I tried setting up an AMD 1800+ on a Soyo Dragon Ultra and didn't see as
much speed increase as I had hoped either.  Turns out that I needed to set
some non-default values (cpu clock multiplier if I remember right) in the
bios setup to get it to run at the 1.56Ghz for the AMD.  You might want to
run the MS NT diag (it's a stand alone diskette created from the NT CDs) or
other to see what it estimates the CPU speed to be.

I assume the program isn't disk bound.  (can't hurt to ask, right?)

You might run the top program to make sure that there isn't another program
hogging CPU.

I would guess that if your program is computational bound, your not doing
much with the kernel, so that wouldn't be my first place to look, but that's
more guess than experience.

You might try booting the new computer with Windows and see if it runs any
better.  It has been my experience that Linux runs about 20% faster than
Windows, but that was done on a CPU bound application that's accessing a
large cached DB (no disk i/o) in Oracle.

Hope this helps a little.
  -----Original Message-----
  From: tclug-list-admin at mn-linux.org
[mailto:tclug-list-admin at mn-linux.org]On Behalf Of Randy Clarksean
  Sent: Sunday, October 06, 2002 3:24 PM
  To: tclug
  Subject: [TCLUG] Kernel - AMD - Computational Speed?



  I just put together a new AMD XP 2100+ system and installed RedHat 7.2
The intent of this machine is heavy computational work (1.5 GB RAM, 80 GB
HD, ABIT Motherboard with PC2100 RAM).  I have an older Win2000 Machine,
dual PIII 700 MHz 1 GB RAM - TYAN Motherboard  that I have been using for
computational work.

  I have data from an analysis run on the PIII system (1 CPU) and I just ran
that same analysis on the new AMD21+ system.  The speed improvement was NOT
what I hoped for.

  PIII 700 MHz              96,360 seconds
  AMD XP 2100+          65,040 seconds

  I do realize that there are operating system issues, etc. ... but with all
of that I had anticipated a larger reduction in computational time.  The
code I am running is a commercial code that is developed to run on both
operating systems, so I am fairly sure they work to get the best CPU time on
both platforms.  I was hoping for something on the order of 1/2 the CPU
time - on large computational runs like this every little reduction in time
helps.

  Not being a "kernel" expert by any means ... would it make any sense to
recompile the kernel on my new platform, rather than relying on the kernel
as loaded from the RH7.2 distribution CDs?  Any suggestions would be greatly
appreciated!

  Randy Clarksean, Ph.D., P.E.
  Leading Technology Designs, Inc.
  106 North Boardman Ave.
  P.O. Box N
  New York Mills, MN  56567

  "Excellence can be attained if you Care more than others think is wise,
Risk more than others think is safe, Dream more than others think is
practical, and Expect more than others think is possible." - Author Unknown

  ph: 218-385-3750
  fax:218-385-3751
  email: rclark at lakesplus.com

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://shadowknight.real-time.com/pipermail/tclug-list/attachments/20021006/29ce3220/attachment.html