Date: Fri, 22 Jun 2012 17:44:32 +0300 From: Volodymyr Kostyrko <c.kworr@gmail.com> To: Wojciech Puchar <wojtek@wojtek.tensor.gdynia.pl> Cc: freebsd-questions@FreeBSD.org Subject: Re: Why Clang Message-ID: <4FE484D0.7030005@gmail.com> In-Reply-To: <alpine.BSF.2.00.1206221624390.27991@wojtek.tensor.gdynia.pl> References: <201206221240.q5MCeVEF041711@mail.r-bonomi.com> <alpine.BSF.2.00.1206221624390.27991@wojtek.tensor.gdynia.pl>
next in thread | previous in thread | raw e-mail | index | archive | help
Wojciech Puchar wrote:
>> Because it doesn't address an of the *OTHER* valid reasons why GCC is
>> being replaced -- among them:
>> 1) GCC's continuously increasing propensity to generate "bad code",
>
> examples? All test shows that gcc code is not only bad, but very good.
> Why are you just saying things you know isn't true?
0k, what if I add my example?
Hardware:
Processor: Intel Xeon E5620 (16 Cores), Motherboard: Supermicro X8DT3
1234567890, Memory: 24576MB, Disk: SEAGATE ST3146855SS S527 + SEAGATE
ST31000640SS 0001 + SEAGATE ST31000640SS 0001 + SEAGATE ST3146855SS S528
+ TOSHIBA Trans 1.00 + TEAC DV-28S-V 1.0B
Software:
OS: FreeBSD, Kernel: 9.0-RELEASE-p3 (x86_64), Compiler: GCC 4.2.1
20070831 + Clang 3.0 (SVN 142614), File-System: zfs
CPUTYPE=core2
clang 3.0
Test project /tmp/ports/usr/ports/graphics/png/work/libpng-1.5.11
Start 1: pngtest
1/2 Test #1: pngtest .......................... Passed 0.02 sec
Start 2: pngvalid
2/2 Test #2: pngvalid ......................... Passed 14.03 sec
gcc 4.6 (lang/gcc, USE_GCC=4.6+)
Test project /tmp/ports/usr/ports/graphics/png/work/libpng-1.5.11
Start 1: pngtest
1/2 Test #1: pngtest .......................... Passed 0.02 sec
Start 2: pngvalid
2/2 Test #2: pngvalid ......................... Passed 14.40 sec
gcc 4.2.1
Test project /tmp/ports/usr/ports/graphics/png/work/libpng-1.5.11
Start 1: pngtest
1/2 Test #1: pngtest .......................... Passed 0.02 sec
Start 2: pngvalid
2/2 Test #2: pngvalid ......................... Passed 14.96 sec
This one shows that clang is superior to both gcc 4.2.1 and gcc 4.6.
I haven't test data now but a month or so ago I tested them on one of
the Alioth Shootout examples (nestedloop probably). gcc 4.2.1 was
winning, clang was close with fractions of percent drop of speed but gcc
4.6 was off for nearly 7%.
>> 3) The continuously increasing trend of introducing 'non standard'
>> features,
> No need to use them.
There's no 'Unsubscribe me' link included...
--
Sphinx of black quartz judge my vow.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4FE484D0.7030005>