Hi all,
today (a month after this message:
http://www.mail-archive.com/linux-il@cs.huji.ac.il/msg58720.html ) I've
rebuilt LLVM+clang again and benchmarked Freecell Solver trunk after being
compiled with it. It took it 103.167913198471 seconds to solve the Microsoft
32K boards - instead of 105 seconds in the older benchmark, while the
performance of Freecell Solver has improved since then.
On the other hand, with gcc-4.5.0 with "-flto" and "-fwhole-program" Freecell
Solver ran at 85.1303749084473 seconds.
I admit that I ran the gcc benchmark with a good renice and only in the
virtual console, while running the LLVM/clang benchmark without a renice and
in KDE and Compiz, but it still cannot explain the dramatic difference.
I'll keep an eye on clang, but so far it seems that gcc still yields the more
performing code.
Regards,
Shlomi Fish
--
-----------------------------------------------------------------
Shlomi Fish http://www.shlomifish.org/
Why I Love Perl - http://shlom.in/joy-of-perl
God considered inflicting XSLT as the tenth plague of Egypt, but then
decided against it because he thought it would be too evil.
Please reply to list if it's a mailing list post - http://shlom.in/reply .
_______________________________________________
Linux-il mailing list
Linux-il@cs.huji.ac.il
http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
