are not yet generated for quad precision
    floating point values by GCC which I consider to be a compiler bug...)
    
    git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@6534 2fb0f335-1f38-0410-981e-8018bf24f1b0
    4b898fa4
    History
    introducing quad precision floating point support
    Andreas Lauser authored
    this is an aid for debugging precision problems. in order to use this,
    add --enable-quad to the CONFIGURE_FLAGS of your options file, re-run
    dunecontrol, include the header file "dumux/common/quad.hh" after
    config.h and *before* everything else. next, set the Scalar property
    to 'quad'. finally, you need to patch Dune::className() to not access
    the typeid method, for example by returning a constant string like
    "<unknown>" (so far, typeids are not yet generated for quad precision
    floating point values by GCC which I consider to be a compiler bug...)
    
    git-svn-id: svn://svn.iws.uni-stuttgart.de/DUMUX/dumux/trunk@6534 2fb0f335-1f38-0410-981e-8018bf24f1b0