Date: Sun, 5 Jan 2003 02:49:23 -0500 From: Rahul Siddharthan <rsidd@online.fr> To: Greg 'groggy' Lehey <grog@FreeBSD.org> Cc: Brett Glass <brett@lariat.org>, chat@FreeBSD.ORG Subject: Re: Bystander shot by a spam filter Message-ID: <20030105074923.GA4956@papagena.rockefeller.edu> In-Reply-To: <20030105073804.GA72674@wantadilla.lemis.com> References: <4.3.2.7.2.20030104201251.029387d0@localhost> <4.3.2.7.2.20030104112015.026a5530@localhost> <4.3.2.7.2.20030104201251.029387d0@localhost> <4.3.2.7.2.20030104202908.03c3b100@localhost> <20030105073804.GA72674@wantadilla.lemis.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Greg 'groggy' Lehey said on Jan 5, 2003 at 18:08:04: > On Saturday, 4 January 2003 at 20:31:52 -0700, Brett Glass wrote: > > At 08:22 PM 1/4/2003, Rahul Siddharthan wrote: > > > >> So you disassembled and examined it? Or did you benchmark it? > > > > Actually, both. In heavy matrix math (with lots of determinants, > > inversions, etc.), Intel beats GCC by a mile. Ditto on "bit-banging" > > (bit manipulation, especially for graphics). Didn't disassemble the > > code, though; just asked the compilers to output assembler. The > > GCC output was so naive it made me wince. > > I'd be interested to see some examples. I'd be surprised to see them. As I said, my code involves plenty of nested loops, linear algebra and floating point operations and the performance is quite respectable. According to the benchmarks I cited earlier, http://www.coyotegulch.com/reviews/intel_comp/intel_gcc_bench2.html (look at the SciMark benchmark) gcc actually beats intel on the sparse matrix multiply on the Pentium IV (which generally emerges as Intel's strong platform) and runs it pretty close on LU decomposition. - Rahul To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-chat" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20030105074923.GA4956>