From mboxrd@z Thu Jan 1 00:00:00 1970 Received: (from majordomo@localhost) by pauillac.inria.fr (8.7.6/8.7.3) id LAA12632; Thu, 2 Aug 2001 11:42:14 +0200 (MET DST) X-Authentication-Warning: pauillac.inria.fr: majordomo set sender to owner-caml-list@pauillac.inria.fr using -f Received: from concorde.inria.fr (concorde.inria.fr [192.93.2.39]) by pauillac.inria.fr (8.7.6/8.7.3) with ESMTP id LAA12605 for ; Thu, 2 Aug 2001 11:42:13 +0200 (MET DST) Received: from fichte.ai.univie.ac.at (fichte.ai.univie.ac.at [131.130.174.156]) by concorde.inria.fr (8.11.1/8.10.0) with ESMTP id f729gCX18081 for ; Thu, 2 Aug 2001 11:42:12 +0200 (MET DST) Received: from kastanie.ai.univie.ac.at (root@kastanie.ai.univie.ac.at [131.130.174.141]) by fichte.ai.univie.ac.at (8.9.3/8.9.3/Debian 8.9.3-21) with ESMTP id LAA11918; Thu, 2 Aug 2001 11:42:11 +0200 Received: (from markus@localhost) by kastanie.ai.univie.ac.at (8.9.3/8.9.3/Debian 8.9.3-21) id LAA16143; Thu, 2 Aug 2001 11:42:11 +0200 Date: Thu, 2 Aug 2001 11:42:11 +0200 From: Markus Mottl To: John R Harrison Cc: caml-list@inria.fr Subject: Re: [Caml-list] "super-compaction" of values Message-ID: <20010802114211.C15610@kastanie.ai.univie.ac.at> References: <200108011606.JAA29019@dhpc0010.pdx.intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: <200108011606.JAA29019@dhpc0010.pdx.intel.com>; from johnh@ichips.intel.com on Wed, Aug 01, 2001 at 09:06:04 -0700 Sender: owner-caml-list@pauillac.inria.fr Precedence: bulk On Wed, 01 Aug 2001, John R Harrison wrote: > About 5 years ago I became very enthusiastic about using hash-consing in > the HOL Light prover to share common subterms of theorems. I was > convinced it would cut space usage dramatically. Well, it's actually not so much space savings that I am interested in, but mostly exploiting (sub-)program equivalences to speed up expensive computations. > I ended up writing a naive implementation, without making stuff > garbage-collectible, but only using it for structures I knew were > persistent. To my chagrin, it turned out that Malcolm was absolutely > right. Space usage actually went *up*, presumably because the hashing > datastructures were large enough to overwhelm the small amount of > sharing. I can well imagine this. In my case it's not "real" programming languages that I have to deal with, but rather quite confined ones, and there are likely to be hundreds of not-so-large programs that mostly differ by a few subtrees only (= lots of opportunities for sharing). It seems worth a try... Regards, Markus Mottl -- Markus Mottl markus@oefai.at Austrian Research Institute for Artificial Intelligence http://www.oefai.at/~markus ------------------- Bug reports: http://caml.inria.fr/bin/caml-bugs FAQ: http://caml.inria.fr/FAQ/ To unsubscribe, mail caml-list-request@inria.fr Archives: http://caml.inria.fr