Path: csiph.com!v102.xanadu-bbs.net!xanadu-bbs.net!feeder.erje.net!eu.feeder.erje.net!news.mixmin.net!aioe.org!.POSTED!not-for-mail
From: "Ed" <invalid@nospam.com>
Newsgroups: comp.lang.forth
Subject: Re: Portability
Date: Wed, 10 Apr 2013 14:02:17 +1000
Organization: Aioe.org NNTP Server
Lines: 27
Message-ID: <kk2o7m$p45$1@speranza.aioe.org>
References: <Q_ednQPtEZJqe8LMnZ2dnUVZ7vqdnZ2d@giganews.com><516007a4$0$619$e4fe514c@dreader34.news.xs4all.nl> <e8qdnQ9Z-t56G_3MnZ2dnUVZ_qadnZ2d@supernews.com><kjvlfb$bfv$1@speranza.aioe.org> <51636f44$0$596$e4fe514c@dreader34.news.xs4all.nl><2013Apr9.092129@mips.complang.tuwien.ac.at> <ZoKdnReYWZBNU_7MnZ2dnUVZ_judnZ2d@supernews.com><2013Apr9.144119@mips.complang.tuwien.ac.at> <0O6dneB1Wq85xfnMnZ2dnUVZ_jCdnZ2d@supernews.com><kk1u0o$n4e$1@online.de> <K4ednVfAlIoqDfnMnZ2dnUVZ_oGdnZ2d@supernews.com> <c8cb3aa7-150c-461a-8922-3fe690d54599@googlegroups.com>
NNTP-Posting-Host: KKp65Z5aRH6eo+MI4TEhnQ.user.speranza.aioe.org
X-Complaints-To: abuse@aioe.org
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2615.200
X-Notice: Filtered by postfilter v. 0.8.2
X-Newsreader: Microsoft Outlook Express 5.00.2615.200
X-Priority: 3
X-MSMail-Priority: Normal
Xref: csiph.com comp.lang.forth:21572

Chris wrote:
> ...
> for polyFORTH ISD-4 Segmented 8086 model that runs on MS-DOS
>
> : NEXT   CS: LODS   W 0 XCHG   W ) CS: LIP ;
>
> The standalone 8086 model is probably better.

To my knowledge polyForth single-segment 8086 NEXT is coded
in-line as:

     LODSW
     XCHG AX,DI
     JMP WORD PTR [DI]

In FD there was an article:

    "A Faster NEXT Loop"  (FD V9N6)

This was MVP-FORTH (8086, Forth-79) whose original NEXT was 7
instructions (compared to FIG's 5).  It explained how NEXT could be
reduced to 3 instructions and the mods necessary.  It was essentially
as above but used a different register and NEXT was not in-line.
The article claimed a 25% speed improvement over the original.