[Beowulf] x86-64 NUMA vs SMP kernel: appl. performance?

[Robert G. Brown]

On Sun, 26 Sep 2004, Greg Lindahl wrote:

> It looks like Bill covered most of the questions, but I'll clean up a
> loose end:

Thanks Bill and Greg;

I'm getting ready to download a 30 day trial version myself.  We have a
small stack of dual opterons in constant production in our department
(and will be getting a lot more, I'm sure).  If the compiler makes my
code run 30% faster, that's like getting a cluster that is 30% larger
for the cost of the compiler, which appears to be a lot less than the
nodes themselves would cost (depending on how it is licensed).

    rgb

Robert G. Brown	                       http://www.phy.duke.edu/~rgb/
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone: 1-919-660-2567  Fax: 919-660-2525     email:rgb at phy.duke.edu