Double oops...

Robert G. Brown rgb at
Fri Feb 9 11:30:11 PST 2001

On Fri, 9 Feb 2001, Judd Tracy wrote:

> > Again, the order of performance is 466 celeron, 933 PIII, 800 Athlons,
> > and 667 EV67 Alphas, in increasing order and obvious groupings.
> Could you put a key into the graph?

I will in "production" (read, as I make a final copy for
web-publication).  This is just a temporary send in response to the
performance issue I was addressing.  I may start breaking it down into
separate figures by architecture to make it simpler to follow at this
point.  For the moment:

  Open circles at bottom are 466 MHz Celeron (lucifer)
  Open triangles are 933 MHz PIII (ganesh)
  Filled pentagons are 933 MHz Dual PIII (b14 - both processors)
  Open pentagons are 933 MHz Dual PIII (b14 - one processor, other idle)
  Open squares are 750 MHz Athlon (grads-15)
  Filled squares are 800 MHz Athlon Tbird (eve)
  Open hexagons are 667 MHz Alpha EV67 with gcc compiler (qcd4)
  Filled hexagons are 677 MHz Alpha EV67 with ccc compiler (qcd4)

The results for very small vector sizes may still be somewhat
unreliable -- they are "reproducible" but the particular values may
still contain artifacts of the measuring process or interactions with
the kernel and timer.


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

More information about the Beowulf mailing list