Thorsten had it right, although it gets more interesting when you look at
the times.  Here are the times:

3min 53sec        C) desktop p4  2.4ghz (windows)
"5 & 1/2 min"     D) desktop socket-a athlon 2.0ghz (windows)
7min +/- 15sec    A) laptop  p3    1ghz (windows)
"27min"           B) laptop  g4  867mhz (osx)

D) and B) belong to my friend who didn't give me exact seconds.  On his
desktop athlon, he first ran it through VNC from the mac but then re-ran it
without to get the slightly better time reported here. The mac's time is
after he stopped playing mp3s etc.  With that, the two desktop times are
"left-alone" times and the two laptop times include online communications,
pointing, clicking, typing; nothing too heavy.  The p3 consistently got
about 6:45 if left alone.

Somewhere I got the idea that the G4 was pretty efficient per clock speed at
doing floating point so even at a slightly lower clock speed I would have
expected it to be somewhat faster or even just "somewhat" slower than the
p3 1ghz.  Yet it took the g4 roughly 4 times as long as the p3.   The same
thing goes for the athlon.  Shouldn't it at 2ghz have been MORE than twice
as fast as the p3 at 1ghz?  Shouldn't the athlon have met or beaten the
2.4ghz p4?

With this scene, memory shouldn't be an issue. The scene consists of a few
dozen simple primitives and nothing special or fancy, and no functions.

Like I said - it wasn't what I expected.  Any more thoughts?

Thanks,
Charles

Post a reply to this message