Re: [avr-libc-dev] Amazing result!

[Peeter Vois]

My internet connection was down meanwhile, and shortly after writing
this message I wanted to send another. Seems like without using libm gcc
can make the floating point multiplication and that one is taking lots
of memory. After specifying exactly that the old libm must be used, the
code was small too.

Excuseme about the misleading message. I was surprised too and forgot
the little piece of truth that my MUL code is almost he same length in
commands as the slow code.

Peeter

Ühel ilusal kenal päeval Mon, 17 Oct 2005 20:59:39 +0300
Peeter Vois <address@hidden> kirjutas:

> I have made first run my multiplication application with MUL modified
> libm. This works as expected - the calculation is in its 2^-23
> calculation error limit. But the amazing results come from the size of
> application.
> 
> using libm without MUL, flash size 1640 bytes
> using libm with MUL, flash size 468 bytes !!!
> 
> Best Regards ...
> Peeter
> 
> 
> 
> _______________________________________________
> AVR-libc-dev mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/avr-libc-dev