Re[4]: [avr-gcc-list] Re: GCC-AVR Update (20082003)

[Stephan Eisvogel]

On Friday, August 22, 2003, 11:38:49 PM, James Dabbs wrote:
> [..] But the compiler output in my case from IAR was *bigger* than
> GCC, and in fact wouldn't fit into the part.  This was with the IAR
> optimizer on full blast, favoring small code, compiling C++ source.  Also a
> section of code (the XTEA encryption algorithm) caused IAR to emit bad code
> at maximum optimization, while GCC dealt with it.

Well  actually  I  ment:  gcc has to produce code sizes of one half to
maybe  one  third  of  _what IAR's produces_ to start turning heads. I
agree  that  gcc produces pretty tight code already, although there is
certainly room for improvement.

One project of mine involved porting the hash algorithm MD5 to AVR and
in  the  end, using gcc with "-Os" in pure C I managed to cram it into
~1950  bytes of flash while not using more than ~130 bytes of RAM peak
for  everything  from start until finish. The code produced looks very
much ok, considering MD5 was designed for 32-bit machines.

Just my 2 € cents.
Stephan