Andy Ross wrote:
> 
> > 2) Would this increase execution time?
> 
> Probably only very slightly.  Unless you play fun games with the
> assembler, all the calculations are done at full precisions per the C
> language specification.  The only problem that should crop up is space

I believe this is no longer true: in C99 the compiler is allowed to do an operation on 
floats at 'float' precision, and in C++ it is expected to.  In C I presume function 
arguments would still get passed as 'double' even when declared as 'float', but in C++ 
you (and the standard library) may be able to define two versions of functions: both 
'float sin(float)' and 'double sin(double)' and the compiler would not promote the 
'float' argument to 'double'.

Note that I have not checked my facts; I am just writing from memory.  Please correct 
me.

This doesn't change your answer: it still wouldn't make much difference in practice.

- Julian

_______________________________________________
Flightgear-devel mailing list
[EMAIL PROTECTED]
http://mail.flightgear.org/mailman/listinfo/flightgear-devel

Reply via email to