Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/5/84; site terak.UUCP
Path: utzoo!linus!decvax!tektronix!hplabs!hao!noao!terak!doug
From: doug@terak.UUCP (Doug Pardee)
Newsgroups: net.micro
Subject: Re: Intel 80*86 vs IBM 370
Message-ID: <543@terak.UUCP>
Date: Fri, 10-May-85 20:13:11 EDT
Article-I.D.: terak.543
Posted: Fri May 10 20:13:11 1985
Date-Received: Mon, 10-Jun-85 05:16:31 EDT
References: <10441@brl-tgr.ARPA> <898@peora.UUCP>
Organization: Terak Corporation, Scottsdale, AZ, USA
Lines: 19

> When I used to do systems programming under MVS the base displacement
> addressing scheme that IBM uses became tiresome very quickly for large
> programs. Is this better or worse than the problems with segmentation
> on the 80*86?

The most important difference is that 370 addresses are actually linear.
You put an entire address in a register, not part in this register and
part in that register.  You can pass the address of something around
simply.  You can pick up the address and use it without any fuss.  And
you only need to use one register to point to an item, and you've got 15
to choose from so you can point to a bunch of different data items at
the same time without loading and storing pointers all the time.

It's not really fair to call the 370 a "segmented" architecture.  It's
a linear architecture, but there is no absolute addressing mode (except
for the first 4K).  Where this gets annoying is for branching...
-- 
Doug Pardee -- Terak Corp. -- !{ihnp4,seismo,decvax}!noao!terak!doug
               ^^^^^--- soon to be CalComp