[Bug target/19133] march=athlon can produce slower code than march=i686

[ornati at fastwebnet dot it]

------- Additional Comments From ornati at fastwebnet dot it  2005-02-08 16:13 -------
gcc -v

Using built-in specs.
Configured with:
/var/tmp/portage/gcc-4.0.0_alpha20050130/work/gcc-4.0-20050130/configure
--enable-version-specific-runtime-libs --prefix=/usr
--bindir=/usr/i686-pc-linux-gnu/gcc-bin/4.0.0
--includedir=/usr/lib/gcc/i686-pc-linux-gnu/4.0.0/include
--datadir=/usr/share/gcc-data/i686-pc-linux-gnu/4.0.0
--mandir=/usr/share/gcc-data/i686-pc-linux-gnu/4.0.0/man
--infodir=/usr/share/gcc-data/i686-pc-linux-gnu/4.0.0/info
--with-gxx-include-dir=/usr/lib/gcc/i686-pc-linux-gnu/4.0.0/include/g++-v4
--host=i686-pc-linux-gnu --disable-altivec --enable-nls
--without-included-gettext --enable-__cxa_atexit --enable-clocale=gnu
--with-system-zlib --disable-checking --disable-werror
--disable-libunwind-exceptions --enable-shared --enable-threads=posix
--disable-multilib --disable-libgcj --enable-languages=c,c++
Thread model: posix
gcc version 4.0.0 alpha20050130 (Gentoo Linux 4.0.0_alpha20050130)


paolo@tux gcc_test $ time ./test-i686 

real    0m8.539s
user    0m8.376s
sys     0m0.001s
paolo@tux gcc_test $ time ./test-athlon 

real    0m6.168s
user    0m6.031s
sys     0m0.005s


As the numbers show, the problem go away for "-march=athlon"... but now
"-march=i686" version is MUCH slower than before (with GCC 3.3.x).

Seems like I've found a very strange test-case...

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19133