Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!clyde.concordia.ca!uunet!snorkelwacker!apple!well!dsmall
From: dsmall@well.UUCP (David Small)
Newsgroups: comp.sys.atari.st
Subject: Re: GCR vs. Mac Plus
Message-ID: <15094@well.UUCP>
Date: 19 Dec 89 04:25:42 GMT
References: <891206.18562778.024049@SFA.CP6>
Reply-To: dsmall@well.UUCP (David Small)
Organization: Whole Earth 'Lectronic Link, Sausalito, CA
Lines: 42


This must be my night for notes.. *grin*

The basenote discusses speed differences between the Mac Plus and GCR.

	One problem with marketing an "emulator" is people think it's
slow. They see the CP/M emulator, PC-Ditto, and whatnot. The difference
is running native 68000 machine code vs. interpreting 8080 or 8088 code.

	The ST runs at basically 8 mhz; the Mac basically also at 8.
(plus or minus just a hair). Problem with the Mac is that video contention
chops a good 20% off the processor. As I understand it, when the SE was
designed, the PAL's that handle video contention were cleaned up; that's
why the SEis 20% faster -- it got to where it should be. ApplePeople feel free
to correct me; this is second hand. 

	I do know in benchmarks at a certain techie MacMagazine, we were 21%
ahead of the Plus in CPU, 4% ahead of the SE, and way behind the Mac II;
however, the II is 68020/30 and 16 mhz, so that's unfair. Amazingly, the
ST's hard disk kept even with the Mac II; the ST's Megafile 30 is no slouch
at all in raw data transfer rate. The ST also uses DMA, the Mac II a
loop-store scheme, so it's an unfair comparison.

	The thing that screws up most benchmarks is we run on a 70 hz
vertical blank, corresponding to the Atari mono monitor. The benchmarks on
the Mac run at 60 hz. Since most timer-tick programs dervice time-elapsed
from this vertical blank, they're off on the Spectre. We are planning a 60 hz
VBL "option", if you're willing to put up with desynced screen/animation,
which will bring sound back to normal pitch and make benchmarks accurate
once again, but that's in the future.

	Any timings right now should be on a stopwatch, and again,hard disk
access just isn't fair; looping can't keep up with a DMA scheme that plugs in
16 words at a whock. I could make a case that DMA could outrun a RAMDISK.

	-- thanks, Dave / Gadgets

p.s. The T-16 accelerator board takes the Spectre up to a "true" 12 mhz.
This makes an amazing difference! Even with the moniterms, with their huge
display memory, screens snap open and shut very fast. Recommended highly!

	.