From mboxrd@z Thu Jan  1 00:00:00 1970
Received: (from weis@localhost) by pauillac.inria.fr (8.7.6/8.7.3) id MAA29725 for caml-red; Sat, 7 Oct 2000 12:12:12 +0200 (MET DST)
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 KAA20210 for <caml-list@pauillac.inria.fr>; Sat, 7 Oct 2000 10:19:43 +0200 (MET DST)
Received: from suburbia.net (suburbia.net [203.4.184.1])
	by concorde.inria.fr (8.10.0/8.10.0) with ESMTP id e978JbT02741
	for <caml-list@inria.fr>; Sat, 7 Oct 2000 10:19:39 +0200 (MET DST)
Received: by suburbia.net (Postfix, from userid 110)
	id BA6226C4D2; Sat,  7 Oct 2000 19:19:29 +1100 (EST)
Subject: automatic translation in Team PLClub ICFP'2000 entry
In-Reply-To: <20001006152659G.sumii@saul.cis.upenn.edu> "from eijiro_sumii@anet.ne.jp
	at Oct 6, 2000 03:26:59 pm"
To: sumii@saul.cis.upenn.edu
Date: Sat, 7 Oct 2000 19:19:29 +1100 (EST)
Cc: caml-list@inria.fr
X-Mailer: ELM [version 2.4ME+ PL78 (25)]
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=US-ASCII
Message-Id: <20001007081929.BA6226C4D2@suburbia.net>
From: proff@suburbia.net (Julian Assange)
Sender: weis@pauillac.inria.fr

> compilation itself (including ocamlopt.opt) took some time, which
> didn't pay.

There must be an adapative break even point though. e.g if you're
n/m pixils done in s seconds, and the average cost of compilation
is c times the simple rendering speed per line and the average speedup
is q, then subtracting the pixils already done, work out whether

	(m-n) * p < c*lines*p' + (m-n)*p*q
	where
		p = (m-n)/s 		rendering speed
		p' =			simple rendering speed (speed of
					a delay loop or uncomplicated pixel etc)
					

		
And if so, compile.