Path: utzoo!attcan!uunet!tut.cis.ohio-state.edu!VAX1.CC.UAKRON.EDU!mcs.kent.edu!usenet.ins.cwru.edu!gatech!bloom-beacon!mintaka!dcw
From: dcw@lcs.mit.edu (David C. Whitney)
Newsgroups: comp.sys.apple2
Subject: Re: ASIC-65816 News
Message-ID: <1991Feb1.153644.7370@mintaka.lcs.mit.edu>
Date: 1 Feb 91 15:36:44 GMT
References: <1991Jan30.202122.22109@ux1.cso.uiuc.edu> <9242@uwm.edu> <1991Jan31.065813.25807@nntp-server.caltech.edu> <9261@uwm.edu> <43536@ut-emx.uucp>
Sender: daemon@mintaka.lcs.mit.edu (Lucifer Maleficius)
Organization: MIT Spoken Language Systems Group
Lines: 44

In article <43536@ut-emx.uucp> daveh@ccwf.cc.utexas.edu (David H. Huang) writes:
>In article <9261@uwm.edu> mcgu5464@csd4.csd.uwm.edu (Ronald J Mcguire) writes:
>>In article <1991Jan31.065813.25807@nntp-server.caltech.edu> toddpw@nntp-server.caltech.edu (Todd P. Whitesel) writes:
>>>Tony Fadell's project is making the CPU itself run faster. He doesn't make
>>>accelerators.

>>Well, Todd, I trust you know what you are talking about. But I don't 
>>understand the whole workings of an accellerator (as long as it works! :-)
>
>Just plugging a faster CPU into your computer won't make it run faster.
>It requires a lot of support circuitry, cache RAM and other stuff. The
>"ASIC" is just a CPU.

This is the same deal with RAM. The clock rates printed on the chip
are extremes of functionality. It's the clock crystal in the computer
which determines how fast everything goes. If you have 70ns RAM, it'll
work in a device that requires 150ns. The 150ns restriction means that
you can't have RAM slower than that.

A 25mhz CPU means the chip can't be clocked faster than that. Since
the clock in the GS goes at 2.8Mhz, that's as fast as it runs. Period.

Accelerators have their own on-board clock crystal and faster rated
CPU. They also have cache RAM which can be accessed very fast (much
less than even 70ns - that is why you *can't* use regular RAM for
cache memory - it's too slow). Whenever the accelerator has to talk to
the rest of the computer, it slows down and does its thing. If your
cache is small, then the accelerator doesn't do you much good.

This is why accelerators cost as much as they do - it's not as simple
as plug and play.

So, the ASIC 65816 on an accelerator could be clocked as fast as
25MHz, but then you need to get cache ram that can be accessed at that
speed. The original Macs were clocked at 8Mhz, but since RAM came only
as fast as 150ns back then (I mean affordable RAM), the processor
spent a lot of time waiting around for the RAM. These are called wait
states and they effectively slow down the processor during memory
accesses (nearly all the time). It's all a big mess.
--
Dave Whitney
Computer Science MIT 1990	| I wrote Z-Link and BinSCII. Send me bug
dcw@lcs.mit.edu   dcw@mit.edu	| reports. I have a job. Don't send me offers.
"The price of freedom is eternal vigilance" --Binky (aka Matt Groening)