Rereading what Wesley writes, it's possible I misinterpreted what he was
asking:

> So how do I measure the added performance from say a 918 to A400-100-110?
Is
> the
>  A400  "1.7 HP Perf. Units" better than the 918? Does that mean it is more
> than
>  twice as powerful? So then I would calculate about a 130% increase in
> processing
>  power.... is this correct? With these new processors it's difficult to
>  understand what they are trying to tell me (us).

This may be recapitulating the obvious, but when you say that "1.7 is more
than twice as powerful" you may be repeating what you often hear when people
say "a 100% increase."  In this case, that isn't what is meant.

In normal relative statistics, 1.0 means that two things are exactly equal.
When something has a relative performance index of 1.7, that means that it's
slightly less than twice as fast -- or said another way, it's 70% faster than
the reference box, which in this case was a 918.

Similarly, if something had a relative performance of 2.0, it would then be
exactly twice as fast as the reference box and you would experience a
performance *increase* of 100%. "Increase" is the magic word in this second
kind of comparison, and the formula for that kind of measurement is:

               (new box performance - old box performance)
percentage increase =  -----------------------------------------  x 100%
                                  (old box performance)

The units that the performances are measured in don't matter, so long as
they're all the same. They wash out in the division and you're left with
"percentage increase".

In the case of the A400-100-110 vs. 918 comparison, if you substitute in all
of the numbers, you will see that you will get a 70% performance increase
with the new box.

Wirt Atmar

* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *