[CUBE] 1gig G3 in Cube?

[Joseph B. Gurman]

     Robert Jagitsch of PowerLogix posted:

>We have just completed a test of an IBM 750FX PowerPC CPU in a Cube.
>The 750FX is a very low power device which means there are no fans or
>beefy DC/DC boards needed. And it's fast...take a look at the
>benchmarks comparing the 750FX at 900MHz to a 500MHz G4 (these done
>in a 100Mhz bus PowerBook Pismo.)
>
>http://www.powerlogix.com/products2/performance/pismo.html
>
>The clock speed of the 750FX more than makes up for the lack of
>Altivec, see the iTunes rip comparison.
>
>Any interest in this?

     And indeed, many of the benchmarks are quite impressive.

     All of the benchmarks, however, are plotted in a misleading way 
--- but note that I'm not accusing Robert or PowerLogix of doing this 
intentionally; perhaps it was their plotting software --- none of the 
plots start at 0 on the horizontal axis!

     Thus, for instance, the "iTunes CD rip, OS X" plot (second plot), 
which at first glance (i.e., if you don't look at the raw numbers) 
appears to show the G3/900 needing just over half the time of a 
G4/500 to complete the rip, in fact shows a ratio of about 7/8.5, or 
a speed improvement of about 18% --- which is more in line with what 
one would expect from the AltiVec-dependent application. Even in 
cases where the ratio of scores corresponds closely to the ratio of 
clock speeds (e.g. "Cocoa Bench-OS X"), the plot makes it look like 
the ratio is ~ 4:1, not 9:5.

     I apologize if all of this is obvious to everyone on the list, 
it's just been a pet peeve of mine since I first saw it pointed out 
in Edward Tufte's marvelous book, "The Visual Display of Quantitative 
Information,"

http://www.amazon.com/exec/obidos/ASIN/0961392142/qid=1053785788/sr=2-1/ref=sr_2_1/002-9857481-5016012 
.

						Joe Gurman
-- 
"I love deadlines. I love the whooshing sound they make as they go by."
                                                             - Douglas 
Adams, 1952 - 2001

Joseph B. Gurman, NASA Goddard Space Flight Center, Solar Physics
Branch, Greenbelt MD 20771 USA