update: was [[rescue] Did I kill the octane?]

[Joshua Boyd]

On Sun, Dec 28, 2003 at 07:30:27PM -0600, Scott Newell wrote:
> >   For FP, try float.c...it's ancient and well understood.  You can find 
> >a copy on my ftp server...I forget where at the moment, probably 
> >/pub/misc but I can't check due to lack of connectivity (I'm typing 
> >this offline).
> 
> flops.c?
> 
> >   I'm very much interested in learning your results.
> 
> http://www.digital-pickle.dyndns.org/sgi/bench.html

Very interesting.  So, for equal clock rates, the R12k doesn't appear to
be all that much better, and with only -O3 it is worse.

I guess this means it is pointless to do the upgrade without moving the
resistors then...  I wonder how hard a 360 or 400 mhz upgrade would be.
I had been under the impression that moving from a R10k @ 250 to any
R12k was supposed to feel like more of a boost than the clock jump would
lead one to expect.  Guess that was wrong though.

I wonder about the int results though.