NetNews Usenet Archive 1993 #1/NN_1993_1.iso/spool/comp/sys/mac/oop/macapp3/394

Internet Message Format | 1993-01-11 | 1.6 KB

Path: sparky!uunet!olivea!apple!applelink.apple.com
From: [e-mail]
Newsgroups: comp.sys.mac.oop.macapp3
Subject: Re: Total Chaos
Message-ID: <[e-mail]>
Date: 11 Jan 93 08:11:00 GMT
Sender: [e-mail]
Organization: AppleLink Gateway
Lines: 29

From: [e-mail]
To: MACAPP3TECH$@applelink.apple.com

John,
   Emulation isn't necessarily going to produce slow running of Mac OS
applications on the new hardware.  I've seen instances in the past in which
such emulations on new hardware ran faster than the old machines did (NCR
315 emulated on the Criterion, as one example).  In this case, we have a
situation where a typical Mac OS application spends considerable time in
the toolbox routines.  The key ones of those can be (and, it is said, will
be) recoded in native code, giving a boost.

   On top of that, the 68000 is an emulator, anyhow.  The visible
instruction set is emulated by the microcode.  Shortly after the chip came
out, both NCR and IBM took a long look at re-microcoding it to run their
mainframe instruction sets (Criterion and 370).  Both concluded that there
wasn't quite enough silicon available (there was another problem...the
microcode was a one-man show, he was gone, and the "source" was only
slightly more organized than Woz's notes on the Apple ][ ROM).

   I don't know whether the 68020, 30, and 40 are also microcoded, or not.

   We won't know whether Mac OS running on the new machines is bad or not
until we see it.

--------------
John W. Baxter   [e-mail]   [BAXTER.JOHN on AppleLink, from
AppleLink only.]