Re: How fast is that machine???

[Michael Fischer <Fischer@YALE.ARPA>]

    I  have  run a few timing tests on the various machines around here,
    including the 3670 we have on loan from Symbolics.  As expected, the
    conclusion is  that lisp  machines  rate an almost unqualified
    WONDERFUL.  The major qualification is that closures, which T handles
    exquisitely, lose big  in  Zetalisp:   While most  tests  have  the
    3670  being  2  to  3  times a TERN, closure intensive computations
    are equally fast on TERN and 3670. 

    THESE NUMBERS ARE NOT THE LAST WORD:  both  T  and  ZETALISP  are
    subject  to change,  with  T more likely to improve than Zetalisp.
    Tests one and two will improve when a disk appears on the TERN.
    However, they ARE indicitive.  Also, theser  tests do not include
    tests of garbage collection or compilation, which in my opinion are
    significantly superior on the 3670. 

There is more to a computer than how fast it runs, just as the top
speed of a car is of only minor importance compared to comfort, 
maneuverability, etc.  Before one rates a machine as "an almost
unqualified WONDERFUL", one needs to ask some hard questions:

1.  What is the cost per node for all that speed?  If a node that
costs twice as much runs 3-5 times as fast, is that really a good
deal, given that most of the time it sits practically idle doing
text editing?

2.  What is the true cost of switching systems, once you factor in
the additional staff needed to support it and the increased difficulty
of interaction with others in the local community?

3.  How good is the software?  Does it support transparent remote file 
access?  Mail?  TeX?  Scribe?  Graphics?  Yale-style editors?  C?

4.  How good is the Lisp?  Does it have the clean semantics to
facilitate the development of large systems that T has?  How long
does it take to learn to use effectively (i.e. what is its "kludginess
coefficient")?  How many of our faculty and students have the time
to invest in learning to use it effectively?

5.  What is its growth path, i.e. if you take similar benchmarks in
two years' time, what will you expect to see?  Is there any reason to
believe the Symbolics class of machines will remain faster than the
Apollo class?  Maybe I hold out too much hope for the new T compiler,
but I had the impression that it might yield a speed increase of
two to three times.  If so, the Tern running T would already approach
the 3670.

--Mike

-------