Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!usc!cs.utexas.edu!uunet!wrkgrp!ets
From: ets@wrkgrp.COM (Edward T Spire)
Newsgroups: comp.editors
Subject: UNIX vs. mainframe editors
Message-ID: <1991Mar26.230631.2727@wrkgrp.COM>
Date: 26 Mar 91 23:06:31 GMT
Organization: The Workstation Group
Lines: 30

>In article <1991Mar20.140959.2939@scrumpy@.bnr.ca> bnrmtl!lewis@larry.mcrcim.mcgill.edu writes:
>>Pierre LEWIS said...
>>
>>And I, for one, have always liked XEDIT and think it's certainly a very
>>powerful editor (it can sort a file much faster that sort(1) on a Sparc).
>
>Anyway, I want to know what you are saying here... That XEDIT (running
>on a mainframe, say a 3090 or something along that order) can sort
>faster than a Sparc???? I WOULD HOPE SO!!!!  Ignoring all the MipsHype
>that suggests a "15Mips" Sparcstation 1+ might be able to outcompute
>a "3Mips" 3033 (I sincerely doubt it... mips != mips in most cases),
>anyway, ignoring all that hype, a mainframe should run circles around
>a Sparc... So what does that prove???

Trivial comparison, seive for primes up to 1000:

    IBM 4361-5 (1.5 MIPS) 6 seconds
    IBM RS/6000 (27 MIPS) 2 seconds.

That would make one million 370 instructions about the same as six
million IBM RISC instructions (assuming the two version of REXX are
somewhat similar in efficiency).

So a 3 MIPS 3033 would be about the same as a 30 MIPS RISC machine (?)
Might be possible...  It's still a whole bunch cheaper!

Anyway, so maybe CMS XEDIT has a great sort algorithm (or is it that
CMS SORT uses a crummy one?  Comparing Unix XEDIT SORT performance to
CMS XEDIT SORT performance, I suspect it's the former), "SO WHAT?"
This is a minor, little used feature of an editor.