NetNews Usenet Archive 1992 #31/NN_1992_31.iso/spool/comp/lang/lisp/mcl/1887

Internet Message Format | 1992-12-24 | 1.9 KB

Path: sparky!uunet!usc!cs.utexas.edu!sun-barr!ames!agate!apple!goofy!cambridge.apple.com![e-mail]
From: [e-mail] (EHN & DIJ Oakley,GB,IDV)
Newsgroups: comp.lang.lisp.mcl
Subject: still checking out farms
Message-ID: <[e-mail]>
Date: 24 Dec 92 16:04:00 GMT
Sender: [e-mail]
Lines: 31
Approved: [e-mail].C0M

Friends,

1.  A little parenthetical clarification on the genetic programming business:

Why all this eval-ing (or, to be more accurate, fast-eval-ing - a cut-down and
faster route provided by Koza)?  Well, the program actually generates loads of
Lisp S-expressions which are *not* known a priori.  It generates a load (50 -
5000) randomly at the start of a run, and then manipulates them during a run,
by crossing over segments of them to generate new S-expressions, which then
have to be fast-evaled.  Normally, as far as I am aware, there is no way of
avoiding all this fast-eval-ing, since none of the S-expressions is known a
priori, and all are generated and used on the fly.  Any bright ideas are always
welcome, though!

2.  A little price comparison:

I have a friend who recently bought two of the new HP 700 series workstations.
In the UK, they cost =L40 000 apiece, to which you have to add the cost of
Lucid, etc.  Prices for 13 x Quadra 700, or 10 x Quadra 950, or 25 x Duo 230,
or 50 x Classic II - having roughly the same total CPU power, and each suitably
configured for 'farming' - are about =L45 000 after discount.  Fortunately,
power consumption and space do not come into the calculation for us!

Thanks again to all those who have replied,

Howard.
Howard Oakley,                      * [e-mail]
EHN & DIJ Oakley, Brooklands Lodge, * AppleLink UK0392
Park View Close, Wroxall, Ventnor,  * CompuServe 70734,120
Isle of Wight UK PO38 3EQ           * Tel [phone], fax 853253