The G4 revolution

Joe Felsenstein joe at evolution.genetics.washington.edu
Sat Sep 11 09:42:12 EST 1999


In article <7raral$e1h at net.bio.net>,
Hans Sluiman  <h.sluiman at rbge.org.uk> wrote:
>Now that the new generation of G4 Powermacs is about to arrive, can we
>expect a substantial increase in speed and performance compared to G3
>based systems when running phylogenetic applications such as PAUP 4,
>Puzzle, Phylip and fastDNAml? This is particularly relevant for those of
>us doing timeconsuming bootstrap analyses of relatively large data sets
>on Mac platforms.
>
>Obviously, existing softeware needs to be rewritten to make use of the
>new Multiprocessing and AltiVec technologies that are built into the G4.

Not so obvious.  From the point of view of developers like me, we have to
support multiple platforms.  Intel and PowerPC chips leapfog each other
and generally stay within a factor of 2 in speed.  Placing all our
bets and all our efforts into servicing one of these platforms is a
big mistake.  If I had listened to enthusiasts for one chip or another
I would still be trying to perfect my IBM 370 mainframe code and would have
made myself incompatible with everything else.


>However, the extent to which applications will benefit from the enhanced
>FPU speed of the G4 will depend on whether they use floating point
>instructions or integers. How do PAUP et al. work?

PHYLIP uses some floating point but it depends heavily on which program
you are talking about.  DNAML is heavily limited by floating point speed
but (say) DNAPARS uses mostly integers.  I suspect that PAUP* will have
roughly the same properties, using integer operations for the parsimony
runs and floating point for the distance and likelihood runs.


-- 
Joe Felsenstein         joe at genetics.washington.edu
 Dept. of Genetics, Univ. of Washington, Box 357360, Seattle, WA 98195-7360 USA




More information about the Mol-evol mailing list