alan at redhat.com
Thu Jun 9 15:28:46 UTC 2005
On Thu, Jun 09, 2005 at 10:33:03AM -0400, Vladimir Makarov wrote:
> When we use tuning to Intel Nocona, SPECFP2000 is 2.7% and SPECINT2000
> is 1.6% worse on Opteron.
And neither specfp or specint are related to anything but compiler and processor
benchmarketing in my experience.
> Code tuned for Intel Nocona parctically always has smaller size.
That could well make up for some of the other hit on a real systems
More information about the devel