Re: [avr-libc-dev] Huge floating point code (take 2)

[Chris Baugher]

On Tue, 1 Oct 2002, Joerg Wunsch wrote:
> As Chris Baugher wrote:
>
> > Here's a dumb question.  I'm writing a program that does a few floating
> > point divides.  When the divides are compiled in the code gets huge.  I
> > looked in the .ss file and it has lots of stuff that I can't imagine needs
> > to be there just for floating point division.
>
> It seems the bloat mostly comes from pack_f() and unpack_f() which
> come in from libgcc.a.  I haven't found yet their defintions in the
> sources for libgcc to see what they're actually doing.

Hmm, I don't see either of those functions in there.  What I do see is
things like <fp_powerseries>, <fp_powerloop11>, <_log_20> and others.

C|