Date: Tue, 2 Aug 2005 11:13:55 -0700 From: "David O'Brien" <obrien@freebsd.org> To: ray@redshift.com Cc: freebsd-amd64@freebsd.org Subject: Re: Benchmarks: AMD64 vs i386 on Dual 246 Opteron Message-ID: <20050802181355.GP71672@dragon.NUXI.org> In-Reply-To: <3.0.1.32.20050802105941.00a547c8@pop.redshift.com> References: <3.0.1.32.20050729020609.00a6a240@pop.redshift.com> <3.0.1.32.20050728013152.00a4d188@pop.redshift.com> <3.0.1.32.20050728013152.00a4d188@pop.redshift.com> <3.0.1.32.20050729020609.00a6a240@pop.redshift.com> <3.0.1.32.20050802105941.00a547c8@pop.redshift.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, Aug 02, 2005 at 10:59:41AM -0700, ray@redshift.com wrote: > Thanks David - needless to say, I'm going to be re-running all my benchmarks ... > I'll post my results when I have another machine in here and re-run the exact > same benchmarks. This time I will be certain to confirm the situation on the > SMP kernel with regard to the AMD64 branch. You need to be careful about what you are testing. If you dual-boot 64-bit and 32-bit from the same disk, don't forget one of them will be on a faster area of the disk. So you need a 2nd data disk that is newfs'ed between each run to do your benchmarks on. At a minium, you should build a set of 32-bit binaries on FreeBSD/i386 with CPUTYPE=k8. Run benchmarks under the 32-bit OS. Then run the same benchmarks with the 32-bit binaries on the 64-bit OS. (ie, you're just changing the kernel). Then build 64-bit binaries and benchmark them. You should use RELENG_6, not 5.x to make sure you're getting the best K8 GCC we support along with the optimizations we've added since 5.x was released. -- -- David (obrien@FreeBSD.org)
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20050802181355.GP71672>