>From the desk of Keith Lysiak
I have some info that may be of some interest and hopefully
start some discussion for my benefit. I too expected significant
speed increased from the Xeon cache so I recently purchased a dual 550
MHz PIII Xeon with 1GB of RAM. This was to replace my dual Pro 200.
I set up two bench marking programs. The first a NEC4 1000
segment dipole. The second a 2.5 million cell 3D FDTD code requiring
80MB. I will skip the detailed times for various machines and get to
your interest. I recorded these times:
NEC (sec) FDTD (min:sec)
PIII 500 MHz 123.698 2:27
PIII 550 MHz Xeon 111.875 2:17
As you can see, the time reduction is approximately equal to
the clock speed. I see no apparent reduction due to the full speed
caching. I have double checked both computers and with Dell and INTEL
to insure proper setup. I was told there were no setting to
enable/disable the Xeon full-speed caching and that possibly the
software was not compiled to take advantage of the Xeon caching. This
does not make sense to me. Software does not have to be specifically
compiled for a Pro 200 to take advantage of the caching.
End of Message
Received on Tue Jul 20 1999 - 16:37:48 EDT
This archive was generated by hypermail 2.2.0 : Sat Oct 02 2010 - 00:10:39 EDT