Johannes Brunen wrote: > Hello Gerrit, > > "Gerrit Voß" <vo...@vossg.org> schrieb im > Newsbeitrag news:1246336792.3319.14.ca...@feynman.camtech.ntu.edu.sg... >> short question. what are the correct windows debug define settings or >> if this is not well defined, which are the once you expect to be set ? >> >> Debug -> -D _DEBUG >> DebugOpt -> ? never used it, does it need -D _DEBUG or -D NDEBUG ? >> ReleaseNoOpt -> ? So far I did not use -D NDEBUG or -D _DEBUG and that >> seem to work >> Releae -> -D NDEBUG >> > imho the following settings would be the way to go: > > Debug -> /D_DEBUG /Od /Ob0 /RTC1 /MDd /Zi > /PDB:"c:\path\to\the\dll\dir\dllname.pdb" > DebugOpt -> /DNDEBUG /O2 /Ob2 /MDd /Zi > /PDB:"c:\path\to\the\dll\dir\dllname.pdb" > Release -> /DNDEBUG /O2 /Ob2 /MD /Zi > /PDB:"c:\path\to\the\dll\dir\dllname.pdb" > ReleaseNoOpt -> /D_DEBUG /Od /Ob0 /RTC1 /MD /Zi > /PDB:"c:\path\to\the\dll\dir\dllname.pdb" > > Marcus, could you please verify that. I'm unsure with respect to the > ReleaseNoOpt configuration, because this one is new for me. Additionally, I > did never generate debug database files on release builds. Actuallly, I > didn't know that this is possible until recently. At last I'm not completely > confident about the inline optimization /Ob2 against its counterpart /Ob1. > Which one is recommendable on the microsoft compiler platform?
Sorry for the late answer! Relase NoOpt looks good (/Od and /MD). I always let the compiler choose any suitable function for inlining. In general I tend to defer to the compiler when it comes to choosing which code is fastest. :) (Btw, I suspect much OpenSG.inl code is never inlined, as the function body vastly exceeds the size of the function call, but I haven't confirmed this.) /Marcus ------------------------------------------------------------------------------ _______________________________________________ Opensg-users mailing list Opensg-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/opensg-users
