Date: Sat, 30 Jul 2005 00:22:25 -0700 From: ray@redshift.com To: Ken Gunderson <kgunders@teamcool.net>,freebsd-amd64@freebsd.org Subject: Re: Benchmarks: AMD64 vs i386 on Dual 246 Opteron Message-ID: <3.0.1.32.20050730002225.00a718a0@pop.redshift.com> In-Reply-To: <20050729154713.48b3fe08.kgunders@teamcool.net> References: <20050729083844.A78480@cons.org> <3.0.1.32.20050729023234.00a8afb8@pop.redshift.com> <20050729083844.A78480@cons.org>
next in thread | previous in thread | raw e-mail | index | archive | help
At 03:47 PM 7/29/2005 -0600, Ken Gunderson wrote: | On Fri, 29 Jul 2005 08:38:44 -0400 | Martin Cracauer <cracauer@cons.org> wrote: | | > ray@redshift.com wrote on Fri, Jul 29, 2005 at 02:32:34AM -0700: | > > At 07:35 PM 7/28/2005 -0400, Martin Cracauer wrote: | > > | > While this sounded like a long shot, I loaded FreeBSD 5.4 i386 on the machine | > > | > and after applying the exact same configuration to the OS, Apache, PHP and | > > | > MySQL, re-ran the benchmarks. Much to my surprise, just changing the OS | > > from 64 | > > | > bit to 32 bit caused the machine to double in speed. The results are attached | > > | > in an Excel spreadsheet. So the exact same machine, running the identical | > > | > configuration, performed roughly twice as fast when running FreeBSD 5.4 i386 vs | > > | > FreeBSD 5.4 AMD64. Something about this seems so wrong to me :-) | > > | | > > | I'm sorry but I cannot support these findings. I don't have | > > | cut'n'paste numbers handy, but generally 64 bits speed up things quite | > > | a bit for me. I have seen slowdown in 64 bit mode in e.g. bzip2 but | > > | generally there is a speedup. | > > | > > On the AMD64, I am pretty sure that I did *not* include the following lines like | > > I normally do on the i386: | > > | > > options SMP # Symmetric MultiProcessor Kernel | > > device apic # I/O APIC | > | > I occurse to me that not using one of the two CPUs might have | > something to do with a performance drop of approximately 50% :-) | | Not to beat a dead horse here, but for the record, the numbers I | reported were definitely for a SMP kernel. | | -- | Best regards, | | Ken Gunderson | I'm going to have to repeat my benchmarks again on the AMD when our actual machine shows up (I was testing on an evaluation machine that had to be sent back). I'll post the info again and this time will make absolutely sure the AMD kernel is configured specifically with SMP - even though I seem to remember adding that line, I can't be sure now because the machine is gone and I erased the configuration. Ray
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3.0.1.32.20050730002225.00a718a0>