Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!mailrus!usenet.ins.cwru.edu!mephisto!ncsuvx!news
From: rnf@shumv1.uucp (Rick Fincher)
Newsgroups: comp.sys.apple
Subject: Re: Speedy WDC Chips (was Re: Apple II
Message-ID: <1989Dec9.071924.16807@ncsuvx.ncsu.edu>
Date: 9 Dec 89 07:19:24 GMT
References: <7803.infoapple.net@pro-generic> <675@batman.moravian.EDU>
Reply-To: rnf@shumv1.ncsu.edu (Rick Fincher)
Organization: NCSU Computing Center
Lines: 43

In article <675@batman.moravian.EDU> nicholaA@batman.moravian.EDU (Andy Nicholas) writes:
>In article <7803.infoapple.net@pro-generic>, ericmcg@pro-generic.cts.COM (Eric Mcgillicuddy) writes:
>> In-Reply-To: message from gem.mps.ohio-state.edu!brutus.cs.uiuc.edu!jarthur!polyslo!vlsi3b15!batman!nicholaA@tut.cis.ohio-state.edu
>
>> Can you explain why 50ns DRAM is needed? the Mac+uses 100ns without wait
>> states, why would a GS require twice as fast  with only 50% speed increase.

The 68000 processor used in the Mac Plus and SE has a 16 bit data bus, the 
68020 and 68030, used in the Mac II series and the SE 30, have 32 bit data
busses.  This allows them to read 2 and 4 bytes respectivly per memory cycle.
Their memory cycles require from 3 to 6 clock cycles for simple loads and
stores, more for fancier addressing modes. The 65816 can load and store 2 bytes
in as lttle as 2 cycles.  To get the number of nanoseconds per clock cycle
divide 1000 ns by cycles in mhz (million cycles per sec).  A 10 mhz processor
then would have a cycle time of 100 ns.  If you consider that one half 
of the cycle is used to update the video circuitry, the RAM only has
half that time to respond, asuming it loads one byte per cycle.  So it has 
to work twice as fast, ie 50ns. The TransWarp GS doesn't have video circuitry
but it does have to trasnsfer its RAM contents back to the GS' main memory.
In the gs running at 2.8 mhz you would think about 357 ns are needed.  But 
you have to divide that by 2, thus giving your RAM about 178ns to deliver 
its contents.  The idea with a cache is that you take a small amount of
very fast (and expensive) memory and store the most recently used memory in
it.  Since programs tend to operate in small loops, they can run out of the
cache at full speed a significant portion of the time.

The video overhead causes problems with faster processors.  A 16mhz processor
would need 60ns RAM to operate at no wait states.  If you had to share that
with video, 30ns would be needed.  So dual port RAM is used for the video
on machines like the 16mhz Mac II.  This RAM can be read by the CPU and video
system simultaneously.  The RAM in some of these machines can't work fast
enough to keep up with the processor, so a wait state has to be thrown in to 
get good data.  You might wonder why they bother to run the processor at 
16mhz if the memory can't keep up.  This is done because a lot of operations
are register to register and don't require memory access, these go at full
speed.
 
Some of the memory cards on the market for the IIgs require 120ns RAM (the 
AE cards for example) because they do special things to make sure DMA works 
for the whole memory range, even for more than 4 rows of chips.

Rick Fincher
rnf@shumv1.ncsu.edu