Bill Studenmund <wrstuden@nas.nasa.gov> writes:
> > Ignatios Souvatzis wrote:
> > > DELAY tends to be a macro, while delay tends to be a function call.
> > 
> > Yes, I see, but given the nature of that call (take some time), does it make
> > sense to have an inline version at all?
> 
> Yes. vaxen. subroutine calls are hideously slow, and can take up to 100
> cycles, according to Jason. :-)

The i386 may (depending on conversations) be getting an inline DELAY() 
that only calls delay() for large values. (I'm going to have to chat
with some people about this.)

> 
> > Is it a matter of how much microseconds we are dalaying? I.E. should
> > "DELAY(10)" be used, but "delay(1000)"?
> 
> Probably.
> 
> Take care,
> 
> Bill