[Zlib-devel] TR: result of compression test

[Gilles Vollant]

I suggest you made yourself test


to get http://www.kernel.org/pub/linux/kernel/v2.6/linux-2.6.34.1.tar
I just downloaded
http://www.kernel.org/pub/linux/kernel/v2.6/linux-2.6.34.1.tar.bz2
and uncompress with bzip2

the you can get
http://gvollant.free.fr/zlib125_dll_collection.zip


pehaps you can try ohter optimisation on GCC
( -flto for link time optimization, -fprofile-generate...)
-----Message d'origine-----
De : zlib-devel-bounces at madler.net [mailto:zlib-devel-bounces at madler.net] De
la part de Cosmin Truta
Envoyé : lundi 19 juillet 2010 19:02
À : zlib-devel at madler.net
Objet : Re: [Zlib-devel] TR: result of compression test

On Mon, Jul 19, 2010 at 11:31 AM, Stefan Fuhrmann wrote:
> I suspect that most of the ASM code's advantage
> stems from the fact that UNALIGNED_OK is not
> defined in the 1.2.5 code for x86 / x64 CPUs.

The inflation ASM code is vectorized, using MMX, so I am not surprised
to see that ASM has advantages here.
I am, however, surprised to see such a big gain from ASM deflation.

Best regards,
Cosmin

_______________________________________________
Zlib-devel mailing list
Zlib-devel at madler.net
http://mail.madler.net/mailman/listinfo/zlib-devel_madler.net