Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!ucbvax!hplabs!otter!ijd
From: ijd@otter.hpl.hp.com (Ian Dickinson)
Newsgroups: comp.lang.prolog
Subject: Re: Re: SB Prolog
Message-ID: <1600020@otter.hpl.hp.com>
Date: 4 Apr 89 09:10:30 GMT
References: <629@cwjcc.CWRU.Edu>
Organization: Hewlett-Packard Laboratories, Bristol, UK.
Lines: 33

My experience is different.  I have used SBprolog for a couple of medium
sized programs, and a whole host of "back-of-an-envelope" experiments.
My verdict was that it is *very* fast - beating Poplog prolog, a couple of
experimental systems and (unfortunately) HPProlog hands down.  My
reservations were more that it was a bit too flaky, and tended to bomb
for no apparent reason about once per day.

I was running on a dedicated (ie to me!) HP 9000/350 workstation. Differences
in percieved speed could be due to having different versions, or it
could just be superior hardware, plug, plug [no! no! I didn't say that :-) ]

Just for the record, I benchmarked (yes, I know we all hate that word in
this context) SBProlog 2.5, using naive reverse (as supplied with the 
package), 1000 iterations, at
	36K Lips compiled
	13K Lips interpreted

Last time I used Quintus - in a previous life, some 2 years ago - I recall
getting 45 to 50 K Lips, compiled, out of a Sun 3.

Personally, I think the best improvement they could make to SBP is the
addition of a garbage collector, which I understand will happen in next
major release.  That and supplying the documentation in LaTeX format anyway.

Enough rambling, I like it, and I hope Saumya Debray keeps up the good work.

Ian.


| Ian Dickinson,  HP Labs, Information Systems Centre,    Bristol, England  |
| net: ijd%otter@hplabs.hp.com        |                                     |
|  or:       ijd@hplb.uucp            |  ?-  mind(X),  body(X),  spirit(X). |
| These opinions are all my own work. |                                     |