[osg-users] Float vs double in BOUNDINGSPHERE and BOUNDINGBOX

Maxim Stere osgforum at tevs.eu
Thu Jun 28 21:00:30 PDT 2018


Is there a reason why OSG_USE_FLOAT_BOUNDINGBOX and OSG_USE_FLOAT_BOUNDINGSPHERE are set to YES in cmake by default?

Since most CPUs now use doubles natively and emulate floats, making float calculations a little slower, would it not be better for OSG speed and performance to have these options be set to OFF by default?

Doubles would take up more RAM but what is more important for intersection speeds in OSG, RAM use or CPU number crunching?

Thank you,

Read this topic online here:

More information about the osg-users mailing list